Validation matrix

Stephens 1974 weights with later Anderson-Darling calibration

Case STAT-DESC-AD-STEPHENS-001 · status: passed

Reference

Evidence type
R
Source
Stephens (1974) printed inputs; R nortest 1.0-4 statistic and later p calibration
Citation
Stephens (1974), Illustration 2/Table 1.3; nortest 1.0-4 ad.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · nortest 1.0.4, jsonlite 2.0.0
Reference script
generators/r/reference_stephens_ad.R · sha256 dc3c92cceb7d2180
Dataset
reference_data/datasets/DESC_STEPHENS_1974_WEIGHTS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance0.94677187959888620.94677187959888621.28E-1817passed
/p_valuetail_probability0.0104540240051477760.0104540240051477764.36E-1916passed
/observationsexact11110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

3  id <- "STAT-DESC-AD-STEPHENS-001"

Whole script generators/r/reference_stephens_ad.R


3 x 3 crosstab: chi-square family, nominal, ordinal, linear-by-linear and eta

Case STAT-DESC-ASSOCIATION-001 · status: passed

Reference

Evidence type
R
Source
R stats::chisq.test, cor(method = kendall) and aov; DescTools association measures; SPSS CROSSTABS ASE formulas written out
Citation
Goodman & Kruskal (1954) JASA 49:732-764; Agresti (2013) Categorical Data Analysis ch. 2-3; IBM SPSS Statistics Algorithms, CROSSTABS; DescTools 0.99 documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, DescTools 0.99.60, jsonlite 2.0.0
Reference script
generators/r/descriptives_association.R · sha256 16db03dab046251a
Dataset
reference_data/datasets/DESC_ASSOC_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_analysedexact33330.00E+2passed
/missing_countexact000.00E+2passed
/row_categories/0exactLowLowpassed
/row_order_sourceexactdeclareddeclaredpassed
/column_order_sourceexactnumericnumericpassed
/counts/0/0exact770.00E+2passed
/counts/2/2exact880.00E+2passed
/chi_squaretolerance14.98022727272727114.9802272727272731.95E-1515passed
/chi_square_dfexact440.00E+2passed
/chi_square_ptail_probability0.0047424032765409690.0047424032765409688.54E-1915passed
/cells_expected_below_5exact990.00E+2passed
/cells/0/expectedtolerance3.33333333333333353.33333333333333351.86E-1717passed
/cells/0/standardized_residualtolerance2.00831604418560872.00831604418560871.75E-1717passed
/cells/0/adjusted_residualtolerance2.94625775902392742.94625775902392748.31E-1817passed
/cells/8/adjusted_residualtolerance3.07059789431495393.07059789431495344.53E-1615passed
/risks/0absentNonepassed
/measures/0/nameexactPhiPhipassed
/measures/0/directionexactsymmetricsymmetricpassed
/measures/0/estimatetolerance0.673755356929898960.6737553569298993.08E-1817passed
/measures/1/nameexactCramer's VCramer's Vpassed
/measures/1/directionexactsymmetricsymmetricpassed
/measures/1/estimatetolerance0.476416981745894250.47641698174589435.79E-1715passed
/measures/2/nameexactContingency coefficientContingency coefficientpassed
/measures/2/directionexactsymmetricsymmetricpassed
/measures/2/estimatetolerance0.558763527949476920.55876352794947693.60E-1817passed
/measures/3/nameexactLambdaLambdapassed
/measures/3/directionexactsymmetricsymmetricpassed
/measures/3/estimatetolerance0.428571428571428550.428571428571428552.36E-1817passed
/measures/4/nameexactLambdaLambdapassed
/measures/4/directionexactrow dependentrow dependentpassed
/measures/4/estimatetolerance0.428571428571428550.428571428571428552.36E-1817passed
/measures/5/nameexactLambdaLambdapassed
/measures/5/directionexactcolumn dependentcolumn dependentpassed
/measures/5/estimatetolerance0.428571428571428550.428571428571428552.36E-1817passed
/measures/6/nameexactGoodman and Kruskal tauGoodman and Kruskal taupassed
/measures/6/directionexactrow dependentrow dependentpassed
/measures/6/estimatetolerance0.221961325966850840.221961325966850862.42E-1715passed
/measures/7/nameexactGoodman and Kruskal tauGoodman and Kruskal taupassed
/measures/7/directionexactcolumn dependentcolumn dependentpassed
/measures/7/estimatetolerance0.233011049723756890.2330110497237571.07E-1615passed
/measures/8/nameexactUncertainty coefficientUncertainty coefficientpassed
/measures/8/directionexactsymmetricsymmetricpassed
/measures/8/estimatetolerance0.202231014500417780.20223101450041761.69E-1615passed
/measures/9/nameexactUncertainty coefficientUncertainty coefficientpassed
/measures/9/directionexactrow dependentrow dependentpassed
/measures/9/estimatetolerance0.202231014500417780.202231014500417225.58E-1614passed
/measures/10/nameexactUncertainty coefficientUncertainty coefficientpassed
/measures/10/directionexactcolumn dependentcolumn dependentpassed
/measures/10/estimatetolerance0.202231014500417780.202231014500417581.97E-1615passed
/measures/11/nameexactGammaGammapassed
/measures/11/directionexactsymmetricsymmetricpassed
/measures/11/estimatetolerance0.738805970149253750.73880597014925361.06E-1615passed
/measures/11/ase1tolerance0.136446148399816360.136446148399816342.40E-1715passed
/measures/11/ase0tolerance0.170747181920112710.17074718192011273.88E-1816passed
/measures/11/lowertolerance0.471376433456406230.47137643345640611.11E-1615passed
/measures/11/uppertolerance1.00623550684210141.00623550684210122.32E-1615passed
/measures/12/nameexactSomers' dSomers' dpassed
/measures/12/directionexactcolumn dependentcolumn dependentpassed
/measures/12/estimatetolerance0.546961325966850880.54696132596685094.65E-1817passed
/measures/12/ase1tolerance0.125214648987206110.125214648987206165.19E-1715passed
/measures/12/ase0tolerance0.126409515896658030.126409515896658052.34E-1715passed
/measures/13/nameexactSomers' dSomers' dpassed
/measures/13/directionexactrow dependentrow dependentpassed
/measures/13/estimatetolerance0.546961325966850880.54696132596685094.65E-1817passed
/measures/13/ase1tolerance0.127330139284789120.127330139284789038.55E-1715passed
/measures/13/ase0tolerance0.126409515896658030.126409515896658052.34E-1715passed
/measures/14/nameexactSomers' dSomers' dpassed
/measures/14/directionexactsymmetricsymmetricpassed
/measures/14/estimatetolerance0.546961325966850880.54696132596685094.65E-1817passed
/measures/14/ase0tolerance0.126409515896658030.126409515896658052.34E-1715passed
/measures/15/nameexactKendall's tau-bKendall's tau-bpassed
/measures/15/directionexactsymmetricsymmetricpassed
/measures/15/estimatetolerance0.546961325966850880.54696132596685094.65E-1817passed
/measures/15/ase0tolerance0.126409515896658030.126409515896658052.34E-1715passed
/measures/15/ase1tolerance0.126168304120550720.126168304120550742.50E-1715passed
/measures/15/lowertolerance0.299675993900075040.299675993900074931.07E-1615passed
/measures/15/uppertolerance0.794246658033626710.79424665803362681.08E-1615passed
/measures/16/nameexactKendall's tau-cKendall's tau-cpassed
/measures/16/directionexactsymmetricsymmetricpassed
/measures/16/estimatetolerance0.545454545454545410.54545454545454544.17E-1817passed
/measures/16/ase1tolerance0.126061280315675480.126061280315675545.90E-1715passed
/measures/16/ase0tolerance0.126061280315675480.12606128031567553.13E-1715passed
/measures/17/nameexactLinear-by-linear associationLinear-by-linear associationpassed
/measures/17/directionexactsymmetricsymmetricpassed
/measures/17/estimatetolerance11.17537785776381111.1753778577638175.54E-1515passed
/measures/18/nameexactEtaEtapassed
/measures/18/directionexactoutcome dependentoutcome dependentpassed
/measures/18/estimatetolerance0.591050225778773970.59105022577877422.26E-1615passed
/measures/17/p_valuetail_probability0.00082890012581967340.00082890012581967231.08E-1814passed
/measures/11/p_valuetail_probability1.5122261925217157e-051.5122261925217155e-051.59E-2115passed
/measures/15/p_valuetail_probability1.5122261925217098e-051.5122261925217155e-055.74E-2014passed
/measures/19absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

260  m1 <- merge_outputs(m1,
261    measure(0, "Phi", "symmetric", "estimate", cf$phi, "phi = sqrt(chi2/n)"),
262    measure(1, "Cramer's V", "symmetric", "estimate", cf$v, "Cramer's V"),
263    measure(2, "Contingency coefficient", "symmetric", "estimate", cf$c, "contingency coefficient"),
264    measure(3, "Lambda", "symmetric", "estimate", lambda_sym(tab), "symmetric lambda (Goodman-Kruskal 1954)"),
265    measure(4, "Lambda", "row dependent", "estimate", lambda_col(t(tab)), "lambda predicting the row from the column"),
266    measure(5, "Lambda", "column dependent", "estimate", lambda_col(tab), "lambda predicting the column from the row"),
267    measure(6, "Goodman and Kruskal tau", "row dependent", "estimate", gk_tau_col(t(tab)), "GK tau, row predicted"),
268    measure(7, "Goodman and Kruskal tau", "column dependent", "estimate", gk_tau_col(tab), "GK tau, column predicted"),
269    measure(8, "Uncertainty coefficient", "symmetric", "estimate", uncert(tab, "symmetric"), "symmetric U = 2I/(H_R + H_C)"),
270    measure(9, "Uncertainty coefficient", "row dependent", "estimate", uncert(tab, "row"), "U(R|C) = I/H_R"),
271    measure(10, "Uncertainty coefficient", "column dependent", "estimate", uncert(tab, "column"), "U(C|R) = I/H_C"),
272    measure(11, "Gamma", "symmetric", "estimate", ord$gamma, "gamma = (P-Q)/(P+Q)"),
273    measure(11, "Gamma", "symmetric", "ase1", ord$gamma_ase1, "gamma ASE1 (SPSS CROSSTABS; DescTools interval agrees)"),
274    measure(11, "Gamma", "symmetric", "ase0", ord$gamma_ase0, "gamma ASE0 = 2/(P+Q) sqrt(sum n(A-D)^2 - (P-Q)^2/n)"),
275    measure(11, "Gamma", "symmetric", "lower", ord$gamma - z95 * ord$gamma_ase1, "Wald lower bound estimate - z ASE1 (DescTools agrees)"),
276    measure(11, "Gamma", "symmetric", "upper", ord$gamma + z95 * ord$gamma_ase1, "Wald upper bound estimate + z ASE1, not truncated at 1 (DescTools clips it to 1)"),
277    measure(12, "Somers' d", "column dependent", "estimate", ord$d_col, "Somers' d(C|R) = (P-Q)/(n^2 - sum n_i+^2)"),
278    measure(12, "Somers' d", "column dependent", "ase1", ord$d_col_ase1, "Somers' d ASE1 (SPSS CROSSTABS; DescTools interval agrees)"),
279    measure(12, "Somers' d", "column dependent", "ase0", ord$d_col_ase0, "Somers' d ASE0"),
280    measure(13, "Somers' d", "row dependent", "estimate", ord$d_row, "Somers' d(R|C)"),
281    measure(13, "Somers' d", "row dependent", "ase1", ord$d_row_ase1, "Somers' d(R|C) ASE1 (SPSS formula on the transposed table; DescTools agrees)"),
282    measure(13, "Somers' d", "row dependent", "ase0", ord$d_row_ase0, "Somers' d(R|C) ASE0"),
283    measure(14, "Somers' d", "symmetric", "estimate", ord$d_sym, "symmetric Somers' d = 2(P-Q)/(D_r + D_c)"),
284    measure(14, "Somers' d", "symmetric", "ase0", ord$d_sym_ase0, "symmetric Somers' d ASE0"),
285    measure(15, "Kendall's tau-b", "symmetric", "estimate", ord$tau_b, "tau-b (equals cor(method = kendall) on integer scores)"),
286    measure(15, "Kendall's tau-b", "symmetric", "ase0", ord$tau_b_ase0, "tau-b ASE0 = 2 sqrt(S)/sqrt(D_r D_c)"),
287    measure(15, "Kendall's tau-b", "symmetric", "ase1", ord$tau_b_ase1, "tau-b ASE1 by the SAS/SPSS formula (1/w^2) sqrt(sum n (2 w d + tau v)^2 - n^3 tau^2 (w_r + w_c)^2)"),
288    measure(15, "Kendall's tau-b", "symmetric", "lower", ord$tau_b - z95 * ord$tau_b_ase1, "Wald lower bound with the SAS/SPSS ASE1"),
289    measure(15, "Kendall's tau-b", "symmetric", "upper", ord$tau_b + z95 * ord$tau_b_ase1, "Wald upper bound with the SAS/SPSS ASE1"),
290    measure(16, "Kendall's tau-c", "symmetric", "estimate", ord$tau_c, "tau-c = m(P-Q)/(n^2 (m-1))"),
291    measure(16, "Kendall's tau-c", "symmetric", "ase1", ord$tau_c_ase, "tau-c ASE (SAS; DescTools interval agrees)"),
292    measure(16, "Kendall's tau-c", "symmetric", "ase0", ord$tau_c_ase, "tau-c ASE0 equals its ASE1"),
293    measure(17, "Linear-by-linear association", "symmetric", "estimate", lbl$m2, "(n-1) r^2 with integer scores in declared order"),
294    measure(18, "Eta", "outcome dependent", "estimate", eta_outcome, "eta with the numeric outcome as the scale (aov between/total SS)")
295  )
296  m1[["/measures/17/p_value"]] <- prob(lbl$p, "linear-by-linear chi-square tail, 1 df")
297  m1[["/measures/11/p_value"]] <- prob(2 * stats::pnorm(-abs(ord$gamma / ord$gamma_ase0)), "gamma T = estimate/ASE0, normal two-sided")
298  m1[["/measures/15/p_value"]] <- prob(2 * stats::pnorm(-abs(ord$tau_b / ord$tau_b_ase0)), "tau-b T = estimate/ASE0, normal two-sided")
299  m1[["/measures/19"]] <- absent("the row variable is text, so no second eta")
300  emit("STAT-DESC-ASSOCIATION-001", "3 x 3 crosstab: chi-square family, nominal, ordinal, linear-by-linear and eta",
301       "nominal",
302       list(row = "exposure", column = "outcome", confidence = 0.95, haldane_correction = FALSE,
303            row_order = list("Low", "Medium", "High")),
304       m1,
305       c("R: chisq.test(correct = FALSE) for chi-square, expected counts and adjusted residuals; lambda, GK tau and U from the Goodman-Kruskal/Theil definitions; gamma, Somers' d, tau-b, tau-c with the SAS/SPSS CROSSTABS ASE1/ASE0 formulas written out (DescTools reproduces the gamma, Somers' d and tau-c intervals; its tau-b variance code differs from the published formula and is not used); tau-b equals cor(method = 'kendall') on integer scores.",
306         "Intervals are estimate +/- z ASE1 without truncation to [-1, 1]; the gamma upper bound exceeds 1 on this table, where DescTools clips to 1.",
307         "Measure indexes follow the window's list: chi-square family, lambda (symmetric, row, column), GK tau (row, column), U (symmetric, row, column), gamma, Somers' d (column, row, symmetric), tau-b, tau-c, linear-by-linear, eta.",
308         paste0(manual, ":65-82: declared text order for ordinal measures, integer scores, ASE1 for intervals and separate null ASE for tests, eta in the numeric direction.")))

Whole script generators/r/descriptives_association.R


2 x 2 crosstab at 90%: signed phi, odds ratio and relative risk intervals

Case STAT-DESC-ASSOCIATION-002 · status: passed

Reference

Evidence type
R
Source
R stats::chisq.test, cor(method = kendall) and aov; DescTools association measures; SPSS CROSSTABS ASE formulas written out
Citation
Goodman & Kruskal (1954) JASA 49:732-764; Agresti (2013) Categorical Data Analysis ch. 2-3; IBM SPSS Statistics Algorithms, CROSSTABS; DescTools 0.99 documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, DescTools 0.99.60, jsonlite 2.0.0
Reference script
generators/r/descriptives_association.R · sha256 16db03dab046251a
Dataset
reference_data/datasets/DESC_ASSOC_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_analysedexact33330.00E+2passed
/row_categories/0exactAApassed
/column_categories/0exact00passed
/row_order_sourceexactundeclaredundeclaredpassed
/counts/0/0exact880.00E+2passed
/chi_squaretolerance0.609230769230769460.60923076923076954.43E-1817passed
/chi_square_dfexact110.00E+2passed
/chi_square_ptail_probability0.435077586192428090.43507758619242821.11E-1615passed
/measures/0/estimatetolerance-0.13587324409735149-0.13587324409735152.11E-1816passed
/measures/1/estimatetolerance0.135873244097351520.135873244097351521.34E-1918passed
/risks/0/nameexactOdds ratioOdds ratiopassed
/risks/0/estimatetolerance0.57142857142857140.57142857142857143.15E-1817passed
/risks/0/lowertolerance0.174943860579697190.17494386057969718.75E-1715passed
/risks/0/uppertolerance1.86648797598784061.86648797598784148.41E-1615passed
/risks/1/nameexactRelative risk (row 1)Relative risk (row 1)passed
/risks/1/estimatetolerance0.742857142857142880.74285714285714292.52E-1817passed
/risks/1/lowertolerance0.400617133988785180.4006171339887851.68E-1615passed
/risks/1/uppertolerance1.3774666330405271.37746663304052744.11E-1615passed
/measures/11/estimateabsentNonepassed
/measures/17/estimateabsentNonepassed
/measures/18/directionexactpassed dependentpassed dependentpassed
/measures/18/estimatetolerance0.135873244097351460.135873244097351442.34E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

312  tab2 <- table(d$treated, d$passed)
313  stopifnot(all(tab2 == matrix(c(8, 12, 7, 6), 2, byrow = TRUE)))
314  cf2 <- chi_family(tab2)
315  rk <- risks(tab2, 0.90)
316  stopifnot(abs(rk$odds - DescTools::OddsRatio(tab2)) < 1e-12)
317  stopifnot(abs(rk$rr - DescTools::RelRisk(tab2)) < 1e-12)
318  or_dt <- DescTools::OddsRatio(tab2, conf.level = 0.90, method = "wald")
319  stopifnot(abs(or_dt[["lwr.ci"]] - rk$odds_lower) < 1e-9)
320  ord2 <- ordinal(tab2)
321  m2 <- list(
322    "/rows_analysed" = exact(as.integer(sum(tab2)), "33 rows"),
323    "/row_categories/0" = exact("A", "text rows in alphabetical display order"),
324    "/column_categories/0" = exact("0", "numeric columns in numeric order"),
325    "/row_order_source" = exact("undeclared", "no declared row order"),
326    "/counts/0/0" = exact(8L, "A x 0 is cell a"),
327    "/chi_square" = tol(cf2$chi2, "Pearson chi-square without continuity correction"),
328    "/chi_square_df" = exact(1L, "1 df"),
329    "/chi_square_p" = prob(cf2$p, "chi-square tail"),
330    "/measures/0/estimate" = tol(cf2$phi, "signed phi (ad-bc)/sqrt(margins)"),
331    "/measures/1/estimate" = tol(cf2$v, "Cramer's V equals |phi| for 2 x 2"),
332    "/risks/0/name" = exact("Odds ratio", "first risk"),
333    "/risks/0/estimate" = tol(rk$odds, "odds ratio ad/(bc) with a = A x 0"),
334    "/risks/0/lower" = tol(rk$odds_lower, "Woolf log-scale 90% lower (DescTools wald agrees)"),
335    "/risks/0/upper" = tol(rk$odds_upper, "Woolf log-scale 90% upper"),
336    "/risks/1/name" = exact("Relative risk (row 1)", "second risk"),
337    "/risks/1/estimate" = tol(rk$rr, "risk of column 0 in row A over row B"),
338    "/risks/1/lower" = tol(rk$rr_lower, "Katz log-scale 90% lower"),
339    "/risks/1/upper" = tol(rk$rr_upper, "Katz log-scale 90% upper"),
340    "/measures/11/estimate" = absent("gamma is unavailable without a declared order for the text rows"),
341    "/measures/17/estimate" = absent("linear-by-linear is unavailable without a declared order"),
342    "/measures/18/direction" = exact("passed dependent", "eta in the numeric direction only")
343  )
344  m2[["/measures/18/estimate"]] <- tol(sqrt(stats::anova(stats::lm(passed ~ treated, data = d))[["Sum Sq"]][1] /
345                                             sum(stats::anova(stats::lm(passed ~ treated, data = d))[["Sum Sq"]])),
346                                        "eta with passed as the scale")
347  emit("STAT-DESC-ASSOCIATION-002", "2 x 2 crosstab at 90%: signed phi, odds ratio and relative risk intervals",
348       "axis_coverage",
349       list(row = "treated", column = "passed", confidence = 0.90, haldane_correction = FALSE),
350       m2,
351       c("R: table rows A, B (alphabetical) and columns 0, 1; a = A x 0 = 8, b = 12, c = 7, d = 6; odds ratio ad/(bc) with the Woolf interval exp(log OR +/- z sqrt(1/a+1/b+1/c+1/d)); relative risk of column 0 for row A over row B with the Katz interval; z at 90%.",
352         "Without a declared order for the text rows the ordinal measures and the linear-by-linear test are unavailable; the manual says alphabetical order is a display rule.",
353         paste0(manual, ":65-69 and 85-96.")))

Whole script generators/r/descriptives_association.R


2 x 2 crosstab with an empty cell and the 0.5 correction

Case STAT-DESC-ASSOCIATION-003 · status: passed

Reference

Evidence type
R
Source
R stats::chisq.test, cor(method = kendall) and aov; DescTools association measures; SPSS CROSSTABS ASE formulas written out
Citation
Goodman & Kruskal (1954) JASA 49:732-764; Agresti (2013) Categorical Data Analysis ch. 2-3; IBM SPSS Statistics Algorithms, CROSSTABS; DescTools 0.99 documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, DescTools 0.99.60, jsonlite 2.0.0
Reference script
generators/r/descriptives_association.R · sha256 16db03dab046251a
Dataset
reference_data/datasets/DESC_ASSOC_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/1/1exact000.00E+2passed
/risks/0/estimatetolerance0.0826210826210826140.082621082621082613.61E-1917passed
/risks/0/lowertolerance0.0042369572653064040.00423695726530640053.49E-1815passed
/risks/0/uppertolerance1.6111192221302971.6111192221302989.01E-1615passed
/risks/0/infiniteexactFalseFalsepassed
/risks/1/estimatetolerance0.716049382716049340.71604938271604932.97E-1817passed
/risks/1/lowertolerance0.528563846878680520.52856384687868051.08E-1817passed
/risks/1/uppertolerance0.970037435431560510.97003743543156052.99E-1817passed
/chi_squaretolerance4.76666666666666664.76666666666666758.96E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

357  tab3 <- table(d$treated, d$severe)
358  stopifnot(all(tab3 == matrix(c(14, 6, 13, 0), 2, byrow = TRUE)))
359  rk3 <- risks(tab3, 0.95, haldane = TRUE)
360  emit("STAT-DESC-ASSOCIATION-003", "2 x 2 crosstab with an empty cell and the 0.5 correction",
361       "axis_coverage",
362       list(row = "treated", column = "severe", confidence = 0.95, haldane_correction = TRUE),
363       list(
364         "/counts/1/1" = exact(0L, "B x 1 is empty"),
365         "/risks/0/estimate" = tol(rk3$odds, "odds ratio after adding 0.5 to all four cells"),
366         "/risks/0/lower" = tol(rk3$odds_lower, "Woolf interval on the corrected cells"),
367         "/risks/0/upper" = tol(rk3$odds_upper, "Woolf interval on the corrected cells"),
368         "/risks/0/infinite" = exact(FALSE, "finite after correction"),
369         "/risks/1/estimate" = tol(rk3$rr, "relative risk on the corrected cells"),
370         "/risks/1/lower" = tol(rk3$rr_lower, "Katz interval on the corrected cells"),
371         "/risks/1/upper" = tol(rk3$rr_upper, "Katz interval on the corrected cells"),
372         "/chi_square" = tol(chi_family(tab3)$chi2, "chi-square is on the uncorrected counts")
373       ),
374       c("R: cells 14.5, 6.5, 13.5, 0.5; odds ratio ad/(bc) and relative risk (a/(a+b))/(c/(c+d)) with Woolf and Katz log-scale intervals at 95%; the chi-square ignores the correction.",
375         paste0(manual, ":96-98: the 0.5 correction adds 0.5 to all cells only when explicitly selected.")))

Whole script generators/r/descriptives_association.R


2 x 2 crosstab with an empty cell and no correction

Case STAT-DESC-ASSOCIATION-004 · status: passed

Reference

Evidence type
R
Source
R stats::chisq.test, cor(method = kendall) and aov; DescTools association measures; SPSS CROSSTABS ASE formulas written out
Citation
Goodman & Kruskal (1954) JASA 49:732-764; Agresti (2013) Categorical Data Analysis ch. 2-3; IBM SPSS Statistics Algorithms, CROSSTABS; DescTools 0.99 documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, DescTools 0.99.60, jsonlite 2.0.0
Reference script
generators/r/descriptives_association.R · sha256 16db03dab046251a
Dataset
reference_data/datasets/DESC_ASSOC_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/risks/0/estimatetolerance00.00.00E+2passed
/risks/0/lowerabsentNonepassed
/risks/1/estimatetolerance0.699999999999999960.74.41E-1817passed
/risks/1/lowertolerance0.525405836421816490.52540583642181654.40E-1817passed
/measures/0/estimatetolerance-0.38005847503304596-0.380058475033045961.58E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

377  emit("STAT-DESC-ASSOCIATION-004", "2 x 2 crosstab with an empty cell and no correction",
378       "boundary",
379       list(row = "treated", column = "severe", confidence = 0.95, haldane_correction = FALSE),
380       list(
381         "/risks/0/estimate" = tol(0, "odds ratio ad/(bc) = 0 when d = 0"),
382         "/risks/0/lower" = absent("log-scale interval unavailable with a zero cell"),
383         "/risks/1/estimate" = tol((14 / 20) / (13 / 13), "relative risk of column 0: 0.7"),
384         "/risks/1/lower" = tol(exp(log(0.7) - stats::qnorm(0.975) * sqrt(1 / 14 - 1 / 20 + 1 / 13 - 1 / 13)), "Katz lower bound, defined because both row risks are positive"),
385         "/measures/0/estimate" = tol(chi_family(tab3)$phi, "signed phi with a zero cell")
386       ),
387       c("R: a = 14, b = 6, c = 13, d = 0: odds ratio 0 with no log-scale interval; relative risk 0.7 with the Katz interval.",
388         paste0(manual, ":96-98: zero cells may make log-scale intervals unavailable.")))

Whole script generators/r/descriptives_association.R


Case summaries by group: five listed, statistics on every case

Case STAT-DESC-CASESUMMARIES-001 · status: passed

Reference

Evidence type
R
Source
R base mean and sd by group
Citation
R stats::sd documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/listedexact550.00E+2passed
/totalexact24240.00E+2passed
/rows/0/0exact11passed
/rows/0/1exactAApassed
/rows/0/2exact12.512.5passed
/rows/2/2exactpassed
/rows/5absentNonepassed
/statistics/0/groupexactAApassed
/statistics/0/nexact10100.00E+2passed
/statistics/0/meantolerance17.00999999999999817.0099999999999981.05E-1718passed
/statistics/0/sdtolerance2.32687200622065452.32687200622065451.26E-1717passed
/statistics/1/groupexactBBpassed
/statistics/1/nexact10100.00E+2passed
/statistics/1/meantolerance23.8223.822.84E-1616passed
/statistics/1/sdtolerance4.892806511150384.892806511150381.11E-1717passed
/statistics/2/groupexactTotalTotalpassed
/statistics/2/nexact20200.00E+2passed
/statistics/2/meantolerance20.41499999999999920.4151.47E-1617passed
/statistics/2/sdtolerance5.10966833819141285.1096683381914134.91E-1717passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

310  by_group <- function(g) d$score[valid_score & d$group == g]
311  cs_outputs <- list(
312    "/listed" = exact(5L, "limit of 5"),
313    "/total" = exact(nrow(d), "all 24 rows are summarised"),
314    "/rows/0/0" = exact("1", "first listed case number"),
315    "/rows/0/1" = exact("A", "its group"),
316    "/rows/0/2" = exact("12.5", "its score as shown"),
317    "/rows/2/2" = exact("", "the third row's score is system-missing"),
318    "/rows/5" = absent("only five rows are listed"),
319    "/statistics/0/group" = exact("A", "groups in sorted order"),
320    "/statistics/0/n" = exact(length(by_group("A")), "valid scores in A"),
321    "/statistics/0/mean" = tol(mean(by_group("A")), "mean of A"),
322    "/statistics/0/sd" = tol(sd(by_group("A")), "SD of A"),
323    "/statistics/1/group" = exact("B", "second group"),
324    "/statistics/1/n" = exact(length(by_group("B")), "valid scores in B"),
325    "/statistics/1/mean" = tol(mean(by_group("B")), "mean of B"),
326    "/statistics/1/sd" = tol(sd(by_group("B")), "SD of B"),
327    "/statistics/2/group" = exact("Total", "total row"),
328    "/statistics/2/n" = exact(n, "all valid scores"),
329    "/statistics/2/mean" = tol(mean(x), "overall mean"),
330    "/statistics/2/sd" = tol(sd(x), "overall SD")
331  )
332  emit("STAT-DESC-CASESUMMARIES-001", "Case summaries by group: five listed, statistics on every case",
333       "nominal", "tabelz.analysis.report_listings:compute_case_summaries", "caseSummariesMenuItem",
334       "CASESUMMARIES",
335       list(columns = list("score"), group = "group", limit = 5L, statistics = TRUE,
336            metadata = metadata, frequency_weight = NULL),
337       cs_outputs,
338       c("R: mean and sd of the valid scores (NA and 999 excluded) per group and overall.",
339         paste0(manual, ":16-18: summaries use all eligible cases, not just the displayed rows.")),
340       "R base mean and sd by group", "R stats::sd documentation")

Whole script generators/r/descriptives_reports.R


Case summaries under an integer frequency weight

Case STAT-DESC-CASESUMMARIES-002 · status: passed

Reference

Evidence type
R
Source
R base rep/mean/sd (literal replication)
Citation
R stats::sd documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/listedexact24240.00E+2passed
/columns/1exactWeightWeightpassed
/rows/0/1exact11passed
/rows/20/1exactpassed
/statistics/0/nexact18180.00E+2passed
/statistics/0/weight_sumtolerance2828.00.00E+2passed
/statistics/0/meantolerance19.92142857142857119.9214285714285743.37E-1515passed
/statistics/0/sdtolerance4.88931996656485354.88931996656485354.09E-1717passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

342  w <- d$weight
343  w[is.na(w)] <- 0L
344  chosen <- valid_score & w > 0
345  W <- sum(w[chosen])
346  wm <- sum(w[chosen] * d$score[chosen]) / W
347  wsd <- sqrt(sum(w[chosen] * (d$score[chosen] - wm)^2) / (W - 1))
348  stopifnot(abs(wm - mean(rep(d$score[chosen], w[chosen]))) < 1e-12)
349  stopifnot(abs(wsd - sd(rep(d$score[chosen], w[chosen]))) < 1e-12)
350  cw_outputs <- list(
351    "/listed" = exact(nrow(d), "all rows fit within the limit"),
352    "/columns/1" = exact("Weight", "the weight column is shown"),
353    "/rows/0/1" = exact("1", "first row's weight"),
354    "/rows/20/1" = exact("", "the missing weight is shown blank"),
355    "/statistics/0/n" = exact(sum(chosen), "rows with a valid score and a positive weight"),
356    "/statistics/0/weight_sum" = tol(W, "W, the sum of the positive weights of valid scores"),
357    "/statistics/0/mean" = tol(wm, "weighted mean (mean of the replicated values)"),
358    "/statistics/0/sd" = tol(wsd, "SD with denominator W - 1 (sd of the replicated values)")
359  )
360  emit("STAT-DESC-CASESUMMARIES-002", "Case summaries under an integer frequency weight",
361       "axis_coverage", "tabelz.analysis.report_listings:compute_case_summaries", "caseSummariesMenuItem",
362       "CASESUMMARIES",
363       list(columns = list("score"), limit = 100L, statistics = TRUE,
364            metadata = metadata, frequency_weight = "weight"),
365       cw_outputs,
366       c("R: mean and sd of rep(score, weight) over valid scores with positive weight; a missing weight is 0.",
367         paste0(manual, ":19-27: weighted mean over positive weights, squared deviations divided by W - 1, zero-weight rows listed but not summarised.")),
368       "R base rep/mean/sd (literal replication)", "R stats::sd documentation")

Whole script generators/r/descriptives_reports.R


Codebook: scale summary, labelled categories and missing counts

Case STAT-DESC-CODEBOOK-001 · status: passed

Reference

Evidence type
R
Source
R base table and stats; quantile type 6
Citation
R stats::quantile documentation (type 6)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rowsexact24240.00E+2passed
/entries/0/columnexactscorescorepassed
/entries/0/levelexactscalescalepassed
/entries/0/validtolerance2020.00.00E+2passed
/entries/0/user_missing_counttolerance22.00.00E+2passed
/entries/0/system_missing_counttolerance22.00.00E+2passed
/entries/0/summary/0/1tolerance2020.00.00E+2passed
/entries/0/summary/1/1tolerance20.41499999999999920.4151.47E-1617passed
/entries/0/summary/2/1tolerance5.10966833819141285.1096683381914134.91E-1717passed
/entries/0/summary/3/1tolerance12.512.50.00E+1passed
/entries/0/summary/4/1tolerance33.20000000000000333.21.58E-1617passed
/entries/0/summary/5/1tolerance16.59999999999999816.5999999999999981.32E-1617passed
/entries/0/summary/6/1tolerance19.519.50.00E+1passed
/entries/0/summary/7/1tolerance23.24999999999999623.254.00E-1515passed
/entries/1/levelexactnominalnominalpassed
/entries/1/labelexactGradeGradepassed
/entries/1/validtolerance2121.00.00E+2passed
/entries/1/user_missing_counttolerance22.00.00E+2passed
/entries/1/system_missing_counttolerance11.00.00E+2passed
/entries/1/values/[value=2]/counttolerance77.00.00E+2passed
/entries/1/values/[value=2]/labelexactFairFairpassed
/entries/1/values/[value=5]/counttolerance00.00.00E+2passed
/entries/1/values/[value=9]/missingexactuser-missinguser-missingpassed
/entries/1/values/[value=9]/counttolerance22.00.00E+2passed
/entries/2/levelexactnominalnominalpassed
/entries/2/values/[value=A]/counttolerance1212.00.00E+2passed
/entries/2/values/[value=B]/valid_percenttolerance5050.00.00E+2passed
/entries/2/validtolerance2424.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

268  cb_outputs <- list(
269    "/rows" = exact(nrow(d), "all 24 rows"),
270    "/entries/0/column" = exact("score", "first entry"),
271    "/entries/0/level" = exact("scale", "numeric without a declared level is scale"),
272    "/entries/0/valid" = tol(n, "valid scores"),
273    "/entries/0/user_missing_count" = tol(sum(d$score == 999, na.rm = TRUE), "two 999 scores"),
274    "/entries/0/system_missing_count" = tol(sum(is.na(d$score)), "two NA scores"),
275    "/entries/0/summary/0/1" = tol(n, "Valid N"),
276    "/entries/0/summary/1/1" = tol(mean(x), "mean"),
277    "/entries/0/summary/2/1" = tol(sd(x), "sample SD"),
278    "/entries/0/summary/3/1" = tol(min(x), "minimum"),
279    "/entries/0/summary/4/1" = tol(max(x), "maximum"),
280    "/entries/0/summary/5/1" = tol(q6(x, 0.25), "25th percentile type 6"),
281    "/entries/0/summary/6/1" = tol(q6(x, 0.5), "50th percentile type 6"),
282    "/entries/0/summary/7/1" = tol(q6(x, 0.75), "75th percentile type 6"),
283    "/entries/1/level" = exact("nominal", "declared nominal"),
284    "/entries/1/label" = exact("Grade", "variable label"),
285    "/entries/1/valid" = tol(length(grade_valid), "valid grades"),
286    "/entries/1/user_missing_count" = tol(sum(d$grade == 9, na.rm = TRUE), "code 9 count"),
287    "/entries/1/system_missing_count" = tol(sum(is.na(d$grade)), "NA grade count"),
288    "/entries/1/values/[value=2]/count" = tol(as.numeric(counts[["2"]]), "grade 2 count"),
289    "/entries/1/values/[value=2]/label" = exact("Fair", "value label"),
290    "/entries/1/values/[value=5]/count" = tol(0, "labelled code 5 has no cases"),
291    "/entries/1/values/[value=9]/missing" = exact("user-missing", "code 9 listed as user-missing"),
292    "/entries/1/values/[value=9]/count" = tol(sum(d$grade == 9, na.rm = TRUE), "code 9 count"),
293    "/entries/2/level" = exact("nominal", "text is nominal"),
294    "/entries/2/values/[value=A]/count" = tol(sum(d$group == "A"), "group A count"),
295    "/entries/2/values/[value=B]/valid_percent" = tol(100 * sum(d$group == "B") / nrow(d), "group B valid percent"),
296    "/entries/2/valid" = tol(nrow(d), "every group value is valid")
297  )
298  emit("STAT-DESC-CODEBOOK-001", "Codebook: scale summary, labelled categories and missing counts",
299       "nominal", "tabelz.analysis.report_listings:compute_codebook", "codebookMenuItem",
300       "CODEBOOK",
301       list(columns = list("score", "grade", "group"), metadata = metadata),
302       cb_outputs,
303       c("R: mean, sd, min, max and quantile(type = 6) on the 20 valid scores; table() counts for grade and group with the labelled code 5 at 0.",
304         paste0(manual, ":11-14: counts and labels including labelled categories with no cases; scale variables carry numerical summaries.")),
305       "R base table and stats; quantile type 6",
306       "R stats::quantile documentation (type 6)")

Whole script generators/r/descriptives_reports.R


Custom table: counts, column, row and table percents, chi-square

Case STAT-DESC-CUSTOMTABLES-001 · status: passed

Reference

Evidence type
R
Source
R base table, stats::chisq.test, prop.test, t.test and p.adjust
Citation
R stats documentation: chisq.test, prop.test (correct = FALSE), t.test (var.equal = FALSE), p.adjust (holm, bonferroni)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_custom_tables.R · sha256 54816fc0af9ccf42
Dataset
reference_data/datasets/DESC_CUSTOM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/used_rowsexact40400.00E+2passed
/missing_rowsexact220.00E+2passed
/row_keys/0/0exactOpsOpspassed
/column_keys/2/0exactWestWestpassed
/column_letters/2exactCCpassed
/independence/0/variableexactdeptdeptpassed
/independence/0/chi_squaretolerance25.37024513947590925.370245139475912.42E-1617passed
/independence/0/dfexact440.00E+2passed
/independence/0/p_valuetail_probability4.2380878730493824e-054.238087873049383e-056.33E-2115passed
/independence/0/min_expectedtolerance3.89999999999999993.91.12E-1717passed
/independence/0/small_expectedexact880.00E+2passed
/independence/0/cellsexact990.00E+2passed
/cells/0/valuetolerance1010.00.00E+2passed
/cells/1/valuetolerance76.9230769230769276.923076923076922.03E-1617passed
/cells/2/valuetolerance71.42857142857143171.428571428571433.98E-1617passed
/cells/3/valuetolerance2525.00.00E+2passed
/cells/4/valuetolerance11.00.00E+2passed
/cells/5/valuetolerance8.33333333333333398.3333333333333342.55E-1717passed
/cells/6/valuetolerance7.14285714285714327.1428571428571433.78E-1717passed
/cells/7/valuetolerance2.52.50.00E+1passed
/cells/8/valuetolerance33.00.00E+2passed
/cells/9/valuetolerance2020.00.00E+2passed
/cells/10/valuetolerance21.42857142857142721.4285714285714274.88E-1717passed
/cells/11/valuetolerance7.57.50.00E+1passed
/cells/16/valuetolerance11.00.00E+2passed
/cells/17/valuetolerance7.69230769230769257.69230769230769251.27E-1717passed
/cells/18/valuetolerance7.69230769230769257.69230769230769251.27E-1717passed
/cells/19/valuetolerance2.52.50.00E+1passed
/cells/20/valuetolerance99.00.00E+2passed
/cells/21/valuetolerance7575.00.00E+2passed
/cells/22/valuetolerance69.23076923076922669.230769230769233.97E-1617passed
/cells/23/valuetolerance22.522.50.00E+1passed
/cells/24/valuetolerance33.00.00E+2passed
/cells/25/valuetolerance2020.00.00E+2passed
/cells/26/valuetolerance23.07692307692307723.0769230769230773.50E-1616passed
/cells/27/valuetolerance7.57.50.00E+1passed
/cells/32/valuetolerance22.00.00E+2passed
/cells/33/valuetolerance15.38461538461538515.3846153846153852.53E-1717passed
/cells/34/valuetolerance15.38461538461538515.3846153846153852.53E-1717passed
/cells/35/valuetolerance55.00.00E+2passed
/cells/36/valuetolerance22.00.00E+2passed
/cells/37/valuetolerance16.66666666666666816.6666666666666681.49E-1617passed
/cells/38/valuetolerance15.38461538461538515.3846153846153852.53E-1717passed
/cells/39/valuetolerance55.00.00E+2passed
/cells/40/valuetolerance99.00.00E+2passed
/cells/41/valuetolerance6060.00.00E+2passed
/cells/42/valuetolerance69.23076923076922669.230769230769233.97E-1617passed
/cells/43/valuetolerance22.522.50.00E+1passed
/cells/12/valuetolerance1414.00.00E+2passed
/cells/13/valuetolerance3535.00.00E+2passed
/cells/52/valuetolerance1212.00.00E+2passed
/cells/53/valuetolerance100100.00.00E+2passed
/cells/60/valuetolerance4040.00.00E+2passed
/cells/64absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

124  o1[[cell(0, 3, 0, 3, 4)]] <- tol(sum(tab[1, ]), "Ops total column count")
125  o1[[cell(0, 3, 1, 3, 4)]] <- tol(100 * sum(tab[1, ]) / N, "Ops total column percent is its share of all rows")
126  o1[[cell(3, 1, 0, 3, 4)]] <- tol(sum(tab[, 2]), "Total row, North count")
127  o1[[cell(3, 1, 1, 3, 4)]] <- tol(100, "Total row column percent is 100")
128  o1[[cell(3, 3, 0, 3, 4)]] <- tol(N, "grand total")
129  o1[["/cells/64"]] <- absent("4 blocks x 4 columns x 4 statistics = 64 cells")
130  emit("STAT-DESC-CUSTOMTABLES-001", "Custom table: counts, column, row and table percents, chi-square",
131       "nominal",
132       list(rows = list("dept"), columns = list("site"), statistics = as.list(stats1),
133            totals = TRUE, independence = TRUE, compare = "none", adjustment = "holm",
134            alpha = 0.05),
135       o1,
136       c("R: table(dept, site) over the 40 complete rows; column percent divides by the column total (the outer row group is the whole table with one row variable), row percent by the row total, table percent by 40; chisq.test(correct = FALSE).",
137         "Cell order: for each row (Ops, Sales, Tech, Total) the columns East, North, West, Total, each with the four statistics in the requested order.",
138         paste0(manual, ":50-55: column percentages are taken within the outer row group; chi-square with expected-count diagnostics.")))

Whole script generators/r/descriptives_custom_tables.R


Custom table: column proportion comparisons with Holm letters

Case STAT-DESC-CUSTOMTABLES-002 · status: passed

Reference

Evidence type
R
Source
R base table, stats::chisq.test, prop.test, t.test and p.adjust
Citation
R stats documentation: chisq.test, prop.test (correct = FALSE), t.test (var.equal = FALSE), p.adjust (holm, bonferroni)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_custom_tables.R · sha256 54816fc0af9ccf42
Dataset
reference_data/datasets/DESC_CUSTOM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/comparisons/9absentNonepassed
/comparisons/0/subjectexactdept: Opsdept: Opspassed
/comparisons/0/firstexactEastEastpassed
/comparisons/0/secondexactNorthNorthpassed
/comparisons/0/statistictolerance3.45168473924365543.45168473924365542.46E-1717passed
/comparisons/0/p_valuetail_probability0.000557098236277389590.00055709823627738962.33E-2117passed
/comparisons/0/adjusted_ptail_probability0.00501388412649650650.00501388412649650654.78E-2017passed
/comparisons/0/family_sizeexact990.00E+2passed
/comparisons/1/subjectexactdept: Opsdept: Opspassed
/comparisons/1/firstexactEastEastpassed
/comparisons/1/secondexactWestWestpassed
/comparisons/1/statistictolerance3.01208610798122623.01208610798122622.40E-1717passed
/comparisons/1/p_valuetail_probability0.00259458987642414980.00259458987642414643.43E-1814passed
/comparisons/1/adjusted_ptail_probability0.0181621291349690490.0181621291349690252.44E-1714passed
/comparisons/1/family_sizeexact990.00E+2passed
/comparisons/2/subjectexactdept: Opsdept: Opspassed
/comparisons/2/firstexactNorthNorthpassed
/comparisons/2/secondexactWestWestpassed
/comparisons/2/statistictolerance-0.84795149574161077-0.84795149574161082.33E-1817passed
/comparisons/2/p_valuetail_probability0.396464984017817380.396464984017817271.12E-1615passed
/comparisons/2/adjusted_ptail_probability11.00.00E+2passed
/comparisons/2/family_sizeexact990.00E+2passed
/comparisons/3/subjectexactdept: Salesdept: Salespassed
/comparisons/3/firstexactEastEastpassed
/comparisons/3/secondexactNorthNorthpassed
/comparisons/3/statistictolerance-3.4320323649182209-3.4320323649182214.58E-1716passed
/comparisons/3/p_valuetail_probability0.000599076060154916890.0005990760601549168.67E-1914passed
/comparisons/3/adjusted_ptail_probability0.00501388412649650650.00501388412649650654.78E-2017passed
/comparisons/3/family_sizeexact990.00E+2passed
/comparisons/4/subjectexactdept: Salesdept: Salespassed
/comparisons/4/firstexactEastEastpassed
/comparisons/4/secondexactWestWestpassed
/comparisons/4/statistictolerance-0.92819096178451421-0.92819096178451422.42E-1817passed
/comparisons/4/p_valuetail_probability0.353308521937843680.35330852193784371.90E-1817passed
/comparisons/4/adjusted_ptail_probability11.00.00E+2passed
/comparisons/4/family_sizeexact990.00E+2passed
/comparisons/5/subjectexactdept: Salesdept: Salespassed
/comparisons/5/firstexactNorthNorthpassed
/comparisons/5/secondexactWestWestpassed
/comparisons/5/statistictolerance2.85788383248864752.85788383248864752.42E-1717passed
/comparisons/5/p_valuetail_probability0.00426476491403053430.0042647649140305295.20E-1814passed
/comparisons/5/adjusted_ptail_probability0.0255885894841832060.0255885894841831753.14E-1714passed
/comparisons/5/family_sizeexact990.00E+2passed
/comparisons/6/subjectexactdept: Techdept: Techpassed
/comparisons/6/firstexactEastEastpassed
/comparisons/6/secondexactNorthNorthpassed
/comparisons/6/statistictolerance-0.0873570697268264-0.08735706972682641.47E-1917passed
/comparisons/6/p_valuetail_probability0.93038769226888540.93038769226888551.13E-1615passed
/comparisons/6/adjusted_ptail_probability11.00.00E+2passed
/comparisons/6/family_sizeexact990.00E+2passed
/comparisons/7/subjectexactdept: Techdept: Techpassed
/comparisons/7/firstexactEastEastpassed
/comparisons/7/secondexactWestWestpassed
/comparisons/7/statistictolerance-2.4107943187305225-2.41079431873052252.30E-1717passed
/comparisons/7/p_valuetail_probability0.0159178223723700110.015917822372379.97E-1815passed
/comparisons/7/adjusted_ptail_probability0.0795891118618500540.079589111861855.58E-1715passed
/comparisons/7/family_sizeexact990.00E+2passed
/comparisons/8/subjectexactdept: Techdept: Techpassed
/comparisons/8/firstexactNorthNorthpassed
/comparisons/8/secondexactWestWestpassed
/comparisons/8/statistictolerance-2.2771094116724551-2.2771094116724553.62E-1716passed
/comparisons/8/p_valuetail_probability0.0227796928098559050.022779692809855891.36E-1715passed
/comparisons/8/adjusted_ptail_probability0.0911187712394236210.091118771239423575.55E-1715passed
/comparisons/8/family_sizeexact990.00E+2passed
/cells/0/flagsexactBCBCpassed
/cells/2/flagsexactpassed
/cells/4/flagsexactpassed
/cells/8/flagsexactpassed
/cells/10/flagsexactACACpassed
/cells/12/flagsexactpassed
/cells/16/flagsexactpassed
/cells/18/flagsexactpassed
/cells/20/flagsexactpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

182  emit("STAT-DESC-CUSTOMTABLES-002", "Custom table: column proportion comparisons with Holm letters",
183       "axis_coverage",
184       list(rows = list("dept"), columns = list("site"), statistics = list("count", "column_percent"),
185            totals = TRUE, independence = FALSE, compare = "proportions", adjustment = "holm",
186            alpha = 0.05),
187       o2,
188       c("R: for each dept and each pair of sites, the pooled-null two-proportion z of the dept's share of the two site columns (prop.test(correct = FALSE) gives z^2 and the same p); p.adjust(holm) over the 9 comparisons; a count cell carries the letters of the columns that are significantly smaller.",
189         paste0(manual, ":57-61: two-proportion tests, Holm by default over all comparisons in the layer, a letter names a column with a significantly smaller value.")))

Whole script generators/r/descriptives_custom_tables.R


Custom table: scale summaries and Welch mean comparisons with Bonferroni

Case STAT-DESC-CUSTOMTABLES-003 · status: passed

Reference

Evidence type
R
Source
R base table, stats::chisq.test, prop.test, t.test and p.adjust
Citation
R stats documentation: chisq.test, prop.test (correct = FALSE), t.test (var.equal = FALSE), p.adjust (holm, bonferroni)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_custom_tables.R · sha256 54816fc0af9ccf42
Dataset
reference_data/datasets/DESC_CUSTOM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/comparisons/0/family_sizeexact990.00E+2passed
/comparisons/0/statisticabsentNonepassed
/comparisons/0/adjusted_pabsentNonepassed
/comparisons/1/family_sizeexact990.00E+2passed
/comparisons/1/statistictolerance-42.465332301523155-42.4653323015231553.25E-1718passed
/comparisons/1/p_valuetail_probability5.325826372087215e-135.325826372087189e-132.62E-2714passed
/comparisons/1/adjusted_ptail_probability4.7932437348784934e-124.79324373487847e-122.34E-2614passed
/comparisons/2/family_sizeexact990.00E+2passed
/comparisons/2/statisticabsentNonepassed
/comparisons/2/adjusted_pabsentNonepassed
/comparisons/3/family_sizeexact990.00E+2passed
/comparisons/3/statisticabsentNonepassed
/comparisons/3/adjusted_pabsentNonepassed
/comparisons/4/family_sizeexact990.00E+2passed
/comparisons/4/statisticabsentNonepassed
/comparisons/4/adjusted_pabsentNonepassed
/comparisons/5/family_sizeexact990.00E+2passed
/comparisons/5/statistictolerance-11.404537543005951-11.404537543005952.05E-1616passed
/comparisons/5/p_valuetail_probability0.0010152169784332730.00101521697843327582.80E-1814passed
/comparisons/5/adjusted_ptail_probability0.00913695280589945680.0091369528058994832.60E-1714passed
/comparisons/6/family_sizeexact990.00E+2passed
/comparisons/6/statistictolerance-5.9207226111810662-5.9207226111810664.17E-1717passed
/comparisons/6/p_valuetail_probability0.0876425319414304020.087642531941430421.35E-1715passed
/comparisons/6/adjusted_ptail_probability0.788782787472873670.78878278747287381.11E-1615passed
/comparisons/7/family_sizeexact990.00E+2passed
/comparisons/7/statistictolerance-32.456658227793142-32.456658227793143.09E-1617passed
/comparisons/7/p_valuetail_probability1.6867623945864102e-051.686762394586407e-053.10E-2014passed
/comparisons/7/adjusted_ptail_probability0.000151808615512776920.000151808615512776652.74E-1914passed
/comparisons/8/family_sizeexact990.00E+2passed
/comparisons/8/statistictolerance-5.3210890454535509-5.3210890454535511.91E-1717passed
/comparisons/8/p_valuetail_probability0.0972676361691797860.097267636169179834.15E-1715passed
/comparisons/8/adjusted_ptail_probability0.875408725522618060.87540872552261843.32E-1615passed
/cells/0/valuetolerance28.26000000000000228.264.37E-1616passed
/cells/1/valuetolerance2.57172661412090612.57172661412090564.59E-1615passed
/cells/2/valuetolerance28.30000000000000128.32.89E-1616passed
/cells/3/valuetolerance1010.00.00E+2passed
/cells/4/valuetolerance1010.00.00E+2passed
/cells/5/valuetolerance48.79999999999999748.81.58E-1617passed
/cells/6/valueabsentNonepassed
/cells/7/valuetolerance48.79999999999999748.81.58E-1617passed
/cells/8/valuetolerance11.00.00E+2passed
/cells/9/valuetolerance11.00.00E+2passed
/cells/10/valuetolerance68.79999999999999768.81.58E-1617passed
/cells/11/valuetolerance0.86602540378443860.86602540378443863.41E-1817passed
/cells/12/valuetolerance69.29999999999999769.31.58E-1617passed
/cells/13/valuetolerance33.00.00E+2passed
/cells/14/valuetolerance33.00.00E+2passed
/cells/20/valuetolerance22.39999999999999922.44.21E-1616passed
/cells/21/valueabsentNonepassed
/cells/22/valuetolerance22.39999999999999922.44.21E-1616passed
/cells/23/valuetolerance11.00.00E+2passed
/cells/24/valuetolerance11.00.00E+2passed
/cells/25/valuetolerance50.54444444444444450.5444444444444442.87E-1617passed
/cells/26/valuetolerance2.86099244629862292.8609924462986232.65E-1717passed
/cells/27/valuetolerance50.70000000000000350.71.58E-1617passed
/cells/28/valuetolerance99.00.00E+2passed
/cells/29/valuetolerance99.00.00E+2passed
/cells/30/valuetolerance73.66666666666667173.666666666666674.04E-1617passed
/cells/31/valuetolerance3.0989245446337293.0989245446337291.69E-1717passed
/cells/32/valuetolerance75.20000000000000375.21.58E-1617passed
/cells/33/valuetolerance33.00.00E+2passed
/cells/34/valuetolerance33.00.00E+2passed
/cells/40/valuetolerance31.04999999999999731.0499999999999971.58E-1617passed
/cells/41/valuetolerance1.20208152801713021.20208152801713023.58E-1817passed
/cells/42/valuetolerance31.04999999999999731.0499999999999971.58E-1617passed
/cells/43/valuetolerance22.00.00E+2passed
/cells/44/valuetolerance22.00.00E+2passed
/cells/45/valuetolerance51.79999999999999751.81.58E-1617passed
/cells/46/valuetolerance4.80832611206852614.8083261120685261.47E-1717passed
/cells/47/valuetolerance51.79999999999999751.81.58E-1617passed
/cells/48/valuetolerance22.00.00E+2passed
/cells/49/valuetolerance22.00.00E+2passed
/cells/50/valuetolerance70.47142857142857570.471428571428588.28E-1717passed
/cells/51/valuetolerance2.29544061544785372.2954406154478544.60E-1615passed
/cells/52/valuetolerance7171.00.00E+2passed
/cells/53/valuetolerance77.00.00E+2passed
/cells/54/valuetolerance99.00.00E+2passed
/cells/75/valuetolerance49.87105263157894849.871052631578953.32E-1718passed
/cells/78/valuetolerance3838.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

235  all_sales <- dd$sales[!is.na(dd$sales)]
236  o3[[cell(3, 3, 0, 3, 5)]] <- tol(mean(all_sales), "grand mean of sales")
237  o3[[cell(3, 3, 3, 3, 5)]] <- tol(length(all_sales), "valid N over all complete rows")
238  emit("STAT-DESC-CUSTOMTABLES-003", "Custom table: scale summaries and Welch mean comparisons with Bonferroni",
239       "axis_coverage",
240       list(rows = list("dept"), columns = list("site"), summary = "sales",
241            statistics = as.list(stats3), totals = TRUE, independence = FALSE,
242            compare = "means", adjustment = "bonferroni", alpha = 0.05),
243       o3,
244       c("R: mean, sd, median and non-missing count of sales per dept x site cell; t.test(var.equal = FALSE) per dept and site pair; p.adjust(bonferroni) over the 9 planned comparisons with an uncomputable pair (one value) entering as p = 1.",
245         paste0(manual, ":57-61: Welch tests for numerical means; Bonferroni as an option.")))

Whole script generators/r/descriptives_custom_tables.R


Custom table: nested rows with subtotals and outer-group column percents

Case STAT-DESC-CUSTOMTABLES-004 · status: passed

Reference

Evidence type
R
Source
R base table, stats::chisq.test, prop.test, t.test and p.adjust
Citation
R stats documentation: chisq.test, prop.test (correct = FALSE), t.test (var.equal = FALSE), p.adjust (holm, bonferroni)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_custom_tables.R · sha256 54816fc0af9ccf42
Dataset
reference_data/datasets/DESC_CUSTOM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/row_keys/0/0exactOpsOpspassed
/row_keys/0/1exactJuniorJuniorpassed
/row_keys/6absentNonepassed
/cells/120absentNonepassed
/cells/0/valuetolerance55.00.00E+2passed
/cells/1/valuetolerance5050.00.00E+2passed
/cells/2/valuetolerance12.512.50.00E+1passed
/cells/3/valuetolerance11.00.00E+2passed
/cells/4/valuetolerance100100.00.00E+2passed
/cells/5/valuetolerance2.52.50.00E+1passed
/cells/6/valuetolerance11.00.00E+2passed
/cells/7/valuetolerance33.33333333333333633.3333333333333362.98E-1617passed
/cells/8/valuetolerance2.52.50.00E+1passed
/cells/12/valuetolerance55.00.00E+2passed
/cells/13/valuetolerance5050.00.00E+2passed
/cells/14/valuetolerance12.512.50.00E+1passed
/cells/15/valuetolerance00.00.00E+2passed
/cells/16/valuetolerance00.00.00E+2passed
/cells/17/valuetolerance00.00.00E+2passed
/cells/18/valuetolerance22.00.00E+2passed
/cells/19/valuetolerance66.66666666666667166.666666666666674.04E-1617passed
/cells/20/valuetolerance55.00.00E+2passed
/cells/24/valuetolerance1010.00.00E+2passed
/cells/25/valuetolerance76.9230769230769276.923076923076922.03E-1617passed
/cells/27/valuetolerance11.00.00E+2passed
/cells/28/valuetolerance8.33333333333333398.3333333333333342.55E-1717passed
/cells/30/valuetolerance33.00.00E+2passed
/cells/31/valuetolerance2020.00.00E+2passed
/cells/33/valuetolerance1414.00.00E+2passed
/cells/36/valuetolerance11.00.00E+2passed
/cells/37/valuetolerance100100.00.00E+2passed
/cells/38/valuetolerance2.52.50.00E+1passed
/cells/39/valuetolerance44.00.00E+2passed
/cells/40/valuetolerance44.44444444444444344.444444444444441.35E-1617passed
/cells/41/valuetolerance1010.00.00E+2passed
/cells/42/valuetolerance22.00.00E+2passed
/cells/43/valuetolerance66.66666666666667166.666666666666674.04E-1617passed
/cells/44/valuetolerance55.00.00E+2passed
/cells/48/valuetolerance00.00.00E+2passed
/cells/49/valuetolerance00.00.00E+2passed
/cells/50/valuetolerance00.00.00E+2passed
/cells/51/valuetolerance55.00.00E+2passed
/cells/52/valuetolerance55.55555555555555755.555555555555561.35E-1617passed
/cells/53/valuetolerance12.512.50.00E+1passed
/cells/54/valuetolerance11.00.00E+2passed
/cells/55/valuetolerance33.33333333333333633.3333333333333362.98E-1617passed
/cells/56/valuetolerance2.52.50.00E+1passed
/cells/60/valuetolerance11.00.00E+2passed
/cells/61/valuetolerance7.69230769230769257.69230769230769251.27E-1717passed
/cells/63/valuetolerance99.00.00E+2passed
/cells/64/valuetolerance7575.00.00E+2passed
/cells/66/valuetolerance33.00.00E+2passed
/cells/67/valuetolerance2020.00.00E+2passed
/cells/69/valuetolerance1313.00.00E+2passed
/cells/72/valuetolerance11.00.00E+2passed
/cells/73/valuetolerance5050.00.00E+2passed
/cells/74/valuetolerance2.52.50.00E+1passed
/cells/75/valuetolerance11.00.00E+2passed
/cells/76/valuetolerance5050.00.00E+2passed
/cells/77/valuetolerance2.52.50.00E+1passed
/cells/78/valuetolerance55.00.00E+2passed
/cells/79/valuetolerance55.55555555555555755.555555555555561.35E-1617passed
/cells/80/valuetolerance12.512.50.00E+1passed
/cells/84/valuetolerance11.00.00E+2passed
/cells/85/valuetolerance5050.00.00E+2passed
/cells/86/valuetolerance2.52.50.00E+1passed
/cells/87/valuetolerance11.00.00E+2passed
/cells/88/valuetolerance5050.00.00E+2passed
/cells/89/valuetolerance2.52.50.00E+1passed
/cells/90/valuetolerance44.00.00E+2passed
/cells/91/valuetolerance44.44444444444444344.444444444444441.35E-1617passed
/cells/92/valuetolerance1010.00.00E+2passed
/cells/96/valuetolerance22.00.00E+2passed
/cells/97/valuetolerance15.38461538461538515.3846153846153852.53E-1717passed
/cells/99/valuetolerance22.00.00E+2passed
/cells/100/valuetolerance16.66666666666666816.6666666666666681.49E-1617passed
/cells/102/valuetolerance99.00.00E+2passed
/cells/103/valuetolerance6060.00.00E+2passed
/cells/105/valuetolerance1313.00.00E+2passed
/cells/117/valuetolerance4040.00.00E+2passed
/cells/109/valuetolerance100100.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

276  o4[[cell(9, 3, 0, 3, 3)]] <- tol(N, "grand total")
277  o4[[cell(9, 0, 1, 3, 3)]] <- tol(100, "Total row column percent is 100")
278  emit("STAT-DESC-CUSTOMTABLES-004", "Custom table: nested rows with subtotals and outer-group column percents",
279       "axis_coverage",
280       list(rows = list("dept", "level"), columns = list("site"),
281            statistics = list("count", "column_percent", "layer_percent"), totals = TRUE,
282            independence = FALSE, compare = "none", adjustment = "holm", alpha = 0.05),
283       o4,
284       c("R: table(dept, level, site); a leaf row's column percent divides by the dept group's count in that site column; a subtotal's column percent divides by the column total; with no layer the layer percent divides by all 40 rows.",
285         "Cell order: Ops/Junior, Ops/Senior, Ops subtotal, Sales/Junior, Sales/Senior, Sales subtotal, Tech/Junior, Tech/Senior, Tech subtotal, Total; columns East, North, West, Total; statistics count, column %, layer %.",
286         paste0(manual, ":50-55: up to three nested variables, subtotals, column percentages within the outer row group.")))

Whole script generators/r/descriptives_custom_tables.R


Shapiro-Wilk normality

Case STAT-DESC-DIAG-001 · status: passed

Reference

Evidence type
R
Source
R survey/stats/nortest
Citation
R survey::svydesign/svymean, stats::shapiro.test and nortest::ad.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, nortest 1.0.4, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_diagnostics.R · sha256 5427a49203331a27
Dataset
reference_data/datasets/DESC_DIAGNOSTIC_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance0.993347741115420460.99334774111542022.25E-1615passed
/p_valuetail_probability0.999990678520278230.99999067852027821.11E-1817passed
/observationsexact12120.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

84  shapiro <- stats::shapiro.test(diagnostic$value)
85  emit("STAT-DESC-DIAG-001", "Shapiro-Wilk normality", "tabelz.executor.distribution_diagnostics",
86       "distributionDiagnosticsMenuItem", "DESC_DIAGNOSTIC_01",
87       list(column = "value", method = "shapiro_wilk"),
88       list("/statistic" = tol(shapiro$statistic, "Shapiro-Wilk W; owner 1e-9 tier."),
89            "/p_value" = tail(shapiro$p.value, "Calibrated Shapiro-Wilk upper-tail p-value; owner 1e-9 tier."),
90            "/observations" = exact(nrow(diagnostic), "Twelve finite complete cases.")),
91       c("Manual chapter 30 lines 15-17: Shapiro-Wilk applies to 3-5000 finite values."))

Whole script generators/r/descriptives_survey_diagnostics.R


Anderson-Darling fitted-normal diagnostic

Case STAT-DESC-DIAG-002 · status: passed

Reference

Evidence type
R
Source
R survey/stats/nortest
Citation
R survey::svydesign/svymean, stats::shapiro.test and nortest::ad.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, nortest 1.0.4, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_diagnostics.R · sha256 5427a49203331a27
Dataset
reference_data/datasets/DESC_DIAGNOSTIC_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance0.0811063402129512670.081106340212949491.78E-1513passed
/p_valuetail_probability0.998168703122598870.99816870312259912.18E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

93  anderson <- nortest::ad.test(diagnostic$value)
94  emit("STAT-DESC-DIAG-002", "Anderson-Darling fitted-normal diagnostic",
95       "tabelz.executor.distribution_diagnostics", "distributionDiagnosticsMenuItem",
96       "DESC_DIAGNOSTIC_01", list(column = "value", method = "anderson_darling"),
97       list("/statistic" = tol(anderson$statistic, "Fitted-mean-and-variance Anderson-Darling statistic; owner 1e-9 tier."),
98            "/p_value" = tail(anderson$p.value, "Fitted-normal Anderson-Darling p-value; owner 1e-9 tier.")),
99       c("Manual chapter 30 lines 18-20: fitted normal mean and variance, not a known-parameter test."))

Whole script generators/r/descriptives_survey_diagnostics.R


All eight candidates on positive unimodal data

Case STAT-DESC-DISTFIT-001 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=normal]/parameters/0/1tolerance4.65435856000000044.654358563.34E-1717passed
/columns/0/candidates/[key=normal]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=normal]/parameters/1/1tolerance0.951272133532947710.95127213353294774.30E-1817passed
/columns/0/candidates/[key=normal]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=normal]/log_likelihoodtolerance-34.224585774197365-34.2245857741973651.40E-1617passed
/columns/0/candidates/[key=normal]/aictolerance72.4491715483947372.449171548394732.80E-1617passed
/columns/0/candidates/[key=normal]/bictolerance74.88692319813112974.886923198131132.57E-1617passed
/columns/0/candidates/[key=laplace]/parameters/0/1tolerance4.59830800000000034.5983081.60E-1717passed
/columns/0/candidates/[key=laplace]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=laplace]/parameters/1/1tolerance0.819638960000000030.81963895999999991.14E-1615passed
/columns/0/candidates/[key=laplace]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=laplace]/log_likelihoodtolerance-37.356396304897402-37.35639630489742.94E-1617passed
/columns/0/candidates/[key=laplace]/aictolerance78.71279260979480378.71279260979484.11E-1617passed
/columns/0/candidates/[key=laplace]/bictolerance81.15054425953120281.15054425953121.26E-1617passed
/columns/0/candidates/[key=logistic]/parameters/0/1tolerance4.65849369467201594.6584936948506391.79E-1010passed
/columns/0/candidates/[key=logistic]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=logistic]/parameters/1/1tolerance0.576438303683673550.57643830369848551.48E-1110passed
/columns/0/candidates/[key=logistic]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=logistic]/log_likelihoodtolerance-35.311585296391961-35.311585296391962.47E-1617passed
/columns/0/candidates/[key=logistic]/aictolerance74.62317059278392274.623170592783924.94E-1617passed
/columns/0/candidates/[key=logistic]/bictolerance77.0609222425203277.060922242520323.09E-1718passed
/columns/0/candidates/[key=lognormal]/parameters/0/1tolerance0.211591248812241730.211591248812241732.13E-1816passed
/columns/0/candidates/[key=lognormal]/parameters/0/0exactsspassed
/columns/0/candidates/[key=lognormal]/parameters/1/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=lognormal]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=lognormal]/parameters/2/1tolerance4.55384136886642214.5538413688664218.87E-1615passed
/columns/0/candidates/[key=lognormal]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=lognormal]/log_likelihoodtolerance-34.545268245090099-34.545268245090096.69E-1515passed
/columns/0/candidates/[key=lognormal]/aictolerance73.09053649018019973.090536490180181.44E-1415passed
/columns/0/candidates/[key=lognormal]/bictolerance75.52828813991659775.528288139916581.39E-1415passed
/columns/0/candidates/[key=gamma]/parameters/0/1tolerance23.06658295077778623.0665829507774872.99E-1313passed
/columns/0/candidates/[key=gamma]/parameters/0/0exactaapassed
/columns/0/candidates/[key=gamma]/parameters/1/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=gamma]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=gamma]/parameters/2/1tolerance0.201779282606878690.20177928260688132.61E-1513passed
/columns/0/candidates/[key=gamma]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=gamma]/log_likelihoodtolerance-34.323543742469269-34.3235437424692551.41E-1415passed
/columns/0/candidates/[key=gamma]/aictolerance72.64708748493853872.647087484938512.82E-1415passed
/columns/0/candidates/[key=gamma]/bictolerance75.08483913467493775.084839134674912.88E-1415passed
/columns/0/candidates/[key=weibull]/parameters/0/1tolerance5.60776618470200195.6077661847019992.62E-1515passed
/columns/0/candidates/[key=weibull]/parameters/0/0exactccpassed
/columns/0/candidates/[key=weibull]/parameters/1/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=weibull]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=weibull]/parameters/2/1tolerance5.04342556084793265.04342556084793358.75E-1615passed
/columns/0/candidates/[key=weibull]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=weibull]/log_likelihoodtolerance-34.111307853414985-34.1113078534149855.26E-1717passed
/columns/0/candidates/[key=weibull]/aictolerance72.2226157068299772.222615706829971.05E-1617passed
/columns/0/candidates/[key=weibull]/bictolerance74.66036735656636874.660367356566373.58E-1617passed
/columns/0/candidates/[key=exponential]/parameters/0/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=exponential]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=exponential]/parameters/1/1tolerance4.65435856000000044.654358563.34E-1717passed
/columns/0/candidates/[key=exponential]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=exponential]/log_likelihoodtolerance-63.445102632898326-63.445102632898337.18E-1515passed
/columns/0/candidates/[key=exponential]/aictolerance128.89020526579665128.890205265796683.06E-1415passed
/columns/0/candidates/[key=exponential]/bictolerance130.10908109066486130.109081090664861.51E-1516passed
/candidate_keyssequence['normal', 'student_t', 'laplace', 'logistic', 'lognormal', ['normal', 'student_t', 'laplace', 'logistic', 'lognormal', passed
/columns/0/countexact25250.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

195  set.seed(20260924)
196  export_dataset(data.frame(y = round(rgamma(25, shape = 6, rate = 1.5) + 1, 6)),
197                 "DESC_DISTFIT_01", "Constructed: rgamma(25, shape 6, rate 1.5) + 1, seed 20260924, rounded to 6 decimals",
198                 "R stats::rgamma documentation", "Constructed validation data; no third-party license.",
199                 c(y = "Float64"))
200  x <- read_dataset("DESC_DISTFIT_01")$y
201  n <- length(x)
202  if (!is.null(fit_t(x))) stop("expected no finite Student t maximum on DESC_DISTFIT_01")
203  fits <- list(fit_normal(x), fit_laplace(x), fit_logistic(x), fit_lognormal(x, 0),
204               fit_gamma(x, 0), fit_weibull(x, 0), fit_exponential(x, 0))
205  out <- list(); for (f in fits) out <- c(out, candidate_outputs(f, n))
206  gap <- min_gap(fits, n)
207  out[["/candidate_keys"]] <- keys_in_order(unlist(all8), "Selected candidates in the stable user-facing order.")
208  out[["/columns/0/count"]] <- exact(n, "All 25 values are finite.")
209  emit("STAT-DESC-DISTFIT-001", "All eight candidates on positive unimodal data", "DESC_DISTFIT_01",
210       list(columns = list("y"), candidates = all8, origin = 0), out,
211       c(sprintf("Smallest BIC gap between the seven finite candidates: %.6g.", gap),
212         "Gamma and Weibull solve their one-dimensional profile score with uniroot(tol = 1e-15); logistic solves its two score equations with nested uniroot(tol = 1e-15).",
213         "Student t: the df score of the profile likelihood is still positive at df = 1e4, so the likelihood rises toward the normal and no finite maximum exists; its outputs are deferred."),
214       deferred = list(list(path = "/columns/0/candidates/[key=student_t]",
215                            reason = "No finite Student t maximum-likelihood estimate on these data (df diverges); no reference value exists.")))

Whole script generators/r/descriptives_distribution_fit.R


Candidate ranking by BIC then AIC

Case STAT-DESC-DISTFIT-002 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/0/keyexactweibullweibullpassed
/columns/0/candidates/1/keyexactnormalnormalpassed
/columns/0/candidates/2/keyexactgammagammapassed
/columns/0/candidates/3/keyexactlognormallognormalpassed
/columns/0/candidates/4/keyexactlogisticlogisticpassed
/columns/0/candidates/5/keyexactlaplacelaplacepassed
/columns/0/candidates/6/keyexactexponentialexponentialpassed
/columns/0/best_labelexactWeibull (origin fixed at 0)Weibull (origin fixed at 0)passed
/columns/0/bictolerance74.66036735656636874.660367356566373.58E-1617passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

224  emit("STAT-DESC-DISTFIT-002", "Candidate ranking by BIC then AIC", "DESC_DISTFIT_01",
225       list(columns = list("y"), candidates = list("normal", "laplace", "logistic", "lognormal",
226                                                   "gamma", "weibull", "exponential"), origin = 0),
227       c(rank_out, list(
228            "/columns/0/best_label" = exact(
229              c(normal = "Normal", student_t = "Student t", laplace = "Laplace", logistic = "Logistic",
230                lognormal = "Lognormal (origin fixed at 0)", gamma = "Gamma (origin fixed at 0)",
231                weibull = "Weibull (origin fixed at 0)",
232                exponential = "Exponential (origin fixed at 0)")[[ranking[1]]],
233              "The leading label names the exact fixed origin of a positive family (issue #3)."),
234            "/columns/0/bic" = (if (ranking[1] %in% c("normal", "laplace", "lognormal", "exponential")) closed else iterative)(
235              log(n) * fits[[match(ranking[1], vapply(fits, `[[`, "", "key"))]]$k -
236                2 * fits[[match(ranking[1], vapply(fits, `[[`, "", "key"))]]$ll,
237              "BIC of the leading candidate."))),
238       c(sprintf("R ranking by BIC then AIC: %s; smallest BIC gap %.6g.", paste(ranking, collapse = ", "), gap),
239         "Student t is left out: it has no finite maximum on these data (see STAT-DESC-DISTFIT-001)."),
240       kind = "threshold")

Whole script generators/r/descriptives_distribution_fit.R


Positive families at a declared origin of 2.5

Case STAT-DESC-DISTFIT-003 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=lognormal]/parameters/0/1tolerance0.537840758640281020.5378407586402814.76E-1817passed
/columns/0/candidates/[key=lognormal]/parameters/0/0exactsspassed
/columns/0/candidates/[key=lognormal]/parameters/1/1tolerance2.52.50.00E+1passed
/columns/0/candidates/[key=lognormal]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=lognormal]/parameters/2/1tolerance1.9019307009474681.9019307009474683.29E-1716passed
/columns/0/candidates/[key=lognormal]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=lognormal]/log_likelihoodtolerance-36.040382788505106-36.0403827885051062.07E-1617passed
/columns/0/candidates/[key=lognormal]/aictolerance76.08076557701021276.080765577010214.13E-1617passed
/columns/0/candidates/[key=lognormal]/bictolerance78.5185172267466178.518517226746614.93E-1718passed
/columns/0/candidates/[key=gamma]/parameters/0/1tolerance4.17148508795537784.1714850879553779.31E-1615passed
/columns/0/candidates/[key=gamma]/parameters/0/0exactaapassed
/columns/0/candidates/[key=gamma]/parameters/1/1tolerance2.52.50.00E+1passed
/columns/0/candidates/[key=gamma]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=gamma]/parameters/2/1tolerance0.516448822080278090.51644882208027821.13E-1615passed
/columns/0/candidates/[key=gamma]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=gamma]/log_likelihoodtolerance-34.686907703712279-34.686907703712282.72E-1617passed
/columns/0/candidates/[key=gamma]/aictolerance73.37381540742455973.373815407424564.57E-1617passed
/columns/0/candidates/[key=gamma]/bictolerance75.81156705716095775.811567057160966.09E-1819passed
/columns/0/candidates/[key=weibull]/parameters/0/1tolerance2.47023214887089672.47023214887089634.40E-1615passed
/columns/0/candidates/[key=weibull]/parameters/0/0exactccpassed
/columns/0/candidates/[key=weibull]/parameters/1/1tolerance2.52.50.00E+1passed
/columns/0/candidates/[key=weibull]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=weibull]/parameters/2/1tolerance2.43303753177635912.43303753177635864.80E-1615passed
/columns/0/candidates/[key=weibull]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=weibull]/log_likelihoodtolerance-33.672596016577792-33.6725960165777857.11E-1515passed
/columns/0/candidates/[key=weibull]/aictolerance71.34519203315558471.345192033155571.42E-1415passed
/columns/0/candidates/[key=weibull]/bictolerance73.78294368289198273.782943682891971.37E-1415passed
/columns/0/candidates/[key=exponential]/parameters/0/1tolerance2.52.50.00E+1passed
/columns/0/candidates/[key=exponential]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=exponential]/parameters/1/1tolerance2.15435855999999992.154358562.25E-1716passed
/columns/0/candidates/[key=exponential]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=exponential]/log_likelihoodtolerance-44.187325681910316-44.1873256819103161.77E-1617passed
/columns/0/candidates/[key=exponential]/aictolerance90.37465136382063290.374651363820633.55E-1617passed
/columns/0/candidates/[key=exponential]/bictolerance91.59352718868883991.593527188688843.08E-1617passed
/columns/0/candidates/[key=lognormal]/labelexactLognormal (origin fixed at 2.5)Lognormal (origin fixed at 2.5)passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

263  origin <- 2.5
264  positive <- list(fit_lognormal(x, origin), fit_gamma(x, origin), fit_weibull(x, origin),
265                   fit_exponential(x, origin))
266  if (min(x) <= origin) stop("dataset must lie above the origin")
267  out <- list(); for (f in positive) out <- c(out, candidate_outputs(f, n))
268  out[["/columns/0/candidates/[key=lognormal]/label"]] <- exact("Lognormal (origin fixed at 2.5)", "The label names the exact declared origin (issue #3).")
269  emit("STAT-DESC-DISTFIT-003", "Positive families at a declared origin of 2.5", "DESC_DISTFIT_01",
270       list(columns = list("y"), candidates = list("lognormal", "gamma", "weibull", "exponential"), origin = origin),
271       out, c("The origin 2.5 lies below every value, so no family is skipped; the fits use y - 2.5."))

Whole script generators/r/descriptives_distribution_fit.R


Anderson-Darling statistic of the normal fit

Case STAT-DESC-DISTFIT-004 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=normal]/parameters/0/1tolerance4.65435856000000044.654358563.34E-1717passed
/columns/0/candidates/[key=normal]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=normal]/parameters/1/1tolerance0.951272133532947710.95127213353294774.30E-1817passed
/columns/0/candidates/[key=normal]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=normal]/log_likelihoodtolerance-34.224585774197365-34.2245857741973651.40E-1617passed
/columns/0/candidates/[key=normal]/aictolerance72.4491715483947372.449171548394732.80E-1617passed
/columns/0/candidates/[key=normal]/bictolerance74.88692319813112974.886923198131132.57E-1617passed
/columns/0/goodness/valuetolerance0.311446129085929610.31144612908592962.76E-1817passed
/columns/0/goodness/statisticexactAnderson-DarlingAnderson-Darlingpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

274  nf <- fit_normal(x)
275  emit("STAT-DESC-DISTFIT-004", "Anderson-Darling statistic of the normal fit", "DESC_DISTFIT_01",
276       list(columns = list("y"), candidates = list("normal"), origin = 0,
277            calibration = "fixed", mc_samples = 99L, seed = 17L),
278       c(candidate_outputs(nf, n),
279         list("/columns/0/goodness/value" = closed(
280                anderson_darling(x, function(v) pnorm(v, nf$par[["loc"]], nf$par[["scale"]])),
281                "A^2 = -n - (1/n) sum (2i-1)[log F(x_i) + log(1 - F(x_(n+1-i)))] at the MLE."),
282              "/columns/0/goodness/statistic" = exact("Anderson-Darling", "Named statistic."))),
283       c("The Monte Carlo p-value (99 seeded refits) belongs to the V27 simulation studies."),
284       deferred = list(list(path = "/columns/0/goodness/p_value",
285                            reason = "Monte Carlo p-value; covered by V27 simulation studies, not a closed-form reference.")),
286       version = "1.1.0")

Whole script generators/r/descriptives_distribution_fit.R


Smallest accepted sample of eight

Case STAT-DESC-DISTFIT-005 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_MIN8.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=normal]/parameters/0/1tolerance5.17499999999999985.1752.24E-1717passed
/columns/0/candidates/[key=normal]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=normal]/parameters/1/1tolerance0.677772085586298090.6777720855862981.11E-1615passed
/columns/0/candidates/[key=normal]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=normal]/log_likelihoodtolerance-8.2399546295772392-8.2399546295772371.75E-1515passed
/columns/0/candidates/[key=normal]/aictolerance20.47990925915447820.4799092591544753.09E-1515passed
/columns/0/candidates/[key=normal]/bictolerance20.63879234251415120.6387923425141473.71E-1515passed
/columns/0/candidates/[key=exponential]/parameters/0/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=exponential]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=exponential]/parameters/1/1tolerance5.17499999999999985.1752.24E-1717passed
/columns/0/candidates/[key=exponential]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=exponential]/log_likelihoodtolerance-21.150714713211464-21.1507147132114649.79E-1717passed
/columns/0/candidates/[key=exponential]/aictolerance44.30142942642292844.301429426422931.96E-1617passed
/columns/0/candidates/[key=exponential]/bictolerance44.38087096810276244.380870968102762.20E-1617passed
/columns/0/countexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

327  export_dataset(data.frame(y = c(4.1, 5.3, 4.8, 6.2, 5.9, 5.1, 4.4, 5.6)),
328                 "DESC_DISTFIT_MIN8", "Constructed: eight values, the manual's minimum",
329                 "Hand-constructed", "Constructed validation data; no third-party license.", c(y = "Float64"))
330  x8 <- read_dataset("DESC_DISTFIT_MIN8")$y
331  out <- c(candidate_outputs(fit_normal(x8), 8), candidate_outputs(fit_exponential(x8, 0), 8))
332  out[["/columns/0/count"]] <- exact(8L, "Eight finite values, the minimum.")
333  emit("STAT-DESC-DISTFIT-005", "Smallest accepted sample of eight", "DESC_DISTFIT_MIN8",
334       list(columns = list("y"), candidates = list("normal", "exponential"), origin = 0), out,
335       c("distribution_fit.MIN_OBSERVATIONS is 8; Laplace is omitted because its location is not unique for even n."),
336       kind = "boundary")

Whole script generators/r/descriptives_distribution_fit.R


Positive families skipped for a value at the origin

Case STAT-DESC-DISTFIT-006 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_ZERO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=normal]/parameters/0/1tolerance1.92727272727272721.92727272727272724.85E-1716passed
/columns/0/candidates/[key=normal]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=normal]/parameters/1/1tolerance0.819191296441368010.8191912964413689.39E-1917passed
/columns/0/candidates/[key=normal]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=normal]/log_likelihoodtolerance-13.414509723891303-13.4145097238913011.64E-1515passed
/columns/0/candidates/[key=normal]/aictolerance30.82901944778260630.8290194477826033.27E-1515passed
/columns/0/candidates/[key=normal]/bictolerance31.62480999337934831.6248099933793443.94E-1515passed
/columns/0/candidates/1absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

339  export_dataset(data.frame(y = c(0, 1.2, 2.3, 1.9, 3.1, 2.6, 1.4, 2.2, 2.8, 1.7, 2.0)),
340                 "DESC_DISTFIT_ZERO", "Constructed: eleven values including one at the origin 0",
341                 "Hand-constructed", "Constructed validation data; no third-party license.", c(y = "Float64"))
342  xz <- read_dataset("DESC_DISTFIT_ZERO")$y
343  emit("STAT-DESC-DISTFIT-006", "Positive families skipped for a value at the origin", "DESC_DISTFIT_ZERO",
344       list(columns = list("y"), candidates = list("normal", "lognormal", "gamma"), origin = 0),
345       c(candidate_outputs(fit_normal(xz), length(xz)),
346         list("/columns/0/candidates/1" = list(comparison = "absent",
347                rationale = "Lognormal and gamma need every value strictly above the origin, so both are skipped, leaving the normal fit only."))),
348       c("One value equals the origin 0, so both strictly positive families are skipped (issue #3 ruling, item 5)."),
349       kind = "boundary")

Whole script generators/r/descriptives_distribution_fit.R


Missing values and excluded rows

Case STAT-DESC-DISTFIT-007 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=normal]/parameters/0/1tolerance50.16940909090909250.169409090909092.69E-1617passed
/columns/0/candidates/[key=normal]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=normal]/parameters/1/1tolerance3.62441699323166233.62441699323166233.17E-1717passed
/columns/0/candidates/[key=normal]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=normal]/log_likelihoodtolerance-29.772951770358866-29.7729517703588624.02E-1515passed
/columns/0/candidates/[key=normal]/aictolerance63.54590354071773163.5459035407177247.04E-1515passed
/columns/0/candidates/[key=normal]/bictolerance64.34169408631447664.341694086314461.43E-1415passed
/columns/0/candidates/[key=laplace]/parameters/0/1tolerance50.43869999999999750.43872.03E-1617passed
/columns/0/candidates/[key=laplace]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=laplace]/parameters/1/1tolerance2.68805454545454532.68805454545454534.78E-1716passed
/columns/0/candidates/[key=laplace]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=laplace]/log_likelihoodtolerance-29.501613847450319-29.5016138474503223.23E-1515passed
/columns/0/candidates/[key=laplace]/aictolerance63.00322769490063763.0032276949006447.45E-1515passed
/columns/0/candidates/[key=laplace]/bictolerance63.79901824049737563.799018240497391.43E-1415passed
/columns/0/countexact11110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

357  export_dataset(data.frame(id = seq_len(16L), y = raw), "DESC_DISTFIT_MISSING",
358                 "Constructed: rnorm(16, 50, 4), seed 20260925, rounded to 4 decimals, rows 3 and 11 missing",
359                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
360                 c(id = "Int64", y = "Float64"))
361  xm <- read_dataset("DESC_DISTFIT_MISSING")$y
362  if (length(xm) != 16L) stop("DESC_DISTFIT_MISSING lost rows on reading")
363  excluded <- c(0L, 7L, 12L)
364  kept <- xm[-(excluded + 1L)]; kept <- kept[!is.na(kept)]
365  out <- c(candidate_outputs(fit_normal(kept), length(kept)), candidate_outputs(fit_laplace(kept), length(kept)))
366  out[["/columns/0/count"]] <- exact(length(kept), "Excluded rows and missing values are both dropped.")
367  emit("STAT-DESC-DISTFIT-007", "Missing values and excluded rows", "DESC_DISTFIT_MISSING",
368       list(columns = list("y"), candidates = list("normal", "laplace"), origin = 0), out,
369       c(sprintf("Rows 0, 7 and 12 (0-based) excluded and two missing values dropped: %d values remain (odd, so the Laplace location is unique).", length(kept))),
370       kind = "missing_excluded", excluded = as.list(excluded))

Whole script generators/r/descriptives_distribution_fit.R


Large offset with a small spread

Case STAT-DESC-DISTFIT-008 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_OFFSET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=normal]/parameters/0/1tolerance1000000.00034959741000000.00034959783.73E-1015passed
/columns/0/candidates/[key=normal]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=normal]/parameters/1/1tolerance0.000965289082892162680.00096528908404893421.16E-128passed
/columns/0/candidates/[key=normal]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=normal]/log_likelihoodtolerance116.00703241088178116.007032385716042.52E-89passed
/columns/0/candidates/[key=normal]/aictolerance-228.01406482176355-228.01406477143215.03E-89passed
/columns/0/candidates/[key=normal]/bictolerance-225.9250199463167-225.925019895985235.03E-89passed
/columns/0/candidates/[key=laplace]/parameters/0/1tolerance1000000.0006775911000000.00067759117.33E-1116passed
/columns/0/candidates/[key=laplace]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=laplace]/parameters/1/1tolerance0.000763969468174590980.00076396946263100425.54E-128passed
/columns/0/candidates/[key=laplace]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=laplace]/log_likelihoodtolerance115.16054659533098115.160546747713161.52E-78passed
/columns/0/candidates/[key=laplace]/aictolerance-226.32109319066197-226.32109349542633.05E-78passed
/columns/0/candidates/[key=laplace]/bictolerance-224.23204831521511-224.232048619979453.05E-78passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

373  set.seed(20260926)
374  export_dataset(data.frame(y = 1e6 + round(rnorm(21, 0, 1e-3), 9)), "DESC_DISTFIT_OFFSET",
375                 "Constructed: 1e6 + rnorm(21, 0, 0.001), seed 20260926, rounded to 1e-9",
376                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.", c(y = "Float64"))
377  xo <- read_dataset("DESC_DISTFIT_OFFSET")$y
378  emit("STAT-DESC-DISTFIT-008", "Large offset with a small spread", "DESC_DISTFIT_OFFSET",
379       list(columns = list("y"), candidates = list("normal", "laplace"), origin = 0),
380       c(candidate_outputs(fit_normal(xo), length(xo)), candidate_outputs(fit_laplace(xo), length(xo))),
381       c("Values near 1e6 with a spread of 1e-3; the MLE scale is computed from centered values in R."),
382       kind = "hard_numeric")

Whole script generators/r/descriptives_distribution_fit.R


Seven values are below the minimum

Case STAT-DESC-DISTFIT-009 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_SEVEN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

385  export_dataset(data.frame(y = c(1.5, 2.5, 3.5, 2.0, 4.0, 3.0, 2.2)), "DESC_DISTFIT_SEVEN",
386                 "Constructed: seven values, one below the minimum", "Hand-constructed",
387                 "Constructed validation data; no third-party license.", c(y = "Float64"))
388  emit("STAT-DESC-DISTFIT-009", "Seven values are below the minimum", "DESC_DISTFIT_SEVEN",
389       list(columns = list("y"), candidates = list("normal"), origin = 0), list(),
390       c("distribution_fit.MIN_OBSERVATIONS is 8."), kind = "invalid_input",
391       error = list(error_class = "ValueError", message_contains = "needs at least 8 finite observations"))

Whole script generators/r/descriptives_distribution_fit.R


A constant column has no distribution

Case STAT-DESC-DISTFIT-010 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

392  export_dataset(data.frame(y = rep(3.25, 10)), "DESC_DISTFIT_CONSTANT",
393                 "Constructed: ten equal values", "Hand-constructed",
394                 "Constructed validation data; no third-party license.", c(y = "Float64"))
395  emit("STAT-DESC-DISTFIT-010", "A constant column has no distribution", "DESC_DISTFIT_CONSTANT",
396       list(columns = list("y"), candidates = list("normal"), origin = 0), list(),
397       c("A zero-variance column is refused before fitting."), kind = "invalid_input",
398       error = list(error_class = "ValueError", message_contains = "is constant"))

Whole script generators/r/descriptives_distribution_fit.R


An unknown candidate is refused

Case STAT-DESC-DISTFIT-011 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

399  emit("STAT-DESC-DISTFIT-011", "An unknown candidate is refused", "DESC_DISTFIT_01",
400       list(columns = list("y"), candidates = list("normal", "cauchy"), origin = 0), list(),
401       c("Only the eight manual candidates are accepted (manual lines 772-774)."), kind = "invalid_input",
402       error = list(error_class = "ValueError", message_contains = "Unknown distributions: cauchy"))

Whole script generators/r/descriptives_distribution_fit.R


Unrestricted Student t is refused

Case STAT-DESC-DISTFIT-012 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_HEAVY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

255  emit("STAT-DESC-DISTFIT-012", "Unrestricted Student t is refused", "DESC_DISTFIT_HEAVY",
256       list(columns = list("y"), candidates = list("student_t"), origin = 0), list(),
257       c(sprintf("A local profile maximum exists at df = %.10g, but the likelihood is unbounded as the scale tends to zero with df < 1/(N - 1).", th$par[["df"]]),
258         "Until 2026-09-25 this case expected that local maximum; the issue #7 ruling makes the candidate unavailable."),
259       kind = "invalid_input",
260       error = list(error_class = "ValueError", message_contains = "No finite unrestricted maximum-likelihood fit"))

Whole script generators/r/descriptives_distribution_fit.R


Exponential with a value at its origin

Case STAT-DESC-DISTFIT-013 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_ZERO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=exponential]/parameters/0/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=exponential]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=exponential]/parameters/1/1tolerance1.92727272727272721.92727272727272724.85E-1716passed
/columns/0/candidates/[key=exponential]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=exponential]/log_likelihoodtolerance-18.217164997675557-18.217164997675563.30E-1515passed
/columns/0/candidates/[key=exponential]/aictolerance38.43432999535111338.434329995351127.60E-1515passed
/columns/0/candidates/[key=exponential]/bictolerance38.83222526814948638.832225268149497.05E-1515passed
/columns/0/goodness/valueabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

406  emit("STAT-DESC-DISTFIT-013", "Exponential with a value at its origin", "DESC_DISTFIT_ZERO",
407       list(columns = list("y"), candidates = list("exponential"), origin = 0),
408       c(candidate_outputs(fit_exponential(xz, 0), length(xz)),
409         list("/columns/0/goodness/value" = list(comparison = "absent",
410                rationale = "At a value on the origin F(x) = 0, so the Anderson-Darling statistic is undefined and is reported unavailable, not approximated (issue #3)."))),
411       c("Until 2026-09-24 the product refused this valid fit; the statistician's ruling on issue #3 allows x = origin for the exponential only."),
412       kind = "boundary")

Whole script generators/r/descriptives_distribution_fit.R


Fixed-origin positive families on 1 to 8

Case STAT-DESC-DISTFIT-014 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_ONE_TO_EIGHT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=lognormal]/parameters/0/1tolerance0.657965568323904180.65796556832390421.90E-1817passed
/columns/0/candidates/[key=lognormal]/parameters/0/0exactsspassed
/columns/0/candidates/[key=lognormal]/parameters/1/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=lognormal]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=lognormal]/parameters/2/1tolerance3.76435059950312793.7643505995031287.88E-1817passed
/columns/0/candidates/[key=lognormal]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=lognormal]/log_likelihoodtolerance-18.607289753973134-18.6072897539731343.39E-1616passed
/columns/0/candidates/[key=lognormal]/aictolerance41.21457950794626941.214579507946273.23E-1617passed
/columns/0/candidates/[key=lognormal]/bictolerance41.37346259130593841.373462591305944.92E-1616passed
/columns/0/candidates/[key=gamma]/parameters/0/1tolerance2.95723801289341642.9572380128934181.74E-1515passed
/columns/0/candidates/[key=gamma]/parameters/0/0exactaapassed
/columns/0/candidates/[key=gamma]/parameters/1/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=gamma]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=gamma]/parameters/2/1tolerance1.52169016507302261.52169016507302179.17E-1615passed
/columns/0/candidates/[key=gamma]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=gamma]/log_likelihoodtolerance-18.066674545012052-18.0666745450120563.83E-1515passed
/columns/0/candidates/[key=gamma]/aictolerance40.13334909002410540.133349090024116.66E-1515passed
/columns/0/candidates/[key=gamma]/bictolerance40.29223217338377340.292232173383787.49E-1515passed
/columns/0/candidates/[key=weibull]/parameters/0/1tolerance2.06176772918195632.06176772918195549.11E-1615passed
/columns/0/candidates/[key=weibull]/parameters/0/0exactccpassed
/columns/0/candidates/[key=weibull]/parameters/1/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=weibull]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=weibull]/parameters/2/1tolerance5.07827032880672395.0782703288067221.77E-1515passed
/columns/0/candidates/[key=weibull]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=weibull]/log_likelihoodtolerance-17.754362454758262-17.7543624547582653.24E-1515passed
/columns/0/candidates/[key=weibull]/aictolerance39.50872490951652339.508724909516537.49E-1515passed
/columns/0/candidates/[key=weibull]/bictolerance39.66760799287619239.66760799287627.32E-1515passed
/columns/0/candidates/[key=exponential]/parameters/0/1tolerance00.00.00E+2passed
/columns/0/candidates/[key=exponential]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=exponential]/parameters/1/1tolerance4.54.50.00E+1passed
/columns/0/candidates/[key=exponential]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=exponential]/log_likelihoodtolerance-20.032619174210193-20.0326191742101933.36E-1616passed
/columns/0/candidates/[key=exponential]/aictolerance42.06523834842038742.065238348420393.29E-1617passed
/columns/0/candidates/[key=exponential]/bictolerance42.14467989010022142.144679890100228.69E-1717passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

423  emit("STAT-DESC-DISTFIT-014", "Fixed-origin positive families on 1 to 8", "DESC_DISTFIT_ONE_TO_EIGHT",
424       list(columns = list("y"), candidates = list("lognormal", "gamma", "weibull", "exponential"), origin = 0),
425       out,
426       c("Reproduces the statistician's issue #3 table (lognormal, gamma, Weibull, exponential at origin 0).",
427         "Weibull shape from its score equation, which SciPy's fit missed by 7.3e-6 relative."))

Whole script generators/r/descriptives_distribution_fit.R


Hybrid normality test with separate sample-SD estimators

Case STAT-DESC-DISTFIT-015 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=normal]/parameters/0/1tolerance4.65435856000000044.654358563.34E-1717passed
/columns/0/candidates/[key=normal]/parameters/0/0exactloclocpassed
/columns/0/candidates/[key=normal]/parameters/1/1tolerance0.951272133532947710.95127213353294774.30E-1817passed
/columns/0/candidates/[key=normal]/parameters/1/0exactscalescalepassed
/columns/0/candidates/[key=normal]/log_likelihoodtolerance-34.224585774197365-34.2245857741973651.40E-1617passed
/columns/0/candidates/[key=normal]/aictolerance72.4491715483947372.449171548394732.80E-1617passed
/columns/0/candidates/[key=normal]/bictolerance74.88692319813112974.886923198131132.57E-1617passed
/columns/0/goodness/valuetolerance0.280209894845544480.280209894845548043.56E-1513passed
/columns/0/goodness/modified_statistictolerance0.289624947312354770.28962494731235853.72E-1513passed
/columns/0/goodness/p_valuetolerance0.613632635005638140.61363263500562611.20E-1413passed
/columns/0/goodness/calibrationexactpublished_tablepublished_tablepassed
/columns/0/goodness/monte_carlo_intervalabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

314  emit("STAT-DESC-DISTFIT-015", "Hybrid normality test with separate sample-SD estimators", "DESC_DISTFIT_01",
315       list(columns = list("y"), candidates = list("normal"), origin = 0, calibration = "hybrid"),
316       c(candidate_outputs(nf, n), normality_outputs(x)),
317       c("Owner-approved #201 amendment: test mean/sample SD differ from the retained ML distribution fit.",
318         "D'Agostino and Stephens (1986), chapter 4, Case 3 p. 122; Tables 4.7 p. 123 and 4.9 p. 127."))

Whole script generators/r/descriptives_distribution_fit.R


Hybrid fixed-origin lognormal normality test

Case STAT-DESC-DISTFIT-016 · status: passed

Reference

Evidence type
R
Source
R stats (optimize, uniroot, nlminb)
Citation
R stats::nlminb, stats::uniroot and stats::dt/dlogis/dgamma/dweibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_distribution_fit.R · sha256 5d3685be492c803d
Dataset
reference_data/datasets/DESC_DISTFIT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/candidates/[key=lognormal]/parameters/0/1tolerance0.537840758640281020.5378407586402814.76E-1817passed
/columns/0/candidates/[key=lognormal]/parameters/0/0exactsspassed
/columns/0/candidates/[key=lognormal]/parameters/1/1tolerance2.52.50.00E+1passed
/columns/0/candidates/[key=lognormal]/parameters/1/0exactloclocpassed
/columns/0/candidates/[key=lognormal]/parameters/2/1tolerance1.9019307009474681.9019307009474683.29E-1716passed
/columns/0/candidates/[key=lognormal]/parameters/2/0exactscalescalepassed
/columns/0/candidates/[key=lognormal]/log_likelihoodtolerance-36.040382788505106-36.0403827885051062.07E-1617passed
/columns/0/candidates/[key=lognormal]/aictolerance76.08076557701021276.080765577010214.13E-1617passed
/columns/0/candidates/[key=lognormal]/bictolerance78.5185172267466178.518517226746614.93E-1718passed
/columns/0/goodness/valuetolerance0.619805458034143440.61980545803414341.64E-1817passed
/columns/0/goodness/modified_statistictolerance0.640630921424090730.64063092142409073.18E-1817passed
/columns/0/goodness/p_valuetolerance0.0947959737107281070.094795973710728113.54E-2018passed
/columns/0/goodness/calibrationexactpublished_tablepublished_tablepassed
/columns/0/goodness/monte_carlo_intervalabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

319  lnf <- fit_lognormal(x, 2.5)
320  emit("STAT-DESC-DISTFIT-016", "Hybrid fixed-origin lognormal normality test", "DESC_DISTFIT_01",
321       list(columns = list("y"), candidates = list("lognormal"), origin = 2.5, calibration = "hybrid"),
322       c(candidate_outputs(lnf, n), normality_outputs(log(x - 2.5))),
323       c("Owner-approved #201 amendment: normality of log(x - origin), with origin fixed at 2.5 and sample SD for the test only.",
324         "D'Agostino and Stephens (1986), chapter 4, Case 3 p. 122; Tables 4.7 p. 123 and 4.9 p. 127."))

Whole script generators/r/descriptives_distribution_fit.R


Explore one dependent: descriptives, M-estimators, extremes, normality, box

Case STAT-DESC-EXPLORE-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS, nortest and e1071
Citation
R stats::t.test, shapiro.test, quantile (type 6), fivenum; nortest::lillie.test (Dallal-Wilkinson/Stephens); MASS::huber; e1071 type-2 moments
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/0/nexact20200.00E+2passed
/cells/0/descriptives/0/1tolerance20.41499999999999920.4151.47E-1617passed
/cells/0/descriptives/0/2tolerance1.14255657466743821.14255657466743827.27E-1817passed
/cells/0/descriptives/1/1tolerance20.14444444444444620.1444444444444462.92E-1616passed
/cells/0/descriptives/2/1tolerance19.519.50.00E+1passed
/cells/0/descriptives/3/1tolerance26.10871052631579326.1087105263157933.24E-1717passed
/cells/0/descriptives/4/1tolerance5.10966833819141285.1096683381914134.91E-1717passed
/cells/0/descriptives/5/1tolerance12.512.50.00E+1passed
/cells/0/descriptives/6/1tolerance33.20000000000000333.21.58E-1617passed
/cells/0/descriptives/7/1tolerance20.70000000000000320.7000000000000031.58E-1617passed
/cells/0/descriptives/9/1tolerance0.830239578983928080.83023957898392832.26E-1615passed
/cells/0/descriptives/9/2tolerance0.512103336707757430.51210333670775741.69E-1817passed
/cells/0/descriptives/10/1tolerance0.721872896982482360.72187289698248246.37E-1918passed
/cells/0/descriptives/10/2tolerance0.992383612541845040.9923836125418456.55E-1918passed
/cells/0/mean_lowertolerance18.02360160569868718.0236016056986873.29E-1717passed
/cells/0/mean_uppertolerance22.80639839430131122.806398394301313.28E-1616passed
/cells/0/normality/[test=Shapiro-Wilk]/statistictolerance0.959744168874605120.95974416887460483.34E-1615passed
/cells/0/normality/[test=Shapiro-Wilk]/p_valuetail_probability0.538750826579401760.53875082657939615.66E-1513passed
/cells/0/normality/[test=Kolmogorov-Smirnov (Lilliefors)]/statistictolerance0.124578359462727160.124578359462727161.77E-1816passed
/cells/0/normality/[test=Kolmogorov-Smirnov (Lilliefors)]/p_valuetail_probability0.573500203473542140.57350020347354174.48E-1615passed
/cells/0/m_estimates/0/valuetolerance20.01486143999785120.0148614384714171.53E-910passed
/cells/0/m_estimates/1/valuetolerance19.99215720290708419.992157199098513.81E-99passed
/cells/0/m_estimates/2/valuetolerance20.14059377777773520.1405937767932789.84E-1010passed
/cells/0/m_estimates/3/valuetolerance19.99455887869761719.9945588748364143.86E-99passed
/cells/0/extremes/0/valuetolerance33.20000000000000333.21.58E-1617passed
/cells/0/extremes/0/caseexact23230.00E+2passed
/cells/0/extremes/5/valuetolerance12.512.50.00E+1passed
/cells/0/extremes/5/caseexact110.00E+2passed
/cells/0/extremes/1/valuetolerance28.69999999999999928.72.89E-1616passed
/cells/0/extremes/1/caseexact22220.00E+2passed
/cells/0/extremes/6/valuetolerance14.114.13.55E-1616passed
/cells/0/extremes/6/caseexact220.00E+2passed
/cells/0/extremes/2/valuetolerance26.30000000000000126.32.89E-1616passed
/cells/0/extremes/2/caseexact20200.00E+2passed
/cells/0/extremes/7/valuetolerance15.30000000000000115.32.89E-1616passed
/cells/0/extremes/7/caseexact24240.00E+2passed
/cells/0/extremes/3/valuetolerance24.89999999999999924.94.21E-1616passed
/cells/0/extremes/3/caseexact19190.00E+2passed
/cells/0/extremes/8/valuetolerance15.915.93.55E-1616passed
/cells/0/extremes/8/caseexact440.00E+2passed
/cells/0/extremes/4/valuetolerance23.39999999999999923.44.21E-1616passed
/cells/0/extremes/4/caseexact18180.00E+2passed
/cells/0/extremes/9/valuetolerance16.39999999999999916.44.21E-1616passed
/cells/0/extremes/9/caseexact550.00E+2passed
/cells/0/descriptives/8/1tolerance6.64999999999999866.6500000000000023.53E-1515passed
/cells/0/box/q1tolerance16.59999999999999816.5999999999999981.32E-1617passed
/cells/0/box/q3tolerance23.24999999999999623.254.00E-1515passed
/cells/0/box/mediantolerance19.519.50.00E+1passed
/cells/0/box/whisker_lowtolerance12.512.50.00E+1passed
/cells/0/box/whisker_hightolerance33.20000000000000333.21.58E-1617passed
/cells/0/box/outliers/0absentNonepassed
/cells/0/box/extremes/0absentNonepassed
/cells/0/missingexact440.00E+2passed
/cells/0/groupexactpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

458  positions_all <- which(valid_score)
459  ex1 <- explore_cell(0L, x, positions_all, 0.95, desc_index_h, "haverage")
460  ex1[["/cells/0/missing"]] <- exact(as.integer(total - n), "user- and system-missing scores")
461  ex1[["/cells/0/group"]] <- exact("", "no factor")
462  emit("STAT-DESC-EXPLORE-001", "Explore one dependent: descriptives, M-estimators, extremes, normality, box",
463       "nominal", "tabelz.analysis.report_explore:compute_explore", "exploreMenuItem", "EXPLORE",
464       list(dependents = list("score"), confidence_level = 0.95, quartiles = "haverage",
465            metadata = metadata),
466       ex1,
467       c("R: t.test interval; SPSS trimmed mean formula (equals mean(trim = 0.05) here); e1071 type-2 G1/G2 with closed-form SEs; quantile(type = 6) quartiles and 1.5/3 IQR fences; shapiro.test; nortest::lillie.test; MASS::huber(k = 1.339) and the documented fixed-MAD weighted-mean iteration for the Tukey, Hampel and Andrews weights (tolerance 1e-6 for the iterative estimators).",
468         paste0(manual, ":35-46: trimmed mean with fractional boundary weight, whiskers within 1.5 IQR, extremes beyond 3 IQR, fixed 1.4826 MAD scale, fitted-normal K-S approximation."),
469         "Extreme case numbers are 1-based row positions in the dataset before exclusions."),
470       "R stats, MASS, nortest and e1071", explore_cite)

Whole script generators/r/descriptives_reports.R


Explore by factor with Tukey's hinges, a 90% interval and exclusions

Case STAT-DESC-EXPLORE-002 · status: passed

Reference

Evidence type
R
Source
R stats, MASS, nortest and e1071
Citation
R stats::t.test, shapiro.test, quantile (type 6), fivenum; nortest::lillie.test (Dallal-Wilkinson/Stephens); MASS::huber; e1071 type-2 moments
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/0/nexact880.00E+2passed
/cells/0/descriptives/0/1tolerance17.44999999999999917.452.89E-1616passed
/cells/0/descriptives/0/2tolerance0.827431481841763740.82743148184176381.10E-1615passed
/cells/0/descriptives/1/1tolerance17.60555555555555417.6055555555555542.92E-1616passed
/cells/0/descriptives/2/1tolerance18.05000000000000118.052.89E-1616passed
/cells/0/descriptives/3/1tolerance5.47714285714285695.4771428571428591.75E-1515passed
/cells/0/descriptives/4/1tolerance2.34032964711017932.34032964711017984.67E-1615passed
/cells/0/descriptives/5/1tolerance12.512.50.00E+1passed
/cells/0/descriptives/6/1tolerance19.60000000000000119.64.21E-1616passed
/cells/0/descriptives/7/1tolerance7.10000000000000147.1000000000000012.11E-1717passed
/cells/0/descriptives/13/1tolerance-1.5449316965811812-1.54493169658118066.31E-1615passed
/cells/0/descriptives/13/2tolerance0.752101433090354930.75210143309035491.64E-1817passed
/cells/0/descriptives/14/1tolerance2.49373703532269622.493737035322696.22E-1514passed
/cells/0/descriptives/14/2tolerance1.48088047897424621.48088047897424623.93E-1716passed
/cells/0/mean_lowertolerance15.88236601732467315.8823660173246741.35E-1516passed
/cells/0/mean_uppertolerance19.01763398267532819.0176339826753243.77E-1515passed
/cells/0/normality/[test=Shapiro-Wilk]/statistictolerance0.85603889123347010.85603889123347043.34E-1615passed
/cells/0/normality/[test=Shapiro-Wilk]/p_valuetail_probability0.109607307670804440.109607307670805298.47E-1614passed
/cells/0/normality/[test=Kolmogorov-Smirnov (Lilliefors)]/statistictolerance0.207464875496917940.207464875496917944.28E-1816passed
/cells/0/normality/[test=Kolmogorov-Smirnov (Lilliefors)]/p_valuetail_probability0.39436206041879540.39436206041879541.52E-1817passed
/cells/0/m_estimates/0/valuetolerance17.84518263714323217.845182637448743.06E-1010passed
/cells/0/m_estimates/1/valuetolerance17.97897681427340917.9789768175770423.30E-99passed
/cells/0/m_estimates/2/valuetolerance17.76107685714286117.761076857302711.60E-1011passed
/cells/0/m_estimates/3/valuetolerance17.97844544317088417.9784454478956064.72E-99passed
/cells/0/extremes/0/valuetolerance19.60000000000000119.64.21E-1616passed
/cells/0/extremes/0/caseexact12120.00E+2passed
/cells/0/extremes/5/valuetolerance12.512.50.00E+1passed
/cells/0/extremes/5/caseexact110.00E+2passed
/cells/0/extremes/1/valuetolerance19.39999999999999919.44.21E-1616passed
/cells/0/extremes/1/caseexact11110.00E+2passed
/cells/0/extremes/6/valuetolerance15.915.93.55E-1616passed
/cells/0/extremes/6/caseexact440.00E+2passed
/cells/0/extremes/2/valuetolerance18.89999999999999918.94.21E-1616passed
/cells/0/extremes/2/caseexact10100.00E+2passed
/cells/0/extremes/7/valuetolerance17.19999999999999917.22.89E-1616passed
/cells/0/extremes/7/caseexact770.00E+2passed
/cells/0/extremes/3/valuetolerance18.30000000000000118.32.89E-1616passed
/cells/0/extremes/3/caseexact990.00E+2passed
/cells/0/extremes/8/valuetolerance17.80000000000000117.82.89E-1616passed
/cells/0/extremes/8/caseexact880.00E+2passed
/cells/0/extremes/4/valuetolerance17.80000000000000117.82.89E-1616passed
/cells/0/extremes/4/caseexact880.00E+2passed
/cells/0/extremes/9/valuetolerance18.30000000000000118.32.89E-1616passed
/cells/0/extremes/9/caseexact990.00E+2passed
/cells/0/descriptives/8/1tolerance16.22500000000000116.2254.21E-1616passed
/cells/0/descriptives/9/1tolerance19.27499999999999919.2754.21E-1616passed
/cells/0/descriptives/10/1tolerance16.55000000000000116.552.89E-1616passed
/cells/0/descriptives/11/1tolerance19.14999999999999919.154.21E-1616passed
/cells/0/descriptives/12/1tolerance2.59999999999999792.5999999999999983.16E-1716passed
/cells/0/box/q1tolerance16.55000000000000116.552.89E-1616passed
/cells/0/box/q3tolerance19.14999999999999919.154.21E-1616passed
/cells/0/box/mediantolerance18.05000000000000118.052.89E-1616passed
/cells/0/box/whisker_lowtolerance15.915.93.55E-1616passed
/cells/0/box/whisker_hightolerance19.60000000000000119.64.21E-1616passed
/cells/0/box/outliers/1absentNonepassed
/cells/0/box/extremes/0absentNonepassed
/cells/0/box/outliers/0/0exact110.00E+2passed
/cells/0/box/outliers/0/1tolerance12.512.50.00E+1passed
/cells/0/groupexactAApassed
/cells/0/missingexact220.00E+2passed
/cells/1/nexact10100.00E+2passed
/cells/1/descriptives/0/1tolerance23.8223.822.84E-1616passed
/cells/1/descriptives/0/2tolerance1.54724127257372351.54724127257372354.16E-1716passed
/cells/1/descriptives/1/1tolerance23.77222222222222223.7722222222222221.43E-1617passed
/cells/1/descriptives/2/1tolerance23.10000000000000123.14.21E-1616passed
/cells/1/descriptives/3/1tolerance23.93955555555555723.9395555555555543.08E-1515passed
/cells/1/descriptives/4/1tolerance4.892806511150384.892806511150381.11E-1717passed
/cells/1/descriptives/5/1tolerance15.30000000000000115.32.89E-1616passed
/cells/1/descriptives/6/1tolerance33.20000000000000333.21.58E-1617passed
/cells/1/descriptives/7/1tolerance17.90000000000000217.9000000000000021.32E-1617passed
/cells/1/descriptives/13/1tolerance0.318279858734953010.3182798587349533.94E-1816passed
/cells/1/descriptives/13/2tolerance0.687042918621516670.68704291862151673.92E-1817passed
/cells/1/descriptives/14/1tolerance0.935001900233118910.93500190023311893.36E-2019passed
/cells/1/descriptives/14/2tolerance1.33424876998998211.3342487699899822.57E-1817passed
/cells/1/mean_lowertolerance20.98373201330561220.9837320133056124.40E-1616passed
/cells/1/mean_uppertolerance26.65626798669438826.6562679866943881.28E-1617passed
/cells/1/normality/[test=Shapiro-Wilk]/statistictolerance0.976868181439327850.97686818143932793.97E-1817passed
/cells/1/normality/[test=Shapiro-Wilk]/p_valuetail_probability0.946233496097867110.94623349609786674.44E-1615passed
/cells/1/normality/[test=Kolmogorov-Smirnov (Lilliefors)]/statistictolerance0.13420331676258440.13420331676258441.46E-1816passed
/cells/1/normality/[test=Kolmogorov-Smirnov (Lilliefors)]/p_valuetail_probability0.876990970388383580.87699097038838363.87E-1817passed
/cells/1/m_estimates/0/valuetolerance23.68064047999817823.6806404712303058.77E-99passed
/cells/1/m_estimates/1/valuetolerance23.57264400957005223.572644002827336.74E-99passed
/cells/1/m_estimates/2/valuetolerance23.71249999999997723.7124999974536622.55E-99passed
/cells/1/m_estimates/3/valuetolerance23.57442169980117823.5744216935660456.24E-99passed
/cells/1/extremes/0/valuetolerance33.20000000000000333.21.58E-1617passed
/cells/1/extremes/0/caseexact23230.00E+2passed
/cells/1/extremes/5/valuetolerance15.30000000000000115.32.89E-1616passed
/cells/1/extremes/5/caseexact24240.00E+2passed
/cells/1/extremes/1/valuetolerance28.69999999999999928.72.89E-1616passed
/cells/1/extremes/1/caseexact22220.00E+2passed
/cells/1/extremes/6/valuetolerance20.10000000000000120.14.21E-1616passed
/cells/1/extremes/6/caseexact13130.00E+2passed
/cells/1/extremes/2/valuetolerance26.30000000000000126.32.89E-1616passed
/cells/1/extremes/2/caseexact20200.00E+2passed
/cells/1/extremes/7/valuetolerance21.521.50.00E+1passed
/cells/1/extremes/7/caseexact15150.00E+2passed
/cells/1/extremes/3/valuetolerance24.89999999999999924.94.21E-1616passed
/cells/1/extremes/3/caseexact19190.00E+2passed
/cells/1/extremes/8/valuetolerance2222.00.00E+2passed
/cells/1/extremes/8/caseexact16160.00E+2passed
/cells/1/extremes/4/valuetolerance23.39999999999999923.44.21E-1616passed
/cells/1/extremes/4/caseexact18180.00E+2passed
/cells/1/extremes/9/valuetolerance22.80000000000000122.82.89E-1616passed
/cells/1/extremes/9/caseexact17170.00E+2passed
/cells/1/descriptives/8/1tolerance21.14999999999999921.154.21E-1616passed
/cells/1/descriptives/9/1tolerance26.90000000000000226.93.42E-1515passed
/cells/1/descriptives/10/1tolerance21.521.50.00E+1passed
/cells/1/descriptives/11/1tolerance26.30000000000000126.32.89E-1616passed
/cells/1/descriptives/12/1tolerance4.80000000000000074.8000000000000011.05E-1717passed
/cells/1/box/q1tolerance21.521.50.00E+1passed
/cells/1/box/q3tolerance26.30000000000000126.32.89E-1616passed
/cells/1/box/mediantolerance23.10000000000000123.14.21E-1616passed
/cells/1/box/whisker_lowtolerance15.30000000000000115.32.89E-1616passed
/cells/1/box/whisker_hightolerance33.20000000000000333.21.58E-1617passed
/cells/1/box/outliers/0absentNonepassed
/cells/1/box/extremes/0absentNonepassed
/cells/1/groupexactBBpassed
/cells/1/missingexact220.00E+2passed
/cells/2absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

482  ex2[["/cells/2"]] <- absent("two levels give two cells")
483  emit("STAT-DESC-EXPLORE-002", "Explore by factor with Tukey's hinges, a 90% interval and exclusions",
484       "axis_coverage", "tabelz.analysis.report_explore:compute_explore", "exploreMenuItem", "EXPLORE",
485       list(dependents = list("score"), factor = "group", confidence_level = 0.90,
486            quartiles = "tukey_hinges", metadata = metadata),
487       ex2,
488       c("R: per level of group after excluding rows 2 and 5 (1-based); fivenum hinges give the box, fences and hinge spread; the type-6 quartiles are reported beside them; t.test at 0.90.",
489         paste0(manual, ":35-41."),
490         "Case numbers in extremes and outliers are positions before exclusion (#131)."),
491       "R stats, MASS, nortest and e1071", explore_cite, excluded = as.list(excluded2))

Whole script generators/r/descriptives_reports.R


Frequencies with labels, user-missing codes and type-6 percentiles

Case STAT-DESC-FREQUENCIES-001 · status: passed

Reference

Evidence type
R
Source
R base table and stats::quantile type 6; e1071 type-2 moments
Citation
R stats::quantile documentation (type 6, Weibull); e1071::skewness and kurtosis documentation (type 2, SPSS G1/G2)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/tables/0/values/[value=1]/counttolerance33.00.00E+2passed
/tables/0/values/[value=1]/percenttolerance12.512.50.00E+1passed
/tables/0/values/[value=1]/valid_percenttolerance14.28571428571428614.2857142857142864.76E-1616passed
/tables/0/values/[value=1]/cumulative_percenttolerance14.28571428571428614.2857142857142864.76E-1616passed
/tables/0/values/[value=2]/counttolerance77.00.00E+2passed
/tables/0/values/[value=2]/percenttolerance29.16666666666666829.1666666666666681.49E-1617passed
/tables/0/values/[value=2]/valid_percenttolerance33.33333333333333633.3333333333333362.98E-1617passed
/tables/0/values/[value=2]/cumulative_percenttolerance47.6190476190476247.619047619047624.01E-1617passed
/tables/0/values/[value=3]/counttolerance77.00.00E+2passed
/tables/0/values/[value=3]/percenttolerance29.16666666666666829.1666666666666681.49E-1617passed
/tables/0/values/[value=3]/valid_percenttolerance33.33333333333333633.3333333333333362.98E-1617passed
/tables/0/values/[value=3]/cumulative_percenttolerance80.95238095238094980.952380952380952.58E-1819passed
/tables/0/values/[value=4]/counttolerance44.00.00E+2passed
/tables/0/values/[value=4]/percenttolerance16.66666666666666816.6666666666666681.49E-1617passed
/tables/0/values/[value=4]/valid_percenttolerance19.04761904761904719.0476190476190474.50E-1616passed
/tables/0/values/[value=4]/cumulative_percenttolerance100100.00.00E+2passed
/tables/0/values/[value=5]/counttolerance00.00.00E+2passed
/tables/0/values/[value=5]/percenttolerance00.00.00E+2passed
/tables/0/values/[value=5]/valid_percenttolerance00.00.00E+2passed
/tables/0/values/[value=5]/cumulative_percenttolerance100100.00.00E+2passed
/tables/0/values/[value=5]/labelexactExcellentExcellentpassed
/tables/0/values/[value=9]/counttolerance22.00.00E+2passed
/tables/0/values/[value=9]/missingexactuser-missinguser-missingpassed
/tables/0/values/[value=9]/percenttolerance8.33333333333333398.3333333333333342.55E-1717passed
/tables/0/values/[missing=system-missing]/counttolerance11.00.00E+2passed
/tables/0/validtolerance2121.00.00E+2passed
/tables/0/missingtolerance33.00.00E+2passed
/tables/1/validtolerance2020.00.00E+2passed
/tables/1/missingtolerance44.00.00E+2passed
/tables/1/values/[value=12.5]/valid_percenttolerance55.00.00E+2passed
/tables/1/statistics/0/1tolerance2020.00.00E+2passed
/tables/1/statistics/1/1tolerance20.41499999999999920.4151.47E-1617passed
/tables/1/statistics/2/1tolerance1.14255657466743821.14255657466743827.27E-1817passed
/tables/1/statistics/3/1tolerance19.519.50.00E+1passed
/tables/1/statistics/4/1tolerance12.512.50.00E+1passed
/tables/1/statistics/5/1tolerance5.10966833819141285.1096683381914134.91E-1717passed
/tables/1/statistics/6/1tolerance26.10871052631579326.1087105263157933.24E-1717passed
/tables/1/statistics/7/1tolerance0.830239578983928080.83023957898392832.26E-1615passed
/tables/1/statistics/8/1tolerance0.512103336707757430.51210333670775741.69E-1817passed
/tables/1/statistics/9/1tolerance0.721872896982482360.72187289698248246.37E-1918passed
/tables/1/statistics/10/1tolerance0.992383612541845040.9923836125418456.55E-1918passed
/tables/1/statistics/11/1tolerance20.70000000000000320.7000000000000031.58E-1617passed
/tables/1/statistics/12/1tolerance12.512.50.00E+1passed
/tables/1/statistics/13/1tolerance33.20000000000000333.21.58E-1617passed
/tables/1/statistics/14/1tolerance408.30000000000001408.31.37E-1517passed
/tables/1/percentiles/0/0exact1010passed
/tables/1/percentiles/0/1tolerance14.21999999999999914.221.64E-1515passed
/tables/1/percentiles/1/0exact2525passed
/tables/1/percentiles/1/1tolerance16.59999999999999816.5999999999999981.32E-1617passed
/tables/1/percentiles/2/0exact5050passed
/tables/1/percentiles/2/1tolerance19.519.50.00E+1passed
/tables/1/percentiles/3/0exact7575passed
/tables/1/percentiles/3/1tolerance23.24999999999999623.254.00E-1515passed
/tables/1/percentiles/4/0exact9090passed
/tables/1/percentiles/4/1tolerance28.46000000000000828.4599999999999971.07E-1415passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

205  emit("STAT-DESC-FREQUENCIES-001", "Frequencies with labels, user-missing codes and type-6 percentiles",
206       "nominal", "tabelz.analysis.report_listings:compute_frequencies", "frequenciesMenuItem",
207       "FREQUENCIES",
208       list(columns = list("grade", "score"), quartiles = TRUE, percentiles = list(10, 90),
209            statistics = TRUE, user_missing_valid = FALSE, metadata = metadata),
210       c(grade_outputs, score_outputs),
211       c("R: table() over valid codes, labelled code 5 listed with count 0; percent of all rows, valid percent and cumulative valid percent by hand; quantile(type = 6).",
212         paste0(manual, ":28-33: percent of all rows versus valid percent, user-missing listed separately, percentile at (n+1)p."),
213         "Statistics order follows the engine's statistics list; skewness and kurtosis are the SPSS G1 and G2 (e1071 type 2)."),
214       freq_source, freq_cite)

Whole script generators/r/descriptives_reports.R


Frequencies treating user-missing values as valid

Case STAT-DESC-FREQUENCIES-002 · status: passed

Reference

Evidence type
R
Source
R base table and stats::quantile type 6; e1071 type-2 moments
Citation
R stats::quantile documentation (type 6, Weibull); e1071::skewness and kurtosis documentation (type 2, SPSS G1/G2)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/tables/0/values/[value=9]/valid_percenttolerance8.6956521739130438.6956521739130431.49E-1717passed
/tables/0/values/[value=9]/missingexactpassed
/tables/0/validtolerance2323.00.00E+2passed
/tables/0/missingtolerance11.00.00E+2passed
/tables/1/validtolerance2222.00.00E+2passed
/tables/1/values/[value=999]/counttolerance22.00.00E+2passed
/tables/1/statistics/1/1tolerance109.37727272727273109.377272727272749.42E-1516passed
/tables/1/statistics/13/1tolerance999999.00.00E+2passed
/tables/1/statistics/5/1tolerance287.98470351543756287.98470351543755.79E-1415passed
/tables/1/percentiles/0/1tolerance19.85000000000000119.854.21E-1616passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

217  x2 <- d$score[!is.na(d$score)]
218  stopifnot(length(x2) == 22)
219  grade_valid2 <- d$grade[!is.na(d$grade)]
220  um_outputs <- list(
221    "/tables/0/values/[value=9]/valid_percent" = tol(100 * sum(grade_valid2 == 9) / length(grade_valid2), "code 9 valid percent when user-missing counts as valid"),
222    "/tables/0/values/[value=9]/missing" = exact("", "Code 9 is a valid category here."),
223    "/tables/0/valid" = tol(length(grade_valid2), "valid grades including code 9"),
224    "/tables/0/missing" = tol(sum(is.na(d$grade)), "only the system-missing grade is missing"),
225    "/tables/1/valid" = tol(length(x2), "valid scores including 999"),
226    "/tables/1/values/[value=999]/count" = tol(sum(x2 == 999), "999 counted as a valid score value")
227  )
228  um_outputs[[stat_path(1, "Mean")]] <- tol(mean(x2), "mean including the two 999 values")
229  um_outputs[[stat_path(1, "Maximum")]] <- tol(max(x2), "maximum including 999")
230  um_outputs[[stat_path(1, "Std. deviation")]] <- tol(sd(x2), "SD including 999")
231  um_outputs[["/tables/1/percentiles/0/1"]] <- tol(q6(x2, 0.5), "median (quartiles on) including 999")
232  emit("STAT-DESC-FREQUENCIES-002", "Frequencies treating user-missing values as valid",
233       "axis_coverage", "tabelz.analysis.report_listings:compute_frequencies", "frequenciesMenuItem",
234       "FREQUENCIES",
235       list(columns = list("grade", "score"), quartiles = FALSE, percentiles = list(50),
236            statistics = TRUE, user_missing_valid = TRUE, metadata = metadata),
237       um_outputs,
238       c("R: the same tabulation with the user-missing codes 9 and 999 kept as valid values.",
239         paste0(manual, ":28-33.")),
240       freq_source, freq_cite)

Whole script generators/r/descriptives_reports.R


Frequencies cut points after row exclusions, no statistics

Case STAT-DESC-FREQUENCIES-003 · status: passed

Reference

Evidence type
R
Source
R base table and stats::quantile type 6; e1071 type-2 moments
Citation
R stats::quantile documentation (type 6, Weibull); e1071::skewness and kurtosis documentation (type 2, SPSS G1/G2)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/tables/0/validtolerance1818.00.00E+2passed
/tables/0/missingtolerance44.00.00E+2passed
/tables/0/statistics/0absentNonepassed
/tables/0/percentiles/0/0exact2525passed
/tables/0/percentiles/0/1tolerance17.65000000000000217.653.42E-1515passed
/tables/0/percentiles/1/1tolerance19.85000000000000119.854.21E-1616passed
/tables/0/percentiles/2/1tolerance23.77499999999999923.7754.21E-1616passed
/tables/0/percentiles/3absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

243  keep <- valid_score & !(seq_len(nrow(d)) %in% c(1, 5))
244  x3 <- d$score[keep]
245  stopifnot(length(x3) == 18)
246  cut_outputs <- list(
247    "/tables/0/valid" = tol(length(x3), "valid scores after excluding two rows"),
248    "/tables/0/missing" = tol(nrow(d) - 2 - length(x3), "missing among the 22 remaining rows"),
249    "/tables/0/statistics/0" = absent("Statistics were not requested."),
250    "/tables/0/percentiles/0/0" = exact("25", "first cut point of four"),
251    "/tables/0/percentiles/0/1" = tol(q6(x3, 0.25), "25th percentile type 6 after exclusions"),
252    "/tables/0/percentiles/1/1" = tol(q6(x3, 0.5), "50th percentile type 6 after exclusions"),
253    "/tables/0/percentiles/2/1" = tol(q6(x3, 0.75), "75th percentile type 6 after exclusions"),
254    "/tables/0/percentiles/3" = absent("Four cut points give three percentiles.")
255  )
256  emit("STAT-DESC-FREQUENCIES-003", "Frequencies cut points after row exclusions, no statistics",
257       "axis_coverage", "tabelz.analysis.report_listings:compute_frequencies", "frequenciesMenuItem",
258       "FREQUENCIES",
259       list(columns = list("score"), quartiles = FALSE, percentiles = list(), cut_points = 4L,
260            statistics = FALSE, user_missing_valid = FALSE, metadata = metadata),
261       cut_outputs,
262       c("R: quantile(type = 6) at 25, 50 and 75 on the 18 valid scores of the 22 non-excluded rows.",
263         paste0(manual, ":28-33.")),
264       freq_source, freq_cite, excluded = list(0L, 4L))

Whole script generators/r/descriptives_reports.R


Multiple-response crosstab against a column on the cases base

Case STAT-DESC-MRCROSS-001 · status: passed

Reference

Evidence type
R
Source
R base table() and aggregate() on a long (row, category) layout
Citation
R base::table documentation; manual chapter 80 multiple-response definitions
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_multiple_response.R · sha256 ee4f1ee9a38ca2f3
Dataset
reference_data/datasets/DESC_MR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_analysedexact14140.00E+2passed
/valid_casesexact10100.00E+2passed
/left_out_casesexact440.00E+2passed
/base_totalexact10100.00E+2passed
/row_totals/s1exact770.00E+2passed
/column_totals/Northexact550.00E+2passed
/cells/6absentNonepassed
/cells/0/rowexacts1s1passed
/cells/0/columnexactNorthNorthpassed
/cells/0/countexact220.00E+2passed
/cells/0/row_percenttolerance28.57142857142857328.5714285714285734.88E-1717passed
/cells/0/column_percenttolerance4040.00.00E+2passed
/cells/0/total_percenttolerance2020.00.00E+2passed
/cells/1/rowexacts1s1passed
/cells/1/columnexactSouthSouthpassed
/cells/1/countexact550.00E+2passed
/cells/1/row_percenttolerance71.42857142857143171.428571428571433.98E-1617passed
/cells/1/column_percenttolerance100100.00.00E+2passed
/cells/1/total_percenttolerance5050.00.00E+2passed
/cells/2/rowexacts2s2passed
/cells/2/columnexactNorthNorthpassed
/cells/2/countexact330.00E+2passed
/cells/2/row_percenttolerance5050.00.00E+2passed
/cells/2/column_percenttolerance6060.00.00E+2passed
/cells/2/total_percenttolerance3030.00.00E+2passed
/cells/3/rowexacts2s2passed
/cells/3/columnexactSouthSouthpassed
/cells/3/countexact330.00E+2passed
/cells/3/row_percenttolerance5050.00.00E+2passed
/cells/3/column_percenttolerance6060.00.00E+2passed
/cells/3/total_percenttolerance3030.00.00E+2passed
/cells/4/rowexacts3s3passed
/cells/4/columnexactNorthNorthpassed
/cells/4/countexact440.00E+2passed
/cells/4/row_percenttolerance8080.00.00E+2passed
/cells/4/column_percenttolerance8080.00.00E+2passed
/cells/4/total_percenttolerance4040.00.00E+2passed
/cells/5/rowexacts3s3passed
/cells/5/columnexactSouthSouthpassed
/cells/5/countexact110.00E+2passed
/cells/5/row_percenttolerance2020.00.00E+2passed
/cells/5/column_percenttolerance2020.00.00E+2passed
/cells/5/total_percenttolerance1010.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

168  emit("STAT-DESC-MRCROSS-001", "Multiple-response crosstab against a column on the cases base",
169       "nominal", "tabelz.analysis.multiple_response:compute_mr_crosstab",
170       "multipleResponseCrosstabMenuItem", "MULTIPLE-RESPONSE-CROSSTAB",
171       list(set_name = "features", response_set = features, by_column = "region",
172            base = "cases"),
173       o3,
174       c("R: (row, dichotomy) pairs joined to the row's region, rows without a region dropped; each cell counts distinct cases; row and column totals count distinct cases; percents divide by those totals.",
175         paste0(manual, ":160-165: on a case base a case counts once for each category pair; denominators match the selected base.")))

Whole script generators/r/descriptives_multiple_response.R


Multiple-response crosstab of two sets on the responses base

Case STAT-DESC-MRCROSS-002 · status: passed

Reference

Evidence type
R
Source
R base table() and aggregate() on a long (row, category) layout
Citation
R base::table documentation; manual chapter 80 multiple-response definitions
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_multiple_response.R · sha256 ee4f1ee9a38ca2f3
Dataset
reference_data/datasets/DESC_MR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valid_casesexact10100.00E+2passed
/base_totalexact29290.00E+2passed
/row_totals/s2exact10100.00E+2passed
/column_totals/Aexact14140.00E+2passed
/cells/0/rowexacts1s1passed
/cells/0/columnexactAApassed
/cells/0/countexact440.00E+2passed
/cells/0/responsesexact440.00E+2passed
/cells/0/casesexact330.00E+2passed
/cells/0/row_percenttolerance36.36363636363636736.363636363636371.34E-1617passed
/cells/0/total_percenttolerance13.79310344827586113.7931034482758612.11E-1616passed
/cells/1/rowexacts1s1passed
/cells/1/columnexactBBpassed
/cells/1/countexact440.00E+2passed
/cells/1/responsesexact440.00E+2passed
/cells/1/casesexact440.00E+2passed
/cells/1/row_percenttolerance36.36363636363636736.363636363636371.34E-1617passed
/cells/1/total_percenttolerance13.79310344827586113.7931034482758612.11E-1616passed
/cells/2/rowexacts1s1passed
/cells/2/columnexactCCpassed
/cells/2/countexact330.00E+2passed
/cells/2/responsesexact330.00E+2passed
/cells/2/casesexact220.00E+2passed
/cells/2/row_percenttolerance27.27272727272727327.2727272727272733.73E-1616passed
/cells/2/total_percenttolerance10.34482758620689710.3448275862068972.03E-1616passed
/cells/3/rowexacts2s2passed
/cells/3/columnexactAApassed
/cells/3/countexact550.00E+2passed
/cells/3/responsesexact550.00E+2passed
/cells/3/casesexact440.00E+2passed
/cells/3/row_percenttolerance5050.00.00E+2passed
/cells/3/total_percenttolerance17.24137931034482917.241379310344831.79E-1616passed
/cells/4/rowexacts2s2passed
/cells/4/columnexactBBpassed
/cells/4/countexact550.00E+2passed
/cells/4/responsesexact550.00E+2passed
/cells/4/casesexact440.00E+2passed
/cells/4/row_percenttolerance5050.00.00E+2passed
/cells/4/total_percenttolerance17.24137931034482917.241379310344831.79E-1616passed
/cells/5/rowexacts2s2passed
/cells/5/columnexactCCpassed
/cells/5/countexact000.00E+2passed
/cells/5/responsesexact000.00E+2passed
/cells/5/casesexact000.00E+2passed
/cells/5/row_percenttolerance00.00.00E+2passed
/cells/5/total_percenttolerance00.00.00E+2passed
/cells/6/rowexacts3s3passed
/cells/6/columnexactAApassed
/cells/6/countexact550.00E+2passed
/cells/6/responsesexact550.00E+2passed
/cells/6/casesexact440.00E+2passed
/cells/6/row_percenttolerance62.562.50.00E+1passed
/cells/6/total_percenttolerance17.24137931034482917.241379310344831.79E-1616passed
/cells/7/rowexacts3s3passed
/cells/7/columnexactBBpassed
/cells/7/countexact220.00E+2passed
/cells/7/responsesexact220.00E+2passed
/cells/7/casesexact220.00E+2passed
/cells/7/row_percenttolerance2525.00.00E+2passed
/cells/7/total_percenttolerance6.89655172413793066.8965517241379315.71E-1818passed
/cells/8/rowexacts3s3passed
/cells/8/columnexactCCpassed
/cells/8/countexact110.00E+2passed
/cells/8/responsesexact110.00E+2passed
/cells/8/casesexact110.00E+2passed
/cells/8/row_percenttolerance12.512.50.00E+1passed
/cells/8/total_percenttolerance3.44827586206896533.44827586206896532.85E-1818passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

205  emit("STAT-DESC-MRCROSS-002", "Multiple-response crosstab of two sets on the responses base",
206       "axis_coverage", "tabelz.analysis.multiple_response:compute_mr_crosstab",
207       "multipleResponseCrosstabMenuItem", "MULTIPLE-RESPONSE-CROSSTAB",
208       list(set_name = "features", response_set = features, by_set_name = "choices",
209            by_set = choices, base = "responses"),
210       o4,
211       c("R: the (row, dichotomy) pairs joined to the (row, code) pairs of the same row; a cell counts the joined records (occurrence pairs, repeated codes multiplying them) and, separately, the distinct cases; totals sum the cells.",
212         paste0(manual, ":160-165: on a response base repetitions multiply the paired response count.")))

Whole script generators/r/descriptives_multiple_response.R


Multiple-dichotomy frequencies: case base, responses and percents

Case STAT-DESC-MRFREQ-001 · status: passed

Reference

Evidence type
R
Source
R base table() and aggregate() on a long (row, category) layout
Citation
R base::table documentation; manual chapter 80 multiple-response definitions
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_multiple_response.R · sha256 ee4f1ee9a38ca2f3
Dataset
reference_data/datasets/DESC_MR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_analysedexact14140.00E+2passed
/valid_casesexact11110.00E+2passed
/cases_without_responseexact330.00E+2passed
/all_missing_rowsexact110.00E+2passed
/total_responsesexact19190.00E+2passed
/categories/3absentNonepassed
/categories/0/categoryexacts1s1passed
/categories/0/responsesexact770.00E+2passed
/categories/0/casesexact770.00E+2passed
/categories/0/percent_of_responsestolerance36.84210526315789736.84210526315793.55E-1617passed
/categories/0/percent_of_casestolerance63.63636363636363363.636363636363631.34E-1617passed
/categories/1/categoryexacts2s2passed
/categories/1/responsesexact770.00E+2passed
/categories/1/casesexact770.00E+2passed
/categories/1/percent_of_responsestolerance36.84210526315789736.84210526315793.55E-1617passed
/categories/1/percent_of_casestolerance63.63636363636363363.636363636363631.34E-1617passed
/categories/2/categoryexacts3s3passed
/categories/2/responsesexact550.00E+2passed
/categories/2/casesexact550.00E+2passed
/categories/2/percent_of_responsestolerance26.31578947368420926.315789473684211.57E-1617passed
/categories/2/percent_of_casestolerance45.45454545454545345.454545454545452.54E-1617passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

106  emit("STAT-DESC-MRFREQ-001", "Multiple-dichotomy frequencies: case base, responses and percents",
107       "nominal", "tabelz.analysis.multiple_response:compute_mr_frequencies",
108       "multipleResponseFrequenciesMenuItem", "MULTIPLE-RESPONSE-FREQUENCIES",
109       list(set_name = "features", response_set = features),
110       o1,
111       c("R: table() of the (row, column) pairs where the column holds the counted value 1; a case is a row with at least one such pair; percent of responses divides by all pairs, percent of cases by the cases.",
112         paste0(manual, ":111-117: the case base consists of rows with at least one counted response; a row answering No to every dichotomy is outside that base; rows with no counted response and rows missing all responses are counted separately.")))

Whole script generators/r/descriptives_multiple_response.R


Multiple-category frequencies: repeated codes count as responses once per case

Case STAT-DESC-MRFREQ-002 · status: passed

Reference

Evidence type
R
Source
R base table() and aggregate() on a long (row, category) layout
Citation
R base::table documentation; manual chapter 80 multiple-response definitions
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_multiple_response.R · sha256 ee4f1ee9a38ca2f3
Dataset
reference_data/datasets/DESC_MR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valid_casesexact13130.00E+2passed
/cases_without_responseexact110.00E+2passed
/all_missing_rowsexact110.00E+2passed
/total_responsesexact24240.00E+2passed
/categories/0/categoryexactAApassed
/categories/0/responsesexact10100.00E+2passed
/categories/0/casesexact880.00E+2passed
/categories/0/percent_of_responsestolerance41.66666666666666441.6666666666666642.98E-1617passed
/categories/0/percent_of_casestolerance61.5384615384615461.538461538461541.01E-1617passed
/categories/1/categoryexactBBpassed
/categories/1/responsesexact880.00E+2passed
/categories/1/casesexact770.00E+2passed
/categories/1/percent_of_responsestolerance33.33333333333333633.3333333333333362.98E-1617passed
/categories/1/percent_of_casestolerance53.84615384615384753.846153846153853.00E-1617passed
/categories/2/categoryexactCCpassed
/categories/2/responsesexact660.00E+2passed
/categories/2/casesexact550.00E+2passed
/categories/2/percent_of_responsestolerance2525.00.00E+2passed
/categories/2/percent_of_casestolerance38.4615384615384638.461538461538461.01E-1617passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

130  emit("STAT-DESC-MRFREQ-002", "Multiple-category frequencies: repeated codes count as responses once per case",
131       "axis_coverage", "tabelz.analysis.multiple_response:compute_mr_frequencies",
132       "multipleResponseFrequenciesMenuItem", "MULTIPLE-RESPONSE-FREQUENCIES",
133       list(set_name = "choices", response_set = choices),
134       o2,
135       c("R: table() of every non-missing code across c1 and c2 (occurrences) and of the distinct (row, code) pairs (cases).",
136         paste0(manual, ":119-158: the worked example distinguishes mentions from technicians; case percentages can total more than 100.")))

Whole script generators/r/descriptives_multiple_response.R


Missing value analysis with the default settings

Case STAT-DESC-MVA-001 · status: passed

Reference

Evidence type
R
Source
Base R (stats) with a hand-written multivariate-normal EM
Citation
Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. JASA 83:1198-1202; Dempster, Laird and Rubin (1977) JRSS B 39:1-38; R stats: quantile (type 7), t.test, p.adjust, pchisq, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_missing_values.R · sha256 1c2458fc190a7995
Dataset
reference_data/datasets/DESC_MVA_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/casesexact24240.00E+2passed
/univariate/0/presentexact19190.00E+2passed
/univariate/0/missingexact550.00E+2passed
/univariate/0/percent_missingtolerance20.83333333333333220.8333333333333321.49E-1617passed
/univariate/0/meantolerance5.84736842105263135.8473684210526329.15E-1615passed
/univariate/0/sdtolerance0.963424690623467470.96342469062346761.13E-1615passed
/univariate/0/low_extremesexact000.00E+2passed
/univariate/0/high_extremesexact000.00E+2passed
/univariate/3/missingexact330.00E+2passed
/univariate/3/percent_missingtolerance12.512.50.00E+1passed
/patterns_totalexact660.00E+2passed
/patterns/0/countexact14140.00E+2passed
/patterns/0/percenttolerance58.33333333333333658.3333333333333362.98E-1617passed
/patterns/0/variables_missingexact000.00E+2passed
/welch_filtered/0/0exactzzpassed
/welch_filtered/0/1tolerance4.16666666666666614.1666666666666662.55E-1717passed
/welch/0/present_countexact15150.00E+2passed
/welch/0/missing_countexact550.00E+2passed
/welch/0/differencetolerance0.346666666666667570.346666666666669341.77E-1514passed
/welch/0/t_valuetolerance0.654874889476169230.65487488947617273.44E-1514passed
/welch/0/dftolerance11.54355010106741711.5435501010674173.02E-1616passed
/welch/0/p_valuetail_probability0.525382781221379050.5253827812213772.00E-1514passed
/welch/0/adjusted_p_valuetail_probability11.00.00E+2passed
/welch/1/present_countexact19190.00E+2passed
/welch/1/missing_countexact440.00E+2passed
/welch/1/differencetolerance0.334210526315789290.33421052631578931.37E-1817passed
/welch/1/t_valuetolerance1.50880735918865641.50880735918865641.71E-1716passed
/welch/1/dftolerance7.87854683910377767.878546839103782.69E-1515passed
/welch/1/p_valuetail_probability0.170357881652794890.170357881652794835.53E-1715passed
/welch/1/adjusted_p_valuetail_probability11.00.00E+2passed
/welch/2/present_countexact15150.00E+2passed
/welch/2/missing_countexact440.00E+2passed
/welch/2/differencetolerance0.376666666666666930.376666666666666048.91E-1614passed
/welch/2/t_valuetolerance0.861069987657915650.86106998765791352.11E-1514passed
/welch/2/dftolerance7.01177186837399987.0117718683740021.77E-1515passed
/welch/2/p_valuetail_probability0.417668880203446390.41766888020344741.00E-1514passed
/welch/2/adjusted_p_valuetail_probability11.00.00E+2passed
/welch/3/present_countexact19190.00E+2passed
/welch/3/missing_countexact440.00E+2passed
/welch/3/differencetolerance0.0618421052631576980.0618421052631572544.44E-1614passed
/welch/3/t_valuetolerance0.215869800489001920.215869800489000371.55E-1514passed
/welch/3/dftolerance5.12146606852280955.1214660685228098.88E-1615passed
/welch/3/p_valuetail_probability0.837424328154707020.83742432815470821.22E-1514passed
/welch/3/adjusted_p_valuetail_probability11.00.00E+2passed
/welch/4/present_countexact16160.00E+2passed
/welch/4/missing_countexact330.00E+2passed
/welch/4/differencetolerance0.293749999999999290.29374999999999935.43E-1917passed
/welch/4/t_valuetolerance0.64338503846391470.64338503846391474.64E-1817passed
/welch/4/dftolerance4.11691086554316944.11691086554316858.59E-1615passed
/welch/4/p_valuetail_probability0.554064914114529520.55406491411452941.15E-1615passed
/welch/4/adjusted_p_valuetail_probability11.00.00E+2passed
/welch/5/present_countexact19190.00E+2passed
/welch/5/missing_countexact110.00E+2passed
/welch/5/statusexactnot computed: fewer than two cases in a groupnot computed: fewer than two cases in a grouppassed
/welch/6/present_countexact20200.00E+2passed
/welch/6/missing_countexact330.00E+2passed
/welch/6/differencetolerance-0.0083333333333333037-0.0083333333333333042.74E-2017passed
/welch/6/t_valuetolerance-0.0259768624588647-0.0259768624588647076.52E-1815passed
/welch/6/dftolerance2.95552895666481462.95552895666481378.56E-1615passed
/welch/6/p_valuetail_probability0.980929941441967430.98092994144196744.56E-1817passed
/welch/6/adjusted_p_valuetail_probability11.00.00E+2passed
/em/statusexactconvergedconvergedpassed
/em/cases_usedexact24240.00E+2passed
/em/means/0tolerance5.73983313473121815.7398308997584072.23E-66passed
/em/means/1tolerance10.64218747357821310.6421901466667622.67E-66passed
/em/means/2tolerance2.60164391199237822.60164383485211877.71E-87passed
/em/covariance/0/0tolerance0.793356541359734990.79335831560715061.77E-65passed
/em/covariance/0/1tolerance1.00897790386203921.00897767702484492.27E-76passed
/em/covariance/1/2tolerance0.643958964341699410.64395919605063772.32E-76passed
/em/correlations/0/2tolerance0.98023139318666430.98023171408566763.21E-76passed
/em/log_likelihoodtolerance-23.746389060077693-23.7463890637310833.65E-99passed
/little/statusexactcomputedcomputedpassed
/little/statistictolerance13.60224719886801713.6022457124837361.49E-66passed
/little/dfexact550.00E+2passed
/little/p_valuetail_probability0.0183435106473492410.0183435216791290741.10E-86passed
/little/patternsexact440.00E+2passed
/crosstabs/0/missingexact220.00E+2passed
/crosstabs/0/presentexact990.00E+2passed
/crosstabs/2/missingexact000.00E+2passed
/crosstabs/2/percent_missingtolerance00.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

262  gx_a_missing <- sum(d$g == "A" & is.na(d$x), na.rm = TRUE)
263  gx_null_missing <- sum(is.na(d$g) & is.na(d$x))
264  outputs[["/crosstabs/0/missing"]] <- exact(as.integer(gx_a_missing), "g = A cases missing x")
265  outputs[["/crosstabs/0/present"]] <- exact(as.integer(sum(d$g == "A" & !is.na(d$x), na.rm = TRUE)), "g = A cases present on x")
266  outputs[["/crosstabs/2/missing"]] <- exact(as.integer(gx_null_missing), "g missing cases missing x")
267  outputs[["/crosstabs/2/percent_missing"]] <- tol(100 * gx_null_missing / sum(is.na(d$g)), "percent of g-missing cases missing x")
268  emit("STAT-DESC-MVA-001", "Missing value analysis with the default settings", "nominal",
269       list(quantitative = list("x", "y", "z"), categorical = list("g")),
270       outputs,
271       list("Univariate: available-case mean and SD; fences from quantile(type = 7) (numpy linear).",
272            "Welch rows are listed per indicator (x, y, g; z is filtered at 4.17% < 5%) and then per other quantitative variable; t.test(var.equal = FALSE), difference present minus missing; p.adjust(method = 'holm') over the six computed tests (g->y has one missing case and is not computed).",
273            "EM is iterated in R to a scaled change of 1e-13 (tolerance 1e-13, 20000 iterations); the product stops at 1e-6, so EM estimates, Little's statistic and the log-likelihood carry relative tolerances 1e-4 and 1e-7.",
274            "Little's test uses the EM maximum-likelihood mean and n-denominator covariance over the patterns of the three quantitative variables (every row has at least one value, so 24 cases enter)."))

Whole script generators/r/descriptives_missing_values.R


Raw p-values, 1% indicator threshold, SPSS-style EM

Case STAT-DESC-MVA-002 · status: passed

Reference

Evidence type
R
Source
Base R (stats) with a hand-written multivariate-normal EM
Citation
Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. JASA 83:1198-1202; Dempster, Laird and Rubin (1977) JRSS B 39:1-38; R stats: quantile (type 7), t.test, p.adjust, pchisq, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_missing_values.R · sha256 1c2458fc190a7995
Dataset
reference_data/datasets/DESC_MVA_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/casesexact24240.00E+2passed
/em/statusexactconvergedconvergedpassed
/em/tolerancetolerance0.00010.00014.79E-2116passed
/em/max_iterationsexact25250.00E+2passed
/em/means/0tolerance5.73983313473121815.7395753804317942.58E-44passed
/little/dfexact550.00E+2passed
/welch/0/present_countexact15150.00E+2passed
/welch/0/missing_countexact550.00E+2passed
/welch/0/differencetolerance0.346666666666667570.346666666666669341.77E-1514passed
/welch/0/t_valuetolerance0.654874889476169230.65487488947617273.44E-1514passed
/welch/0/dftolerance11.54355010106741711.5435501010674173.02E-1616passed
/welch/0/p_valuetail_probability0.525382781221379050.5253827812213772.00E-1514passed
/welch/1/present_countexact19190.00E+2passed
/welch/1/missing_countexact440.00E+2passed
/welch/1/differencetolerance0.334210526315789290.33421052631578931.37E-1817passed
/welch/1/t_valuetolerance1.50880735918865641.50880735918865641.71E-1716passed
/welch/1/dftolerance7.87854683910377767.878546839103782.69E-1515passed
/welch/1/p_valuetail_probability0.170357881652794890.170357881652794835.53E-1715passed
/welch/2/present_countexact15150.00E+2passed
/welch/2/missing_countexact440.00E+2passed
/welch/2/differencetolerance0.376666666666666930.376666666666666048.91E-1614passed
/welch/2/t_valuetolerance0.861069987657915650.86106998765791352.11E-1514passed
/welch/2/dftolerance7.01177186837399987.0117718683740021.77E-1515passed
/welch/2/p_valuetail_probability0.417668880203446390.41766888020344741.00E-1514passed
/welch/3/present_countexact19190.00E+2passed
/welch/3/missing_countexact440.00E+2passed
/welch/3/differencetolerance0.0618421052631576980.0618421052631572544.44E-1614passed
/welch/3/t_valuetolerance0.215869800489001920.215869800489000371.55E-1514passed
/welch/3/dftolerance5.12146606852280955.1214660685228098.88E-1615passed
/welch/3/p_valuetail_probability0.837424328154707020.83742432815470821.22E-1514passed
/welch/4/present_countexact19190.00E+2passed
/welch/4/missing_countexact000.00E+2passed
/welch/4/statusexactnot computed: fewer than two cases in a groupnot computed: fewer than two cases in a grouppassed
/welch/5/present_countexact19190.00E+2passed
/welch/5/missing_countexact110.00E+2passed
/welch/5/statusexactnot computed: fewer than two cases in a groupnot computed: fewer than two cases in a grouppassed
/welch/6/present_countexact16160.00E+2passed
/welch/6/missing_countexact330.00E+2passed
/welch/6/differencetolerance0.293749999999999290.29374999999999935.43E-1917passed
/welch/6/t_valuetolerance0.64338503846391470.64338503846391474.64E-1817passed
/welch/6/dftolerance4.11691086554316944.11691086554316858.59E-1615passed
/welch/6/p_valuetail_probability0.554064914114529520.55406491411452941.15E-1615passed
/welch/7/present_countexact19190.00E+2passed
/welch/7/missing_countexact110.00E+2passed
/welch/7/statusexactnot computed: fewer than two cases in a groupnot computed: fewer than two cases in a grouppassed
/welch/8/present_countexact20200.00E+2passed
/welch/8/missing_countexact330.00E+2passed
/welch/8/differencetolerance-0.0083333333333333037-0.0083333333333333042.74E-2017passed
/welch/8/t_valuetolerance-0.0259768624588647-0.0259768624588647076.52E-1815passed
/welch/8/dftolerance2.95552895666481462.95552895666481378.56E-1615passed
/welch/8/p_valuetail_probability0.980929941441967430.98092994144196744.56E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

277  w2 <- welch_rows(d, quant, "g", 1.0, "none")
278  stopifnot(identical(w2$indicators, c("x", "y", "z", "g")), length(w2$filtered) == 0)
279  not_computed <- which(vapply(w2$rows, function(r) r$status != "", TRUE))
280  stopifnot(identical(not_computed, c(5L, 6L, 8L)))  # z->x, z->y and g->y have under two missing cases
281  outputs <- c(
282    list("/cases" = exact(nrow(d), "24 rows"),
283         "/em/status" = exact("converged", "EM converges under the SPSS-style preset (1e-4, 25 iterations)"),
284         "/em/tolerance" = tol(1e-4, "SPSS-style tolerance"),
285         "/em/max_iterations" = exact(25L, "SPSS-style iteration cap"),
286         "/em/means/0" = tol(em$mu[1], "EM mean of x (SPSS-style preset)", "1e-3"),
287         "/little/df" = exact(as.integer(lt$df), "Little df does not depend on the EM tolerance")),
288    welch_outputs(w2, seq_along(w2$rows))
289  )
290  emit("STAT-DESC-MVA-002", "Raw p-values, 1% indicator threshold, SPSS-style EM", "axis_coverage",
291       list(quantitative = list("x", "y", "z"), categorical = list("g"), min_percent = 1.0,
292            adjustment = "none", em_preset = "spss"),
293       outputs,
294       list("With the threshold at 1% the indicator z (one missing case) enters: its two comparisons have under two cases in the missing group and are reported as not computed, as is g->y; the nine rows are x->y, x->z, y->x, y->z, z->x, z->y, g->x, g->y, g->z.",
295            "adjustment = none leaves the raw Welch p-values; no adjusted value is compared.",
296            "The SPSS-style preset stops EM at a scaled change of 1e-4, so only the mean of x is compared, at relative tolerance 1e-3; the R EM is run to 1e-13."))

Whole script generators/r/descriptives_missing_values.R


EM to a custom tolerance of 1e-10

Case STAT-DESC-MVA-003 · status: passed

Reference

Evidence type
R
Source
Base R (stats) with a hand-written multivariate-normal EM
Citation
Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. JASA 83:1198-1202; Dempster, Laird and Rubin (1977) JRSS B 39:1-38; R stats: quantile (type 7), t.test, p.adjust, pchisq, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_missing_values.R · sha256 1c2458fc190a7995
Dataset
reference_data/datasets/DESC_MVA_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/casesexact24240.00E+2passed
/em/tolerancetolerance1e-101e-103.64E-27passed
/em/max_iterationsexact500050000.00E+2passed
/em/statusexactconvergedconvergedpassed
/em/cases_usedexact24240.00E+2passed
/em/means/0tolerance5.73983313473121815.7398331345058932.25E-1010passed
/em/means/1tolerance10.64218747357821310.6421874738476972.69E-1010passed
/em/means/2tolerance2.60164391199237822.60164391198467.78E-1211passed
/em/covariance/0/0tolerance0.793356541359734990.79335654153858571.79E-109passed
/em/covariance/0/1tolerance1.00897790386203921.00897790383916692.29E-1110passed
/em/covariance/1/2tolerance0.643958964341699410.64395896436506212.34E-1110passed
/em/correlations/0/2tolerance0.98023139318666430.9802313932190443.24E-1110passed
/em/log_likelihoodtolerance-23.746389060077693-23.746389060077693.76E-1515passed
/little/statusexactcomputedcomputedpassed
/little/statistictolerance13.60224719886801713.6022471987241521.44E-1010passed
/little/dfexact550.00E+2passed
/little/p_valuetail_probability0.0183435106473492410.0183435106484169941.07E-1210passed
/little/patternsexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

299  outputs <- c(
300    list("/cases" = exact(nrow(d), "24 rows"),
301         "/em/tolerance" = tol(1e-10, "custom tolerance"),
302         "/em/max_iterations" = exact(5000L, "custom iteration cap")),
303    em_outputs(em, lt, "1e-7", "1e-6")
304  )
305  emit("STAT-DESC-MVA-003", "EM to a custom tolerance of 1e-10", "axis_coverage",
306       list(quantitative = list("x", "y", "z"), categorical = list("g"), em_preset = "custom",
307            em_tolerance = 1e-10, em_max_iterations = 5000L),
308       outputs,
309       list("Both EM runs stop near the same maximum (R at a scaled change of 1e-13, the product at 1e-10), so the EM estimates and Little's statistic are compared at relative tolerance 1e-7 and the p-value at 1e-6.",
310            "The convergence check compares the largest scaled parameter change and the observed log-likelihood change with the tolerance."))

Whole script generators/r/descriptives_missing_values.R


Excluded rows leave the analysis population

Case STAT-DESC-MVA-004 · status: passed

Reference

Evidence type
R
Source
Base R (stats) with a hand-written multivariate-normal EM
Citation
Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. JASA 83:1198-1202; Dempster, Laird and Rubin (1977) JRSS B 39:1-38; R stats: quantile (type 7), t.test, p.adjust, pchisq, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_missing_values.R · sha256 1c2458fc190a7995
Dataset
reference_data/datasets/DESC_MVA_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/casesexact22220.00E+2passed
/excluded_countexact220.00E+2passed
/univariate/0/presentexact17170.00E+2passed
/univariate/0/meantolerance5.77058823529411765.77058823529411858.62E-1615passed
/univariate/0/high_extremesexact000.00E+2passed
/patterns_totalexact660.00E+2passed
/patterns/0/countexact12120.00E+2passed
/welch/0/present_countexact13130.00E+2passed
/welch/0/missing_countexact550.00E+2passed
/welch/0/differencetolerance0.179999999999999720.18000000000000151.77E-1514passed
/welch/0/t_valuetolerance0.344346343953780780.344346343953784163.38E-1514passed
/welch/0/dftolerance10.88667023131080710.886670231310812.24E-1515passed
/welch/0/p_valuetail_probability0.737145423202670620.7371454232026682.66E-1514passed
/welch/0/adjusted_p_valuetail_probability11.00.00E+2passed
/welch/1/present_countexact17170.00E+2passed
/welch/1/missing_countexact440.00E+2passed
/welch/1/differencetolerance0.28529411764705870.285294117647059144.42E-1614passed
/welch/1/t_valuetolerance1.30137854918445361.30137854918445562.01E-1514passed
/welch/1/dftolerance7.55118583594585867.55118583594585958.59E-1615passed
/welch/1/p_valuetail_probability0.231426994325301920.231426994325301424.98E-1614passed
/welch/1/adjusted_p_valuetail_probability11.00.00E+2passed
/em/statusexactconvergedconvergedpassed
/em/cases_usedexact22220.00E+2passed
/em/means/0tolerance5.68144950512342025.6814470146175662.49E-66passed
/em/means/1tolerance10.51661695152159310.5166199441797782.99E-66passed
/em/means/2tolerance2.55707858145128552.55707848806388869.34E-87passed
/em/covariance/0/0tolerance0.619987611345976570.61998899528983741.38E-65passed
/em/covariance/0/1tolerance0.797996504192973080.79799582952092156.75E-76passed
/em/covariance/1/2tolerance0.520627338443350850.52062754092291642.02E-76passed
/em/correlations/0/2tolerance0.97416649484752960.97416704397877435.49E-76passed
/em/log_likelihoodtolerance-19.9840296172089-19.9840296207953333.59E-99passed
/little/statusexactcomputedcomputedpassed
/little/statistictolerance13.66419389478232913.6642105835780151.67E-55passed
/little/dfexact550.00E+2passed
/little/p_valuetail_probability0.0178892570013537180.017889136095754241.21E-75passed
/little/patternsexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

313  kept <- d[-c(1, 16), , drop = FALSE]  # zero-based source rows 0 and 15 are excluded
314  ux4 <- univariate(kept$x)
315  w4 <- welch_rows(kept, quant, "g", 5.0, "holm")
316  stopifnot(identical(w4$indicators, c("x", "y", "g")))
317  pt4 <- pattern_table(kept, c(quant, "g"))
318  em4 <- em_mvn(as.matrix(kept[quant]), 1e-13, 20000)
319  lt4 <- little(em4)
320  outputs <- c(
321    list("/cases" = exact(nrow(kept), "22 rows after excluding two"),
322         "/excluded_count" = exact(2L, "two excluded rows"),
323         "/univariate/0/present" = exact(as.integer(ux4$present), "x present without the excluded rows"),
324         "/univariate/0/mean" = tol(ux4$mean, "x mean without the excluded rows"),
325         "/univariate/0/high_extremes" = exact(as.integer(ux4$high), "x high extremes without row 15 (7.9)"),
326         "/patterns_total" = exact(as.integer(pt4$total), "patterns without the excluded rows"),
327         "/patterns/0/count" = exact(pt4$first_count, "most frequent pattern count without the excluded rows")),
328    welch_outputs(w4, 1:2),
329    em_outputs(em4, lt4, "1e-4", "1e-4")
330  )
331  emit("STAT-DESC-MVA-004", "Excluded rows leave the analysis population", "missing_excluded",
332       list(quantitative = list("x", "y", "z"), categorical = list("g")),
333       outputs,
334       list("Zero-based source rows 0 and 15 are excluded before every count, test and EM; the reference drops the same rows from the re-read CSV."),
335       excluded = list(0L, 15L))

Whole script generators/r/descriptives_missing_values.R


Probability calculator: normal

Case STAT-DESC-PROBCALC-001 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/values/0/densitytolerance0.0647587978329458720.064758797832945871.14E-1917passed
/values/0/cdftail_probability0.0668072012688580710.066807201268858073.08E-1917passed
/values/0/sftail_probability0.933192798731141910.93319279873114194.81E-1817passed
/values/0/log10_cdftolerance-1.1751767216148163-1.17517672161481634.42E-1716passed
/values/0/log10_sftolerance-0.030028621232115648-0.0300286212321156481.54E-2018passed
/values/0/massabsentNonepassed
/values/1/densitytolerance0.199471140200716350.199471140200716351.43E-1817passed
/values/1/cdftail_probability0.50.50.00E+1passed
/values/1/sftail_probability0.50.50.00E+1passed
/values/1/log10_cdftolerance-0.30102999566398114-0.301029995663981142.51E-1817passed
/values/1/log10_sftolerance-0.30102999566398114-0.301029995663981142.51E-1817passed
/values/1/massabsentNonepassed
/values/2/densitytolerance0.0431386594132557660.0431386594132557662.49E-1917passed
/values/2/cdftail_probability0.959940843136182890.95994084313618293.60E-1817passed
/values/2/sftail_probability0.04005915686381710.0400591568638170861.42E-1715passed
/values/2/log10_cdftolerance-0.01775552976369445-0.0177555297636944446.45E-1815passed
/values/2/log10_sftolerance-1.3972981956626405-1.39729819566264071.85E-1615passed
/values/2/massabsentNonepassed
/values/3/densitytolerance7.4335975736714884e-077.433597573671489e-071.05E-2215passed
/values/3/cdftail_probability0.999999713348428080.99999971334842813.54E-1817passed
/values/3/sftail_probability2.8665157187919391e-072.866515718791933e-076.33E-2214passed
/values/3/log10_cdftolerance-1.2449121373882915e-07-1.244912137388289e-072.63E-2214passed
/values/3/log10_sftolerance-6.5426456723906536-6.54264567239065459.32E-1615passed
/values/3/massabsentNonepassed
/quantiles/0/lower_quantiletolerance6.08007203091989196.0800720309198918.87E-1615passed
/quantiles/0/upper_quantiletolerance13.91992796908010813.919927969080111.88E-1515passed
/quantiles/1/lower_quantiletolerance1010.00.00E+2passed
/quantiles/1/upper_quantiletolerance1010.00.00E+2passed
/quantiles/2/lower_quantiletolerance13.28970725390294313.2897072539029451.83E-1515passed
/quantiles/2/upper_quantiletolerance6.71029274609705696.7102927460970551.73E-1515passed
/interval/probabilitytolerance0.682689492137085850.68268949213708591.55E-1817passed
/interval/complementtolerance0.317310507862914090.317310507862914155.84E-1715passed
/distribution/meantolerance1010.00.00E+2passed
/distribution/variancetolerance44.00.00E+2passed
/distribution/kindexactcontinuouscontinuouspassed
/distribution/support_upperabsentNonepassed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Probability calculator: t

Case STAT-DESC-PROBCALC-002 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/values/0/densitytolerance0.0184029466507277390.0184029466507277353.53E-1815passed
/values/0/cdftail_probability0.0171904339443798050.017190433944379813.48E-1815passed
/values/0/sftail_probability0.98280956605562020.98280956605562021.55E-1817passed
/values/0/log10_cdftolerance-1.7647131601209227-1.76471316012092271.92E-1716passed
/values/0/log10_sftolerance-0.0075306250196421074-0.0075306250196421072.25E-2017passed
/values/0/massabsentNonepassed
/values/1/densitytolerance0.377546820831865580.37754682083186561.61E-1817passed
/values/1/cdftail_probability0.50.50.00E+1passed
/values/1/sftail_probability0.50.50.00E+1passed
/values/1/log10_cdftolerance-0.30102999566398114-0.301029995663981142.51E-1817passed
/values/1/log10_sftolerance-0.30102999566398114-0.301029995663981142.51E-1817passed
/values/1/massabsentNonepassed
/values/2/densitytolerance0.166346211195196630.166346211195196662.82E-1715passed
/values/2/cdftail_probability0.863817480791585270.86381748079158531.79E-1817passed
/values/2/sftail_probability0.136182519208414680.13618251920841472.40E-1715passed
/values/2/log10_cdftolerance-0.06357801152432975-0.063578011524329753.84E-2119passed
/values/2/log10_sftolerance-0.86587863617182126-0.86587863617182131.25E-1817passed
/values/2/massabsentNonepassed
/values/3/densitytolerance0.000897023698102736910.00089702369810273636.53E-1915passed
/values/3/cdftail_probability0.998679262097224240.99867926209722426.91E-1918passed
/values/3/sftail_probability0.00132073790277579080.00132073790277579146.02E-1915passed
/values/3/log10_cdftolerance-0.0005739682975473237-0.00057396829754732392.15E-1915passed
/values/3/log10_sftolerance-2.8791833585185782-2.8791833585185786.29E-1817passed
/values/3/massabsentNonepassed
/quantiles/0/lower_quantiletolerance-3.5270508906260951-3.52705089062609242.65E-1515passed
/quantiles/0/upper_quantiletolerance3.52705089062609513.52705089062609242.65E-1515passed
/quantiles/1/lower_quantiletolerance00.00.00E+2passed
/quantiles/1/upper_quantiletolerance00.00.00E+2passed
/quantiles/2/lower_quantiletolerance2.65891234720440252.65891234720440348.92E-1615passed
/quantiles/2/upper_quantiletolerance-2.6589123472044025-2.65891234720440348.92E-1615passed
/interval/probabilitytolerance0.76186850446934340.76186850446934331.12E-1615passed
/interval/complementtolerance0.238131495530656630.238131495530656741.09E-1615passed
/distribution/meantolerance00.00.00E+2passed
/distribution/variancetolerance1.81.84.44E-1716passed
/distribution/kindexactcontinuouscontinuouspassed
/distribution/support_upperabsentNonepassed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Probability calculator: chi_square

Case STAT-DESC-PROBCALC-003 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/values/0/densitytolerance0.00366159407889768590.00366159407889768943.48E-1815passed
/values/0/cdftail_probability0.000553518609575034350.0005535186095750356.52E-1914passed
/values/0/sftail_probability0.999446481390425130.99944648139042492.18E-1615passed
/values/0/log10_cdftolerance-3.256867773340117-3.2568677733401171.16E-1717passed
/values/0/log10_sftolerance-0.00024045663252066145-0.000240456632520661537.89E-2015passed
/values/0/massabsentNonepassed
/values/1/densitytolerance0.104119774808171970.104119774808171925.35E-1715passed
/values/1/cdftail_probability0.571120142446945310.57112014244694541.12E-1615passed
/values/1/sftail_probability0.428879857553054690.428879857553054861.65E-1615passed
/values/1/log10_cdftolerance-0.24327252272911862-0.24327252272911878.09E-1715passed
/values/1/log10_sftolerance-0.36766435002806902-0.367664350028068965.53E-1715passed
/values/1/massabsentNonepassed
/values/2/densitytolerance0.0177859686563992710.0177859686563992646.97E-1815passed
/values/2/cdftail_probability0.94881864658693460.94881864658693462.04E-1918passed
/values/2/sftail_probability0.0511813534130654560.0511813534130654144.19E-1715passed
/values/2/log10_cdftolerance-0.022816788956374855-0.0228167889563748451.01E-1715passed
/values/2/log10_sftolerance-1.2908882340397962-1.29088823403979652.67E-1615passed
/values/2/massabsentNonepassed
/values/3/densitytolerance4.0105572021449396e-054.0105572021449416e-052.04E-2015passed
/values/3/cdftail_probability0.999905040274918670.99990504027491873.88E-1817passed
/values/3/sftail_probability9.4959725081341848e-059.495972508134177e-058.15E-2015passed
/values/3/log10_cdftolerance-4.1242442822385615e-05-4.124244282238553e-058.82E-2014passed
/values/3/log10_sftolerance-4.0224605513810658-4.0224605513810679.29E-1615passed
/values/3/massabsentNonepassed
/quantiles/0/lower_quantiletolerance2.16734990929805752.1673499092980574.58E-1615passed
/quantiles/0/upper_quantiletolerance14.06714044934016714.0671404493401675.21E-1717passed
/quantiles/1/lower_quantiletolerance6.34581119552151666.3458111955215151.75E-1515passed
/quantiles/1/upper_quantiletolerance6.34581119552151666.3458111955215151.75E-1515passed
/quantiles/2/lower_quantiletolerance14.06714044934016714.0671404493401691.83E-1515passed
/quantiles/2/upper_quantiletolerance2.16734990929805752.1673499092980584.31E-1615passed
/interval/probabilitytolerance0.859281500221742770.85928150022174283.32E-1817passed
/interval/complementtolerance0.140718499778257290.14071849977825728.44E-1715passed
/distribution/meantolerance77.00.00E+2passed
/distribution/variancetolerance1414.00.00E+2passed
/distribution/kindexactcontinuouscontinuouspassed
/distribution/support_lowertolerance00.00.00E+2passed
/distribution/support_upperabsentNonepassed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Probability calculator: f

Case STAT-DESC-PROBCALC-004 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/values/0/densitytolerance0.697942432708171580.69794243270817247.76E-1614passed
/values/0/cdftail_probability0.140193771573581070.140193771573581181.07E-1615passed
/values/0/sftail_probability0.859806228426418960.85980622842641891.10E-1615passed
/values/0/log10_cdftolerance-0.85327128045859713-0.85327128045859683.37E-1615passed
/values/0/log10_sftolerance-0.065599413183390506-0.065599413183390554.21E-1715passed
/values/0/massabsentNonepassed
/values/1/densitytolerance0.412564201662374010.412564201662374454.42E-1614passed
/values/1/cdftail_probability0.573778620735210350.573778620735213.29E-1615passed
/values/1/sftail_probability0.426221379264789650.42622137926478992.74E-1615passed
/values/1/log10_cdftolerance-0.24125563780787154-0.241255637807871741.97E-1615passed
/values/1/log10_sftolerance-0.37036476986382483-0.370364769863824552.80E-1615passed
/values/1/massabsentNonepassed
/values/2/densitytolerance0.0372019457103187760.037201945710318787.11E-1815passed
/values/2/cdftail_probability0.949989033528223060.94998903352822321.10E-1615passed
/values/2/sftail_probability0.0500109664717769130.050010966471776927.06E-1815passed
/values/2/log10_cdftolerance-0.022281408085539842-0.0222814080855398453.12E-1815passed
/values/2/log10_sftolerance-1.3009347525448358-1.30093475254483581.52E-1716passed
/values/2/massabsentNonepassed
/values/3/densitytolerance0.00095568041573822410.00095568041573822652.39E-1814passed
/values/3/cdftail_probability0.997865328375137310.99786532837513741.07E-1615passed
/values/3/sftail_probability0.00213467162486269010.002134671624862691.40E-2017passed
/values/3/log10_cdftolerance-0.00092806701931578397-0.00092806701931578382.12E-1915passed
/values/3/log10_sftolerance-2.6706689227366072-2.6706689227366073.52E-1716passed
/values/3/massabsentNonepassed
/quantiles/0/lower_quantiletolerance0.191731839424087450.191731839424087472.46E-1715passed
/quantiles/0/upper_quantiletolerance2.60552492083067832.60552492083067832.60E-1717passed
/quantiles/1/lower_quantiletolerance0.83530589856828730.83530589856828795.52E-1615passed
/quantiles/1/upper_quantiletolerance0.83530589856828730.83530589856828795.52E-1615passed
/quantiles/2/lower_quantiletolerance5.95254468154586655.9525446815458681.74E-1515passed
/quantiles/2/upper_quantiletolerance0.0369660905439079810.036966090543908171.88E-1614passed
/interval/probabilitytolerance0.654679006706004760.65467900670600513.35E-1615passed
/interval/complementtolerance0.345320993293995180.34532099329399515.35E-1715passed
/distribution/meantolerance1.21.24.44E-1716passed
/distribution/variancetolerance1.56000000000000011.564.67E-1716passed
/distribution/kindexactcontinuouscontinuouspassed
/distribution/support_lowertolerance00.00.00E+2passed
/distribution/support_upperabsentNonepassed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Probability calculator: exponential

Case STAT-DESC-PROBCALC-005 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/values/0/densitytolerance0.384315775660929280.38431577566092931.89E-1817passed
/values/0/cdftail_probability0.0392105608476767950.039210560847676781.36E-1715passed
/values/0/sftail_probability0.960789439152323180.96078943915232323.04E-1817passed
/values/0/log10_cdftolerance-1.406596945730672-1.40659694573067221.77E-1615passed
/values/0/log10_sftolerance-0.017371779276130071-0.017371779276130074.15E-1916passed
/values/0/massabsentNonepassed
/values/1/densitytolerance0.268128018414255710.26812801841425571.19E-1817passed
/values/1/cdftail_probability0.329679953964360730.32967995396436075.75E-1715passed
/values/1/sftail_probability0.670320046035639330.67032004603563932.46E-1817passed
/values/1/log10_cdftolerance-0.4819074591271999-0.48190745912721.11E-1615passed
/values/1/log10_sftolerance-0.17371779276130073-0.173717792761300733.33E-1816passed
/values/1/massabsentNonepassed
/values/2/densitytolerance0.147151776468576940.147151776468576948.39E-1917passed
/values/2/cdftail_probability0.632120558828557670.63212055882855774.02E-1817passed
/values/2/sftail_probability0.367879441171442330.367879441171442334.02E-1816passed
/values/2/log10_cdftolerance-0.19920008462778144-0.199200084627781442.57E-1816passed
/values/2/log10_sftolerance-0.43429448190325176-0.434294481903251761.16E-1817passed
/values/2/massabsentNonepassed
/values/3/densitytolerance0.00329189881960801220.00329189881960801223.46E-2016passed
/values/3/cdftail_probability0.991770252950979980.991770252950982.55E-1918passed
/values/3/sftail_probability0.00822974704902003020.008229747049020034.71E-2017passed
/values/3/log10_cdftolerance-0.0035889220309802627-0.00358892203098026274.16E-2016passed
/values/3/log10_sftolerance-2.0846135131356087-2.08461351313560872.00E-1717passed
/values/3/massabsentNonepassed
/quantiles/0/lower_quantiletolerance0.128233235968876340.128233235968876341.24E-1817passed
/quantiles/0/upper_quantiletolerance7.48933068388497697.4893306838849771.36E-1717passed
/quantiles/1/lower_quantiletolerance1.73286795139986331.73286795139986332.89E-1716passed
/quantiles/1/upper_quantiletolerance1.73286795139986331.73286795139986332.89E-1716passed
/quantiles/2/lower_quantiletolerance5.75646273248511525.7564627324851153.18E-1818passed
/quantiles/2/upper_quantiletolerance0.263401289144565730.263401289144565731.17E-1817passed
/interval/probabilitytolerance0.468423528040983920.468423528040983975.19E-1715passed
/interval/complementtolerance0.531576471959016140.5315764719590161.12E-1615passed
/distribution/meantolerance2.52.50.00E+1passed
/distribution/variancetolerance6.256.250.00E+0passed
/distribution/kindexactcontinuouscontinuouspassed
/distribution/support_lowertolerance00.00.00E+2passed
/distribution/support_upperabsentNonepassed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Probability calculator: weibull

Case STAT-DESC-PROBCALC-006 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/values/0/densitytolerance0.0842780322702397360.084278032270239749.01E-2017passed
/values/0/cdftail_probability0.00996487775773988170.009964877757739881.72E-1815passed
/values/0/sftail_probability0.990035122242260090.99003512224226015.30E-1918passed
/values/0/log10_cdftolerance-2.0015280245565346-2.00152802455653464.26E-1716passed
/values/0/log10_sftolerance-0.0043493982050363681-0.0043493982050363682.84E-2017passed
/values/0/massabsentNonepassed
/values/1/densitytolerance0.256411124565949330.256411124565949333.22E-1816passed
/values/1/cdftail_probability0.264927469243213250.264927469243213253.86E-1816passed
/values/1/sftail_probability0.735072530756786690.73507253075678676.31E-1918passed
/values/1/log10_cdftolerance-0.57687300915367401-0.5768730091536741.08E-1918passed
/values/1/log10_sftolerance-0.13366980628542832-0.133669806285428321.99E-1816passed
/values/1/massabsentNonepassed
/values/2/densitytolerance0.208465016663817310.20846501666381733.39E-1917passed
/values/2/cdftail_probability0.632120558828557670.63212055882855774.02E-1817passed
/values/2/sftail_probability0.367879441171442330.367879441171442334.02E-1816passed
/values/2/log10_cdftolerance-0.19920008462778144-0.199200084627781442.57E-1816passed
/values/2/log10_sftolerance-0.43429448190325176-0.434294481903251761.16E-1817passed
/values/2/massabsentNonepassed
/values/3/densitytolerance0.00188851726162933570.00188851726162933574.76E-2016passed
/values/3/cdftail_probability0.998455427089103510.99845542708910351.18E-1817passed
/values/3/sftail_probability0.00154457291089652090.00154457291089652091.27E-2017passed
/values/3/log10_cdftolerance-0.00067131807552373403-0.0006713180755237344.66E-2117passed
/values/3/log10_sftolerance-2.8111915861953847-2.81119158619538472.71E-1717passed
/values/3/massabsentNonepassed
/quantiles/0/lower_quantiletolerance0.798409865090239030.7984098650902394.83E-1817passed
/quantiles/0/upper_quantiletolerance4.89992358159544144.8999235815954419.15E-1817passed
/quantiles/1/lower_quantiletolerance2.99942173889321672.99942173889321672.68E-1717passed
/quantiles/1/upper_quantiletolerance1.89719898854811441.89719898854811441.78E-1717passed
/quantiles/2/lower_quantiletolerance7.36657771595677877.3665777159567791.75E-1717passed
/quantiles/2/upper_quantiletolerance0.200416446509158960.200416446509158963.09E-1816passed
/interval/probabilitytolerance0.7645902364225110.7645902364225114.83E-1817passed
/interval/complementtolerance0.2354097635774890.2354097635774894.83E-1816passed
/distribution/meantolerance2.67673350749797262.6767335074979724.80E-1615passed
/distribution/variancetolerance2.62669344521268312.6266934452126848.97E-1615passed
/distribution/kindexactcontinuouscontinuouspassed
/distribution/support_lowertolerance00.00.00E+2passed
/distribution/support_upperabsentNonepassed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Binomial(10, 1/2): P(2 < X <= 5)

Case STAT-DESC-PROBCALC-007 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/interval/probabilitytolerance0.5683593750.5683593750.00E-7passed
/interval/integer_lowerexact330.00E+2passed
/interval/integer_upperexact550.00E+2passed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Binomial(10, 1/2): P(2 <= X < 5)

Case STAT-DESC-PROBCALC-008 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/interval/probabilitytolerance0.366210937499999940.36621093756.00E-1715passed
/interval/integer_lowerexact220.00E+2passed
/interval/integer_upperexact440.00E+2passed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Poisson(3.7) values and discrete quantiles

Case STAT-DESC-PROBCALC-009 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/values/0/masstolerance0.0247235264703393880.0247235264703393881.06E-1917passed
/values/0/cdftail_probability0.0247235264703393880.024723526470339393.58E-1815passed
/values/0/sftail_probability0.975276473529660630.97527647352966064.23E-1817passed
/values/0/densityabsentNonepassed
/values/0/in_supportexactTrueTruepassed
/values/1/masstolerance0.169232538689473130.169232538689473133.16E-1816passed
/values/1/cdftail_probability0.285433113100068250.28543311310006835.31E-1715passed
/values/1/sftail_probability0.714566886899931750.71456688689993182.45E-1817passed
/values/1/densityabsentNonepassed
/values/1/in_supportexactTrueTruepassed
/values/2/masstolerance00.00.00E+2passed
/values/2/cdftail_probability0.285433113100068250.28543311310006835.31E-1715passed
/values/2/sftail_probability0.714566886899931750.71456688689993182.45E-1817passed
/values/2/densityabsentNonepassed
/values/2/in_supportexactFalseFalsepassed
/values/3/masstolerance0.00885447672496403330.0088544767249640428.64E-1815passed
/values/3/cdftail_probability0.995151662610845640.99515166261084564.25E-1817passed
/values/3/sftail_probability0.00484833738915437830.0048483373891543878.67E-1814passed
/values/3/densityabsentNonepassed
/values/3/in_supportexactTrueTruepassed
/quantiles/0/lower_quantiletolerance11.00.00E+2passed
/quantiles/0/upper_quantiletolerance77.00.00E+2passed
/quantiles/0/achieved_cdftail_probability0.116200574410595160.116200574410595133.17E-1715passed
/quantiles/0/achieved_sftail_probability0.0352407304721390340.0352407304721390349.07E-2017passed
/quantiles/1/lower_quantiletolerance44.00.00E+2passed
/quantiles/1/upper_quantiletolerance44.00.00E+2passed
/quantiles/1/achieved_cdftail_probability0.687219365371992350.68721936537199242.29E-1817passed
/quantiles/1/achieved_sftail_probability0.312780634628007650.312780634628007652.29E-1817passed
/quantiles/2/lower_quantiletolerance77.00.00E+2passed
/quantiles/2/upper_quantiletolerance11.00.00E+2passed
/quantiles/2/achieved_cdftail_probability0.964759269527860950.9647592695278612.03E-1817passed
/quantiles/2/achieved_sftail_probability0.883799425589404830.88379942558940481.06E-1918passed
/distribution/meantolerance3.70000000000000023.72.24E-1717passed
/distribution/variancetolerance3.70000000000000023.72.24E-1717passed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Quantile at a cumulative step returns that step

Case STAT-DESC-PROBCALC-010 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/quantiles/0/lower_quantiletolerance11.00.00E+2passed
/quantiles/0/achieved_cdftail_probability0.312500000000000060.31256.00E-1715passed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Upper tail below the double range keeps its logarithm

Case STAT-DESC-PROBCALC-011 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/values/0/sfrepresentation_bound0.0passed
/values/0/log10_sftolerance-349.43700645934581-349.437006459345875.65E-1415passed
/values/0/cdftail_probability11.00.00E+2passed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Student t with one degree of freedom has no mean

Case STAT-DESC-PROBCALC-012 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/distribution/meanabsentNonepassed
/distribution/varianceabsentNonepassed
/values/0/cdftail_probability0.750000000000000220.752.20E-1615passed
/warningscontains['The mean is undefined']['The mean is undefined for these parameters.']passed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


Binomial with p = 0 is degenerate

Case STAT-DESC-PROBCALC-013 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/values/0/masstolerance11.00.00E+2passed
/values/1/masstolerance00.00.00E+2passed
/warningscontains['degenerate']['The binomial distribution is degenerate at a single value.passed

Convention

Reference code

Whole script generators/r/descriptives_probability.R


A quantile probability of 1 is refused

Case STAT-DESC-PROBCALC-014 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/descriptives_probability.R


A negative standard deviation is refused

Case STAT-DESC-PROBCALC-015 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/descriptives_probability.R


An empty request is refused

Case STAT-DESC-PROBCALC-016 · status: passed

Reference

Evidence type
R
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/descriptives_probability.R


OLS profile, desirability and the best corner

Case STAT-DESC-PROFILER-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm/glm coefficients with hand-coded desirability
Citation
R stats::lm, stats::glm and stats::qt/qnorm documentation; Derringer-Suich desirability as defined in manual chapter 54
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_profiler.R · sha256 f581b11f336b6b96
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/charts/0/points/0/predictor_valuetolerance11.00.00E+2passed
/charts/0/points/0/predictiontolerance2.39211048443573532.3921104844357371.74E-1515passed
/charts/0/points/0/lowertolerance1.46483047360106511.46483047360106621.12E-1515passed
/charts/0/points/0/uppertolerance3.31939049527040543.3193904952704082.68E-1515passed
/charts/0/points/0/desirabilitytolerance0.119605524221786760.119605524221786869.75E-1715passed
/charts/0/points/37/predictor_valuetolerance3.963.963.55E-1717passed
/charts/0/points/37/predictiontolerance5.1817811333527045.1817811333527072.63E-1515passed
/charts/0/points/37/lowertolerance4.53484634749073034.5348463474907332.71E-1515passed
/charts/0/points/37/uppertolerance5.82871591921467765.828715919214682.64E-1515passed
/charts/0/points/37/desirabilitytolerance0.259089056667635210.25908905666763531.10E-1615passed
/charts/0/points/100/predictor_valuetolerance99.00.00E+2passed
/charts/0/points/100/predictiontolerance9.9317608869140299.9317608869140345.36E-1515passed
/charts/0/points/100/lowertolerance9.05539549766269949.0553954976627055.28E-1515passed
/charts/0/points/100/uppertolerance10.80812627616535910.8081262761653645.03E-1515passed
/charts/0/points/100/desirabilitytolerance0.496588044345701470.49658804434570172.26E-1615passed
/charts/1/points/0/predictor_valueexactAApassed
/charts/1/points/0/predictiontolerance6.16193568567488196.1619356856748852.69E-1515passed
/charts/1/points/0/desirabilitytolerance0.30809678428374410.30809678428374421.07E-1615passed
/charts/1/points/1/predictor_valueexactBBpassed
/charts/1/points/1/predictiontolerance7.63437342147069137.6343734214706953.56E-1515passed
/charts/1/points/1/desirabilitytolerance0.381718671073534580.381718671073534741.63E-1615passed
/charts/1/points/2/predictor_valueexactCCpassed
/charts/1/points/2/predictiontolerance5.08606675176480575.0860667517648058.73E-1615passed
/charts/1/points/2/desirabilitytolerance0.254303337588240310.254303337588240255.75E-1715passed
/optimum_settings/x1tolerance99.00.00E+2passed
/optimum_settings/grpexactBBpassed
/optimum_outcomes/0/predictiontolerance11.40419862270983811.4041986227098445.75E-1515passed
/optimum_outcomes/0/lowertolerance10.43395343411600810.4339534341160178.72E-1515passed
/optimum_outcomes/0/uppertolerance12.37444381130366912.374443811303671.77E-1515passed
/combined_desirabilitytolerance0.57020993113549190.57020993113549223.32E-1615passed
/categorical_combinationsexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

126  goal_y <- list(response_name = "y", goal = "maximize", weight = 2, low = 0, high = 20)
127  config1 <- list(models = list(list(artifact = artifact_of(ols, "ols", "identity"))),
128                  predictors = predictors_spec(), goals = list(goal_y))
129  best <- profile_point(ols, "ols", 9, "B", 0.95)
130  out <- c(chart_outputs(0L, ols, "ols", goal_y, 0.95), level_outputs(1L, ols, "ols", goal_y, 0.95),
131           list("/optimum_settings/x1" = tol(9, "The all-upper corner is always evaluated and is best for an increasing goal."),
132                "/optimum_settings/grp" = exact("B", "B has the largest level effect."),
133                "/optimum_outcomes/0/prediction" = tol(best[["fit"]], "Prediction at the best setting."),
134                "/optimum_outcomes/0/lower" = tol(best[["lo"]], "Mean-response lower limit there."),
135                "/optimum_outcomes/0/upper" = tol(best[["hi"]], "Mean-response upper limit there."),
136                "/combined_desirability" = tol(desire(best[["fit"]], goal_y), "D equals d for one response."),
137                "/categorical_combinations" = exact(3L, "Three permitted levels.")))
138  emit_profiler("STAT-DESC-PROFILER-001", "OLS profile, desirability and the best corner", config1, out,
139                c("The high threshold 20 lies above every prediction, so d rises strictly with x1 and with the B effect."))

Whole script generators/r/descriptives_profiler.R


Two models with weighted joint desirability

Case STAT-DESC-PROFILER-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm/glm coefficients with hand-coded desirability
Citation
R stats::lm, stats::glm and stats::qt/qnorm documentation; Derringer-Suich desirability as defined in manual chapter 54
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_profiler.R · sha256 f581b11f336b6b96
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/charts/1/points/0/predictor_valuetolerance11.00.00E+2passed
/charts/1/points/0/predictiontolerance0.0162316612446395810.016231661244639553.14E-1714passed
/charts/1/points/0/lowertolerance0.000434870699235466530.00043487069923546422.33E-1814passed
/charts/1/points/0/uppertolerance0.384894016766274990.38489401676627555.03E-1614passed
/charts/1/points/0/desirabilitytolerance0.0162316612446395810.016231661244639553.14E-1714passed
/charts/1/points/37/predictor_valuetolerance3.963.963.55E-1717passed
/charts/1/points/37/predictiontolerance0.208444777412293340.2084447774122933.37E-1614passed
/charts/1/points/37/lowertolerance0.0351691734896649110.035169173489664622.92E-1614passed
/charts/1/points/37/uppertolerance0.6554610731710020.65546107317100298.84E-1614passed
/charts/1/points/37/desirabilitytolerance0.208444777412293340.2084447774122933.37E-1614passed
/charts/1/points/100/predictor_valuetolerance99.00.00E+2passed
/charts/1/points/100/predictiontolerance0.967147822537425590.96714782253742564.37E-1817passed
/charts/1/points/100/lowertolerance0.635402699130083510.63540269913008112.44E-1514passed
/charts/1/points/100/uppertolerance0.997993192162744960.9979931921627452.41E-1817passed
/charts/1/points/100/desirabilitytolerance0.967147822537425590.96714782253742564.37E-1817passed
/optimum_settings/x1tolerance99.00.00E+2passed
/optimum_settings/grpexactBBpassed
/combined_desirabilitytolerance0.68713538672138630.68713538672138652.21E-1615passed
/categorical_combinationsexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

141  goal_b <- list(response_name = "yb", goal = "maximize", weight = 1, low = 0, high = 1)
142  config2 <- list(models = list(list(artifact = artifact_of(ols, "ols", "identity")),
143                                list(artifact = artifact_of(logit, "logistic", "logit", extra = list(event_value = "yes", reference_value = "no")))),
144                  predictors = predictors_spec(), goals = list(goal_y, goal_b))
145  bl <- profile_point(logit, "logistic", 9, "B", 0.95)
146  dy <- desire(best[["fit"]], goal_y); db <- desire(bl[["fit"]], goal_b)
147  out <- c(chart_outputs(1L, logit, "logistic", goal_b, 0.95), list(
148    "/optimum_settings/x1" = tol(9, "Both responses increase with x1."),
149    "/optimum_settings/grp" = exact("B", "B is best for both responses."),
150    "/combined_desirability" = tol((dy^2 * db^1)^(1 / 3), "D = (d_y^2 d_yb^1)^(1/3) with weights 2 and 1."),
151    "/categorical_combinations" = exact(3L, "Three permitted levels.")))
152  emit_profiler("STAT-DESC-PROFILER-002", "Two models with weighted joint desirability", config2, out,
153                c("Chart 1 is x1 against the logistic model (charts run predictor by predictor, then model by model).",
154                  "Logistic limits are Wald limits on the logit scale (line 72-74)."), kind = "axis_coverage")

Whole script generators/r/descriptives_profiler.R


Target goal at a 90% confidence level

Case STAT-DESC-PROFILER-003 · status: passed

Reference

Evidence type
R
Source
R stats::lm/glm coefficients with hand-coded desirability
Citation
R stats::lm, stats::glm and stats::qt/qnorm documentation; Derringer-Suich desirability as defined in manual chapter 54
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_profiler.R · sha256 f581b11f336b6b96
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/charts/0/points/0/predictor_valuetolerance11.00.00E+2passed
/charts/0/points/0/predictiontolerance2.39211048443573532.3921104844357371.74E-1515passed
/charts/0/points/0/lowertolerance1.62268078258802051.6226807825880221.56E-1515passed
/charts/0/points/0/uppertolerance3.161540186283453.16154018628345222.24E-1515passed
/charts/0/points/0/desirabilitytolerance00.00.00E+2passed
/charts/0/points/20/predictor_valuetolerance2.60000000000000012.61.12E-1717passed
/charts/0/points/20/predictiontolerance3.90004056493139383.90004056493139652.70E-1515passed
/charts/0/points/20/lowertolerance3.278056002327783.27805600232778272.71E-1515passed
/charts/0/points/20/uppertolerance4.52202512753500764.522025127535012.69E-1515passed
/charts/0/points/20/desirabilitytolerance00.00.00E+2passed
/charts/0/points/45/predictor_valuetolerance4.59999999999999964.64.47E-1717passed
/charts/0/points/45/predictiontolerance5.78495316555096655.784953165550973.56E-1515passed
/charts/0/points/45/lowertolerance5.2694165812102925.26941658121029553.55E-1515passed
/charts/0/points/45/uppertolerance6.3004897498916416.3004897498916453.56E-1515passed
/charts/0/points/45/desirabilitytolerance0.594984388516988870.594984388516991.11E-1514passed
/charts/0/points/70/predictor_valuetolerance6.60000000000000056.60000000000000053.29E-1717passed
/charts/0/points/70/predictiontolerance7.66986576617054187.66986576617054452.70E-1515passed
/charts/0/points/70/lowertolerance7.12664308801131077.1266430880113143.60E-1515passed
/charts/0/points/70/uppertolerance8.2130884443297728.2130884443297741.81E-1515passed
/charts/0/points/70/desirabilitytolerance0.866026846765891590.86602684676589114.45E-1615passed
/charts/0/points/100/predictor_valuetolerance99.00.00E+2passed
/charts/0/points/100/predictiontolerance9.9317608869140299.9317608869140345.36E-1515passed
/charts/0/points/100/lowertolerance9.20457864176420599.2045786417642115.36E-1515passed
/charts/0/points/100/uppertolerance10.65894313206385210.6589431320638575.46E-1515passed
/charts/0/points/100/desirabilitytolerance0.413647822617194180.413647822617193131.05E-1514passed
/charts/1/points/0/predictor_valueexactAApassed
/charts/1/points/0/predictiontolerance6.16193568567488196.1619356856748852.69E-1515passed
/charts/1/points/0/desirabilitytolerance0.720645228558293940.72064522855829488.86E-1614passed
/charts/1/points/1/predictor_valueexactBBpassed
/charts/1/points/1/predictiontolerance7.63437342147069137.6343734214706953.56E-1515passed
/charts/1/points/1/desirabilitytolerance0.873125315705861720.8731253157058616.71E-1615passed
/charts/1/points/2/predictor_valueexactCCpassed
/charts/1/points/2/predictiontolerance5.08606675176480575.0860667517648058.73E-1615passed
/charts/1/points/2/desirabilitytolerance0.362022250588268590.362022250588268263.33E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

156  goal_t <- list(response_name = "y", goal = "target", weight = 1, low = 4, high = 12, target = 7)
157  config3 <- list(models = list(list(artifact = artifact_of(ols, "ols", "identity"))),
158                  predictors = predictors_spec(), goals = list(goal_t), confidence_level = 0.9)
159  emit_profiler("STAT-DESC-PROFILER-003", "Target goal at a 90% confidence level", config3,
160                c(chart_outputs(0L, ols, "ols", goal_t, 0.9, points = c(0L, 20L, 45L, 70L, 100L)),
161                  level_outputs(1L, ols, "ols", goal_t, 0.9)),
162                c("Target 7 inside (4, 12): d rises linearly to 1 at 7 and falls to 0 at 12 (line 82-83)."),
163                deferred = list(list(path = "/optimum_settings",
164                                     reason = "Many settings reach the target; the best setting found depends on the seeded search order (line 96-99).")))

Whole script generators/r/descriptives_profiler.R


Poisson profile at an entered exposure of 2.5

Case STAT-DESC-PROFILER-004 · status: passed

Reference

Evidence type
R
Source
R stats::lm/glm coefficients with hand-coded desirability
Citation
R stats::lm, stats::glm and stats::qt/qnorm documentation; Derringer-Suich desirability as defined in manual chapter 54
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_profiler.R · sha256 f581b11f336b6b96
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/charts/0/points/0/predictor_valuetolerance11.00.00E+2passed
/charts/0/points/0/predictiontolerance3.54336058043626513.54336058043626428.71E-1615passed
/charts/0/points/0/lowertolerance2.29250191602010482.2925019160201049.09E-1615passed
/charts/0/points/0/uppertolerance5.47672571841790215.4767257184179022.23E-1717passed
/charts/0/points/0/desirabilitytolerance0.940943990326062240.94094399032606226.48E-1918passed
/charts/0/points/37/predictor_valuetolerance3.963.963.55E-1717passed
/charts/0/points/37/predictiontolerance5.95080838780449465.9508083878044886.20E-1514passed
/charts/0/points/37/lowertolerance4.40533696105916484.40533696105915955.29E-1514passed
/charts/0/points/37/uppertolerance8.03845898313537438.0384589831353658.86E-1514passed
/charts/0/points/37/desirabilitytolerance0.900819860203258440.90081986020325851.09E-1615passed
/charts/0/points/100/predictor_valuetolerance99.00.00E+2passed
/charts/0/points/100/predictiontolerance14.38657975925318614.386579759253153.56E-1414passed
/charts/0/points/100/lowertolerance9.96854929998166389.9685492999816481.60E-1414passed
/charts/0/points/100/uppertolerance20.76266776046682620.7626677604667626.40E-1414passed
/charts/0/points/100/desirabilitytolerance0.760223670679113560.76022367067911415.56E-1615passed
/charts/1/points/0/predictor_valueexactAApassed
/charts/1/points/0/predictiontolerance7.13980669249805197.1398066924980438.89E-1514passed
/charts/1/points/0/desirabilitytolerance0.88100322179169910.88100322179169932.20E-1615passed
/charts/1/points/1/predictor_valueexactBBpassed
/charts/1/points/1/predictiontolerance9.8777631719964569.8777631719964497.07E-1515passed
/charts/1/points/1/desirabilitytolerance0.835370613800059060.83537061380005921.10E-1615passed
/charts/1/points/2/predictor_valueexactCCpassed
/charts/1/points/2/predictiontolerance5.85898958676834265.85898958676833557.12E-1514passed
/charts/1/points/2/desirabilitytolerance0.902350173553860930.9023501735538611.16E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

166  goal_p <- list(response_name = "cnt", goal = "minimize", weight = 1, low = 0, high = 60)
167  config4 <- list(models = list(list(artifact = artifact_of(pois, "poisson", "log", extra = list(exposure_column = "expo")), exposure = 2.5)),
168                  predictors = predictors_spec(), goals = list(goal_p))
169  emit_profiler("STAT-DESC-PROFILER-004", "Poisson profile at an entered exposure of 2.5", config4,
170                c(chart_outputs(0L, pois, "poisson", goal_p, 0.95, exposure = 2.5), level_outputs(1L, pois, "poisson", goal_p, 0.95, exposure = 2.5)),
171                c("Expected counts are multiplied by the entered exposure; Wald limits include log(2.5) (line 62-63, 74)."),
172                deferred = list(list(path = "/optimum_settings", reason = "The minimizing corner depends on the signs of the fitted Poisson effects; charts carry the reference.")))

Whole script generators/r/descriptives_profiler.R


An unattainable goal gives joint desirability zero

Case STAT-DESC-PROFILER-005 · status: passed

Reference

Evidence type
R
Source
R stats::lm/glm coefficients with hand-coded desirability
Citation
R stats::lm, stats::glm and stats::qt/qnorm documentation; Derringer-Suich desirability as defined in manual chapter 54
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_profiler.R · sha256 f581b11f336b6b96
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/combined_desirabilitytolerance00.00.00E+2passed
/charts/0/points/100/desirabilitytolerance00.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

174  goal_zero <- list(response_name = "y", goal = "maximize", weight = 1, low = 100, high = 200)
175  config5 <- list(models = list(list(artifact = artifact_of(ols, "ols", "identity"))),
176                  predictors = predictors_spec(), goals = list(goal_zero))
177  emit_profiler("STAT-DESC-PROFILER-005", "An unattainable goal gives joint desirability zero", config5,
178                list("/combined_desirability" = tol(0, "Every prediction lies below the lower threshold 100 (line 99)."),
179                     "/charts/0/points/100/desirability" = tol(0, "d is 0 at and below the lower threshold.")),
180                c(sprintf("The largest prediction in the region is %.6g, far below 100.", best[["fit"]])), kind = "boundary")

Whole script generators/r/descriptives_profiler.R


A target on a bound is refused

Case STAT-DESC-PROFILER-006 · status: passed

Reference

Evidence type
R
Source
R stats::lm/glm coefficients with hand-coded desirability
Citation
R stats::lm, stats::glm and stats::qt/qnorm documentation; Derringer-Suich desirability as defined in manual chapter 54
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_profiler.R · sha256 f581b11f336b6b96
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

182  bad_target <- goal_t; bad_target$target <- 12
183  emit_profiler("STAT-DESC-PROFILER-006", "A target on a bound is refused",
184                list(models = config1$models, predictors = predictors_spec(), goals = list(bad_target)), list(),
185                c("Line 43-44: an explicit target must lie strictly between the bounds."), kind = "invalid_input",
186                error = list(error_class = "ValueError", message_contains = "strictly inside its desirability bounds"))

Whole script generators/r/descriptives_profiler.R


Predictors must cover every retained role

Case STAT-DESC-PROFILER-007 · status: passed

Reference

Evidence type
R
Source
R stats::lm/glm coefficients with hand-coded desirability
Citation
R stats::lm, stats::glm and stats::qt/qnorm documentation; Derringer-Suich desirability as defined in manual chapter 54
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_profiler.R · sha256 f581b11f336b6b96
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

187  emit_profiler("STAT-DESC-PROFILER-007", "Predictors must cover every retained role",
188                list(models = config1$models, predictors = predictors_spec()[1], goals = list(goal_y)), list(),
189                c("Line 19: all predictors must be specified."), kind = "invalid_input",
190                error = list(error_class = "ValueError", message_contains = "exactly cover the retained model roles"))

Whole script generators/r/descriptives_profiler.R


Seeded normal table: shape, names and theoretical moments

Case STAT-DESC-RANDOM-001 · status: passed

Reference

Evidence type
Analytical
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/summary/rowsexact50500.00E+2passed
/summary/column_namessequence['a', 'b']['a', 'b']passed
/summary/dtypeexactfloat64float64passed
/summary/algorithmexactpcg64-inverse-cdf-v1pcg64-inverse-cdf-v1passed
/summary/seedexact12345123450.00E+2passed
/summary/distribution/meantolerance55.00.00E+2passed
/summary/distribution/standard_deviationtolerance22.00.00E+2passed
/summary/columns/0/countexact50500.00E+2passed
/summary/columns/1/countexact50500.00E+2passed
/frame/50absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

251  gen <- "tabelz.analysis.distribution_sampling:generate_random_data"
252  normal_request <- list(distribution = list(family = "normal", mean = 5, standard_deviation = 2),
253                         rows = 50L, column_names = list("a", "b"), seed = 12345L)
254  emit("STAT-DESC-RANDOM-001", gen, menu, "Seeded normal table: shape, names and theoretical moments",
255       normal_request,
256       list("/summary/rows" = exact(50L, "Requested rows."),
257            "/summary/column_names" = list(comparison = "sequence", values = list("a", "b"), rationale = "Requested names in order."),
258            "/summary/dtype" = exact("float64", "Continuous columns are floating point (line 103)."),
259            "/summary/algorithm" = exact("pcg64-inverse-cdf-v1", "The documented algorithm (line 106)."),
260            "/summary/seed" = exact(12345L, "The requested seed is recorded."),
261            "/summary/distribution/mean" = tol(5, "Theoretical mean."),
262            "/summary/distribution/standard_deviation" = tol(2, "Theoretical standard deviation."),
263            "/summary/columns/0/count" = exact(50L, "Every row is drawn."),
264            "/summary/columns/1/count" = exact(50L, "Every row is drawn."),
265            "/frame/50" = absent("Exactly 50 rows.")),
266       c("Draws are covered by the V27 simulation studies; repeat 2 checks that the same seed recreates the same table (line 113)."),
267       kind = "determinism", repeat_n = 2L, oracle_class = "analytic")

Whole script generators/r/descriptives_probability.R


Seeded Poisson table uses integer columns

Case STAT-DESC-RANDOM-002 · status: passed

Reference

Evidence type
Analytical
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/summary/dtypeexactint64int64passed
/summary/distribution/meantolerance3.53.50.00E+1passed
/summary/distribution/standard_deviationtolerance1.87082869338697071.87082869338697073.53E-1716passed
/summary/columns/0/countexact20200.00E+2passed
/frame/20absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

268  emit("STAT-DESC-RANDOM-002", gen, menu, "Seeded Poisson table uses integer columns",
269       list(distribution = list(family = "poisson", mean = 3.5), rows = 20L, column_names = list("count"), seed = 7L),
270       list("/summary/dtype" = exact("int64", "Poisson columns are integers (line 103-104)."),
271            "/summary/distribution/mean" = tol(3.5, "Theoretical mean."),
272            "/summary/distribution/standard_deviation" = tol(sqrt(3.5), "Theoretical standard deviation sqrt(lambda)."),
273            "/summary/columns/0/count" = exact(20L, "Every row is drawn."),
274            "/frame/20" = absent("Exactly 20 rows.")),
275       c("A discrete family produces int64 values (line 103-104)."), oracle_class = "analytic")

Whole script generators/r/descriptives_probability.R


One row is the smallest table

Case STAT-DESC-RANDOM-003 · status: passed

Reference

Evidence type
Analytical
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/summary/rowsexact110.00E+2passed
/frame/1absentNonepassed
/summary/distribution/meantolerance22.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

276  emit("STAT-DESC-RANDOM-003", gen, menu, "One row is the smallest table",
277       list(distribution = list(family = "exponential", scale = 2), rows = 1L, column_names = list("x"), seed = 0L),
278       list("/summary/rows" = exact(1L, "A single row."), "/frame/1" = absent("Exactly one row."),
279            "/summary/distribution/mean" = tol(2, "The exponential mean is its scale.")),
280       c("rows has a minimum of 1."), kind = "boundary", oracle_class = "analytic")

Whole script generators/r/descriptives_probability.R


Column names that differ only in case are refused

Case STAT-DESC-RANDOM-004 · status: passed

Reference

Evidence type
Analytical
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

281  emit("STAT-DESC-RANDOM-004", gen, menu, "Column names that differ only in case are refused",
282       list(distribution = list(family = "normal"), rows = 5L, column_names = list("Value", "value"), seed = 1L),
283       list(), c("Line 94: names must be distinct ignoring case."), kind = "invalid_input",
284       error = list(error_class = "ValueError", message_contains = "is repeated"), oracle_class = "analytic")

Whole script generators/r/descriptives_probability.R


More than 1,000,000 cells is refused

Case STAT-DESC-RANDOM-005 · status: passed

Reference

Evidence type
Analytical
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

285  emit("STAT-DESC-RANDOM-005", gen, menu, "More than 1,000,000 cells is refused",
286       list(distribution = list(family = "normal"), rows = 500001L, column_names = list("a", "b"), seed = 1L),
287       list(), c("Line 93: at most 1,000,000 cells; 500,001 x 2 = 1,000,002."), kind = "invalid_input",
288       error = list(error_class = "ValueError", message_contains = "the maximum is 1,000,000"), oracle_class = "analytic")

Whole script generators/r/descriptives_probability.R


Replaying the recorded operation recreates the table

Case STAT-DESC-RANDTABLE-001 · status: passed

Reference

Evidence type
Analytical
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/frame/50absentNonepassed
/row_ids_uniqueexactTrueTruepassed
/receipt/row_identityexactgeneratedgeneratedpassed
/receipt/seedexact12345123450.00E+2passed
/receipt/algorithmexactpcg64-inverse-cdf-v1pcg64-inverse-cdf-v1passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

291  op <- "operation:tabelz.random_table"
292  emit("STAT-DESC-RANDTABLE-001", op, menu, "Replaying the recorded operation recreates the table",
293       c(normal_request, list(recorded = TRUE)),
294       list("/frame/50" = absent("Exactly the 50 requested rows."),
295            "/row_ids_unique" = exact(TRUE, "Generated row identities are stable and distinct (line 103)."),
296            "/receipt/row_identity" = exact("generated", "The table's rows have generated identities."),
297            "/receipt/seed" = exact(12345L, "The recorded seed (line 115-116)."),
298            "/receipt/algorithm" = exact("pcg64-inverse-cdf-v1", "The recorded algorithm.")),
299       c("Lines 115-121: creating the tab records one operation; replay regenerates the table and compares its fingerprint."),
300       kind = "determinism", repeat_n = 2L, oracle_class = "analytic")

Whole script generators/r/descriptives_probability.R


A different recorded fingerprint is reported, not accepted

Case STAT-DESC-RANDTABLE-002 · status: passed

Reference

Evidence type
Analytical
Source
R stats distribution functions
Citation
R stats::Normal, TDist, Chisquare, FDist, Binomial, Poisson, Exponential and Weibull documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_probability.R · sha256 b73a1190ef5b8394
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

301  emit("STAT-DESC-RANDTABLE-002", op, menu, "A different recorded fingerprint is reported, not accepted",
302       c(normal_request, list(recorded = FALSE)), list(),
303       c("Lines 119-121: replay reports different values instead of accepting new numbers."),
304       kind = "invalid_input", oracle_class = "analytic",
305       error = list(error_class = "ValueError", message_contains = "produced different values than were recorded"))

Whole script generators/r/descriptives_probability.R


Rater agreement: unweighted Cohen's kappa and nominal alpha

Case STAT-DESC-RATER-AGREE-001 · status: passed

Reference

Evidence type
R
Source
irr::kappa2, irr::kappam.fleiss, irr::kripp.alpha and base R (Fleiss, Cohen and Everitt 1969 variance)
Citation
Gamer, M., Lemon, J. and Fellows Puspendra Singh, I. (2019). irr: various coefficients of interrater reliability and agreement. R package version 0.85; Cohen (1960); Fleiss, Cohen and Everitt (1969) Psychological Bulletin 72:323-327; Fleiss, Nee and Landis (1979); Krippendorff (2004).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, psych 2.6.5, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/descriptives_reliability_agreement.R · sha256 6f498a9f8544f55b
Dataset
reference_data/datasets/DESC_RATERS_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cohen/unitsexact29290.00E+2passed
/cohen/kappatolerance0.341977309562398750.341977309562398754.22E-1816passed
/cohen/observed_agreementtolerance0.517241379310344860.51724137931034492.04E-1817passed
/cohen/expected_agreementtolerance0.266349583828775270.266349583828775271.50E-1817passed
/cohen/standard_errortolerance0.121096843021858410.121096843021858431.74E-1715passed
/cohen/lowertolerance0.104631858598055740.104631858598055658.52E-1715passed
/cohen/uppertolerance0.57932276052674170.57932276052674181.09E-1615passed
/cohen/null_standard_errortolerance0.106108799578780390.106108799578780391.63E-1816passed
/cohen/z_valuetolerance3.222893020371023.222893020371021.02E-1717passed
/cohen/p_valuetail_probability0.00126902939636419630.00126902939636411877.76E-1713passed
/table/0/0exact660.00E+2passed
/table/1/2exact110.00E+2passed
/complete_unitsexact29290.00E+2passed
/percent_agreementtolerance0.517241379310344860.51724137931034492.04E-1817passed
/krippendorff/alphatolerance0.334999999999999960.334999999999999964.47E-1816passed
/krippendorff/pairable_valuesexact58580.00E+2passed
/krippendorff/unitsexact29290.00E+2passed
/unitsexact30300.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

403  outputs <- c(cohen_outputs(rd$r1, rd$r2, "none", 0.95), alpha_outputs(c("r1", "r2"), "nominal"),
404               list("/units" = exact(nrow(rd), "rows")))
405  emit("STAT-DESC-RATER-AGREE-001", "Rater agreement: unweighted Cohen's kappa and nominal alpha",
406       RATER, "raterAgreementMenuItem", RATER_REQ, "DESC_RATERS_B7",
407       list(raters = list("r1", "r2"), categories = list(), weights = "none", alpha_metric = "nominal",
408            confidence_level = 0.95, replicates = 500L, seed = 20260928L, kappa_bootstrap = FALSE),
409       outputs, c(agreement_notes, "Without declared categories the observed codes sort numerically: 1, 2, 3, 4."),
410       RATER_SOURCE, RATER_CITATION)

Whole script generators/r/descriptives_reliability_agreement.R


Rater agreement: linear-weighted kappa and ordinal alpha

Case STAT-DESC-RATER-AGREE-002 · status: passed

Reference

Evidence type
R
Source
irr::kappa2, irr::kappam.fleiss, irr::kripp.alpha and base R (Fleiss, Cohen and Everitt 1969 variance)
Citation
Gamer, M., Lemon, J. and Fellows Puspendra Singh, I. (2019). irr: various coefficients of interrater reliability and agreement. R package version 0.85; Cohen (1960); Fleiss, Cohen and Everitt (1969) Psychological Bulletin 72:323-327; Fleiss, Nee and Landis (1979); Krippendorff (2004).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, psych 2.6.5, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/descriptives_reliability_agreement.R · sha256 6f498a9f8544f55b
Dataset
reference_data/datasets/DESC_RATERS_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cohen/unitsexact29290.00E+2passed
/cohen/kappatolerance0.545861297539149690.54586129753914972.43E-1817passed
/cohen/observed_agreementtolerance0.839080459770115030.83908045977011491.13E-1615passed
/cohen/expected_agreementtolerance0.645659928656361570.64565992865636162.76E-1918passed
/cohen/standard_errortolerance0.0948312480805268420.094831248080526971.25E-1614passed
/cohen/lowertolerance0.359995466692334290.359995466692333735.58E-1614passed
/cohen/uppertolerance0.731727128385965760.73172712838596571.07E-1615passed
/cohen/null_standard_errortolerance0.12393343336160750.123933433361607381.21E-1615passed
/cohen/z_valuetolerance4.40447168074865044.404471680748653.60E-1717passed
/cohen/p_valuetail_probability1.0604199179198304e-051.0604199179241333e-054.30E-1711passed
/table/0/0exact660.00E+2passed
/table/1/2exact110.00E+2passed
/complete_unitsexact29290.00E+2passed
/percent_agreementtolerance0.517241379310344860.51724137931034492.04E-1817passed
/krippendorff/alphatolerance0.740263613187853720.74026361318785374.34E-1918passed
/krippendorff/pairable_valuesexact58580.00E+2passed
/krippendorff/unitsexact29290.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

413  outputs <- c(cohen_outputs(rd$r1, rd$r2, "linear", 0.95), alpha_outputs(c("r1", "r2"), "ordinal"))
414  emit("STAT-DESC-RATER-AGREE-002", "Rater agreement: linear-weighted kappa and ordinal alpha",
415       RATER, "raterAgreementMenuItem", RATER_REQ, "DESC_RATERS_B7",
416       list(raters = list("r1", "r2"), categories = as.list(CATS), weights = "linear",
417            alpha_metric = "ordinal", confidence_level = 0.95, replicates = 500L,
418            seed = 20260928L, kappa_bootstrap = FALSE),
419       outputs, agreement_notes, RATER_SOURCE, RATER_CITATION, kind = "axis_coverage")

Whole script generators/r/descriptives_reliability_agreement.R


Rater agreement: quadratic-weighted kappa at 90% and interval alpha

Case STAT-DESC-RATER-AGREE-003 · status: passed

Reference

Evidence type
R
Source
irr::kappa2, irr::kappam.fleiss, irr::kripp.alpha and base R (Fleiss, Cohen and Everitt 1969 variance)
Citation
Gamer, M., Lemon, J. and Fellows Puspendra Singh, I. (2019). irr: various coefficients of interrater reliability and agreement. R package version 0.85; Cohen (1960); Fleiss, Cohen and Everitt (1969) Psychological Bulletin 72:323-327; Fleiss, Nee and Landis (1979); Krippendorff (2004).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, psych 2.6.5, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/descriptives_reliability_agreement.R · sha256 6f498a9f8544f55b
Dataset
reference_data/datasets/DESC_RATERS_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cohen/unitsexact29290.00E+2passed
/cohen/kappatolerance0.733595800524934450.73359580052493386.67E-1615passed
/cohen/observed_agreementtolerance0.946360153256704970.94636015325670491.09E-1615passed
/cohen/expected_agreementtolerance0.798652397938961590.79865239793896169.59E-1917passed
/cohen/standard_errortolerance0.0671507019685509690.067150701968550972.65E-1917passed
/cohen/lowertolerance0.623142724839625960.62314272483962545.54E-1615passed
/cohen/uppertolerance0.844048876210242720.84404887621024225.60E-1615passed
/cohen/null_standard_errortolerance0.180732089891868150.180732089891868422.75E-1614passed
/cohen/z_valuetolerance4.05902350248835394.05902350248836451.06E-1414passed
/cohen/p_valuetail_probability4.9278354256276913e-054.927835425624207e-053.48E-1712passed
/table/0/0exact660.00E+2passed
/table/1/2exact110.00E+2passed
/complete_unitsexact29290.00E+2passed
/percent_agreementtolerance0.517241379310344860.51724137931034492.04E-1817passed
/krippendorff/alphatolerance0.736807387862796760.73680738786279681.22E-1817passed
/krippendorff/pairable_valuesexact58580.00E+2passed
/krippendorff/unitsexact29290.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

422  outputs <- c(cohen_outputs(rd$r1, rd$r2, "quadratic", 0.90), alpha_outputs(c("r1", "r2"), "interval"))
423  emit("STAT-DESC-RATER-AGREE-003", "Rater agreement: quadratic-weighted kappa at 90% and interval alpha",
424       RATER, "raterAgreementMenuItem", RATER_REQ, "DESC_RATERS_B7",
425       list(raters = list("r1", "r2"), categories = as.list(CATS), weights = "quadratic",
426            alpha_metric = "interval", confidence_level = 0.90, replicates = 500L,
427            seed = 20260928L, kappa_bootstrap = FALSE),
428       outputs, c(agreement_notes, "Interval alpha uses the numeric codes' squared differences; the 90% kappa interval uses qnorm(0.95)."),
429       RATER_SOURCE, RATER_CITATION, kind = "axis_coverage")

Whole script generators/r/descriptives_reliability_agreement.R


Rater agreement: Fleiss' kappa for three raters with missing ratings

Case STAT-DESC-RATER-AGREE-004 · status: passed

Reference

Evidence type
R
Source
irr::kappa2, irr::kappam.fleiss, irr::kripp.alpha and base R (Fleiss, Cohen and Everitt 1969 variance)
Citation
Gamer, M., Lemon, J. and Fellows Puspendra Singh, I. (2019). irr: various coefficients of interrater reliability and agreement. R package version 0.85; Cohen (1960); Fleiss, Cohen and Everitt (1969) Psychological Bulletin 72:323-327; Fleiss, Nee and Landis (1979); Krippendorff (2004).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, psych 2.6.5, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/descriptives_reliability_agreement.R · sha256 6f498a9f8544f55b
Dataset
reference_data/datasets/DESC_RATERS_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fleiss/unitsexact25250.00E+2passed
/fleiss/kappatolerance0.422187981510015410.42218798151001541.25E-1817passed
/fleiss/z_valuetolerance5.62847108287739855.62847108287739851.41E-1818passed
/fleiss/p_valuetail_probability1.8181397470801036e-081.8181397470801e-083.67E-2314passed
/category_kappas/[category=1]/kappatolerance0.699999999999999960.74.41E-1817passed
/category_kappas/[category=1]/null_standard_errortolerance0.115470053837925160.115470053837925162.87E-1816passed
/category_kappas/[category=1]/z_valuetolerance6.06217782649106956.06217782649106959.98E-1817passed
/category_kappas/[category=1]/p_valuetail_probability1.342906179143749e-091.3429061791437452e-093.76E-2414passed
/category_kappas/[category=2]/kappatolerance0.305555555555555580.30555555555555562.27E-1918passed
/category_kappas/[category=2]/null_standard_errortolerance0.115470053837925160.115470053837925162.87E-1816passed
/category_kappas/[category=2]/z_valuetolerance2.64618873378578462.64618873378578463.63E-1716passed
/category_kappas/[category=2]/p_valuetail_probability0.00814043841567800570.0081404384156780064.58E-2017passed
/category_kappas/[category=3]/kappatolerance0.295112781954887220.29511278195488724.46E-1816passed
/category_kappas/[category=3]/null_standard_errortolerance0.115470053837925160.115470053837925162.87E-1816passed
/category_kappas/[category=3]/z_valuetolerance2.55575166154430192.5557516615443021.07E-1717passed
/category_kappas/[category=3]/p_valuetail_probability0.0105958709750488420.0105958709750488348.37E-1815passed
/category_kappas/[category=4]/kappatolerance0.207746478873239380.207746478873239381.89E-1817passed
/category_kappas/[category=4]/null_standard_errortolerance0.115470053837925160.115470053837925162.87E-1816passed
/category_kappas/[category=4]/z_valuetolerance1.79913728250992481.79913728250992482.66E-1716passed
/category_kappas/[category=4]/p_valuetail_probability0.0719969673985771250.071996967398577123.14E-1917passed
/krippendorff/alphatolerance0.43779527559055120.43779527559055122.78E-1817passed
/krippendorff/pairable_valuesexact85850.00E+2passed
/krippendorff/unitsexact30300.00E+2passed
/complete_unitsexact25250.00E+2passed
/percent_agreementtolerance0.440.442.22E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

432  complete3 <- stats::complete.cases(rd)
433  outputs <- c(fleiss_outputs(c("r1", "r2", "r3")), alpha_outputs(c("r1", "r2", "r3"), "nominal"),
434               list("/complete_units" = exact(sum(complete3), "units rated by all three raters"),
435                    "/percent_agreement" = tol(mean(apply(rd[complete3, ], 1, function(v) length(unique(v)) == 1)), "share of complete units where all three agree")))
436  emit("STAT-DESC-RATER-AGREE-004", "Rater agreement: Fleiss' kappa for three raters with missing ratings",
437       RATER, "raterAgreementMenuItem", RATER_REQ, "DESC_RATERS_B7",
438       list(raters = list("r1", "r2", "r3"), categories = as.list(CATS), weights = "none",
439            alpha_metric = "nominal", confidence_level = 0.95, replicates = 500L,
440            seed = 20260928L, kappa_bootstrap = FALSE),
441       outputs, c(agreement_notes, "Fleiss' kappa uses the 25 units with all three ratings; Krippendorff's alpha also uses the units with two ratings."),
442       RATER_SOURCE, RATER_CITATION, kind = "axis_coverage")

Whole script generators/r/descriptives_reliability_agreement.R


Scale reliability with a reverse-coded item (halves split, ICC(3,1))

Case STAT-DESC-SCALE-REL-001 · status: passed

Reference

Evidence type
R
Source
psych::alpha, psych::ICC, stats::friedman.test and base R covariance algebra
Citation
Revelle, W. (2024). psych: procedures for psychological research. R package version 2.6.5; Cronbach (1951); Guttman (1945) Psychometrika 10:255-282; McGraw and Wong (1996) Psychological Methods 1:30-46; Shrout and Fleiss (1979).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, psych 2.6.5, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/descriptives_reliability_agreement.R · sha256 6f498a9f8544f55b
Dataset
reference_data/datasets/DESC_SCALE_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact19190.00E+2passed
/kexact550.00E+2passed
/alphatolerance0.819765240439227490.81976524043922753.85E-1817passed
/standardized_alphatolerance0.824835561491332970.82483556149133311.10E-1615passed
/guttman/0tolerance0.6558121923513820.6558121923513824.92E-1817passed
/guttman/1tolerance0.828882635859399470.82888263585939954.75E-1817passed
/guttman/2tolerance0.819765240439227490.81976524043922753.85E-1817passed
/guttman/3tolerance0.854221885649375290.85422188564937512.22E-1615passed
/guttman/4tolerance0.810307017342470130.81030701734247013.87E-1817passed
/guttman/5tolerance0.840807426306520940.84080742630652081.07E-1615passed
/split_half/correlationtolerance0.792481949744770820.79248194974477085.78E-1918passed
/split_half/spearman_brown_equaltolerance0.884228652743321960.8842286527433223.39E-1817passed
/split_half/spearman_brown_unequaltolerance0.887927684118524850.88792768411852491.44E-1817passed
/split_half/guttman_split_halftolerance0.854221885649375290.85422188564937531.27E-1918passed
/summaries/0/meantolerance2.79999999999999982.82.24E-1717passed
/summaries/3/meantolerance0.485009731990154960.485009731990154962.49E-1817passed
/split_half/first_alphatolerance0.657453109575518280.65745310957551812.22E-1615passed
/split_half/second_alphatolerance0.63999999999999990.63999999999999992.30E-1817passed
/item_totals/0/meantolerance2.89473684210526332.89473684210526332.52E-1717passed
/item_totals/0/sdtolerance1.14962490704602651.14962490704602651.04E-1717passed
/item_totals/0/corrected_item_totaltolerance0.49687565555181190.496875655551811731.65E-1615passed
/item_totals/0/scale_mean_if_deletedtolerance11.10526315789473611.1052631578947362.81E-1616passed
/item_totals/0/scale_variance_if_deletedtolerance10.43274853801169610.4327485380116981.55E-1515passed
/item_totals/0/alpha_if_deletedtolerance0.822869955156950720.82286995515695073.54E-1817passed
/item_totals/1/meantolerance33.00.00E+2passed
/item_totals/1/sdtolerance11.00.00E+2passed
/item_totals/1/corrected_item_totaltolerance0.499999999999999940.499999999999999944.49E-1817passed
/item_totals/1/scale_mean_if_deletedtolerance1111.00.00E+2passed
/item_totals/1/scale_variance_if_deletedtolerance11.11111111111111111.111111111111112.84E-1616passed
/item_totals/1/alpha_if_deletedtolerance0.815438596491227960.8154385964912282.74E-1817passed
/item_totals/2/meantolerance2.68421052631578942.68421052631578941.98E-1717passed
/item_totals/2/sdtolerance1.00291971424255811.0029197142425582.41E-1817passed
/item_totals/2/corrected_item_totaltolerance0.764004008182177840.76400400818217753.32E-1615passed
/item_totals/2/scale_mean_if_deletedtolerance11.31578947368421111.315789473684213.80E-1616passed
/item_totals/2/scale_variance_if_deletedtolerance9.67251461988304059.672514619883049.76E-1817passed
/item_totals/2/alpha_if_deletedtolerance0.739218057234985790.73921805723498582.55E-1817passed
/item_totals/3/meantolerance2.7368421052631582.7368421052631583.51E-1716passed
/item_totals/3/sdtolerance1.04573765900534931.04573765900534931.82E-1716passed
/item_totals/3/corrected_item_totaltolerance0.620579922173684410.62057992217368443.76E-1817passed
/item_totals/3/scale_mean_if_deletedtolerance11.26315789473684211.2631578947368424.79E-1616passed
/item_totals/3/scale_variance_if_deletedtolerance10.204678362573110.20467836257319.66E-1717passed
/item_totals/3/alpha_if_deletedtolerance0.781661891117478410.78166189111747843.45E-1817passed
/item_totals/4/meantolerance2.68421052631578942.68421052631578941.98E-1717passed
/item_totals/4/sdtolerance0.945905302926917320.94590530292691734.06E-1817passed
/item_totals/4/corrected_item_totaltolerance0.71428877058078910.71428877058078883.32E-1615passed
/item_totals/4/scale_mean_if_deletedtolerance11.31578947368421111.315789473684213.80E-1616passed
/item_totals/4/scale_variance_if_deletedtolerance10.22807017543859610.2280701754385961.32E-1616passed
/item_totals/4/alpha_if_deletedtolerance0.757004002287021140.75700400228702111.75E-1817passed
/tests/[name=Friedman's chi-square]/statistictolerance3.31182795698924753.311827956989213.73E-1413passed
/tests/[name=Friedman's chi-square]/p_valuetail_probability0.507060407766553590.50706040776655976.11E-1513passed
/tests/[name=Hotelling's T-squared (equal item means)]/statistictolerance6.05233514429546566.0523351442954747.99E-1514passed
/tests/[name=Hotelling's T-squared (equal item means)]/p_valuetail_probability0.328357717337528050.328357717337527673.84E-1614passed
/anova/0/1tolerance55.59999999999999455.600000000000011.45E-1415passed
/anova/1/1tolerance1.5157894736842091.51578947368421152.46E-1514passed
/anova/2/1tolerance40.08421052631578640.08421052631581.44E-1415passed
/anova/2/3tolerance0.55672514619883040.55672514619883051.12E-1615passed
/icc/0/estimatetolerance0.476347634763476140.476347634763476362.17E-1615passed
/icc/0/lowertolerance0.270390403950215960.270390403950216242.80E-1614passed
/icc/0/uppertolerance0.701680663807433320.70168066380743352.27E-1615passed
/icc/0/f_valuetolerance5.54831932773108875.5483193277310923.57E-1515passed
/icc/0/df1tolerance1818.00.00E+2passed
/icc/0/df2tolerance7272.00.00E+2passed
/icc/0/p_valuetail_probability6.7726694227101891e-086.772669422710107e-088.20E-2213passed
/icc/1/estimatetolerance0.819765240439227380.81976524043922762.25E-1615passed
/icc/1/lowertolerance0.649489311048295550.64948931104829582.20E-1615passed
/icc/1/uppertolerance0.921633551164351080.92163355116435113.61E-1817passed
/missing_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

210  m <- adjusted(ITEMS, "i4")
211  outputs <- c(scale_core(m, c("i1", "i2", "i3"), c("i4", "i5")),
212               icc_outputs(m, "two_way_mixed", "consistency", 0.95),
213               list("/missing_count" = exact(1L, "row 7 has a missing i2")))
214  emit("STAT-DESC-SCALE-REL-001", "Scale reliability with a reverse-coded item (halves split, ICC(3,1))",
215       SCALE, "scaleReliabilityMenuItem", SCALE_REQ, "DESC_SCALE_B7",
216       list(items = as.list(ITEMS), reverse_items = list("i4"), scale_low = 1, scale_high = 5,
217            split = "halves", first_half = list(), icc_model = "two_way_mixed",
218            icc_type = "consistency", confidence_level = 0.95),
219       outputs,
220        c(paste0(MANUAL_SCALE, ":14-17: reversal uses x* = L + U - x once; incomplete cases are omitted."),
221          paste0(MANUAL_SCALE, ":19-29: Cronbach's alpha from the covariance matrix; standardized alpha from the mean inter-item correlation; Guttman's coefficients are covariance-based summaries."),
222          paste0(MANUAL_SCALE, ":33-38: first/second half split; Spearman-Brown 2r/(1+r) and Guttman's split-half value."),
223          paste0(MANUAL_SCALE, ":54-72: ICC(C,1) = (MSR - MSE)/(MSR + (k-1) MSE), ICC(C,k) = (MSR - MSE)/MSR; consistency intervals use the F pivot."),
224          "psych::ICC ICC3 / ICC3k are the two-way consistency coefficients ICC(C,1) and ICC(C,k) (McGraw and Wong).",
225          ICC_PSYCH_NOTE,
226          GUTTMAN_L6_NOTE,
227          "Halves: first i1, i2, i3 ((k+1)/2 = 3 items); second i4, i5."),
228       SCALE_SOURCE, SCALE_CITATION)

Whole script generators/r/descriptives_reliability_agreement.R


Scale reliability with an odd/even split and ICC(2,1) absolute agreement

Case STAT-DESC-SCALE-REL-002 · status: passed

Reference

Evidence type
R
Source
psych::alpha, psych::ICC, stats::friedman.test and base R covariance algebra
Citation
Revelle, W. (2024). psych: procedures for psychological research. R package version 2.6.5; Cronbach (1951); Guttman (1945) Psychometrika 10:255-282; McGraw and Wong (1996) Psychological Methods 1:30-46; Shrout and Fleiss (1979).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, psych 2.6.5, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/descriptives_reliability_agreement.R · sha256 6f498a9f8544f55b
Dataset
reference_data/datasets/DESC_SCALE_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact19190.00E+2passed
/kexact550.00E+2passed
/alphatolerance0.819765240439227490.81976524043922753.85E-1817passed
/standardized_alphatolerance0.824835561491332970.82483556149133311.10E-1615passed
/guttman/0tolerance0.6558121923513820.6558121923513824.92E-1817passed
/guttman/1tolerance0.828882635859399470.82888263585939954.75E-1817passed
/guttman/2tolerance0.819765240439227490.81976524043922753.85E-1817passed
/guttman/3tolerance0.795910639909125360.79591063990912562.23E-1615passed
/guttman/4tolerance0.810307017342470130.81030701734247013.87E-1817passed
/guttman/5tolerance0.840807426306520940.84080742630652081.07E-1615passed
/split_half/correlationtolerance0.723403855826929680.72340385582692952.26E-1615passed
/split_half/spearman_brown_equaltolerance0.839505903832185150.83950590383218483.31E-1615passed
/split_half/spearman_brown_unequaltolerance0.844198420661896430.84419842066189595.54E-1615passed
/split_half/guttman_split_halftolerance0.795910639909125360.79591063990912541.41E-1817passed
/summaries/0/meantolerance2.79999999999999982.82.24E-1717passed
/summaries/3/meantolerance0.485009731990154960.485009731990154962.49E-1817passed
/split_half/first_alphatolerance0.760733452593917780.76073345259391762.22E-1615passed
/split_half/second_alphatolerance0.483050847457626940.48305084745762672.20E-1615passed
/item_totals/0/meantolerance2.89473684210526332.89473684210526332.52E-1717passed
/item_totals/0/sdtolerance1.14962490704602651.14962490704602651.04E-1717passed
/item_totals/0/corrected_item_totaltolerance0.49687565555181190.496875655551811731.65E-1615passed
/item_totals/0/scale_mean_if_deletedtolerance11.10526315789473611.1052631578947362.81E-1616passed
/item_totals/0/scale_variance_if_deletedtolerance10.43274853801169610.4327485380116981.55E-1515passed
/item_totals/0/alpha_if_deletedtolerance0.822869955156950720.82286995515695073.54E-1817passed
/item_totals/1/meantolerance33.00.00E+2passed
/item_totals/1/sdtolerance11.00.00E+2passed
/item_totals/1/corrected_item_totaltolerance0.499999999999999940.499999999999999944.49E-1817passed
/item_totals/1/scale_mean_if_deletedtolerance1111.00.00E+2passed
/item_totals/1/scale_variance_if_deletedtolerance11.11111111111111111.111111111111112.84E-1616passed
/item_totals/1/alpha_if_deletedtolerance0.815438596491227960.8154385964912282.74E-1817passed
/item_totals/2/meantolerance2.68421052631578942.68421052631578941.98E-1717passed
/item_totals/2/sdtolerance1.00291971424255811.0029197142425582.41E-1817passed
/item_totals/2/corrected_item_totaltolerance0.764004008182177840.76400400818217753.32E-1615passed
/item_totals/2/scale_mean_if_deletedtolerance11.31578947368421111.315789473684213.80E-1616passed
/item_totals/2/scale_variance_if_deletedtolerance9.67251461988304059.672514619883049.76E-1817passed
/item_totals/2/alpha_if_deletedtolerance0.739218057234985790.73921805723498582.55E-1817passed
/item_totals/3/meantolerance2.7368421052631582.7368421052631583.51E-1716passed
/item_totals/3/sdtolerance1.04573765900534931.04573765900534931.82E-1716passed
/item_totals/3/corrected_item_totaltolerance0.620579922173684410.62057992217368443.76E-1817passed
/item_totals/3/scale_mean_if_deletedtolerance11.26315789473684211.2631578947368424.79E-1616passed
/item_totals/3/scale_variance_if_deletedtolerance10.204678362573110.20467836257319.66E-1717passed
/item_totals/3/alpha_if_deletedtolerance0.781661891117478410.78166189111747843.45E-1817passed
/item_totals/4/meantolerance2.68421052631578942.68421052631578941.98E-1717passed
/item_totals/4/sdtolerance0.945905302926917320.94590530292691734.06E-1817passed
/item_totals/4/corrected_item_totaltolerance0.71428877058078910.71428877058078883.32E-1615passed
/item_totals/4/scale_mean_if_deletedtolerance11.31578947368421111.315789473684213.80E-1616passed
/item_totals/4/scale_variance_if_deletedtolerance10.22807017543859610.2280701754385961.32E-1616passed
/item_totals/4/alpha_if_deletedtolerance0.757004002287021140.75700400228702111.75E-1817passed
/tests/[name=Friedman's chi-square]/statistictolerance3.31182795698924753.311827956989213.73E-1413passed
/tests/[name=Friedman's chi-square]/p_valuetail_probability0.507060407766553590.50706040776655976.11E-1513passed
/tests/[name=Hotelling's T-squared (equal item means)]/statistictolerance6.05233514429546566.0523351442954747.99E-1514passed
/tests/[name=Hotelling's T-squared (equal item means)]/p_valuetail_probability0.328357717337528050.328357717337527673.84E-1614passed
/anova/0/1tolerance55.59999999999999455.600000000000011.45E-1415passed
/anova/1/1tolerance1.5157894736842091.51578947368421152.46E-1514passed
/anova/2/1tolerance40.08421052631578640.08421052631581.44E-1415passed
/anova/2/3tolerance0.55672514619883040.55672514619883051.12E-1615passed
/icc/0/estimatetolerance0.480577136514983180.48057713651498331.14E-1615passed
/icc/0/lowertolerance0.275083426166203070.275083426166203241.68E-1615passed
/icc/0/uppertolerance0.704759970312641370.70475997031264146.04E-1918passed
/icc/0/f_valuetolerance5.54831932773108875.5483193277310923.57E-1515passed
/icc/0/df1tolerance1818.00.00E+2passed
/icc/0/df2tolerance7272.00.00E+2passed
/icc/0/p_valuetail_probability6.7726694227101891e-086.772669422710107e-088.20E-2213passed
/icc/1/estimatetolerance0.822255981769844160.82225598176984442.20E-1615passed
/icc/1/lowertolerance0.654856477223706320.65485647722370652.26E-1615passed
/icc/1/uppertolerance0.922692603673287540.92269260367328753.09E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

231  outputs <- c(scale_core(m, c("i1", "i3", "i5"), c("i2", "i4")),
232               icc_outputs(m, "two_way_random", "agreement", 0.95))
233  emit("STAT-DESC-SCALE-REL-002", "Scale reliability with an odd/even split and ICC(2,1) absolute agreement",
234       SCALE, "scaleReliabilityMenuItem", SCALE_REQ, "DESC_SCALE_B7",
235       list(items = as.list(ITEMS), reverse_items = list("i4"), scale_low = 1, scale_high = 5,
236            split = "odd_even", first_half = list(), icc_model = "two_way_random",
237            icc_type = "agreement", confidence_level = 0.95),
238       outputs,
239        c(paste0(MANUAL_SCALE, ":54-72: ICC(A,1) = (MSR - MSE)/(MSR + (k-1) MSE + k (MSC - MSE)/n); agreement intervals use an approximate df calculation."),
240          "psych::ICC ICC2 / ICC2k are the two-way random absolute-agreement coefficients ICC(A,1) and ICC(A,k); their bounds follow Shrout and Fleiss (1979) / McGraw and Wong (1996).",
241          ICC_PSYCH_NOTE,
242          GUTTMAN_L6_NOTE,
243          "Odd/even split: first i1, i3, i5; second i2, i4."),
244       SCALE_SOURCE, SCALE_CITATION, kind = "axis_coverage")

Whole script generators/r/descriptives_reliability_agreement.R


Scale reliability with a declared split and one-way ICC at 90%

Case STAT-DESC-SCALE-REL-003 · status: passed

Reference

Evidence type
R
Source
psych::alpha, psych::ICC, stats::friedman.test and base R covariance algebra
Citation
Revelle, W. (2024). psych: procedures for psychological research. R package version 2.6.5; Cronbach (1951); Guttman (1945) Psychometrika 10:255-282; McGraw and Wong (1996) Psychological Methods 1:30-46; Shrout and Fleiss (1979).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, psych 2.6.5, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/descriptives_reliability_agreement.R · sha256 6f498a9f8544f55b
Dataset
reference_data/datasets/DESC_SCALE_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact19190.00E+2passed
/kexact550.00E+2passed
/alphatolerance0.819765240439227490.81976524043922753.85E-1817passed
/standardized_alphatolerance0.824835561491332970.82483556149133311.10E-1615passed
/guttman/0tolerance0.6558121923513820.6558121923513824.92E-1817passed
/guttman/1tolerance0.828882635859399470.82888263585939954.75E-1817passed
/guttman/2tolerance0.819765240439227490.81976524043922753.85E-1817passed
/guttman/3tolerance0.645967436577054020.6459674365770541.64E-1817passed
/guttman/4tolerance0.810307017342470130.81030701734247013.87E-1817passed
/guttman/5tolerance0.840807426306520940.84080742630652081.07E-1615passed
/split_half/correlationtolerance0.516345339264743460.51634533926474361.15E-1615passed
/split_half/spearman_brown_equaltolerance0.681039240724824250.68103924072482423.66E-1817passed
/split_half/spearman_brown_unequaltolerance0.687741345122960030.68774134512296011.16E-1615passed
/split_half/guttman_split_halftolerance0.645967436577054020.64596743657705382.20E-1615passed
/summaries/0/meantolerance2.79999999999999982.82.24E-1717passed
/summaries/3/meantolerance0.485009731990154960.485009731990154962.49E-1817passed
/split_half/first_alphatolerance0.826086956521739020.8260869565217394.24E-1817passed
/split_half/second_alphatolerance0.790048543689320540.79004854368932052.90E-1817passed
/item_totals/0/meantolerance2.89473684210526332.89473684210526332.52E-1717passed
/item_totals/0/sdtolerance1.14962490704602651.14962490704602651.04E-1717passed
/item_totals/0/corrected_item_totaltolerance0.49687565555181190.496875655551811731.65E-1615passed
/item_totals/0/scale_mean_if_deletedtolerance11.10526315789473611.1052631578947362.81E-1616passed
/item_totals/0/scale_variance_if_deletedtolerance10.43274853801169610.4327485380116981.55E-1515passed
/item_totals/0/alpha_if_deletedtolerance0.822869955156950720.82286995515695073.54E-1817passed
/item_totals/1/meantolerance33.00.00E+2passed
/item_totals/1/sdtolerance11.00.00E+2passed
/item_totals/1/corrected_item_totaltolerance0.499999999999999940.499999999999999944.49E-1817passed
/item_totals/1/scale_mean_if_deletedtolerance1111.00.00E+2passed
/item_totals/1/scale_variance_if_deletedtolerance11.11111111111111111.111111111111112.84E-1616passed
/item_totals/1/alpha_if_deletedtolerance0.815438596491227960.8154385964912282.74E-1817passed
/item_totals/2/meantolerance2.68421052631578942.68421052631578941.98E-1717passed
/item_totals/2/sdtolerance1.00291971424255811.0029197142425582.41E-1817passed
/item_totals/2/corrected_item_totaltolerance0.764004008182177840.76400400818217753.32E-1615passed
/item_totals/2/scale_mean_if_deletedtolerance11.31578947368421111.315789473684213.80E-1616passed
/item_totals/2/scale_variance_if_deletedtolerance9.67251461988304059.672514619883049.76E-1817passed
/item_totals/2/alpha_if_deletedtolerance0.739218057234985790.73921805723498582.55E-1817passed
/item_totals/3/meantolerance2.7368421052631582.7368421052631583.51E-1716passed
/item_totals/3/sdtolerance1.04573765900534931.04573765900534931.82E-1716passed
/item_totals/3/corrected_item_totaltolerance0.620579922173684410.62057992217368443.76E-1817passed
/item_totals/3/scale_mean_if_deletedtolerance11.26315789473684211.2631578947368424.79E-1616passed
/item_totals/3/scale_variance_if_deletedtolerance10.204678362573110.20467836257319.66E-1717passed
/item_totals/3/alpha_if_deletedtolerance0.781661891117478410.78166189111747843.45E-1817passed
/item_totals/4/meantolerance2.68421052631578942.68421052631578941.98E-1717passed
/item_totals/4/sdtolerance0.945905302926917320.94590530292691734.06E-1817passed
/item_totals/4/corrected_item_totaltolerance0.71428877058078910.71428877058078883.32E-1615passed
/item_totals/4/scale_mean_if_deletedtolerance11.31578947368421111.315789473684213.80E-1616passed
/item_totals/4/scale_variance_if_deletedtolerance10.22807017543859610.2280701754385961.32E-1616passed
/item_totals/4/alpha_if_deletedtolerance0.757004002287021140.75700400228702111.75E-1817passed
/tests/[name=Friedman's chi-square]/statistictolerance3.31182795698924753.311827956989213.73E-1413passed
/tests/[name=Friedman's chi-square]/p_valuetail_probability0.507060407766553590.50706040776655976.11E-1513passed
/tests/[name=Hotelling's T-squared (equal item means)]/statistictolerance6.05233514429546566.0523351442954747.99E-1514passed
/tests/[name=Hotelling's T-squared (equal item means)]/p_valuetail_probability0.328357717337528050.328357717337527673.84E-1614passed
/anova/0/1tolerance55.59999999999999455.600000000000011.45E-1415passed
/anova/1/1tolerance1.5157894736842091.51578947368421152.46E-1514passed
/anova/2/1tolerance40.08421052631578640.08421052631581.44E-1415passed
/anova/2/3tolerance0.55672514619883040.55672514619883051.12E-1615passed
/icc/0/estimatetolerance0.481497894970086270.481497894970086441.67E-1615passed
/icc/0/lowertolerance0.309485900572547660.309485900572547932.74E-1615passed
/icc/0/uppertolerance0.672487742360179610.67248774236017971.09E-1615passed
/icc/0/f_valuetolerance5.64316239316239095.6431623931623931.75E-1515passed
/icc/0/df1tolerance1818.00.00E+2passed
/icc/0/df2tolerance7676.00.00E+2passed
/icc/0/p_valuetail_probability3.4707763146305469e-083.470776314630516e-083.11E-2214passed
/icc/1/estimatetolerance0.822794396062097590.82279439606209771.09E-1615passed
/icc/1/lowertolerance0.691451518831476490.69145151883147694.40E-1615passed
/icc/1/uppertolerance0.911242111249662260.91124211124966233.39E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

247  outputs <- c(scale_core(m, c("i2", "i5"), c("i1", "i3", "i4")),
248               icc_outputs(m, "one_way", "consistency", 0.90))
249  emit("STAT-DESC-SCALE-REL-003", "Scale reliability with a declared split and one-way ICC at 90%",
250       SCALE, "scaleReliabilityMenuItem", SCALE_REQ, "DESC_SCALE_B7",
251       list(items = as.list(ITEMS), reverse_items = list("i4"), scale_low = 1, scale_high = 5,
252            split = "custom", first_half = list("i2", "i5"), icc_model = "one_way",
253            icc_type = "consistency", confidence_level = 0.90),
254       outputs,
255        c(paste0(MANUAL_SCALE, ":66-72: one-way ICC(1,1) = (MSR - MSW)/(MSR + (k-1) MSW) and ICC(1,k) = (MSR - MSW)/MSR; F-pivot intervals."),
256          "psych::ICC(alpha = 0.10, lmer = FALSE) ICC1 / ICC1k; the declared first half i2, i5 keeps the items' declared order within each half.",
257          ICC_PSYCH_NOTE,
258          GUTTMAN_L6_NOTE),
259       SCALE_SOURCE, SCALE_CITATION, kind = "axis_coverage")

Whole script generators/r/descriptives_reliability_agreement.R


Scale reliability of four items without reversal or ICC

Case STAT-DESC-SCALE-REL-004 · status: passed

Reference

Evidence type
R
Source
psych::alpha, psych::ICC, stats::friedman.test and base R covariance algebra
Citation
Revelle, W. (2024). psych: procedures for psychological research. R package version 2.6.5; Cronbach (1951); Guttman (1945) Psychometrika 10:255-282; McGraw and Wong (1996) Psychological Methods 1:30-46; Shrout and Fleiss (1979).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, psych 2.6.5, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/descriptives_reliability_agreement.R · sha256 6f498a9f8544f55b
Dataset
reference_data/datasets/DESC_SCALE_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact19190.00E+2passed
/kexact440.00E+2passed
/alphatolerance0.781661891117478410.78166189111747843.45E-1817passed
/standardized_alphatolerance0.792421204769963920.79242120476996393.91E-1817passed
/guttman/0tolerance0.586246418338108870.58624641833810894.40E-1817passed
/guttman/1tolerance0.795659994976847250.79565999497684731.28E-1817passed
/guttman/2tolerance0.781661891117478410.78166189111747843.45E-1817passed
/guttman/3tolerance0.887106017191977130.88710601719197714.66E-1817passed
/guttman/4tolerance0.790692849461649640.79069284946164961.66E-1817passed
/guttman/5tolerance0.806317079547831720.80631707954783161.14E-1615passed
/split_half/correlationtolerance0.801463856276855240.80146385627685511.16E-1615passed
/split_half/spearman_brown_equaltolerance0.88979176960371220.88979176960371211.10E-1615passed
/split_half/spearman_brown_unequaltolerance0.889791769603712310.88979176960371212.20E-1615passed
/split_half/guttman_split_halftolerance0.887106017191977130.88710601719197692.27E-1615passed
/summaries/0/meantolerance2.81578947368421062.81578947368421061.98E-1717passed
/summaries/3/meantolerance0.488324050509884590.488324050509884645.40E-1715passed
/split_half/first_alphatolerance0.174712643678160750.174712643678160972.23E-1614passed
/split_half/second_alphatolerance0.787313432835820890.78731343283582092.21E-1817passed
/item_totals/0/meantolerance2.89473684210526332.89473684210526332.52E-1717passed
/item_totals/0/sdtolerance1.14962490704602651.14962490704602651.04E-1717passed
/item_totals/0/corrected_item_totaltolerance0.390693103699399140.3906931036993991.15E-1615passed
/item_totals/0/scale_mean_if_deletedtolerance8.36842105263157888.3684210526315793.96E-1717passed
/item_totals/0/scale_variance_if_deletedtolerance6.57894736842105226.5789473684210521.09E-1717passed
/item_totals/0/alpha_if_deletedtolerance0.838666666666666670.83866666666666671.40E-1817passed
/item_totals/1/meantolerance33.00.00E+2passed
/item_totals/1/sdtolerance11.00.00E+2passed
/item_totals/1/corrected_item_totaltolerance0.521453864094656770.52145386409465671.09E-1615passed
/item_totals/1/scale_mean_if_deletedtolerance8.26315789473684258.2631578947368422.08E-1717passed
/item_totals/1/scale_variance_if_deletedtolerance6.53801169590643296.5380116959064334.07E-1717passed
/item_totals/1/alpha_if_deletedtolerance0.760733452593917780.76073345259391762.22E-1615passed
/item_totals/2/meantolerance2.68421052631578942.68421052631578941.98E-1717passed
/item_totals/2/sdtolerance1.00291971424255811.0029197142425582.41E-1817passed
/item_totals/2/corrected_item_totaltolerance0.760785100763442410.76078510076344222.20E-1615passed
/item_totals/2/scale_mean_if_deletedtolerance8.57894736842105318.5789473684210539.59E-1918passed
/item_totals/2/scale_variance_if_deletedtolerance5.59064327485380115.59064327485389.07E-1615passed
/item_totals/2/alpha_if_deletedtolerance0.637029288702928960.63702928870292854.40E-1615passed
/item_totals/3/meantolerance2.68421052631578942.68421052631578941.98E-1717passed
/item_totals/3/sdtolerance0.945905302926917320.94590530292691734.06E-1817passed
/item_totals/3/corrected_item_totaltolerance0.735356829100904160.73535682910090423.53E-1817passed
/item_totals/3/scale_mean_if_deletedtolerance8.57894736842105318.5789473684210539.59E-1918passed
/item_totals/3/scale_variance_if_deletedtolerance5.92397660818713415.9239766081871341.86E-1717passed
/item_totals/3/alpha_if_deletedtolerance0.657453109575518280.65745310957551812.22E-1615passed
/tests/[name=Friedman's chi-square]/statistictolerance3.88372093023255793.88372093023255532.63E-1515passed
/tests/[name=Friedman's chi-square]/p_valuetail_probability0.274297015089319090.27429701508931987.17E-1614passed
/tests/[name=Hotelling's T-squared (equal item means)]/statistictolerance5.81333333333333265.8133333333333374.42E-1515passed
/tests/[name=Hotelling's T-squared (equal item means)]/p_valuetail_probability0.202589059666855840.202589059666855424.20E-1614passed
/anova/0/1tolerance45.92105263157890245.921052631578955.02E-1414passed
/anova/1/1tolerance1.42105263157894511.42105263157894873.58E-1514passed
/anova/2/1tolerance30.07894736842105830.0789473684210441.38E-1415passed
/anova/2/3tolerance0.557017543859649190.5570175438596492.23E-1615passed
/icc/0absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

262  m4 <- adjusted(c("i1", "i2", "i3", "i5"))
263  outputs <- c(scale_core(m4, c("i1", "i2"), c("i3", "i5")),
264               list("/icc/0" = list(comparison = "absent", rationale = "icc_model none leaves the ICC table empty")))
265  emit("STAT-DESC-SCALE-REL-004", "Scale reliability of four items without reversal or ICC",
266       SCALE, "scaleReliabilityMenuItem", SCALE_REQ, "DESC_SCALE_B7",
267       list(items = list("i1", "i2", "i3", "i5"), reverse_items = list(), split = "halves",
268            first_half = list(), icc_model = "none", icc_type = "consistency",
269            confidence_level = 0.95),
270       outputs,
271        c("No reverse coding; halves i1, i2 and i3, i5; icc_model none leaves the ICC table empty.",
272          GUTTMAN_L6_NOTE),
273       SCALE_SOURCE, SCALE_CITATION, kind = "axis_coverage")

Whole script generators/r/descriptives_reliability_agreement.R


Score a target with a frozen OLS artifact (mean)

Case STAT-DESC-SCORE-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_TARGET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/predictionssequence['7.2197878156661623', '7.547346447130268', '9.4486969658988[7.219787815666163, 7.547346447130272, 9.44869696589881, 10.passed
/lowersequence['6.4521248456942617', '6.8999033557116123', '8.398392801404[6.452124845694263, 6.899903355711616, 8.398392801404185, 9.passed
/uppersequence['7.987450785638063', '8.1947895385489247', '10.499001130393[7.987450785638064, 8.194789538548928, 10.499001130393435, 1passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

314  score_case("STAT-DESC-SCORE-001", "Score a target with a frozen OLS artifact (mean)", ols, "ols", "identity")

Whole script generators/r/descriptives_model_scoring.R


Score a target with a frozen OLS artifact (individual)

Case STAT-DESC-SCORE-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_TARGET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/predictionssequence['7.2197878156661623', '7.547346447130268', '9.4486969658988[7.219787815666163, 7.547346447130272, 9.44869696589881, 10.passed
/lowersequence['5.134655536448621', '5.5034157381788589', '7.2437912934666[5.134655536448621, 5.503415738178862, 7.243791293466627, 7.passed
/uppersequence['9.3049200948837036', '9.5912771560816772', '11.65360263833[9.304920094883705, 9.59127715608168, 11.653602638330993, 12passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

315  score_case("STAT-DESC-SCORE-002", "Score a target with a frozen OLS artifact (individual)", ols, "ols", "identity", "individual")

Whole script generators/r/descriptives_model_scoring.R


Score a target with a frozen WLS artifact

Case STAT-DESC-SCORE-003 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_TARGET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/predictionssequence['7.2306945355702243', '7.5060744815531635', '9.393682040293[7.23069453557023, 7.506074481553167, 9.393682040293195, 9.8passed
/lowersequence['6.5148130933061825', '6.8133408296695643', '8.385372989442[6.514813093306186, 6.813340829669567, 8.385372989442352, 9.passed
/uppersequence['7.9465759778342662', '8.1988081334367617', '10.40199109114[7.946575977834273, 8.198808133436767, 10.401991091144037, 1passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

316  score_case("STAT-DESC-SCORE-003", "Score a target with a frozen WLS artifact", wls, "wls", "identity",
317             extra = list(weight_column = "w"), notes = "The target needs no weight column to be scored.")

Whole script generators/r/descriptives_model_scoring.R


Score a target with a frozen logistic artifact

Case STAT-DESC-SCORE-004 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_TARGET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/predictionssequence['0.56530586934621607', '0.73392828917432984', '0.9224419489[0.5653058693462151, 0.7339282891743292, 0.9224419489983366,passed
/lowersequence['0.1383143552738369', '0.34076889503247665', '0.35531187244[0.13831435527383562, 0.34076889503247354, 0.355311872440902passed
/uppersequence['0.91331580226586473', '0.93638399592839727', '0.9961189705[0.9133158022658652, 0.9363839959283979, 0.996118970513382, passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

318  score_case("STAT-DESC-SCORE-004", "Score a target with a frozen logistic artifact", logit, "logistic", "logit",
319             extra = list(event_value = "yes", reference_value = "no"))

Whole script generators/r/descriptives_model_scoring.R


Score a target with a frozen Poisson artifact and exposure

Case STAT-DESC-SCORE-005 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_TARGET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/predictionssequence['11.825295455790158', '8.2537107396756717', '20.60416446653[11.825295455790132, 8.253710739675657, 20.604164466535728, passed
/lowersequence['8.6346675577179006', '6.2328133322615829', '13.43601252975[8.634667557717886, 6.232813332261573, 13.43601252975934, 13passed
/uppersequence['16.194904051834708', '10.929854199486954', '31.59654640271[16.194904051834666, 10.929854199486929, 31.596546402718896,passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

320  score_case("STAT-DESC-SCORE-005", "Score a target with a frozen Poisson artifact and exposure", pois, "poisson", "log",
321             extra = list(exposure_column = "expo"), offset = log(t$expo[scored]),
322             notes = "Predictions are expected counts at the target's entered exposure.")

Whole script generators/r/descriptives_model_scoring.R


An unseen categorical level is rejected

Case STAT-DESC-SCORE-006 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_TARGET_UNSEEN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

324  unseen <- read_dataset("DESC_SCORE_TARGET")
325  unseen$grp[3] <- "D"
326  export_dataset(unseen, "DESC_SCORE_TARGET_UNSEEN", "DESC_SCORE_TARGET with an unseen level D in row 3 (1-based)",
327                 "Hand-constructed", "Constructed validation data; no third-party license.",
328                 c(x1 = "Float64", grp = "String", expo = "Float64"))
329  emit("STAT-DESC-SCORE-006", score_entry, "An unseen categorical level is rejected", "DESC_SCORE_TARGET_UNSEEN",
330       list(artifact = artifact_of(ols, "ols", "identity")), list(),
331       c("Manual line 94-95: an unknown categorical level in a scoreable row is rejected explicitly."),
332       kind = "invalid_input", error = list(error_class = "ValueError", message_contains = "D"))

Whole script generators/r/descriptives_model_scoring.R


A target without a fitted predictor is refused

Case STAT-DESC-SCORE-007 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_TARGET_NO_X1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

333  no_x <- read_dataset("DESC_SCORE_TARGET")[c("grp", "expo")]
334  export_dataset(no_x, "DESC_SCORE_TARGET_NO_X1", "DESC_SCORE_TARGET without the x1 column",
335                 "Hand-constructed", "Constructed validation data; no third-party license.",
336                 c(grp = "String", expo = "Float64"))
337  emit("STAT-DESC-SCORE-007", score_entry, "A target without a fitted predictor is refused", "DESC_SCORE_TARGET_NO_X1",
338       list(artifact = artifact_of(ols, "ols", "identity")), list(),
339       c("Manual line 90-91: the target needs every fitted predictor."), kind = "invalid_input",
340       error = list(error_class = "ValueError", message_contains = "x1"))

Whole script generators/r/descriptives_model_scoring.R


OLS fit with mean-response intervals and influence

Case STAT-DESC-SCOREFIT-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficient_namessequence['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']passed
/artifact/coefficientssequence['1.4496541841259485', '0.94245630030978667', '1.47243773579[1.4496541841259485, 0.9424563003097871, 1.4724377357958096,passed
/artifact/covariance/0sequence['0.26438303400691682', '-0.033670376837922074', '-0.1145835[0.2643830340069169, -0.03367037683792208, -0.11458352745500passed
/artifact/covariance/1sequence['-0.033670376837922074', '0.0064623962300721788', '0.004919[-0.03367037683792208, 0.006462396230072183, 0.0049191760103passed
/artifact/covariance/2sequence['-0.11458352745500151', '0.0049191760103309487', '0.1816517[-0.11458352745500165, 0.004919176010330953, 0.1816517499910passed
/artifact/covariance/3sequence['-0.11394379029508098', '0.004796390481959574', '0.09260464[-0.11394379029508112, 0.004796390481959576, 0.0926046490408passed
/artifact/design_rankexact440.00E+2passed
/artifact/residual_dfexact26260.00E+2passed
/artifact/residual_variancetolerance0.889536366059752730.88953636605975346.62E-1615passed
/predictionssequence['5.3894425141327797', '6.8716052798081506', '1.577301945711[5.3894425141327815, 6.871605279808154, 1.57730194571147, 4.passed
/lowersequence['4.7057843798594607', '6.2520151581426155', '0.768674159642[4.7057843798594625, 6.252015158142619, 0.7686741596423843, passed
/uppersequence['6.0731006484060988', '7.4911954014736857', '2.385929731780[6.0731006484061005, 7.491195401473689, 2.385929731780556, 5passed
/influence/0/leveragetolerance0.124356033540771210.124356033540771174.13E-1715passed
/influence/0/studentized_residualtolerance-1.8349316966202853-1.83493169662028934.02E-1514passed
/influence/0/cooks_distancetolerance0.119541614202626790.119541614202627274.77E-1614passed
/influence/0/dfbetassequence['-0.26199469388673396', '0.32163114812840865', '-0.40987677[-0.26199469388673424, 0.32163114812840915, -0.4098767733542passed
/influence/1/leveragetolerance0.102140472008479350.102140472008479269.50E-1715passed
/influence/1/studentized_residualtolerance0.014540551564178470.0145405515641766661.80E-1512passed
/influence/1/cooks_distancetolerance6.0130004334173899e-066.0130004334158915e-061.50E-1812passed
/influence/1/dfbetassequence['0.0021930242248506481', '0.0006961720057348849', '-0.00322[0.0021930242248503744, 0.0006961720057347992, -0.0032298914passed
/influence/2/leveragetolerance0.1739747461013430.173974746101343111.14E-1615passed
/influence/2/studentized_residualtolerance1.20486085498235721.20486085498235674.67E-1615passed
/influence/2/cooks_distancetolerance0.0764375369521398950.076437536952139872.77E-1715passed
/influence/2/dfbetassequence['0.29639820587478877', '-0.36386574798313431', '-0.05224165[0.29639820587478877, -0.3638657479831343, -0.05224165822026passed
/influence/3/leveragetolerance0.156469691311965890.156469691311965955.72E-1715passed
/influence/3/studentized_residualtolerance0.534487669015347060.53448766901534932.22E-1514passed
/influence/3/cooks_distancetolerance0.013247835387782540.0132478353877826531.13E-1614passed
/influence/3/dfbetassequence['0.11107386992797576', '-0.1363570222816993', '0.1074016662[0.1110738699279762, -0.13635702228169988, 0.107401666206341passed
/influence/4/leveragetolerance0.103991174315749320.103991174315749364.29E-1715passed
/influence/4/studentized_residualtolerance-1.1544912631849256-1.1544912631849257.01E-1615passed
/influence/4/cooks_distancetolerance0.0386727900153034720.0386727900153034452.74E-1715passed
/influence/4/dfbetassequence['-0.28840892032022025', '0.077570282418737938', '0.28284799[-0.28840892032022025, 0.07757028241873792, 0.28284799037219passed
/influence/5/leveragetolerance0.12033395025541280.120333950255412822.49E-1715passed
/influence/5/studentized_residualtolerance-1.1184316704427275-1.11843167044272641.10E-1515passed
/influence/5/cooks_distancetolerance0.0427788656424586970.042778865642458627.67E-1714passed
/influence/5/dfbetassequence['-0.13921499629656628', '0.1709037630926768', '0.0245373356[-0.13921499629656603, 0.17090376309267657, 0.02453733562318passed
/influence/6/leveragetolerance0.109096829353573440.109096829353573446.91E-1917passed
/influence/6/studentized_residualtolerance0.269833898633412510.26983389863341232.26E-1615passed
/influence/6/cooks_distancetolerance0.00222902351232399760.0022290235123239934.79E-1814passed
/influence/6/dfbetassequence['-0.021809863956073289', '0.026774326917280513', '0.0659646[-0.021809863956073268, 0.02677432691728048, 0.0659646908510passed
/influence/7/leveragetolerance0.127914550862687880.127914550862687881.33E-1816passed
/influence/7/studentized_residualtolerance0.534037347127885090.53403734712788521.08E-1615passed
/influence/7/cooks_distancetolerance0.0104578926221587460.010457892622158753.75E-1815passed
/influence/7/dfbetassequence['0.18016711217113252', '-0.094207575303941249', '-0.1383021[0.18016711217113254, -0.09420757530394126, -0.1383021902592passed
/influence/8/leveragetolerance0.179980228069976490.179980228069976545.44E-1715passed
/influence/8/studentized_residualtolerance1.11530529363459821.1153052936345982.39E-1615passed
/influence/8/cooks_distancetolerance0.0682539845028327650.068253984502832724.14E-1715passed
/influence/8/dfbetassequence['-0.28512562958559567', '0.35002725530043266', '0.050254810[-0.28512562958559556, 0.3500272553004325, 0.050254810573784passed
/influence/9/leveragetolerance0.107707336565510360.107707336565510381.54E-1715passed
/influence/9/studentized_residualtolerance-0.61021641484846989-0.61021641484846917.77E-1614passed
/influence/9/cooks_distancetolerance0.0112368800545255240.0112368800545254982.59E-1714passed
/influence/9/dfbetassequence['-0.045628113508801985', '0.056014197528419581', '-0.133149[-0.04562811350880194, 0.056014197528419546, -0.133149144099passed
/influence/10/leveragetolerance0.119108406297620870.11910840629762093.01E-1715passed
/influence/10/studentized_residualtolerance0.811886115942271560.81188611594227193.31E-1615passed
/influence/10/cooks_distancetolerance0.0222817956037430630.0222817956037430761.34E-1715passed
/influence/10/dfbetassequence['0.060853251156399862', '0.11876968598699685', '-0.17307985[0.06085325115639993, 0.11876968598699686, -0.17307985782612passed
/influence/11/leveragetolerance0.107091592369123910.107091592369123942.91E-1715passed
/influence/11/studentized_residualtolerance1.23119527115794151.23119527115794152.68E-1716passed
/influence/11/cooks_distancetolerance0.0454508856384894380.045450885638489462.11E-1715passed
/influence/11/dfbetassequence['0.090314436206185345', '-0.11087222942814014', '-0.0159183[0.0903144362061853, -0.11087222942814018, -0.01591836865109passed
/influence/12/leveragetolerance0.167271913270623880.167271913270623881.29E-1817passed
/influence/12/studentized_residualtolerance-1.1715081999470958-1.1715081999470961.85E-1615passed
/influence/12/cooks_distancetolerance0.0689208278881501820.068920827888150212.81E-1715passed
/influence/12/dfbetassequence['0.27327773338541284', '-0.33548248570514283', '-0.33439647[0.27327773338541284, -0.33548248570514283, -0.3343964759665passed
/influence/13/leveragetolerance0.174681166760074890.174681166760074955.63E-1715passed
/influence/13/studentized_residualtolerance-1.7265305745225181-1.72653057452251611.97E-1514passed
/influence/13/cooks_distancetolerance0.157729484613046810.157729484613046562.49E-1614passed
/influence/13/dfbetassequence['-0.80417975460810887', '0.54124618883967346', '0.515988585[-0.8041797546081082, 0.5412461888396729, 0.515988585569437,passed
/influence/14/leveragetolerance0.172223381184271940.172223381184272028.30E-1715passed
/influence/14/studentized_residualtolerance-2.2758913109707288-2.2758913109707293.14E-1716passed
/influence/14/cooks_distancetolerance0.269415141617150080.26941514161715021.09E-1615passed
/influence/14/dfbetassequence['0.60005983798191775', '-0.73664825715633919', '-0.10576353[0.6000598379819181, -0.7366482571563392, -0.105763531445932passed
/influence/15/leveragetolerance0.182506787294603620.182506787294603685.84E-1715passed
/influence/15/studentized_residualtolerance0.55291156817421050.55291156817421159.98E-1614passed
/influence/15/cooks_distancetolerance0.0170626399369898410.0170626399369899036.23E-1714passed
/influence/15/dfbetassequence['0.14113818734189515', '-0.17326471985406383', '0.108607124[0.14113818734189537, -0.17326471985406422, 0.10860712404405passed
/influence/16/leveragetolerance0.108707619681721990.1087076196817221.36E-1715passed
/influence/16/studentized_residualtolerance1.08663936920407011.08663936920407011.58E-1716passed
/influence/16/cooks_distancetolerance0.036003993314768980.036003993314768996.68E-1815passed
/influence/16/dfbetassequence['0.12408461614468061', '0.10779560383624696', '-0.240154275[0.12408461614468067, 0.10779560383624696, -0.24015427547930passed
/influence/17/leveragetolerance0.106162389363432060.106162389363432115.36E-1715passed
/influence/17/studentized_residualtolerance-0.24557763787301159-0.245577637873011322.73E-1614passed
/influence/17/cooks_distancetolerance0.0017907283275799310.00179072832757992773.30E-1814passed
/influence/17/dfbetassequence['-0.016306324009364093', '0.020018045537796167', '0.0028740[-0.016306324009364055, 0.020018045537796143, 0.002874070722passed
/influence/18/leveragetolerance0.158678367400802210.158678367400802275.71E-1715passed
/influence/18/studentized_residualtolerance1.56015658985830831.56015658985830834.14E-1817passed
/influence/18/cooks_distancetolerance0.114770970991251210.114770970991251265.16E-1715passed
/influence/18/dfbetassequence['0.34569130136507559', '-0.42437916778620233', '0.326753500[0.34569130136507564, -0.4243791677862025, 0.326753500080154passed
/influence/19/leveragetolerance0.124912541454195590.124912541454195612.36E-1715passed
/influence/19/studentized_residualtolerance-0.53132717108369842-0.53132717108369813.29E-1615passed
/influence/19/cooks_distancetolerance0.010074387337155870.0100743873371558618.86E-1815passed
/influence/19/dfbetassequence['-0.030719882408610271', '-0.088389225098364374', '0.111232[-0.030719882408610316, -0.08838922509836432, 0.111232734887passed
/influence/20/leveragetolerance0.117162383801421620.117162383801421631.23E-1715passed
/influence/20/studentized_residualtolerance0.0928305818346956690.092830581834695343.33E-1614passed
/influence/20/cooks_distancetolerance0.000285910117192680210.00028591011719267812.12E-1814passed
/influence/20/dfbetassequence['-0.010340205539492416', '0.012693891354106833', '0.0018225[-0.010340205539492385, 0.012693891354106788, 0.001822512663passed
/influence/21/leveragetolerance0.17773041083025840.177730410830258373.26E-1715passed
/influence/21/studentized_residualtolerance0.26687822767441660.26687822767441597.20E-1614passed
/influence/21/cooks_distancetolerance0.00384869782587452270.0038486978258745012.17E-1714passed
/influence/21/dfbetassequence['-0.065631888837938773', '0.080571325501379576', '0.0755189[-0.06563188883793858, 0.08057132550137931, 0.07551890974105passed
/influence/22/leveragetolerance0.108392380970956940.108392380970956962.41E-1715passed
/influence/22/studentized_residualtolerance0.393239634919591740.39323963491959191.67E-1615passed
/influence/22/cooks_distancetolerance0.00469980200664483520.00469980200664483954.31E-1815passed
/influence/22/dfbetassequence['0.044565013765092712', '0.037522492855971465', '-0.0852509[0.04456501376509277, 0.03752249285597148, -0.08525092020644passed
/influence/23/leveragetolerance0.11018434914542380.110184349145423855.44E-1715passed
/influence/23/studentized_residualtolerance0.179674107071266080.17967410707126654.12E-1614passed
/influence/23/cooks_distancetolerance0.000999380496136746290.0009993804961367514.77E-1814passed
/influence/23/dfbetassequence['-0.015363480991025974', '0.018860588193930869', '0.0027078[-0.015363480991026014, 0.01886058819393091, 0.0027078899504passed
/influence/24/leveragetolerance0.126976978397087550.126976978397087552.37E-1816passed
/influence/24/studentized_residualtolerance-0.65293521414606248-0.65293521414606195.54E-1615passed
/influence/24/cooks_distancetolerance0.0155017055711315560.0155017055711315272.92E-1714passed
/influence/24/dfbetassequence['0.092440379272324444', '-0.11348208957100837', '-0.1691878[0.09244037927232437, -0.11348208957100826, -0.1691878828404passed
/influence/25/leveragetolerance0.134563749813749630.13456374981374975.90E-1715passed
/influence/25/studentized_residualtolerance-0.66178705043461827-0.66178705043461784.40E-1615passed
/influence/25/cooks_distancetolerance0.0170243107034649090.0170243107034648951.42E-1715passed
/influence/25/dfbetassequence['-0.23558532632282225', '0.13079267350834636', '0.174459225[-0.23558532632282217, 0.13079267350834634, 0.17445922522080passed
/influence/26/leveragetolerance0.102994331858032810.102994331858032853.76E-1715passed
/influence/26/studentized_residualtolerance-0.77975119470783782-0.77975119470783726.61E-1615passed
/influence/26/cooks_distancetolerance0.0174530062261411540.0174530062261411272.73E-1714passed
/influence/26/dfbetassequence['-0.036413671249881423', '0.044702320943694508', '0.0064180[-0.036413671249881346, 0.04470232094369448, 0.0064180906978passed
/influence/27/leveragetolerance0.145441981265533150.145441981265533151.02E-1817passed
/influence/27/studentized_residualtolerance1.14023276543672241.14023276543672042.04E-1514passed
/influence/27/cooks_distancetolerance0.0553191209380563820.05531912093805621.80E-1614passed
/influence/27/dfbetassequence['-0.21548073135107013', '0.26452946048577636', '0.312583942[-0.21548073135106977, 0.2645294604857759, 0.312583942023539passed
/influence/28/leveragetolerance0.126303658674774440.12630365867477455.94E-1715passed
/influence/28/studentized_residualtolerance1.17930673793309841.17930673793309759.18E-1615passed
/influence/28/cooks_distancetolerance0.0502630666713385630.050263066671338524.14E-1715passed
/influence/28/dfbetassequence['0.065255586140101934', '0.20624125543150101', '-0.25179263[0.06525558614010198, 0.20624125543150093, -0.25179263862810passed
/influence/29/leveragetolerance0.122940597780820590.122940597780820633.91E-1715passed
/influence/29/studentized_residualtolerance0.60350890778529820.6035089077852997.76E-1614passed
/influence/29/cooks_distancetolerance0.0127636148277392460.0127636148277392853.86E-1714passed
/influence/29/dfbetassequence['0.078515012125579736', '-0.096386965402373778', '-0.013838[0.07851501212557978, -0.09638696540237389, -0.0138386614605passed
/influence_eligibleexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

138  ols <- lm(y ~ x1 + grp, data = d)
139  emit("STAT-DESC-SCOREFIT-001", fit_entry, "OLS fit with mean-response intervals and influence", "DESC_SCORE_FIT",
140       config("ols", "y"), c(fit_outputs(ols, closed_tol, "mean"), influence_outputs(ols)),
141       c("lm(y ~ x1 + grp) with levels A < B < C; predict(interval = 'confidence')."))

Whole script generators/r/descriptives_model_scoring.R


OLS individual prediction intervals

Case STAT-DESC-SCOREFIT-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficient_namessequence['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']passed
/artifact/coefficientssequence['1.4496541841259485', '0.94245630030978667', '1.47243773579[1.4496541841259485, 0.9424563003097871, 1.4724377357958096,passed
/artifact/covariance/0sequence['0.26438303400691682', '-0.033670376837922074', '-0.1145835[0.2643830340069169, -0.03367037683792208, -0.11458352745500passed
/artifact/covariance/1sequence['-0.033670376837922074', '0.0064623962300721788', '0.004919[-0.03367037683792208, 0.006462396230072183, 0.0049191760103passed
/artifact/covariance/2sequence['-0.11458352745500151', '0.0049191760103309487', '0.1816517[-0.11458352745500165, 0.004919176010330953, 0.1816517499910passed
/artifact/covariance/3sequence['-0.11394379029508098', '0.004796390481959574', '0.09260464[-0.11394379029508112, 0.004796390481959576, 0.0926046490408passed
/artifact/design_rankexact440.00E+2passed
/artifact/residual_dfexact26260.00E+2passed
/artifact/residual_variancetolerance0.889536366059752730.88953636605975346.62E-1615passed
/predictionssequence['5.3894425141327797', '6.8716052798081506', '1.577301945711[5.3894425141327815, 6.871605279808154, 1.57730194571147, 4.passed
/lowersequence['3.3337531818340156', '4.8363259119486974', '-0.52325728524[3.333753181834016, 4.8363259119487, -0.5232572852486708, 2.passed
/uppersequence['7.4451318464315435', '8.9068846476676029', '3.677861176671[7.445131846431547, 8.906884647667608, 3.6778611766716107, 6passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

142  emit("STAT-DESC-SCOREFIT-002", fit_entry, "OLS individual prediction intervals", "DESC_SCORE_FIT",
143       config("ols", "y", "individual"), fit_outputs(ols, closed_tol, "individual"),
144       c("predict(interval = 'prediction') adds s^2 to the mean variance (manual line 62)."))

Whole script generators/r/descriptives_model_scoring.R


WLS fit with mean-response intervals

Case STAT-DESC-SCOREFIT-003 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficient_namessequence['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']passed
/artifact/coefficientssequence['1.5887260014019653', '0.91458245442831509', '1.37239543550[1.5887260014019673, 0.9145824544283148, 1.3723954355092067,passed
/artifact/covariance/0sequence['0.28642861716034251', '-0.0355477060164456', '-0.109239580[0.2864286171603421, -0.03554770601644556, -0.10923958035671passed
/artifact/covariance/1sequence['-0.0355477060164456', '0.0068592489046931217', '0.00135749[-0.03554770601644556, 0.006859248904693112, 0.0013574911526passed
/artifact/covariance/2sequence['-0.10923958035671742', '0.0013574911526055733', '0.1885365[-0.10923958035671738, 0.001357491152605578, 0.1885365493556passed
/artifact/covariance/3sequence['-0.11692142475736224', '0.0028397719717045355', '0.1027664[-0.11692142475736217, 0.0028397719717045363, 0.102766461841passed
/artifact/design_rankexact440.00E+2passed
/artifact/residual_dfexact26260.00E+2passed
/artifact/residual_variancetolerance1.46918899763085741.46918899763085686.48E-1615passed
/predictionssequence['5.3554983026044987', '6.8503188617280619', '1.755089380907[5.355498302604502, 6.850318861728063, 1.7550893809079033, 4passed
/lowersequence['4.6302755556634665', '6.1860382293504825', '0.927269524330[4.63027555566347, 6.186038229350483, 0.9272695243302116, 3.passed
/uppersequence['6.0807210495455308', '7.5145994941056413', '2.582909237485[6.080721049545534, 7.514599494105642, 2.582909237485595, 5.passed
/influence_eligibleexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

145  wls <- lm(y ~ x1 + grp, data = d, weights = w)
146  emit("STAT-DESC-SCOREFIT-003", fit_entry, "WLS fit with mean-response intervals", "DESC_SCORE_FIT",
147       config("wls", "y", extra = list(weight_column = "w")),
148       c(fit_outputs(wls, closed_tol, "mean", family_name = "wls"),
149         list("/influence_eligible" = exact(FALSE, "WLS fits do not report OLS influence (line 80-81)."))),
150       c("lm(y ~ x1 + grp, weights = w): weights are relative precisions (line 64-65)."))

Whole script generators/r/descriptives_model_scoring.R


Binary logistic fit with Wald probability limits

Case STAT-DESC-SCOREFIT-004 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficient_namessequence['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']passed
/artifact/coefficientssequence['-5.0402717125000986', '0.93584500396446113', '2.7494072306[-5.040271712499659, 0.9358450039643886, 2.749407230610945, passed
/artifact/covariance/0sequence['4.7688071082740757', '-0.72834854040780017', '-2.956039349[4.768807108274074, -0.7283485404078, -2.956039349125172, -0passed
/artifact/covariance/1sequence['-0.72834854040780017', '0.12866312609528396', '0.408122228[-0.7283485404078, 0.1286631260952839, 0.40812222863918957, passed
/artifact/covariance/2sequence['-2.9560393491251755', '0.40812222863919012', '2.8023027095[-2.956039349125172, 0.40812222863918957, 2.802302709578001,passed
/artifact/covariance/3sequence['-0.19292380414514609', '-0.079983484568603477', '0.3919926[-0.19292380414514648, -0.07998348456860331, 0.3919926208536passed
/artifact/design_rankexact440.00E+2passed
/predictionssequence['0.53971062251102653', '0.58507556616313416', '0.0047721264[0.5397106225110119, 0.5850755661631395, 0.00477212644588699passed
/lowersequence['0.15938340707420148', '0.22572206756121405', '8.4761279114[0.1593834070741935, 0.22572206756121796, 8.476127911444656epassed
/uppersequence['0.87880673205963689', '0.87212902533358416', '0.2133629108[0.8788067320596307, 0.8721290253335866, 0.2133629108055825,passed
/influence_eligibleexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

151  logit <- glm(I(yb == "yes") ~ x1 + grp, data = d, family = binomial(),
152               control = glm.control(epsilon = 1e-14, maxit = 100))
153  emit("STAT-DESC-SCOREFIT-004", fit_entry, "Binary logistic fit with Wald probability limits", "DESC_SCORE_FIT",
154       config("logistic", "yb", extra = list(event_value = "yes")),
155       c(fit_outputs(logit, glm_tol, "mean", family_name = "logistic"),
156         list("/influence_eligible" = exact(FALSE, "Logistic fits do not report OLS influence."))),
157       c("glm(binomial) with epsilon = 1e-14, maxit = 100; event 'yes'."))

Whole script generators/r/descriptives_model_scoring.R


Poisson fit with an exposure offset

Case STAT-DESC-SCOREFIT-005 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficient_namessequence['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']passed
/artifact/coefficientssequence['0.17363233427100305', '0.17515252719200125', '0.3246003844[0.1736323342995671, 0.1751525271881041, 0.32460038442272976passed
/artifact/covariance/0sequence['0.063524746393292753', '-0.0077718074270232262', '-0.02644[0.0635247464277806, -0.00777180743198422, -0.02644219232530passed
/artifact/covariance/1sequence['-0.0077718074270232262', '0.0013753472171571914', '0.00120[-0.00777180743198422, 0.0013753472179059397, 0.001209448796passed
/artifact/covariance/2sequence['-0.026442192323744628', '0.0012094487963528382', '0.037064[-0.02644219232530354, 0.0012094487961028923, 0.037064847289passed
/artifact/covariance/3sequence['-0.023978983751600978', '0.00077354413880870339', '0.02028[-0.023978983756166233, 0.0007735441391317236, 0.02028808010passed
/artifact/design_rankexact440.00E+2passed
/predictionssequence['6.1203715437494219', '4.0992661975501754', '1.506598556255[6.120371543836556, 4.09926619757536, 1.5065985563060778, 5.passed
/lowersequence['4.5533423563872724', '3.1150886315618789', '1.021991677870[4.553342355827524, 3.115088631533335, 1.021991677806478, 3.passed
/uppersequence['8.2266925923089342', '5.3943837064925173', '2.220995785837[8.226692593554494, 5.39438370660823, 2.2209957861255396, 7.passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

158  pois <- glm(cnt ~ x1 + grp + offset(log(expo)), data = d, family = poisson(),
159              control = glm.control(epsilon = 1e-14, maxit = 100))
160  emit("STAT-DESC-SCOREFIT-005", fit_entry, "Poisson fit with an exposure offset", "DESC_SCORE_FIT",
161       config("poisson", "cnt", extra = list(exposure_column = "expo")),
162       fit_outputs(pois, glm_tol, "mean", family_name = "poisson"),
163       c("glm(poisson) with offset log(expo), epsilon = 1e-14, maxit = 100; predictions are expected counts at the entered exposure."))

Whole script generators/r/descriptives_model_scoring.R


Poisson fit without exposure

Case STAT-DESC-SCOREFIT-006 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficient_namessequence['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']passed
/artifact/coefficientssequence['1.1035973738352829', '0.098058806231247159', '0.2385285152[1.1035973738136189, 0.09805880623531776, 0.2385285152998247passed
/artifact/covariance/0sequence['0.060587330549453491', '-0.0074440842341056976', '-0.02296[0.06058733054819236, -0.007444084233863689, -0.022969698311passed
/artifact/covariance/1sequence['-0.0074440842341056976', '0.0013522470285454438', '0.00061[-0.007444084233863689, 0.0013522470284986542, 0.00061069414passed
/artifact/covariance/2sequence['-0.022969698311983144', '0.00061069414684996205', '0.03627[-0.022969698311764312, 0.0006106941468115813, 0.03627708399passed
/artifact/covariance/3sequence['-0.024106532400062237', '0.00081720450718561456', '0.01997[-0.024106532399851128, 0.000817204507148207, 0.019976904409passed
/artifact/design_rankexact440.00E+2passed
/predictionssequence['4.9473092541844244', '5.3001211886710449', '3.723543078153[4.947309254511192, 5.300121188680342, 3.723543078102783, 4.passed
/lowersequence['3.6422530542224143', '4.0256978394467264', '2.554192482287[3.6422530544940677, 4.025697839454496, 2.5541924822585185, passed
/uppersequence['6.7199803232134858', '6.9779913284451842', '5.428241274302[6.719980323599987, 6.977991328456198, 5.428241274215703, 6.passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

164  pois0 <- glm(cnt ~ x1 + grp, data = d, family = poisson(), control = glm.control(epsilon = 1e-14, maxit = 100))
165  emit("STAT-DESC-SCOREFIT-006", fit_entry, "Poisson fit without exposure", "DESC_SCORE_FIT",
166       config("poisson", "cnt"), fit_outputs(pois0, glm_tol, "mean", family_name = "poisson"),
167       c("No exposure is a zero offset (model_scoring.ModelScoringConfig)."))

Whole script generators/r/descriptives_model_scoring.R


OLS at a 90% confidence level

Case STAT-DESC-SCOREFIT-007 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficient_namessequence['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']passed
/artifact/coefficientssequence['1.4496541841259485', '0.94245630030978667', '1.47243773579[1.4496541841259485, 0.9424563003097871, 1.4724377357958096,passed
/artifact/covariance/0sequence['0.26438303400691682', '-0.033670376837922074', '-0.1145835[0.2643830340069169, -0.03367037683792208, -0.11458352745500passed
/artifact/covariance/1sequence['-0.033670376837922074', '0.0064623962300721788', '0.004919[-0.03367037683792208, 0.006462396230072183, 0.0049191760103passed
/artifact/covariance/2sequence['-0.11458352745500151', '0.0049191760103309487', '0.1816517[-0.11458352745500165, 0.004919176010330953, 0.1816517499910passed
/artifact/covariance/3sequence['-0.11394379029508098', '0.004796390481959574', '0.09260464[-0.11394379029508112, 0.004796390481959576, 0.0926046490408passed
/artifact/design_rankexact440.00E+2passed
/artifact/residual_dfexact26260.00E+2passed
/artifact/residual_variancetolerance0.889536366059752730.88953636605975346.62E-1615passed
/predictionssequence['5.3894425141327797', '6.8716052798081506', '1.577301945711[5.3894425141327815, 6.871605279808154, 1.57730194571147, 4.passed
/lowersequence['4.8221630858334539', '6.357487605523569', '0.9063263729922[4.822163085833456, 6.3574876055235725, 0.9063263729922104, passed
/uppersequence['5.9567219424321056', '7.3857229540927323', '2.248277518430[5.956721942432107, 7.385722954092736, 2.2482775184307298, 5passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

168  emit("STAT-DESC-SCOREFIT-007", fit_entry, "OLS at a 90% confidence level", "DESC_SCORE_FIT",
169       config("ols", "y", level = 0.9), fit_outputs(ols, closed_tol, "mean", level = 0.9),
170       c("A non-default confidence level of 0.90."))

Whole script generators/r/descriptives_model_scoring.R


Missing values and excluded rows are left out of the fit

Case STAT-DESC-SCOREFIT-008 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficient_namessequence['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']['Intercept', 'x1', 'grp = B (vs A)', 'grp = C (vs A)']passed
/artifact/coefficientssequence['1.7057540434087008', '0.87100058102761524', '1.68060597399[1.7057540434087053, 0.8710005810276148, 1.6806059739958343,passed
/artifact/covariance/0sequence['0.30495205974652789', '-0.04168424601215151', '-0.11147165[0.30495205974652817, -0.041684246012151566, -0.111471654434passed
/artifact/covariance/1sequence['-0.04168424601215151', '0.0081917639607258872', '0.0036615[-0.041684246012151566, 0.008191763960725901, 0.003661588462passed
/artifact/covariance/2sequence['-0.11147165443498296', '0.003661588462445103', '0.21384114[-0.11147165443498296, 0.0036615884624450973, 0.213841147630passed
/artifact/covariance/3sequence['-0.13407444059404808', '0.0081034749491491788', '0.0964615[-0.13407444059404822, 0.008103474949149186, 0.0964615827979passed
/artifact/design_rankexact440.00E+2passed
/artifact/residual_dfexact21210.00E+2passed
/artifact/residual_variancetolerance0.835555123100247710.83555512310024879.99E-1614passed
/predictionssequence['6.7166203860605709', '1.6731349157019493', '5.598255640021[6.716620386060573, 1.6731349157019524, 5.598255640021116, 2passed
/lowersequence['6.0707470614718915', '0.84626546588029639', '4.95396502712[6.070747061471892, 0.8462654658802989, 4.953965027125841, 2passed
/uppersequence['7.3624937106492503', '2.5000043655236022', '6.242546252916[7.362493710649253, 2.5000043655236057, 6.242546252916391, 3passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

173  m <- read_dataset("DESC_SCORE_FIT")
174  m$y[c(4, 17)] <- NA
175  m$x1[9] <- NA
176  export_dataset(m, "DESC_SCORE_FIT_MISSING", "DESC_SCORE_FIT with y missing in rows 4 and 17 and x1 in row 9 (1-based)",
177                 "R stats::lm documentation", "Constructed validation data; no third-party license.",
178                 c(x1 = "Float64", grp = "String", w = "Float64", expo = "Float64", y = "Float64",
179                   yb = "String", cnt = "Int64"))
180  mm <- read_dataset("DESC_SCORE_FIT_MISSING")
181  excluded <- c(0L, 21L)
182  kept <- mm[-(excluded + 1L), ]
183  kept <- kept[stats::complete.cases(kept[c("y", "x1", "grp")]), ]
184  kept$grp <- factor(kept$grp, levels = sort(unique(kept$grp)))
185  olsm <- lm(y ~ x1 + grp, data = kept)
186  emit("STAT-DESC-SCOREFIT-008", fit_entry, "Missing values and excluded rows are left out of the fit",
187       "DESC_SCORE_FIT_MISSING", config("ols", "y"), fit_outputs(olsm, closed_tol, "mean", data = kept),
188       c("Rows 0 and 21 (0-based) are excluded; rows with missing y or x1 are omitted and counted (line 50-53).",
189         sprintf("%d rows remain for the fit.", nrow(kept))),
190       kind = "missing_excluded", excluded = excluded)

Whole script generators/r/descriptives_model_scoring.R


Predictor at a 1e6 offset

Case STAT-DESC-SCOREFIT-009 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT_OFFSET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficientssequence['-942454.85065262951', '0.9424563003068136', '1.47243773580[-942454.8506508748, 0.9424563003050592, 1.4724377359817167,passed
/predictionssequence['5.3894425140886852', '6.8716052797390148', '1.577301945681[5.389442514628172, 6.871605280088261, 1.57730194600299, 4.4passed
/lowersequence['4.705784379815702', '6.2520151580608756', '0.7686741595957[4.705784380381216, 6.252015158416386, 0.7686741599606299, 3passed
/uppersequence['6.0731006483616685', '7.491195401417154', '2.3859297317674[6.073100648875128, 7.491195401760136, 2.38592973204535, 5.2passed
/artifact/residual_variancetolerance0.889536366095421080.88953636606677582.86E-1110passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

193  h <- read_dataset("DESC_SCORE_FIT")
194  h$x1 <- h$x1 + 1e6
195  export_dataset(h, "DESC_SCORE_FIT_OFFSET", "DESC_SCORE_FIT with x1 shifted by 1e6",
196                 "R stats::lm documentation", "Constructed validation data; no third-party license.",
197                 c(x1 = "Float64", grp = "String", w = "Float64", expo = "Float64", y = "Float64",
198                   yb = "String", cnt = "Int64"))
199  ho <- read_dataset("DESC_SCORE_FIT_OFFSET")
200  ho$grp <- factor(ho$grp, levels = sort(unique(ho$grp)))
201  olsh <- lm(y ~ x1 + grp, data = ho)
202  out <- fit_outputs(olsh, closed_tol, "mean", data = ho)
203  emit("STAT-DESC-SCOREFIT-009", fit_entry, "Predictor at a 1e6 offset", "DESC_SCORE_FIT_OFFSET",
204       config("ols", "y"), out[c("/artifact/coefficients", "/predictions", "/lower", "/upper", "/artifact/residual_variance")],
205       c("x1 + 1e6 leaves the slope and fitted values unchanged and moves the intercept; lm uses a QR decomposition."),
206       kind = "hard_numeric")

Whole script generators/r/descriptives_model_scoring.R


Three rows leave one residual df and no deletion diagnostics

Case STAT-DESC-SCOREFIT-010 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/coefficientssequence['-0.050000000000001779', '2.0500000000000003'][-0.05000000000000076, 2.05]passed
/artifact/residual_dfexact110.00E+2passed
/predictionssequence['1.9999999999999984', '4.0499999999999989', '8.149999999999[1.9999999999999991, 4.049999999999999, 8.149999999999999]passed
/lowersequence['-0.0090273691365292397', '2.6294031236941908', '5.85935636[-0.009027369136545005, 2.6294031236941793, 5.85935636535893passed
/uppersequence['4.0090273691365264', '5.470596876305807', '10.440643634641[4.009027369136543, 5.470596876305819, 10.44064363464106]passed
/influence/0/leveragetolerance0.714285714285714410.71428571428571431.08E-1615passed
/influence/0/studentized_residualtolerance1.00000000000000071.00000000000000042.56E-1615passed
/influence/0/cooks_distancetolerance1.25000000000000271.25000000000000091.81E-1514passed
/influence/0/deleted_studentized_residualabsentNonepassed
/influence/0/dffitsabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

209  export_dataset(data.frame(x = c(1, 2, 4), y = c(2.1, 3.9, 8.2)), "DESC_SCORE_MIN",
210                 "Constructed: three points for a straight line", "Hand-constructed",
211                 "Constructed validation data; no third-party license.", c(x = "Float64", y = "Float64"))
212  s <- read_dataset("DESC_SCORE_MIN")
213  olss <- lm(y ~ x, data = s)
214  ps <- predict(olss, interval = "confidence")
215  emit("STAT-DESC-SCOREFIT-010", fit_entry, "Three rows leave one residual df and no deletion diagnostics", "DESC_SCORE_MIN",
216       list(family = "ols", response = "y", predictors = list("x")),
217       list("/artifact/coefficients" = seq_tol(coef(olss), "Least-squares line."),
218            "/artifact/residual_df" = exact(1L, "n - p = 3 - 2."),
219            "/predictions" = seq_tol(ps[, "fit"], "Fitted values."),
220            "/lower" = seq_tol(ps[, "lwr"], "t limits with 1 df."),
221            "/upper" = seq_tol(ps[, "upr"], "t limits with 1 df."),
222            "/influence/0/leverage" = tol(hatvalues(olss)[[1]], "Leverage needs no deletion (issue #8)."),
223            "/influence/0/studentized_residual" = tol(rstandard(olss)[[1]], "Internally standardized residual; defined with d = 1 (issue #8)."),
224            "/influence/0/cooks_distance" = tol(cooks.distance(olss)[[1]], "Cook's distance uses the full-fit variance; defined with d = 1 (issue #8)."),
225            "/influence/0/deleted_studentized_residual" = absent("The deleted fit has no residual df."),
226            "/influence/0/dffits" = absent("The deleted fit has no residual df.")),
227       c("With n = p + 1 the full-fit diagnostics are defined (standardized magnitudes are all 1); deleted-variance diagnostics are not (statistician's ruling on issue #8).",
228         "Until 2026-09-25 this case expected every diagnostic but leverage to be absent."),
229       kind = "boundary")

Whole script generators/r/descriptives_model_scoring.R


WLS does not offer individual prediction intervals

Case STAT-DESC-SCOREFIT-011 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

232  emit("STAT-DESC-SCOREFIT-011", fit_entry, "WLS does not offer individual prediction intervals", "DESC_SCORE_FIT",
233       config("wls", "y", "individual", extra = list(weight_column = "w")), list(),
234       c("Manual line 34-36: WLS reports mean uncertainty only."), kind = "invalid_input",
235       error = list(error_class = "ValueError", message_contains = "WLS supports mean-response intervals only"))

Whole script generators/r/descriptives_model_scoring.R


A constant numeric predictor stops the fit

Case STAT-DESC-SCOREFIT-012 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_FIT_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

236  c0 <- read_dataset("DESC_SCORE_FIT")
237  c0$x1 <- 5
238  export_dataset(c0, "DESC_SCORE_FIT_CONSTANT", "DESC_SCORE_FIT with a constant x1",
239                 "Hand-constructed", "Constructed validation data; no third-party license.",
240                 c(x1 = "Float64", grp = "String", w = "Float64", expo = "Float64", y = "Float64",
241                   yb = "String", cnt = "Int64"))
242  emit("STAT-DESC-SCOREFIT-012", fit_entry, "A constant numeric predictor stops the fit", "DESC_SCORE_FIT_CONSTANT",
243       config("ols", "y"), list(),
244       c("Manual line 54-55: a constant numeric predictor stops the fit with an error.",
245         "The message fragment is the product's wording for this error (model_scoring._predictor_specs)."),
246       kind = "invalid_input",
247       error = list(error_class = "ModelScoringExecutionError", message_contains = "must be finite and vary"))

Whole script generators/r/descriptives_model_scoring.R


Score columns from the frozen OLS artifact

Case STAT-DESC-SCOREOP-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_TARGET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pred/0tolerance7.21978781566616237.2197878156661639.33E-1615passed
/lo/0tolerance6.45212484569426176.4521248456942638.89E-1615passed
/hi/0tolerance7.9874507856380637.9874507856380648.77E-1615passed
/pred/1tolerance7.5473464471302687.5473464471302723.59E-1515passed
/lo/1tolerance6.89990335571161236.8999033557116163.54E-1515passed
/hi/1tolerance8.19478953854892478.1947895385489283.52E-1515passed
/pred/2absentNonepassed
/lo/2absentNonepassed
/hi/2absentNonepassed
/pred/3tolerance9.4486969658988089.448696965898811.81E-1515passed
/lo/3tolerance8.39839280140418338.3983928014041851.79E-1515passed
/hi/3tolerance10.49900113039343310.4990011303934351.52E-1515passed
/pred/4absentNonepassed
/lo/4absentNonepassed
/hi/4absentNonepassed
/pred/5tolerance10.00842084195104310.0084208419510495.53E-1515passed
/lo/5tolerance9.21268646152992869.2126864615299388.93E-1515passed
/hi/5tolerance10.80415522237215810.804155222372161.53E-1515passed
/pred/6tolerance6.18968307942756556.18968307942756554.03E-1717passed
/lo/6tolerance5.51511877024381115.5151187702438129.30E-1615passed
/hi/6tolerance6.86424738861131986.8642473886113199.10E-1615passed
/pred/7tolerance5.1553923569440295.15539235694403253.53E-1515passed
/lo/7tolerance4.50696527949508854.5069652794950923.57E-1515passed
/hi/7tolerance5.80381943439296945.8038194343929733.59E-1515passed
/pred/8tolerance10.21764614061981610.217646140619824.02E-1515passed
/lo/8tolerance9.39803788238640979.3980378823864177.10E-1515passed
/hi/8tolerance11.03725439885322311.0372543988532232.36E-1616passed
/pred/9tolerance8.29795782322055868.2979578232205593.74E-1717passed
/lo/9tolerance7.40368842928360327.4036884292836034.37E-1717passed
/hi/9tolerance9.19222721715751419.1922272171575141.86E-1717passed
/pred/10tolerance8.6113796101800188.6113796101800223.52E-1515passed
/lo/10tolerance7.88223017574550247.8822301757455063.60E-1515passed
/hi/10tolerance9.34052904461453449.3405290446145383.52E-1515passed
/pred/11tolerance9.64369025373115649.6436902537311625.33E-1515passed
/lo/11tolerance8.88711823514826988.8871182351482798.90E-1514passed
/hi/11tolerance10.40026227231404310.4002622723140451.76E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

358  emit("STAT-DESC-SCOREOP-001", op_entry, "Score columns from the frozen OLS artifact", "DESC_SCORE_TARGET",
359       list(artifact = artifact_of(ols, "ols", "identity"), names = list("pred", "lo", "hi")), op_out,
360       c("Three columns: prediction, lower and upper; the excluded row 2 and the row with a missing x1 stay null."),
361       kind = "row_alignment", excluded = target_excluded)

Whole script generators/r/descriptives_model_scoring.R


Repeated score column names are refused

Case STAT-DESC-SCOREOP-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm, glm, predict and influence measures
Citation
R stats::lm, stats::glm, stats::predict.lm, stats::predict.glm and stats::influence.measures documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_model_scoring.R · sha256 13a39b7b09c45297
Dataset
reference_data/datasets/DESC_SCORE_TARGET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

362  emit("STAT-DESC-SCOREOP-002", op_entry, "Repeated score column names are refused", "DESC_SCORE_TARGET",
363       list(artifact = artifact_of(ols, "ols", "identity"), names = list("pred", "pred")), list(),
364       c("Score names must be distinct (ModelScoringScoreParameters)."), kind = "invalid_input",
365       error = list(error_class = "ValueError", message_contains = "must be distinct"))

Whole script generators/r/descriptives_model_scoring.R


Descriptive summary of a column with user-missing values masked

Case STAT-DESC-SUMMARY-001 · status: passed

Reference

Evidence type
R
Source
R base mean, median, sd, min, max
Citation
R stats::median and sd documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columnexactscorescorepassed
/statusexactokokpassed
/sample_countexact20200.00E+2passed
/values/meantolerance20.41499999999999920.4151.47E-1617passed
/values/mediantolerance19.519.50.00E+1passed
/values/stdevtolerance5.10966833819141285.1096683381914134.91E-1717passed
/values/sample_variancetolerance26.10871052631579326.1087105263157933.24E-1717passed
/values/mintolerance12.512.50.00E+1passed
/values/maxtolerance33.20000000000000333.21.58E-1617passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

495  ds_source <- "R base mean, median, sd, min, max"
496  ds_cite <- "R stats::median and sd documentation"
497  ds_manual <- "docs/manual/chapters/06-analysis-foundations.tex:4-5: a descriptive summary characterizes the observed data."
498  emit("STAT-DESC-SUMMARY-001", "Descriptive summary of a column with user-missing values masked",
499       "nominal", "operation:descriptives", "descriptiveSummaryMenuItem", "SUMMARY",
500       list(column = "score", metadata = metadata),
501       list("/column" = exact("score", "the summarised column"),
502            "/status" = exact("ok", "a successful summary"),
503            "/sample_count" = exact(n, "finite values after masking 999 and NA"),
504            "/values/mean" = tol(mean(x), "mean"),
505            "/values/median" = tol(stats::median(x), "median (average of the two middle values)"),
506            "/values/stdev" = tol(sd(x), "sample SD (ddof 1)"),
507            "/values/sample_variance" = tol(var(x), "sample variance shown as stdev squared"),
508            "/values/min" = tol(min(x), "minimum"),
509            "/values/max" = tol(max(x), "maximum")),
510       c("R: mean, median, sd, min and max of the 20 finite scores; the declared user-missing 999 is presented as missing, as the desktop masks it before the engine runs.",
511         ds_manual),
512       ds_source, ds_cite)

Whole script generators/r/descriptives_reports.R


Descriptive summary under a frequency weight after exclusions

Case STAT-DESC-SUMMARY-002 · status: passed

Reference

Evidence type
R
Source
R base rep/mean/median/sd (literal replication)
Citation
R stats::median and sd documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nortest 1.0.4, e1071 1.7.17, jsonlite 2.0.0
Reference script
generators/r/descriptives_reports.R · sha256 9e562f8ee4d0f2d0
Dataset
reference_data/datasets/DESC_REPORTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statusexactokokpassed
/sample_countexact18180.00E+2passed
/values/positive_weight_rowstolerance1616.00.00E+2passed
/values/frequency_totaltolerance2525.00.00E+2passed
/values/meantolerance20.68400000000000120.6845.16E-1717passed
/values/mediantolerance19.60000000000000119.64.21E-1616passed
/values/stdevtolerance4.60096728960335094.6009672896033511.51E-1717passed
/values/mintolerance15.30000000000000115.32.89E-1616passed
/values/maxtolerance33.20000000000000333.21.58E-1617passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

514  excluded3 <- c(0L, 1L)
515  kept3 <- !(seq_len(nrow(d)) %in% (excluded3 + 1L))
516  finite3 <- kept3 & valid_score
517  pos3 <- finite3 & w > 0
518  rep3 <- rep(d$score[pos3], w[pos3])
519  emit("STAT-DESC-SUMMARY-002", "Descriptive summary under a frequency weight after exclusions",
520       "axis_coverage", "operation:descriptives", "descriptiveSummaryMenuItem", "SUMMARY",
521       list(column = "score", frequency_weight = "weight", metadata = metadata),
522       list("/status" = exact("ok", "a successful summary"),
523            "/sample_count" = exact(sum(finite3), "finite scores among the non-excluded rows, whatever their weight"),
524            "/values/positive_weight_rows" = tol(sum(pos3), "finite scores with a positive weight"),
525            "/values/frequency_total" = tol(sum(w[pos3]), "W, the total weight of those rows"),
526            "/values/mean" = tol(mean(rep3), "mean of the replicated values"),
527            "/values/median" = tol(stats::median(rep3), "median of the replicated values"),
528            "/values/stdev" = tol(sd(rep3), "SD of the replicated values (W - 1)"),
529            "/values/min" = tol(min(rep3), "minimum"),
530            "/values/max" = tol(max(rep3), "maximum")),
531       c("R: rep(score, weight) over the finite scores of the non-excluded rows with a positive weight (a missing weight counts 0); rows 1 and 2 excluded.",
532         ds_manual),
533       "R base rep/mean/median/sd (literal replication)", ds_cite, excluded = as.list(excluded3))

Whole script generators/r/descriptives_reports.R


Survey mean

Case STAT-DESC-SURVEY-001 · status: passed

Reference

Evidence type
R
Source
R survey/stats/nortest
Citation
R survey::svydesign/svymean, stats::shapiro.test and nortest::ad.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, nortest 1.0.4, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_diagnostics.R · sha256 5427a49203331a27
Dataset
reference_data/datasets/DESC_SURVEY_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance5.80851063829787245.8085106382978722.88E-1818passed
/standard_errortolerance0.657719451279432880.65771945127943281.15E-1615passed
/lowertolerance2.97857224584110242.9785722458411034.06E-1615passed
/uppertolerance8.63844903075464298.6384490307546432.33E-1717passed
/degrees_of_freedomexact220.00E+2passed
/observationsexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

40  raw <- data.frame(
41    response = c(2.0, 3.0, 4.5, 5.0, 5.5, 7.0, 8.0, 9.5),
42    binary = c(0, 0, 1, 1, 0, 1, 1, 1),
43    weight = c(1, 1.5, 2, 0.75, 1.25, 2.5, 1, 1.75),
44    stratum = rep(c("A", "B"), each = 4),
45    psu = rep(c("1", "2"), each = 2, times = 2)
46  )
47  export_dataset(raw, "DESC_SURVEY_01", "Constructed two-stratum one-stage survey",
48                 "R survey::svydesign and survey::svymean documentation",
49                 "Constructed validation data; no third-party license.",
50                 c(response = "Float64", binary = "Float64", weight = "Float64",
51                   stratum = "String", psu = "String"))
52  d <- read_dataset("DESC_SURVEY_01")
53  design <- survey::svydesign(ids = ~psu, strata = ~stratum, weights = ~weight,
54                              data = d, nest = TRUE)
55  for (mode in c("mean", "proportion")) {
56    response <- if (mode == "mean") "response" else "binary"
57    estimate <- survey::svymean(stats::as.formula(paste0("~", response)), design)
58    point <- as.double(coef(estimate)[[1]])
59    standard_error <- sqrt(as.double(stats::vcov(estimate)[1, 1]))
60    df <- as.integer(survey::degf(design))
61    bounds <- point + c(-1, 1) * stats::qt(0.975, df) * standard_error
62    id <- if (mode == "mean") "STAT-DESC-SURVEY-001" else "STAT-DESC-SURVEY-002"
63    emit(id, paste("Survey", mode), "tabelz.executor.survey_descriptives",
64         "surveyDescriptivesMenuItem", "DESC_SURVEY_01",
65         list(response_column = response, weight_column = "weight",
66              stratum_column = "stratum", psu_column = "psu", mode = mode),
67         list("/estimate" = tol(point, "Design-weighted ratio estimate; owner 1e-9 tier."),
68              "/standard_error" = tol(standard_error, "One-stage with-replacement Taylor SE; owner 1e-9 tier."),
69              "/lower" = tol(bounds[1], "Two-sided t interval lower bound; owner 1e-9 tier."),
70              "/upper" = tol(bounds[2], "Two-sided t interval upper bound; owner 1e-9 tier."),
71              "/degrees_of_freedom" = exact(df, "Four PSUs minus two strata give two design degrees of freedom."),
72              "/observations" = exact(nrow(d), "Every response is observed.")),
73         c("Manual chapter 28 lines 25-43: one-stage sampling with replacement, Taylor ratio variance and t interval.",
74           "R svydesign sets nest=TRUE because PSU labels repeat across strata; four original PSUs in two strata."))

Whole script generators/r/descriptives_survey_diagnostics.R


Survey proportion

Case STAT-DESC-SURVEY-002 · status: passed

Reference

Evidence type
R
Source
R survey/stats/nortest
Citation
R survey::svydesign/svymean, stats::shapiro.test and nortest::ad.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, nortest 1.0.4, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_diagnostics.R · sha256 5427a49203331a27
Dataset
reference_data/datasets/DESC_SURVEY_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.680851063829787220.68085106382978722.49E-1817passed
/standard_errortolerance0.233411697175133770.233411697175133778.96E-2018passed
/lowertolerance-0.32343841217625691-0.32343841217625693.57E-1816passed
/uppertolerance1.68514053983583121.68514053983583123.04E-1716passed
/degrees_of_freedomexact220.00E+2passed
/observationsexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

40  raw <- data.frame(
41    response = c(2.0, 3.0, 4.5, 5.0, 5.5, 7.0, 8.0, 9.5),
42    binary = c(0, 0, 1, 1, 0, 1, 1, 1),
43    weight = c(1, 1.5, 2, 0.75, 1.25, 2.5, 1, 1.75),
44    stratum = rep(c("A", "B"), each = 4),
45    psu = rep(c("1", "2"), each = 2, times = 2)
46  )
47  export_dataset(raw, "DESC_SURVEY_01", "Constructed two-stratum one-stage survey",
48                 "R survey::svydesign and survey::svymean documentation",
49                 "Constructed validation data; no third-party license.",
50                 c(response = "Float64", binary = "Float64", weight = "Float64",
51                   stratum = "String", psu = "String"))
52  d <- read_dataset("DESC_SURVEY_01")
53  design <- survey::svydesign(ids = ~psu, strata = ~stratum, weights = ~weight,
54                              data = d, nest = TRUE)
55  for (mode in c("mean", "proportion")) {
56    response <- if (mode == "mean") "response" else "binary"
57    estimate <- survey::svymean(stats::as.formula(paste0("~", response)), design)
58    point <- as.double(coef(estimate)[[1]])
59    standard_error <- sqrt(as.double(stats::vcov(estimate)[1, 1]))
60    df <- as.integer(survey::degf(design))
61    bounds <- point + c(-1, 1) * stats::qt(0.975, df) * standard_error
62    id <- if (mode == "mean") "STAT-DESC-SURVEY-001" else "STAT-DESC-SURVEY-002"
63    emit(id, paste("Survey", mode), "tabelz.executor.survey_descriptives",
64         "surveyDescriptivesMenuItem", "DESC_SURVEY_01",
65         list(response_column = response, weight_column = "weight",
66              stratum_column = "stratum", psu_column = "psu", mode = mode),
67         list("/estimate" = tol(point, "Design-weighted ratio estimate; owner 1e-9 tier."),
68              "/standard_error" = tol(standard_error, "One-stage with-replacement Taylor SE; owner 1e-9 tier."),
69              "/lower" = tol(bounds[1], "Two-sided t interval lower bound; owner 1e-9 tier."),
70              "/upper" = tol(bounds[2], "Two-sided t interval upper bound; owner 1e-9 tier."),
71              "/degrees_of_freedom" = exact(df, "Four PSUs minus two strata give two design degrees of freedom."),
72              "/observations" = exact(nrow(d), "Every response is observed.")),
73         c("Manual chapter 28 lines 25-43: one-stage sampling with replacement, Taylor ratio variance and t interval.",
74           "R svydesign sets nest=TRUE because PSU labels repeat across strata; four original PSUs in two strata."))

Whole script generators/r/descriptives_survey_diagnostics.R


Survey mean, total, proportions and counts (Taylor)

Case STAT-DESC-SURVEY-EST-001 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimates/0/estimatetolerance5.26304438360459955.2630443836045999.14E-1615passed
/estimates/0/standard_errortolerance0.221971468844489180.221971468844489235.37E-1715passed
/estimates/0/lowertolerance4.71989976587296314.719899765872963.54E-1515passed
/estimates/0/uppertolerance5.80618900133623585.8061890013362381.81E-1515passed
/estimates/0/dfexact660.00E+2passed
/estimates/1/estimatetolerance261.82645830000001261.826458299999965.39E-1415passed
/estimates/1/standard_errortolerance39.6489107932880139.648910793288012.68E-1617passed
/estimates/1/lowertolerance164.80906859491392164.809068594913474.51E-1314passed
/estimates/1/uppertolerance358.8438480050861358.843848005086443.43E-1315passed
/estimates/1/dfexact660.00E+2passed
/estimates/0/rowsexact22220.00E+2passed
/estimates/0/missing_rowsexact220.00E+2passed
/estimates/1/rowsexact22220.00E+2passed
/estimates/2/estimatetolerance0.337429876920483820.33742987692048382.24E-1817passed
/estimates/2/standard_errortolerance0.0680367897562399320.068036789756239938.10E-2017passed
/estimates/2/lowertolerance0.170949849752081720.170949849752081056.69E-1614passed
/estimates/2/uppertolerance0.503909904088885920.50390990408888666.64E-1614passed
/estimates/2/dfexact660.00E+2passed
/estimates/3/estimatetolerance18.417818.41782.73E-1616passed
/estimates/3/standard_errortolerance4.31841657941117684.3184165794111774.77E-1716passed
/estimates/3/lowertolerance7.85101529365787347.8510152936578314.26E-1414passed
/estimates/3/uppertolerance28.98458470634212628.984584706342174.27E-1414passed
/estimates/3/dfexact660.00E+2passed
/estimates/4/estimatetolerance0.344833335165419030.344833335165419034.10E-1917passed
/estimates/4/standard_errortolerance0.0932788632970202870.093278863297020291.07E-1917passed
/estimates/4/lowertolerance0.116588179102608630.116588179102607719.16E-1614passed
/estimates/4/uppertolerance0.57307849122822940.57307849122823038.91E-1614passed
/estimates/4/dfexact660.00E+2passed
/estimates/7/estimatetolerance17.342917.34292.05E-1616passed
/estimates/7/standard_errortolerance6.7615114274349436.7615114274349449.33E-1615passed
/estimates/7/lowertolerance0.798077556557302610.79807755655723516.75E-1413passed
/estimates/7/uppertolerance33.88772244344269833.887722443442777.08E-1414passed
/estimates/7/dfexact660.00E+2passed
/design/design_dfexact660.00E+2passed
/design/psusexact880.00E+2passed
/design/strataexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

120  df_t <- as.integer(degf(taylor))
121  mean_x <- svymean(~x, taylor, na.rm = TRUE)
122  total_x <- svytotal(~x, taylor, na.rm = TRUE)
123  prop_g <- svymean(~g, taylor)
124  count_g <- svytotal(~g, taylor)
125  outputs <- c(
126    estimate_outputs("/estimates/0", coef(mean_x)[[1]], se_of(mean_x, 1), df_t, 0.95, "mean of x"),
127    estimate_outputs("/estimates/1", coef(total_x)[[1]], se_of(total_x, 1), df_t, 0.95, "total of x"),
128    list("/estimates/0/rows" = exact(sum(!is.na(d$x)), "rows with x observed"),
129         "/estimates/0/missing_rows" = exact(sum(is.na(d$x)), "rows with x missing"),
130         "/estimates/1/rows" = exact(sum(!is.na(d$x)), "rows with x observed")),
131    estimate_outputs("/estimates/2", coef(prop_g)[["ga"]], se_of(prop_g, "ga"), df_t, 0.95, "proportion g = a"),
132    estimate_outputs("/estimates/3", coef(count_g)[["ga"]], se_of(count_g, "ga"), df_t, 0.95, "count g = a"),
133    estimate_outputs("/estimates/4", coef(prop_g)[["gb"]], se_of(prop_g, "gb"), df_t, 0.95, "proportion g = b"),
134    estimate_outputs("/estimates/7", coef(count_g)[["gc"]], se_of(count_g, "gc"), df_t, 0.95, "count g = c"),
135    list("/design/design_df" = exact(df_t, "8 PSUs minus 2 strata"),
136         "/design/psus" = exact(8L, "eight PSUs"),
137         "/design/strata" = exact(2L, "two strata"))
138  )
139  emit("STAT-DESC-SURVEY-EST-001", "Survey mean, total, proportions and counts (Taylor)",
140       ESTIMATES, "surveyEstimatesMenuItem", EST_REQ,
141       list(design = design_taylor, numeric_columns = list("x"), categorical_columns = list("g"),
142            design_effect = "none", bounded_interval = "none", weights_population_scale = FALSE,
143            confidence_level = 0.95),
144       outputs,
145       c("svymean/svytotal with na.rm=TRUE keep the design and drop the missing rows' contribution, as the engine's zero-contribution convention.",
146         "Estimates are listed numeric columns first (Mean, Total per column), then each categorical level in natural order (Proportion, Count); g levels a, b, c give indices 2-7.",
147         "Intervals: estimate +/- qt(0.975, D) * SE with D = PSUs - strata = 6."))

Whole script generators/r/descriptives_survey_estimates.R


Survey domain estimates with 90% intervals

Case STAT-DESC-SURVEY-EST-002 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimates/0/estimatetolerance14.29264161985484814.2926416198548461.73E-1515passed
/estimates/0/standard_errortolerance0.322551226691272950.322551226691272951.99E-1817passed
/estimates/0/lowertolerance13.66586643669234613.6658664366923442.09E-1515passed
/estimates/0/uppertolerance14.9194168030173514.9194168030173491.37E-1516passed
/estimates/0/dfexact660.00E+2passed
/estimates/1/estimatetolerance359.00686000000002359.006862.59E-1517passed
/estimates/1/standard_errortolerance44.21842007699256544.218420076992566.83E-1515passed
/estimates/1/lowertolerance273.0824980708461273.082498070846045.70E-1415passed
/estimates/1/uppertolerance444.93122192915393444.9312219291546.18E-1415passed
/estimates/1/dfexact660.00E+2passed
/estimates/2/estimatetolerance0.537229828451766260.53722982845176611.15E-1615passed
/estimates/2/standard_errortolerance0.126185259555074620.126185259555074627.80E-1917passed
/estimates/2/lowertolerance0.292029120392640120.29202912039263983.35E-1614passed
/estimates/2/uppertolerance0.782430536510892340.78243053651089262.20E-1615passed
/estimates/2/dfexact660.00E+2passed
/estimates/5/estimatetolerance11.62400000000000111.6239999999999992.22E-1515passed
/estimates/5/standard_errortolerance3.6110903939576663.6110903939576664.40E-1716passed
/estimates/5/lowertolerance4.60700035530322884.6070003553032217.97E-1514passed
/estimates/5/uppertolerance18.64099964469677318.6409996446967773.73E-1515passed
/estimates/5/dfexact660.00E+2passed
/estimates/0/rowsexact12120.00E+2passed
/design/domain_rowsexact12120.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

150  domain_a <- subset(taylor, dom == "A")
151  mean_y <- svymean(~y, domain_a)
152  total_y <- svytotal(~y, domain_a)
153  prop_b <- svymean(~b, domain_a)
154  count_b <- svytotal(~b, domain_a)
155  outputs <- c(
156    estimate_outputs("/estimates/0", coef(mean_y)[[1]], se_of(mean_y, 1), df_t, 0.90, "domain mean of y"),
157    estimate_outputs("/estimates/1", coef(total_y)[[1]], se_of(total_y, 1), df_t, 0.90, "domain total of y"),
158    estimate_outputs("/estimates/2", coef(prop_b)[["bno"]], se_of(prop_b, "bno"), df_t, 0.90, "domain proportion b = no"),
159    estimate_outputs("/estimates/5", coef(count_b)[["byes"]], se_of(count_b, "byes"), df_t, 0.90, "domain count b = yes"),
160    list("/estimates/0/rows" = exact(sum(d$dom == "A"), "domain rows"),
161         "/design/domain_rows" = exact(sum(d$dom == "A"), "domain rows"))
162  )
163  emit("STAT-DESC-SURVEY-EST-002", "Survey domain estimates with 90% intervals",
164       ESTIMATES, "surveyEstimatesMenuItem", EST_REQ,
165       list(design = design_taylor, numeric_columns = list("y"), categorical_columns = list("b"),
166            design_effect = "none", bounded_interval = "none", weights_population_scale = FALSE,
167            domain_column = "dom", domain_values = list("A"), confidence_level = 0.90),
168       outputs,
169       c("subset(design, dom == 'A') keeps nondomain PSUs in the variance (R survey domain estimation).",
170         "90% intervals use qt(0.95, 6)."),
171       kind = "axis_coverage")

Whole script generators/r/descriptives_survey_estimates.R


Survey mean and total with a first-stage FPC

Case STAT-DESC-SURVEY-EST-003 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimates/0/estimatetolerance13.86857609751092813.868576097510932.01E-1515passed
/estimates/0/standard_errortolerance0.359315081788383890.359315081788383955.98E-1715passed
/estimates/0/lowertolerance12.98936376558780712.9893637655878061.41E-1515passed
/estimates/0/uppertolerance14.74778842943404914.7477884294340545.44E-1515passed
/estimates/0/dfexact660.00E+2passed
/estimates/1/estimatetolerance756.98294169999997756.98294170000011.13E-1315passed
/estimates/1/standard_errortolerance68.62930197005357568.629301970053581.33E-1617passed
/estimates/1/lowertolerance589.05308937366908589.05308937366864.53E-1315passed
/estimates/1/uppertolerance924.91279402633086924.91279402633156.79E-1315passed
/estimates/1/dfexact660.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

174  df_f <- as.integer(degf(with_fpc))
175  mean_yf <- svymean(~y, with_fpc)
176  total_yf <- svytotal(~y, with_fpc)
177  outputs <- c(
178    estimate_outputs("/estimates/0", coef(mean_yf)[[1]], se_of(mean_yf, 1), df_f, 0.95, "FPC mean of y"),
179    estimate_outputs("/estimates/1", coef(total_yf)[[1]], se_of(total_yf, 1), df_f, 0.95, "FPC total of y")
180  )
181  emit("STAT-DESC-SURVEY-EST-003", "Survey mean and total with a first-stage FPC",
182       ESTIMATES, "surveyEstimatesMenuItem", EST_REQ,
183       list(design = design_fpc, numeric_columns = list("y"), categorical_columns = list(),
184            design_effect = "none", bounded_interval = "none", weights_population_scale = FALSE,
185            confidence_level = 0.95),
186       outputs,
187       c("svydesign(fpc = ~N) with N the population count of PSUs per stratum (20 and 30); the (1 - n_h/N_h) factor applies per stratum."),
188       kind = "axis_coverage")

Whole script generators/r/descriptives_survey_estimates.R


Survey mean and proportion from supplied replicate weights

Case STAT-DESC-SURVEY-EST-004 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimates/0/estimatetolerance13.86857609751092813.868576097510932.01E-1515passed
/estimates/0/standard_errortolerance0.120619806586285880.120619806586288072.19E-1513passed
/estimates/0/lowertolerance13.48471003969636713.4847100396963625.49E-1515passed
/estimates/0/uppertolerance14.2524421553254914.2524421553254998.52E-1515passed
/estimates/0/dfexact330.00E+2passed
/estimates/2/estimatetolerance0.394556140601583620.39455614060158364.11E-1816passed
/estimates/2/standard_errortolerance0.0365946961189259480.036594696118926016.29E-1714passed
/estimates/2/lowertolerance0.278095485144927710.27809548514492752.20E-1615passed
/estimates/2/uppertolerance0.511016796058239580.51101679605823971.06E-1615passed
/estimates/2/dfexact330.00E+2passed
/design/design_dfexact330.00E+2passed
/design/replicatesexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

191  df_r <- as.integer(degf(replicate))
192  mean_yr <- svymean(~y, replicate)
193  prop_br <- svymean(~b, replicate)
194  outputs <- c(
195    estimate_outputs("/estimates/0", coef(mean_yr)[[1]], se_of(mean_yr, 1), df_r, 0.95, "replicate mean of y"),
196    estimate_outputs("/estimates/2", coef(prop_br)[["bno"]], se_of(prop_br, "bno"), df_r, 0.95, "replicate proportion b = no"),
197    list("/design/design_df" = exact(df_r, "rank of the four replicate-weight columns minus 1"),
198         "/design/replicates" = exact(4L, "four replicate columns"))
199  )
200  emit("STAT-DESC-SURVEY-EST-004", "Survey mean and proportion from supplied replicate weights",
201       ESTIMATES, "surveyEstimatesMenuItem", EST_REQ,
202       list(design = design_rep, numeric_columns = list("y"), categorical_columns = list("b"),
203            design_effect = "none", bounded_interval = "none", weights_population_scale = FALSE,
204            confidence_level = 0.95),
205       outputs,
206       c("svrepdesign(type = 'other', scale = 0.75, rscales = 1, mse = TRUE, combined.weights = TRUE): V = 0.75 * sum_r (theta_r - theta_full)^2.",
207         "degf = qr(replicate weights, tol = 1e-5)$rank - 1 = 3."),
208       kind = "axis_coverage")

Whole script generators/r/descriptives_survey_estimates.R


Survey proportion with Korn-Graubard limits and SRS design effect

Case STAT-DESC-SURVEY-EST-005 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimates/2/estimatetolerance0.394556140601583670.39455614060158365.41E-1715passed
/estimates/2/standard_errortolerance0.0525330985383096330.0525330985383096266.66E-1815passed
/estimates/2/lowertolerance0.266012279210827370.26601227921082685.51E-1614passed
/estimates/2/uppertolerance0.523100001992339970.52310000199234044.43E-1615passed
/estimates/2/dfexact660.00E+2passed
/estimates/2/extensions/bounded_lowertolerance0.272570926461390420.272570926461390033.92E-1614passed
/estimates/2/extensions/bounded_uppertolerance0.526986394802889670.52698639480288992.20E-1615passed
/estimates/2/extensions/effective_ntolerance61.86674010048246761.8667401004820264.41E-1314passed
/estimates/0/extensions/design_effecttolerance1.1101385556463771.11013855564637722.01E-1615passed
/estimates/0/extensions/deft_wrtolerance1.05363112883322541.05363112883322562.28E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

211  kg <- svyciprop(~I(b == "no"), taylor, method = "beta", level = 0.95)
212  p_no <- as.double(coef(prop_g <- svymean(~b, taylor))[["bno"]])
213  v_no <- as.double(vcov(prop_g)["bno", "bno"])
214  n_eff <- p_no * (1 - p_no) / v_no * (qt(0.025, nrow(d) - 1) / qt(0.025, df_t))^2
215  mean_y_deff <- svymean(~y, taylor, deff = "replace")
216  outputs <- c(
217    estimate_outputs("/estimates/2", p_no, sqrt(v_no), df_t, 0.95, "proportion b = no"),
218    list("/estimates/2/extensions/bounded_lower" = tol(attr(kg, "ci")[[1]], "svyciprop beta (Korn-Graubard) lower limit"),
219         "/estimates/2/extensions/bounded_upper" = tol(attr(kg, "ci")[[2]], "svyciprop beta (Korn-Graubard) upper limit"),
220         "/estimates/2/extensions/effective_n" = tol(n_eff, "p(1-p)/V times [qt(alpha/2, n-1)/qt(alpha/2, D)]^2 with n = 24 rows"),
221         "/estimates/0/extensions/design_effect" = tol(as.double(deff(mean_y_deff)[[1]]), "svymean deff = 'replace' for the mean of y"),
222         "/estimates/0/extensions/deft_wr" = tol(sqrt(as.double(deff(mean_y_deff)[[1]])), "square root of the with-replacement design effect"))
223  )
224  emit("STAT-DESC-SURVEY-EST-005", "Survey proportion with Korn-Graubard limits and SRS design effect",
225       ESTIMATES, "surveyEstimatesMenuItem", EST_REQ,
226       list(design = design_taylor, numeric_columns = list("y"), categorical_columns = list("b"),
227            design_effect = "srs_wr", bounded_interval = "korn_graubard",
228            weights_population_scale = FALSE, confidence_level = 0.95),
229       outputs,
230       c(paste0(MANUAL, ":126-138: Korn-Graubard uses n_e = p(1-p)/V * (t_{n_s-1}/t_D)^2 and beta quantiles; n_s counts domain rows before item missingness."),
231         paste0(MANUAL, ":144-147: with-replacement SRS comparator divides the normalized weighted dispersion by the observed n - 1 (R svymean deff = 'replace')."),
232         "Estimates: y Mean (0), y Total (1), b = no Proportion (2), b = no Count (3), b = yes (4, 5)."),
233       kind = "axis_coverage")

Whole script generators/r/descriptives_survey_estimates.R


Survey proportion with a delta-logit interval

Case STAT-DESC-SURVEY-EST-006 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimates/0/estimatetolerance0.394556140601583670.39455614060158365.41E-1715passed
/estimates/0/extensions/bounded_lowertolerance0.27561712312921910.27561712312921865.00E-1614passed
/estimates/0/extensions/bounded_uppertolerance0.527449028617880340.52744902861788095.60E-1614passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

236  xl <- svyciprop(~I(b == "no"), taylor, method = "xlogit", level = 0.95)
237  emit("STAT-DESC-SURVEY-EST-006", "Survey proportion with a delta-logit interval",
238       ESTIMATES, "surveyEstimatesMenuItem", EST_REQ,
239       list(design = design_taylor, numeric_columns = list(), categorical_columns = list("b"),
240            design_effect = "none", bounded_interval = "delta_logit",
241            weights_population_scale = FALSE, confidence_level = 0.95),
242       list("/estimates/0/estimate" = tol(p_no, "proportion b = no"),
243            "/estimates/0/extensions/bounded_lower" = tol(attr(xl, "ci")[[1]], "svyciprop xlogit lower limit"),
244            "/estimates/0/extensions/bounded_upper" = tol(attr(xl, "ci")[[2]], "svyciprop xlogit upper limit")),
245       c(paste0(MANUAL, ":122-125: delta-logit transforms logit(p) +/- t_{D} sqrt(V)/[p(1-p)] back to the probability scale (svyciprop method 'xlogit').")),
246       kind = "axis_coverage")

Whole script generators/r/descriptives_survey_estimates.R


Survey ratio of weighted totals (Taylor)

Case STAT-DESC-SURVEY-RATIO-001 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/ratios/0/ratiotolerance2.6384504651873832.6384504651873849.09E-1615passed
/ratios/0/standard_errortolerance0.0651138097214457030.065113809721445798.37E-1714passed
/ratios/0/lowertolerance2.47912271250677922.4791227125067791.97E-1717passed
/ratios/0/uppertolerance2.79777821786798692.79777821786798861.74E-1515passed
/ratios/0/dfexact660.00E+2passed
/ratios/0/rowsexact22220.00E+2passed
/ratios/0/missing_rowsexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

362  outputs <- c(ratio_outputs(0, taylor, "y", df_t, 0.95, "y / x"),
363               list("/ratios/0/rows" = exact(sum(!is.na(d$x)), "rows with both observed"),
364                    "/ratios/0/missing_rows" = exact(sum(is.na(d$x)), "rows with x missing")))
365  emit("STAT-DESC-SURVEY-RATIO-001", "Survey ratio of weighted totals (Taylor)",
366       RATIOS, "surveyRatiosMenuItem", RATIO_REQ,
367       list(design = design_taylor, numerators = list("y"), denominator = "x", confidence_level = 0.95),
368       outputs,
369       c("svyratio(~y, ~x, na.rm = TRUE): rows with a missing x stay in the design with zero contribution.",
370         "Linearized contribution w (y - R x) / sum(w x); interval ratio +/- qt(0.975, 6) * SE."))

Whole script generators/r/descriptives_survey_estimates.R


Survey ratios of two numerators in a domain

Case STAT-DESC-SURVEY-RATIO-002 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/ratios/0/ratiotolerance2.31388434368081082.31388434368081034.93E-1615passed
/ratios/0/standard_errortolerance0.136342783652147610.136342783652147538.44E-1715passed
/ratios/0/lowertolerance2.04894573509739342.0489457350973934.27E-1615passed
/ratios/0/uppertolerance2.57882295226422812.57882295226422764.59E-1615passed
/ratios/0/dfexact660.00E+2passed
/ratios/1/ratiotolerance1.39148118313503581.39148118313503552.51E-1615passed
/ratios/1/standard_errortolerance0.05705592898266150.0570559289826614653.51E-1715passed
/ratios/1/lowertolerance1.28061122704944651.2806112270494464.23E-1615passed
/ratios/1/uppertolerance1.5023511392206251.5023511392206252.03E-1716passed
/ratios/1/dfexact660.00E+2passed
/ratios/1/rowsexact12120.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

373  outputs <- c(ratio_outputs(0, domain_a, "y", df_t, 0.90, "domain y / x"),
374               ratio_outputs(1, domain_a, "z", df_t, 0.90, "domain z / x"),
375               list("/ratios/1/rows" = exact(sum(d$dom == "A" & !is.na(d$x)), "domain rows with x observed")))
376  emit("STAT-DESC-SURVEY-RATIO-002", "Survey ratios of two numerators in a domain",
377       RATIOS, "surveyRatiosMenuItem", RATIO_REQ,
378       list(design = design_taylor, numerators = list("y", "z"), denominator = "x",
379            domain_column = "dom", domain_values = list("A"), confidence_level = 0.90),
380       outputs,
381       c("Ratios are listed in numerator order (y then z); subset(design, dom == 'A') keeps the design."),
382       kind = "axis_coverage")

Whole script generators/r/descriptives_survey_estimates.R


Survey ratio from supplied replicate weights

Case STAT-DESC-SURVEY-RATIO-003 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/ratios/0/ratiotolerance2.6384504651873832.6384504651873849.09E-1615passed
/ratios/0/standard_errortolerance0.0382612327740668650.0382612327740680241.16E-1513passed
/ratios/0/lowertolerance2.51668614630995392.51668614630995122.66E-1514passed
/ratios/0/uppertolerance2.76021478406481212.76021478406481664.47E-1514passed
/ratios/0/dfexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

385  outputs <- ratio_outputs(0, replicate, "y", df_r, 0.95, "replicate y / x")
386  emit("STAT-DESC-SURVEY-RATIO-003", "Survey ratio from supplied replicate weights",
387       RATIOS, "surveyRatiosMenuItem", RATIO_REQ,
388       list(design = design_rep, numerators = list("y"), denominator = "x", confidence_level = 0.95),
389       outputs,
390       c("svyratio on svrepdesign(type = 'other', scale = 0.75, mse = TRUE): the ratio is recomputed under each replicate weight column."),
391       kind = "axis_coverage")

Whole script generators/r/descriptives_survey_estimates.R


Survey crosstab with the second-order Rao-Scott F

Case STAT-DESC-SURVEY-XTAB-001 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/0/counttolerance3.63329999999999983.63329999999999984.80E-1716passed
/cells/0/count_setolerance2.64433159229322052.64433159229322054.35E-1716passed
/cells/0/percenttolerance6.65651691198293886.6565169119829392.22E-1717passed
/cells/0/percent_setolerance5.12028286037508455.12028286037508451.05E-1717passed
/cells/0/percent_lowertolerance-5.8723639002833146-5.8723639002833644.97E-1414passed
/cells/0/percent_uppertolerance19.18539772424919219.1853977242492424.99E-1414passed
/cells/0/row_percenttolerance19.72711181574346319.7271118157434631.58E-1617passed
/cells/0/row_percent_setolerance12.56701122923582212.5670112292358222.56E-1616passed
/cells/0/column_percenttolerance16.87089928909402316.8708992890940233.25E-1717passed
/cells/0/column_percent_setolerance10.76148007557678810.7614800755767889.58E-1717passed
/cells/3/counttolerance7.88090000000000057.88090000000000053.96E-1717passed
/cells/3/count_setolerance3.1074586031031853.10745860310318544.46E-1615passed
/cells/3/percenttolerance14.43848405902247114.4384840590224753.78E-1515passed
/cells/3/percent_setolerance5.94074837097700395.9407483709770059.11E-1615passed
/cells/3/percent_lowertolerance-0.098003534591327934-0.0980035345913835.51E-1412passed
/cells/3/percent_uppertolerance28.9749716526362728.9749716526363346.43E-1414passed
/cells/3/row_percenttolerance41.87090570027468341.870905700274681.70E-1617passed
/cells/3/row_percent_setolerance12.89887676499296112.898876764992961.79E-1515passed
/cells/3/column_percenttolerance23.84776694798572923.847766947985734.69E-1717passed
/cells/3/column_percent_setolerance9.61175666860539669.6117566686053981.75E-1515passed
/test/f_statistictolerance1.73071681860325241.7307168186032541.59E-1515passed
/test/numerator_dftolerance1.86445878818852951.8644587881885294.41E-1615passed
/test/denominator_dftolerance11.18675272913117711.1867527291311751.76E-1515passed
/test/p_valuetail_probability0.221522878373364010.22152287837336364.14E-1614passed
/test/pearsontolerance2.52903117866209472.5290311786620961.34E-1515passed
/test/contingency_dfexact220.00E+2passed
/analysis_rowsexact24240.00E+2passed
/cells/0/rowsexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

288  outputs <- c(crosstab_outputs(taylor, df_t, 0.95, 0, "a", "no"),
289               crosstab_outputs(taylor, df_t, 0.95, 3, "b", "yes"),
290               rao_scott_outputs(taylor, nrow(d)),
291               list("/analysis_rows" = exact(nrow(d), "every row has g and b"),
292                    "/cells/0/rows" = exact(sum(d$g == "a" & d$b == "no"), "unweighted rows in cell a/no")))
293  emit("STAT-DESC-SURVEY-XTAB-001", "Survey crosstab with the second-order Rao-Scott F",
294       CROSSTAB, "surveyCrosstabMenuItem", XTAB_REQ,
295       list(design = design_taylor, row_column = "g", column_column = "b", confidence_level = 0.95),
296       outputs,
297       c(paste0(MANUAL, ":73-75: survey crosstabs use design-adjusted Rao-Scott inference; the ordinary Pearson chi-square is an unadjusted comparison."),
298         "Cells are row-major over row levels a, b, c and column levels no, yes: index = row * 2 + column.",
299         "svychisq(statistic = 'F'): F = X2 / tr(Delta) on (d0, d0 * D) with d0 = tr(Delta)^2 / tr(Delta^2)."))

Whole script generators/r/descriptives_survey_estimates.R


Survey one-way frequencies in a domain

Case STAT-DESC-SURVEY-XTAB-002 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/0/counttolerance10.47510.4750000000000011.42E-1515passed
/cells/0/count_setolerance5.47769577285924135.4776957728592414.78E-1717passed
/cells/0/percenttolerance35.55149791442524835.5514979144252556.98E-1515passed
/cells/0/percent_setolerance19.17978637541596519.1797863754159651.71E-1718passed
/cells/0/percent_lowertolerance-11.37974867000888-11.3797486700090641.84E-1313passed
/cells/0/percent_uppertolerance82.48274449885937582.482744498859571.99E-1314passed
/cells/1/counttolerance9.8898000000000019.88981.74E-1515passed
/cells/1/count_setolerance4.10746434190243464.1074643419024359.72E-1817passed
/cells/1/percenttolerance33.56536554406518733.565365544065194.64E-1616passed
/cells/1/percent_setolerance14.50204776750504914.5020477675050533.53E-1515passed
/cells/1/percent_lowertolerance-1.9198670041133568-1.9198670041135061.49E-1313passed
/cells/1/percent_uppertolerance69.05059809224373269.050598092243891.56E-1314passed
/cells/2/counttolerance9.09950000000000089.09951.00E-1717passed
/cells/2/count_setolerance7.81730332339739687.8173033233973988.41E-1615passed
/cells/2/percenttolerance30.88313654150955430.8831365415095583.56E-1515passed
/cells/2/percent_setolerance23.09958952080570823.0995895208057081.79E-1617passed
/cells/2/percent_lowertolerance-25.639522813534061-25.6395228135342852.24E-1314passed
/cells/2/percent_uppertolerance87.40579589655317287.40579589655342.28E-1314passed
/analysis_rowsexact12120.00E+2passed
/cells/0/rowsexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

319  outputs[["/analysis_rows"]] <- exact(sum(d$dom == "B"), "domain rows")
320  outputs[["/cells/0/rows"]] <- exact(sum(d$dom == "B" & d$g == "a"), "unweighted domain rows with g = a")
321  emit("STAT-DESC-SURVEY-XTAB-002", "Survey one-way frequencies in a domain",
322       CROSSTAB, "surveyCrosstabMenuItem", XTAB_REQ,
323       list(design = design_taylor, row_column = "g", domain_column = "dom",
324            domain_values = list("B"), confidence_level = 0.95),
325       outputs,
326       c("One column only: cells are the levels of g in natural order; no test of independence.",
327         "subset(design, dom == 'B') keeps the whole design for the variance."),
328       kind = "axis_coverage")

Whole script generators/r/descriptives_survey_estimates.R


Survey crosstab with a first-stage FPC and 90% intervals

Case STAT-DESC-SURVEY-XTAB-003 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svymean, svytotal, svyratio, svyby, svychisq, svyciprop
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, jsonlite 2.0.0
Reference script
generators/r/descriptives_survey_estimates.R · sha256 10511ae4153cc664
Dataset
reference_data/datasets/DESC_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/1/counttolerance14.784514.7845000000000011.31E-1516passed
/cells/1/count_setolerance3.40465159056122963.40465159056122963.17E-1717passed
/cells/1/percenttolerance27.0864707800654427.0864707800654477.17E-1515passed
/cells/1/percent_setolerance4.28986106553020634.2898610655302058.92E-1615passed
/cells/1/percent_lowertolerance18.75049735137677918.750497351376784.20E-1616passed
/cells/1/percent_uppertolerance35.42244420875410135.4224442087541151.39E-1415passed
/cells/1/row_percenttolerance80.2728881842565380.272888184256541.43E-1415passed
/cells/1/row_percent_setolerance11.51637701618865311.5163770161886541.28E-1515passed
/cells/1/column_percenttolerance44.73820381460175444.738203814601766.66E-1515passed
/cells/1/column_percent_setolerance7.97964401877101547.9796440187710168.56E-1615passed
/cells/4/counttolerance6.96160000000000076.96169.12E-1615passed
/cells/4/count_setolerance3.95138518707891473.95138518707891472.91E-1717passed
/cells/4/percenttolerance12.75424769065599812.7542476906559962.16E-1515passed
/cells/4/percent_setolerance6.52774991919746936.52774991919746751.80E-1515passed
/cells/4/percent_lowertolerance0.0696527715361163760.069652771536105721.07E-1412passed
/cells/4/percent_uppertolerance25.43884260977587925.4388426097758866.96E-1515passed
/cells/4/row_percenttolerance40.14092222177375640.140922221773756.64E-1515passed
/cells/4/row_percent_setolerance12.28087503647723412.2808750364772343.36E-1616passed
/cells/4/column_percenttolerance32.32555871823328232.3255587182332757.35E-1515passed
/cells/4/column_percent_setolerance17.76728367347806917.767283673478071.94E-1616passed
/test/f_statistictolerance2.08456250522739422.0845625052273958.57E-1615passed
/test/numerator_dftolerance1.87211835679406651.8721183567940664.81E-1615passed
/test/denominator_dftolerance11.23271014076439811.2327101407643971.44E-1515passed
/test/p_valuetail_probability0.1711715174634140.171171517463414141.36E-1615passed
/test/pearsontolerance2.52903117866209472.5290311786620961.34E-1515passed
/test/contingency_dfexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

331  outputs <- c(crosstab_outputs(with_fpc, df_f, 0.90, 1, "a", "yes"),
332               crosstab_outputs(with_fpc, df_f, 0.90, 4, "c", "no"),
333               rao_scott_outputs(with_fpc, nrow(d)))
334  emit("STAT-DESC-SURVEY-XTAB-003", "Survey crosstab with a first-stage FPC and 90% intervals",
335       CROSSTAB, "surveyCrosstabMenuItem", XTAB_REQ,
336       list(design = design_fpc, row_column = "g", column_column = "b", confidence_level = 0.90),
337       outputs,
338       c("svydesign(fpc = ~N): finite population corrections enter the cell covariance and so the Rao-Scott design effects.",
339         "90% percent intervals use qt(0.95, 6)."),
340       kind = "axis_coverage")

Whole script generators/r/descriptives_survey_estimates.R


Two-way counts and all percentage denominators

Case STAT-DESC-TABCAT-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::quantile and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_tabulation.R · sha256 bb74c4a5af397aaa
Dataset
reference_data/datasets/DESC_TABULATION_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valid_nexact10100.00E+2passed
/cells/0/countexact330.00E+2passed
/cells/0/row_totalexact550.00E+2passed
/cells/0/column_totalexact550.00E+2passed
/cells/0/overall_totalexact10100.00E+2passed
/cells/0/row_percentagetolerance6060.00.00E+2passed
/cells/0/column_percentagetolerance6060.00.00E+2passed
/cells/0/overall_percentagetolerance3030.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

125  counts <- table(factor(d$row, levels = c("X", "Y")), factor(d$column, levels = c("L", "R")))
126  total <- sum(counts)
127  out <- list(
128    "/valid_n" = exact(total, "Ten rows have complete categorical roles."),
129    "/cells/0/count" = exact(counts["X", "L"], "X/L count is exact."),
130    "/cells/0/row_total" = exact(sum(counts["X", ]), "X row margin is exact."),
131    "/cells/0/column_total" = exact(sum(counts[, "L"]), "L column margin is exact."),
132    "/cells/0/overall_total" = exact(total, "Grand total is exact."),
133    "/cells/0/row_percentage" = tol(100 * counts["X", "L"] / sum(counts["X", ]),
134                                    "Cell / row total in percent units; owner 1e-9 tier."),
135    "/cells/0/column_percentage" = tol(100 * counts["X", "L"] / sum(counts[, "L"]),
136                                       "Cell / column total in percent units; owner 1e-9 tier."),
137    "/cells/0/overall_percentage" = tol(100 * counts["X", "L"] / total,
138                                        "Cell / grand total in percent units; owner 1e-9 tier.")
139  )
140  emit("STAT-DESC-TABCAT-001", "Two-way counts and all percentage denominators",
141       "DESC_TABULATION_01", list(kind = "categorical", row_column = "row",
142       column_column = "column", row_percentages = TRUE, column_percentages = TRUE,
143       overall_percentages = TRUE, include_margins = TRUE), out,
144       c("Cells are ordered by observed X,Y and L,R labels; percentage outputs are in percent units (100 times the cell fraction)."),
145       version = "2.0.0")

Whole script generators/r/descriptives_tabulation.R


One-way table without margins

Case STAT-DESC-TABCAT-002 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::quantile and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_tabulation.R · sha256 bb74c4a5af397aaa
Dataset
reference_data/datasets/DESC_TABULATION_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valid_nexact10100.00E+2passed
/cells/0/countexact550.00E+2passed
/cells/0/row_percentageabsentNonepassed
/cells/0/column_percentageabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

147  out <- list(
148    "/valid_n" = exact(nrow(d), "Every row has a row category."),
149    "/cells/0/count" = exact(sum(d$row == "X"), "One-way X count is exact."),
150    "/cells/0/row_percentage" = absent("Row percentages are disabled."),
151    "/cells/0/column_percentage" = absent("Column percentages are disabled.")
152  )
153  emit("STAT-DESC-TABCAT-002", "One-way table without margins", "DESC_TABULATION_01",
154       list(kind = "categorical", row_column = "row", include_margins = FALSE), out,
155       c("The one-way table counts each observed category once and omits margins."))

Whole script generators/r/descriptives_tabulation.R


Ungrouped numeric summaries

Case STAT-DESC-TABNUM-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::quantile and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_tabulation.R · sha256 bb74c4a5af397aaa
Dataset
reference_data/datasets/DESC_TABULATION_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/valid_nexact990.00E+2passed
/numeric_summaries/0/meantolerance55.00.00E+2passed
/numeric_summaries/0/mediantolerance55.00.00E+2passed
/numeric_summaries/0/first_quartiletolerance33.00.00E+2passed
/numeric_summaries/0/third_quartiletolerance77.00.00E+2passed
/numeric_summaries/0/standard_deviationtolerance2.73861278752583062.73861278752583065.85E-1817passed
/numeric_summaries/0/standard_errortolerance0.91287092917527690.91287092917527691.72E-1817passed
/numeric_summaries/0/confidence_lowertolerance2.8949158624141422.89491586241414244.49E-1615passed
/numeric_summaries/0/confidence_uppertolerance7.10508413758585897.1050841375858571.79E-1515passed
/source_countexact10100.00E+2passed
/missing_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

74  out <- summary_outputs(d$value)
75  out[["/source_count"]] <- exact(nrow(d), "All ten source rows are counted.")
76  out[["/missing_count"]] <- exact(sum(is.na(d$value)), "One response is missing.")
77  emit("STAT-DESC-TABNUM-001", "Ungrouped numeric summaries", "DESC_TABULATION_01",
78       list(kind = "numeric", value_column = "value"), out,
79       c("Manual chapter 29 lines 24-38: sample SD, type-7 quantiles and two-sided t interval."))

Whole script generators/r/descriptives_tabulation.R


Two ordered groups

Case STAT-DESC-TABNUM-002 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::quantile and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_tabulation.R · sha256 bb74c4a5af397aaa
Dataset
reference_data/datasets/DESC_TABULATION_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/valid_nexact550.00E+2passed
/numeric_summaries/0/meantolerance33.00.00E+2passed
/numeric_summaries/0/mediantolerance33.00.00E+2passed
/numeric_summaries/0/first_quartiletolerance22.00.00E+2passed
/numeric_summaries/0/third_quartiletolerance44.00.00E+2passed
/numeric_summaries/0/standard_deviationtolerance1.58113883008418981.58113883008418983.86E-1716passed
/numeric_summaries/0/standard_errortolerance0.707106781186547570.70710678118654762.74E-1817passed
/numeric_summaries/0/confidence_lowertolerance1.03675683852244281.03675683852244282.51E-1716passed
/numeric_summaries/0/confidence_uppertolerance4.9632431614775574.9632431614775574.72E-1717passed
/numeric_summaries/1/meantolerance7.57.50.00E+1passed
/numeric_summaries/1/missing_nexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

81  out <- summary_outputs(d$value[d$group == "A"])
82  out[["/numeric_summaries/1/mean"]] <- tol(mean(d$value[d$group == "B"], na.rm = TRUE),
83                                               "Second group mean; owner 1e-9 tier.")
84  out[["/numeric_summaries/1/missing_n"]] <- exact(1L, "The B group has one missing response.")
85  emit("STAT-DESC-TABNUM-002", "Two ordered groups", "DESC_TABULATION_01",
86       list(kind = "numeric", value_column = "value", group_columns = list("group")), out,
87       c("Group labels are A then B in the manual's sorted string order."))

Whole script generators/r/descriptives_tabulation.R


Frequency-weighted order statistics

Case STAT-DESC-TABNUM-003 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::quantile and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_tabulation.R · sha256 bb74c4a5af397aaa
Dataset
reference_data/datasets/DESC_TABULATION_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/valid_nexact13130.00E+2passed
/numeric_summaries/0/meantolerance5.69230769230769255.69230769230769251.27E-1717passed
/numeric_summaries/0/mediantolerance66.00.00E+2passed
/numeric_summaries/0/first_quartiletolerance44.00.00E+2passed
/numeric_summaries/0/third_quartiletolerance88.00.00E+2passed
/numeric_summaries/0/standard_deviationtolerance2.81023769411697582.81023769411697582.93E-1716passed
/numeric_summaries/0/standard_errortolerance0.779419700183111020.7794197001831115.80E-1918passed
/numeric_summaries/0/confidence_lowertolerance3.99409804985334563.99409804985334561.32E-2020passed
/numeric_summaries/0/confidence_uppertolerance7.39051733476203947.3905173347620392.53E-1717passed
/numeric_summaries/0/countexact10100.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

89  replicated <- rep(d$value[!is.na(d$value)], d$frequency[!is.na(d$value)])
90  out <- summary_outputs(replicated)
91  out[["/numeric_summaries/0/count"]] <- exact(nrow(d), "Source rows include zero-frequency and missing rows.")
92  emit("STAT-DESC-TABNUM-003", "Frequency-weighted order statistics", "DESC_TABULATION_01",
93       list(kind = "numeric", value_column = "value", frequency_column = "frequency"), out,
94       c("Manual chapter 29 lines 28-36: integer frequencies are replication counts; zero-frequency rows remain in source count."))

Whole script generators/r/descriptives_tabulation.R


Explicit row exclusions and missingness

Case STAT-DESC-TABNUM-004 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::quantile and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_tabulation.R · sha256 bb74c4a5af397aaa
Dataset
reference_data/datasets/DESC_TABULATION_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/valid_nexact880.00E+2passed
/numeric_summaries/0/meantolerance5.3755.3750.00E-1passed
/numeric_summaries/0/mediantolerance5.55.50.00E+1passed
/numeric_summaries/0/first_quartiletolerance3.753.750.00E+0passed
/numeric_summaries/0/third_quartiletolerance7.257.250.00E+0passed
/numeric_summaries/0/standard_deviationtolerance2.66926956300782782.6692695630078285.73E-1817passed
/numeric_summaries/0/standard_errortolerance0.943729304408843620.94372930440884363.31E-1817passed
/numeric_summaries/0/confidence_lowertolerance3.14343479985605973.14343479985605973.07E-1717passed
/numeric_summaries/0/confidence_uppertolerance7.60656520014393947.606565200143949.31E-1615passed
/excluded_countexact220.00E+2passed
/missing_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 96  retained <- d[-c(2, 9), ]
 97  out <- summary_outputs(retained$value)
 98  out[["/excluded_count"]] <- exact(2L, "Two zero-based source row indexes are excluded.")
 99  out[["/missing_count"]] <- exact(0L, "The only missing response is excluded.")
100  emit("STAT-DESC-TABNUM-004", "Explicit row exclusions and missingness", "DESC_TABULATION_01",
101       list(kind = "numeric", value_column = "value"), out,
102       c("Excluded rows are zero-based positions 1 and 8 and contribute to neither valid nor missing N."),
103       kind = "missing_excluded", excluded = list(1L, 8L))

Whole script generators/r/descriptives_tabulation.R


One valid value

Case STAT-DESC-TABNUM-005 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::quantile and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_tabulation.R · sha256 bb74c4a5af397aaa
Dataset
reference_data/datasets/DESC_TABULATION_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/valid_nexact110.00E+2passed
/numeric_summaries/0/meantolerance77.00.00E+2passed
/numeric_summaries/0/mediantolerance77.00.00E+2passed
/numeric_summaries/0/first_quartiletolerance77.00.00E+2passed
/numeric_summaries/0/third_quartiletolerance77.00.00E+2passed
/numeric_summaries/0/standard_deviationabsentNonepassed
/numeric_summaries/0/confidence_lowerabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

105  one <- data.frame(value = 7)
106  export_dataset(one, "DESC_TABULATION_ONE", "Constructed one-row boundary",
107                 "R stats::quantile documentation", "Constructed validation data; no third-party license.",
108                 c(value = "Float64"))
109  one <- read_dataset("DESC_TABULATION_ONE")
110  emit("STAT-DESC-TABNUM-005", "One valid value", "DESC_TABULATION_ONE",
111       list(kind = "numeric", value_column = "value"), summary_outputs(one$value),
112       c("For n=1, the mean and quantiles are seven; sample SD and t interval are unavailable."),
113       kind = "boundary")

Whole script generators/r/descriptives_tabulation.R


Large offset and small spread

Case STAT-DESC-TABNUM-006 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::quantile and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/descriptives_tabulation.R · sha256 bb74c4a5af397aaa
Dataset
reference_data/datasets/DESC_TABULATION_OFFSET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/valid_nexact770.00E+2passed
/numeric_summaries/0/meantolerance10000000001000000000.00.00E+2passed
/numeric_summaries/0/mediantolerance10000000001000000000.00.00E+2passed
/numeric_summaries/0/first_quartiletolerance999999998.5999999998.50.00E+1passed
/numeric_summaries/0/third_quartiletolerance1000000001.51000000001.50.00E+1passed
/numeric_summaries/0/standard_deviationtolerance2.16024689946928692.1602468994692874.08E-1716passed
/numeric_summaries/0/standard_errortolerance0.816496580927726030.8164965809277264.46E-1817passed
/numeric_summaries/0/confidence_lowertolerance999999998.002105999999998.00210491.22E-715passed
/numeric_summaries/0/confidence_uppertolerance1000000001.99789521000000001.99789517.84E-816passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

115  hard <- data.frame(value = 1e9 + c(-3, -2, -1, 0, 1, 2, 3))
116  export_dataset(hard, "DESC_TABULATION_OFFSET", "Constructed large-offset small-spread values",
117                 "R stats::sd and stats::quantile documentation",
118                 "Constructed validation data; no third-party license.", c(value = "Float64"))
119  hard <- read_dataset("DESC_TABULATION_OFFSET")
120  emit("STAT-DESC-TABNUM-006", "Large offset and small spread", "DESC_TABULATION_OFFSET",
121       list(kind = "numeric", value_column = "value"), summary_outputs(hard$value),
122       c("Seven values lie at 1e9 plus -3,-2,-1,0,1,2,3; mean is 1e9 and sample SD is sqrt(28/6)."),
123       kind = "hard_numeric")

Whole script generators/r/descriptives_tabulation.R


Existing-table factorial with two-factor interactions

Case STAT-DOE-ANALYZE-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
reference_data/datasets/DOE_FACTORIAL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/payload/terms/[term=A]/coefficienttolerance4.36184999999999874.3618499999999935.36E-1514passed
/payload/terms/[term=A]/effecttolerance8.72369999999999738.7236999999999871.06E-1414passed
/payload/terms/[term=A]/std_errortolerance0.212573427877483210.212573427877482655.59E-1614passed
/payload/terms/[term=A]/p_valuetail_probability7.2415658667940198e-097.241565866793926e-099.34E-2313passed
/payload/terms/[term=B]/coefficienttolerance-2.5754375000000018-2.5754375000000064.02E-1514passed
/payload/terms/[term=B]/effecttolerance-5.1508750000000036-5.1508750000000128.04E-1514passed
/payload/terms/[term=B]/std_errortolerance0.212573427877483210.212573427877482655.59E-1614passed
/payload/terms/[term=B]/p_valuetail_probability7.0961603671585962e-077.096160367158331e-072.65E-2013passed
/payload/terms/[term=C]/coefficienttolerance0.941524999999998720.94152499999999622.55E-1514passed
/payload/terms/[term=C]/effecttolerance1.88304999999999741.88304999999999235.06E-1514passed
/payload/terms/[term=C]/std_errortolerance0.212573427877483260.212573427877482656.09E-1614passed
/payload/terms/[term=C]/p_valuetail_probability0.00164912352079182170.0016491235207918216.22E-1915passed
/payload/terms/[term=A x B]/coefficienttolerance1.80831250000000071.80831250000000473.95E-1514passed
/payload/terms/[term=A x B]/effecttolerance3.61662500000000133.61662500000000938.00E-1514passed
/payload/terms/[term=A x B]/std_errortolerance0.212573427877483210.212573427877482655.59E-1614passed
/payload/terms/[term=A x B]/p_valuetail_probability1.3506812233571844e-051.3506812233571331e-055.13E-1913passed
/payload/terms/[term=A x C]/coefficienttolerance0.265575000000000670.26557500000000524.55E-1513passed
/payload/terms/[term=A x C]/effecttolerance0.531150000000001340.53115000000001049.11E-1513passed
/payload/terms/[term=A x C]/std_errortolerance0.212573427877483210.212573427877482655.59E-1614passed
/payload/terms/[term=A x C]/p_valuetail_probability0.243056854330212980.24305685433020438.68E-1513passed
/payload/terms/[term=B x C]/coefficienttolerance-0.15436249999999851-0.154362499999996182.33E-1513passed
/payload/terms/[term=B x C]/effecttolerance-0.30872499999999703-0.308724999999992374.66E-1513passed
/payload/terms/[term=B x C]/std_errortolerance0.212573427877483210.212573427877482655.59E-1614passed
/payload/terms/[term=B x C]/p_valuetail_probability0.486194787227997470.48619478722800265.11E-1513passed
/payload/terms/[term=A]/lower_citolerance3.88097549750638263.88097549750637864.00E-1514passed
/payload/terms/[term=A]/upper_citolerance4.84272450249361524.8427245024936087.11E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

102  map <- c(a = "A", b = "B", c = "C", "a:b" = "A x B", "a:c" = "A x C", "b:c" = "B x C")
103  existing <- list(factors = list(list(name = "A", low = 10, high = 20), list(name = "B", low = 1, high = 3), list(name = "C", low = 100, high = 200)),
104                   response_name = "y", model = "interactions")
105  emit("STAT-DOE-ANALYZE-001", "tabelz.analysis.doe_advanced:analyze_catalog_design", "Existing-table factorial with two-factor interactions", "DOE_FACTORIAL",
106       list(response = "y", design = existing),
107       c(term_out("/payload/terms/", map, " x "),
108         list("/payload/terms/[term=A]/lower_ci" = tol(ci["a", 1], "t interval lower limit."),
109              "/payload/terms/[term=A]/upper_ci" = tol(ci["a", 2], "t interval upper limit."))),
110       c("lm(y ~ (a + b + c)^2) on factors coded to -1/+1 from the declared Low/High (line 33-34)."))

Whole script generators/r/design_doe.R


Main-effects model at 90% confidence

Case STAT-DOE-ANALYZE-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
reference_data/datasets/DOE_FACTORIAL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/payload/terms/[term=B]/coefficienttolerance-2.5754375000000014-2.57543750000000543.98E-1514passed
/payload/terms/[term=B]/lower_citolerance-3.5745575413773891-3.57455754137739132.25E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

111  fit3 <- lm(y ~ a + b + c, data = d); s3 <- summary(fit3)$coefficients
112  emit("STAT-DOE-ANALYZE-002", "tabelz.analysis.doe_advanced:analyze_catalog_design", "Main-effects model at 90% confidence", "DOE_FACTORIAL",
113       list(response = "y", design = modifyList(existing, list(model = "main_effects")), confidence_level = 0.9),
114       list("/payload/terms/[term=B]/coefficient" = tol(s3["b", 1], "Main-effects coefficient."),
115            "/payload/terms/[term=B]/lower_ci" = tol(confint(fit3, level = 0.9)["b", 1], "90% t interval.")),
116       c("model = main_effects."), kind = "axis_coverage")

Whole script generators/r/design_doe.R


Append two centre points

Case STAT-DOE-AUGMENT-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
reference_data/datasets/DOE_AUGMENT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/added_runsexact220.00E+2passed
/frame/4/Xtolerance550.00E+2passed
/frame/5/Ytolerance60600.00E+2passed
/frame/0/ytolerance3.10000000000000013.11.12E-1717passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

195  aug <- data.frame(RunOrder = 1:4, StdOrder = 1:4, Block = rep(1L, 4), PointType = rep("Factorial", 4),
196                    X = c(0, 10, 0, 10), Y = c(50, 50, 70, 70), y = c(3.1, 4.2, 3.9, 5.3))
197  export_dataset(aug, "DOE_AUGMENT", "Constructed executed 2^2 run sheet with bookkeeping columns", "Hand-constructed",
198                 "Constructed validation data; no third-party license.",
199                 c(RunOrder = "Int64", StdOrder = "Int64", Block = "Int64", PointType = "String", X = "Float64", Y = "Float64", y = "Float64"))
200  aug_design <- list(family = "full_factorial", factors = list(list(name = "X", low = 0, high = 10), list(name = "Y", low = 50, high = 70)),
201                     response_name = "y")
202  emit("STAT-DOE-AUGMENT-001", "tabelz.analysis.doe_augmentation:plan_design_augmentation", "Append two centre points", "DOE_AUGMENT",
203       list(design = aug_design, kind = "center", center_count = 2L),
204       list("/added_runs" = exact(2L, "Two appended centre runs."),
205            "/frame/4/X" = tol(5, "Centre of X."), "/frame/5/Y" = tol(60, "Centre of Y."),
206            "/frame/0/y" = tol(3.1, "The original runs and responses are unchanged (line 148-151).")),
207       c("Centre points sit at each factor's midpoint (line 156-157)."), kind = "nominal")

Whole script generators/r/design_doe.R


Zero centre points is an explicit no-op

Case STAT-DOE-AUGMENT-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
reference_data/datasets/DOE_AUGMENT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/no_opexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

208  emit("STAT-DOE-AUGMENT-002", "tabelz.analysis.doe_augmentation:plan_design_augmentation", "Zero centre points is an explicit no-op", "DOE_AUGMENT",
209       list(design = aug_design, kind = "center", center_count = 0L),
210       list("/no_op" = exact(TRUE, "Line 157-158: requesting zero centre points is an explicit no-op.")),
211       c("Nothing is appended."), kind = "boundary")

Whole script generators/r/design_doe.R


Binary response with a logit link

Case STAT-DOE-BINARY-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
reference_data/datasets/DOE_BINARY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/payload/terms/[term=A]/coefficienttolerance0.45771162602248150.45771162602248132.23E-1615passed
/payload/terms/[term=A]/std_errortolerance0.347326384073221530.347326384079544646.32E-1210passed
/payload/terms/[term=A]/p_valuetail_probability0.187565867599966160.187565867607999318.03E-1210passed
/payload/terms/[term=B]/coefficienttolerance-0.87228630644909266-0.87228630644909215.52E-1615passed
/payload/terms/[term=B]/std_errortolerance0.361185797287026430.36118579729603219.01E-1210passed
/payload/terms/[term=B]/p_valuetail_probability0.0157325060500487260.015732506052649752.60E-129passed
/payload/terms/[term=C]/coefficienttolerance0.872286306449092770.87228630644909243.29E-1615passed
/payload/terms/[term=C]/std_errortolerance0.361185797287026430.361185797296032239.01E-1210passed
/payload/terms/[term=C]/p_valuetail_probability0.0157325060500487050.015732506052649752.60E-129passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

142  emit("STAT-DOE-BINARY-001", "tabelz.analysis.doe_binomial:analyze_doe_binomial", "Binary response with a logit link", "DOE_BINARY",
143       list(response = "event", design = modifyList(existing, list(response_name = "event", model = "main_effects", response_kind = "binary"))),
144       bin_out, c("glm(binomial) with epsilon 1e-14; one trial per run (line 110-111)."))

Whole script generators/r/design_doe.R


Binary response with a probit link

Case STAT-DOE-BINARY-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
reference_data/datasets/DOE_BINARY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/payload/terms/[term=A]/coefficienttolerance0.299710318137441370.29971032188364333.75E-97passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

145  gp <- glm(event ~ a + b + c, family = binomial(link = "probit"), data = b, control = glm.control(epsilon = 1e-14, maxit = 100))
146  emit("STAT-DOE-BINARY-002", "tabelz.analysis.doe_binomial:analyze_doe_binomial", "Binary response with a probit link", "DOE_BINARY",
147       list(response = "event", design = modifyList(existing, list(response_name = "event", model = "main_effects", response_kind = "binary", link = "probit"))),
148       list("/payload/terms/[term=A]/coefficient" = tol(coef(gp)[["a"]], "Probit coefficient.", "1e-6")),
149       c("Explicit probit link (line 113-114)."), kind = "axis_coverage")

Whole script generators/r/design_doe.R


Catalog full factorial with two centre points

Case STAT-DOE-CATALOG-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact660.00E+2passed
/design_pointssequence['0|50', '0|70', '10|50', '10|70', '5|60', '5|60']['0|50', '0|70', '10|50', '10|70', '5|60', '5|60']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

63  catalog <- "tabelz.analysis.doe_designs:generate_catalog_design"
64  emit("STAT-DOE-CATALOG-001", catalog, "Catalog full factorial with two centre points", NULL,
65       list(family = "full_factorial", factors = list(list(name = "X", low = 0, high = 10), list(name = "Y", low = 50, high = 70)), center_points = 2L),
66       list("/runs" = exact(6L, "4 corners + 2 centres."),
67            "/design_points" = points_of(rbind(corners(c(0, 50), c(10, 70)), c(5, 60), c(5, 60)))), c("Catalog family full_factorial."))

Whole script generators/r/design_doe.R


Face-centred central composite for two factors

Case STAT-DOE-CATALOG-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact11110.00E+2passed
/design_pointssequence['-1|-1', '-1|0', '-1|1', '0|-1', '0|0', '0|0', '0|0', '0|1'['-1|-1', '-1|0', '-1|1', '0|-1', '0|0', '0|0', '0|0', '0|1'passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

68  face <- rbind(corners(c(-1, -1), c(1, 1)), c(-1, 0), c(1, 0), c(0, -1), c(0, 1), c(0, 0), c(0, 0), c(0, 0))
69  emit("STAT-DOE-CATALOG-002", catalog, "Face-centred central composite for two factors", NULL,
70       list(family = "central_composite", factors = list(list(name = "X", low = -1, high = 1), list(name = "Y", low = -1, high = 1)),
71            center_points = 3L, axial_type = "face_centered"),
72       list("/runs" = exact(11L, "4 corners + 4 face points + 3 centres."), "/design_points" = points_of(face)),
73       c("Face-centred axial points sit at +/-1 on one factor with the other at 0."), kind = "axis_coverage")

Whole script generators/r/design_doe.R


Full 2^3 factorial with two replicates and three centre points

Case STAT-DOE-CREATE-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact19190.00E+2passed
/design_pointssequence['10|1|100', '10|1|100', '10|1|200', '10|1|200', '10|3|100',['10|1|100', '10|1|100', '10|1|200', '10|1|200', '10|3|100',passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

48  create <- "tabelz.analysis.doe:create_factorial_design"
49  lows <- c(10, 1, 100); highs <- c(20, 3, 200)
50  pts <- rbind(corners(lows, highs), corners(lows, highs), matrix(rep((lows + highs) / 2, 3), 3, byrow = TRUE))
51  emit("STAT-DOE-CREATE-001", create, "Full 2^3 factorial with two replicates and three centre points", NULL,
52       list(factors = list(list(name = "A", low = 10, high = 20), list(name = "B", low = 1, high = 3), list(name = "C", low = 100, high = 200)),
53            replicates = 2L, center_points = 3L, seed = 7L),
54       list("/runs" = exact(19L, "8 corners x 2 replicates + 3 centres."), "/design_points" = points_of(pts)),
55       c("Each corner appears twice; centres sit at each factor's midpoint."))

Whole script generators/r/design_doe.R


The smallest design: two factors, one replicate

Case STAT-DOE-CREATE-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact440.00E+2passed
/design_pointssequence['0|-5', '0|5', '1|-5', '1|5']['0|-5', '0|5', '1|-5', '1|5']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

56  emit("STAT-DOE-CREATE-002", create, "The smallest design: two factors, one replicate", NULL,
57       list(factors = list(list(name = "A", low = 0, high = 1), list(name = "B", low = -5, high = 5)), randomize = FALSE),
58       list("/runs" = exact(4L, "Four corners."), "/design_points" = points_of(corners(c(0, -5), c(1, 5)))),
59       c("Randomization off."), kind = "boundary")

Whole script generators/r/design_doe.R


A factor with low above high is refused

Case STAT-DOE-CREATE-003 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

60  emit("STAT-DOE-CREATE-003", create, "A factor with low above high is refused", NULL,
61       list(factors = list(list(name = "A", low = 5, high = 1), list(name = "B", low = 0, high = 1))), list(),
62       c("Low and High must be ordered."), kind = "invalid_input", error = list(error_class = "ValueError", message_contains = "A"))

Whole script generators/r/design_doe.R


Legacy factorial analysis with interactions

Case STAT-DOE-LEGACY-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
reference_data/datasets/DOE_FACTORIAL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/terms/[term=A]/coefficienttolerance4.36184999999999874.3618499999999935.36E-1514passed
/terms/[term=A]/effecttolerance8.72369999999999738.7236999999999871.06E-1414passed
/terms/[term=A]/std_errortolerance0.212573427877483210.212573427877482655.59E-1614passed
/terms/[term=A]/p_valuetail_probability7.2415658667940198e-097.241565866793926e-099.34E-2313passed
/terms/[term=B]/coefficienttolerance-2.5754375000000018-2.5754375000000064.02E-1514passed
/terms/[term=B]/effecttolerance-5.1508750000000036-5.1508750000000128.04E-1514passed
/terms/[term=B]/std_errortolerance0.212573427877483210.212573427877482655.59E-1614passed
/terms/[term=B]/p_valuetail_probability7.0961603671585962e-077.096160367158331e-072.65E-2013passed
/terms/[term=C]/coefficienttolerance0.941524999999998720.94152499999999622.55E-1514passed
/terms/[term=C]/effecttolerance1.88304999999999741.88304999999999235.06E-1514passed
/terms/[term=C]/std_errortolerance0.212573427877483260.212573427877482656.09E-1614passed
/terms/[term=C]/p_valuetail_probability0.00164912352079182170.0016491235207918216.22E-1915passed
/terms/[term=A × B]/coefficienttolerance1.80831250000000071.80831250000000473.95E-1514passed
/terms/[term=A × B]/effecttolerance3.61662500000000133.61662500000000938.00E-1514passed
/terms/[term=A × B]/std_errortolerance0.212573427877483210.212573427877482655.59E-1614passed
/terms/[term=A × B]/p_valuetail_probability1.3506812233571844e-051.3506812233571331e-055.13E-1913passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

117  legacy_map <- c(a = "A", b = "B", c = "C", "a:b" = "A × B")
118  emit("STAT-DOE-LEGACY-001", "tabelz.analysis.doe:analyze_factorial", "Legacy factorial analysis with interactions", "DOE_FACTORIAL",
119       list(response = "y", factors = list("A", "B", "C")),
120       term_out("/terms/", legacy_map, " × "),
121       c("The legacy analysis codes each factor from its observed minimum and maximum, which here equal the declared levels; interactions are named with a multiplication sign."))

Whole script generators/r/design_doe.R


Two maximized responses that both peak at the high corner

Case STAT-DOE-OPT-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
reference_data/datasets/DOE_FACTORIAL_SHEET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/settings/[factor=A]/valuetolerance2020.00.00E+2passed
/settings/[factor=B]/valuetolerance33.00.00E+2passed
/settings/[factor=C]/valuetolerance200200.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

179  sheet <- read_dataset("DOE_FACTORIAL")
180  export_dataset(data.frame(RunOrder = seq_len(16), StdOrder = seq_len(16), Block = rep(1L, 16), PointType = rep("Factorial", 16),
181                            A = sheet$A, B = sheet$B, C = sheet$C, y2 = sheet$y2, y3 = sheet$y3),
182                 "DOE_FACTORIAL_SHEET", "DOE_FACTORIAL's responses on a catalog run sheet with bookkeeping columns",
183                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
184                 c(RunOrder = "Int64", StdOrder = "Int64", Block = "Int64", PointType = "String", A = "Float64", B = "Float64",
185                   C = "Float64", y2 = "Float64", y3 = "Float64"))
186  emit("STAT-DOE-OPT-001", "tabelz.analysis.doe_advanced:optimize_multiple_responses", "Two maximized responses that both peak at the high corner", "DOE_FACTORIAL_SHEET",
187       list(design = cat_design, goals = list(list(response = "y2", goal = "maximize", weight = 1),
188                                              list(response = "y3", goal = "maximize", weight = 1))),
189       list("/settings/[factor=A]/value" = tol(20, "A at its high level."),
190            "/settings/[factor=B]/value" = tol(3, "B at its high level."),
191            "/settings/[factor=C]/value" = tol(200, "C at its high level.")),
192       c("Under the interaction model both predictions are multilinear, so their maxima lie at vertices; R confirms the high corner is the best vertex for each, so any weighted geometric desirability of two increasing goals peaks there (line 101-105)."),
193       kind = "axis_coverage")

Whole script generators/r/design_doe.R


One-sample t power at n = 30

Case STAT-DOE-POWER-001 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/achieved_powertolerance0.753964715743851380.753964715743796.14E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

48  p30 <- t_power(0.5, 30)
49  emit("STAT-DOE-POWER-001", pw, pmenu, "One-sample t power at n = 30",
50       list(family = "raw_t", solve_for = "power", sample_size = 30L, mean_difference = 1, standard_deviation = 2),
51       list("/achieved_power" = tol(p30, "Two-sided noncentral t power with d = 0.5, ncp = d sqrt(n).")),
52       c("d = 1 / 2 = 0.5."))

Whole script generators/r/design_planning.R


One-sample t sample size for 80% power

Case STAT-DOE-POWER-002 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sample_sizeexact34340.00E+2passed
/achieved_powertolerance0.807777501279440040.80777750127927381.66E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

53  n <- 2L; while (t_power(0.5, n) < 0.8) n <- n + 1L
54  emit("STAT-DOE-POWER-002", pw, pmenu, "One-sample t sample size for 80% power",
55       list(family = "raw_t", solve_for = "sample_size", mean_difference = 1, standard_deviation = 2),
56       list("/sample_size" = exact(n, "Smallest n with power >= 0.8."), "/achieved_power" = tol(t_power(0.5, n), "Power at that n.")),
57       c("Linear search over n in R."), kind = "axis_coverage")

Whole script generators/r/design_planning.R


Independent t power with allocation ratio 1.5

Case STAT-DOE-POWER-003 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/achieved_powertolerance0.5306481191331810.53064811913287573.05E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

58  n2 <- ceiling(20 * 1.5)
59  emit("STAT-DOE-POWER-003", pw, pmenu, "Independent t power with allocation ratio 1.5",
60       list(family = "raw_t", solve_for = "power", design = "independent", sample_size = 20L, ratio = 1.5, mean_difference = 1.2, standard_deviation = 2),
61       list("/achieved_power" = tol(t_power(0.6, 20, n2), "Pooled-df noncentral t with n1 = 20, n2 = ceiling(20 x 1.5) = 30.")),
62       c("Line 34-37."), kind = "axis_coverage")

Whole script generators/r/design_planning.R


Known-sigma Z power at n = 25

Case STAT-DOE-POWER-004 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/achieved_powertolerance0.705418001113800370.70541800111380031.06E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

63  z <- qnorm(0.975); zp <- pnorm(0.5 * sqrt(25) - z) + pnorm(-0.5 * sqrt(25) - z)
64  emit("STAT-DOE-POWER-004", pw, pmenu, "Known-sigma Z power at n = 25",
65       list(family = "z", solve_for = "power", sample_size = 25L, mean_difference = 1, known_sigma = 2),
66       list("/achieved_power" = tol(zp, "Two-sided normal power.")), c("Line 48-49."), kind = "axis_coverage")

Whole script generators/r/design_planning.R


One-way ANOVA power from group means

Case STAT-DOE-POWER-005 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/achieved_powertolerance0.472533034760505950.4725330347605066.00E-1715passed
/effect_sizetolerance0.415739709641548990.4157397096415493.44E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

78  fp <- ncf_upper(qf(0.95, 2, N - 3), 2, N - 3, f^2 * N)
79  emit("STAT-DOE-POWER-005", pw, pmenu, "One-way ANOVA power from group means",
80       list(family = "raw_anova", solve_for = "power", group_means = list(10, 12, 13), group_sizes = list(15L, 15L, 15L), standard_deviation = 3),
81       list("/achieved_power" = tol(fp, "Noncentral F with lambda = f^2 N, as a Poisson mixture of pbeta tails."), "/effect_size" = tol(f, "Cohen's f.")),
82       c("Manual chapter 36 lines 44-45: 'the supplied group sizes determine the weighting and the solve's N is total observations'; N is the entered sample size, 30 by default, and lambda = f^2 N with N - 3 denominator df.",
83         "Corrected before first commit: the first draft took N as the sum of the group sizes, which the quoted line does not say."),
84       kind = "axis_coverage")

Whole script generators/r/design_planning.R


Normal variance test power for a true ratio of 2

Case STAT-DOE-POWER-006 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/achieved_powertolerance0.628923163612745140.62892316361274564.44E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

85  nv <- 20; lo <- qchisq(0.025, nv - 1); hi <- qchisq(0.975, nv - 1)
86  vp <- pchisq(lo / 2, nv - 1) + pchisq(hi / 2, nv - 1, lower.tail = FALSE)
87  emit("STAT-DOE-POWER-006", pw, pmenu, "Normal variance test power for a true ratio of 2",
88       list(family = "normal_variance", solve_for = "power", sample_size = 20L, variance_ratio = 2),
89       list("/achieved_power" = tol(vp, "(n-1) s^2 / sigma0^2 ~ ratio x chi2(n - 1) under the alternative.")),
90       c("Line 49-51."), kind = "axis_coverage")

Whole script generators/r/design_planning.R


Poisson rate power with an integer cutoff

Case STAT-DOE-POWER-007 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/achieved_powertolerance0.919430978905020350.91943097890502021.15E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

91  mu0 <- 0.1 * 10 * 15; mu1 <- 0.2 * 10 * 15
92  cut <- 0L; while (ppois(cut - 1, mu0, lower.tail = FALSE) > 0.05) cut <- cut + 1L
93  emit("STAT-DOE-POWER-007", pw, pmenu, "Poisson rate power with an integer cutoff",
94       list(family = "poisson_rate", solve_for = "power", sample_size = 15L, poisson_rate = 0.2, null_rate = 0.1, exposure = 10, alternative = "greater"),
95       list("/achieved_power" = tol(ppois(cut - 1, mu1, lower.tail = FALSE), "P(X >= c) under the true rate, with c the smallest cutoff whose null tail is at most alpha.")),
96       c(sprintf("Total exposure 15 x 10; null mean %g, true mean %g, cutoff %d.", mu0, mu1, cut)), kind = "axis_coverage")

Whole script generators/r/design_planning.R


One-sample proportion power with the arcsine effect

Case STAT-DOE-POWER-008 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/achieved_powertolerance0.296285448079648350.29628544807964821.66E-1615passed
/effect_sizetolerance0.201357920790330660.201357920790330662.18E-1816passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 97  h <- 2 * (asin(sqrt(0.6)) - asin(sqrt(0.5))); prp <- pnorm(abs(h) * sqrt(50) - z) + pnorm(-abs(h) * sqrt(50) - z)
 98  emit("STAT-DOE-POWER-008", pw, pmenu, "One-sample proportion power with the arcsine effect",
 99       list(family = "raw_proportion", solve_for = "power", sample_size = 50L, proportion = 0.6, null_proportion = 0.5),
100       list("/achieved_power" = tol(prp, "Two-sided normal power for h = 2 (asin sqrt 0.6 - asin sqrt 0.5)."), "/effect_size" = tol(h, "Cohen's h.")),
101       c("Line 40-42."), kind = "axis_coverage")

Whole script generators/r/design_planning.R


Paired proportion planning is refused

Case STAT-DOE-POWER-009 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

102  emit("STAT-DOE-POWER-009", pw, pmenu, "Paired proportion planning is refused",
103       list(family = "raw_proportion", solve_for = "power", design = "paired", proportion = 0.6, null_proportion = 0.5), list(),
104       c("Line 23-25: no paired binary covariance model is declared."), kind = "invalid_input",
105       error = list(error_class = "ValueError", message_contains = "aired"))

Whole script generators/r/design_planning.R


POWER-WINDOW 1

Case STAT-DOE-POWER-WINDOW-001 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_runsexact16160.00E+2passed
/corner_runsexact16160.00E+2passed
/rankexact880.00E+2passed
/error_dfexact880.00E+2passed
/noncentralitytolerance22.00.00E+2passed
/critical_ttolerance2.30600413520416622.3060041352041663.94E-1615passed
/powertolerance0.421051938048901180.42105193804838215.19E-1311passed
/solved_replicatesexact220.00E+2passed
/target_metexactFalseFalsepassed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


POWER-WINDOW 2

Case STAT-DOE-POWER-WINDOW-002 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_runsexact16160.00E+2passed
/corner_runsexact16160.00E+2passed
/rankexact440.00E+2passed
/error_dfexact12120.00E+2passed
/noncentralitytolerance22.00.00E+2passed
/critical_ttolerance2.17881282966722842.17881282966722844.56E-1716passed
/powertolerance0.452193700936374120.45219370093548078.93E-1311passed
/solved_replicatesexact220.00E+2passed
/target_metexactFalseFalsepassed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


POWER-WINDOW 3

Case STAT-DOE-POWER-WINDOW-003 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_runsexact16160.00E+2passed
/corner_runsexact16160.00E+2passed
/rankexact880.00E+2passed
/error_dfexact880.00E+2passed
/noncentralitytolerance22.00.00E+2passed
/critical_ttolerance1.85954803753089681.85954803753089734.58E-1615passed
/powertolerance0.572749254653879380.57274925465366812.11E-1312passed
/solved_replicatesexact220.00E+2passed
/target_metexactFalseFalsepassed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


POWER-WINDOW 4

Case STAT-DOE-POWER-WINDOW-004 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_runsexact40400.00E+2passed
/corner_runsexact40400.00E+2passed
/rankexact880.00E+2passed
/error_dfexact32320.00E+2passed
/noncentralitytolerance3.16227766016837953.1622776601683794.21E-1615passed
/critical_ttolerance2.03693334346010162.03693334346010162.47E-1716passed
/powertolerance0.865659663272845270.86565966327277217.32E-1413passed
/solved_replicatesexact550.00E+2passed
/target_metexactTrueTruepassed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


Distribution-free full-range coverage

Case STAT-DOE-PREC-DF-001 · status: passed

Reference

Evidence type
R
Source
R stats analytic formulas
Citation
R stats::qt, stats::qnorm, stats::qchisq and stats::pbinom documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/design_precision_planning.R · sha256 bc33fc829853bb1d
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact46460.00E+2passed
/achieved_confidencetolerance0.951996200375517580.95199620037551763.88E-1817passed
/previous_nexact45450.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 92  range_confidence <- stats::pbinom(n - 2L, size = n, prob = content)
 93  range_n <- which(range_confidence >= confidence)[1] + 1L
 94  stopifnot(!is.na(range_n), range_n >= 3L)
 95  emit("STAT-DOE-PREC-DF-001", "Distribution-free full-range coverage",
 96       list(mode = "distribution_free_tolerance", criterion = "coverage",
 97            confidence = confidence, content = content, max_n = 500L),
 98       list("/n" = exact(range_n, "First full-range order-statistic n reaching 95% confidence."),
 99            "/achieved_confidence" = tol(range_confidence[range_n - 1L], "P(Binomial(n,P)<=n-2); owner 1e-9 tier."),
100            "/previous_n" = exact(range_n - 1L, "Immediate predecessor has inadequate confidence.")),
101       c("Manual chapter 35: maximum full-range confidence is P(Binomial(n,P)<=n-2), with P=0.9."))

Whole script generators/r/design_precision_planning.R


Student-t mean half-width

Case STAT-DOE-PREC-MEAN-001 · status: passed

Reference

Evidence type
R
Source
R stats analytic formulas
Citation
R stats::qt, stats::qnorm, stats::qchisq and stats::pbinom documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/design_precision_planning.R · sha256 bc33fc829853bb1d
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact42420.00E+2passed
/achieved_precisiontolerance0.74789264909951580.74789264909951591.16E-1615passed
/previous_nexact41410.00E+2passed
/previous_achieved_precisiontolerance0.757533472235072390.75753347223507222.21E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

45  n <- 2L:500L
46  confidence <- 0.95
47  sd0 <- 2.4
48  mean_half <- stats::qt((1 + confidence) / 2, df = n - 1) * sd0 / sqrt(n)
49  mean_target <- 0.75
50  mean_n <- first(mean_half, mean_target)
51  stopifnot(mean_half[mean_n - 1L] <= mean_target, mean_half[mean_n - 2L] > mean_target)
52  emit("STAT-DOE-PREC-MEAN-001", "Student-t mean half-width", 
53       list(mode = "mean", criterion = "absolute", target = mean_target,
54            confidence = confidence, assumed_sd = sd0, max_n = 500L),
55       list("/n" = exact(mean_n, "First integer meeting the Student-t half-width target."),
56            "/achieved_precision" = tol(mean_half[mean_n - 1L], "t_(0.975,n-1)*2.4/sqrt(n); owner 1e-9 tier."),
57            "/previous_n" = exact(mean_n - 1L, "Immediate predecessor does not meet the target."),
58            "/previous_achieved_precision" = tol(mean_half[mean_n - 2L], "Predecessor t half-width; owner 1e-9 tier.")),
59       c("Manual chapter 35: h_n=t_(0.975,n-1)*s0/sqrt(n); search smallest n with h_n<=0.75.",
60         "R verified h_(n-1)>0.75 and h_n<=0.75 with assumed SD 2.4."))

Whole script generators/r/design_precision_planning.R


Wilson proportion half-width

Case STAT-DOE-PREC-PROP-001 · status: passed

Reference

Evidence type
R
Source
R stats analytic formulas
Citation
R stats::qt, stats::qnorm, stats::qchisq and stats::pbinom documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/design_precision_planning.R · sha256 bc33fc829853bb1d
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact84840.00E+2passed
/achieved_precisiontolerance0.0999599852736740010.099959985273674021.42E-1715passed
/previous_nexact83830.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

62  p0 <- 0.35
63  z <- stats::qnorm((1 + confidence) / 2)
64  denominator <- 1 + z^2 / n
65  proportion_half <- z / denominator * sqrt(p0 * (1 - p0) / n + z^2 / (4 * n^2))
66  proportion_target <- 0.1
67  proportion_n <- first(proportion_half, proportion_target)
68  emit("STAT-DOE-PREC-PROP-001", "Wilson proportion half-width",
69       list(mode = "proportion", criterion = "absolute", target = proportion_target,
70            confidence = confidence, assumed_proportion = p0, max_n = 500L),
71       list("/n" = exact(proportion_n, "First integer meeting Wilson half-width target."),
72            "/achieved_precision" = tol(proportion_half[proportion_n - 1L], "Wilson half-width at entered p0; owner 1e-9 tier."),
73            "/previous_n" = exact(proportion_n - 1L, "Immediate predecessor misses the target.")),
74       c("Manual chapter 35: Wilson score half-width uses entered p0=0.35, z=qnorm(0.975), and D=1+z^2/n."))

Whole script generators/r/design_precision_planning.R


Howe two-sided normal tolerance factor

Case STAT-DOE-PREC-TOL-001 · status: passed

Reference

Evidence type
R
Source
R tolerance::K.factor 3.0.0, explicit HE2
Citation
Howe (1969), equation (2.13); R tolerance 3.0.0 K.factor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/design_precision_planning.R · sha256 bc33fc829853bb1d
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact15150.00E+2passed
/achieved_precisiontolerance2.47971680122487692.47971680122487831.35E-1515passed
/previous_nexact14140.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

11  emit <- function(id, title, config, outputs, notes) {
12    write_expected("design", id, list(
13      schema_version = "2.0.0", case_id = id, outputs = outputs,
14      expected_error = NULL, deferred_outputs = list(), generator_run = run
15    ))
16    write_case("design", id, list(
17      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "design",
18      entry_point = "tabelz.executor.precision_planning",
19      menu_items = list("precisionPlanningMenuItem"), title = title,
20      case_kind = "nominal", dataset = NULL, excluded_rows = list(),
21      configuration = config, "repeat" = 1L,
22      expected = list(path = paste0("reference_data/expected/design/", id, ".json"), sha256 = ""),
23      oracle = list(
24        oracle_class = "independent_implementation",
25        source = if (id == "STAT-DOE-PREC-TOL-001")
26          "R tolerance::K.factor 3.0.0, explicit HE2" else "R stats analytic formulas",
27        citation = if (id == "STAT-DOE-PREC-TOL-001")
28          "Howe (1969), equation (2.13); R tolerance 3.0.0 K.factor documentation" else
29          "R stats::qt, stats::qnorm, stats::qchisq and stats::pbinom documentation",
30        generator = list(path = "generators/r/design_precision_planning.R", sha256 = ""),
31        independent = TRUE,
32        convention_notes = as.list(notes)
33      ), supported = TRUE, unsupported_reason = NULL,
34      requirement_ids = list("REQ-STAT-DOE-001"), risk_ids = list("RISK-REFERENCE-001"),
35      review_reference = NULL
36    ))

Whole script generators/r/design_precision_planning.R


Zero-failure binomial demonstration

Case STAT-DOE-RELPLAN-001 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chosen_nexact29290.00E+2passed
/achieved_confidencetolerance0.952898713027537590.95289871302753751.09E-1615passed
/achieved_criteriontolerance0.0471012869724624220.0471012869724624856.27E-1714passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

108  rp <- "tabelz.executor.reliability_planning"; rmenu <- "reliabilityPlanningMenuItem"
109  binom_n <- function(R, C, f) { n <- f + 1L; while (pbinom(f, n, 1 - R) > 1 - C) n <- n + 1L; n }
110  n0 <- binom_n(0.9, 0.95, 0)
111  emit("STAT-DOE-RELPLAN-001", rp, rmenu, "Zero-failure binomial demonstration",
112       list(mode = "binomial", target_reliability = 0.9, mission_time = 1000, confidence = 0.95),
113       list("/chosen_n" = exact(n0, "Smallest n with 0.9^n <= 0.05; the manual's worked example gives 29."),
114            "/achieved_confidence" = tol(1 - pbinom(0, n0, 0.1), "Achieved confidence 1 - P(X <= 0) (manual chapter 58)."),
115            "/achieved_criterion" = tol(pbinom(0, n0, 0.1), "Criterion P(X <= 0) at the chosen n.")),
116       c(sprintf("ceiling(log(0.05) / log(0.9)) = %d.", ceiling(log(0.05) / log(0.9))),
117         "Until 2026-09-25 the achieved confidence was compared with the criterion field."))

Whole script generators/r/design_planning.R


One allowed failure at 90% confidence

Case STAT-DOE-RELPLAN-002 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chosen_nexact77770.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

118  n1 <- binom_n(0.95, 0.90, 1)
119  emit("STAT-DOE-RELPLAN-002", rp, rmenu, "One allowed failure at 90% confidence",
120       list(mode = "binomial", target_reliability = 0.95, mission_time = 1000, confidence = 0.9, allowed_failures = 1L),
121       list("/chosen_n" = exact(n1, "Smallest n with P(X <= 1) <= 0.10; the manual's worked example gives 77.")),
122       c("Line 131-132."), kind = "axis_coverage")

Whole script generators/r/design_planning.R


Duration for ten units with an exponential model

Case STAT-DOE-RELPLAN-003 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chosen_test_durationtolerance2843.31588057434192843.31588057466433.22E-1012passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

123  t_needed <- uniroot(function(t) pbinom(0, 10, 1 - 0.9^(t / 1000)) - 0.05, c(1000, 10000), tol = 1e-12)$root
124  emit("STAT-DOE-RELPLAN-003", rp, rmenu, "Duration for ten units with an exponential model",
125       list(mode = "duration", target_reliability = 0.9, mission_time = 1000, confidence = 0.95, search_dimension = "test_duration", sample_size = 10L),
126       list("/chosen_test_duration" = tol(1000 * log(0.05) / (10 * log(0.9)), "t = t0 ln(1-C) / (n ln R0); the manual gives about 2843 h.")),
127       c(sprintf("Root of the binomial rule: %.10g.", t_needed)), kind = "axis_coverage")

Whole script generators/r/design_planning.R


Failure-terminated exponential chi-square rule

Case STAT-DOE-RELPLAN-004 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_test_timetolerance45025.06929443067945025.069294430682.92E-1317passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

128  theta0 <- -1000 / log(0.9)
129  emit("STAT-DOE-RELPLAN-004", rp, rmenu, "Failure-terminated exponential chi-square rule",
130       list(mode = "duration", target_reliability = 0.9, mission_time = 1000, confidence = 0.95, allowed_failures = 1L,
131            scheme = "failure_terminated", sample_size = 10L),
132       list("/total_test_time" = tol(theta0 / 2 * qchisq(0.95, 4), "theta0 / 2 chi2_(0.95, 4); the manual gives about 45025.")),
133       c("Line 81-92."), kind = "axis_coverage")

Whole script generators/r/design_planning.R


A Weibull plan without a shape is refused

Case STAT-DOE-RELPLAN-005 · status: passed

Reference

Evidence type
R
Source
R stats noncentral t, F, chi-square, binomial and Poisson functions
Citation
R stats::TDist, stats::FDist, stats::Chisquare, stats::Binomial and stats::Poisson documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_planning.R · sha256 bae913190e25a0a5
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

134  emit("STAT-DOE-RELPLAN-005", rp, rmenu, "A Weibull plan without a shape is refused",
135       list(mode = "duration", target_reliability = 0.9, mission_time = 1000, confidence = 0.95, distribution = "weibull", sample_size = 10L,
136            search_dimension = "test_duration"), list(),
137       c("Line 112-113."), kind = "invalid_input", error = list(error_class = "ValueError", message_contains = "shape"))

Whole script generators/r/design_planning.R


ROBUST-ANALYSIS 1

Case STAT-DOE-ROBUST-ANALYSIS-001 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_ROBUST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/run_countexact12120.00E+2passed
/residual_dfexact880.00E+2passed
/rankexact440.00E+2passed
/term_countexact440.00E+2passed
/residual_variancetolerance0.999999999999995231.04.77E-1514passed
/r_squaredtolerance0.95454545454545470.95454545454545461.14E-1615passed
/coefficients/0/estimatetolerance10.00000000000000410.04.00E-1515passed
/coefficients/0/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/0/t_valuetolerance34.64101615137764234.641016151377559.28E-1414passed
/coefficients/0/p_valuetolerance5.2735611365661297e-105.273561136566244e-101.15E-2313passed
/coefficients/0/lowertolerance9.33431394589374729.3343139458937425.34E-1515passed
/coefficients/0/uppertolerance10.6656860541062610.6656860541062581.86E-1515passed
/coefficients/1/estimatetolerance1.99999999999999932.07.00E-1615passed
/coefficients/1/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/1/t_valuetolerance6.9282032302755236.928203230275511.33E-1414passed
/coefficients/1/p_valuetolerance0.000121040205555298510.000121040205555300141.63E-1813passed
/coefficients/1/lowertolerance1.33431394589374321.33431394589374238.96E-1615passed
/coefficients/1/uppertolerance2.66568605410625552.66568605410625772.20E-1515passed
/coefficients/2/estimatetolerance2.99999999999999782.99999999999999961.76E-1515passed
/coefficients/2/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/2/t_valuetolerance10.3923048454132810.3923048454132621.77E-1414passed
/coefficients/2/p_valuetolerance6.3640368981107531e-066.364036898110831e-067.80E-2013passed
/coefficients/2/lowertolerance2.33431394589374142.3343139458937424.60E-1615passed
/coefficients/2/uppertolerance3.66568605410625413.66568605410625733.15E-1515passed
/coefficients/3/estimatetolerance1.00000000000000110.99999999999999961.54E-1514passed
/coefficients/3/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/3/t_valuetolerance3.46410161513776683.4641016151377531.37E-1414passed
/coefficients/3/p_valuetolerance0.00851626337090111850.0085162633709012921.73E-1613passed
/coefficients/3/lowertolerance0.334313945893744860.334313945893741972.89E-1514passed
/coefficients/3/uppertolerance1.66568605410625731.66568605410625734.77E-1716passed
/settings/0/rankexact110.00E+2passed
/settings/0/controls/0/valuetolerance-1-1.00.00E+2passed
/settings/0/meantolerance8.00000000000000368.03.60E-1515passed
/settings/0/noise_variancetolerance3.99999999999998764.01.24E-1414passed
/settings/0/residual_variancetolerance0.999999999999995231.04.77E-1514passed
/settings/0/total_sdtolerance2.23606797749978582.236067977499794.01E-1514passed
/settings/0/objectivetolerance8.9999999999999689.03.20E-1414passed
/settings/1/rankexact220.00E+2passed
/settings/1/controls/0/valuetolerance11.00.00E+2passed
/settings/1/meantolerance12.00000000000000412.04.00E-1515passed
/settings/1/noise_variancetolerance1616.00.00E+2passed
/settings/1/residual_variancetolerance0.999999999999995231.04.77E-1514passed
/settings/1/total_sdtolerance4.12310562561765974.1231056256176618.86E-1615passed
/settings/1/objectivetolerance21.00000000000001121.01.10E-1415passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-ANALYSIS 2

Case STAT-DOE-ROBUST-ANALYSIS-002 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_ROBUST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/run_countexact12120.00E+2passed
/residual_dfexact880.00E+2passed
/rankexact440.00E+2passed
/term_countexact440.00E+2passed
/residual_variancetolerance0.999999999999995231.04.77E-1514passed
/r_squaredtolerance0.95454545454545470.95454545454545461.14E-1615passed
/coefficients/0/estimatetolerance10.00000000000000410.04.00E-1515passed
/coefficients/0/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/0/t_valuetolerance34.64101615137764234.641016151377559.28E-1414passed
/coefficients/0/p_valuetolerance5.2735611365661297e-105.273561136566244e-101.15E-2313passed
/coefficients/0/lowertolerance9.33431394589374729.3343139458937425.34E-1515passed
/coefficients/0/uppertolerance10.6656860541062610.6656860541062581.86E-1515passed
/coefficients/1/estimatetolerance1.99999999999999932.07.00E-1615passed
/coefficients/1/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/1/t_valuetolerance6.9282032302755236.928203230275511.33E-1414passed
/coefficients/1/p_valuetolerance0.000121040205555298510.000121040205555300141.63E-1813passed
/coefficients/1/lowertolerance1.33431394589374321.33431394589374238.96E-1615passed
/coefficients/1/uppertolerance2.66568605410625552.66568605410625772.20E-1515passed
/coefficients/2/estimatetolerance2.99999999999999782.99999999999999961.76E-1515passed
/coefficients/2/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/2/t_valuetolerance10.3923048454132810.3923048454132621.77E-1414passed
/coefficients/2/p_valuetolerance6.3640368981107531e-066.364036898110831e-067.80E-2013passed
/coefficients/2/lowertolerance2.33431394589374142.3343139458937424.60E-1615passed
/coefficients/2/uppertolerance3.66568605410625413.66568605410625733.15E-1515passed
/coefficients/3/estimatetolerance1.00000000000000110.99999999999999961.54E-1514passed
/coefficients/3/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/3/t_valuetolerance3.46410161513776683.4641016151377531.37E-1414passed
/coefficients/3/p_valuetolerance0.00851626337090111850.0085162633709012921.73E-1613passed
/coefficients/3/lowertolerance0.334313945893744860.334313945893741972.89E-1514passed
/coefficients/3/uppertolerance1.66568605410625731.66568605410625734.77E-1716passed
/settings/0/rankexact110.00E+2passed
/settings/0/controls/0/valuetolerance-1-1.00.00E+2passed
/settings/0/meantolerance7.00000000000000447.04.40E-1515passed
/settings/0/noise_variancetolerance2.99999999999999073.09.30E-1514passed
/settings/0/residual_variancetolerance0.999999999999995231.04.77E-1514passed
/settings/0/total_sdtolerance1.99999999999999642.03.60E-1514passed
/settings/0/objectivetolerance1010.00.00E+2passed
/settings/1/rankexact220.00E+2passed
/settings/1/controls/0/valuetolerance11.00.00E+2passed
/settings/1/meantolerance10.00000000000000410.04.00E-1515passed
/settings/1/noise_variancetolerance1212.00.00E+2passed
/settings/1/residual_variancetolerance0.999999999999995231.04.77E-1514passed
/settings/1/total_sdtolerance3.60555127546398873.6055512754639894.25E-1615passed
/settings/1/objectivetolerance15.40832691319598715.4083269131959833.53E-1515passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-ANALYSIS 3

Case STAT-DOE-ROBUST-ANALYSIS-003 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_ROBUST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/run_countexact12120.00E+2passed
/residual_dfexact880.00E+2passed
/rankexact440.00E+2passed
/term_countexact440.00E+2passed
/residual_variancetolerance0.999999999999995231.04.77E-1514passed
/r_squaredtolerance0.95454545454545470.95454545454545461.14E-1615passed
/coefficients/0/estimatetolerance10.00000000000000410.04.00E-1515passed
/coefficients/0/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/0/t_valuetolerance34.64101615137764234.641016151377559.28E-1414passed
/coefficients/0/p_valuetolerance5.2735611365661297e-105.273561136566244e-101.15E-2313passed
/coefficients/0/lowertolerance9.33431394589374729.3343139458937425.34E-1515passed
/coefficients/0/uppertolerance10.6656860541062610.6656860541062581.86E-1515passed
/coefficients/1/estimatetolerance1.99999999999999932.07.00E-1615passed
/coefficients/1/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/1/t_valuetolerance6.9282032302755236.928203230275511.33E-1414passed
/coefficients/1/p_valuetolerance0.000121040205555298510.000121040205555300141.63E-1813passed
/coefficients/1/lowertolerance1.33431394589374321.33431394589374238.96E-1615passed
/coefficients/1/uppertolerance2.66568605410625552.66568605410625772.20E-1515passed
/coefficients/2/estimatetolerance2.99999999999999782.99999999999999961.76E-1515passed
/coefficients/2/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/2/t_valuetolerance10.3923048454132810.3923048454132621.77E-1414passed
/coefficients/2/p_valuetolerance6.3640368981107531e-066.364036898110831e-067.80E-2013passed
/coefficients/2/lowertolerance2.33431394589374142.3343139458937424.60E-1615passed
/coefficients/2/uppertolerance3.66568605410625413.66568605410625733.15E-1515passed
/coefficients/3/estimatetolerance1.00000000000000110.99999999999999961.54E-1514passed
/coefficients/3/std_errortolerance0.28867513459481220.288675134594812876.66E-1614passed
/coefficients/3/t_valuetolerance3.46410161513776683.4641016151377531.37E-1414passed
/coefficients/3/p_valuetolerance0.00851626337090111850.0085162633709012921.73E-1613passed
/coefficients/3/lowertolerance0.334313945893744860.334313945893741972.89E-1514passed
/coefficients/3/uppertolerance1.66568605410625731.66568605410625734.77E-1716passed
/settings/0/rankexact110.00E+2passed
/settings/0/controls/0/valuetolerance11.00.00E+2passed
/settings/0/meantolerance12.00000000000000412.04.00E-1515passed
/settings/0/noise_variancetolerance1616.00.00E+2passed
/settings/0/residual_variancetolerance0.999999999999995231.04.77E-1514passed
/settings/0/total_sdtolerance4.12310562561765974.1231056256176618.86E-1615passed
/settings/0/objectivetolerance5.8153415615735145.815341561573514.43E-1515passed
/settings/1/rankexact220.00E+2passed
/settings/1/controls/0/valuetolerance-1-1.00.00E+2passed
/settings/1/meantolerance8.00000000000000368.03.60E-1515passed
/settings/1/noise_variancetolerance3.99999999999998764.01.24E-1414passed
/settings/1/residual_variancetolerance0.999999999999995231.04.77E-1514passed
/settings/1/total_sdtolerance2.23606797749978582.236067977499794.01E-1514passed
/settings/1/objectivetolerance4.64589803375032514.6458980337503159.81E-1514passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-CONFIRM 1

Case STAT-DOE-ROBUST-CONFIRM-001 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_CONFIRM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/run_countexact12120.00E+2passed
/residual_dfexact10100.00E+2passed
/mean_errortolerance00.00.00E+2passed
/rmsetolerance0.816496580927726030.8164965809277264.46E-1817passed
/observed_meantolerance88.00.00E+2passed
/observed_noise_variancetolerance44.00.00E+2passed
/observed_residual_variancetolerance0.800000000000000040.84.41E-1817passed
/observed_objectivetolerance8.80000000000000078.81.05E-1717passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-CONFIRM 2

Case STAT-DOE-ROBUST-CONFIRM-002 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_CONFIRM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/run_countexact12120.00E+2passed
/residual_dfexact10100.00E+2passed
/mean_errortolerance00.00.00E+2passed
/rmsetolerance0.816496580927726030.8164965809277264.46E-1817passed
/observed_meantolerance77.00.00E+2passed
/observed_noise_variancetolerance33.00.00E+2passed
/observed_residual_variancetolerance0.800000000000000040.84.41E-1817passed
/observed_objectivetolerance9.92403830344268919.924038303442691.29E-1717passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-CONFIRM 3

Case STAT-DOE-ROBUST-CONFIRM-003 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_CONFIRM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/run_countexact12120.00E+2passed
/residual_dfexact10100.00E+2passed
/mean_errortolerance00.00.00E+2passed
/rmsetolerance0.816496580927726030.8164965809277264.46E-1817passed
/observed_meantolerance88.00.00E+2passed
/observed_noise_variancetolerance44.00.00E+2passed
/observed_residual_variancetolerance0.800000000000000040.84.41E-1817passed
/observed_objectivetolerance4.71366465496900314.7136646549690039.39E-1817passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-GENERATE 1

Case STAT-DOE-ROBUST-GENERATE-001 · status: passed

Reference

Evidence type
Analytical
Source
Hand-checkable full crossing and fixed-input arithmetic
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact12120.00E+2passed
/design_pointssequence['-1|-1', '-1|-1', '-1|-1', '-1|1', '-1|1', '-1|1', '1|-1', ['-1|-1', '-1|-1', '-1|-1', '-1|1', '-1|1', '-1|1', '1|-1', passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-GENERATE 2

Case STAT-DOE-ROBUST-GENERATE-002 · status: passed

Reference

Evidence type
Analytical
Source
Hand-checkable full crossing and fixed-input arithmetic
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact880.00E+2passed
/design_pointssequence['-1|-1', '-1|-1', '-1|1', '-1|1', '1|-1', '1|-1', '1|1', '1['-1|-1', '-1|-1', '-1|1', '-1|1', '1|-1', '1|-1', '1|1', '1passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-GENERATE 3

Case STAT-DOE-ROBUST-GENERATE-003 · status: passed

Reference

Evidence type
Analytical
Source
Hand-checkable full crossing and fixed-input arithmetic
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact36360.00E+2passed
/design_pointssequence['-1|-1|1', '-1|-1|1', '-1|-1|1', '-1|-1|2', '-1|-1|2', '-1|['-1|-1|1', '-1|-1|1', '-1|-1|1', '-1|-1|2', '-1|-1|2', '-1|passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-PLAN 1

Case STAT-DOE-ROBUST-PLAN-001 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_ROBUST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact440.00E+2passed
/design_pointssequence['-1|-1|6', '-1|-1|6', '-1|1|10', '-1|1|10']['-1|-1|6', '-1|-1|6', '-1|1|10', '-1|1|10']passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-PLAN 2

Case STAT-DOE-ROBUST-PLAN-002 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_ROBUST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact660.00E+2passed
/design_pointssequence['-1|-1|6', '-1|-1|6', '-1|-1|6', '-1|1|10', '-1|1|10', '-1|['-1|-1|6', '-1|-1|6', '-1|-1|6', '-1|1|10', '-1|1|10', '-1|passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


ROBUST-PLAN 3

Case STAT-DOE-ROBUST-PLAN-003 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_ROBUST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact880.00E+2passed
/design_pointssequence['-1|-1|6', '-1|-1|6', '-1|-1|6', '-1|-1|6', '-1|1|10', '-1|['-1|-1|6', '-1|-1|6', '-1|-1|6', '-1|-1|6', '-1|1|10', '-1|passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


SIMULATION-FIXED 1

Case STAT-DOE-SIMULATION-FIXED-001 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/drawsexact1001000.00E+2passed
/meantolerance12.00000000000000412.04.00E-1515passed
/standard_deviationtolerance00.00.00E+2passed
/mean_monte_carlo_standard_errortolerance00.00.00E+2passed
/minimumtolerance12.00000000000000412.04.00E-1515passed
/maximumtolerance12.00000000000000412.04.00E-1515passed
/percentiles/0/valuetolerance12.00000000000000412.04.00E-1515passed
/thresholds/0/exceedancesexact1001000.00E+2passed
/thresholds/0/probabilitytolerance11.00.00E+2passed
/thresholds/0/lowertolerance0.963783307354823540.96378330735482361.08E-1615passed
/thresholds/0/uppertolerance11.00.00E+2passed
/percentiles/1/valuetolerance12.00000000000000412.04.00E-1515passed
/thresholds/1/exceedancesexact000.00E+2passed
/thresholds/1/probabilitytolerance00.00.00E+2passed
/thresholds/1/lowertolerance00.00.00E+2passed
/thresholds/1/uppertolerance0.0362166926451764630.036216692645176415.59E-1714passed
/percentiles/2/valuetolerance12.00000000000000412.04.00E-1515passed
/thresholds/2/exceedancesexact000.00E+2passed
/thresholds/2/probabilitytolerance00.00.00E+2passed
/thresholds/2/lowertolerance00.00.00E+2passed
/thresholds/2/uppertolerance0.0362166926451764630.036216692645176415.59E-1714passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


SIMULATION-FIXED 2

Case STAT-DOE-SIMULATION-FIXED-002 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/drawsexact2002000.00E+2passed
/meantolerance8.00000000000000368.03.60E-1515passed
/standard_deviationtolerance00.00.00E+2passed
/mean_monte_carlo_standard_errortolerance00.00.00E+2passed
/minimumtolerance8.00000000000000368.03.60E-1515passed
/maximumtolerance8.00000000000000368.03.60E-1515passed
/percentiles/0/valuetolerance8.00000000000000368.03.60E-1515passed
/thresholds/0/exceedancesexact2002000.00E+2passed
/thresholds/0/probabilitytolerance11.00.00E+2passed
/thresholds/0/lowertolerance0.981724659644863770.98172465964486382.99E-1817passed
/thresholds/0/uppertolerance11.00.00E+2passed
/percentiles/1/valuetolerance8.00000000000000368.03.60E-1515passed
/thresholds/1/exceedancesexact000.00E+2passed
/thresholds/1/probabilitytolerance00.00.00E+2passed
/thresholds/1/lowertolerance00.00.00E+2passed
/thresholds/1/uppertolerance0.0182753403551362270.0182753403551362371.04E-1715passed
/percentiles/2/valuetolerance8.00000000000000368.03.60E-1515passed
/thresholds/2/exceedancesexact000.00E+2passed
/thresholds/2/probabilitytolerance00.00.00E+2passed
/thresholds/2/lowertolerance00.00.00E+2passed
/thresholds/2/uppertolerance0.0182753403551362270.0182753403551362371.04E-1715passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


SIMULATION-FIXED 3

Case STAT-DOE-SIMULATION-FIXED-003 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/drawsexact1001000.00E+2passed
/meantolerance12.00000000000000412.04.00E-1515passed
/standard_deviationtolerance00.00.00E+2passed
/mean_monte_carlo_standard_errortolerance00.00.00E+2passed
/minimumtolerance12.00000000000000412.04.00E-1515passed
/maximumtolerance12.00000000000000412.04.00E-1515passed
/percentiles/0/valuetolerance12.00000000000000412.04.00E-1515passed
/thresholds/0/exceedancesexact1001000.00E+2passed
/thresholds/0/probabilitytolerance11.00.00E+2passed
/thresholds/0/lowertolerance0.970486950392960070.97048695039296011.74E-1817passed
/thresholds/0/uppertolerance11.00.00E+2passed
/percentiles/1/valuetolerance12.00000000000000412.04.00E-1515passed
/thresholds/1/exceedancesexact000.00E+2passed
/thresholds/1/probabilitytolerance00.00.00E+2passed
/thresholds/1/lowertolerance00.00.00E+2passed
/thresholds/1/uppertolerance0.0295130496070399320.029513049607039946.68E-1815passed
/percentiles/2/valuetolerance12.00000000000000412.04.00E-1515passed
/thresholds/2/exceedancesexact000.00E+2passed
/thresholds/2/probabilitytolerance00.00.00E+2passed
/thresholds/2/lowertolerance00.00.00E+2passed
/thresholds/2/uppertolerance0.0295130496070399320.029513049607039946.68E-1815passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


SIMULATION-FIXED 4

Case STAT-DOE-SIMULATION-FIXED-004 · status: passed

Reference

Evidence type
Analytical
Source
Hand-checkable full crossing and fixed-input arithmetic
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/drawsexact1001000.00E+2passed
/inputs/0/nameexactCCpassed
/inputs/0/roleexactrandomrandompassed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


TAGUCHI-ANALYSIS 1

Case STAT-DOE-TAGUCHI-ANALYSIS-001 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_TAGUCHI.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runs/0/countexact330.00E+2passed
/runs/0/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/0/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/0/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/0/sntolerance6.22477896799175226.2247789679917539.20E-1615passed
/runs/1/countexact330.00E+2passed
/runs/1/meantolerance4.3333333333333334.3333333333333333.73E-1717passed
/runs/1/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/1/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/1/sntolerance12.44086112631650612.4408611263165063.53E-1616passed
/runs/2/countexact330.00E+2passed
/runs/2/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/2/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/2/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/2/sntolerance6.22477896799175226.2247789679917539.20E-1615passed
/runs/3/countexact330.00E+2passed
/runs/3/meantolerance8.33333333333333398.3333333333333342.55E-1717passed
/runs/3/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/3/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/3/sntolerance18.34021218059519518.3402121805951952.04E-1717passed
/runs/4/countexact330.00E+2passed
/runs/4/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/4/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/4/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/4/sntolerance9.00418520032060789.0041852003206083.74E-1717passed
/runs/5/countexact330.00E+2passed
/runs/5/meantolerance5.03333333333333325.0333333333333349.03E-1615passed
/runs/5/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/5/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/5/sntolerance13.82125334560380713.8212533456038073.61E-1616passed
/runs/6/countexact330.00E+2passed
/runs/6/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/6/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/6/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/6/sntolerance9.00418520032060789.0041852003206083.74E-1717passed
/runs/7/countexact330.00E+2passed
/runs/7/meantolerance9.03333333333333329.0333333333333331.49E-1717passed
/runs/7/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/7/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/7/sntolerance19.05238195713640419.0523819571364043.67E-1616passed
/runs/8/countexact330.00E+2passed
/runs/8/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/8/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/8/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/8/sntolerance10.77644158744544710.7764415874454471.12E-1616passed
/runs/9/countexact330.00E+2passed
/runs/9/meantolerance5.63333333333333295.6333333333333334.04E-1717passed
/runs/9/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/9/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/9/sntolerance14.84459255808879214.8445925580887923.66E-1616passed
/runs/10/countexact330.00E+2passed
/runs/10/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/10/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/10/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/10/sntolerance10.77644158744544710.7764415874454471.12E-1616passed
/runs/11/countexact330.00E+2passed
/runs/11/meantolerance9.63333333333333469.6333333333333353.60E-1717passed
/runs/11/sdtolerance0.763762615825973380.7637626158259746.65E-1615passed
/runs/11/log_sdtolerance-0.26949825036634345-0.269498250366342568.88E-1614passed
/runs/11/sntolerance19.61889921213200719.6188992121320076.27E-1717passed
/tables/[summary=mean]/averages/0/0tolerance33.00.00E+2passed
/tables/[summary=mean]/averages/0/1tolerance77.00.00E+2passed
/tables/[summary=mean]/averages/1/0tolerance44.00.00E+2passed
/tables/[summary=mean]/averages/1/1tolerance66.00.00E+2passed
/predictions/[summary=mean]/valuetolerance66.00.00E+2passed
/tables/[summary=sd]/averages/0/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/0/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/tables/[summary=sd]/averages/1/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/1/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/predictions/[summary=sd]/valuetolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=log_sd]/averages/0/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/0/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/tables/[summary=log_sd]/averages/1/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/1/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/predictions/[summary=log_sd]/valuetolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=sn]/averages/0/0tolerance8.66846858525260278.6684685852526033.50E-1717passed
/tables/[summary=sn]/averages/0/1tolerance16.35303339664545116.353033396645453.63E-1616passed
/tables/[summary=sn]/averages/1/0tolerance11.18535213096115111.1853521309611512.18E-1616passed
/tables/[summary=sn]/averages/1/1tolerance13.83614985093690213.83614985093691.69E-1515passed
/predictions/[summary=sn]/valuetolerance15.02763453665757315.0276345366575734.61E-1616passed
/anova/[source=A]/dfexact110.00E+2passed
/anova/[source=A]/sstolerance48.00000000000000748.07.00E-1515passed
/anova/[source=A]/ftolerance28.07625649913344628.0762564991334182.80E-1415passed
/anova/[source=A]/ptolerance0.000494704573913236380.00049470457391323811.73E-1814passed
/anova/[source=B]/dfexact110.00E+2passed
/anova/[source=B]/sstolerance12.00000000000000512.05.00E-1515passed
/anova/[source=B]/ftolerance7.01906412478336347.01906412478335458.89E-1514passed
/anova/[source=B]/ptolerance0.0265043503601723570.026504350360172461.04E-1614passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


TAGUCHI-ANALYSIS 2

Case STAT-DOE-TAGUCHI-ANALYSIS-002 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_TAGUCHI.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runs/0/countexact330.00E+2passed
/runs/0/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/0/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/0/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/0/sntolerance-7.659167939666319-7.6591679396663193.60E-1717passed
/runs/1/countexact330.00E+2passed
/runs/1/meantolerance4.3333333333333334.3333333333333333.73E-1717passed
/runs/1/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/1/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/1/sntolerance-12.825465899699681-12.8254658996996813.16E-1616passed
/runs/2/countexact330.00E+2passed
/runs/2/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/2/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/2/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/2/sntolerance-7.659167939666319-7.6591679396663193.60E-1717passed
/runs/3/countexact330.00E+2passed
/runs/3/meantolerance8.33333333333333398.3333333333333342.55E-1717passed
/runs/3/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/3/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/3/sntolerance-18.440627725826516-18.4406277258265163.25E-1616passed
/runs/4/countexact330.00E+2passed
/runs/4/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/4/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/4/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/4/sntolerance-9.8181860717066378-9.8181860717066381.86E-1717passed
/runs/5/countexact330.00E+2passed
/runs/5/meantolerance5.03333333333333325.0333333333333349.03E-1615passed
/runs/5/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/5/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/5/sntolerance-14.103272455302914-14.1032724553029146.79E-1717passed
/runs/6/countexact330.00E+2passed
/runs/6/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/6/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/6/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/6/sntolerance-9.8181860717066378-9.8181860717066381.86E-1717passed
/runs/7/countexact330.00E+2passed
/runs/7/meantolerance9.03333333333333329.0333333333333331.49E-1717passed
/runs/7/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/7/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/7/sntolerance-19.13760886412323-19.137608864123233.47E-1717passed
/runs/8/countexact330.00E+2passed
/runs/8/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/8/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/8/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/8/sntolerance-11.332194567324942-11.3321945673249421.39E-1717passed
/runs/9/countexact330.00E+2passed
/runs/9/meantolerance5.63333333333333295.6333333333333334.04E-1717passed
/runs/9/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/9/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/9/sntolerance-15.068206042642258-15.0682060426422582.86E-1616passed
/runs/10/countexact330.00E+2passed
/runs/10/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/10/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/10/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/10/sntolerance-11.332194567324942-11.3321945673249421.39E-1717passed
/runs/11/countexact330.00E+2passed
/runs/11/meantolerance9.63333333333333469.6333333333333353.60E-1717passed
/runs/11/sdtolerance0.763762615825973380.7637626158259746.65E-1615passed
/runs/11/log_sdtolerance-0.26949825036634345-0.269498250366342568.88E-1614passed
/runs/11/sntolerance-19.693693117335275-19.6936931173352753.89E-1616passed
/tables/[summary=mean]/averages/0/0tolerance33.00.00E+2passed
/tables/[summary=mean]/averages/0/1tolerance77.00.00E+2passed
/tables/[summary=mean]/averages/1/0tolerance44.00.00E+2passed
/tables/[summary=mean]/averages/1/1tolerance66.00.00E+2passed
/predictions/[summary=mean]/valuetolerance66.00.00E+2passed
/tables/[summary=sd]/averages/0/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/0/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/tables/[summary=sd]/averages/1/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/1/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/predictions/[summary=sd]/valuetolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=log_sd]/averages/0/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/0/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/tables/[summary=log_sd]/averages/1/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/1/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/predictions/[summary=log_sd]/valuetolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=sn]/averages/0/0tolerance-9.6031828595659654-9.6031828595659671.78E-1515passed
/tables/[summary=sn]/averages/0/1tolerance-16.544812350821644-16.5448123508216443.14E-1717passed
/tables/[summary=sn]/averages/1/0tolerance-11.801082162723793-11.8010821627237932.66E-1616passed
/tables/[summary=sn]/averages/1/1tolerance-14.34691304766382-14.3469130476638181.53E-1515passed
/predictions/[summary=sn]/valuetolerance-15.271896908351632-15.2718969083516324.95E-1717passed
/anova/[source=A]/dfexact110.00E+2passed
/anova/[source=A]/sstolerance48.00000000000000748.07.00E-1515passed
/anova/[source=A]/ftolerance28.07625649913344628.0762564991334182.80E-1415passed
/anova/[source=A]/ptolerance0.000494704573913236380.00049470457391323811.73E-1814passed
/anova/[source=B]/dfexact110.00E+2passed
/anova/[source=B]/sstolerance12.00000000000000512.05.00E-1515passed
/anova/[source=B]/ftolerance7.01906412478336347.01906412478335458.89E-1514passed
/anova/[source=B]/ptolerance0.0265043503601723570.026504350360172461.04E-1614passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


TAGUCHI-ANALYSIS 3

Case STAT-DOE-TAGUCHI-ANALYSIS-003 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_TAGUCHI.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runs/0/countexact330.00E+2passed
/runs/0/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/0/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/0/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/0/sntolerance2.34083206033367962.34083206033368059.20E-1615passed
/runs/1/countexact330.00E+2passed
/runs/1/meantolerance4.3333333333333334.3333333333333333.73E-1717passed
/runs/1/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/1/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/1/sntolerance2.34083206033367962.34083206033367963.17E-1716passed
/runs/2/countexact330.00E+2passed
/runs/2/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/2/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/2/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/2/sntolerance2.34083206033367962.34083206033368059.20E-1615passed
/runs/3/countexact330.00E+2passed
/runs/3/meantolerance8.33333333333333398.3333333333333342.55E-1717passed
/runs/3/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/3/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/3/sntolerance2.34083206033367962.34083206033368059.20E-1615passed
/runs/4/countexact330.00E+2passed
/runs/4/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/4/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/4/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/4/sntolerance2.34083206033367962.34083206033368059.20E-1615passed
/runs/5/countexact330.00E+2passed
/runs/5/meantolerance5.03333333333333325.0333333333333349.03E-1615passed
/runs/5/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/5/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/5/sntolerance2.34083206033367962.34083206033368059.20E-1615passed
/runs/6/countexact330.00E+2passed
/runs/6/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/6/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/6/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/6/sntolerance2.34083206033367962.34083206033368059.20E-1615passed
/runs/7/countexact330.00E+2passed
/runs/7/meantolerance9.03333333333333329.0333333333333331.49E-1717passed
/runs/7/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/7/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/7/sntolerance2.34083206033367962.34083206033368059.20E-1615passed
/runs/8/countexact330.00E+2passed
/runs/8/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/8/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/8/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/8/sntolerance2.34083206033367962.34083206033367963.17E-1716passed
/runs/9/countexact330.00E+2passed
/runs/9/meantolerance5.63333333333333295.6333333333333334.04E-1717passed
/runs/9/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/9/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/9/sntolerance2.34083206033367962.34083206033367963.17E-1716passed
/runs/10/countexact330.00E+2passed
/runs/10/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/10/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/10/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/10/sntolerance2.34083206033367962.34083206033367963.17E-1716passed
/runs/11/countexact330.00E+2passed
/runs/11/meantolerance9.63333333333333469.6333333333333353.60E-1717passed
/runs/11/sdtolerance0.763762615825973380.7637626158259746.65E-1615passed
/runs/11/log_sdtolerance-0.26949825036634345-0.269498250366342568.88E-1614passed
/runs/11/sntolerance2.34083206033367962.3408320603336718.41E-1514passed
/tables/[summary=mean]/averages/0/0tolerance33.00.00E+2passed
/tables/[summary=mean]/averages/0/1tolerance77.00.00E+2passed
/tables/[summary=mean]/averages/1/0tolerance44.00.00E+2passed
/tables/[summary=mean]/averages/1/1tolerance66.00.00E+2passed
/predictions/[summary=mean]/valuetolerance66.00.00E+2passed
/tables/[summary=sd]/averages/0/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/0/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/tables/[summary=sd]/averages/1/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/1/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/predictions/[summary=sd]/valuetolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=log_sd]/averages/0/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/0/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/tables/[summary=log_sd]/averages/1/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/1/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/predictions/[summary=log_sd]/valuetolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=sn]/averages/0/0tolerance2.34083206033367962.34083206033368059.20E-1615passed
/tables/[summary=sn]/averages/0/1tolerance2.34083206033367962.34083206033367878.56E-1615passed
/tables/[summary=sn]/averages/1/0tolerance2.34083206033367962.340832060333684.76E-1615passed
/tables/[summary=sn]/averages/1/1tolerance2.34083206033367962.34083206033367878.56E-1615passed
/predictions/[summary=sn]/valuetolerance2.34083206033367962.340832060333684.76E-1615passed
/anova/[source=A]/dfexact110.00E+2passed
/anova/[source=A]/sstolerance48.00000000000000748.07.00E-1515passed
/anova/[source=A]/ftolerance28.07625649913344628.0762564991334182.80E-1415passed
/anova/[source=A]/ptolerance0.000494704573913236380.00049470457391323811.73E-1814passed
/anova/[source=B]/dfexact110.00E+2passed
/anova/[source=B]/sstolerance12.00000000000000512.05.00E-1515passed
/anova/[source=B]/ftolerance7.01906412478336347.01906412478335458.89E-1514passed
/anova/[source=B]/ptolerance0.0265043503601723570.026504350360172461.04E-1614passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


TAGUCHI-ANALYSIS 4

Case STAT-DOE-TAGUCHI-ANALYSIS-004 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_TAGUCHI.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runs/0/countexact330.00E+2passed
/runs/0/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/0/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/0/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/0/sntolerance9.70036776622556839.700367766225571.77E-1515passed
/runs/1/countexact330.00E+2passed
/runs/1/meantolerance4.3333333333333334.3333333333333333.73E-1717passed
/runs/1/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/1/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/1/sntolerance15.07727401207716715.0772740120771671.13E-1617passed
/runs/2/countexact330.00E+2passed
/runs/2/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/2/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/2/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/2/sntolerance9.70036776622556839.700367766225571.77E-1515passed
/runs/3/countexact330.00E+2passed
/runs/3/meantolerance8.33333333333333398.3333333333333342.55E-1717passed
/runs/3/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/3/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/3/sntolerance20.75720713938118320.7572071393811832.25E-1616passed
/runs/4/countexact330.00E+2passed
/runs/4/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/4/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/4/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/4/sntolerance11.97923481236230411.9792348123623051.36E-1515passed
/runs/5/countexact330.00E+2passed
/runs/5/meantolerance5.03333333333333325.0333333333333349.03E-1615passed
/runs/5/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/5/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/5/sntolerance16.3779459118038216.377945911803824.22E-1616passed
/runs/6/countexact330.00E+2passed
/runs/6/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/6/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/6/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/6/sntolerance11.97923481236230411.9792348123623051.36E-1515passed
/runs/7/countexact330.00E+2passed
/runs/7/meantolerance9.03333333333333329.0333333333333331.49E-1717passed
/runs/7/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/7/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/7/sntolerance21.45779278342854521.4577927834285453.19E-1616passed
/runs/8/countexact330.00E+2passed
/runs/8/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/8/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/8/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/8/sntolerance13.54693692475290413.5469369247529021.66E-1515passed
/runs/9/countexact330.00E+2passed
/runs/9/meantolerance5.63333333333333295.6333333333333334.04E-1717passed
/runs/9/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/9/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/9/sntolerance17.356141058213917.35614105821394.08E-1616passed
/runs/10/countexact330.00E+2passed
/runs/10/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/10/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/10/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/10/sntolerance13.54693692475290413.5469369247529021.66E-1515passed
/runs/11/countexact330.00E+2passed
/runs/11/meantolerance9.63333333333333469.6333333333333353.60E-1717passed
/runs/11/sdtolerance0.763762615825973380.7637626158259746.65E-1615passed
/runs/11/log_sdtolerance-0.26949825036634345-0.269498250366342568.88E-1614passed
/runs/11/sntolerance22.01636382107139122.016363821071381.11E-1415passed
/tables/[summary=mean]/averages/0/0tolerance33.00.00E+2passed
/tables/[summary=mean]/averages/0/1tolerance77.00.00E+2passed
/tables/[summary=mean]/averages/1/0tolerance44.00.00E+2passed
/tables/[summary=mean]/averages/1/1tolerance66.00.00E+2passed
/predictions/[summary=mean]/valuetolerance66.00.00E+2passed
/tables/[summary=sd]/averages/0/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/0/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/tables/[summary=sd]/averages/1/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/1/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/predictions/[summary=sd]/valuetolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=log_sd]/averages/0/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/0/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/tables/[summary=log_sd]/averages/1/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/1/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/predictions/[summary=log_sd]/valuetolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=sn]/averages/0/0tolerance11.74217983444692611.7421798344469241.85E-1515passed
/tables/[summary=sn]/averages/0/1tolerance18.84045412099618.8404541209959973.41E-1515passed
/tables/[summary=sn]/averages/1/0tolerance14.0063167475726114.0063167475726121.80E-1515passed
/tables/[summary=sn]/averages/1/1tolerance16.57631720787031416.5763172078703142.63E-1616passed
/predictions/[summary=sn]/valuetolerance17.55545389084714717.5554538908471443.42E-1515passed
/anova/[source=A]/dfexact110.00E+2passed
/anova/[source=A]/sstolerance48.00000000000000748.07.00E-1515passed
/anova/[source=A]/ftolerance28.07625649913344628.0762564991334182.80E-1415passed
/anova/[source=A]/ptolerance0.000494704573913236380.00049470457391323811.73E-1814passed
/anova/[source=B]/dfexact110.00E+2passed
/anova/[source=B]/sstolerance12.00000000000000512.05.00E-1515passed
/anova/[source=B]/ftolerance7.01906412478336347.01906412478335458.89E-1514passed
/anova/[source=B]/ptolerance0.0265043503601723570.026504350360172461.04E-1614passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


TAGUCHI-ANALYSIS 5

Case STAT-DOE-TAGUCHI-ANALYSIS-005 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_TAGUCHI.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runs/0/countexact330.00E+2passed
/runs/0/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/0/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/0/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/0/sntolerance-4.2596873227227814-4.2596873227228082.66E-1414passed
/runs/0/slopetolerance-0.50000000000000122-0.50000000000000038.87E-1614passed
/runs/1/countexact330.00E+2passed
/runs/1/meantolerance4.3333333333333334.3333333333333333.73E-1717passed
/runs/1/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/1/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/1/sntolerance-4.2596873227228009-4.2596873227228161.51E-1414passed
/runs/1/slopetolerance-0.50000000000000044-0.499999999999999836.07E-1614passed
/runs/2/countexact330.00E+2passed
/runs/2/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/2/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/2/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/2/sntolerance-4.2596873227227814-4.2596873227228082.66E-1414passed
/runs/2/slopetolerance-0.50000000000000122-0.50000000000000038.87E-1614passed
/runs/3/countexact330.00E+2passed
/runs/3/meantolerance8.33333333333333398.3333333333333342.55E-1717passed
/runs/3/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/3/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/3/sntolerance-4.2596873227227441-4.2596873227228136.93E-1413passed
/runs/3/slopetolerance-0.50000000000000266-0.49999999999999943.27E-1514passed
/runs/4/countexact330.00E+2passed
/runs/4/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/4/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/4/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/4/sntolerance-4.259687322722792-4.25968732272281252.05E-1414passed
/runs/4/slopetolerance-0.50000000000000078-0.499999999999999839.47E-1614passed
/runs/5/countexact330.00E+2passed
/runs/5/meantolerance5.03333333333333325.0333333333333349.03E-1615passed
/runs/5/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/5/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/5/sntolerance-4.2596873227227814-4.2596873227228133.20E-1414passed
/runs/5/slopetolerance-0.50000000000000122-0.499999999999999671.55E-1514passed
/runs/6/countexact330.00E+2passed
/runs/6/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/6/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/6/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/6/sntolerance-4.259687322722792-4.25968732272281252.05E-1414passed
/runs/6/slopetolerance-0.50000000000000078-0.499999999999999839.47E-1614passed
/runs/7/countexact330.00E+2passed
/runs/7/meantolerance9.03333333333333329.0333333333333331.49E-1717passed
/runs/7/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/7/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/7/sntolerance-4.2596873227227814-4.2596873227227875.32E-1514passed
/runs/7/slopetolerance-0.50000000000000122-0.5000000000000012.21E-1615passed
/runs/8/countexact330.00E+2passed
/runs/8/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/8/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/8/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/8/sntolerance-4.2596873227227814-4.25968732272281253.11E-1414passed
/runs/8/slopetolerance-0.50000000000000122-0.51.22E-1514passed
/runs/9/countexact330.00E+2passed
/runs/9/meantolerance5.63333333333333295.6333333333333334.04E-1717passed
/runs/9/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/9/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/9/sntolerance-4.259687322722761-4.2596873227228044.26E-1414passed
/runs/9/slopetolerance-0.500000000000002-0.50000000000000041.56E-1514passed
/runs/10/countexact330.00E+2passed
/runs/10/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/10/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/10/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/10/sntolerance-4.2596873227227814-4.25968732272281253.11E-1414passed
/runs/10/slopetolerance-0.50000000000000122-0.51.22E-1514passed
/runs/11/countexact330.00E+2passed
/runs/11/meantolerance9.63333333333333469.6333333333333353.60E-1717passed
/runs/11/sdtolerance0.763762615825973380.7637626158259746.65E-1615passed
/runs/11/log_sdtolerance-0.26949825036634345-0.269498250366342568.88E-1614passed
/runs/11/sntolerance-4.2596873227227441-4.2596873227228227.81E-1413passed
/runs/11/slopetolerance-0.50000000000000266-0.50000000000000022.44E-1514passed
/tables/[summary=mean]/averages/0/0tolerance33.00.00E+2passed
/tables/[summary=mean]/averages/0/1tolerance77.00.00E+2passed
/tables/[summary=mean]/averages/1/0tolerance44.00.00E+2passed
/tables/[summary=mean]/averages/1/1tolerance66.00.00E+2passed
/predictions/[summary=mean]/valuetolerance66.00.00E+2passed
/tables/[summary=sd]/averages/0/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/0/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/tables/[summary=sd]/averages/1/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/1/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/predictions/[summary=sd]/valuetolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=log_sd]/averages/0/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/0/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/tables/[summary=log_sd]/averages/1/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/1/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/predictions/[summary=log_sd]/valuetolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=sn]/averages/0/0tolerance-4.2596873227227849-4.2596873227228112.58E-1414passed
/tables/[summary=sn]/averages/0/1tolerance-4.259687322722769-4.259687322722814.08E-1414passed
/tables/[summary=sn]/averages/1/0tolerance-4.2596873227227832-4.2596873227228112.75E-1414passed
/tables/[summary=sn]/averages/1/1tolerance-4.2596873227227707-4.2596873227228093.82E-1414passed
/predictions/[summary=sn]/valuetolerance-4.2596873227227752-4.259687322722813.46E-1414passed
/anova/[source=A]/dfexact110.00E+2passed
/anova/[source=A]/sstolerance48.00000000000000748.07.00E-1515passed
/anova/[source=A]/ftolerance28.07625649913344628.0762564991334182.80E-1415passed
/anova/[source=A]/ptolerance0.000494704573913236380.00049470457391323811.73E-1814passed
/anova/[source=B]/dfexact110.00E+2passed
/anova/[source=B]/sstolerance12.00000000000000512.05.00E-1515passed
/anova/[source=B]/ftolerance7.01906412478336347.01906412478335458.89E-1514passed
/anova/[source=B]/ptolerance0.0265043503601723570.026504350360172461.04E-1614passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


TAGUCHI-ANALYSIS 6

Case STAT-DOE-TAGUCHI-ANALYSIS-006 · status: passed

Reference

Evidence type
R
Source
R stats: lm, anova, pf, pt, qt, binom.test; base mean/var/log10
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
reference_data/datasets/DOE_B10_TAGUCHI.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runs/0/countexact330.00E+2passed
/runs/0/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/0/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/0/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/0/sntolerance-4.9802492768137476-4.9802492768137461.75E-1515passed
/runs/0/slopetolerance0.92857142857142860.92857142857142863.15E-1817passed
/runs/1/countexact330.00E+2passed
/runs/1/meantolerance4.3333333333333334.3333333333333333.73E-1717passed
/runs/1/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/1/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/1/sntolerance-3.0449052777348786-3.0449052777348771.75E-1515passed
/runs/1/slopetolerance1.78571428571428541.78571428571428584.09E-1615passed
/runs/2/countexact330.00E+2passed
/runs/2/meantolerance2.33333333333333352.33333333333333351.86E-1717passed
/runs/2/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/2/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/2/sntolerance-4.9802492768137476-4.9802492768137461.75E-1515passed
/runs/2/slopetolerance0.92857142857142860.92857142857142863.15E-1817passed
/runs/3/countexact330.00E+2passed
/runs/3/meantolerance8.33333333333333398.3333333333333342.55E-1717passed
/runs/3/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/3/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/3/sntolerance-1.9061751225227783-1.9061751225227771.31E-1515passed
/runs/3/slopetolerance3.53.50.00E+1passed
/runs/4/countexact330.00E+2passed
/runs/4/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/4/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/4/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/4/sntolerance-4.0318132596348093-4.0318132596348072.67E-1515passed
/runs/4/slopetolerance1.22857142857142861.22857142857142892.70E-1615passed
/runs/5/countexact330.00E+2passed
/runs/5/meantolerance5.03333333333333325.0333333333333349.03E-1615passed
/runs/5/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/5/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/5/sntolerance-2.7177165343131948-2.7177165343131948.68E-1615passed
/runs/5/slopetolerance2.08571428571428542.0857142857142864.54E-1615passed
/runs/6/countexact330.00E+2passed
/runs/6/meantolerance3.03333333333333373.03333333333333374.10E-1716passed
/runs/6/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/6/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/6/sntolerance-4.0318132596348093-4.0318132596348072.67E-1515passed
/runs/6/slopetolerance1.22857142857142861.22857142857142892.70E-1615passed
/runs/7/countexact330.00E+2passed
/runs/7/meantolerance9.03333333333333329.0333333333333331.49E-1717passed
/runs/7/sdtolerance0.763762615825973380.76376261582597331.12E-1615passed
/runs/7/log_sdtolerance-0.26949825036634345-0.26949825036634361.67E-1615passed
/runs/7/sntolerance-1.8097606428057389-1.80976064280573761.34E-1515passed
/runs/7/slopetolerance3.79999999999999943.84.22E-1615passed
/runs/8/countexact330.00E+2passed
/runs/8/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/8/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/8/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/8/sntolerance-3.4929082175886195-3.49290821758861953.17E-1717passed
/runs/8/slopetolerance1.48571428571428531.48571428571428552.43E-1615passed
/runs/9/countexact330.00E+2passed
/runs/9/meantolerance5.63333333333333295.6333333333333334.04E-1717passed
/runs/9/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/9/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/9/sntolerance-2.5005354267411701-2.50053542674116841.75E-1515passed
/runs/9/slopetolerance2.34285714285714252.34285714285714252.72E-1716passed
/runs/10/countexact330.00E+2passed
/runs/10/meantolerance3.63333333333333333.6333333333333334.40E-1615passed
/runs/10/sdtolerance0.763762615825973380.76376261582597341.15E-1817passed
/runs/10/log_sdtolerance-0.26949825036634345-0.269498250366343453.24E-1917passed
/runs/10/sntolerance-3.4929082175886195-3.49290821758861953.17E-1717passed
/runs/10/slopetolerance1.48571428571428531.48571428571428552.43E-1615passed
/runs/11/countexact330.00E+2passed
/runs/11/meantolerance9.63333333333333469.6333333333333353.60E-1717passed
/runs/11/sdtolerance0.763762615825973380.7637626158259746.65E-1615passed
/runs/11/log_sdtolerance-0.26949825036634345-0.269498250366342568.88E-1614passed
/runs/11/sntolerance-1.738256528936456-1.73825652893645243.56E-1514passed
/runs/11/slopetolerance4.05714285714285694.0571428571428589.28E-1615passed
/tables/[summary=mean]/averages/0/0tolerance33.00.00E+2passed
/tables/[summary=mean]/averages/0/1tolerance77.00.00E+2passed
/tables/[summary=mean]/averages/1/0tolerance44.00.00E+2passed
/tables/[summary=mean]/averages/1/1tolerance66.00.00E+2passed
/predictions/[summary=mean]/valuetolerance66.00.00E+2passed
/tables/[summary=sd]/averages/0/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/0/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/tables/[summary=sd]/averages/1/0tolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=sd]/averages/1/1tolerance0.763762615825973380.76376261582597351.10E-1615passed
/predictions/[summary=sd]/valuetolerance0.763762615825973380.76376261582597331.12E-1615passed
/tables/[summary=log_sd]/averages/0/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/0/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/tables/[summary=log_sd]/averages/1/0tolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=log_sd]/averages/1/1tolerance-0.26949825036634345-0.26949825036634345.52E-1715passed
/predictions/[summary=log_sd]/valuetolerance-0.26949825036634345-0.269498250366343561.11E-1615passed
/tables/[summary=sn]/averages/0/0tolerance-4.1683235846790589-4.1683235846790571.73E-1515passed
/tables/[summary=sn]/averages/0/1tolerance-2.2862249221757027-2.2862249221757011.75E-1515passed
/tables/[summary=sn]/averages/1/0tolerance-3.4613546654710698-3.4613546654710698.41E-1615passed
/tables/[summary=sn]/averages/1/1tolerance-2.9931938413836914-2.993193841383691.35E-1515passed
/predictions/[summary=sn]/valuetolerance-2.5203053342193913-2.52030533421939048.97E-1615passed
/anova/[source=A]/dfexact110.00E+2passed
/anova/[source=A]/sstolerance48.00000000000000748.07.00E-1515passed
/anova/[source=A]/ftolerance28.07625649913344628.0762564991334182.80E-1415passed
/anova/[source=A]/ptolerance0.000494704573913236380.00049470457391323811.73E-1814passed
/anova/[source=B]/dfexact110.00E+2passed
/anova/[source=B]/sstolerance12.00000000000000512.05.00E-1515passed
/anova/[source=B]/ftolerance7.01906412478336347.01906412478335458.89E-1514passed
/anova/[source=B]/ptolerance0.0265043503601723570.026504350360172461.04E-1614passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


TAGUCHI-DESIGN 1

Case STAT-DOE-TAGUCHI-DESIGN-001 · status: passed

Reference

Evidence type
Analytical
Source
Hand-checkable full crossing and fixed-input arithmetic
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact440.00E+2passed
/rows/0/2exact110.00E+2passed
/rows/0/3exact110.00E+2passed
/rows/0/4exact110.00E+2passed
/rows/1/2exact110.00E+2passed
/rows/1/3exact220.00E+2passed
/rows/1/4exact220.00E+2passed
/rows/2/2exact220.00E+2passed
/rows/2/3exact110.00E+2passed
/rows/2/4exact220.00E+2passed
/rows/3/2exact220.00E+2passed
/rows/3/3exact220.00E+2passed
/rows/3/4exact110.00E+2passed
/response_columnssequence['Y_R1', 'Y_R2']['Y_R1', 'Y_R2']passed
/assignment/0/levelsexact220.00E+2passed
/assignment/1/levelsexact220.00E+2passed
/assignment/2/levelsexact220.00E+2passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


TAGUCHI-DESIGN 2

Case STAT-DOE-TAGUCHI-DESIGN-002 · status: passed

Reference

Evidence type
Analytical
Source
Hand-checkable full crossing and fixed-input arithmetic
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact990.00E+2passed
/rows/0/2exact110.00E+2passed
/rows/0/3exact110.00E+2passed
/rows/0/4exact110.00E+2passed
/rows/1/2exact110.00E+2passed
/rows/1/3exact220.00E+2passed
/rows/1/4exact220.00E+2passed
/rows/2/2exact110.00E+2passed
/rows/2/3exact330.00E+2passed
/rows/2/4exact330.00E+2passed
/rows/3/2exact220.00E+2passed
/rows/3/3exact110.00E+2passed
/rows/3/4exact220.00E+2passed
/rows/4/2exact220.00E+2passed
/rows/4/3exact220.00E+2passed
/rows/4/4exact330.00E+2passed
/rows/5/2exact220.00E+2passed
/rows/5/3exact330.00E+2passed
/rows/5/4exact110.00E+2passed
/rows/6/2exact330.00E+2passed
/rows/6/3exact110.00E+2passed
/rows/6/4exact330.00E+2passed
/rows/7/2exact330.00E+2passed
/rows/7/3exact220.00E+2passed
/rows/7/4exact110.00E+2passed
/rows/8/2exact330.00E+2passed
/rows/8/3exact330.00E+2passed
/rows/8/4exact220.00E+2passed
/response_columnssequence['Y_R1', 'Y_R2']['Y_R1', 'Y_R2']passed
/assignment/0/levelsexact330.00E+2passed
/assignment/1/levelsexact330.00E+2passed
/assignment/2/levelsexact330.00E+2passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


TAGUCHI-DESIGN 3

Case STAT-DOE-TAGUCHI-DESIGN-003 · status: passed

Reference

Evidence type
Analytical
Source
Hand-checkable full crossing and fixed-input arithmetic
Citation
R Core Team, stats package documentation: lm, anova, FDist, TDistr, binom.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_window_coverage.R · sha256 994610a525207dcf
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact440.00E+2passed
/rows/0/2exact110.00E+2passed
/rows/0/3exact110.00E+2passed
/rows/0/4exact110.00E+2passed
/rows/1/2exact110.00E+2passed
/rows/1/3exact220.00E+2passed
/rows/1/4exact220.00E+2passed
/rows/2/2exact220.00E+2passed
/rows/2/3exact110.00E+2passed
/rows/2/4exact220.00E+2passed
/rows/3/2exact220.00E+2passed
/rows/3/3exact220.00E+2passed
/rows/3/4exact110.00E+2passed
/response_columnssequence['Y_M1_O1', 'Y_M1_O2', 'Y_M1_O3', 'Y_M1_O4', 'Y_M2_O1', 'Y_M['Y_M1_O1', 'Y_M1_O2', 'Y_M1_O3', 'Y_M1_O4', 'Y_M2_O1', 'Y_Mpassed
/assignment/0/levelsexact220.00E+2passed
/assignment/1/levelsexact220.00E+2passed
/assignment/2/levelsexact220.00E+2passed

Convention

Reference code

Whole script generators/r/design_window_coverage.R


Log-variance model slopes with equal replication

Case STAT-DOE-VAR-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::glm on coded factors; enumerated design points
Citation
R stats::lm and stats::glm documentation; coding rules from manual chapter 13
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/design_doe.R · sha256 ec14480ae0666aea
Dataset
reference_data/datasets/DOE_VARIABILITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/payload/variability/terms/[term=A]/coefficienttolerance0.692428637275692260.69242863727569241.11E-1615passed
/payload/variability/terms/[term=B]/coefficienttolerance-0.70814748653806969-0.70814748653806952.19E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

152  set.seed(20261103)
153  dv <- data.frame(A = rep(c(10, 20, 10, 20), each = 5), B = rep(c(1, 1, 3, 3), each = 5))
154  sdv <- exp(0.2 + 0.35 * code(dv$A, 10, 20) - 0.25 * code(dv$B, 1, 3))
155  dv$y <- round(30 + rnorm(20, 0, sdv), 4)
156  export_dataset(dv, "DOE_VARIABILITY", "Constructed 2^2 design with five runs per setting and a settings-dependent SD, seed 20261103",
157                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.", c(A = "Float64", B = "Float64", y = "Float64"))
158  v <- read_dataset("DOE_VARIABILITY")
159  agg <- aggregate(y ~ A + B, data = v, FUN = var)
160  agg$a <- code(agg$A, 10, 20); agg$b <- code(agg$B, 1, 3)
161  lv <- coef(lm(log(y) ~ a + b, data = agg))
162  emit("STAT-DOE-VAR-001", "tabelz.analysis.doe_dispersion:analyze_doe_variability", "Log-variance model slopes with equal replication", "DOE_VARIABILITY",
163       list(response = "y", design = list(factors = list(list(name = "A", low = 10, high = 20), list(name = "B", low = 1, high = 3)),
164                                          response_name = "y", model = "main_effects")),
165       list("/payload/variability/terms/[term=A]/coefficient" = tol(lv[["a"]], "Slope of log s^2 on coded A."),
166            "/payload/variability/terms/[term=B]/coefficient" = tol(lv[["b"]], "Slope of log s^2 on coded B.")),
167       c("Five runs at every setting: equal df, so weighting and the digamma bias correction shift only the intercept."),
168       deferred = list(list(path = "/payload/variability/terms/[term=Intercept]/coefficient",
169                            reason = "The manual does not give the bias correction or weights that set the intercept.")))

Whole script generators/r/design_doe.R


Crossed worksheet: 3 parts x 2 operators x 2 trials

Case STAT-DOE-WORKSHEET-001 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation of the worksheet design
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/design_worksheets.R · sha256 536f8a134b709491
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rowsexact12120.00E+2passed
/cellssequence['P1|Ann|1', 'P1|Ann|2', 'P1|Bob|1', 'P1|Bob|2', 'P2|Ann|1',['P1|Ann|1', 'P1|Ann|2', 'P1|Bob|1', 'P1|Bob|2', 'P2|Ann|1',passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

40  parts <- c("P1", "P2", "P3"); ops <- c("Ann", "Bob")
41  crossed <- list(layout = "crossed", part_labels = as.list(parts), operator_labels = as.list(ops), trials = 2L, seed = 11L)
42  emit("STAT-DOE-WORKSHEET-001", ws, "Crossed worksheet: 3 parts x 2 operators x 2 trials", crossed,
43       list("/rows" = exact(12L, "3 x 2 x 2 rows."), "/cells" = cells_of(parts, ops, 2)),
44       c("Line 17-18: crossed assigns every operator to every part for every trial."))

Whole script generators/r/design_worksheets.R


The same seed reproduces the sheet

Case STAT-DOE-WORKSHEET-002 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation of the worksheet design
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/design_worksheets.R · sha256 536f8a134b709491
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rowsexact12120.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

45  emit("STAT-DOE-WORKSHEET-002", ws, "The same seed reproduces the sheet", crossed,
46       list("/rows" = exact(12L, "Twelve rows.")), c("Line 43-45: the same design and seed reproduce the exact run order."),
47       kind = "determinism", repeat_n = 2L)

Whole script generators/r/design_worksheets.R


Nested worksheet: two parts per operator

Case STAT-DOE-WORKSHEET-003 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation of the worksheet design
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/design_worksheets.R · sha256 536f8a134b709491
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rowsexact18180.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

48  emit("STAT-DOE-WORKSHEET-003", ws, "Nested worksheet: two parts per operator",
49       list(layout = "nested", part_count = 2L, operator_labels = list("Ann", "Bob", "Cy"), trials = 3L, seed = 5L),
50       list("/rows" = exact(18L, "3 operators x 2 own parts x 3 trials.")),
51       c("Line 19-21: each operator gets a distinct set of parts."), kind = "axis_coverage")

Whole script generators/r/design_worksheets.R


Attribute agreement worksheet with standards

Case STAT-DOE-WORKSHEET-004 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation of the worksheet design
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/design_worksheets.R · sha256 536f8a134b709491
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rowsexact660.00E+2passed
/cellssequence['P1|Ann|1', 'P1|Bob|1', 'P2|Ann|1', 'P2|Bob|1', 'P3|Ann|1',['P1|Ann|1', 'P1|Bob|1', 'P2|Ann|1', 'P2|Bob|1', 'P3|Ann|1',passed
/columnsset['Part', 'Operator', 'Trial', 'TrialID', 'Value', 'StandardV['Part', 'Operator', 'Trial', 'TrialID', 'Value', 'StandardVpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

52  emit("STAT-DOE-WORKSHEET-004", ws, "Attribute agreement worksheet with standards",
53       list(layout = "attribute_agreement", part_labels = as.list(parts), operator_labels = as.list(ops), trials = 1L,
54            standard_values = list("G", "NG", "G")),
55       list("/rows" = exact(6L, "3 x 2 x 1 rating rows."), "/cells" = cells_of(parts, ops, 1),
56            "/columns" = list(comparison = "set", values = list("Part", "Operator", "Trial", "TrialID", "Value", "StandardValue"),
57                              rationale = "Line 45-48: Value replaces Measurement, plus StandardValue when declared.")),
58       c("One rating row per part/operator/trial."), kind = "axis_coverage")

Whole script generators/r/design_worksheets.R


Part labels are refused for nested designs

Case STAT-DOE-WORKSHEET-005 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation of the worksheet design
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/design_worksheets.R · sha256 536f8a134b709491
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

59  emit("STAT-DOE-WORKSHEET-005", ws, "Part labels are refused for nested designs",
60       list(layout = "nested", part_labels = list("P1", "P2"), operator_count = 2L, trials = 2L), list(),
61       c("Line 20-21: part labels cannot be declared for nested designs."), kind = "invalid_input",
62       error = list(error_class = "ValueError", message_contains = "ested"))

Whole script generators/r/design_worksheets.R


The recorded operation recreates the worksheet

Case STAT-DOE-WORKSHEET-006 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation of the worksheet design
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/design_worksheets.R · sha256 536f8a134b709491
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rowsexact12120.00E+2passed
/cellssequence['P1|Ann|1', 'P1|Ann|2', 'P1|Bob|1', 'P1|Bob|2', 'P2|Ann|1',['P1|Ann|1', 'P1|Ann|2', 'P1|Bob|1', 'P1|Bob|2', 'P2|Ann|1',passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

63  emit("STAT-DOE-WORKSHEET-006", op, "The recorded operation recreates the worksheet", crossed,
64       list("/rows" = exact(12L, "Twelve rows."), "/cells" = cells_of(parts, ops, 2)),
65       c("Line 50-54: the adopted sheet is a separate table with its own recorded operation."), kind = "determinism", repeat_n = 2L)

Whole script generators/r/design_worksheets.R


One categorical column separates the two modes

Case STAT-EXP-MODEEXPLAIN-001 · status: passed

Reference

Evidence type
Analytical
Source
R 4.6.1 base and stats
Citation
R stats::lm, stats::quantile (type 7), stats::t.test, stats::ecdf documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/explore_plots.R · sha256 15466ea60c37fc33
Dataset
reference_data/datasets/EXP_MODES_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/mode_countexact220.00E+2passed
/root/split_featureexactdriverdriverpassed
/candidates/0/featureexactdriverdriverpassed
/candidates/0/gain_normalizedtolerance11.00.00E+2passed
/rules/[mode=0]/conditions/[feature=driver]/labelexact= A= Apassed
/rules/[mode=0]/supportexact48480.00E+2passed
/rules/[mode=0]/puritytolerance11.00.00E+2passed
/rules/[mode=0]/validation_puritytolerance11.00.00E+2passed
/rules/[mode=1]/conditions/[feature=driver]/labelexact= B= Bpassed
/rules/[mode=1]/supportexact48480.00E+2passed
/rules/[mode=1]/puritytolerance11.00.00E+2passed
/rules/2absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

258  me <- "tabelz.analysis.mode_explainer:explain_modes"; mm <- "explainMultimodalityMenuItem"
259  cells1 <- expand.grid(i = 1:12, driver = c("A", "B"), kind = c("CAT", "DOG"), shift = c("day", "night"),
260                        stringsAsFactors = FALSE)
261  single <- data.frame(value = ifelse(cells1$driver == "B", 12, -12) + 1.2 * qnorm(ppoints(12))[cells1$i],
262                       driver = cells1$driver, unrelated = c("p", "q", "r")[(cells1$i - 1) %% 3 + 1])
263  export_dataset(single, "EXP_MODES_01",
264                 "Constructed: 96 rows, two modes at -12 and 12 separated exactly by driver; unrelated is balanced within each mode",
265                 "Hand-constructed", "Constructed validation data; no third-party license.",
266                 c(value = "Float64", driver = "String", unrelated = "String"))
267  emit("STAT-EXP-MODEEXPLAIN-001", me, mm, "One categorical column separates the two modes",
268       list(column = "value"),
269       list("/mode_count" = exact(2L, "Two constructed modes 24 units apart."),
270            "/root/split_feature" = exact("driver", "driver alone separates the modes."),
271            "/candidates/0/feature" = exact("driver", "Ranked first by its own gain."),
272            "/candidates/0/gain_normalized" = num(1, "Perfect separation removes all mode entropy."),
273            "/rules/[mode=0]/conditions/[feature=driver]/label" = exact("= A", "Lower mode is driver A."),
274            "/rules/[mode=0]/support" = exact(48L, "All 48 A rows."),
275            "/rules/[mode=0]/purity" = num(1, "Pure on training rows."),
276            "/rules/[mode=0]/validation_purity" = num(1, "Pure on held-out rows."),
277            "/rules/[mode=1]/conditions/[feature=driver]/label" = exact("= B", "Upper mode is driver B."),
278            "/rules/[mode=1]/support" = exact(48L, "All 48 B rows."),
279            "/rules/[mode=1]/purity" = num(1, "Pure on training rows."),
280            "/rules/2" = absent("One rule per driver level.")),
281       c("Chapter 11: rows are labelled by density basin, modes numbered from the lowest values upward; each leaf is one rule.",
282         "A column that separates the modes exactly has normalized information gain 1 and yields pure rules."),
283       oracle_class = "analytic", dataset = "EXP_MODES_01")

Whole script generators/r/explore_plots.R


Only a two-column conjunction identifies the upper mode

Case STAT-EXP-MODEEXPLAIN-002 · status: passed

Reference

Evidence type
Analytical
Source
R 4.6.1 base and stats
Citation
R stats::lm, stats::quantile (type 7), stats::t.test, stats::ecdf documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/explore_plots.R · sha256 15466ea60c37fc33
Dataset
reference_data/datasets/EXP_MODES_02.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/mode_countexact220.00E+2passed
/rules/[mode=1]/conditions/[feature=kind]/labelexact= CAT= CATpassed
/rules/[mode=1]/conditions/[feature=shift]/labelexact= night= nightpassed
/rules/[mode=1]/conditions/2absentNonepassed
/rules/[mode=1]/supportexact40400.00E+2passed
/rules/[mode=1]/puritytolerance11.00.00E+2passed
/rules/[mode=1]/validation_puritytolerance11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

284  per <- 40
285  cells2 <- expand.grid(i = 1:per, kind = c("CAT", "DOG"), shift = c("day", "night"), stringsAsFactors = FALSE)
286  upper <- cells2$kind == "CAT" & cells2$shift == "night"
287  inter_value <- numeric(nrow(cells2))
288  inter_value[upper] <- 12 + 1.2 * qnorm(ppoints(sum(upper)))
289  inter_value[!upper] <- -12 + 1.2 * qnorm(ppoints(sum(!upper)))
290  inter <- data.frame(value = inter_value, kind = cells2$kind, shift = cells2$shift)
291  export_dataset(inter, "EXP_MODES_02",
292                 "Constructed: 160 rows; only kind = CAT AND shift = night (40 rows) forms the upper mode at 12",
293                 "Hand-constructed", "Constructed validation data; no third-party license.",
294                 c(value = "Float64", kind = "String", shift = "String"))
295  emit("STAT-EXP-MODEEXPLAIN-002", me, mm, "Only a two-column conjunction identifies the upper mode",
296       list(column = "value"),
297       list("/mode_count" = exact(2L, "Upper mode of 40 rows, lower mode of 120."),
298            "/rules/[mode=1]/conditions/[feature=kind]/label" = exact("= CAT", "Part of the conjunction."),
299            "/rules/[mode=1]/conditions/[feature=shift]/label" = exact("= night", "Part of the conjunction."),
300            "/rules/[mode=1]/conditions/2" = absent("Exactly two conditions."),
301            "/rules/[mode=1]/support" = exact(as.integer(sum(upper)), "The 40 constructed rows."),
302            "/rules/[mode=1]/purity" = num(1, "Pure on training rows."),
303            "/rules/[mode=1]/validation_purity" = num(1, "Pure on held-out rows.")),
304       c("Chapter 11: a mode identified only by a combination of columns appears as one rule with both conditions.",
305         "Neither column separates the modes alone; the split order is not asserted."),
306       oracle_class = "analytic", dataset = "EXP_MODES_02")

Whole script generators/r/explore_plots.R


Retained single-Gaussian fit: exact MLE, likelihood, BIC and row allocations

Case STAT-EXP-MODEFIT-001 · status: passed

Reference

Evidence type
R
Source
R diptest::dip and base/stats closed-form likelihood calculations
Citation
https://stat.ethz.ch/R-manual/R-devel/library/stats/html/Normal.html; one-Gaussian MLE mu=mean(x), variance=mean((x-mu)^2).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, diptest 0.77.2, jsonlite 2.0.0
Reference script
generators/r/explore_mode_screen_fit.R · sha256 6b0f4af3271659ab
Dataset
reference_data/datasets/EXP_MODE_CURRENT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/variableexactxxpassed
/mode_countexact110.00E+2passed
/row_countexact32320.00E+2passed
/rootabsentNonepassed
/training_countexact000.00E+2passed
/validation_countexact000.00E+2passed
/component_fit/selected_keyexactgaussian:1gaussian:1passed
/component_fit/candidates/0/parameter_countexact220.00E+2passed
/component_fit/candidates/0/log_likelihoodtolerance-90.007421827624981-90.007421827624981.08E-1617passed
/component_fit/candidates/0/bictolerance186.94631546084941186.94631546084944.90E-1516passed
/component_fit/candidates/0/components/0/familyexactgaussiangaussianpassed
/component_fit/candidates/0/components/0/weighttolerance11.00.00E+2passed
/component_fit/candidates/0/components/0/meantolerance4.163336342344337e-17-1.6820483361490517e-162.10E-16passed
/component_fit/candidates/0/components/0/variancetolerance16.24177777777777716.2417777777777772.03E-1616passed
/investigation/source_row_countexact34340.00E+2passed
/investigation/eligible_countexact32320.00E+2passed
/investigation/sampledexactFalseFalsepassed
/component_fit/posteriors/0/0tolerance11.00.00E+2passed
/component_fit/labels/0exact000.00E+2passed
/component_fit/posteriors/1/0tolerance11.00.00E+2passed
/component_fit/labels/1exact000.00E+2passed
/component_fit/posteriors/2/0tolerance11.00.00E+2passed
/component_fit/labels/2exact000.00E+2passed
/component_fit/posteriors/3/0tolerance11.00.00E+2passed
/component_fit/labels/3exact000.00E+2passed
/component_fit/posteriors/4/0tolerance11.00.00E+2passed
/component_fit/labels/4exact000.00E+2passed
/component_fit/posteriors/5/0tolerance11.00.00E+2passed
/component_fit/labels/5exact000.00E+2passed
/component_fit/posteriors/6/0tolerance11.00.00E+2passed
/component_fit/labels/6exact000.00E+2passed
/component_fit/posteriors/7/0tolerance11.00.00E+2passed
/component_fit/labels/7exact000.00E+2passed
/component_fit/posteriors/8/0tolerance11.00.00E+2passed
/component_fit/labels/8exact000.00E+2passed
/component_fit/posteriors/9/0tolerance11.00.00E+2passed
/component_fit/labels/9exact000.00E+2passed
/component_fit/posteriors/10/0tolerance11.00.00E+2passed
/component_fit/labels/10exact000.00E+2passed
/component_fit/posteriors/11/0tolerance11.00.00E+2passed
/component_fit/labels/11exact000.00E+2passed
/component_fit/posteriors/12/0tolerance11.00.00E+2passed
/component_fit/labels/12exact000.00E+2passed
/component_fit/posteriors/13/0tolerance11.00.00E+2passed
/component_fit/labels/13exact000.00E+2passed
/component_fit/posteriors/14/0tolerance11.00.00E+2passed
/component_fit/labels/14exact000.00E+2passed
/component_fit/posteriors/15/0tolerance11.00.00E+2passed
/component_fit/labels/15exact000.00E+2passed
/component_fit/posteriors/16/0tolerance11.00.00E+2passed
/component_fit/labels/16exact000.00E+2passed
/component_fit/posteriors/17/0tolerance11.00.00E+2passed
/component_fit/labels/17exact000.00E+2passed
/component_fit/posteriors/18/0tolerance11.00.00E+2passed
/component_fit/labels/18exact000.00E+2passed
/component_fit/posteriors/19/0tolerance11.00.00E+2passed
/component_fit/labels/19exact000.00E+2passed
/component_fit/posteriors/20/0tolerance11.00.00E+2passed
/component_fit/labels/20exact000.00E+2passed
/component_fit/posteriors/21/0tolerance11.00.00E+2passed
/component_fit/labels/21exact000.00E+2passed
/component_fit/posteriors/22/0tolerance11.00.00E+2passed
/component_fit/labels/22exact000.00E+2passed
/component_fit/posteriors/23/0tolerance11.00.00E+2passed
/component_fit/labels/23exact000.00E+2passed
/component_fit/posteriors/24/0tolerance11.00.00E+2passed
/component_fit/labels/24exact000.00E+2passed
/component_fit/posteriors/25/0tolerance11.00.00E+2passed
/component_fit/labels/25exact000.00E+2passed
/component_fit/posteriors/26/0tolerance11.00.00E+2passed
/component_fit/labels/26exact000.00E+2passed
/component_fit/posteriors/27/0tolerance11.00.00E+2passed
/component_fit/labels/27exact000.00E+2passed
/component_fit/posteriors/28/0tolerance11.00.00E+2passed
/component_fit/labels/28exact000.00E+2passed
/component_fit/posteriors/29/0tolerance11.00.00E+2passed
/component_fit/labels/29exact000.00E+2passed
/component_fit/posteriors/30/0tolerance11.00.00E+2passed
/component_fit/labels/30exact000.00E+2passed
/component_fit/posteriors/31/0tolerance11.00.00E+2passed
/component_fit/labels/31exact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

157  x <- finite("x")
158  variance <- mean((x - mean(x))^2)
159  ll <- sum(dnorm(x, mean(x), sqrt(variance), log = TRUE))
160  fit_notes <- c(paste0(manual, ":98-117: declared Gaussian/count candidates, likelihood, posteriors and BIC; fixed K is distinct from BIC selection."),
161                 paste0(manual, ":119-124: every included finite observation is fitted unless sampling is explicitly enabled."),
162                 "em-v2 is the recorded NumPy implementation of the same fit contract; these closed-form one-component cases do not qualify multi-component starts, local optima or BIC comparisons.")
163  emit("STAT-EXP-MODEFIT-001", "Retained single-Gaussian fit: exact MLE, likelihood, BIC and row allocations", "nominal",
164       fit_entry, list(column = "x", implementation = "em-v2", config = list(build_tree = FALSE,
165                         component_config = list(gaussian_k = 1L))),
166       fit_outputs("x", "gaussian", variance, ll, 2L), list(), fit_notes,
167       "https://stat.ethz.ch/R-manual/R-devel/library/stats/html/Normal.html; one-Gaussian MLE mu=mean(x), variance=mean((x-mu)^2).")

Whole script generators/r/explore_mode_screen_fit.R


Retained Poisson fit: exact rate, likelihood, BIC and row allocations

Case STAT-EXP-MODEFIT-002 · status: passed

Reference

Evidence type
R
Source
R diptest::dip and base/stats closed-form likelihood calculations
Citation
https://stat.ethz.ch/R-manual/R-devel/library/stats/html/Poisson.html; one-Poisson MLE lambda=sum(y)/n, variance=lambda.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, diptest 0.77.2, jsonlite 2.0.0
Reference script
generators/r/explore_mode_screen_fit.R · sha256 6b0f4af3271659ab
Dataset
reference_data/datasets/EXP_MODE_CURRENT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/variableexactcountcountpassed
/mode_countexact110.00E+2passed
/row_countexact32320.00E+2passed
/rootabsentNonepassed
/training_countexact000.00E+2passed
/validation_countexact000.00E+2passed
/component_fit/selected_keyexactpoissonpoissonpassed
/component_fit/candidates/0/parameter_countexact110.00E+2passed
/component_fit/candidates/0/log_likelihoodtolerance-73.910006712492816-73.910006712492824.99E-1718passed
/component_fit/candidates/0/bictolerance151.28574932778537151.285749327785372.23E-1516passed
/component_fit/candidates/0/components/0/familyexactpoissonpoissonpassed
/component_fit/candidates/0/components/0/weighttolerance11.00.00E+2passed
/component_fit/candidates/0/components/0/meantolerance3.53.50.00E+1passed
/component_fit/candidates/0/components/0/variancetolerance3.53.50.00E+1passed
/investigation/source_row_countexact34340.00E+2passed
/investigation/eligible_countexact32320.00E+2passed
/investigation/sampledexactFalseFalsepassed
/component_fit/posteriors/0/0tolerance11.00.00E+2passed
/component_fit/labels/0exact000.00E+2passed
/component_fit/posteriors/1/0tolerance11.00.00E+2passed
/component_fit/labels/1exact000.00E+2passed
/component_fit/posteriors/2/0tolerance11.00.00E+2passed
/component_fit/labels/2exact000.00E+2passed
/component_fit/posteriors/3/0tolerance11.00.00E+2passed
/component_fit/labels/3exact000.00E+2passed
/component_fit/posteriors/4/0tolerance11.00.00E+2passed
/component_fit/labels/4exact000.00E+2passed
/component_fit/posteriors/5/0tolerance11.00.00E+2passed
/component_fit/labels/5exact000.00E+2passed
/component_fit/posteriors/6/0tolerance11.00.00E+2passed
/component_fit/labels/6exact000.00E+2passed
/component_fit/posteriors/7/0tolerance11.00.00E+2passed
/component_fit/labels/7exact000.00E+2passed
/component_fit/posteriors/8/0tolerance11.00.00E+2passed
/component_fit/labels/8exact000.00E+2passed
/component_fit/posteriors/9/0tolerance11.00.00E+2passed
/component_fit/labels/9exact000.00E+2passed
/component_fit/posteriors/10/0tolerance11.00.00E+2passed
/component_fit/labels/10exact000.00E+2passed
/component_fit/posteriors/11/0tolerance11.00.00E+2passed
/component_fit/labels/11exact000.00E+2passed
/component_fit/posteriors/12/0tolerance11.00.00E+2passed
/component_fit/labels/12exact000.00E+2passed
/component_fit/posteriors/13/0tolerance11.00.00E+2passed
/component_fit/labels/13exact000.00E+2passed
/component_fit/posteriors/14/0tolerance11.00.00E+2passed
/component_fit/labels/14exact000.00E+2passed
/component_fit/posteriors/15/0tolerance11.00.00E+2passed
/component_fit/labels/15exact000.00E+2passed
/component_fit/posteriors/16/0tolerance11.00.00E+2passed
/component_fit/labels/16exact000.00E+2passed
/component_fit/posteriors/17/0tolerance11.00.00E+2passed
/component_fit/labels/17exact000.00E+2passed
/component_fit/posteriors/18/0tolerance11.00.00E+2passed
/component_fit/labels/18exact000.00E+2passed
/component_fit/posteriors/19/0tolerance11.00.00E+2passed
/component_fit/labels/19exact000.00E+2passed
/component_fit/posteriors/20/0tolerance11.00.00E+2passed
/component_fit/labels/20exact000.00E+2passed
/component_fit/posteriors/21/0tolerance11.00.00E+2passed
/component_fit/labels/21exact000.00E+2passed
/component_fit/posteriors/22/0tolerance11.00.00E+2passed
/component_fit/labels/22exact000.00E+2passed
/component_fit/posteriors/23/0tolerance11.00.00E+2passed
/component_fit/labels/23exact000.00E+2passed
/component_fit/posteriors/24/0tolerance11.00.00E+2passed
/component_fit/labels/24exact000.00E+2passed
/component_fit/posteriors/25/0tolerance11.00.00E+2passed
/component_fit/labels/25exact000.00E+2passed
/component_fit/posteriors/26/0tolerance11.00.00E+2passed
/component_fit/labels/26exact000.00E+2passed
/component_fit/posteriors/27/0tolerance11.00.00E+2passed
/component_fit/labels/27exact000.00E+2passed
/component_fit/posteriors/28/0tolerance11.00.00E+2passed
/component_fit/labels/28exact000.00E+2passed
/component_fit/posteriors/29/0tolerance11.00.00E+2passed
/component_fit/labels/29exact000.00E+2passed
/component_fit/posteriors/30/0tolerance11.00.00E+2passed
/component_fit/labels/30exact000.00E+2passed
/component_fit/posteriors/31/0tolerance11.00.00E+2passed
/component_fit/labels/31exact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

169  y <- finite("count")
170  lambda <- mean(y)
171  ll <- sum(dpois(y, lambda, log = TRUE))
172  emit("STAT-EXP-MODEFIT-002", "Retained Poisson fit: exact rate, likelihood, BIC and row allocations", "axis_coverage",
173       fit_entry, list(column = "count", implementation = "em-v2", config = list(build_tree = FALSE,
174                         component_config = list(measurement = "count", count_candidates = list("poisson")))),
175       fit_outputs("count", "poisson", lambda, ll, 1L), list(), fit_notes,
176       "https://stat.ethz.ch/R-manual/R-devel/library/stats/html/Poisson.html; one-Poisson MLE lambda=sum(y)/n, variance=lambda.")

Whole script generators/r/explore_mode_screen_fit.R


Current screen: independent dip and bandwidth on two continuous columns

Case STAT-EXP-MODESCREEN-001 · status: passed

Reference

Evidence type
R
Source
R diptest::dip and base/stats closed-form likelihood calculations
Citation
https://cran.r-project.org/web/packages/diptest/diptest.pdf (dip, min.is.0=FALSE); Hartigan and Hartigan (1985), Annals of Statistics 13:70-84. Bandwidth formula: manual section 110.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, diptest 0.77.2, jsonlite 2.0.0
Reference script
generators/r/explore_mode_screen_fit.R · sha256 6b0f4af3271659ab
Dataset
reference_data/datasets/EXP_MODE_CURRENT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact34340.00E+2passed
/excluded_countexact110.00E+2passed
/family_sizeexact220.00E+2passed
/family_alphatolerance0.0500000000000000030.052.24E-1917passed
/columns/[column=x]/observationsexact32320.00E+2passed
/columns/[column=x]/missingexact110.00E+2passed
/columns/[column=x]/dip/availableexactTrueTruepassed
/columns/[column=x]/dip/statistictolerance0.200000000000000010.21.10E-1817passed
/columns/[column=x]/dip/adjusted_p_valueabsentNonepassed
/columns/[column=x]/kde/availableexactTrueTruepassed
/columns/[column=x]/kde/standard_deviationtolerance4.09459474102016154.09459474102016152.13E-1717passed
/columns/[column=x]/kde/interquartile_rangetolerance7.94666666666666727.9466666666666671.14E-1717passed
/columns/[column=x]/kde/reference_bandwidthtolerance1.84256763345907281.84256763345907281.78E-1717passed
/columns/[column=z]/observationsexact32320.00E+2passed
/columns/[column=z]/missingexact110.00E+2passed
/columns/[column=z]/dip/availableexactTrueTruepassed
/columns/[column=z]/dip/statistictolerance0.0156250.0156250.00E-4passed
/columns/[column=z]/dip/adjusted_p_valueabsentNonepassed
/columns/[column=z]/kde/availableexactTrueTruepassed
/columns/[column=z]/kde/standard_deviationtolerance0.996071354748200410.99607135474820044.34E-1918passed
/columns/[column=z]/kde/interquartile_rangetolerance1.30140571026581451.30140571026581453.04E-1716passed
/columns/[column=z]/kde/reference_bandwidthtolerance0.434123476367395570.434123476367395573.44E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 99  screen_notes <- c(paste0(manual, ":3-27: finite included population; dip on continuous measurements and new unadjusted screens."),
100                    paste0(manual, ":37-44: h0 uses sample SD, type-7 IQR and multiplier 0.9; bounded support prevents automatic KDE interpretation."))
101  dip_citation <- "https://cran.r-project.org/web/packages/diptest/diptest.pdf (dip, min.is.0=FALSE); Hartigan and Hartigan (1985), Annals of Statistics 13:70-84. Bandwidth formula: manual section 110."
102  emit("STAT-EXP-MODESCREEN-001", "Current screen: independent dip and bandwidth on two continuous columns", "nominal",
103       screen_entry, list(columns = list("x", "z"), measurement = "continuous", replicates = 999L, seed = 17L),
104       c(screen_base(c("x", "z")), screen_column("x"), screen_column("z")),
105       screen_deferred(c("x", "z")), screen_notes, dip_citation)

Whole script generators/r/explore_mode_screen_fit.R


Bounded support: dip remains available and automatic KDE signal is refused

Case STAT-EXP-MODESCREEN-002 · status: passed

Reference

Evidence type
R
Source
R diptest::dip and base/stats closed-form likelihood calculations
Citation
https://cran.r-project.org/web/packages/diptest/diptest.pdf (dip, min.is.0=FALSE); Hartigan and Hartigan (1985), Annals of Statistics 13:70-84. Bandwidth formula: manual section 110.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, diptest 0.77.2, jsonlite 2.0.0
Reference script
generators/r/explore_mode_screen_fit.R · sha256 6b0f4af3271659ab
Dataset
reference_data/datasets/EXP_MODE_CURRENT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact34340.00E+2passed
/excluded_countexact110.00E+2passed
/family_sizeexact110.00E+2passed
/family_alphatolerance0.0500000000000000030.052.24E-1917passed
/columns/[column=x]/observationsexact32320.00E+2passed
/columns/[column=x]/missingexact110.00E+2passed
/columns/[column=x]/dip/availableexactTrueTruepassed
/columns/[column=x]/dip/statistictolerance0.200000000000000010.21.10E-1817passed
/columns/[column=x]/dip/adjusted_p_valueabsentNonepassed
/columns/[column=x]/kde/availableexactFalseFalsepassed
/columns/[column=x]/kde/standard_deviationtolerance4.09459474102016154.09459474102016152.13E-1717passed
/columns/[column=x]/kde/interquartile_rangetolerance7.94666666666666727.9466666666666671.14E-1717passed
/columns/[column=x]/kde/reference_bandwidthtolerance1.84256763345907281.84256763345907281.78E-1717passed
/columns/0/available_methodsexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

107  emit("STAT-EXP-MODESCREEN-002", "Bounded support: dip remains available and automatic KDE signal is refused", "axis_coverage",
108       screen_entry, list(columns = list("x"), measurement = "continuous", support = "bounded", replicates = 999L, seed = 17L),
109       c(screen_base("x"), screen_column("x", restricted = TRUE),
110         list("/columns/0/available_methods" = exact(1L, "Only dip is available under the declared restricted support"))),
111       screen_deferred("x"), screen_notes, dip_citation)

Whole script generators/r/explore_mode_screen_fit.R


Declared discrete counts refuse both continuous screen methods

Case STAT-EXP-MODESCREEN-003 · status: passed

Reference

Evidence type
R
Source
R diptest::dip and base/stats closed-form likelihood calculations
Citation
Analytic eligibility contract in docs/manual/chapters/110-multimodality.tex; counts constructed in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, diptest 0.77.2, jsonlite 2.0.0
Reference script
generators/r/explore_mode_screen_fit.R · sha256 6b0f4af3271659ab
Dataset
reference_data/datasets/EXP_MODE_CURRENT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact34340.00E+2passed
/excluded_countexact110.00E+2passed
/family_sizeexact110.00E+2passed
/family_alphatolerance0.0500000000000000030.052.24E-1917passed
/columns/0/observationsexact32320.00E+2passed
/columns/0/missingexact110.00E+2passed
/columns/0/dip/availableexactFalseFalsepassed
/columns/0/dip/statisticabsentNonepassed
/columns/0/dip/p_valueabsentNonepassed
/columns/0/kde/availableexactFalseFalsepassed
/columns/0/available_methodsexact000.00E+2passed
/columns/0/positive_methodsexact000.00E+2passed
/columns/0/explain_positiveexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

113  emit("STAT-EXP-MODESCREEN-003", "Declared discrete counts refuse both continuous screen methods", "boundary",
114       screen_entry, list(columns = list("count"), measurement = "discrete"),
115       c(screen_base("count"), list(
116         "/columns/0/observations" = exact(length(finite("count")), "Included finite count observations"),
117         "/columns/0/missing" = exact(1L, "One included missing count"),
118         "/columns/0/dip/available" = exact(FALSE, "Known discrete data are ineligible"),
119         "/columns/0/dip/statistic" = exact(NULL, "No continuous dip statistic is reported"),
120         "/columns/0/dip/p_value" = exact(NULL, "No continuous p-value is reported"),
121         "/columns/0/kde/available" = exact(FALSE, "Known discrete data are ineligible"),
122         "/columns/0/available_methods" = exact(0L, "Neither continuous method is eligible"),
123         "/columns/0/positive_methods" = exact(0L, "Unavailable methods cannot be positive"),
124         "/columns/0/explain_positive" = exact(FALSE, "No positive screen gate")
125       )), list(), c(screen_notes, paste0(manual, ":10-17,92-94: known discrete data are ineligible for the continuous screen.")),
126       "Analytic eligibility contract in docs/manual/chapters/110-multimodality.tex; counts constructed in R.")

Whole script generators/r/explore_mode_screen_fit.R


Bimodality coefficient of two separated modes

Case STAT-EXP-MULTIMODAL-001 · status: passed

Reference

Evidence type
Analytical
Source
R 4.6.1 base and stats
Citation
R stats::lm, stats::quantile (type 7), stats::t.test, stats::ecdf documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/explore_plots.R · sha256 15466ea60c37fc33
Dataset
reference_data/datasets/EXP_PLOTS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/columns/0/finite_countexact24240.00E+2passed
/columns/0/bc_valuetolerance0.599046444914183240.59904644491418293.33E-1615passed
/columns/0/passedcontains['BC']['BC']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

196  w <- d$m
197  n <- length(w)
198  m2 <- mean((w - mean(w))^2); m3 <- mean((w - mean(w))^3); m4 <- mean((w - mean(w))^4)
199  skew <- m3 / m2^1.5 * sqrt(n * (n - 1)) / (n - 2)
200  kurt <- ((n + 1) * (m4 / m2^2 - 3) + 6) * (n - 1) / ((n - 2) * (n - 3))
201  bc <- (skew^2 + 1) / (kurt + 3 * (n - 1)^2 / ((n - 2) * (n - 3)))
202  stopifnot(bc > 0.555)
203  emit("STAT-EXP-MULTIMODAL-001", "tabelz.analysis.multimodality:compute_multimodality", "multimodalityMenuItem",
204       "Bimodality coefficient of two separated modes", list(columns = list("m"), active_methods = list("BC")),
205       list("/columns/0/finite_count" = exact(24L, "All values are finite."),
206            "/columns/0/bc_value" = num(bc, "Sarle's coefficient with adjusted skewness G1 and excess kurtosis G2."),
207            "/columns/0/passed" = contains("BC", "BC exceeds the 5/9 uniform benchmark.")),
208       c("Sarle's bimodality coefficient (SAS definition); the manual does not define it.",
209         "The column is two normal-score blocks of 12, 6 units apart."), oracle_class = "analytic")

Whole script generators/r/explore_plots.R


test six mp table matches the published greenwood log log values

Case STAT-HARV-GEHAN-KM-001 · status: imprecise_reference

Reference

Evidence type
Published example
Source
Freireich et al. (1963); Gehan (1965), 6-MP/placebo remission data.
Citation
Freireich et al. (1963); Gehan (1965), 6-MP/placebo remission data.
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/HARVEST_GEHAN_MP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/curves/0/nexact21210.00E+2passed
/curves/0/eventsexact990.00E+2passed
/curves/0/censoredexact12120.00E+2passed
/curves/0/median/timetolerance23.023.00.00E+115passed
/curves/0/at_risk/0exact21210.00E+2passed
/curves/0/event_counts/0exact330.00E+2passed
/curves/0/survival/0tolerance0.85710.85714285714285724.29E-54passed
/curves/0/standard_error/0tolerance0.07640.07636035483212125imprecise_reference
/curves/0/lower/0tolerance0.61970.61971795526598381.80E-54passed
/curves/0/upper/0tolerance0.95160.9515517476469974.83E-54passed
/curves/0/at_risk/1exact17170.00E+2passed
/curves/0/event_counts/1exact110.00E+2passed
/curves/0/survival/1tolerance0.80670.80672268907563032.27E-54passed
/curves/0/standard_error/1tolerance0.08690.0869352851800572imprecise_reference
/curves/0/lower/1tolerance0.56310.56314656462223344.66E-54passed
/curves/0/upper/1tolerance0.92280.92280901920288059.02E-65passed
/curves/0/at_risk/3exact15150.00E+2passed
/curves/0/event_counts/3exact110.00E+2passed
/curves/0/survival/3tolerance0.75290.75294117647058824.12E-54passed
/curves/0/standard_error/3tolerance0.09630.09634965299432051imprecise_reference
/curves/0/lower/3tolerance0.50320.50319951079995654.89E-76passed
/curves/0/upper/3tolerance0.88940.88936183516926553.82E-54passed
/curves/0/at_risk/5exact12120.00E+2passed
/curves/0/event_counts/5exact110.00E+2passed
/curves/0/survival/5tolerance0.69020.69019607843137253.92E-65passed
/curves/0/standard_error/5tolerance0.10680.10681470777500983imprecise_reference
/curves/0/lower/5tolerance0.43160.431610222486184141.02E-54passed
/curves/0/upper/5tolerance0.84910.84906596334945073.40E-54passed
/curves/0/at_risk/6exact11110.00E+2passed
/curves/0/event_counts/6exact110.00E+2passed
/curves/0/survival/6tolerance0.62750.62745098039215694.90E-54passed
/curves/0/standard_error/6tolerance0.11410.11405386525675254imprecise_reference
/curves/0/lower/6tolerance0.36750.36751085598750211.09E-54passed
/curves/0/upper/6tolerance0.80490.80491218947184131.22E-54passed
/curves/0/at_risk/10exact770.00E+2passed
/curves/0/event_counts/10exact110.00E+2passed
/curves/0/survival/10tolerance0.53780.53781512605042031.51E-54passed
/curves/0/standard_error/10tolerance0.12820.12823375169303403imprecise_reference
/curves/0/lower/10tolerance0.26780.26777893676713512.11E-54passed
/curves/0/upper/10tolerance0.74680.74679071763539469.28E-64passed
/curves/0/at_risk/11exact660.00E+2passed
/curves/0/event_counts/11exact110.00E+2passed
/curves/0/survival/11tolerance0.44820.448179271708683552.07E-54passed
/curves/0/standard_error/11tolerance0.13460.134591456755760488.54E-64passed
/curves/0/lower/11tolerance0.18810.18805200595923244imprecise_reference
/curves/0/upper/11tolerance0.68010.68014262849524154.26E-54passed

Convention


test gehan log rank matches the published chi square

Case STAT-HARV-GEHAN-LOGRANK-001 · status: imprecise_reference

Reference

Evidence type
Published example
Source
Freireich et al. (1963); Gehan (1965), 6-MP/placebo remission data.
Citation
Freireich et al. (1963); Gehan (1965), 6-MP/placebo remission data.
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/HARVEST_GEHAN_BOTH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/log_rank/observed/0exact990.00E+2passed
/log_rank/observed/1exact21210.00E+2passed
/log_rank/expected/0tolerance19.2519.250500948031135.01E-44passed
/log_rank/expected/1tolerance10.7510.7494990519688685.01E-44passed
/log_rank/statistictolerance16.7916.79294098921654imprecise_reference
/log_rank/degrees_of_freedomexact110.00E+2passed
/log_rank/p_valuetolerance4.17e-54.1688091093345274e-05imprecise_reference
/curves/1/median/timetolerance8.08.00.00E+115passed

Convention


test grubbs matches nist fixture and preserves candidate

Case STAT-HARV-NIST-GRUBBS-001 · status: passed

Reference

Evidence type
Analytical
Source
NIST e-Handbook 1.3.5.17.1, Grubbs' Test for Outliers, https://www.itl.nist.gov/div898/handbook/eda/section3/eda35h1.htm
Citation
NIST e-Handbook 1.3.5.17.1, Grubbs' Test for Outliers, https://www.itl.nist.gov/div898/handbook/eda/section3/eda35h1.htm
Reference software
not recorded
Reference script
generators/build_harvested_cases.py · sha256 24024e79e196b541
Dataset
reference_data/datasets/HARVEST_NIST_GRUBBS_PRIMARY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance2.46876461121245051092e+02.468764611212453.45E-1615passed
/critical_valuetolerance2.12664508719546529412e+02.12664508719546671.45E-1515passed
/p_valuetolerance3.00263868207083747264e-73.002638682071362e-075.25E-2012passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

184  def _grubbs() -> None:
185      dataset_id = "HARVEST_NIST_GRUBBS_PRIMARY"
186      source = (
187          "NIST e-Handbook 1.3.5.17.1, Grubbs' Test for Outliers, "
188          "https://www.itl.nist.gov/div898/handbook/eda/section3/eda35h1.htm"
189      )
190      values = (199.31, 199.53, 200.19, 200.82, 201.92, 201.95, 202.18, 245.57)
191      _dataset(dataset_id, [("x", "Float64")], [(x,) for x in values], source, source)
192      g, critical, p_value = _grubbs_exact(values, "0.05")
193      _case(
194          "STAT-HARV-NIST-GRUBBS-001",
195          dataset_id,
196          "tabelz.executor.distribution_diagnostics",
197          "distributionDiagnosticsMenuItem",
198          {"column": "x", "method": "grubbs"},
199          {
200              "/statistic": _relative(g, "G = max|x - mean| / s in 60-digit decimal."),
201              "/critical_value": _relative(
202                  critical, "(n-1)/sqrt(n) sqrt(t^2/(n-2+t^2)), t the alpha/(2n) upper quantile."
203              ),
204              "/p_value": _relative(p_value, "min(1, 2n P(T > t_G)) with n - 2 = 6 df."),
205          },
206          source,
207          "test_grubbs_matches_nist_fixture_and_preserves_candidate",
208          ["tests/unit/analysis/test_distribution_diagnostics.py"],
209          oracle_class="analytic",
210          oracle_notes=[
211              "Eight uranium-isotope measurements verified against NIST e-Handbook 1.3.5.17.1.",
212              "This case uses NIST's two-sided critical-value formula; its worked "
213              "example instead prints the upper one-sided critical value 2.032.",
214              "Expected values are computed here in 60-digit decimal from the exact even-df "

Whole script generators/build_harvested_cases.py


ANOM normal two groups

Case STAT-HYP-ANALYSIS-OF-MEANS-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::qt
Citation
Nelson LS (1982), Journal of Quality Technology 14, 172-173 (two-group ANOM reduces to the t quantile); R Core Team, stats package help: qt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_analysis_of_means.R · sha256 d1cae03fb86acee1
Dataset
reference_data/datasets/HYP_ANOM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/panels/0/centertolerance11.3437511.343750.00E-3passed
/panels/0/critical_valuetolerance2.14478668791780392.1447866879177949.76E-1514passed
/panels/0/critical_dfexact14140.00E+2passed
/panels/0/groups/0/labelexactAApassed
/panels/0/groups/0/sizetolerance77.00.00E+2passed
/panels/0/groups/0/rowsexact770.00E+2passed
/panels/0/groups/0/statistictolerance10.21428571428571410.2142857142857144.76E-1616passed
/panels/0/groups/0/standard_errortolerance0.25017486595749650.25017486595749661.06E-1615passed
/panels/0/groups/0/lowertolerance10.80717827784274110.8071782778427431.99E-1515passed
/panels/0/groups/0/uppertolerance11.88032172215725911.8803217221572571.99E-1515passed
/panels/0/groups/0/decisionexactbelowbelowpassed
/panels/0/groups/1/labelexactBBpassed
/panels/0/groups/1/sizetolerance99.00.00E+2passed
/panels/0/groups/1/rowsexact990.00E+2passed
/panels/0/groups/1/statistictolerance12.22222222222222112.2222222222222214.33E-1616passed
/panels/0/groups/1/standard_errortolerance0.194580451300275040.194580451300275151.13E-1615passed
/panels/0/groups/1/lowertolerance10.92641643832213210.9264164383221341.83E-1515passed
/panels/0/groups/1/uppertolerance11.76108356167786811.7610835616778661.83E-1515passed
/panels/0/groups/1/decisionexactaboveabovepassed
/panels/0/signalsexact220.00E+2passed
/pooled_sdtolerance0.882533972803272730.88253397280327324.40E-1615passed
/residual_dfexact14140.00E+2passed
/observationsexact16160.00E+2passed
/missing_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

46  emit <- function(id, title, kind, config, outputs, notes, source_ref, citation, oracle_class = "independent_implementation") {
47    write_expected("hypothesis", id, list(
48      schema_version = "2.0.0", case_id = id, outputs = outputs, expected_error = NULL,
49      deferred_outputs = list(), generator_run = run))
50    write_case("hypothesis", id, list(
51      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "hypothesis",
52      entry_point = entry, menu_items = list(menu), title = title, case_kind = kind,
53      dataset = list(path = "reference_data/datasets/HYP_ANOM.json", sha256 = ""),
54      excluded_rows = list(), configuration = config, "repeat" = 1L,
55      expected = list(path = paste0("reference_data/expected/hypothesis/", id, ".json"), sha256 = ""),
56      oracle = list(oracle_class = oracle_class, source = source_ref, citation = citation,
57                    generator = list(path = "generators/r/hypothesis_analysis_of_means.R", sha256 = ""),
58                    independent = TRUE,
59                    convention_notes = c(notes, notes_common, "Expected computed from the exported CSV.")),
60      supported = TRUE, unsupported_reason = NULL,
61      requirement_ids = list("REQ-STAT-HYP-ANALYSIS-OF-MEANS-001"),
62      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/hypothesis_analysis_of_means.R


ANOM normal three groups

Case STAT-HYP-ANALYSIS-OF-MEANS-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::integrate, uniroot
Citation
Nelson LS (1982), Journal of Quality Technology 14, 172-173 (the max |T_i| distribution of ANOM); R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_analysis_of_means.R · sha256 d1cae03fb86acee1
Dataset
reference_data/datasets/HYP_ANOM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/panels/0/centertolerance11.3437511.343750.00E-3passed
/panels/0/critical_valuetolerance2.64010545244306052.64010545244309963.91E-1413passed
/panels/0/critical_dfexact13130.00E+2passed
/panels/0/groups/0/labelexactAApassed
/panels/0/groups/0/sizetolerance55.00.00E+2passed
/panels/0/groups/0/rowsexact550.00E+2passed
/panels/0/groups/0/statistictolerance10.0810.087.11E-1717passed
/panels/0/groups/0/standard_errortolerance0.389578996959275230.38957899695927541.62E-1615passed
/panels/0/groups/0/lowertolerance10.3152203659705210.3152203659705041.63E-1414passed
/panels/0/groups/0/uppertolerance12.3722796340294812.3722796340294961.63E-1414passed
/panels/0/groups/0/decisionexactbelowbelowpassed
/panels/0/groups/1/labelexactBBpassed
/panels/0/groups/1/sizetolerance55.00.00E+2passed
/panels/0/groups/1/rowsexact550.00E+2passed
/panels/0/groups/1/statistictolerance11.62000000000000111.6200000000000015.24E-1818passed
/panels/0/groups/1/standard_errortolerance0.389578996959275230.38957899695927541.62E-1615passed
/panels/0/groups/1/lowertolerance10.3152203659705210.3152203659705041.63E-1414passed
/panels/0/groups/1/uppertolerance12.3722796340294812.3722796340294961.63E-1414passed
/panels/0/groups/1/decisionexactwithinwithinpassed
/panels/0/groups/2/labelexactCCpassed
/panels/0/groups/2/sizetolerance66.00.00E+2passed
/panels/0/groups/2/rowsexact660.00E+2passed
/panels/0/groups/2/statistictolerance12.16666666666666612.1666666666666667.45E-1717passed
/panels/0/groups/2/standard_errortolerance0.339084993001212810.3390849930012131.65E-1615passed
/panels/0/groups/2/lowertolerance10.44852986113588110.4485298611358671.38E-1414passed
/panels/0/groups/2/uppertolerance12.23897013886411912.2389701388641331.38E-1414passed
/panels/0/groups/2/decisionexactwithinwithinpassed
/panels/0/signalsexact110.00E+2passed
/pooled_sdtolerance1.05061642467404461.0506164246740454.35E-1615passed
/residual_dfexact13130.00E+2passed
/observationsexact16160.00E+2passed
/missing_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

159  normal_one_way("STAT-HYP-ANALYSIS-OF-MEANS-002", "ANOM normal three groups", "axis_coverage", "g3", 0.05)

Whole script generators/r/hypothesis_analysis_of_means.R


ANOM normal two groups alpha 0.01

Case STAT-HYP-ANALYSIS-OF-MEANS-003 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::qt
Citation
Nelson LS (1982), Journal of Quality Technology 14, 172-173 (two-group ANOM reduces to the t quantile); R Core Team, stats package help: qt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_analysis_of_means.R · sha256 d1cae03fb86acee1
Dataset
reference_data/datasets/HYP_ANOM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/panels/0/centertolerance11.3437511.343750.00E-3passed
/panels/0/critical_valuetolerance2.97684273437083392.9768427343707815.28E-1413passed
/panels/0/critical_dfexact14140.00E+2passed
/panels/0/groups/0/labelexactAApassed
/panels/0/groups/0/sizetolerance77.00.00E+2passed
/panels/0/groups/0/rowsexact770.00E+2passed
/panels/0/groups/0/statistictolerance10.21428571428571410.2142857142857144.76E-1616passed
/panels/0/groups/0/standard_errortolerance0.25017486595749650.25017486595749661.06E-1615passed
/panels/0/groups/0/lowertolerance10.59901876795222910.5990187679522411.24E-1414passed
/panels/0/groups/0/uppertolerance12.08848123204777112.0884812320477591.24E-1414passed
/panels/0/groups/0/decisionexactbelowbelowpassed
/panels/0/groups/1/labelexactBBpassed
/panels/0/groups/1/sizetolerance99.00.00E+2passed
/panels/0/groups/1/rowsexact990.00E+2passed
/panels/0/groups/1/statistictolerance12.22222222222222112.2222222222222214.33E-1616passed
/panels/0/groups/1/standard_errortolerance0.194580451300275040.194580451300275151.13E-1615passed
/panels/0/groups/1/lowertolerance10.76451459729617910.7645145972961888.76E-1515passed
/panels/0/groups/1/uppertolerance11.92298540270382111.9229854027038128.76E-1515passed
/panels/0/groups/1/decisionexactaboveabovepassed
/panels/0/signalsexact220.00E+2passed
/pooled_sdtolerance0.882533972803272730.88253397280327324.40E-1615passed
/residual_dfexact14140.00E+2passed
/observationsexact16160.00E+2passed
/missing_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

160  normal_one_way("STAT-HYP-ANALYSIS-OF-MEANS-003", "ANOM normal two groups alpha 0.01", "axis_coverage", "g2", 0.01)

Whole script generators/r/hypothesis_analysis_of_means.R


ANOM normal two-way 2 x 2

Case STAT-HYP-ANALYSIS-OF-MEANS-004 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::qt
Citation
Nelson PR, Wludyka PS, Copeland KAF (2005), The Analysis of Means, SIAM, ch. 6 (two-way ANOM; a 2 x 2 interaction reduces to a t quantile); R Core Team, stats package help: qt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_analysis_of_means.R · sha256 d1cae03fb86acee1
Dataset
reference_data/datasets/HYP_ANOM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/panels/0/centertolerance11.3437511.3437499999999981.78E-1515passed
/panels/0/critical_valuetolerance2.17881282966722842.1788128296672181.06E-1414passed
/panels/0/critical_dfexact12120.00E+2passed
/panels/0/groups/0/labelexacthighhighpassed
/panels/0/groups/0/sizetolerance88.00.00E+2passed
/panels/0/groups/0/rowsexact880.00E+2passed
/panels/0/groups/0/statistictolerance12.212512.2124999999999991.42E-1515passed
/panels/0/groups/0/standard_errortolerance0.262177381111847070.26217738111184732.26E-1615passed
/panels/0/groups/0/lowertolerance10.77251455838495410.7725145583849544.69E-1616passed
/panels/0/groups/0/uppertolerance11.91498544161504611.9149854416150433.08E-1515passed
/panels/0/groups/0/decisionexactaboveabovepassed
/panels/0/groups/1/labelexactlowlowpassed
/panels/0/groups/1/sizetolerance88.00.00E+2passed
/panels/0/groups/1/rowsexact880.00E+2passed
/panels/0/groups/1/statistictolerance10.47500000000000110.4750000000000014.21E-1616passed
/panels/0/groups/1/standard_errortolerance0.262177381111847070.26217738111184732.26E-1615passed
/panels/0/groups/1/lowertolerance10.77251455838495410.7725145583849544.69E-1616passed
/panels/0/groups/1/uppertolerance11.91498544161504611.9149854416150433.08E-1515passed
/panels/0/groups/1/decisionexactbelowbelowpassed
/panels/0/signalsexact220.00E+2passed
/panels/1/centertolerance11.3437511.3437499999999981.78E-1515passed
/panels/1/critical_valuetolerance2.17881282966722842.1788128296672181.06E-1414passed
/panels/1/critical_dfexact12120.00E+2passed
/panels/1/groups/0/labelexactcoldcoldpassed
/panels/1/groups/0/sizetolerance88.00.00E+2passed
/panels/1/groups/0/rowsexact880.00E+2passed
/panels/1/groups/0/statistictolerance11.087511.08753.55E-1616passed
/panels/1/groups/0/standard_errortolerance0.262177381111847070.26217738111184732.26E-1615passed
/panels/1/groups/0/lowertolerance10.77251455838495410.7725145583849544.69E-1616passed
/panels/1/groups/0/uppertolerance11.91498544161504611.9149854416150433.08E-1515passed
/panels/1/groups/0/decisionexactwithinwithinpassed
/panels/1/groups/1/labelexactwarmwarmpassed
/panels/1/groups/1/sizetolerance88.00.00E+2passed
/panels/1/groups/1/rowsexact880.00E+2passed
/panels/1/groups/1/statistictolerance11.60000000000000111.61.36E-1515passed
/panels/1/groups/1/standard_errortolerance0.262177381111847070.26217738111184732.26E-1615passed
/panels/1/groups/1/lowertolerance10.77251455838495410.7725145583849544.69E-1616passed
/panels/1/groups/1/uppertolerance11.91498544161504611.9149854416150433.08E-1515passed
/panels/1/groups/1/decisionexactwithinwithinpassed
/panels/1/signalsexact000.00E+2passed
/panels/2/centertolerance00.00.00E+2passed
/panels/2/critical_valuetolerance2.17881282966722842.17881282966722939.34E-1615passed
/panels/2/critical_dfexact12120.00E+2passed
/panels/2/groups/0/labelexacthigh / coldhigh / coldpassed
/panels/2/groups/0/sizetolerance44.00.00E+2passed
/panels/2/groups/0/rowsexact440.00E+2passed
/panels/2/groups/0/statistictolerance0.168749999999999290.16874999999999935.43E-1917passed
/panels/2/groups/0/standard_errortolerance0.262177381111847070.26217738111184732.26E-1615passed
/panels/2/groups/0/lowertolerance-0.57123544161504691-0.57123544161504766.69E-1614passed
/panels/2/groups/0/uppertolerance0.571235441615046910.57123544161504766.69E-1614passed
/panels/2/groups/0/decisionexactwithinwithinpassed
/panels/2/groups/1/labelexacthigh / warmhigh / warmpassed
/panels/2/groups/1/sizetolerance44.00.00E+2passed
/panels/2/groups/1/rowsexact440.00E+2passed
/panels/2/groups/1/statistictolerance-0.16875000000000107-0.168750000000001074.19E-1816passed
/panels/2/groups/1/standard_errortolerance0.262177381111847070.26217738111184732.26E-1615passed
/panels/2/groups/1/lowertolerance-0.57123544161504691-0.57123544161504766.69E-1614passed
/panels/2/groups/1/uppertolerance0.571235441615046910.57123544161504766.69E-1614passed
/panels/2/groups/1/decisionexactwithinwithinpassed
/panels/2/groups/2/labelexactlow / coldlow / coldpassed
/panels/2/groups/2/sizetolerance44.00.00E+2passed
/panels/2/groups/2/rowsexact440.00E+2passed
/panels/2/groups/2/statistictolerance-0.16875000000000107-0.168750000000002841.77E-1513passed
/panels/2/groups/2/standard_errortolerance0.262177381111847070.26217738111184732.26E-1615passed
/panels/2/groups/2/lowertolerance-0.57123544161504691-0.57123544161504766.69E-1614passed
/panels/2/groups/2/uppertolerance0.571235441615046910.57123544161504766.69E-1614passed
/panels/2/groups/2/decisionexactwithinwithinpassed
/panels/2/groups/3/labelexactlow / warmlow / warmpassed
/panels/2/groups/3/sizetolerance44.00.00E+2passed
/panels/2/groups/3/rowsexact440.00E+2passed
/panels/2/groups/3/statistictolerance0.168749999999997510.16874999999999753.10E-1816passed
/panels/2/groups/3/standard_errortolerance0.262177381111847070.26217738111184732.26E-1615passed
/panels/2/groups/3/lowertolerance-0.57123544161504691-0.57123544161504766.69E-1614passed
/panels/2/groups/3/uppertolerance0.571235441615046910.57123544161504766.69E-1614passed
/panels/2/groups/3/decisionexactwithinwithinpassed
/panels/2/signalsexact000.00E+2passed
/pooled_sdtolerance1.04870952444738831.04870952444738928.85E-1615passed
/residual_dfexact12120.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

172  df <- a * b * (r - 1)
173  root <- sqrt(cell_ss / df)
174  grand <- mean(cell_mean)
175  row_means <- rowMeans(cell_mean); col_means <- colMeans(cell_mean)
176  h <- stats::qt(0.975, df)
177  outputs <- panel_outputs(0, grand, h, df,
178    data.frame(label = levels_a, size = b * r, statistic = row_means, se = root * sqrt((a - 1) / (a * b * r)), rows = b * r), "1e-9", "0")
179  outputs <- c(outputs, panel_outputs(1, grand, h, df,
180    data.frame(label = levels_b, size = a * r, statistic = col_means, se = root * sqrt((b - 1) / (a * b * r)), rows = a * r), "1e-9", "0"))
181  contrast <- cell_mean - row_means - rep(col_means, each = a) + grand
182  inter <- data.frame(label = as.vector(t(outer(levels_a, levels_b, paste, sep = " / "))), size = r,
183                      statistic = as.vector(t(contrast)), se = root * sqrt((a - 1) * (b - 1) / (a * b * r)), rows = r)
184  outputs <- c(outputs, panel_outputs(2, 0, h, df, inter, "1e-9", "0"))
185  outputs[["/pooled_sd"]] <- tol(root, "root residual mean square")
186  outputs[["/residual_df"]] <- exact(df, "ab(r - 1)")
187  emit("STAT-HYP-ANALYSIS-OF-MEANS-004", "ANOM normal two-way 2 x 2", "axis_coverage",
188       list(analysis = "normal_two_way", response = "y", factor_a = "fa", factor_b = "fb", alpha = "0.05"),
189       outputs, c("Balanced 2 x 2 x 4: main-effect panels use the residual mean square on ab(r - 1) = 12 df with SE = s sqrt((a - 1)/(abr)); the interaction panel shows d_ij = m_ij - m_i. - m_.j + m with SE = s sqrt((a-1)(b-1)/(abr)) centred at zero, cells in row-major order 'fa / fb'. Every panel has two levels or a 2 x 2 contrast set, so h = qt(0.975, 12) exactly.",
190                  "Panel order: factor A (fa), factor B (fb), interaction; levels sorted ('high' before 'low', 'cold' before 'warm')."),
191       "R 4.6.1 stats::qt", "Nelson PR, Wludyka PS, Copeland KAF (2005), The Analysis of Means, SIAM, ch. 6 (two-way ANOM; a 2 x 2 interaction reduces to a t quantile); R Core Team, stats package help: qt.")

Whole script generators/r/hypothesis_analysis_of_means.R


ANOM binomial three groups

Case STAT-HYP-ANALYSIS-OF-MEANS-005 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::integrate, uniroot
Citation
Nelson PR, Wludyka PS, Copeland KAF (2005), The Analysis of Means, SIAM, ch. 2; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_analysis_of_means.R · sha256 d1cae03fb86acee1
Dataset
reference_data/datasets/HYP_ANOM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/panels/0/centertolerance0.315624999999999990.3156251.10E-1817passed
/panels/0/critical_valuetolerance2.34342734448502952.34342734448494967.99E-1413passed
/panels/0/critical_dfabsentNonepassed
/panels/0/groups/0/labelexactAApassed
/panels/0/groups/0/sizetolerance200200.00.00E+2passed
/panels/0/groups/0/rowsexact550.00E+2passed
/panels/0/groups/0/statistictolerance0.270000000000000020.272.24E-1817passed
/panels/0/groups/0/standard_errortolerance0.0272492227705959070.0272492227705959074.44E-1916passed
/panels/0/groups/0/lowertolerance0.251768426243421440.25176842624342362.16E-1514passed
/panels/0/groups/0/uppertolerance0.379481573756578540.37948157375657642.17E-1514passed
/panels/0/groups/0/decisionexactwithinwithinpassed
/panels/0/groups/1/labelexactBBpassed
/panels/0/groups/1/sizetolerance200200.00.00E+2passed
/panels/0/groups/1/rowsexact550.00E+2passed
/panels/0/groups/1/statistictolerance0.330000000000000020.334.46E-1816passed
/panels/0/groups/1/standard_errortolerance0.0272492227705959070.0272492227705959074.44E-1916passed
/panels/0/groups/1/lowertolerance0.251768426243421440.25176842624342362.16E-1514passed
/panels/0/groups/1/uppertolerance0.379481573756578540.37948157375657642.17E-1514passed
/panels/0/groups/1/decisionexactwithinwithinpassed
/panels/0/groups/2/labelexactCCpassed
/panels/0/groups/2/sizetolerance240240.00.00E+2passed
/panels/0/groups/2/rowsexact660.00E+2passed
/panels/0/groups/2/statistictolerance0.341666666666666670.34166666666666674.07E-1816passed
/panels/0/groups/2/standard_errortolerance0.0237174041325998070.0237174041325998079.28E-2017passed
/panels/0/groups/2/lowertolerance0.260044986615463380.260044986615465271.89E-1514passed
/panels/0/groups/2/uppertolerance0.37120501338453660.37120501338453471.89E-1514passed
/panels/0/groups/2/decisionexactwithinwithinpassed
/panels/0/signalsexact000.00E+2passed
/panels/0/groups/0/eventstolerance5454.00.00E+2passed
/panels/0/groups/1/eventstolerance6666.00.00E+2passed
/panels/0/groups/2/eventstolerance8282.00.00E+2passed
/observationsexact16160.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

194  labels <- sort(unique(d$g3))
195  x <- as.vector(tapply(d$events, factor(d$g3, levels = labels), sum))
196  n <- as.vector(tapply(d$trials, factor(d$g3, levels = labels), sum))
197  rows <- as.vector(table(factor(d$g3, levels = labels)))
198  N <- sum(n); p_bar <- sum(x) / N
199  h <- anom_h(n, 0.05, NULL)
200  groups <- data.frame(label = labels, size = n, statistic = x / n,
201                       se = sqrt(p_bar * (1 - p_bar)) / sqrt(n) * sqrt((N - n) / N), rows = rows)
202  outputs <- panel_outputs(0, p_bar, h, NULL, groups, "1e-6", "0")
203  for (g in seq_along(labels)) outputs[[paste0("/panels/0/groups/", g - 1, "/events")]] <- tol(x[g], "pooled successes of the group")
204  outputs[["/observations"]] <- exact(nrow(d), "analysed rows")
205  emit("STAT-HYP-ANALYSIS-OF-MEANS-005", "ANOM binomial three groups", "axis_coverage",
206       list(analysis = "binomial", group = "g3", events = "events", trials = "trials", alpha = "0.05"),
207       outputs, "Binomial: p_bar = sum x / sum n; SE_i = sqrt(p_bar (1 - p_bar) / n_i) sqrt((N - n_i)/N); h solves P(max |Z_i| <= h) = 0.95 over the deviation correlation with weights n_i (infinite df) by the same deterministic quadrature as the three-group normal case; 1e-6 relative tolerance on h and the limits.",
208       "R 4.6.1 stats::integrate, uniroot", "Nelson PR, Wludyka PS, Copeland KAF (2005), The Analysis of Means, SIAM, ch. 2; R Core Team, stats package help: integrate, uniroot.")

Whole script generators/r/hypothesis_analysis_of_means.R


ANOM Poisson rates two groups with exposure

Case STAT-HYP-ANALYSIS-OF-MEANS-006 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::qnorm
Citation
Nelson PR, Wludyka PS, Copeland KAF (2005), The Analysis of Means, SIAM, ch. 2; R Core Team, stats package help: qnorm.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_analysis_of_means.R · sha256 d1cae03fb86acee1
Dataset
reference_data/datasets/HYP_ANOM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/panels/0/centertolerance4.44444444444444464.4444444444444454.18E-1717passed
/panels/0/critical_valuetolerance1.95996398454005361.95996398454005278.83E-1615passed
/panels/0/critical_dfabsentNonepassed
/panels/0/groups/0/labelexactAApassed
/panels/0/groups/0/sizetolerance11.69999999999999911.72.89E-1616passed
/panels/0/groups/0/rowsexact770.00E+2passed
/panels/0/groups/0/statistictolerance4.35897435897435954.35897435897435952.46E-1717passed
/panels/0/groups/0/standard_errortolerance0.496387194033293340.496387194033293344.12E-1817passed
/panels/0/groups/0/lowertolerance3.4715434217522943.4715434217522959.11E-1615passed
/panels/0/groups/0/uppertolerance5.41734546713659535.4173454671365949.28E-1615passed
/panels/0/groups/0/decisionexactwithinwithinpassed
/panels/0/groups/1/labelexactBBpassed
/panels/0/groups/1/sizetolerance21.60000000000000121.64.21E-1616passed
/panels/0/groups/1/rowsexact990.00E+2passed
/panels/0/groups/1/statistictolerance4.49074074074074054.49074074074074052.24E-1717passed
/panels/0/groups/1/standard_errortolerance0.268876396768033820.26887639676803387.17E-1917passed
/panels/0/groups/1/lowertolerance3.91745639048619683.91745639048619684.00E-1716passed
/panels/0/groups/1/uppertolerance4.97143249840269254.97143249840269252.36E-1717passed
/panels/0/groups/1/decisionexactwithinwithinpassed
/panels/0/signalsexact000.00E+2passed
/panels/0/groups/0/eventstolerance5151.00.00E+2passed
/panels/0/groups/1/eventstolerance9797.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

211  labels <- sort(unique(d$g2))
212  c_ <- as.vector(tapply(d$counts, factor(d$g2, levels = labels), sum))
213  e <- as.vector(tapply(d$exposure, factor(d$g2, levels = labels), sum))
214  rows <- as.vector(table(factor(d$g2, levels = labels)))
215  E <- sum(e); rate <- sum(c_) / E
216  h <- stats::qnorm(0.975)
217  groups <- data.frame(label = labels, size = e, statistic = c_ / e, se = sqrt(rate * (1 / e - 1 / E)), rows = rows)
218  outputs <- panel_outputs(0, rate, h, NULL, groups, "1e-9", "0")
219  for (g in seq_along(labels)) outputs[[paste0("/panels/0/groups/", g - 1, "/events")]] <- tol(c_[g], "pooled counts of the group")
220  emit("STAT-HYP-ANALYSIS-OF-MEANS-006", "ANOM Poisson rates two groups with exposure", "axis_coverage",
221       list(analysis = "poisson", group = "g2", events = "counts", exposure = "exposure", exposure_unit = "hour", alpha = "0.05"),
222       outputs, "Poisson: lambda_hat = sum c / sum e; rate_i = c_i / e_i; SE_i = sqrt(lambda_hat (1/e_i - 1/E)); two groups give h = qnorm(0.975) exactly.",
223       "R 4.6.1 stats::qnorm", "Nelson PR, Wludyka PS, Copeland KAF (2005), The Analysis of Means, SIAM, ch. 2; R Core Team, stats package help: qnorm.")

Whole script generators/r/hypothesis_analysis_of_means.R


One-way ANOVA Tukey-Kramer

Case STAT-HYP-ANOVA-POST-HOC-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::aov, TukeyHSD, qtukey
Citation
R Core Team, stats package help: aov, TukeyHSD, ptukey, qtukey.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_anova_post_hoc.R · sha256 057b84f90d3c52ae
Dataset
reference_data/datasets/HYP_POSTHOC.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/between_sstolerance28.20666666666664228.2066666666665964.61E-1414passed
/anova/within_sstolerance28.53333333333331728.5333333333333462.86E-1414passed
/anova/between_dfexact220.00E+2passed
/anova/within_dfexact15150.00E+2passed
/anova/msetolerance1.90222222222222111.90222222222222312.05E-1514passed
/anova/f_valuetolerance7.41413551401868987.4141355140186692.05E-1414passed
/anova/p_valuetail_probability0.00576731720866283020.00576731720866288665.64E-1714passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/groups/0/nexact550.00E+2passed
/groups/0/meantolerance10.910.8999999999999991.42E-1515passed
/groups/0/variancetolerance0.874999999999999670.87500000000000036.63E-1615passed
/groups/1/nexact660.00E+2passed
/groups/1/meantolerance13.66666666666666613.6666666666666641.70E-1515passed
/groups/1/variancetolerance1.35066666666666731.35066666666666851.16E-1515passed
/groups/2/nexact770.00E+2passed
/groups/2/meantolerance11.111.1000000000000033.20E-1515passed
/groups/2/variancetolerance3.04666666666666643.0466666666666674.56E-1615passed
/comparisons/critical_valuetolerance3.67337762497445123.6733776588970963.39E-88passed
/comparisons/rows/0/differencetolerance-2.7666666666666657-2.76666666666666571.93E-1717passed
/comparisons/rows/0/standard_errortolerance0.83515356760387560.8351535676038764.46E-1615passed
/comparisons/rows/0/statistictolerance-3.3127639921415417-3.312763992141541.76E-1515passed
/comparisons/rows/0/dftolerance1515.00.00E+2passed
/comparisons/rows/0/p_adjustedtail_probability0.0123756432555284230.0123756432550542474.74E-1310passed
/comparisons/rows/0/lowertolerance-4.9359531947252275-4.9359532147580032.00E-88passed
/comparisons/rows/0/uppertolerance-0.59738013860810035-0.59738011857532892.00E-87passed
/comparisons/rows/0/significantexactTrueTruepassed
/comparisons/rows/1/differencetolerance-0.19999999999999929-0.200000000000004625.33E-1513passed
/comparisons/rows/1/standard_errortolerance0.807583107915511020.80758310791551154.47E-1615passed
/comparisons/rows/1/statistictolerance-0.24765253017268807-0.24765253017269456.44E-1513passed
/comparisons/rows/1/dftolerance1515.00.00E+2passed
/comparisons/rows/1/p_adjustedtail_probability0.966829550984284140.96682955098427151.27E-1413passed
/comparisons/rows/1/lowertolerance-2.2976730798325713-2.297673099204021.94E-88passed
/comparisons/rows/1/uppertolerance1.89767307983257631.8976730992040111.94E-87passed
/comparisons/rows/1/significantexactFalseFalsepassed
/comparisons/rows/2/differencetolerance2.56666666666666642.5666666666666615.30E-1514passed
/comparisons/rows/2/standard_errortolerance0.767322011142042350.76732201114204273.31E-1615passed
/comparisons/rows/2/statistictolerance3.34496681888034563.3449668188803378.40E-1514passed
/comparisons/rows/2/dftolerance1515.00.00E+2passed
/comparisons/rows/2/p_adjustedtail_probability0.011605771298891510.0116057712984184444.73E-1310passed
/comparisons/rows/2/lowertolerance0.573570587069064390.5735705686633571.84E-87passed
/comparisons/rows/2/uppertolerance4.55976274626426834.5597627646699661.84E-88passed
/comparisons/rows/2/significantexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

44  emit <- function(id, title, kind, config, outputs, notes, source_ref, citation) {
45    write_expected("hypothesis", id, list(
46      schema_version = "2.0.0", case_id = id, outputs = outputs, expected_error = NULL,
47      deferred_outputs = list(), generator_run = run))
48    write_case("hypothesis", id, list(
49      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "hypothesis",
50      entry_point = entry, menu_items = list(menu), title = title, case_kind = kind,
51      dataset = list(path = "reference_data/datasets/HYP_POSTHOC.json", sha256 = ""),
52      excluded_rows = list(), configuration = config, "repeat" = 1L,
53      expected = list(path = paste0("reference_data/expected/hypothesis/", id, ".json"), sha256 = ""),
54      oracle = list(oracle_class = "independent_implementation", source = source_ref,
55                    citation = citation,
56                    generator = list(path = "generators/r/hypothesis_anova_post_hoc.R", sha256 = ""),
57                    independent = TRUE,
58                    convention_notes = c(notes, notes_common, "Expected computed from the exported CSV.")),
59      supported = TRUE, unsupported_reason = NULL,
60      requirement_ids = list("REQ-STAT-HYP-ANOVA-POST-HOC-001"),
61      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/hypothesis_anova_post_hoc.R


One-way ANOVA Bonferroni

Case STAT-HYP-ANOVA-POST-HOC-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::pairwise.t.test
Citation
R Core Team, stats package help: pairwise.t.test, p.adjust, qt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_anova_post_hoc.R · sha256 057b84f90d3c52ae
Dataset
reference_data/datasets/HYP_POSTHOC.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/between_sstolerance28.20666666666664228.2066666666665964.61E-1414passed
/anova/within_sstolerance28.53333333333331728.5333333333333462.86E-1414passed
/anova/between_dfexact220.00E+2passed
/anova/within_dfexact15150.00E+2passed
/anova/msetolerance1.90222222222222111.90222222222222312.05E-1514passed
/anova/f_valuetolerance7.41413551401868987.4141355140186692.05E-1414passed
/anova/p_valuetail_probability0.00576731720866283020.00576731720866288665.64E-1714passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/groups/0/nexact550.00E+2passed
/groups/0/meantolerance10.910.8999999999999991.42E-1515passed
/groups/0/variancetolerance0.874999999999999670.87500000000000036.63E-1615passed
/groups/1/nexact660.00E+2passed
/groups/1/meantolerance13.66666666666666613.6666666666666641.70E-1515passed
/groups/1/variancetolerance1.35066666666666731.35066666666666851.16E-1515passed
/groups/2/nexact770.00E+2passed
/groups/2/meantolerance11.111.1000000000000033.20E-1515passed
/groups/2/variancetolerance3.04666666666666643.0466666666666674.56E-1615passed
/comparisons/critical_valuetolerance2.69373931919648652.69373931919648821.74E-1515passed
/comparisons/rows/0/differencetolerance-2.7666666666666657-2.76666666666666571.93E-1717passed
/comparisons/rows/0/standard_errortolerance0.83515356760387560.8351535676038764.46E-1615passed
/comparisons/rows/0/statistictolerance-3.3127639921415417-3.312763992141541.76E-1515passed
/comparisons/rows/0/dftolerance1515.00.00E+2passed
/comparisons/rows/0/p_adjustedtail_probability0.0142001484441106330.0142001484441106855.20E-1714passed
/comparisons/rows/0/significantexactTrueTruepassed
/comparisons/rows/0/lowertolerance-5.0163526692884464-5.0163526692884492.63E-1515passed
/comparisons/rows/0/uppertolerance-0.51698066404488507-0.51698066404488242.66E-1514passed
/comparisons/rows/1/differencetolerance-0.19999999999999929-0.200000000000004625.33E-1513passed
/comparisons/rows/1/standard_errortolerance0.807583107915511020.80758310791551154.47E-1615passed
/comparisons/rows/1/statistictolerance-0.24765253017268807-0.24765253017269456.44E-1513passed
/comparisons/rows/1/dftolerance1515.00.00E+2passed
/comparisons/rows/1/p_adjustedtail_probability11.00.00E+2passed
/comparisons/rows/1/significantexactFalseFalsepassed
/comparisons/rows/1/lowertolerance-2.3754183713109107-2.37541837131091878.01E-1514passed
/comparisons/rows/1/uppertolerance1.97541837131091211.97541837131090952.63E-1514passed
/comparisons/rows/2/differencetolerance2.56666666666666642.5666666666666615.30E-1514passed
/comparisons/rows/2/standard_errortolerance0.767322011142042350.76732201114204273.31E-1615passed
/comparisons/rows/2/statistictolerance3.34496681888034563.3449668188803378.40E-1514passed
/comparisons/rows/2/dftolerance1515.00.00E+2passed
/comparisons/rows/2/p_adjustedtail_probability0.0132925596081008830.0132925596081011012.18E-1613passed
/comparisons/rows/2/significantexactTrueTruepassed
/comparisons/rows/2/lowertolerance0.499701194768422670.49970119476841477.99E-1513passed
/comparisons/rows/2/uppertolerance4.63363213856491024.63363213856490752.68E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

162  simple_case("STAT-HYP-ANOVA-POST-HOC-002", "One-way ANOVA Bonferroni", "bonferroni", "bonferroni")

Whole script generators/r/hypothesis_anova_post_hoc.R


One-way ANOVA Holm

Case STAT-HYP-ANOVA-POST-HOC-003 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::pairwise.t.test
Citation
R Core Team, stats package help: pairwise.t.test, p.adjust, qt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_anova_post_hoc.R · sha256 057b84f90d3c52ae
Dataset
reference_data/datasets/HYP_POSTHOC.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/between_sstolerance28.20666666666664228.2066666666665964.61E-1414passed
/anova/within_sstolerance28.53333333333331728.5333333333333462.86E-1414passed
/anova/between_dfexact220.00E+2passed
/anova/within_dfexact15150.00E+2passed
/anova/msetolerance1.90222222222222111.90222222222222312.05E-1514passed
/anova/f_valuetolerance7.41413551401868987.4141355140186692.05E-1414passed
/anova/p_valuetail_probability0.00576731720866283020.00576731720866288665.64E-1714passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/groups/0/nexact550.00E+2passed
/groups/0/meantolerance10.910.8999999999999991.42E-1515passed
/groups/0/variancetolerance0.874999999999999670.87500000000000036.63E-1615passed
/groups/1/nexact660.00E+2passed
/groups/1/meantolerance13.66666666666666613.6666666666666641.70E-1515passed
/groups/1/variancetolerance1.35066666666666731.35066666666666851.16E-1515passed
/groups/2/nexact770.00E+2passed
/groups/2/meantolerance11.111.1000000000000033.20E-1515passed
/groups/2/variancetolerance3.04666666666666643.0466666666666674.56E-1615passed
/comparisons/rows/0/differencetolerance-2.7666666666666657-2.76666666666666571.93E-1717passed
/comparisons/rows/0/standard_errortolerance0.83515356760387560.8351535676038764.46E-1615passed
/comparisons/rows/0/statistictolerance-3.3127639921415417-3.312763992141541.76E-1515passed
/comparisons/rows/0/dftolerance1515.00.00E+2passed
/comparisons/rows/0/p_adjustedtail_probability0.0132925596081008830.0132925596081011012.18E-1613passed
/comparisons/rows/0/significantexactTrueTruepassed
/comparisons/rows/0/lowerabsentNonepassed
/comparisons/rows/1/differencetolerance-0.19999999999999929-0.200000000000004625.33E-1513passed
/comparisons/rows/1/standard_errortolerance0.807583107915511020.80758310791551154.47E-1615passed
/comparisons/rows/1/statistictolerance-0.24765253017268807-0.24765253017269456.44E-1513passed
/comparisons/rows/1/dftolerance1515.00.00E+2passed
/comparisons/rows/1/p_adjustedtail_probability0.807761238629737480.80776123862973255.00E-1514passed
/comparisons/rows/1/significantexactFalseFalsepassed
/comparisons/rows/1/lowerabsentNonepassed
/comparisons/rows/2/differencetolerance2.56666666666666642.5666666666666615.30E-1514passed
/comparisons/rows/2/standard_errortolerance0.767322011142042350.76732201114204273.31E-1615passed
/comparisons/rows/2/statistictolerance3.34496681888034563.3449668188803378.40E-1514passed
/comparisons/rows/2/dftolerance1515.00.00E+2passed
/comparisons/rows/2/p_adjustedtail_probability0.0132925596081008830.0132925596081011012.18E-1613passed
/comparisons/rows/2/significantexactTrueTruepassed
/comparisons/rows/2/lowerabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

163  simple_case("STAT-HYP-ANOVA-POST-HOC-003", "One-way ANOVA Holm", "holm", "holm")

Whole script generators/r/hypothesis_anova_post_hoc.R


One-way ANOVA Scheffe

Case STAT-HYP-ANOVA-POST-HOC-004 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::pf, qf
Citation
Scheffe H (1959), The Analysis of Variance, Wiley, section 3.5; R Core Team, stats package help: pf, qf.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_anova_post_hoc.R · sha256 057b84f90d3c52ae
Dataset
reference_data/datasets/HYP_POSTHOC.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/between_sstolerance28.20666666666664228.2066666666665964.61E-1414passed
/anova/within_sstolerance28.53333333333331728.5333333333333462.86E-1414passed
/anova/between_dfexact220.00E+2passed
/anova/within_dfexact15150.00E+2passed
/anova/msetolerance1.90222222222222111.90222222222222312.05E-1514passed
/anova/f_valuetolerance7.41413551401868987.4141355140186692.05E-1414passed
/anova/p_valuetail_probability0.00576731720866283020.00576731720866288665.64E-1714passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/groups/0/nexact550.00E+2passed
/groups/0/meantolerance10.910.8999999999999991.42E-1515passed
/groups/0/variancetolerance0.874999999999999670.87500000000000036.63E-1615passed
/groups/1/nexact660.00E+2passed
/groups/1/meantolerance13.66666666666666613.6666666666666641.70E-1515passed
/groups/1/variancetolerance1.35066666666666731.35066666666666851.16E-1515passed
/groups/2/nexact770.00E+2passed
/groups/2/meantolerance11.111.1000000000000033.20E-1515passed
/groups/2/variancetolerance3.04666666666666643.0466666666666674.56E-1615passed
/comparisons/critical_valuetolerance2.71378714849681612.71378714849681664.70E-1615passed
/comparisons/rows/0/differencetolerance-2.7666666666666657-2.76666666666666571.93E-1717passed
/comparisons/rows/0/standard_errortolerance0.83515356760387560.8351535676038764.46E-1615passed
/comparisons/rows/0/statistictolerance-3.3127639921415417-3.312763992141541.76E-1515passed
/comparisons/rows/0/dftolerance1515.00.00E+2passed
/comparisons/rows/0/p_adjustedtail_probability0.0162776741413452120.016277674141345242.80E-1714passed
/comparisons/rows/0/lowertolerance-5.0330956854513307-5.03309568545133251.76E-1515passed
/comparisons/rows/0/uppertolerance-0.5002376478820012-0.50023764788199941.77E-1514passed
/comparisons/rows/0/significantexactTrueTruepassed
/comparisons/rows/1/differencetolerance-0.19999999999999929-0.200000000000004625.33E-1513passed
/comparisons/rows/1/standard_errortolerance0.807583107915511020.80758310791551154.47E-1615passed
/comparisons/rows/1/statistictolerance-0.24765253017268807-0.24765253017269456.44E-1513passed
/comparisons/rows/1/dftolerance1515.00.00E+2passed
/comparisons/rows/1/p_adjustedtail_probability0.969860177183219820.96986017718321821.67E-1514passed
/comparisons/rows/1/lowertolerance-2.3916086596042305-2.3916086596042376.69E-1514passed
/comparisons/rows/1/uppertolerance1.99160865960423191.9916086596042283.95E-1514passed
/comparisons/rows/1/significantexactFalseFalsepassed
/comparisons/rows/2/differencetolerance2.56666666666666642.5666666666666615.30E-1514passed
/comparisons/rows/2/standard_errortolerance0.767322011142042350.76732201114204273.31E-1615passed
/comparisons/rows/2/statistictolerance3.34496681888034563.3449668188803378.40E-1514passed
/comparisons/rows/2/dftolerance1515.00.00E+2passed
/comparisons/rows/2/p_adjustedtail_probability0.0153044270180666560.0153044270180668752.19E-1613passed
/comparisons/rows/2/lowertolerance0.484318054070661310.484318054070654656.66E-1513passed
/comparisons/rows/2/uppertolerance4.6490152792626724.64901527926266764.45E-1515passed
/comparisons/rows/2/significantexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

182  emit("STAT-HYP-ANOVA-POST-HOC-004", "One-way ANOVA Scheffe", "axis_coverage",
183       list(response = "y", group = "group", procedure = "scheffe", alpha = 0.05),
184       outputs, "Scheffe (1959): a pairwise contrast is significant when t^2/(k-1) exceeds F_{k-1, N-k}(alpha); p = pf(t^2/(k-1), k-1, df, lower.tail = FALSE); interval half-width sqrt((k-1) F) SE.",
185       "R 4.6.1 stats::pf, qf", "Scheffe H (1959), The Analysis of Variance, Wiley, section 3.5; R Core Team, stats package help: pf, qf.")

Whole script generators/r/hypothesis_anova_post_hoc.R


One-way ANOVA LSD alpha 0.10

Case STAT-HYP-ANOVA-POST-HOC-005 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::pairwise.t.test
Citation
R Core Team, stats package help: pairwise.t.test, p.adjust, qt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_anova_post_hoc.R · sha256 057b84f90d3c52ae
Dataset
reference_data/datasets/HYP_POSTHOC.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/between_sstolerance28.20666666666664228.2066666666665964.61E-1414passed
/anova/within_sstolerance28.53333333333331728.5333333333333462.86E-1414passed
/anova/between_dfexact220.00E+2passed
/anova/within_dfexact15150.00E+2passed
/anova/msetolerance1.90222222222222111.90222222222222312.05E-1514passed
/anova/f_valuetolerance7.41413551401868987.4141355140186692.05E-1414passed
/anova/p_valuetail_probability0.00576731720866283020.00576731720866288665.64E-1714passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/groups/0/nexact550.00E+2passed
/groups/0/meantolerance10.910.8999999999999991.42E-1515passed
/groups/0/variancetolerance0.874999999999999670.87500000000000036.63E-1615passed
/groups/1/nexact660.00E+2passed
/groups/1/meantolerance13.66666666666666613.6666666666666641.70E-1515passed
/groups/1/variancetolerance1.35066666666666731.35066666666666851.16E-1515passed
/groups/2/nexact770.00E+2passed
/groups/2/meantolerance11.111.1000000000000033.20E-1515passed
/groups/2/variancetolerance3.04666666666666643.0466666666666674.56E-1615passed
/comparisons/critical_valuetolerance1.75305035569257271.7530503556925726.40E-1615passed
/comparisons/rows/0/differencetolerance-2.7666666666666657-2.76666666666666571.93E-1717passed
/comparisons/rows/0/standard_errortolerance0.83515356760387560.8351535676038764.46E-1615passed
/comparisons/rows/0/statistictolerance-3.3127639921415417-3.312763992141541.76E-1515passed
/comparisons/rows/0/dftolerance1515.00.00E+2passed
/comparisons/rows/0/p_adjustedtail_probability0.00473338281470354460.0047333828147035621.73E-1714passed
/comparisons/rows/0/significantexactTrueTruepassed
/comparisons/rows/0/lowertolerance-4.2307329254125605-4.2307329254125619.07E-1615passed
/comparisons/rows/0/uppertolerance-1.3026004079207705-1.30260040792077032.46E-1615passed
/comparisons/rows/1/differencetolerance-0.19999999999999929-0.200000000000004625.33E-1513passed
/comparisons/rows/1/standard_errortolerance0.807583107915511020.80758310791551154.47E-1615passed
/comparisons/rows/1/statistictolerance-0.24765253017268807-0.24765253017269456.44E-1513passed
/comparisons/rows/1/dftolerance1515.00.00E+2passed
/comparisons/rows/1/p_adjustedtail_probability0.807761238629737480.80776123862973255.00E-1514passed
/comparisons/rows/1/significantexactFalseFalsepassed
/comparisons/rows/1/lowertolerance-1.6157338545825992-1.61573385458260485.57E-1514passed
/comparisons/rows/1/uppertolerance1.21573385458260061.21573385458259555.07E-1514passed
/comparisons/rows/2/differencetolerance2.56666666666666642.5666666666666615.30E-1514passed
/comparisons/rows/2/standard_errortolerance0.767322011142042350.76732201114204273.31E-1615passed
/comparisons/rows/2/statistictolerance3.34496681888034563.3449668188803378.40E-1514passed
/comparisons/rows/2/dftolerance1515.00.00E+2passed
/comparisons/rows/2/p_adjustedtail_probability0.00443085320270029430.0044308532027003677.28E-1713passed
/comparisons/rows/2/significantexactTrueTruepassed
/comparisons/rows/2/lowertolerance1.22151254210336881.22151254210336355.34E-1514passed
/comparisons/rows/2/uppertolerance3.91182079122996383.91182079122995855.28E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

164  simple_case("STAT-HYP-ANOVA-POST-HOC-005", "One-way ANOVA LSD alpha 0.10", "lsd", "none", alpha = 0.10)

Whole script generators/r/hypothesis_anova_post_hoc.R


One-way ANOVA Dunnett against control A

Case STAT-HYP-ANOVA-POST-HOC-006 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::integrate, uniroot
Citation
Dunnett CW (1955), A multiple comparison procedure for comparing several treatments with a control, JASA 50, 1096-1121; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_anova_post_hoc.R · sha256 057b84f90d3c52ae
Dataset
reference_data/datasets/HYP_POSTHOC.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/between_sstolerance28.20666666666664228.2066666666665964.61E-1414passed
/anova/within_sstolerance28.53333333333331728.5333333333333462.86E-1414passed
/anova/between_dfexact220.00E+2passed
/anova/within_dfexact15150.00E+2passed
/anova/msetolerance1.90222222222222111.90222222222222312.05E-1514passed
/anova/f_valuetolerance7.41413551401868987.4141355140186692.05E-1414passed
/anova/p_valuetail_probability0.00576731720866283020.00576731720866288665.64E-1714passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/groups/0/nexact550.00E+2passed
/groups/0/meantolerance10.910.8999999999999991.42E-1515passed
/groups/0/variancetolerance0.874999999999999670.87500000000000036.63E-1615passed
/groups/1/nexact660.00E+2passed
/groups/1/meantolerance13.66666666666666613.6666666666666641.70E-1515passed
/groups/1/variancetolerance1.35066666666666731.35066666666666851.16E-1515passed
/groups/2/nexact770.00E+2passed
/groups/2/meantolerance11.111.1000000000000033.20E-1515passed
/groups/2/variancetolerance3.04666666666666643.0466666666666674.56E-1615passed
/comparisons/critical_valuetolerance2.42815185291658552.4281329517735831.89E-55passed
/comparisons/rows/0/differencetolerance2.76666666666666572.76666666666666571.93E-1717passed
/comparisons/rows/0/standard_errortolerance0.83515356760387560.83515356760387593.35E-1615passed
/comparisons/rows/0/statistictolerance3.31276399214154173.31276399214154041.32E-1515passed
/comparisons/rows/0/dftolerance1515.00.00E+2passed
/comparisons/rows/0/p_adjustedtolerance0.00870014381869288480.008699326532524888.17E-74passed
/comparisons/rows/0/lowertolerance0.738786984019418110.73880276937642811.58E-54passed
/comparisons/rows/0/uppertolerance4.79454634931391334.7945305639569031.58E-55passed
/comparisons/rows/0/significantexactTrueTruepassed
/comparisons/rows/1/differencetolerance0.199999999999999290.200000000000004625.33E-1513passed
/comparisons/rows/1/standard_errortolerance0.807583107915511020.80758310791551143.36E-1615passed
/comparisons/rows/1/statistictolerance0.247652530172688070.247652530172694566.49E-1513passed
/comparisons/rows/1/dftolerance1515.00.00E+2passed
/comparisons/rows/1/p_adjustedtolerance0.954270253487924490.95425438167121221.59E-54passed
/comparisons/rows/1/lowertolerance-1.7609344198691836-1.76091915562536981.53E-55passed
/comparisons/rows/1/uppertolerance2.16093441986918222.1609191556253791.53E-55passed
/comparisons/rows/1/significantexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

221  emit("STAT-HYP-ANOVA-POST-HOC-006", "One-way ANOVA Dunnett against control A", "axis_coverage",
222       list(response = "y", group = "group", procedure = "dunnett", alpha = 0.05, control = "A"),
223       outputs, c("Dunnett (1955): rows are (B, A) and (C, A); critical value and adjusted p from the bivariate t with correlation sqrt(n_i n_j / ((n_i + n_0)(n_j + n_0))) on N - k df, computed by deterministic quadrature (product-correlation integral over W, then over the chi scale; rel.tol 1e-11; uniroot to 1e-12).",
224                     "The window calibrates the maximum multivariate t by replicated quasi-Monte Carlo with a stated target half-width of 1e-4 on the probability (docs/manual/chapters/83-repeated-anova-and-contrasts.tex:149-153), so the comparison allows 1e-4 absolute and relative; the reference itself is accurate to about 1e-10."),
225       "R 4.6.1 stats::integrate, uniroot", "Dunnett CW (1955), A multiple comparison procedure for comparing several treatments with a control, JASA 50, 1096-1121; R Core Team, stats package help: integrate, uniroot.")

Whole script generators/r/hypothesis_anova_post_hoc.R


One-way ANOVA Games-Howell

Case STAT-HYP-ANOVA-POST-HOC-007 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::ptukey, qtukey
Citation
Games PA, Howell JF (1976), Journal of Educational Statistics 1(2), 113-125; R Core Team, stats package help: ptukey, qtukey.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_anova_post_hoc.R · sha256 057b84f90d3c52ae
Dataset
reference_data/datasets/HYP_POSTHOC.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/between_sstolerance28.20666666666664228.2066666666665964.61E-1414passed
/anova/within_sstolerance28.53333333333331728.5333333333333462.86E-1414passed
/anova/between_dfexact220.00E+2passed
/anova/within_dfexact15150.00E+2passed
/anova/msetolerance1.90222222222222111.90222222222222312.05E-1514passed
/anova/f_valuetolerance7.41413551401868987.4141355140186692.05E-1414passed
/anova/p_valuetail_probability0.00576731720866283020.00576731720866288665.64E-1714passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/groups/0/nexact550.00E+2passed
/groups/0/meantolerance10.910.8999999999999991.42E-1515passed
/groups/0/variancetolerance0.874999999999999670.87500000000000036.63E-1615passed
/groups/1/nexact660.00E+2passed
/groups/1/meantolerance13.66666666666666613.6666666666666641.70E-1515passed
/groups/1/variancetolerance1.35066666666666731.35066666666666851.16E-1515passed
/groups/2/nexact770.00E+2passed
/groups/2/meantolerance11.111.1000000000000033.20E-1515passed
/groups/2/variancetolerance3.04666666666666643.0466666666666674.56E-1615passed
/comparisons/rows/0/differencetolerance-2.7666666666666657-2.76666666666666571.93E-1717passed
/comparisons/rows/0/standard_errortolerance0.632543366980566120.63254336698056632.26E-1615passed
/comparisons/rows/0/statistictolerance-4.3738766558778366-4.3738766558778351.79E-1515passed
/comparisons/rows/0/dftolerance8.99818051102863958.9981805110286381.80E-1515passed
/comparisons/rows/0/p_adjustedtail_probability0.00456559032921477480.00456559048753424351.58E-107passed
/comparisons/rows/0/lowertolerance-4.5327970806466436-4.5327970811791885.33E-109passed
/comparisons/rows/0/uppertolerance-1.0005362526866879-1.00053625215414285.33E-109passed
/comparisons/rows/1/differencetolerance-0.19999999999999929-0.200000000000004625.33E-1513passed
/comparisons/rows/1/standard_errortolerance0.781177377577010450.78117737757701061.07E-1615passed
/comparisons/rows/1/statistictolerance-0.25602379912785272-0.25602379912785956.78E-1513passed
/comparisons/rows/1/dftolerance9.49290924695950989.4929092469595121.75E-1515passed
/comparisons/rows/1/p_adjustedtail_probability0.964644379538921350.96464437953908111.60E-1312passed
/comparisons/rows/1/lowertolerance-2.3603428415803225-2.360342842010284.30E-109passed
/comparisons/rows/1/uppertolerance1.96034284158032391.96034284201027064.30E-109passed
/comparisons/rows/2/differencetolerance2.56666666666666642.5666666666666615.30E-1514passed
/comparisons/rows/2/standard_errortolerance0.8126187336932410.81261873369324111.13E-1615passed
/comparisons/rows/2/statistictolerance3.15851279357235273.15851279357234567.10E-1514passed
/comparisons/rows/2/dftolerance10.45533605364699710.4553360536469972.99E-1717passed
/comparisons/rows/2/p_adjustedtail_probability0.0240470527621211570.0240470527523226629.80E-129passed
/comparisons/rows/2/lowertolerance0.354875836742392540.3548758359675957.75E-108passed
/comparisons/rows/2/uppertolerance4.77845749659094034.7784574973657277.75E-109passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

246  emit("STAT-HYP-ANOVA-POST-HOC-007", "One-way ANOVA Games-Howell", "axis_coverage",
247       list(response = "y", group = "group", procedure = "games_howell", alpha = 0.05),
248       outputs, "Games and Howell (1976): SE = sqrt(s_i^2/n_i + s_j^2/n_j), Welch-Satterthwaite df per pair, p = ptukey(|t| sqrt(2), k, df_ij, lower = FALSE), half-width qtukey(0.95, k, df_ij) SE / sqrt(2). R documents qtukey as accurate to the 4th decimal place, so ptukey/qtukey-derived values are compared at 1e-6 relative.",
249       "R 4.6.1 stats::ptukey, qtukey", "Games PA, Howell JF (1976), Journal of Educational Statistics 1(2), 113-125; R Core Team, stats package help: ptukey, qtukey.")

Whole script generators/r/hypothesis_anova_post_hoc.R


One-way ANOVA Helmert contrasts pooled

Case STAT-HYP-ANOVA-POST-HOC-008 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::pt, qt
Citation
R Core Team, stats package help: pt, qt; contrast definition as in the SPSS ONEWAY Helmert family.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_anova_post_hoc.R · sha256 057b84f90d3c52ae
Dataset
reference_data/datasets/HYP_POSTHOC.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/between_sstolerance28.20666666666664228.2066666666665964.61E-1414passed
/anova/within_sstolerance28.53333333333331728.5333333333333462.86E-1414passed
/anova/between_dfexact220.00E+2passed
/anova/within_dfexact15150.00E+2passed
/anova/msetolerance1.90222222222222111.90222222222222312.05E-1514passed
/anova/f_valuetolerance7.41413551401868987.4141355140186692.05E-1414passed
/anova/p_valuetail_probability0.00576731720866283020.00576731720866288665.64E-1714passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/groups/0/nexact550.00E+2passed
/groups/0/meantolerance10.910.8999999999999991.42E-1515passed
/groups/0/variancetolerance0.874999999999999670.87500000000000036.63E-1615passed
/groups/1/nexact660.00E+2passed
/groups/1/meantolerance13.66666666666666613.6666666666666641.70E-1515passed
/groups/1/variancetolerance1.35066666666666731.35066666666666851.16E-1515passed
/groups/2/nexact770.00E+2passed
/groups/2/meantolerance11.111.1000000000000033.20E-1515passed
/groups/2/variancetolerance3.04666666666666643.0466666666666674.56E-1615passed
/contrasts/rows/0/estimatetolerance-1.4833333333333325-1.48333333333333522.67E-1514passed
/contrasts/rows/0/standard_errortolerance0.726388471577166020.72638847157716654.41E-1615passed
/contrasts/rows/0/dftolerance1515.00.00E+2passed
/contrasts/rows/0/t_valuetolerance-2.0420661827309226-2.04206618273092482.17E-1514passed
/contrasts/rows/0/p_valuetail_probability0.0591367577431764490.0591367577431762062.43E-1614passed
/contrasts/rows/0/lowertolerance-3.0315937109763427-3.03159371097634623.52E-1514passed
/contrasts/rows/0/uppertolerance0.0649270443096776620.064927044309676111.55E-1513passed
/contrasts/rows/1/estimatetolerance2.56666666666666642.5666666666666615.30E-1514passed
/contrasts/rows/1/standard_errortolerance0.767322011142042350.76732201114204273.31E-1615passed
/contrasts/rows/1/dftolerance1515.00.00E+2passed
/contrasts/rows/1/t_valuetolerance3.34496681888034563.3449668188803378.40E-1514passed
/contrasts/rows/1/p_valuetail_probability0.00443085320270029430.0044308532027003677.28E-1713passed
/contrasts/rows/1/lowertolerance0.931158514719947710.93115851471994136.44E-1514passed
/contrasts/rows/1/uppertolerance4.20217481861338544.2021748186133814.46E-1514passed
/comparisonsabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

277  helmert <- rbind(c(1, -1/2, -1/2), c(0, 1, -1))
278  outputs <- c(anova_outputs(), contrast_rows(helmert, c("A vs later", "B vs later"), "pooled"))
279  outputs[["/comparisons"]] <- list(comparison = "absent", rationale = "procedure none: no pairwise table")
280  emit("STAT-HYP-ANOVA-POST-HOC-008", "One-way ANOVA Helmert contrasts pooled", "axis_coverage",
281       list(response = "y", group = "group", procedure = "none", alpha = 0.05,
282            contrast_family = "helmert", contrast_variance = "pooled", contrast_adjustment = "none"),
283       outputs, "Helmert: each level against the mean of the later levels, rows (1, -1/2, -1/2) and (0, 1, -1); estimate = c'means, SE = sqrt(MSE sum(c_i^2 / n_i)), df = N - k, individual t intervals.",
284       "R 4.6.1 stats::pt, qt", "R Core Team, stats package help: pt, qt; contrast definition as in the SPSS ONEWAY Helmert family.")

Whole script generators/r/hypothesis_anova_post_hoc.R


One-way ANOVA custom contrasts Welch Bonferroni

Case STAT-HYP-ANOVA-POST-HOC-009 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::pt, qt
Citation
Welch BL (1947), Biometrika 34, 28-35 (Satterthwaite df); R Core Team, stats package help: pt, qt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_anova_post_hoc.R · sha256 057b84f90d3c52ae
Dataset
reference_data/datasets/HYP_POSTHOC.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/between_sstolerance28.20666666666664228.2066666666665964.61E-1414passed
/anova/within_sstolerance28.53333333333331728.5333333333333462.86E-1414passed
/anova/between_dfexact220.00E+2passed
/anova/within_dfexact15150.00E+2passed
/anova/msetolerance1.90222222222222111.90222222222222312.05E-1514passed
/anova/f_valuetolerance7.41413551401868987.4141355140186692.05E-1414passed
/anova/p_valuetail_probability0.00576731720866283020.00576731720866288665.64E-1714passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/groups/0/nexact550.00E+2passed
/groups/0/meantolerance10.910.8999999999999991.42E-1515passed
/groups/0/variancetolerance0.874999999999999670.87500000000000036.63E-1615passed
/groups/1/nexact660.00E+2passed
/groups/1/meantolerance13.66666666666666613.6666666666666641.70E-1515passed
/groups/1/variancetolerance1.35066666666666731.35066666666666851.16E-1515passed
/groups/2/nexact770.00E+2passed
/groups/2/meantolerance11.111.1000000000000033.20E-1515passed
/groups/2/variancetolerance3.04666666666666643.0466666666666674.56E-1615passed
/contrasts/rows/0/estimatetolerance-1.4833333333333325-1.48333333333333522.67E-1514passed
/contrasts/rows/0/standard_errortolerance0.583170045173190930.5831700451731911.13E-1615passed
/contrasts/rows/0/dftolerance11.26961433182945711.2696143318294533.51E-1515passed
/contrasts/rows/0/t_valuetolerance-2.5435691452444362-2.543569145244444.02E-1514passed
/contrasts/rows/0/p_valuetail_probability0.0268619487525262840.0268619487525261461.38E-1614passed
/contrasts/rows/0/lowertolerance-2.9899676416378771-2.98996764163787982.66E-1515passed
/contrasts/rows/0/uppertolerance0.0233009749712118630.023300974971209422.44E-1512passed
/contrasts/rows/1/estimatetolerance-0.19999999999999929-0.200000000000004625.33E-1513passed
/contrasts/rows/1/standard_errortolerance0.781177377577010450.78117737757701061.07E-1615passed
/contrasts/rows/1/dftolerance9.49290924695950989.4929092469595121.75E-1515passed
/contrasts/rows/1/t_valuetolerance-0.25602379912785272-0.25602379912785956.78E-1513passed
/contrasts/rows/1/p_valuetail_probability0.803394546912140180.80339454691213515.10E-1514passed
/contrasts/rows/1/lowertolerance-2.276517475567617-2.2765174755676236.18E-1514passed
/contrasts/rows/1/uppertolerance1.87651747556761841.8765174755676144.45E-1514passed
/contrasts/rows/0/p_adjustedtail_probability0.0537238975050525690.053723897505052292.78E-1614passed
/contrasts/rows/1/p_adjustedtail_probability11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

300  emit("STAT-HYP-ANOVA-POST-HOC-009", "One-way ANOVA custom contrasts Welch Bonferroni", "axis_coverage",
301       list(response = "y", group = "group", procedure = "none", alpha = 0.05,
302            contrast_family = "custom", coefficients = list(1, -0.5, -0.5, 1, 0, -1),
303            contrast_variance = "welch", contrast_adjustment = "bonferroni"),
304       outputs, "Custom rows (1, -0.5, -0.5) and (1, 0, -1): SE = sqrt(sum c_i^2 s_i^2 / n_i), Welch-Satterthwaite df, Bonferroni over m = 2 contrasts (p times 2, qt(1 - alpha/(2m))).",
305       "R 4.6.1 stats::pt, qt", "Welch BL (1947), Biometrika 34, 28-35 (Satterthwaite df); R Core Team, stats package help: pt, qt.")

Whole script generators/r/hypothesis_anova_post_hoc.R


Configuration axis: STAT-HYP-AXIS-BASIC-POISSON-001

Case STAT-HYP-AXIS-BASIC-POISSON-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_RATES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.149999999999999990.154.45E-1816passed
/p_valuetail_probability11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

74  one_rate <- stats::poisson.test(sum(r$count), T = sum(r$exposure), r = 0.15)
75  emit("STAT-HYP-AXIS-BASIC-POISSON-001", "tabelz.executor.additional_basic_tests",
76       "additionalBasicTestsMenuItem", "HYP_RATES",
77       list(method = "one_poisson_rate", input_kind = "raw", count_column = "count",
78            exposure_column = "exposure", null_rate = 0.15, exposure_unit = "hours"),
79       list("/estimate" = tol(sum(r$count) / sum(r$exposure), "events per exposure"),
80            "/p_value" = prob(one_rate$p.value, "exact one-rate Poisson tail")),
81       list("Hours are display units; exact Poisson test uses total count and total exposure."))

Whole script generators/r/hypothesis_axes.R


Configuration axis: STAT-HYP-AXIS-BASIC-SD-001

Case STAT-HYP-AXIS-BASIC-SD-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance13.90937500000000113.9093750000000043.26E-1515passed
/p_valuetail_probability0.935173631038799910.93517363103880077.80E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

107  variance <- stats::var(d$x)
108  chi <- (nrow(d) - 1) * variance / 4
109  emit("STAT-HYP-AXIS-BASIC-SD-001", "tabelz.executor.additional_basic_tests",
110       "additionalBasicTestsMenuItem", "HYP_CORE",
111       list(method = "one_variance_chi_square", input_kind = "raw", response_column = "x",
112            null_variance = 4, variance_display = "sd"),
113       list("/statistic" = tol(chi, "variance chi-square"),
114            "/p_value" = prob(min(1, 2 * min(stats::pchisq(chi, nrow(d) - 1),
115                                             stats::pchisq(chi, nrow(d) - 1, lower.tail = FALSE))),
116                              "two-sided variance tail")),
117       list("Null variance 4; sd display is a presentation axis only."))

Whole script generators/r/hypothesis_axes.R


Configuration axis: STAT-HYP-AXIS-BASIC-SUMMARY-001

Case STAT-HYP-AXIS-BASIC-SUMMARY-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.6251.6250.00E-1passed
/p_valuetail_probability0.104162558830439120.104162558830439056.68E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 97  z_summary <- (mean(d$x) - 6) / (2 / sqrt(nrow(d)))
 98  emit("STAT-HYP-AXIS-BASIC-SUMMARY-001", "tabelz.executor.additional_basic_tests",
 99       "additionalBasicTestsMenuItem", NULL,
100       list(method = "known_sigma_z", input_kind = "summary", summary_mean = mean(d$x),
101            sample_size = nrow(d), known_sigma = 2, null_mean = 6),
102       list("/statistic" = tol(z_summary, "summary known-sigma z"),
103            "/p_value" = prob(2 * stats::pnorm(-abs(z_summary)), "normal tail")),
104       list("Summary mean and sample size are derived from the exported CSV.",
105            "docs/manual/chapters/31-additional-basic-tests.tex:6-8: 'Summary mode is source independent: it does not claim a table population'."))

Whole script generators/r/hypothesis_axes.R


Additional configuration coverage: STAT-HYP-AXIS-BASIC-SUMMARY-TWORATE-001

Case STAT-HYP-AXIS-BASIC-SUMMARY-TWORATE-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats/multcomp
Citation
R stats::pnorm, pt, qt, binom.test, power.t.test; multcomp::glht and mvtnorm documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, multcomp 1.4.32, mvtnorm 1.4.2, jsonlite 2.0.0
Reference script
generators/r/hypothesis_completion.R · sha256 733ecb7b7daec34d
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/p_valuetail_probability0.899674257844005230.89967425784400531.13E-1615passed
/estimatetolerance0.552631578947368470.55263157894736853.64E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

119  count_a <- sum(r$count[r$g == "A"])
120  count_b <- sum(r$count[r$g == "B"])
121  exposure_a <- sum(r$exposure[r$g == "A"])
122  exposure_b <- sum(r$exposure[r$g == "B"])
123  null_probability <- exposure_a / (exposure_a + exposure_b)
124  conditional <- stats::binom.test(count_a, count_a + count_b,
125                                    p = null_probability, alternative = "greater")
126  emit("STAT-HYP-AXIS-BASIC-SUMMARY-TWORATE-001", "tabelz.executor.additional_basic_tests",
127       "additionalBasicTestsMenuItem", NULL,
128       list(method = "two_poisson_rate", input_kind = "summary", summary_count = count_a,
129            summary_exposure = exposure_a, summary_count_2 = count_b,
130            summary_exposure_2 = exposure_b, null_ratio = 1, alternative = "greater"),
131       list("/p_value" = prob(conditional$p.value, "conditional binomial right tail"),
132            "/estimate" = tol((count_a / exposure_a) / (count_b / exposure_b), "A/B rate ratio")),
133       "axis_coverage", list("Summary two-rate mode is source independent; totals derived from the R CSV."))

Whole script generators/r/hypothesis_completion.R


Configuration axis: STAT-HYP-AXIS-BASIC-TWORATE-001

Case STAT-HYP-AXIS-BASIC-TWORATE-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_RATES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.552631578947368470.55263157894736853.64E-1817passed
/p_valuetail_probability0.394913125858860240.39491312585886035.22E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

83  count_a <- sum(r$count[r$g == "A"])
84  count_b <- sum(r$count[r$g == "B"])
85  exposure_a <- sum(r$exposure[r$g == "A"])
86  exposure_b <- sum(r$exposure[r$g == "B"])
87  two_rate <- stats::poisson.test(c(count_a, count_b), T = c(exposure_a, exposure_b), r = 1)
88  emit("STAT-HYP-AXIS-BASIC-TWORATE-001", "tabelz.executor.additional_basic_tests",
89       "additionalBasicTestsMenuItem", "HYP_RATES",
90       list(method = "two_poisson_rate", input_kind = "raw", count_column = "count",
91            exposure_column = "exposure", group_column = "g", group_1 = "A", group_2 = "B",
92            null_ratio = 1, exposure_unit = "days"),
93       list("/estimate" = tol((count_a / exposure_a) / (count_b / exposure_b), "A/B rate ratio"),
94            "/p_value" = prob(two_rate$p.value, "conditional exact two-rate tail")),
95       list("Days are display units; conditional Poisson test compares A/B rate ratio."))

Whole script generators/r/hypothesis_axes.R


Configuration axis: STAT-HYP-AXIS-EQUIV-PAIRED-001

Case STAT-HYP-AXIS-EQUIV-PAIRED-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.562500000000000220.56250000000000022.04E-1817passed
/p_lowertail_probability6.7332568764626561e-206.733256876462766e-201.09E-3313passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

48  paired <- d$x - d$y
49  paired_effect <- mean(paired)
50  paired_se <- sd(paired) / sqrt(length(paired))
51  emit("STAT-HYP-AXIS-EQUIV-PAIRED-001", "tabelz.executor.equivalence_extensions",
52       "equivalenceMenuItem", "HYP_CORE",
53       list(mode = "paired_difference", first_column = "x", second_column = "y",
54            lower_bound = -2, upper_bound = 2),
55       list("/estimate" = tol(paired_effect, "paired difference mean"),
56            "/p_lower" = prob(stats::pt((paired_effect + 2) / paired_se, length(paired) - 1,
57                                         lower.tail = FALSE), "lower paired TOST")),
58       list("Paired difference is x minus y; Student t tails."))

Whole script generators/r/hypothesis_axes.R


Configuration axis: STAT-HYP-AXIS-EQUIV-RATIO-001

Case STAT-HYP-AXIS-EQUIV-RATIO-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance1.10014382872363271.10014382872363273.93E-1918passed
/p_lowertail_probability4.0254058583697218e-144.02540585836975e-142.84E-2814passed
/transformed_estimatetolerance0.0954409246419667680.09544092464196682.76E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

60  log_ratio <- log(d$x / d$y)
61  log_effect <- mean(log_ratio)
62  log_se <- sd(log_ratio) / sqrt(length(log_ratio))
63  emit("STAT-HYP-AXIS-EQUIV-RATIO-001", "tabelz.executor.equivalence_extensions",
64       "equivalenceMenuItem", "HYP_CORE",
65       list(mode = "paired_log_ratio", first_column = "x", second_column = "y",
66            lower_bound = 0.8, upper_bound = 1.25),
67       list("/estimate" = tol(exp(log_effect), "geometric mean ratio"),
68            "/p_lower" = prob(stats::pt((log_effect - log(0.8)) / log_se,
69                                         length(log_ratio) - 1, lower.tail = FALSE), "log-ratio lower TOST"),
70            "/transformed_estimate" = tol(log_effect, "mean log ratio")),
71       list("TOST on log(x/y); display estimate is exp(mean log ratio)."))

Whole script generators/r/hypothesis_axes.R


Configuration axis: STAT-HYP-AXIS-PERM-GREATER-001

Case STAT-HYP-AXIS-PERM-GREATER-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-1.5749999999999993-1.57500000000000028.78E-1615passed
/p_valuetail_probability0.949961149961150.949961149961152.38E-1817passed

Convention

Reference code

Whole script generators/r/hypothesis_axes.R


Configuration axis: STAT-HYP-AXIS-PERM-LESS-001

Case STAT-HYP-AXIS-PERM-LESS-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-1.5749999999999993-1.57500000000000028.78E-1615passed
/p_valuetail_probability0.05330225330225330.05330225330225333.95E-1917passed

Convention

Reference code

Whole script generators/r/hypothesis_axes.R


Configuration axis: STAT-HYP-AXIS-PERM-MC-001

Case STAT-HYP-AXIS-PERM-MC-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-1.5749999999999993-1.57500000000000028.78E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

158  emit("STAT-HYP-AXIS-PERM-MC-001", "tabelz.executor.permutation_tests",
159       "permutationTestsMenuItem", "HYP_CORE",
160       list(mode = "independent", response_column = "x", group_column = "g",
161            group_a = "A", group_b = "B", inference = "monte_carlo", seed = 77L,
162            resamples = 100L),
163       list("/statistic" = tol(observed, "observed A-B mean")),
164       list("Monte Carlo p and histogram are deferred to V27; compare only the observed statistic."),
165       deferred = list(list(path = "/p_value", reason = "Seeded random permutation p-values are V27 simulation evidence.")),
166       repetitions = 2L)

Whole script generators/r/hypothesis_axes.R


Configuration axis: STAT-HYP-AXIS-PERM-PAIRED-001

Case STAT-HYP-AXIS-PERM-PAIRED-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_PAIRS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance0.560000000000000160.56000000000000024.31E-1817passed
/p_valuetail_probability0.06250.06250.00E-2passed
/assignment_countexact32320.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

120  pair_df <- data.frame(x = d$x[1:5], y = d$y[1:5])
121  export_dataset(pair_df, "HYP_PAIRS", "First five pairs from exported HYP_CORE",
122                 "Constructed by hypothesis_axes.R; exact sign-flip enumeration.",
123                 "Constructed test data; no third-party dataset license.",
124                 c(x = "Float64", y = "Float64"))
125  pairs <- read_dataset("HYP_PAIRS")
126  differences <- pairs$x - pairs$y
127  observed_pair <- mean(differences)
128  signs <- as.matrix(expand.grid(rep(list(c(-1, 1)), length(differences))))
129  null_pair <- as.vector(signs %*% differences) / length(differences)
130  emit("STAT-HYP-AXIS-PERM-PAIRED-001", "tabelz.executor.permutation_tests",
131       "permutationTestsMenuItem", "HYP_PAIRS",
132       list(mode = "paired", response_column = "x", second_response_column = "y",
133            inference = "exact"),
134       list("/statistic" = tol(observed_pair, "mean paired difference"),
135            "/p_value" = prob(mean(abs(null_pair) >= abs(observed_pair) - 1e-12),
136                              "all paired sign flips"),
137            "/assignment_count" = exact(nrow(signs), "Every sign pattern enumerated.")),
138       list("Five pairs give 2^5 exact sign patterns; paired difference x-y."))

Whole script generators/r/hypothesis_axes.R


Additional configuration coverage: STAT-HYP-AXIS-PERM-PAIRED-MC-001

Case STAT-HYP-AXIS-PERM-PAIRED-MC-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats/multcomp
Citation
R stats::pnorm, pt, qt, binom.test, power.t.test; multcomp::glht and mvtnorm documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, multcomp 1.4.32, mvtnorm 1.4.2, jsonlite 2.0.0
Reference script
generators/r/hypothesis_completion.R · sha256 733ecb7b7daec34d
Dataset
reference_data/datasets/HYP_PAIRS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance0.560000000000000160.56000000000000024.31E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

107  paired <- pairs$x - pairs$y
108  emit("STAT-HYP-AXIS-PERM-PAIRED-MC-001", "tabelz.executor.permutation_tests",
109       "permutationTestsMenuItem", "HYP_PAIRS",
110       list(mode = "paired", response_column = "x", second_response_column = "y",
111            inference = "monte_carlo", resamples = 100L, seed = 42L,
112            histogram_bins = 5L, max_rows = 100L, max_cells = 1000L),
113       list("/statistic" = tol(mean(paired), "paired mean difference")), "axis_coverage",
114       list("Mode and inference axes interact; random p-value belongs to V27."),
115       deferred = list(list(path = "/p_value", reason = "Seeded Monte Carlo p-value is V27 simulation evidence.")),
116       repetitions = 2L)

Whole script generators/r/hypothesis_completion.R


Configuration axis: STAT-HYP-AXIS-RANK-GREATER-001

Case STAT-HYP-AXIS-RANK-GREATER-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact440.00E+2passed
/p_valuetail_probability0.998001998001997960.9980019980019981.37E-1817passed

Convention

Reference code

Whole script generators/r/hypothesis_axes.R


Configuration axis: STAT-HYP-AXIS-RANK-LESS-001

Case STAT-HYP-AXIS-RANK-LESS-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats help: t.test, poisson.test, binom.test, pnorm, pt, ptukey, power.t.test; exact enumeration by combn.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_axes.R · sha256 583517bdbb860353
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact440.00E+2passed
/p_valuetail_probability0.0132367632367632360.0132367632367632361.01E-1917passed

Convention

Reference code

Whole script generators/r/hypothesis_axes.R


Known-sigma one-sample Z

Case STAT-HYP-BASIC-Z-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 independent implementation
Citation
R stats::Normal documentation (pnorm and qnorm); closed-form Z statistic.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_secondary.R · sha256 13b3dbafe11fd4e1
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.6251.6250.00E-1passed
/p_valuetail_probability0.104162558830439120.104162558830439056.68E-1715passed
/estimatetolerance6.81256.81250.00E-2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

105  sigma <- 2
106  null_mean <- 6
107  z <- (mean(d$x) - null_mean) / (sigma / sqrt(length(d$x)))
108  emit("STAT-HYP-BASIC-Z-001", "tabelz.executor.additional_basic_tests", "additionalBasicTestsMenuItem",
109       "Known-sigma one-sample Z", "nominal", "HYP_CORE",
110       list(method = "known_sigma_z", input_kind = "raw", response_column = "x",
111            known_sigma = sigma, null_mean = null_mean),
112       list("/statistic" = tol(z, "known-sigma Z"),
113            "/p_value" = prob(2 * stats::pnorm(-abs(z)), "two-sided normal tail"),
114            "/estimate" = tol(mean(d$x), "sample mean")),
115       list("Z uses known sigma 2 and n=16; normal two-sided tail."),
116       "R stats::Normal documentation (pnorm and qnorm); closed-form Z statistic.")

Whole script generators/r/hypothesis_secondary.R


Three-level factor, scale 0.5

Case STAT-HYP-BAYES-ANOVA-001 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, pt, contr.helmert, determinant, solve
Citation
Rouder, Morey, Speckman and Province (2012), Default Bayes factors for ANOVA designs, Journal of Mathematical Psychology 56(5), 356-374, eqs. 7-10; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.519740507879017040.5197405078790174.07E-1817passed
/bayes_factors/0/bf01tolerance1.92403706241957131.92403706241957131.32E-1717passed
/bayes_factors/0/log_bf10tolerance-0.65442561526981691-0.65442561526981692.88E-1817passed
/posteriors/0/meantolerance0.709192751317726190.70919275131772729.98E-1614passed
/posteriors/0/sdtolerance0.702868999788100910.70286899978810063.31E-1615passed
/posteriors/0/mediantolerance0.668116515609980530.66811651560998148.91E-1614passed
/posteriors/0/lowertolerance-0.57828476358183456-0.57828476358183591.33E-1514passed
/posteriors/0/uppertolerance2.2001668273997412.20016682739975741.64E-1414passed
/posteriors/1/meantolerance-0.58157145568226587-0.58157145568226667.77E-1614passed
/posteriors/1/sdtolerance0.70647927693896240.70647927693896262.19E-1615passed
/posteriors/1/mediantolerance-0.54354830948573329-0.54354830948573966.33E-1513passed
/posteriors/1/lowertolerance-2.0777456794515672-2.07774567945158942.22E-1413passed
/posteriors/1/uppertolerance0.729029305959557730.72902930595955881.11E-1514passed
/posteriors/2/meantolerance-0.12762129563546043-0.12762129563546031.39E-1614passed
/posteriors/2/sdtolerance0.658404715039778040.65840471503977832.26E-1615passed
/posteriors/2/mediantolerance-0.12013040778427868-0.120130407784278293.90E-1614passed
/posteriors/2/lowertolerance-1.4652511847112863-1.46525118471128611.75E-1615passed
/posteriors/2/uppertolerance1.17515375295738991.17515375295739088.74E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

484    emit("hypothesis", "bayesAnovaMenuItem", "tabelz.analysis.bayesian.linear:compute_bayes_anova",
485         id, title, kind,
486         list(response = "y", factor = factor, prior_scale = r, confidence_level = level),
487         outputs,
488         c(sprintf("%d levels (%s, in code-point order), n = %d; level effects coded by normalized Helmert contrasts Q (any orthonormal basis of the sum-to-zero space gives the same Bayes factor and level effects Q theta).", a, paste(levels, collapse = ", "), n),
489           sprintf("BF10(g) = |I + g X'PX|^(-1/2) ((y'Py - y'PX (X'PX + I/g)^-1 X'Py)/y'Py)^(-(n-1)/2) with P the centring matrix, averaged over g ~ Inverse-Gamma(1/2, r^2/2), r = %g (Rouder, Morey, Speckman and Province 2012, eqs. 7-10), by R integrate on u = log g.", r),
490           "Posterior of each level effect: given g, Student-t on n - 1 df with centre Q (X'X + I/g)^-1 X'y and scale from s^2 Q (X'X + I/g)^-1 Q', s^2 = residual/(n - 1), mixed over the posterior of g by R integrate; quantiles by uniroot on the mixture CDF; SD by the law of total variance.",
491           "Tolerance 1e-6: both sides integrate numerically.",
492           paste0(manual, ":31-32: 'One-way ANOVA & Shared scale for fixed level effects, default 0.5; at most ten levels.'")),
493         "R stats: integrate, uniroot, pt, contr.helmert, determinant, solve",
494         "Rouder, Morey, Speckman and Province (2012), Default Bayes factors for ANOVA designs, Journal of Mathematical Psychology 56(5), 356-374, eqs. 7-10; R Core Team, stats package help: integrate, uniroot.",
495         "REQ-STAT-HYP-BAYES-ANOVA-001")

Whole script generators/r/hypothesis_bayesian.R


Three-level factor, scale 1

Case STAT-HYP-BAYES-ANOVA-002 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, pt, contr.helmert, determinant, solve
Citation
Rouder, Morey, Speckman and Province (2012), Default Bayes factors for ANOVA designs, Journal of Mathematical Psychology 56(5), 356-374, eqs. 7-10; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.273916692510961510.27391669251096152.50E-1817passed
/bayes_factors/0/bf01tolerance3.65074501605988243.65074501605988244.18E-1716passed
/bayes_factors/0/log_bf10tolerance-1.294931260754983-1.2949312607549836.57E-1817passed
/posteriors/0/meantolerance0.943938055291683510.94393805529168415.54E-1615passed
/posteriors/0/sdtolerance0.767645210372692070.76764521037269254.40E-1615passed
/posteriors/0/mediantolerance0.931596694826925640.93159669482695412.84E-1413passed
/posteriors/0/lowertolerance-0.54266377923398423-0.54266377923398212.11E-1514passed
/posteriors/0/uppertolerance2.49505889153968142.49505889153968239.00E-1615passed
/posteriors/1/meantolerance-0.7810869148387144-0.78108691483871551.11E-1514passed
/posteriors/1/sdtolerance0.786705628561587770.78670562856158791.07E-1615passed
/posteriors/1/mediantolerance-0.76919754303930776-0.76919754303933192.41E-1413passed
/posteriors/1/lowertolerance-2.3701548522339544-2.3701548522339554.34E-1615passed
/posteriors/1/uppertolerance0.74677917192533860.74677917192534031.67E-1514passed
/posteriors/2/meantolerance-0.16285114045296889-0.162851140452968751.39E-1615passed
/posteriors/2/sdtolerance0.750521451765776560.75052145176577666.16E-1918passed
/posteriors/2/mediantolerance-0.16086418268917485-0.160864182689174325.27E-1614passed
/posteriors/2/lowertolerance-1.6578006305300588-1.65780063053005834.72E-1615passed
/posteriors/2/uppertolerance1.32217108269246931.3221710826924692.20E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

498  anova_case("STAT-HYP-BAYES-ANOVA-002", "Three-level factor, scale 1", "axis_coverage", "g3", 1, 0.95)

Whole script generators/r/hypothesis_bayesian.R


Two-level factor, 90% interval

Case STAT-HYP-BAYES-ANOVA-003 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, pt, contr.helmert, determinant, solve
Citation
Rouder, Morey, Speckman and Province (2012), Default Bayes factors for ANOVA designs, Journal of Mathematical Psychology 56(5), 356-374, eqs. 7-10; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.760435699753336250.76043569975333611.14E-1615passed
/bayes_factors/0/bf01tolerance1.31503557805659521.31503557805659521.13E-1717passed
/bayes_factors/0/log_bf10tolerance-0.27386372082007909-0.273863720820079155.84E-1715passed
/posteriors/0/meantolerance-0.57715857206188637-0.57715857206188576.70E-1614passed
/posteriors/0/sdtolerance0.543308826908605820.54330882690860667.81E-1614passed
/posteriors/0/mediantolerance-0.55639619102158921-0.55639619102161282.35E-1413passed
/posteriors/0/lowertolerance-1.4985352915013004-1.49853529150131661.62E-1413passed
/posteriors/0/uppertolerance0.273928026459559940.2739280264595581.95E-1514passed
/posteriors/1/meantolerance0.577158572061886370.57715857206188612.26E-1615passed
/posteriors/1/sdtolerance0.543308826908605820.54330882690860656.70E-1614passed
/posteriors/1/mediantolerance0.556396191021589210.55639619102161282.35E-1413passed
/posteriors/1/lowertolerance-0.27392802645955938-0.27392802645955935.38E-1715passed
/posteriors/1/uppertolerance1.49853529150129951.49853529150131661.71E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

499  anova_case("STAT-HYP-BAYES-ANOVA-003", "Two-level factor, 90% interval", "axis_coverage", "g", 0.5, 0.9)

Whole script generators/r/hypothesis_bayesian.R


Beta(1, 1) prior against p = 0.5

Case STAT-HYP-BAYES-BINOMIAL-001 · status: passed

Reference

Evidence type
R
Source
R stats: lbeta, qbeta, pbeta, dbeta, uniroot
Citation
R Core Team, stats package help: Beta (qbeta, pbeta, dbeta), Special (lbeta), uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.297286201930164970.2972862019301652.75E-1817passed
/bayes_factors/0/bf01tolerance3.3637619018554673.3637619018554672.64E-1717passed
/bayes_factors/0/log_bf10tolerance-1.2130599613226458-1.21305996132264581.89E-1716passed
/posteriors/0/meantolerance0.545454545454545410.54545454545454544.17E-1817passed
/posteriors/0/sdtolerance0.10382549831275650.10382549831275653.43E-1816passed
/posteriors/0/mediantolerance0.546856501535443740.54685650153544391.11E-1615passed
/posteriors/0/lowertolerance0.340206309280274680.34020630928027465.86E-1715passed
/posteriors/0/uppertolerance0.742869373593593170.74286937359359256.62E-1615passed
/posteriors/0/hdi_lowertolerance0.343101178692305110.343101178692020842.84E-1312passed
/posteriors/0/hdi_uppertolerance0.745609846926174490.74560984692642772.53E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

122  binomial_case <- function(id, title, kind, event, p0, a, b, level, excluded = list()) {
123    v <- dat$outcome[rows_kept(excluded)]
124    n <- length(v)
125    x <- sum(v == event)
126    A <- a + x
127    B <- b + n - x
128    log_bf <- lbeta(A, B) - lbeta(a, b) - x * log(p0) - (n - x) * log1p(-p0)
129    tail <- (1 - level) / 2
130    hdi <- hdi_closed(function(p) stats::dbeta(p, A, B), function(p) stats::pbeta(p, A, B),
131                      function(q) stats::qbeta(q, A, B), level, stats::qbeta(1e-12, A, B))
132    outputs <- c(
133      list("/n" = count_of(n, "Number of analysed trials.")),
134      bf_outputs(log_bf, "beta-binomial"),
135      posterior_outputs("/posteriors/0", A / (A + B), sqrt(A * B / ((A + B)^2 * (A + B + 1))),
136                        stats::qbeta(0.5, A, B), stats::qbeta(tail, A, B), stats::qbeta(1 - tail, A, B),
137                        hdi, sprintf("Beta(%g, %g)", A, B)))
138    emit("hypothesis", "bayesBinomialMenuItem",
139         "tabelz.analysis.bayesian.conjugate:compute_bayes_binomial", id, title, kind,
140         list(column = "outcome", event_value = event, null_proportion = p0, prior_a = a,
141              prior_b = b, confidence_level = level),
142         outputs,
143         c(sprintf("x = %d events (%s) in n = %d trials; prior Beta(%g, %g) gives posterior Beta(%g, %g).", x, event, n, a, b, A, B),
144           "log BF10 = lbeta(a + x, b + n - x) - lbeta(a, b) - x log p0 - (n - x) log(1 - p0): the beta-binomial marginal likelihood against the point null.",
145           "Equal-tailed interval: qbeta at (1 - level)/2 and 1 - (1 - level)/2; HDI by equal densities at both ends.",
146           paste0(manual, ":25-26: 'Binomial proportion & Beta(a,b), default Beta(1,1); name the event and point-null proportion.'"),
147           paste0(manual, ":62-64: 'Posterior mean is 7/12; SD is sqrt(7*5/(12^2*13)) ... BF10 = B(7,5)/0.5^10 ... equal-tailed limits are the 0.025 and 0.975 quantiles'")),
148         "R stats: lbeta, qbeta, pbeta, dbeta, uniroot",
149         "R Core Team, stats package help: Beta (qbeta, pbeta, dbeta), Special (lbeta), uniroot.",
150         "REQ-STAT-HYP-BAYES-BINOMIAL-001", excluded)

Whole script generators/r/hypothesis_bayesian.R


Informative Beta(2, 5) prior against p = 0.3

Case STAT-HYP-BAYES-BINOMIAL-002 · status: passed

Reference

Evidence type
R
Source
R stats: lbeta, qbeta, pbeta, dbeta, uniroot
Citation
R Core Team, stats package help: Beta (qbeta, pbeta, dbeta), Special (lbeta), uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance3.10386873510721453.1038687351072172.67E-1515passed
/bayes_factors/0/bf01tolerance0.322178573046343010.322178573046342732.79E-1615passed
/bayes_factors/0/log_bf10tolerance1.13264931247935151.13264931247935249.10E-1615passed
/posteriors/0/meantolerance0.481481481481481450.481481481481481454.75E-1817passed
/posteriors/0/sdtolerance0.0944262872887552670.094426287288755273.35E-1917passed
/posteriors/0/mediantolerance0.481017543145880910.481017543145880741.66E-1615passed
/posteriors/0/lowertolerance0.299272179441596450.29927217944159655.86E-1715passed
/posteriors/0/uppertolerance0.66629178152470780.66629178152470771.06E-1615passed
/posteriors/0/hdi_lowertolerance0.29834961429604540.29834961429582982.16E-1312passed
/posteriors/0/hdi_uppertolerance0.665350770029124350.66535077002935132.27E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

154  binomial_case("STAT-HYP-BAYES-BINOMIAL-002", "Informative Beta(2, 5) prior against p = 0.3",
155                "axis_coverage", "pass", 0.3, 2, 5, 0.95)

Whole script generators/r/hypothesis_bayesian.R


90% credible interval, event 'fail'

Case STAT-HYP-BAYES-BINOMIAL-003 · status: passed

Reference

Evidence type
R
Source
R stats: lbeta, qbeta, pbeta, dbeta, uniroot
Citation
R Core Team, stats package help: Beta (qbeta, pbeta, dbeta), Special (lbeta), uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.297286201930165190.29728620193016524.80E-1816passed
/bayes_factors/0/bf01tolerance3.36376190185546393.3637619018554643.50E-1716passed
/bayes_factors/0/log_bf10tolerance-1.2130599613226449-1.2130599613226453.07E-1716passed
/posteriors/0/meantolerance0.454545454545454530.454545454545454533.15E-1918passed
/posteriors/0/sdtolerance0.10382549831275650.10382549831275653.43E-1816passed
/posteriors/0/mediantolerance0.453143498464556260.453143498464556262.52E-1918passed
/posteriors/0/lowertolerance0.285800478087175790.28580047808717623.90E-1614passed
/posteriors/0/uppertolerance0.628098862918754830.62809886291875491.11E-1615passed
/posteriors/0/hdi_lowertolerance0.282956103025027880.282956103024932249.56E-1412passed
/posteriors/0/hdi_uppertolerance0.625121657361201070.62512165736129799.68E-1412passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

156  binomial_case("STAT-HYP-BAYES-BINOMIAL-003", "90% credible interval, event 'fail'",
157                "axis_coverage", "fail", 0.5, 1, 1, 0.9)

Whole script generators/r/hypothesis_bayesian.R


Three excluded rows

Case STAT-HYP-BAYES-BINOMIAL-004 · status: passed

Reference

Evidence type
R
Source
R stats: lbeta, qbeta, pbeta, dbeta, uniroot
Citation
R Core Team, stats package help: Beta (qbeta, pbeta, dbeta), Special (lbeta), uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact17170.00E+2passed
/bayes_factors/0/bf10tolerance0.374422962658258030.37442296265825671.33E-1514passed
/bayes_factors/0/bf01tolerance2.67077636718749072.67077636718750049.74E-1514passed
/bayes_factors/0/log_bf10tolerance-0.98236920437305031-0.98236920437305393.55E-1514passed
/posteriors/0/meantolerance0.578947368421052650.57894736842105274.95E-1817passed
/posteriors/0/sdtolerance0.110400931386331740.110400931386331741.81E-1816passed
/posteriors/0/mediantolerance0.58177374444000440.58177374444000443.34E-1817passed
/posteriors/0/lowertolerance0.357451205893062560.35745120589306271.13E-1615passed
/posteriors/0/uppertolerance0.784698492612700190.78469849261270011.07E-1615passed
/posteriors/0/hdi_lowertolerance0.363425716316358840.36342571631612192.37E-1312passed
/posteriors/0/hdi_uppertolerance0.790078578219072350.79007857821926281.90E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

158  binomial_case("STAT-HYP-BAYES-BINOMIAL-004", "Three excluded rows", "missing_excluded",
159                "pass", 0.5, 1, 1, 0.95, excluded = list(0L, 7L, 12L))

Whole script generators/r/hypothesis_bayesian.R


Uniform prior (width 1) on x1 and y

Case STAT-HYP-BAYES-CORRELATION-001 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, dbeta, cor
Citation
Ly, Verhagen and Wagenmakers (2016), Harold Jeffreys's default Bayes factor hypothesis tests: explanation, extension, and application in psychology, Journal of Mathematical Psychology 72, 19-32, eqs. 7 and 9; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance251.74340431116534251.743404311164231.11E-1214passed
/bayes_factors/0/bf01tolerance0.00397229870922043410.0039722987092204521.82E-1714passed
/bayes_factors/0/log_bf10tolerance5.52841033189564265.5284103318956384.47E-1515passed
/posteriors/0/meantolerance0.68852797484619710.68852797484619692.23E-1615passed
/posteriors/0/sdtolerance0.121320508303337830.121320508303338516.85E-1614passed
/posteriors/0/mediantolerance0.707239861963051330.70723986196305112.21E-1615passed
/posteriors/0/lowertolerance0.40074136167785390.400741361677853573.31E-1615passed
/posteriors/0/uppertolerance0.870086336629167790.87008633662916788.30E-1918passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

375    log_prior <- function(rho) stats::dbeta((rho + 1) / 2, alpha, alpha, log = TRUE) - log(2)
376    log_l0 <- log_lik(0)
377    bf <- stats::integrate(function(rho) vapply(rho, function(v) exp(log_prior(v) + log_lik(v) - log_l0), 0),
378                           -1, 1, rel.tol = 1e-11, abs.tol = 0, subdivisions = 2000L)$value
379    log_bf <- log(bf)
380    s <- numeric_posterior(function(rho) log_prior(rho) + log_lik(rho), -1, 1, level)
381    outputs <- c(
382      list("/n" = count_of(n, "Number of complete pairs.")),
383      bf_outputs(log_bf, "correlation", kind = quad),
384      numeric_outputs("/posteriors/0", s, "posterior of rho"))
385    emit("hypothesis", "bayesCorrelationMenuItem",
386         "tabelz.analysis.bayesian.jzs:compute_bayes_correlation", id, title, kind,
387         list(first = first, second = second, prior_width = width, confidence_level = level),
388         outputs,
389         c(sprintf("r = %s from n = %d pairs.", format(r, digits = 17), n),
390           "Likelihood of rho (Ly, Verhagen and Wagenmakers 2016, eq. 7): L(rho) proportional to (1 - rho^2)^((n-1)/2) integral_0^inf (cosh w - rho r)^-(n-1) dw, by R integrate.",
391           sprintf("Prior: stretched Beta(1/width, 1/width) on (-1, 1), width %g, density dbeta((rho + 1)/2, 1/width, 1/width)/2; BF10 = integral prior(rho) L(rho)/L(0) drho by R integrate.", width),
392           "Posterior moments and quantiles of rho by R integrate and uniroot on the normalized prior times likelihood.",
393           "Tolerance 1e-6: both sides integrate numerically (nested R integrate rel.tol 1e-11/1e-12; HypoNum adaptive quadrature with estimated error below 1e-6).",
394           paste0(manual, ":23-24: 'Pearson correlation & Stretched beta prior on correlation in (-1,1); default width 1 is uniform.'")),
395         "R stats: integrate, uniroot, dbeta, cor",
396         "Ly, Verhagen and Wagenmakers (2016), Harold Jeffreys's default Bayes factor hypothesis tests: explanation, extension, and application in psychology, Journal of Mathematical Psychology 72, 19-32, eqs. 7 and 9; R Core Team, stats package help: integrate, uniroot.",
397         "REQ-STAT-HYP-BAYES-CORRELATION-001")

Whole script generators/r/hypothesis_bayesian.R


Stretched Beta(2, 2) prior (width 0.5)

Case STAT-HYP-BAYES-CORRELATION-002 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, dbeta, cor
Citation
Ly, Verhagen and Wagenmakers (2016), Harold Jeffreys's default Bayes factor hypothesis tests: explanation, extension, and application in psychology, Journal of Mathematical Psychology 72, 19-32, eqs. 7 and 9; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance193.04083084157321193.040830841572536.81E-1314passed
/bayes_factors/0/bf01tolerance0.00518025122270992810.0051802512227099461.82E-1714passed
/bayes_factors/0/log_bf10tolerance5.26290172529411395.262901725294113.52E-1515passed
/posteriors/0/meantolerance0.652219799857066350.6522197998570663.32E-1615passed
/posteriors/0/sdtolerance0.128623104493866240.12862310449386696.68E-1614passed
/posteriors/0/mediantolerance0.670820109023817610.67082010902381715.53E-1615passed
/posteriors/0/lowertolerance0.350621582483061190.3506215824830611.67E-1615passed
/posteriors/0/uppertolerance0.848650350291641040.84865035029164082.20E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

401  correlation_case("STAT-HYP-BAYES-CORRELATION-002", "Stretched Beta(2, 2) prior (width 0.5)",
402                   "axis_coverage", "x1", "y", 0.5, 0.95)

Whole script generators/r/hypothesis_bayesian.R


Weak correlation x2 and d, 90% interval

Case STAT-HYP-BAYES-CORRELATION-003 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, dbeta, cor
Citation
Ly, Verhagen and Wagenmakers (2016), Harold Jeffreys's default Bayes factor hypothesis tests: explanation, extension, and application in psychology, Journal of Mathematical Psychology 72, 19-32, eqs. 7 and 9; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.342290519889434930.34229051988943481.11E-1615passed
/bayes_factors/0/bf01tolerance2.9214948761158072.92149487611580841.38E-1515passed
/bayes_factors/0/log_bf10tolerance-1.0720954291569085-1.0720954291569094.19E-1615passed
/posteriors/0/meantolerance-0.13496237220944055-0.134962372209440468.79E-1715passed
/posteriors/0/sdtolerance0.209873026172502560.209873026172502592.59E-1715passed
/posteriors/0/mediantolerance-0.14129037598631206-0.141290375986297821.42E-1412passed
/posteriors/0/lowertolerance-0.46984475651198532-0.46984475651198521.13E-1615passed
/posteriors/0/uppertolerance0.221670142865304230.221670142865304765.29E-1614passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

403  correlation_case("STAT-HYP-BAYES-CORRELATION-003", "Weak correlation x2 and d, 90% interval",
404                   "axis_coverage", "x2", "d", 1, 0.9)

Whole script generators/r/hypothesis_bayesian.R


Normal(0, 1) prior, known SD 1, null 0

Case STAT-HYP-BAYES-NORMAL-001 · status: passed

Reference

Evidence type
R
Source
R stats: dnorm, qnorm
Citation
R Core Team, stats package help: Normal (dnorm, qnorm).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance22.004005289999722.0040052899996821.77E-1415passed
/bayes_factors/0/bf01tolerance0.0454462715683165290.045446271568316574.18E-1715passed
/bayes_factors/0/log_bf10tolerance3.09122449542400363.09122449542400279.17E-1615passed
/posteriors/0/meantolerance0.662857142857142920.66285714285714292.49E-1817passed
/posteriors/0/sdtolerance0.218217890235992360.218217890235992362.24E-1816passed
/posteriors/0/mediantolerance0.662857142857142920.66285714285714292.49E-1817passed
/posteriors/0/lowertolerance0.235157937212283290.235157937212283171.15E-1615passed
/posteriors/0/uppertolerance1.09055634850200271.09055634850200272.96E-1716passed
/posteriors/0/hdi_lowertolerance0.235157937212283290.235157937211969543.14E-1311passed
/posteriors/0/hdi_uppertolerance1.09055634850200271.09055634850231623.13E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

218  normal_case <- function(id, title, kind, sigma, m0, tau, mu0, level) {
219    v <- dat$d
220    n <- length(v)
221    ybar <- mean(v)
222    precision <- 1 / tau^2 + n / sigma^2
223    post_mean <- (m0 / tau^2 + n * ybar / sigma^2) / precision
224    post_sd <- sqrt(1 / precision)
225    log_bf <- stats::dnorm(ybar, m0, sqrt(tau^2 + sigma^2 / n), log = TRUE) -
226      stats::dnorm(ybar, mu0, sigma / sqrt(n), log = TRUE)
227    tail <- (1 - level) / 2
228    lower <- stats::qnorm(tail, post_mean, post_sd)
229    upper <- stats::qnorm(1 - tail, post_mean, post_sd)
230    outputs <- c(
231      list("/n" = count_of(n, "Number of observations.")),
232      bf_outputs(log_bf, "normal-normal"),
233      posterior_outputs("/posteriors/0", post_mean, post_sd, post_mean, lower, upper,
234                        c(lower, upper), "Normal posterior"))
235    emit("hypothesis", "bayesNormalMenuItem",
236         "tabelz.analysis.bayesian.conjugate:compute_bayes_normal", id, title, kind,
237         list(column = "d", known_sd = sigma, prior_mean = m0, prior_sd = tau, null_value = mu0,
238              confidence_level = level),
239         outputs,
240         c(sprintf("ybar = %s from n = %d; posterior precision 1/tau^2 + n/sigma^2, mean (m0/tau^2 + n ybar/sigma^2)/precision.", format(ybar, digits = 17), n),
241           "BF10 = N(ybar; m0, tau^2 + sigma^2/n) / N(ybar; mu0, sigma^2/n) (prior predictive against the point null).",
242           "The posterior is symmetric: its HDI equals the equal-tailed interval from qnorm.",
243           paste0(manual, ":29-30: 'Normal mean, known SD & Normal prior on the mean; supply the known observation SD, prior mean and prior SD.'")),
244         "R stats: dnorm, qnorm",
245         "R Core Team, stats package help: Normal (dnorm, qnorm).",
246         "REQ-STAT-HYP-BAYES-NORMAL-001")

Whole script generators/r/hypothesis_bayesian.R


Informative Normal(1, 0.5) prior, known SD 1.2, null 0.5

Case STAT-HYP-BAYES-NORMAL-002 · status: passed

Reference

Evidence type
R
Source
R stats: dnorm, qnorm
Citation
R Core Team, stats package help: Normal (dnorm, qnorm).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.534903971309437790.53490397130943791.08E-1615passed
/bayes_factors/0/bf01tolerance1.86949443944492181.86949443944492162.48E-1615passed
/bayes_factors/0/log_bf10tolerance-0.62566804107646079-0.62566804107646062.22E-1615passed
/posteriors/0/meantolerance0.76397515527950310.76397515527950314.21E-1817passed
/posteriors/0/sdtolerance0.236433121871730180.236433121871730184.13E-1816passed
/posteriors/0/mediantolerance0.76397515527950310.76397515527950314.21E-1817passed
/posteriors/0/lowertolerance0.300574751658542690.300574751658542645.18E-1715passed
/posteriors/0/uppertolerance1.22737555890046361.22737555890046362.98E-1716passed
/posteriors/0/hdi_lowertolerance0.300574751658542690.30057475165820283.40E-1311passed
/posteriors/0/hdi_uppertolerance1.22737555890046361.22737555890080353.40E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

250  normal_case("STAT-HYP-BAYES-NORMAL-002", "Informative Normal(1, 0.5) prior, known SD 1.2, null 0.5",
251              "axis_coverage", 1.2, 1, 0.5, 0.5, 0.95)

Whole script generators/r/hypothesis_bayesian.R


90% credible interval

Case STAT-HYP-BAYES-NORMAL-003 · status: passed

Reference

Evidence type
R
Source
R stats: dnorm, qnorm
Citation
R Core Team, stats package help: Normal (dnorm, qnorm).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance22.004005289999722.0040052899996821.77E-1415passed
/bayes_factors/0/bf01tolerance0.0454462715683165290.045446271568316574.18E-1715passed
/bayes_factors/0/log_bf10tolerance3.09122449542400363.09122449542400279.17E-1615passed
/posteriors/0/meantolerance0.662857142857142920.66285714285714292.49E-1817passed
/posteriors/0/sdtolerance0.218217890235992360.218217890235992362.24E-1816passed
/posteriors/0/mediantolerance0.662857142857142920.66285714285714292.49E-1817passed
/posteriors/0/lowertolerance0.303920654636772540.30392065463677255.96E-1715passed
/posteriors/0/uppertolerance1.02179363107751311.0217936310775131.30E-1716passed
/posteriors/0/hdi_lowertolerance0.303920654636772540.303920654636550052.22E-1312passed
/posteriors/0/hdi_uppertolerance1.02179363107751311.02179363107773582.23E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

252  normal_case("STAT-HYP-BAYES-NORMAL-003", "90% credible interval", "axis_coverage",
253              1, 0, 1, 0, 0.9)

Whole script generators/r/hypothesis_bayesian.R


Gamma(1, 1) prior, unit exposure, rate 1

Case STAT-HYP-BAYES-POISSON-001 · status: passed

Reference

Evidence type
R
Source
R stats: lgamma, qgamma, pgamma, dgamma, uniroot
Citation
R Core Team, stats package help: GammaDist (qgamma, pgamma, dgamma), Special (lgamma), uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance97960.94911164870497960.94911164872.75E-1317passed
/bayes_factors/0/bf01tolerance1.0208149360213663e-051.0208149360213663e-052.67E-2216passed
/bayes_factors/0/log_bf10tolerance11.49232419978650911.4923241997865099.80E-1717passed
/posteriors/0/meantolerance2.33333333333333352.3333333333333334.63E-1615passed
/posteriors/0/sdtolerance0.333333333333333310.33333333333333334.83E-1816passed
/posteriors/0/mediantolerance2.31747965712986482.31747965712986444.32E-1615passed
/posteriors/0/lowertolerance1.726212845182341.726212845182342.78E-1716passed
/posteriors/0/uppertolerance3.03052553248225023.03052553248225074.50E-1615passed
/posteriors/0/hdi_lowertolerance1.69765952392196891.69765952392193793.10E-1413passed
/posteriors/0/hdi_uppertolerance2.99603427895694722.99603427895698944.22E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

181    tail <- (1 - level) / 2
182    hdi <- hdi_closed(function(v) stats::dgamma(v, shape, rate = rate),
183                      function(v) stats::pgamma(v, shape, rate = rate),
184                      function(q) stats::qgamma(q, shape, rate = rate), level,
185                      stats::qgamma(1e-12, shape, rate = rate))
186    outputs <- c(
187      list("/n" = count_of(length(rows), "Number of count rows.")), bf,
188      posterior_outputs("/posteriors/0", shape / rate, sqrt(shape) / rate,
189                        stats::qgamma(0.5, shape, rate = rate), stats::qgamma(tail, shape, rate = rate),
190                        stats::qgamma(1 - tail, shape, rate = rate), hdi,
191                        sprintf("Gamma(%g, rate %g)", shape, rate)))
192    configuration <- list(column = "count", null_rate = lambda0, prior = prior, prior_shape = a,
193                          prior_rate = b, confidence_level = level)
194    if (with_exposure) configuration$exposure <- "exposure"
195    emit("hypothesis", "bayesPoissonMenuItem",
196         "tabelz.analysis.bayesian.conjugate:compute_bayes_poisson", id, title, kind,
197         configuration, outputs,
198         c(sprintf("X = %d events over total exposure T = %s (%s).", X, format(Tt, digits = 17),
199                   if (with_exposure) "sum of the exposure column" else "one unit per row"),
200           prior_note,
201           "Equal-tailed interval: qgamma at (1 - level)/2 and 1 - (1 - level)/2; HDI by equal densities at both ends.",
202           paste0(manual, ":27-28: 'Poisson rate & Gamma shape and rate, both positive, per unit of exposure. The Jeffreys option supplies a posterior only.'"),
203           paste0(manual, ":86-88: 'An improper prior's arbitrary normalization prevents a point-null Bayes factor from being defined.'")),
204         "R stats: lgamma, qgamma, pgamma, dgamma, uniroot",
205         "R Core Team, stats package help: GammaDist (qgamma, pgamma, dgamma), Special (lgamma), uniroot.",
206         "REQ-STAT-HYP-BAYES-POISSON-001")

Whole script generators/r/hypothesis_bayesian.R


Gamma(2, 0.5) prior with an exposure column

Case STAT-HYP-BAYES-POISSON-002 · status: passed

Reference

Evidence type
R
Source
R stats: lgamma, qgamma, pgamma, dgamma, uniroot
Citation
R Core Team, stats package help: GammaDist (qgamma, pgamma, dgamma), Special (lgamma), uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.423016169502700050.423016169502697053.00E-1514passed
/bayes_factors/0/bf01tolerance2.36397582904597982.36397582904599671.69E-1414passed
/bayes_factors/0/log_bf10tolerance-0.86034487489174083-0.86034487489174797.10E-1514passed
/posteriors/0/meantolerance1.93648334624322231.93648334624322251.92E-1616passed
/posteriors/0/sdtolerance0.273860101156679910.27386010115667993.71E-1816passed
/posteriors/0/mediantolerance1.92358886979063671.92358886979063652.20E-1615passed
/posteriors/0/lowertolerance1.43729526481262071.4372952648126212.32E-1615passed
/posteriors/0/uppertolerance2.50893100669706782.50893100669706834.56E-1615passed
/posteriors/0/hdi_lowertolerance1.41404413525688551.414044135256662.26E-1312passed
/posteriors/0/hdi_uppertolerance2.4808991483506262.48089914835095373.28E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

210  poisson_case("STAT-HYP-BAYES-POISSON-002", "Gamma(2, 0.5) prior with an exposure column",
211               "axis_coverage", TRUE, "gamma", 2, 0.5, 1.5, 0.95)

Whole script generators/r/hypothesis_bayesian.R


Jeffreys prior with an exposure column (posterior only)

Case STAT-HYP-BAYES-POISSON-003 · status: passed

Reference

Evidence type
R
Source
R stats: lgamma, qgamma, pgamma, dgamma, uniroot
Citation
R Core Team, stats package help: GammaDist (qgamma, pgamma, dgamma), Special (lgamma), uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10absentNonepassed
/bayes_factors/0/bf01absentNonepassed
/bayes_factors/0/log_bf10absentNonepassed
/posteriors/0/meantolerance1.9154818325434441.91548183254344443.96E-1615passed
/posteriors/0/sdtolerance0.275047161871724290.27504716187172441.09E-1615passed
/posteriors/0/mediantolerance1.9023332155631331.90233321556313343.95E-1615passed
/posteriors/0/lowertolerance1.41472188517534741.4147218851753486.26E-1615passed
/posteriors/0/uppertolerance2.4909446571898892.49094465718988944.44E-1615passed
/posteriors/0/hdi_lowertolerance1.39105487325601131.39105487325598332.80E-1413passed
/posteriors/0/hdi_uppertolerance2.46232739820786952.46232739820790954.00E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

212  poisson_case("STAT-HYP-BAYES-POISSON-003", "Jeffreys prior with an exposure column (posterior only)",
213               "axis_coverage", TRUE, "jeffreys", 1, 1, 1, 0.95)

Whole script generators/r/hypothesis_bayesian.R


99% credible interval

Case STAT-HYP-BAYES-POISSON-004 · status: passed

Reference

Evidence type
R
Source
R stats: lgamma, qgamma, pgamma, dgamma, uniroot
Citation
R Core Team, stats package help: GammaDist (qgamma, pgamma, dgamma), Special (lgamma), uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.168850686386725010.1688506863867252.96E-1816passed
/bayes_factors/0/bf01tolerance5.92239227094204875.9223922709420492.46E-1717passed
/bayes_factors/0/log_bf10tolerance-1.7787404670908629-1.77874046709086293.54E-1716passed
/posteriors/0/meantolerance2.33333333333333352.3333333333333334.63E-1615passed
/posteriors/0/sdtolerance0.333333333333333310.33333333333333334.83E-1816passed
/posteriors/0/mediantolerance2.31747965712986482.31747965712986444.32E-1615passed
/posteriors/0/lowertolerance1.56413263084747011.56413263084747011.79E-1716passed
/posteriors/0/uppertolerance3.28102740228364323.28102740228364364.46E-1615passed
/posteriors/0/hdi_lowertolerance1.53667469962231281.53667469962144538.67E-1312passed
/posteriors/0/hdi_uppertolerance3.24579024904355283.2457902490449871.43E-1212passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

214  poisson_case("STAT-HYP-BAYES-POISSON-004", "99% credible interval", "axis_coverage",
215               FALSE, "gamma", 1, 1, 2, 0.99)

Whole script generators/r/hypothesis_bayesian.R


One-sample JZS t, default scale

Case STAT-HYP-BAYES-TTEST-001 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, dnorm, dchisq, dcauchy
Citation
Rouder, Speckman, Sun, Morey and Iverson (2009), Bayesian t tests for accepting and rejecting the null hypothesis, Psychonomic Bulletin & Review 16(2), 225-237, eq. 1; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance14.74789352191707414.7478935219171012.70E-1414passed
/bayes_factors/0/bf01tolerance0.0678062937268894970.067806293726889371.25E-1614passed
/bayes_factors/0/log_bf10tolerance2.6911002605137952.6911002605137971.80E-1515passed
/posteriors/0/meantolerance0.689996461376162550.68999646137616282.23E-1615passed
/posteriors/0/sdtolerance0.251662728924378850.25166272892437862.74E-1614passed
/posteriors/0/mediantolerance0.687000762673327880.68700076267329583.21E-1413passed
/posteriors/0/lowertolerance0.206051669483269060.206051669483268978.68E-1715passed
/posteriors/0/uppertolerance1.19103389812652741.19103389812652851.09E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

327    stopifnot(abs(nct_density(t, nu, t) - stats::dt(t, nu, ncp = t)) < 1e-9)
328    s <- numeric_posterior(log_post, -limit, limit, level)
329    outputs <- c(
330      list("/n" = count_of(n, "Number of analysed observations.")),
331      bf_outputs(log_bf, "JZS", kind = quad),
332      numeric_outputs("/posteriors/0", s, "posterior of delta"))
333    configuration <- list(design = design, column = column, null_value = null_value,
334                          prior_scale = r, confidence_level = level)
335    if (!is.null(second)) configuration$second <- second
336    if (!is.null(group)) configuration$group <- group
337    if (!is.null(first_group)) configuration$first_group <- first_group
338    emit("hypothesis", "bayesTTestMenuItem", "tabelz.analysis.bayesian.jzs:compute_bayes_ttest",
339         id, title, kind, configuration, outputs,
340         c(definition,
341           sprintf("BF10 = integral over g of (1 + N g r^2)^(-1/2) (1 + t^2/((1 + N g r^2) nu))^(-(nu+1)/2) (2 pi)^(-1/2) g^(-3/2) exp(-1/(2g)) dg divided by (1 + t^2/nu)^(-(nu+1)/2) (Rouder et al. 2009, eq. 1), r = %s, by R integrate on u = log g.", format(r, digits = 17)),
342           "Posterior of delta proportional to the noncentral t density of t with ncp delta sqrt(N) (written as an integral over the chi-square denominator) times the Cauchy(0, r) prior; moments and quantiles by R integrate and uniroot.",
343           "Tolerance 1e-6: both sides integrate numerically (R integrate rel.tol 1e-12; HypoNum adaptive quadrature with estimated error below 1e-6).",
344           paste0(manual, ":21-22: 't tests & Cauchy prior on standardized effect, default scale 1/sqrt(2); declare the null mean or paired difference and the comparison direction.'"),
345           paste0(manual, ":38-40: 'For independent-group effects, verify which group is subtracted from which.'")),
346         "R stats: integrate, uniroot, dnorm, dchisq, dcauchy",
347         "Rouder, Speckman, Sun, Morey and Iverson (2009), Bayesian t tests for accepting and rejecting the null hypothesis, Psychonomic Bulletin & Review 16(2), 225-237, eq. 1; R Core Team, stats package help: integrate, uniroot.",
348         "REQ-STAT-HYP-BAYES-TTEST-001")

Whole script generators/r/hypothesis_bayesian.R


Paired JZS t

Case STAT-HYP-BAYES-TTEST-002 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, dnorm, dchisq, dcauchy
Citation
Rouder, Speckman, Sun, Morey and Iverson (2009), Bayesian t tests for accepting and rejecting the null hypothesis, Psychonomic Bulletin & Review 16(2), 225-237, eq. 1; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance10.49289682367320610.4928968236732241.78E-1414passed
/bayes_factors/0/bf01tolerance0.0953025667558154850.095302566755815321.67E-1614passed
/bayes_factors/0/log_bf10tolerance2.35069853525534532.3506985352553471.81E-1515passed
/posteriors/0/meantolerance0.652747731026565470.65274773102656572.22E-1615passed
/posteriors/0/sdtolerance0.247892046691643910.24789204669164381.07E-1615passed
/posteriors/0/mediantolerance0.649746301999889920.649746301999891.10E-1615passed
/posteriors/0/lowertolerance0.176138346846253650.17613834684625351.37E-1615passed
/posteriors/0/uppertolerance1.14640176543127061.146401765431276.74E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

352  ttest_case("STAT-HYP-BAYES-TTEST-002", "Paired JZS t", "axis_coverage", "paired", "d",
353             second = "d2")

Whole script generators/r/hypothesis_bayesian.R


Independent-samples JZS t, code-point group order

Case STAT-HYP-BAYES-TTEST-003 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, dnorm, dchisq, dcauchy
Citation
Rouder, Speckman, Sun, Morey and Iverson (2009), Bayesian t tests for accepting and rejecting the null hypothesis, Psychonomic Bulletin & Review 16(2), 225-237, eq. 1; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.760435699753335360.76043569975333711.78E-1514passed
/bayes_factors/0/bf01tolerance1.31503557805659681.31503557805659373.14E-1514passed
/bayes_factors/0/log_bf10tolerance-0.27386372082008026-0.27386372082007792.33E-1514passed
/posteriors/0/meantolerance-0.43918011588724809-0.43918011588724887.22E-1614passed
/posteriors/0/sdtolerance0.405638067581540530.405638067581540361.68E-1615passed
/posteriors/0/mediantolerance-0.42091394658236875-0.42091394658236957.75E-1614passed
/posteriors/0/lowertolerance-1.2812368009452408-1.28123680094524179.27E-1615passed
/posteriors/0/uppertolerance0.310863284980642470.31086328498063834.16E-1513passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

354  ttest_case("STAT-HYP-BAYES-TTEST-003", "Independent-samples JZS t, code-point group order",
355             "axis_coverage", "independent", "y", group = "g")

Whole script generators/r/hypothesis_bayesian.R


Independent-samples JZS t, group B first

Case STAT-HYP-BAYES-TTEST-004 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, dnorm, dchisq, dcauchy
Citation
Rouder, Speckman, Sun, Morey and Iverson (2009), Bayesian t tests for accepting and rejecting the null hypothesis, Psychonomic Bulletin & Review 16(2), 225-237, eq. 1; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.760435699753335360.76043569975333711.78E-1514passed
/bayes_factors/0/bf01tolerance1.31503557805659681.31503557805659373.14E-1514passed
/bayes_factors/0/log_bf10tolerance-0.27386372082008026-0.27386372082007792.33E-1514passed
/posteriors/0/meantolerance0.439180115887248090.43918011588724887.22E-1614passed
/posteriors/0/sdtolerance0.405638067581540530.405638067581540361.68E-1615passed
/posteriors/0/mediantolerance0.420913946582368690.420913946582369366.68E-1614passed
/posteriors/0/lowertolerance-0.31086328498064203-0.31086328498064031.72E-1514passed
/posteriors/0/uppertolerance1.28123680094524021.28123680094524151.31E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

356  ttest_case("STAT-HYP-BAYES-TTEST-004", "Independent-samples JZS t, group B first",
357             "axis_coverage", "independent", "y", group = "g", first_group = "B")

Whole script generators/r/hypothesis_bayesian.R


One-sample JZS t, scale 1, null 0.3, 90% interval

Case STAT-HYP-BAYES-TTEST-005 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, dnorm, dchisq, dcauchy
Citation
Rouder, Speckman, Sun, Morey and Iverson (2009), Bayesian t tests for accepting and rejecting the null hypothesis, Psychonomic Bulletin & Review 16(2), 225-237, eq. 1; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/0/bf10tolerance0.903740064082585560.90374006408258582.21E-1615passed
/bayes_factors/0/bf01tolerance1.10651285667536591.10651285667536572.03E-1615passed
/bayes_factors/0/log_bf10tolerance-0.10121349966908855-0.101213499669088322.25E-1614passed
/posteriors/0/meantolerance0.406336614594509650.406336614594509761.10E-1615passed
/posteriors/0/sdtolerance0.226962193714500630.226962193714500382.48E-1614passed
/posteriors/0/mediantolerance0.405109223978896760.40510922397887951.73E-1413passed
/posteriors/0/lowertolerance0.0351663117670635430.0351663117670629545.89E-1613passed
/posteriors/0/uppertolerance0.781719944659826860.7817199446598281.11E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

358  ttest_case("STAT-HYP-BAYES-TTEST-005", "One-sample JZS t, scale 1, null 0.3, 90% interval",
359             "axis_coverage", "one_sample", "d", null_value = 0.3, r = 1, level = 0.9)

Whole script generators/r/hypothesis_bayesian.R


Bootstrap estimate mean

Case STAT-HYP-BOOTSTRAP-ESTIMATE-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 base and stats
Citation
R Core Team, base and stats package help: mean, median, sd, quantile, cor.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bootstrap_estimate.R · sha256 8eb6eded9e3eb420
Dataset
reference_data/datasets/HYP_BOOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_usedexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/unitsexact19190.00E+2passed
/components/0/estimatetolerance12.512.50.00E+1passed
/components/0/attemptedexact2002000.00E+2passed
/components/0/successfulexact2002000.00E+2passed
/components/0/failedexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

44  emit <- function(id, title, kind, config, outputs, notes, extra_deferred = list()) {
45    write_expected("hypothesis", id, list(
46      schema_version = "2.0.0", case_id = id, outputs = outputs, expected_error = NULL,
47      deferred_outputs = c(deferred(), extra_deferred), generator_run = run))
48    write_case("hypothesis", id, list(
49      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "hypothesis",
50      entry_point = entry, menu_items = list(menu), title = title, case_kind = kind,
51      dataset = list(path = "reference_data/datasets/HYP_BOOT.json", sha256 = ""),
52      excluded_rows = list(), configuration = c(config, list(resamples = resamples, seed = 20260928L,
53                                                             confidence_level = 0.95, intervals = "percentile", scheme = "rows")),
54      "repeat" = 1L,
55      expected = list(path = paste0("reference_data/expected/hypothesis/", id, ".json"), sha256 = ""),
56      oracle = list(oracle_class = "independent_implementation", source = "R 4.6.1 base and stats",
57                    citation = "R Core Team, base and stats package help: mean, median, sd, quantile, cor.",
58                    generator = list(path = "generators/r/hypothesis_bootstrap_estimate.R", sha256 = ""),
59                    independent = TRUE,
60                    convention_notes = c(notes, notes_common, "Expected computed from the exported CSV.")),
61      supported = TRUE, unsupported_reason = NULL,
62      requirement_ids = list("REQ-STAT-HYP-BOOTSTRAP-ESTIMATE-001"),
63      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/hypothesis_bootstrap_estimate.R


Bootstrap estimate median

Case STAT-HYP-BOOTSTRAP-ESTIMATE-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 base and stats
Citation
R Core Team, base and stats package help: mean, median, sd, quantile, cor.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bootstrap_estimate.R · sha256 8eb6eded9e3eb420
Dataset
reference_data/datasets/HYP_BOOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_usedexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/unitsexact19190.00E+2passed
/components/0/estimatetolerance12.412.43.55E-1616passed
/components/0/attemptedexact2002000.00E+2passed
/components/0/successfulexact2002000.00E+2passed
/components/0/failedexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

87  emit("STAT-HYP-BOOTSTRAP-ESTIMATE-002", "Bootstrap estimate median", "axis_coverage",
88       list(statistic = "median", column = "v"),
89       counts(length(v), stats::median(v), "median of the 19 finite values"),
90       "Median of an odd count of values is the middle order statistic.")

Whole script generators/r/hypothesis_bootstrap_estimate.R


Bootstrap estimate standard deviation

Case STAT-HYP-BOOTSTRAP-ESTIMATE-003 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 base and stats
Citation
R Core Team, base and stats package help: mean, median, sd, quantile, cor.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bootstrap_estimate.R · sha256 8eb6eded9e3eb420
Dataset
reference_data/datasets/HYP_BOOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_usedexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/unitsexact19190.00E+2passed
/components/0/estimatetolerance2.17587785604901152.17587785604901151.61E-1717passed
/components/0/attemptedexact2002000.00E+2passed
/components/0/successfulexact2002000.00E+2passed
/components/0/failedexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

92  emit("STAT-HYP-BOOTSTRAP-ESTIMATE-003", "Bootstrap estimate standard deviation", "axis_coverage",
93       list(statistic = "sd", column = "v"),
94       counts(length(v), stats::sd(v), "sample SD (n - 1) of the 19 finite values"),
95       "Standard deviation with the n - 1 divisor.")

Whole script generators/r/hypothesis_bootstrap_estimate.R


Bootstrap estimate 10 percent trimmed mean

Case STAT-HYP-BOOTSTRAP-ESTIMATE-004 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 base and stats
Citation
R Core Team, base and stats package help: mean, median, sd, quantile, cor.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bootstrap_estimate.R · sha256 8eb6eded9e3eb420
Dataset
reference_data/datasets/HYP_BOOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_usedexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/unitsexact19190.00E+2passed
/components/0/estimatetolerance12.49411764705882412.4941176470588251.33E-1515passed
/components/0/attemptedexact2002000.00E+2passed
/components/0/successfulexact2002000.00E+2passed
/components/0/failedexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 97  emit("STAT-HYP-BOOTSTRAP-ESTIMATE-004", "Bootstrap estimate 10 percent trimmed mean", "axis_coverage",
 98       list(statistic = "trimmed_mean", column = "v", trim = 0.1),
 99       counts(length(v), mean(v, trim = 0.1), "mean(trim = 0.1): floor(0.1 n) values cut from each end"),
100       "Trimmed mean: R's mean(trim = 0.1) and SciPy's trim_mean both drop floor(0.1 n) = 1 value from each end of the 19 sorted values.")

Whole script generators/r/hypothesis_bootstrap_estimate.R


Bootstrap estimate first quartile

Case STAT-HYP-BOOTSTRAP-ESTIMATE-005 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 base and stats
Citation
R Core Team, base and stats package help: mean, median, sd, quantile, cor.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bootstrap_estimate.R · sha256 8eb6eded9e3eb420
Dataset
reference_data/datasets/HYP_BOOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_usedexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/unitsexact19190.00E+2passed
/components/0/estimatetolerance11.05000000000000111.0500000000000021.49E-1515passed
/components/0/attemptedexact2002000.00E+2passed
/components/0/successfulexact2002000.00E+2passed
/components/0/failedexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

102  emit("STAT-HYP-BOOTSTRAP-ESTIMATE-005", "Bootstrap estimate first quartile", "axis_coverage",
103       list(statistic = "quantile", column = "v", probability = 0.25),
104       counts(length(v), stats::quantile(v, 0.25, type = 7, names = FALSE), "quantile(0.25, type = 7)"),
105       "Quantile with R's type 7 linear interpolation (NumPy's default 'linear' method).")

Whole script generators/r/hypothesis_bootstrap_estimate.R


Bootstrap estimate Pearson correlation

Case STAT-HYP-BOOTSTRAP-ESTIMATE-006 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 base and stats
Citation
R Core Team, base and stats package help: mean, median, sd, quantile, cor.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bootstrap_estimate.R · sha256 8eb6eded9e3eb420
Dataset
reference_data/datasets/HYP_BOOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_usedexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/unitsexact19190.00E+2passed
/components/0/estimatetolerance0.99338563686847170.99338563686847181.11E-1615passed
/components/0/attemptedexact2002000.00E+2passed
/components/0/successfulexact2002000.00E+2passed
/components/0/failedexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

107  complete <- is.finite(d$v) & is.finite(d$w)
108  emit("STAT-HYP-BOOTSTRAP-ESTIMATE-006", "Bootstrap estimate Pearson correlation", "axis_coverage",
109       list(statistic = "pearson", column = "v", second = "w"),
110       counts(sum(complete), stats::cor(d$v[complete], d$w[complete]), "Pearson r on the complete pairs"),
111       "Pearson correlation of v and w on the 19 rows complete on both.")

Whole script generators/r/hypothesis_bootstrap_estimate.R


Bootstrap estimate mean difference B minus A

Case STAT-HYP-BOOTSTRAP-ESTIMATE-007 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 base and stats
Citation
R Core Team, base and stats package help: mean, median, sd, quantile, cor.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bootstrap_estimate.R · sha256 8eb6eded9e3eb420
Dataset
reference_data/datasets/HYP_BOOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_usedexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/unitsexact19190.00E+2passed
/components/0/estimatetolerance-0.86555555555555586-0.86555555555555594.44E-1817passed
/components/0/attemptedexact2002000.00E+2passed
/components/0/successfulexact2002000.00E+2passed
/components/0/failedexact000.00E+2passed
/components/0/nameexactMean of v: B minus AMean of v: B minus Apassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

113  a <- d$v[is.finite(d$v) & d$g == "A"]
114  b <- d$v[is.finite(d$v) & d$g == "B"]
115  emit("STAT-HYP-BOOTSTRAP-ESTIMATE-007", "Bootstrap estimate mean difference B minus A", "axis_coverage",
116       list(statistic = "mean_difference", column = "v", group = "g", first_group = "B"),
117       counts(length(a) + length(b), mean(b) - mean(a), "mean of group B minus mean of group A", "Mean of v: B minus A"),
118       "Difference of two group means with the declared first group B: mean(B) - mean(A) on the finite rows.")

Whole script generators/r/hypothesis_bootstrap_estimate.R


Bootstrap estimate mean minus target

Case STAT-HYP-BOOTSTRAP-ESTIMATE-008 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 base and stats
Citation
R Core Team, base and stats package help: mean, median, sd, quantile, cor.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bootstrap_estimate.R · sha256 8eb6eded9e3eb420
Dataset
reference_data/datasets/HYP_BOOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_usedexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/unitsexact19190.00E+2passed
/components/0/estimatetolerance0.50.50.00E+1passed
/components/0/attemptedexact2002000.00E+2passed
/components/0/successfulexact2002000.00E+2passed
/components/0/failedexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

120  emit("STAT-HYP-BOOTSTRAP-ESTIMATE-008", "Bootstrap estimate mean minus target", "axis_coverage",
121       list(statistic = "mean_minus_target", column = "v", target = 12),
122       counts(length(v), mean(v) - 12, "mean of v minus the fixed target 12"),
123       "Mean minus target: the target 12 is fixed in every replicate.")

Whole script generators/r/hypothesis_bootstrap_estimate.R


Bootstrap estimate mean of paired differences

Case STAT-HYP-BOOTSTRAP-ESTIMATE-009 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 base and stats
Citation
R Core Team, base and stats package help: mean, median, sd, quantile, cor.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bootstrap_estimate.R · sha256 8eb6eded9e3eb420
Dataset
reference_data/datasets/HYP_BOOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_usedexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/unitsexact19190.00E+2passed
/components/0/estimatetolerance-8.594736842105263-8.5947368421052638.31E-1818passed
/components/0/attemptedexact2002000.00E+2passed
/components/0/successfulexact2002000.00E+2passed
/components/0/failedexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

125  emit("STAT-HYP-BOOTSTRAP-ESTIMATE-009", "Bootstrap estimate mean of paired differences", "axis_coverage",
126       list(statistic = "paired_mean_difference", column = "v", second = "w"),
127       counts(sum(complete), mean(d$v[complete] - d$w[complete]), "mean of v minus w over the complete pairs"),
128       "Mean of paired differences v - w; each complete row is one pair.")

Whole script generators/r/hypothesis_bootstrap_estimate.R


Boundary: STAT-HYP-BOUNDARY-BASIC-MIN-001

Case STAT-HYP-BOUNDARY-BASIC-MIN-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_TINY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance22.00.00E+2passed
/p_valuetail_probability0.0455002638963584240.045500263896358393.46E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

143  z_min <- (tiny$x[[2]] - 0) / 1
144  emit("STAT-HYP-BOUNDARY-BASIC-MIN-001", "tabelz.executor.additional_basic_tests",
145       "additionalBasicTestsMenuItem", "HYP_TINY",
146       list(method = "known_sigma_z", response_column = "x", known_sigma = 1,
147            null_mean = 0, alpha = 0.1, max_rows = 100L),
148       list("/statistic" = tol(z_min, "one-observation known-sigma z"),
149            "/p_value" = prob(2 * stats::pnorm(-abs(z_min)), "normal tail")),
150       notes = list("One finite response is sufficient when population sigma is supplied."),
151       excluded = list(0L))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-BASIC-ZEROEVENT-001

Case STAT-HYP-BOUNDARY-BASIC-ZEROEVENT-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_ZERO_RATE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance00.00.00E+215passed
/p_valuetail_probability0.0447415437123311820.0447415437123311757.11E-1815passed
/uppertolerance0.0970757751082614580.097075775108261463.90E-1917passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

152  count <- sum(z$count)
153  exposure <- sum(z$exposure)
154  rate_test <- stats::poisson.test(count, T = exposure, r = 0.1)
155  expected_count <- 0.1 * exposure
156  p_zero <- min(1, 2 * min(stats::ppois(count, expected_count),
157                           stats::ppois(count - 1, expected_count, lower.tail = FALSE)))
158  emit("STAT-HYP-BOUNDARY-BASIC-ZEROEVENT-001", "tabelz.executor.additional_basic_tests",
159       "additionalBasicTestsMenuItem", "HYP_ZERO_RATE",
160       list(method = "one_poisson_rate", count_column = "count", exposure_column = "exposure",
161            null_rate = 0.1),
162       list("/estimate" = tol(0, "zero event rate"),
163            "/p_value" = prob(p_zero, "doubled minimum Poisson tail"),
164            "/upper" = tol(rate_test$conf.int[[2]], "positive exact upper rate bound")),
165       notes = list("Zero observed events; manual chapter 31 specifies doubled minimum-tail p-value."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-EQUIV-MIN-001

Case STAT-HYP-BOUNDARY-EQUIV-MIN-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_EDGE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/difference_a_minus_btolerance-0.25-0.250.00E+0passed
/p_lowertail_probability0.033412710221407620.033412710221407617.29E-1815passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

103  a <- e$x[e$g == "A"]
104  b <- e$x[e$g == "B"]
105  delta <- mean(a) - mean(b)
106  va <- var(a) / length(a)
107  vb <- var(b) / length(b)
108  se <- sqrt(va + vb)
109  df <- (va + vb)^2 / (va^2 / (length(a) - 1) + vb^2 / (length(b) - 1))
110  emit("STAT-HYP-BOUNDARY-EQUIV-MIN-001", "tabelz.executor.equivalence", "equivalenceMenuItem",
111       "HYP_EDGE",
112       list(response_column = "x", group_column = "g", group_a = "A", group_b = "B",
113            lower_bound = -3, upper_bound = 3, alpha = 0.1, max_rows = 100L),
114       list("/difference_a_minus_b" = tol(delta, "n=2/group A-B mean"),
115            "/p_lower" = prob(stats::pt((delta + 3) / se, df, lower.tail = FALSE), "Welch lower TOST")),
116       notes = list("Two finite observations per selected group; Welch-Satterthwaite df."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-EQUIV-ZEROSE-001

Case STAT-HYP-BOUNDARY-EQUIV-ZEROSE-001 · status: passed

Reference

Evidence type
Analytical
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

117  emit("STAT-HYP-BOUNDARY-EQUIV-ZEROSE-001", "tabelz.executor.equivalence", "equivalenceMenuItem",
118       "HYP_CONSTANT",
119       list(response_column = "x", group_column = "g", group_a = "A", group_b = "B",
120            lower_bound = -1, upper_bound = 1),
121       error = list(error_class = "EquivalenceExecutionError", message_contains = "zero standard error"),
122       notes = list("Both groups are constant at the same value; Welch standard error is zero."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-EQUIVEXT-MIN-001

Case STAT-HYP-BOUNDARY-EQUIVEXT-MIN-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_TINY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance00.00.00E+215passed
/p_lowertail_probability0.0525684567112534240.052568456711253441.36E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

125  mini_effect <- mean(tiny$x - 1.5)
126  mini_se <- sd(tiny$x - 1.5) / sqrt(nrow(tiny))
127  emit("STAT-HYP-BOUNDARY-EQUIVEXT-MIN-001", "tabelz.executor.equivalence_extensions",
128       "equivalenceMenuItem", "HYP_TINY",
129       list(mode = "one_sample_difference", response_column = "x", target = 1.5,
130            lower_bound = -3, upper_bound = 3, alpha = 0.1, max_rows = 100L),
131       list("/estimate" = tol(mini_effect, "n=2 mean-target effect"),
132            "/p_lower" = prob(stats::pt((mini_effect + 3) / mini_se, 1,
133                                         lower.tail = FALSE), "n=2 lower TOST")),
134       notes = list("Two finite nonconstant responses are the one-sample TOST minimum."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-EQUIVEXT-ZEROSE-001

Case STAT-HYP-BOUNDARY-EQUIVEXT-ZEROSE-001 · status: passed

Reference

Evidence type
Analytical
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

135  emit("STAT-HYP-BOUNDARY-EQUIVEXT-ZEROSE-001", "tabelz.executor.equivalence_extensions",
136       "equivalenceMenuItem", "HYP_CONSTANT",
137       list(mode = "one_sample_difference", response_column = "x", target = 1,
138            lower_bound = -1, upper_bound = 1),
139       error = list(error_class = "EquivalenceExtensionsExecutionError", message_contains = "zero standard error"),
140       notes = list("Four identical responses make the one-sample standard error zero."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-PERM-LIMIT-001

Case STAT-HYP-BOUNDARY-PERM-LIMIT-001 · status: passed

Reference

Evidence type
Analytical
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_PERM_LIMIT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

180  emit("STAT-HYP-BOUNDARY-PERM-LIMIT-001", "tabelz.executor.permutation_tests",
181       "permutationTestsMenuItem", "HYP_PERM_LIMIT",
182       list(mode = "independent", response_column = "x", group_column = "g",
183            group_a = "A", group_b = "B", inference = "exact"),
184       error = list(error_class = "PermutationTestsExecutionError", message_contains = "above the bounded budget"),
185       notes = list("choose(22,11) exact assignments exceed the documented 20,000 budget."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-PERM-MIN-001

Case STAT-HYP-BOUNDARY-PERM-MIN-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_EDGE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-0.25-0.24999999999999991.11E-1615passed
/p_valuetail_probability0.666666666666666630.66666666666666663.41E-1918passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

168  idx <- utils::combn(seq_len(nrow(e)), 2)
169  sa <- colSums(matrix(e$x[idx], nrow = 2))
170  null <- sa / 2 - (sum(e$x) - sa) / 2
171  p_perm <- mean(abs(null) >= abs(delta) - 1e-12)
172  emit("STAT-HYP-BOUNDARY-PERM-MIN-001", "tabelz.executor.permutation_tests",
173       "permutationTestsMenuItem", "HYP_EDGE",
174       list(mode = "independent", response_column = "x", group_column = "g",
175            group_a = "A", group_b = "B", inference = "exact", max_rows = 100L,
176            max_cells = 1000L),
177       list("/statistic" = tol(delta, "n=2/group A-B mean"),
178            "/p_value" = prob(p_perm, "all six assignments")),
179       notes = list("Two complete responses per group; exactly choose(4,2)=6 label assignments."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-POWER-MIN-001

Case STAT-HYP-BOUNDARY-POWER-MIN-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.123566742658828940.123566742658729549.94E-1412passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

85  p_min <- stats::power.t.test(n = 2, delta = 0.5, sd = 1, sig.level = 0.1,
86                               type = "one.sample", alternative = "two.sided", strict = TRUE)
87  emit("STAT-HYP-BOUNDARY-POWER-MIN-001", "tabelz.analysis.hypothesis_tests:run_power_analysis",
88       "hypothesisTestsMenuItem", NULL,
89       list(analysis = "one_sample_t", solve_for = "power", sample_size = 2L,
90            effect_size = 0.5, alpha = 0.1),
91       list("/value" = tol(p_min$power, "n=2 noncentral t power")),
92       notes = list("Sample size two is the smallest accepted planned t sample."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-POWER-NEARNULL-001

Case STAT-HYP-BOUNDARY-POWER-NEARNULL-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.0500000000000003770.0500000000000003156.20E-1714passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 93  p_near <- stats::power.t.test(n = 30, delta = 1e-8, sd = 1, sig.level = 0.05,
 94                                type = "one.sample", alternative = "two.sided", strict = TRUE)
 95  emit("STAT-HYP-BOUNDARY-POWER-NEARNULL-001", "tabelz.analysis.hypothesis_tests:run_power_analysis",
 96       "hypothesisTestsMenuItem", NULL,
 97       list(analysis = "one_sample_t", solve_for = "power", sample_size = 30L,
 98            effect_size = 1e-8),
 99       list("/value" = tol(p_near$power, "tiny-positive-effect noncentral t power")),
100       notes = list("Effect size is positive but near zero, testing the no-effect boundary."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-RANK-MIN-001

Case STAT-HYP-BOUNDARY-RANK-MIN-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_EDGE_BLOCKS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/friedman_statistictolerance44.00.00E+2passed
/p_valuetail_probability0.13533528323661270.13533528323661272.32E-1816passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

188  friedman <- stats::friedman.test(x ~ treatment | block, data = bl)
189  emit("STAT-HYP-BOUNDARY-RANK-MIN-001", "tabelz.executor.blocked_rank_tests",
190       "blockedRankTestsMenuItem", "HYP_EDGE_BLOCKS",
191       list(method = "friedman", response_column = "x", treatment_column = "treatment",
192            block_column = "block", max_rows = 100L, max_blocks = 10L),
193       list("/friedman_statistic" = tol(friedman$statistic, "two-block Friedman chi-square"),
194            "/p_value" = prob(friedman$p.value, "Friedman chi-square tail")),
195       notes = list("Two complete blocks with three treatments each are the minimum design."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-RANK-TIES-001

Case STAT-HYP-BOUNDARY-RANK-TIES-001 · status: passed

Reference

Evidence type
Analytical
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

196  emit("STAT-HYP-BOUNDARY-RANK-TIES-001", "tabelz.executor.blocked_rank_tests",
197       "blockedRankTestsMenuItem", "HYP_CONSTANT",
198       list(method = "runs", response_column = "x"),
199       error = list(error_class = "BlockedRankTestsExecutionError", message_contains = "at least one observation in each category"),
200       notes = list("All numeric values equal the median and are omitted; no runs sequence remains."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-TEST-CONSTANT-001

Case STAT-HYP-BOUNDARY-TEST-CONSTANT-001 · status: passed

Reference

Evidence type
Analytical
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

78  emit("STAT-HYP-BOUNDARY-TEST-CONSTANT-001", "tabelz.analysis.hypothesis_tests:run_hypothesis_test",
79       "hypothesisTestsMenuItem", "HYP_CONSTANT",
80       list(test = "one_sample_t", response = "x", target = 0),
81       error = list(error_class = "ValueError", message_contains = "no sample variation"),
82       notes = list("Zero sample variance makes a finite t statistic undefined."))

Whole script generators/r/hypothesis_boundaries.R


Boundary: STAT-HYP-BOUNDARY-TEST-MIN-001

Case STAT-HYP-BOUNDARY-TEST-MIN-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and manual-defined design boundaries
Citation
R stats help: t.test, poisson.test, friedman.test, ptukey, power.t.test; Tabelz manual chapters 07, 16, 31-34, 62.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_boundaries.R · sha256 39e5c59beef75737
Dataset
reference_data/datasets/HYP_TINY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance33.00.00E+2passed
/p_valuetail_probability0.204832764699133450.204832764699133453.92E-1816passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

71  t_min <- stats::t.test(tiny$x, mu = 0)
72  emit("STAT-HYP-BOUNDARY-TEST-MIN-001", "tabelz.analysis.hypothesis_tests:run_hypothesis_test",
73       "hypothesisTestsMenuItem", "HYP_TINY",
74       list(test = "one_sample_t", response = "x", target = 0, alpha = 0.1),
75       list("/statistic" = tol(t_min$statistic, "n=2 Student t"),
76            "/p_value" = prob(t_min$p.value, "n=2 Student t tail")),
77       notes = list("Two nonconstant finite observations are the minimum for a Student t statistic."))

Whole script generators/r/hypothesis_boundaries.R


Two-way categorical association

Case STAT-HYP-CHI-ASSOC-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and car
Citation
R Core Team, stats help: oneway.test, chisq.test, binom.test, wilcox.test, aov and kruskal.test; car::Anova and car::leveneTest help.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/hypothesis_methods.R · sha256 e2faa3576983ee8d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance00.00.00E+2passed
/p_valuetail_probability11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

62  association <- stats::chisq.test(table(d$g, d$h), correct = FALSE)
63  emit("STAT-HYP-CHI-ASSOC-001", "Two-way categorical association",
64       list(test = "chi_square_association", group = "g", second_group = "h"),
65       list("/statistic" = tol(association$statistic, "Pearson association chi-square"),
66            "/p_value" = prob(association$p.value, "association chi-square tail")),
67       list("Yates continuity correction disabled, as the manual's Pearson test specifies."))

Whole script generators/r/hypothesis_methods.R


Two-category chi-square goodness of fit

Case STAT-HYP-CHI-GOF-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and car
Citation
R Core Team, stats help: oneway.test, chisq.test, binom.test, wilcox.test, aov and kruskal.test; car::Anova and car::leveneTest help.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/hypothesis_methods.R · sha256 e2faa3576983ee8d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance00.00.00E+2passed
/p_valuetail_probability11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

54  observed <- as.integer(table(d$outcome))
55  gof <- stats::chisq.test(observed, p = rep(0.5, 2))
56  emit("STAT-HYP-CHI-GOF-001", "Two-category chi-square goodness of fit",
57       list(test = "chi_square_gof", response = "outcome"),
58       list("/statistic" = tol(gof$statistic, "Pearson GOF chi-square"),
59            "/p_value" = prob(gof$p.value, "GOF chi-square tail")),
60       list("Blank expected proportions means equal category probabilities."))

Whole script generators/r/hypothesis_methods.R


One-sample difference TOST

Case STAT-HYP-EQUIV-ONE-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 independent implementation
Citation
R stats::t.test documentation; analytic one-sample TOST.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_secondary.R · sha256 13b3dbafe11fd4e1
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.812500000000000110.81250000000000021.12E-1615passed
/standard_errortolerance0.481479923430527530.481479923430527651.15E-1615passed
/p_lowertail_probability1.6227244268166216e-051.6227244268166257e-054.08E-2014passed
/p_uppertail_probability0.0130928772292040130.0130928772292040463.27E-1714passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 86  target <- 6
 87  delta <- d$x - target
 88  effect <- mean(delta)
 89  se1 <- sd(delta) / sqrt(length(delta))
 90  df1 <- length(delta) - 1
 91  lo1 <- -2
 92  hi1 <- 2
 93  emit("STAT-HYP-EQUIV-ONE-001", "tabelz.executor.equivalence_extensions", "equivalenceMenuItem",
 94       "One-sample difference TOST", "nominal", "HYP_CORE",
 95       list(mode = "one_sample_difference", response_column = "x", target = target,
 96            lower_bound = lo1, upper_bound = hi1),
 97       list("/estimate" = tol(effect, "mean(x-target)"),
 98            "/standard_error" = tol(se1, "one-sample SE"),
 99            "/p_lower" = prob(stats::pt((effect - lo1) / se1, df1, lower.tail = FALSE), "lower TOST"),
100            "/p_upper" = prob(stats::pt((effect - hi1) / se1, df1), "upper TOST")),
101       list("Student one-sample TOST on x-target."),
102       "R stats::t.test documentation; analytic one-sample TOST.")

Whole script generators/r/hypothesis_secondary.R


Welch two-group TOST

Case STAT-HYP-EQUIV-WELCH-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 independent implementation
Citation
R stats::t.test documentation (Welch variance and Student t quantiles); analytic TOST from the two one-sided tests.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_secondary.R · sha256 13b3dbafe11fd4e1
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/difference_a_minus_btolerance-1.5749999999999993-1.57500000000000028.78E-1615passed
/standard_errortolerance0.903514961533170010.90351496153317022.26E-1615passed
/degrees_of_freedomtolerance13.9994558941084713.9994558941084721.91E-1515passed
/p_lowertail_probability0.0685392586923868760.068539258692387071.94E-1614passed
/p_uppertail_probability8.6476893043807419e-058.647689304380732e-059.51E-2014passed
/interval_lowertolerance-3.1663744299504435-3.1663744299504451.32E-1515passed
/interval_uppertolerance0.0163744299504449130.016374429950444474.44E-1613passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

59  a <- d$x[d$g == "A"]
60  bb <- d$x[d$g == "B"]
61  diff <- mean(a) - mean(bb)
62  term_a <- var(a) / length(a)
63  term_b <- var(bb) / length(bb)
64  se <- sqrt(term_a + term_b)
65  df <- (term_a + term_b)^2 / (term_a^2 / (length(a) - 1) + term_b^2 / (length(bb) - 1))
66  lower <- -3
67  upper <- 3
68  p_lower <- stats::pt((diff - lower) / se, df, lower.tail = FALSE)
69  p_upper <- stats::pt((diff - upper) / se, df)
70  crit <- stats::qt(0.95, df)
71  emit("STAT-HYP-EQUIV-WELCH-001", "tabelz.executor.equivalence", "equivalenceMenuItem",
72       "Welch two-group TOST", "nominal", "HYP_CORE",
73       list(response_column = "x", group_column = "g", group_a = "A", group_b = "B",
74            lower_bound = lower, upper_bound = upper),
75       list("/difference_a_minus_b" = tol(diff, "A minus B"),
76            "/standard_error" = tol(se, "Welch standard error"),
77            "/degrees_of_freedom" = tol(df, "Welch-Satterthwaite df"),
78            "/p_lower" = prob(p_lower, "lower TOST"),
79            "/p_upper" = prob(p_upper, "upper TOST"),
80            "/interval_lower" = tol(diff - crit * se, "90% interval lower"),
81            "/interval_upper" = tol(diff + crit * se, "90% interval upper")),
82       list("Two one-sided Welch t tails; strict equivalence requires both p-values below alpha."),
83       "R stats::t.test documentation (Welch variance and Student t quantiles); analytic TOST from the two one-sided tests.")

Whole script generators/r/hypothesis_secondary.R


Complete-block Friedman

Case STAT-HYP-FRIEDMAN-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 independent implementation
Citation
R stats::friedman.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_secondary.R · sha256 13b3dbafe11fd4e1
Dataset
reference_data/datasets/HYP_BLOCKS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/friedman_statistictolerance88.00.00E+2passed
/p_valuetail_probability0.0183156388887341790.0183156388887341823.14E-1815passed
/block_countexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

137  friedman <- stats::friedman.test(x ~ treatment | block, data = b)
138  emit("STAT-HYP-FRIEDMAN-001", "tabelz.executor.blocked_rank_tests", "blockedRankTestsMenuItem",
139       "Complete-block Friedman", "nominal", "HYP_BLOCKS",
140       list(method = "friedman", response_column = "x", treatment_column = "treatment",
141            block_column = "block"),
142       list("/friedman_statistic" = tol(friedman$statistic, "Friedman chi-square"),
143            "/p_value" = prob(friedman$p.value, "Friedman chi-square tail"),
144            "/block_count" = exact(length(unique(b$block)), "Four complete blocks.")),
145       list("No within-block ties; R friedman.test chi-square approximation."),
146       "R stats::friedman.test documentation.")

Whole script generators/r/hypothesis_secondary.R


Large-offset numerical case: STAT-HYP-HARD-BASIC-001

Case STAT-HYP-HARD-BASIC-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, power.t.test; analytic TOST and runs formulae.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_hard.R · sha256 892a9672937f8cfa
Dataset
reference_data/datasets/HYP_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.6251.6250.00E-1passed
/p_valuetail_probability0.104162558830439120.104162558830439056.68E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

76  z <- (mean(d$x) - center) / (2 / sqrt(nrow(d)))
77  emit("STAT-HYP-HARD-BASIC-001", "tabelz.executor.additional_basic_tests",
78       "additionalBasicTestsMenuItem",
79       list(method = "known_sigma_z", response_column = "x", known_sigma = 2, null_mean = center),
80       list("/statistic" = tol(z, "known-sigma z"),
81            "/p_value" = prob(2 * stats::pnorm(-abs(z)), "normal tail")),
82       list("Known sigma 2 and large mean offset."))

Whole script generators/r/hypothesis_hard.R


Large-offset numerical case: STAT-HYP-HARD-EQUIV-001

Case STAT-HYP-HARD-EQUIV-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, power.t.test; analytic TOST and runs formulae.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_hard.R · sha256 892a9672937f8cfa
Dataset
reference_data/datasets/HYP_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/difference_a_minus_btolerance-1.5750000476837158-1.57500004768371582.03E-1716passed
/standard_errortolerance0.903514954300059790.90351495430005981.59E-1817passed
/p_lowertail_probability0.0685392633050430490.06853926330504315.59E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

51  a <- d$x[d$g == "A"]
52  b <- d$x[d$g == "B"]
53  delta <- mean(a) - mean(b)
54  va <- var(a) / length(a)
55  vb <- var(b) / length(b)
56  se <- sqrt(va + vb)
57  df <- (va + vb)^2 / (va^2 / (length(a) - 1) + vb^2 / (length(b) - 1))
58  emit("STAT-HYP-HARD-EQUIV-001", "tabelz.executor.equivalence", "equivalenceMenuItem",
59       list(response_column = "x", group_column = "g", group_a = "A", group_b = "B",
60            lower_bound = -3, upper_bound = 3),
61       list("/difference_a_minus_b" = tol(delta, "small A-B difference"),
62            "/standard_error" = tol(se, "Welch SE"),
63            "/p_lower" = prob(stats::pt((delta + 3) / se, df, lower.tail = FALSE), "lower TOST")),
64       list("Welch TOST on differences of means near one billion."))

Whole script generators/r/hypothesis_hard.R


Large-offset numerical case: STAT-HYP-HARD-EQUIVEXT-001

Case STAT-HYP-HARD-EQUIVEXT-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, power.t.test; analytic TOST and runs formulae.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_hard.R · sha256 892a9672937f8cfa
Dataset
reference_data/datasets/HYP_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.81250000745058060.81250000745058063.08E-1817passed
/standard_errortolerance0.481479920425928020.4814799204259287.55E-1917passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

66  one_effect <- mean(d$x - center)
67  one_se <- sd(d$x - center) / sqrt(nrow(d))
68  emit("STAT-HYP-HARD-EQUIVEXT-001", "tabelz.executor.equivalence_extensions",
69       "equivalenceMenuItem",
70       list(mode = "one_sample_difference", response_column = "x", target = center,
71            lower_bound = -2, upper_bound = 2),
72       list("/estimate" = tol(one_effect, "small mean-target effect"),
73            "/standard_error" = tol(one_se, "one-sample SE")),
74       list("Subtract the large target from each observed response before computing the t standard error."))

Whole script generators/r/hypothesis_hard.R


Large-offset numerical case: STAT-HYP-HARD-PERM-001

Case STAT-HYP-HARD-PERM-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, power.t.test; analytic TOST and runs formulae.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_hard.R · sha256 892a9672937f8cfa
Dataset
reference_data/datasets/HYP_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-1.5750000029802322-1.57500000298023223.88E-1716passed
/p_valuetail_probability0.103185703185703190.103185703185703192.55E-1816passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 88  centred <- d$x - stats::median(d$x)
 89  perm_delta <- mean(centred[d$g == "A"]) - mean(centred[d$g == "B"])
 90  sum_a <- colSums(matrix(centred[assign], nrow = length(a)))
 91  null <- sum_a / length(a) - (sum(centred) - sum_a) / length(b)
 92  p_exact <- mean(abs(null) >= abs(perm_delta))
 93  emit("STAT-HYP-HARD-PERM-001", "tabelz.executor.permutation_tests",
 94       "permutationTestsMenuItem",
 95       list(mode = "independent", response_column = "x", group_column = "g",
 96            group_a = "A", group_b = "B", inference = "exact"),
 97       list("/statistic" = tol(perm_delta, "centred large-offset mean difference"),
 98            "/p_value" = prob(p_exact, "all exact assignments, centred enumeration")),
 99       list("All choose(16,8) assignments enumerated with no Monte Carlo; the pooled responses are measured from their median so the exact tail is decided on the low-order bits; two-sided tail uses the absolute mean difference with no tie tolerance."),
100       version = "2.0.0")

Whole script generators/r/hypothesis_hard.R


Large-offset numerical case: STAT-HYP-HARD-POWER-001

Case STAT-HYP-HARD-POWER-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, power.t.test; analytic TOST and runs formulae.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_hard.R · sha256 892a9672937f8cfa
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.0500000321567815040.050000032156776295.21E-1512passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

131  power <- stats::power.t.test(n = 30, delta = 1e-4, sd = 1, sig.level = 0.05,
132                               type = "one.sample", alternative = "two.sided", strict = TRUE)
133  emit("STAT-HYP-HARD-POWER-001", "tabelz.analysis.hypothesis_tests:run_power_analysis",
134       "hypothesisTestsMenuItem",
135       list(analysis = "one_sample_t", solve_for = "power", effect_size = 1e-4,
136            sample_size = 30L),
137       list("/value" = tol(power$power, "near-null noncentral-t power")),
138       list("Strict two-sided power includes both rejection tails for a tiny effect."), dataset = NULL)

Whole script generators/r/hypothesis_hard.R


Large-offset numerical case: STAT-HYP-HARD-RANK-001

Case STAT-HYP-HARD-RANK-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, power.t.test; analytic TOST and runs formulae.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_hard.R · sha256 892a9672937f8cfa
Dataset
reference_data/datasets/HYP_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/expected_runstolerance99.00.00E+2passed
/p_valuetail_probability0.00966062299748495120.009660622997484951.74E-1815passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

102  sequence <- d$x > stats::median(d$x)
103  n1 <- sum(sequence)
104  n2 <- length(sequence) - n1
105  runs <- 1L + sum(sequence[-1] != sequence[-length(sequence)])
106  expected_runs <- 1 + 2 * n1 * n2 / length(sequence)
107  runs_var <- 2 * n1 * n2 * (2 * n1 * n2 - length(sequence)) /
108    (length(sequence)^2 * (length(sequence) - 1))
109  runs_p <- 2 * stats::pnorm(-abs((runs - expected_runs) / sqrt(runs_var)))
110  emit("STAT-HYP-HARD-RANK-001", "tabelz.executor.blocked_rank_tests",
111       "blockedRankTestsMenuItem",
112       list(method = "runs", response_column = "x", inference = "asymptotic"),
113       list("/expected_runs" = tol(expected_runs, "runs null expectation"),
114            "/p_value" = prob(runs_p, "asymptotic no-correction runs tail")),
115       list("Numeric median reference, source row order, no continuity correction."))

Whole script generators/r/hypothesis_hard.R


Large-offset numerical case: STAT-HYP-HARD-TEST-001

Case STAT-HYP-HARD-TEST-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, power.t.test; analytic TOST and runs formulae.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_hard.R · sha256 892a9672937f8cfa
Dataset
reference_data/datasets/HYP_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.68750547121725061.68750546068663981.05E-88passed
/p_valuetail_probability0.112183892625371180.112183894682387152.06E-97passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

43  center <- 1e9 + 6
44  t <- stats::t.test(d$x, mu = center)
45  emit("STAT-HYP-HARD-TEST-001", "tabelz.analysis.hypothesis_tests:run_hypothesis_test",
46       "hypothesisTestsMenuItem", list(test = "one_sample_t", response = "x", target = center),
47       list("/statistic" = tol(t$statistic, "one-sample t statistic"),
48            "/p_value" = prob(t$p.value, "Student t tail")),
49       list("Large common offset, small sample spread; null mean uses the same offset."))

Whole script generators/r/hypothesis_hard.R


Interacting axes: STAT-HYP-INTERACT-BASIC-ONERATE-GREATER-001

Case STAT-HYP-INTERACT-BASIC-ONERATE-GREATER-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
reference_data/datasets/HYP_RATES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.149999999999999990.154.45E-1816passed
/p_valuetail_probability0.538402666936381950.53840266693638172.26E-1615passed

Convention

Reference code

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-BASIC-ONERATE-LESS-001

Case STAT-HYP-INTERACT-BASIC-ONERATE-LESS-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
reference_data/datasets/HYP_RATES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.149999999999999990.154.45E-1816passed
/p_valuetail_probability0.575965248573064640.57596524857306451.15E-1615passed

Convention

Reference code

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-BASIC-SUMMARY-ONERATE-001

Case STAT-HYP-INTERACT-BASIC-SUMMARY-ONERATE-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.149999999999999990.154.45E-1816passed
/p_valuetail_probability11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

50  count <- sum(r$count)
51  exposure <- sum(r$exposure)
52  expected <- 0.15 * exposure
53  p_rate <- min(1, 2 * min(stats::ppois(count, expected),
54                           stats::ppois(count - 1, expected, lower.tail = FALSE)))
55  emit("STAT-HYP-INTERACT-BASIC-SUMMARY-ONERATE-001", basic, menu, NULL,
56       list(method = "one_poisson_rate", input_kind = "summary",
57            summary_count = count, summary_exposure = exposure, null_rate = 0.15),
58       list("/estimate" = tol(count / exposure, "summary event rate"),
59            "/p_value" = prob(p_rate, "doubled minimum Poisson tail")),
60       list("Summary count/exposure totals are from exported HYP_RATES; no table source is claimed."))

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-BASIC-SUMMARY-VARIANCE-001

Case STAT-HYP-INTERACT-BASIC-SUMMARY-VARIANCE-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance13.90937500000000113.9093750000000131.21E-1415passed
/p_valuetail_probability0.935173631038799910.93517363103880171.78E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

38  n <- nrow(d)
39  sample_var <- stats::var(d$x)
40  chi <- (n - 1) * sample_var / 4
41  two_tail_chi <- min(1, 2 * min(stats::pchisq(chi, n - 1),
42                                 stats::pchisq(chi, n - 1, lower.tail = FALSE)))
43  emit("STAT-HYP-INTERACT-BASIC-SUMMARY-VARIANCE-001", basic, menu, NULL,
44       list(method = "one_variance_chi_square", input_kind = "summary",
45            summary_variance = sample_var, sample_size = n, null_variance = 4),
46       list("/statistic" = tol(chi, "summary variance chi-square"),
47            "/p_value" = prob(two_tail_chi, "two-sided variance chi-square tail")),
48       list("Summary variance and n come from the R CSV; no source table is passed."))

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-BASIC-TWORATE-GREATER-001

Case STAT-HYP-INTERACT-BASIC-TWORATE-GREATER-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
reference_data/datasets/HYP_RATES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.552631578947368470.55263157894736853.64E-1817passed
/p_valuetail_probability0.899674257844005230.89967425784400531.13E-1615passed

Convention

Reference code

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-BASIC-TWORATE-LESS-001

Case STAT-HYP-INTERACT-BASIC-TWORATE-LESS-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
reference_data/datasets/HYP_RATES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.552631578947368470.55263157894736853.64E-1817passed
/p_valuetail_probability0.245755762435790590.24575576243579063.23E-1816passed

Convention

Reference code

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-BASIC-VARIANCE-GREATER-001

Case STAT-HYP-INTERACT-BASIC-VARIANCE-GREATER-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance13.90937500000000113.9093750000000043.26E-1515passed
/p_valuetail_probability0.532413184480600040.53241318448059973.30E-1615passed

Convention

Reference code

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-BASIC-VARIANCE-LESS-001

Case STAT-HYP-INTERACT-BASIC-VARIANCE-LESS-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance13.90937500000000113.9093750000000043.26E-1515passed
/p_valuetail_probability0.467586815519399960.467586815519400353.85E-1615passed

Convention

Reference code

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-POWER-ANOVA-001

Case STAT-HYP-INTERACT-POWER-ANOVA-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valueexact42420.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

142  anova <- pwr::pwr.anova.test(k = 3, f = 0.5, sig.level = 0.05, power = 0.8)
143  emit("STAT-HYP-INTERACT-POWER-ANOVA-001", power_entry, power_menu, NULL,
144       list(analysis = "one_way_anova", solve_for = "sample_size", effect_size = 0.5,
145            target_power = 0.8, groups = 3L),
146       list("/value" = exact(3 * ceiling(anova$n), "Three equal groups rounded upward.")),
147       list("R noncentral F ANOVA with three equal-size groups; total is a multiple of three."))

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-POWER-ONE-PROP-001

Case STAT-HYP-INTERACT-POWER-ONE-PROP-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valueexact32320.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

127  one_prop <- pwr::pwr.chisq.test(w = 0.5, df = 1, sig.level = 0.05, power = 0.8)
128  emit("STAT-HYP-INTERACT-POWER-ONE-PROP-001", power_entry, power_menu, NULL,
129       list(analysis = "one_proportion", solve_for = "sample_size", effect_size = 0.5,
130            target_power = 0.8),
131       list("/value" = exact(ceiling(one_prop$N), "Required one-proportion chi-square n.")),
132       list("Cohen h=0.5 enters the one-degree chi-square approximation as w=0.5."))

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-POWER-PAIRED-001

Case STAT-HYP-INTERACT-POWER-PAIRED-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valueexact34340.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

107  power_entry <- "tabelz.analysis.hypothesis_tests:run_power_analysis"
108  power_menu <- "hypothesisTestsMenuItem"
109  paired <- stats::power.t.test(power = 0.8, delta = 0.5, sd = 1,
110                                sig.level = 0.05, type = "paired",
111                                alternative = "two.sided", strict = TRUE)
112  emit("STAT-HYP-INTERACT-POWER-PAIRED-001", power_entry, power_menu, NULL,
113       list(analysis = "paired_t", solve_for = "sample_size", effect_size = 0.5,
114            target_power = 0.8),
115       list("/value" = exact(ceiling(paired$n), "Round required pairs upward.")),
116       list("R noncentral t power for paired differences, strict two-sided."))

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-POWER-TWO-PROP-001

Case STAT-HYP-INTERACT-POWER-TWO-PROP-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valueexact1261260.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

134  two_prop <- pwr::pwr.2p.test(h = 0.5, sig.level = 0.05, power = 0.8,
135                               alternative = "two.sided")
136  emit("STAT-HYP-INTERACT-POWER-TWO-PROP-001", power_entry, power_menu, NULL,
137       list(analysis = "two_proportion", solve_for = "sample_size", effect_size = 0.5,
138            target_power = 0.8, ratio = 1),
139       list("/value" = exact(2 * ceiling(two_prop$n), "Required total across equal groups.")),
140       list("Cohen h=0.5; normal two-group approximation, round each group upward."))

Whole script generators/r/hypothesis_interactions.R


Interacting axes: STAT-HYP-INTERACT-POWER-TWO-T-001

Case STAT-HYP-INTERACT-POWER-TWO-T-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test, pchisq, ppois, pbinom; pwr::pwr.chisq.test, pwr.2p.test, pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_interactions.R · sha256 ab0cee41d6980e30
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valueexact1281280.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

118  two_t <- stats::power.t.test(power = 0.8, delta = 0.5, sd = 1,
119                               sig.level = 0.05, type = "two.sample",
120                               alternative = "two.sided", strict = TRUE)
121  emit("STAT-HYP-INTERACT-POWER-TWO-T-001", power_entry, power_menu, NULL,
122       list(analysis = "two_sample_t", solve_for = "sample_size", effect_size = 0.5,
123            target_power = 0.8, ratio = 1),
124       list("/value" = exact(2 * ceiling(two_t$n), "Required total across equal groups.")),
125       list("R two-sample noncentral t with equal group sizes; round each group upward."))

Whole script generators/r/hypothesis_interactions.R


Invalid request: Known sigma must be positive.

Case STAT-HYP-INVALID-BASIC-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Alpha must be strictly positive.

Case STAT-HYP-INVALID-BASIC-002 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Lower bound must precede upper bound.

Case STAT-HYP-INVALID-EQUIV-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Compared groups must differ.

Case STAT-HYP-INVALID-EQUIV-002 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Lower bound must precede upper bound.

Case STAT-HYP-INVALID-EQUIVEXT-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Paired mode needs both columns.

Case STAT-HYP-INVALID-EQUIVEXT-002 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Independent labels must differ.

Case STAT-HYP-INVALID-PERM-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Inference is exact or Monte Carlo.

Case STAT-HYP-INVALID-PERM-002 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Planning analysis must be listed.

Case STAT-HYP-INVALID-POWER-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Standardized effect size is positive.

Case STAT-HYP-INVALID-POWER-002 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Friedman needs treatment and block columns.

Case STAT-HYP-INVALID-RANK-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Friedman is two-sided.

Case STAT-HYP-INVALID-RANK-002 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Alpha must be strictly positive.

Case STAT-HYP-INVALID-TEST-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Invalid request: Only listed hypothesis tests are accepted.

Case STAT-HYP-INVALID-TEST-002 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz user manual: hypothesis and feature setup rules
Citation
docs/manual/chapters/07-models-and-tests.tex and feature chapters 16, 31-34, 62
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_invalid.R · sha256 65212f2bd73aed4f
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_invalid.R


Kruskal-Wallis rank test

Case STAT-HYP-KRUSKAL-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance2.1617647058823552.1617647058823553.10E-1716passed
/p_valuetail_probability0.141482121482793380.1414821214827894.39E-1513passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

155  kruskal <- stats::kruskal.test(x ~ g, data = d)
156  emit("STAT-HYP-KRUSKAL-001", "Kruskal-Wallis rank test", "nominal",
157       list(test = "kruskal_wallis", response = "x", group = "g"),
158       list("/statistic" = tol(kruskal$statistic, "Kruskal-Wallis H"),
159            "/p_value" = prob(kruskal$p.value, "Kruskal-Wallis chi-square tail")),
160       list("R kruskal.test tie-corrected chi-square."))

Whole script generators/r/hypothesis_tests.R


Two-sample exact rank test

Case STAT-HYP-MANN-WHITNEY-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1818.00.00E+2passed
/p_valuetail_probability0.16052836052836050.16052836052836053.07E-1816passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

148  mann <- stats::wilcox.test(x ~ g, data = d, exact = TRUE, correct = FALSE)
149  emit("STAT-HYP-MANN-WHITNEY-001", "Two-sample exact rank test", "nominal",
150       list(test = "mann_whitney", response = "x", group = "g"),
151       list("/statistic" = tol(mann$statistic, "Mann-Whitney U"),
152            "/p_value" = prob(mann$p.value, "exact rank tail")),
153       list("No ties and n=8/group: R wilcox.test exact p. R W equals U for the first group."))

Whole script generators/r/hypothesis_tests.R


Common-effect meta-analysis of generic effects

Case STAT-HYP-META-001 · status: passed

Reference

Evidence type
R
Source
Base R (stats) inverse-variance pooling written from the manual formulas
Citation
DerSimonian and Laird (1986) Controlled Clinical Trials 7:177-188; Hartung and Knapp (2001) Statistics in Medicine 20:3875-3889; Viechtbauer (2007) Statistics in Medicine 26:37-52 (Q-profile interval); Egger et al. (1997) BMJ 315:629-634; Begg and Mazumdar (1994) Biometrics 50:1088-1101; R stats: lm, cor.test (exact Kendall), uniroot, pt, pnorm, pchisq.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_meta_analysis.R · sha256 cb731b4b15b1a99e
Dataset
reference_data/datasets/HYP_META_STUDIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pooled/0/estimatetolerance0.338701680095177930.338701680095177875.94E-1715passed
/pooled/0/standard_errortolerance0.0487848028932759140.0487848028932759143.18E-1917passed
/pooled/0/statistictolerance6.94277028926690056.9427702892668991.74E-1515passed
/pooled/0/p_valuetail_probability3.8448435076081302e-123.844843507608176e-124.60E-2613passed
/pooled/0/lowertolerance0.243085223431471740.243085223431471658.69E-1715passed
/pooled/0/uppertolerance0.434318136758884090.43431813675888412.01E-1817passed
/heterogeneity/qtolerance12.21153962817890112.2115396281789014.02E-1717passed
/heterogeneity/dfexact770.00E+2passed
/heterogeneity/p_valuetail_probability0.093813386519811220.093813386519811165.51E-1715passed
/kexact880.00E+2passed
/heterogeneity/i2tolerance42.67717083071936242.677170830719361.58E-1718passed
/heterogeneity/h2tolerance1.74450566116841441.74450566116841443.43E-1716passed
/studies/0/weight_percenttolerance19.66906606062629919.66906606062634.75E-1616passed
/studies/4/weight_percenttolerance29.38218510291089229.3821851029108963.58E-1515passed
/studies/0/lowertolerance0.0944039617005940910.094403961700594062.76E-1715passed
/bias/0/statistictolerance-0.71722519962852016-0.71722519962852031.11E-1615passed
/bias/0/p_valuetail_probability0.500191106807762690.50019110680776261.15E-1615passed
/bias/1/statistictolerance-0.21428571428571427-0.214285714285714273.82E-1816passed
/bias/1/p_valuetail_probability0.548412698412698460.54841269841269852.75E-1817passed
/cumulative/7/estimatetolerance0.338701680095177930.338701680095177875.94E-1715passed
/leave_one_out/0/estimatetolerance0.34572929964543830.345729299645438255.17E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

212  y <- d$effect; v <- d$se^2; k <- length(y)
213  cp <- common_pool(y, v)
214  eg <- egger(y, v); bg <- begg(y, v)
215  outputs <- common_outputs(cp, 0.95)
216  outputs[["/k"]] <- exact(k, "eight studies")
217  outputs[["/heterogeneity/i2"]] <- tol(max(0, 100 * (cp$q - cp$df) / cp$q), "common-effect I^2 from Q")
218  outputs[["/heterogeneity/h2"]] <- tol(cp$q / cp$df, "common-effect H^2 = Q/df")
219  outputs[["/studies/0/weight_percent"]] <- tol(cp$weights[1], "inverse-variance weight of study 1 (%)")
220  outputs[["/studies/4/weight_percent"]] <- tol(cp$weights[5], "inverse-variance weight of study 5 (%)")
221  outputs[["/studies/0/lower"]] <- tol(y[1] - stats::qnorm(0.975) * sqrt(v[1]), "study 1 normal CI lower")
222  outputs[["/bias/0/statistic"]] <- tol(eg$t, "Egger intercept t (lm of y/SE on 1/SE)")
223  outputs[["/bias/0/p_value"]] <- prob(eg$p, "Egger intercept p on k - 2 df")
224  outputs[["/bias/1/statistic"]] <- tol(bg$tau, "Begg Kendall tau (cor.test)")
225  outputs[["/bias/1/p_value"]] <- prob(bg$p, "Begg exact Kendall p (cor.test exact=TRUE)")
226  outputs[["/cumulative/7/estimate"]] <- tol(cp$mu, "cumulative estimate after all studies = pooled")
227  loo <- common_pool(y[-1], v[-1])
228  outputs[["/leave_one_out/0/estimate"]] <- tol(loo$mu, "common-effect estimate without study 1")
229  emit("STAT-HYP-META-001", "Common-effect meta-analysis of generic effects", "nominal",
230       list(data = "generic", study = "study", effect = "effect", standard_error = "se",
231            model = "fixed", measure = "hedges_g"),
232       outputs,
233       list("Common effect: mu = sum(w y)/sum(w), w = 1/se^2, normal z inference; Q = sum w (y - mu)^2 on k - 1 df; I^2 = max(0, 100 (Q - df)/Q), H^2 = Q/df.",
234            "Egger: lm(y/se ~ 1/se), intercept t on k - 2 df. Begg: Kendall tau between (y - mu)/sqrt(v - 1/sum w) and v, exact p (no ties)."))

Whole script generators/r/hypothesis_meta_analysis.R


Random effects, DerSimonian-Laird, normal inference

Case STAT-HYP-META-002 · status: passed

Reference

Evidence type
R
Source
Base R (stats) inverse-variance pooling written from the manual formulas
Citation
DerSimonian and Laird (1986) Controlled Clinical Trials 7:177-188; Hartung and Knapp (2001) Statistics in Medicine 20:3875-3889; Viechtbauer (2007) Statistics in Medicine 26:37-52 (Q-profile interval); Egger et al. (1997) BMJ 315:629-634; Begg and Mazumdar (1994) Biometrics 50:1088-1101; R stats: lm, cor.test (exact Kendall), uniroot, pt, pnorm, pchisq.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_meta_analysis.R · sha256 cb731b4b15b1a99e
Dataset
reference_data/datasets/HYP_META_STUDIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pooled/0/estimatetolerance0.338701680095177930.338701680095177875.94E-1715passed
/pooled/0/standard_errortolerance0.0487848028932759140.0487848028932759143.18E-1917passed
/pooled/0/statistictolerance6.94277028926690056.9427702892668991.74E-1515passed
/pooled/0/p_valuetail_probability3.8448435076081302e-123.844843507608176e-124.60E-2613passed
/pooled/0/lowertolerance0.243085223431471740.243085223431471658.69E-1715passed
/pooled/0/uppertolerance0.434318136758884090.43431813675888412.01E-1817passed
/heterogeneity/qtolerance12.21153962817890112.2115396281789014.02E-1717passed
/heterogeneity/dfexact770.00E+2passed
/heterogeneity/p_valuetail_probability0.093813386519811220.093813386519811165.51E-1715passed
/pooled/1/estimatetolerance0.325567272679882720.32556727267988274.34E-1816passed
/pooled/1/standard_errortolerance0.0683365094560687140.068336509456068711.86E-1917passed
/pooled/1/statistictolerance4.76417767414911894.7641776741491194.26E-1717passed
/pooled/1/p_valuetail_probability1.8962536666438452e-061.8962536666438446e-066.20E-2215passed
/pooled/1/lowertolerance0.191630175316807230.19163017531680722.47E-1715passed
/pooled/1/uppertolerance0.459504370042958190.459504370042958245.34E-1715passed
/pooled/1/prediction_lowertolerance-0.018013793646925069-0.01801379364692641.33E-1513passed
/pooled/1/prediction_uppertolerance0.669148339006690570.66914833900669191.34E-1514passed
/heterogeneity/tau2tolerance0.0150462571329575550.015046257132957555.13E-1815passed
/heterogeneity/i2tolerance42.67717083071936242.6771708307193557.09E-1515passed
/heterogeneity/h2tolerance1.74450566116841421.74450566116841442.34E-1615passed
/heterogeneity/tau2_uppertolerance0.2033449636432540.203344963643253973.24E-1715passed
/studies/0/weight_percenttolerance17.20266076375513417.2026607637551343.56E-1616passed
/leave_one_out/0/estimatetolerance0.325170517784165390.32517051778416541.54E-1817passed
/leave_one_out/0/tau2tolerance0.0230428352467389780.0230428352467389856.51E-1815passed
/cumulative/2/estimatetolerance0.33851613461300190.33851613461300194.44E-1816passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

237  t_dl <- tau2_dl(y, v)
238  stopifnot(t_dl > 0)
239  rp <- random_pool(y, v, t_dl, "normal", 0.95)
240  prof <- qprofile(y, v, 0.95)
241  outputs <- c(common_outputs(cp, 0.95), random_outputs(rp, t_dl, prof, cp, 0.95))
242  loo <- random_pool(y[-1], v[-1], tau2_dl(y[-1], v[-1]), "normal", 0.95)
243  outputs[["/leave_one_out/0/estimate"]] <- tol(loo$mu, "DL random-effects estimate without study 1", "1e-8")
244  outputs[["/leave_one_out/0/tau2"]] <- tol(tau2_dl(y[-1], v[-1]), "DL tau^2 without study 1", "1e-8")
245  cum3 <- random_pool(y[1:3], v[1:3], tau2_dl(y[1:3], v[1:3]), "normal", 0.95)
246  outputs[["/cumulative/2/estimate"]] <- tol(cum3$mu, "DL cumulative estimate after three studies", "1e-8")
247  emit("STAT-HYP-META-002", "Random effects, DerSimonian-Laird, normal inference", "axis_coverage",
248       list(data = "generic", study = "study", effect = "effect", standard_error = "se",
249            model = "random", measure = "hedges_g", tau2 = "dl", inference = "normal"),
250       outputs,
251       list("DL tau^2 = max(0, (Q - (k - 1)) / (sum w - sum w^2 / sum w)); random-effects weights 1/(v + tau^2); normal interval.",
252            "Typical within-study variance s^2 = (k - 1)/(sum w - sum w^2/sum w); I^2 = 100 tau^2/(tau^2 + s^2); H^2 = (tau^2 + s^2)/s^2.",
253            "Q-profile limits: tau^2 where the generalized Q equals chi-square(k - 1) quantiles 0.975 (lower) and 0.025 (upper), by uniroot to 1e-14; 0 when Q(0) is below the quantile.",
254            "Prediction interval: mu +- t(k - 2, 0.975) sqrt(tau^2 + 1/sum w*); relative tolerance 1e-8 because the Q-profile limits come from a root search."))

Whole script generators/r/hypothesis_meta_analysis.R


Hedges' g from two arms, DL, modified Hartung-Knapp, 90%

Case STAT-HYP-META-003 · status: passed

Reference

Evidence type
R
Source
Base R (stats) inverse-variance pooling written from the manual formulas
Citation
DerSimonian and Laird (1986) Controlled Clinical Trials 7:177-188; Hartung and Knapp (2001) Statistics in Medicine 20:3875-3889; Viechtbauer (2007) Statistics in Medicine 26:37-52 (Q-profile interval); Egger et al. (1997) BMJ 315:629-634; Begg and Mazumdar (1994) Biometrics 50:1088-1101; R stats: lm, cor.test (exact Kendall), uniroot, pt, pnorm, pchisq.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_meta_analysis.R · sha256 cb731b4b15b1a99e
Dataset
reference_data/datasets/HYP_META_STUDIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pooled/0/estimatetolerance0.416062151633170170.41606215163317024.14E-1817passed
/pooled/0/standard_errortolerance0.103450675942602250.103450675942602252.08E-1816passed
/pooled/0/statistictolerance4.02184082261593724.0218408226159372.76E-1717passed
/pooled/0/p_valuetail_probability5.7745077414981249e-055.774507741498131e-056.07E-2014passed
/pooled/0/lowertolerance0.245900932098399530.245900932098399448.60E-1715passed
/pooled/0/uppertolerance0.586223371167940770.58622337116794091.07E-1615passed
/heterogeneity/qtolerance15.49505366024962115.4950536602496192.06E-1515passed
/heterogeneity/dfexact770.00E+2passed
/heterogeneity/p_valuetail_probability0.0301519868842121740.0301519868842122063.16E-1714passed
/pooled/1/estimatetolerance0.409065523603523440.40906552360352355.54E-1715passed
/pooled/1/standard_errortolerance0.161158378072198630.161158378072198652.47E-1715passed
/pooled/1/statistictolerance2.53828270361633292.53828270361633254.41E-1615passed
/pooled/1/p_valuetail_probability0.0387611057766928040.038761105776692854.90E-1714passed
/pooled/1/lowertolerance0.10373830847692950.103738308476929555.27E-1715passed
/pooled/1/uppertolerance0.714392738730117440.71439273873011741.90E-1817passed
/pooled/1/prediction_lowertolerance-0.29483029557110885-0.294830295571109354.96E-1614passed
/pooled/1/prediction_uppertolerance1.11296134277815571.11296134277815646.93E-1615passed
/heterogeneity/tau2tolerance0.10524512157772120.10524512157772121.24E-1816passed
/heterogeneity/i2tolerance54.82429326490482454.824293264904826.86E-1515passed
/heterogeneity/h2tolerance2.2135790943213742.2135790943213747.85E-1817passed
/heterogeneity/tau2_uppertolerance0.614173501498964660.61417350149896451.13E-1615passed
/studies/0/weight_percenttolerance12.925198116973412.92519811697341.80E-1616passed
/heterogeneity/tau2_lowertolerance0.00968431474818456060.0096843147481845528.67E-1815passed
/heterogeneity/hk_scaletolerance1.06646998295109261.06646998295109292.54E-1615passed
/studies/0/effecttolerance0.384062778753969990.384062778753971.11E-1817passed
/studies/0/standard_errortolerance0.288395235308329060.288395235308329063.49E-1816passed
/studies/3/effecttolerance-0.29410290839615849-0.29410290839615851.12E-1817passed
/pooled/1/distributionexactt(7)t(7)passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

257  m <- d$n1 + d$n2 - 2
258  sp <- sqrt(((d$n1 - 1) * d$sd1^2 + (d$n2 - 1) * d$sd2^2) / m)
259  dd <- (d$mean1 - d$mean2) / sp
260  J <- exp(lgamma(m / 2) - 0.5 * log(m / 2) - lgamma((m - 1) / 2))
261  g <- J * dd
262  vg <- 1 / d$n1 + 1 / d$n2 + g^2 / (2 * (d$n1 + d$n2))
263  cpg <- common_pool(g, vg)
264  t_g <- tau2_dl(g, vg)
265  rpg <- random_pool(g, vg, t_g, "modified_hk", 0.90)
266  profg <- qprofile(g, vg, 0.90)
267  outputs <- c(common_outputs(cpg, 0.90), random_outputs(rpg, t_g, profg, cpg, 0.90))
268  outputs[["/studies/0/effect"]] <- tol(g[1], "Hedges' g of study 1 (exact J)")
269  outputs[["/studies/0/standard_error"]] <- tol(sqrt(vg[1]), "SE of Hedges' g of study 1")
270  outputs[["/studies/3/effect"]] <- tol(g[4], "Hedges' g of study 4")
271  outputs[["/pooled/1/distribution"]] <- list(comparison = "exact", value = "t(7)",
272                                              rationale = "Hartung-Knapp uses t on k - 1 = 7 df")
273  emit("STAT-HYP-META-003", "Hedges' g from two arms, DL, modified Hartung-Knapp, 90%", "axis_coverage",
274       list(data = "continuous", study = "study", n1 = "n1", mean1 = "mean1", sd1 = "sd1",
275            n2 = "n2", mean2 = "mean2", sd2 = "sd2", measure = "hedges_g", model = "random",
276            tau2 = "dl", inference = "modified_hk", confidence_level = 0.90),
277       outputs,
278       list("Hedges' g = J d with d = (m1 - m2)/s_pooled, J = Gamma(m/2)/(sqrt(m/2) Gamma((m - 1)/2)), m = n1 + n2 - 2; variance 1/n1 + 1/n2 + g^2/(2 (n1 + n2)) (metafor escalc SMD convention).",
279            "Modified Hartung-Knapp: Var(mu) = max(1, sum w (y - mu)^2/(k - 1)) / sum w with t(k - 1); the same scaled variance enters the prediction interval.",
280            "tau^2 = 0 would make the modified HK scale floor at 1; this data set has tau^2 > 0 (asserted in the generator)."))

Whole script generators/r/hypothesis_meta_analysis.R


Log odds ratio from events, REML, ordinary Hartung-Knapp

Case STAT-HYP-META-004 · status: passed

Reference

Evidence type
R
Source
Base R (stats) inverse-variance pooling written from the manual formulas
Citation
DerSimonian and Laird (1986) Controlled Clinical Trials 7:177-188; Hartung and Knapp (2001) Statistics in Medicine 20:3875-3889; Viechtbauer (2007) Statistics in Medicine 26:37-52 (Q-profile interval); Egger et al. (1997) BMJ 315:629-634; Begg and Mazumdar (1994) Biometrics 50:1088-1101; R stats: lm, cor.test (exact Kendall), uniroot, pt, pnorm, pchisq.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_meta_analysis.R · sha256 cb731b4b15b1a99e
Dataset
reference_data/datasets/HYP_META_STUDIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pooled/0/estimatetolerance-0.93203759368255279-0.93203759368255291.11E-1615passed
/pooled/0/standard_errortolerance0.233032183828386390.23303218382838641.11E-1817passed
/pooled/0/statistictolerance-3.9996088882252425-3.9996088882252434.64E-1615passed
/pooled/0/p_valuetail_probability6.3447250747710346e-056.344725074771012e-052.30E-1914passed
/pooled/0/lowertolerance-1.3887722812249073-1.38877228122490751.79E-1615passed
/pooled/0/uppertolerance-0.47530290614019838-0.47530290614019842.08E-1817passed
/heterogeneity/qtolerance12.97228838380294612.9722883838029442.09E-1515passed
/heterogeneity/dfexact770.00E+2passed
/heterogeneity/p_valuetail_probability0.0727864647965180830.072786464796518082.35E-1917passed
/pooled/1/estimatetolerance-1.0001121575329686-1.00011215753296881.82E-1615passed
/pooled/1/standard_errortolerance0.324103884449534310.324103884449534365.29E-1715passed
/pooled/1/statistictolerance-3.08577652264669-3.085776522646691.04E-1818passed
/pooled/1/p_valuetail_probability0.017671446837023670.017671446837023689.91E-1815passed
/pooled/1/lowertolerance-1.7664960627377626-1.7664960627377634.51E-1615passed
/pooled/1/uppertolerance-0.23372825232817451-0.23372825232817441.08E-1615passed
/pooled/1/prediction_lowertolerance-2.6741891205495505-2.6741891205495587.54E-1514passed
/pooled/1/prediction_uppertolerance0.673964805483613370.67396480548362057.11E-1513passed
/heterogeneity/tau2tolerance0.363030320003336960.36303032000333722.21E-1615passed
/heterogeneity/i2tolerance44.8325511083890444.8325511083890551.47E-1415passed
/heterogeneity/h2tolerance1.81266311945061691.81266311945061756.33E-1615passed
/heterogeneity/tau2_uppertolerance3.50970760033842223.5097076003384264.01E-1514passed
/studies/0/weight_percenttolerance13.58000406163392313.5800040616339234.84E-1616passed
/heterogeneity/hk_scaletolerance0.994235395803059750.99423539580305943.31E-1615passed
/studies/3/effecttolerance-2.95441412422326-2.954414124223263.57E-1817passed
/studies/3/standard_errortolerance1.52105239164090181.52105239164090182.03E-1716passed
/studies/0/effecttolerance-1.5761631235492282-1.57616312354922823.78E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

284  a <- d$events1; b <- d$total1 - d$events1; cc <- d$events2; dd2 <- d$total2 - d$events2
285  zero <- (a == 0 | b == 0 | cc == 0 | dd2 == 0)
286  stopifnot(sum(zero) == 1L, !any((a == 0 & cc == 0) | (b == 0 & dd2 == 0)))
287  a[zero] <- a[zero] + 0.5; b[zero] <- b[zero] + 0.5; cc[zero] <- cc[zero] + 0.5; dd2[zero] <- dd2[zero] + 0.5
288  lor <- log(a * dd2 / (b * cc))
289  vlor <- 1 / a + 1 / b + 1 / cc + 1 / dd2
290  cpl <- common_pool(lor, vlor)
291  t_reml <- tau2_reml(lor, vlor)
292  stopifnot(t_reml > 0)
293  rpl <- random_pool(lor, vlor, t_reml, "hk", 0.95)
294  profl <- qprofile(lor, vlor, 0.95)
295  outputs <- c(common_outputs(cpl, 0.95), random_outputs(rpl, t_reml, profl, cpl, 0.95))
296  outputs[[paste0("/studies/", which(zero) - 1, "/effect")]] <- tol(lor[zero], "log OR of the zero-cell study after adding 0.5 to each cell")
297  outputs[[paste0("/studies/", which(zero) - 1, "/standard_error")]] <- tol(sqrt(vlor[zero]), "SE of the corrected log OR")
298  outputs[["/studies/0/effect"]] <- tol(lor[1], "log OR of study 1")
299  emit("STAT-HYP-META-004", "Log odds ratio from events, REML, ordinary Hartung-Knapp", "axis_coverage",
300       list(data = "binary", study = "study", events1 = "events1", total1 = "total1",
301            events2 = "events2", total2 = "total2", measure = "log_or", correction = 0.5,
302            model = "random", tau2 = "reml", inference = "hk"),
303       outputs,
304       list("Log OR = log(a d/(b c)) with variance 1/a + 1/b + 1/c + 1/d; the one study with a zero cell gets 0.5 added to all four cells (manual chapter 78 lines 17-20).",
305            "REML tau^2: the root of the restricted score -tr(P)/2 + y'PPy/2 (one sign change on a 4000-point log grid, asserted), found by uniroot to 1e-15 and compared with the boundary 0 by the restricted likelihood.",
306            "Ordinary Hartung-Knapp: Var(mu) = (sum w (y - mu)^2/(k - 1)) / sum w without the floor at 1, t(k - 1).",
307            "Relative tolerance 1e-8: tau^2 is an iterative root in both implementations."))

Whole script generators/r/hypothesis_meta_analysis.R


Paule-Mandel tau^2 with normal inference, 99%

Case STAT-HYP-META-005 · status: passed

Reference

Evidence type
R
Source
Base R (stats) inverse-variance pooling written from the manual formulas
Citation
DerSimonian and Laird (1986) Controlled Clinical Trials 7:177-188; Hartung and Knapp (2001) Statistics in Medicine 20:3875-3889; Viechtbauer (2007) Statistics in Medicine 26:37-52 (Q-profile interval); Egger et al. (1997) BMJ 315:629-634; Begg and Mazumdar (1994) Biometrics 50:1088-1101; R stats: lm, cor.test (exact Kendall), uniroot, pt, pnorm, pchisq.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_meta_analysis.R · sha256 cb731b4b15b1a99e
Dataset
reference_data/datasets/HYP_META_STUDIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pooled/0/estimatetolerance0.338701680095177930.338701680095177875.94E-1715passed
/pooled/0/standard_errortolerance0.0487848028932759140.0487848028932759143.18E-1917passed
/pooled/0/statistictolerance6.94277028926690056.9427702892668991.74E-1515passed
/pooled/0/p_valuetail_probability3.8448435076081302e-123.844843507608176e-124.60E-2613passed
/pooled/0/lowertolerance0.213040355234820680.21304035523482068.58E-1715passed
/pooled/0/uppertolerance0.464363004955535170.46436300495553524.64E-1816passed
/heterogeneity/qtolerance12.21153962817890112.2115396281789014.02E-1717passed
/heterogeneity/dfexact770.00E+2passed
/heterogeneity/p_valuetail_probability0.093813386519811220.093813386519811165.51E-1715passed
/pooled/1/estimatetolerance0.323079151925690810.323079151925623366.74E-1412passed
/pooled/1/standard_errortolerance0.0747843362082120660.074784336208408021.96E-1311passed
/pooled/1/statistictolerance4.32014467610148414.3201446760892631.22E-1111passed
/pooled/1/p_valuetail_probability1.5592693918690127e-051.5592693919553694e-058.64E-1610passed
/pooled/1/lowertolerance0.130447467274125150.130447467273552915.72E-1311passed
/pooled/1/uppertolerance0.515710836577256470.51571083657769384.37E-1312passed
/pooled/1/prediction_lowertolerance-0.28972670441955839-0.28972670442237342.82E-1211passed
/pooled/1/prediction_uppertolerance0.935885008270940010.93588500827362012.68E-1211passed
/heterogeneity/tau2tolerance0.0217285400119379830.0217285400121532552.15E-1311passed
/heterogeneity/i2tolerance51.81077065051919651.810770650766562.47E-1011passed
/heterogeneity/h2tolerance2.07515250502916352.0751525050398161.07E-1111passed
/heterogeneity/tau2_uppertolerance0.375176294757212590.37517629475721261.69E-1817passed
/studies/0/weight_percenttolerance16.53248097650461216.5324809764860451.86E-1111passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

310  t_pm <- tau2_pm(y, v)
311  stopifnot(t_pm > 0)
312  rpp <- random_pool(y, v, t_pm, "normal", 0.99)
313  profp <- qprofile(y, v, 0.99)
314  outputs <- c(common_outputs(cp, 0.99), random_outputs(rpp, t_pm, profp, cp, 0.99))
315  emit("STAT-HYP-META-005", "Paule-Mandel tau^2 with normal inference, 99%", "axis_coverage",
316       list(data = "generic", study = "study", effect = "effect", standard_error = "se",
317            model = "random", measure = "hedges_g", tau2 = "pm", inference = "normal",
318            confidence_level = 0.99),
319       outputs,
320       list("Paule-Mandel tau^2: the root of the generalized Q(tau^2) = sum (y - mu*)^2/(v + tau^2) = k - 1, by uniroot to 1e-14 (0 when Q(0) <= k - 1).",
321            "Relative tolerance 1e-8: tau^2 and the Q-profile limits are iterative roots in both implementations."))

Whole script generators/r/hypothesis_meta_analysis.R


Below minimum: One retained response is below the sample-variance minimum.

Case STAT-HYP-MINIMUM-BASIC-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz manual sample/design requirements
Citation
docs/manual/chapters/07-models-and-tests.tex; 16, 31-34 and 62 feature chapters.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_minimum.R · sha256 26fd44b917263966
Dataset
reference_data/datasets/HYP_TINY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_minimum.R


Below minimum: One response in each group is below the Welch TOST minimum.

Case STAT-HYP-MINIMUM-EQUIV-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz manual sample/design requirements
Citation
docs/manual/chapters/07-models-and-tests.tex; 16, 31-34 and 62 feature chapters.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_minimum.R · sha256 26fd44b917263966
Dataset
reference_data/datasets/HYP_TINY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_minimum.R


Below minimum: One retained response is below the one-sample TOST minimum.

Case STAT-HYP-MINIMUM-EQUIVEXT-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz manual sample/design requirements
Citation
docs/manual/chapters/07-models-and-tests.tex; 16, 31-34 and 62 feature chapters.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_minimum.R · sha256 26fd44b917263966
Dataset
reference_data/datasets/HYP_TINY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_minimum.R


Below minimum: One response per group is below the independent permutation minimum.

Case STAT-HYP-MINIMUM-PERM-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz manual sample/design requirements
Citation
docs/manual/chapters/07-models-and-tests.tex; 16, 31-34 and 62 feature chapters.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_minimum.R · sha256 26fd44b917263966
Dataset
reference_data/datasets/HYP_TINY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_minimum.R


Below minimum: Prospective t power needs at least two planned observations.

Case STAT-HYP-MINIMUM-POWER-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz manual sample/design requirements
Citation
docs/manual/chapters/07-models-and-tests.tex; 16, 31-34 and 62 feature chapters.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_minimum.R · sha256 26fd44b917263966
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_minimum.R


Below minimum: One retained response cannot populate both runs categories.

Case STAT-HYP-MINIMUM-RANK-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz manual sample/design requirements
Citation
docs/manual/chapters/07-models-and-tests.tex; 16, 31-34 and 62 feature chapters.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_minimum.R · sha256 26fd44b917263966
Dataset
reference_data/datasets/HYP_TINY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_minimum.R


Below minimum: One remaining response is below the one-sample t minimum.

Case STAT-HYP-MINIMUM-TEST-001 · status: passed

Reference

Evidence type
Analytical
Source
Tabelz manual sample/design requirements
Citation
docs/manual/chapters/07-models-and-tests.tex; 16, 31-34 and 62 feature chapters.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0, stats 4.6.1
Reference script
generators/r/hypothesis_minimum.R · sha256 26fd44b917263966
Dataset
reference_data/datasets/HYP_TINY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/hypothesis_minimum.R


Missing and excluded population: STAT-HYP-MISSING-BASIC-001

Case STAT-HYP-MISSING-BASIC-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, binom.test; complete-case population calculated in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_population.R · sha256 4ad84083bcd91622
Dataset
reference_data/datasets/HYP_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance2.36526199092495572.36526199092495573.97E-1817passed
/p_valuetail_probability0.018017320102923180.018017320102923182.58E-1916passed
/sample_sizeexact14140.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

83  z <- (mean(x) - 6) / (2 / sqrt(length(x)))
84  emit("STAT-HYP-MISSING-BASIC-001", "tabelz.executor.additional_basic_tests",
85       "additionalBasicTestsMenuItem",
86       list(method = "known_sigma_z", response_column = "x", known_sigma = 2, null_mean = 6),
87       list("/statistic" = tol(z, "known-sigma Z"),
88            "/p_value" = prob(2 * stats::pnorm(-abs(z)), "normal tail"),
89            "/sample_size" = exact(length(x), "Finite responses.")),
90       list("Known sigma 2; incomplete response removed."))

Whole script generators/r/hypothesis_population.R


Missing and excluded population: STAT-HYP-MISSING-EQUIV-001

Case STAT-HYP-MISSING-EQUIV-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, binom.test; complete-case population calculated in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_population.R · sha256 4ad84083bcd91622
Dataset
reference_data/datasets/HYP_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/difference_a_minus_btolerance-0.85999999999999943-0.86000000000000038.90E-1614passed
/p_lowertail_probability0.016141097430740280.016141097430740313.16E-1714passed
/analyzed_countexact13130.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

55  grouped <- included[is.finite(included$x) & !is.na(included$g), , drop = FALSE]
56  a <- grouped$x[grouped$g == "A"]
57  b <- grouped$x[grouped$g == "B"]
58  delta <- mean(a) - mean(b)
59  va <- var(a) / length(a)
60  vb <- var(b) / length(b)
61  se <- sqrt(va + vb)
62  df <- (va + vb)^2 / (va^2 / (length(a) - 1) + vb^2 / (length(b) - 1))
63  emit("STAT-HYP-MISSING-EQUIV-001", "tabelz.executor.equivalence", "equivalenceMenuItem",
64       list(response_column = "x", group_column = "g", group_a = "A", group_b = "B",
65            lower_bound = -3, upper_bound = 3),
66       list("/difference_a_minus_b" = tol(delta, "A-B mean"),
67            "/p_lower" = prob(stats::pt((delta + 3) / se, df, lower.tail = FALSE), "lower TOST"),
68            "/analyzed_count" = exact(nrow(grouped), "Complete x/group rows.")),
69       list("Welch two-sample TOST, complete x/group rows."))

Whole script generators/r/hypothesis_population.R


Missing and excluded population: STAT-HYP-MISSING-EQUIVEXT-001

Case STAT-HYP-MISSING-EQUIVEXT-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, binom.test; complete-case population calculated in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_population.R · sha256 4ad84083bcd91622
Dataset
reference_data/datasets/HYP_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance1.26428571428571451.26428571428571454.30E-1716passed
/p_lowertail_probability1.5495157642539877e-061.5495157642539961e-068.42E-2114passed
/analyzed_countexact14140.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

71  effect <- mean(x - 6)
72  se_one <- sd(x) / sqrt(length(x))
73  emit("STAT-HYP-MISSING-EQUIVEXT-001", "tabelz.executor.equivalence_extensions",
74       "equivalenceMenuItem",
75       list(mode = "one_sample_difference", response_column = "x", target = 6,
76            lower_bound = -2, upper_bound = 2),
77       list("/estimate" = tol(effect, "mean(x-6)"),
78            "/p_lower" = prob(stats::pt((effect + 2) / se_one, length(x) - 1,
79                                         lower.tail = FALSE), "lower one-sample TOST"),
80            "/analyzed_count" = exact(length(x), "Complete response count.")),
81       list("One-sample difference TOST on finite x after exclusion."))

Whole script generators/r/hypothesis_population.R


Missing and excluded population: STAT-HYP-MISSING-PERM-001

Case STAT-HYP-MISSING-PERM-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, binom.test; complete-case population calculated in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_population.R · sha256 4ad84083bcd91622
Dataset
reference_data/datasets/HYP_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-0.85999999999999943-0.86000000000000038.90E-1614passed
/p_valuetail_probability0.383838383838383810.38383838383838382.03E-1817passed
/analyzed_countexact13130.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 92  assignments <- utils::combn(seq_len(nrow(grouped)), length(a))
 93  pooled <- grouped$x
 94  sum_a <- colSums(matrix(pooled[assignments], nrow = length(a)))
 95  null <- sum_a / length(a) - (sum(pooled) - sum_a) / length(b)
 96  p_perm <- mean(abs(null) >= abs(delta) - 1e-12)
 97  emit("STAT-HYP-MISSING-PERM-001", "tabelz.executor.permutation_tests",
 98       "permutationTestsMenuItem",
 99       list(mode = "independent", response_column = "x", group_column = "g",
100            group_a = "A", group_b = "B", inference = "exact"),
101       list("/statistic" = tol(delta, "included A-B mean"),
102            "/p_value" = prob(p_perm, "all complete-case label assignments"),
103            "/analyzed_count" = exact(nrow(grouped), "Complete x/group rows.")),
104       list("All choose(n,nA) complete-case assignments enumerated."))

Whole script generators/r/hypothesis_population.R


Missing and excluded population: STAT-HYP-MISSING-RANK-001

Case STAT-HYP-MISSING-RANK-001 · status: passed

Reference

Evidence type
Analytical
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, binom.test; complete-case population calculated in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_population.R · sha256 4ad84083bcd91622
Dataset
reference_data/datasets/HYP_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

109  stopifnot(any(!is.finite(included$x)))
110  emit("STAT-HYP-MISSING-RANK-001", "tabelz.executor.blocked_rank_tests",
111       "blockedRankTestsMenuItem",
112       list(method = "runs", response_column = "x", reference_mode = "median",
113            inference = "asymptotic"),
114       structure(list(), names = character()),
115       list("Chapter 32 and the issue #12 ruling: a missing value in the retained runs sequence is refused, not dropped.",
116            "Until 2026-09-25 this case expected the missing value to be dropped and the neighbours joined."),
117       error = list(error_class = "BlockedRankTestsExecutionError",
118                    message_contains = "missing or non-finite required values"))

Whole script generators/r/hypothesis_population.R


Missing and excluded population: STAT-HYP-MISSING-TEST-001

Case STAT-HYP-MISSING-TEST-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R stats package help: t.test, pnorm, ptukey, binom.test; complete-case population calculated in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_population.R · sha256 4ad84083bcd91622
Dataset
reference_data/datasets/HYP_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance3.00705194568602073.0070519456860182.62E-1515passed
/p_valuetail_probability0.0101009769868535930.010100976986853666.76E-1714passed
/sample_countexact14140.00E+2passed
/dropped_rowsexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

45  x <- included$x[is.finite(included$x)]
46  t <- stats::t.test(x, mu = 6)
47  emit("STAT-HYP-MISSING-TEST-001", "tabelz.analysis.hypothesis_tests:run_hypothesis_test",
48       "hypothesisTestsMenuItem", list(test = "one_sample_t", response = "x", target = 6),
49       list("/statistic" = tol(t$statistic, "one-sample t"),
50            "/p_value" = prob(t$p.value, "Student t tail"),
51            "/sample_count" = exact(length(x), "Finite selected responses after exclusion."),
52            "/dropped_rows" = exact(nrow(included) - length(x), "Missing selected response.")),
53       list("One-sample response x; R t.test removes the single missing response in included rows."))

Whole script generators/r/hypothesis_population.R


Mood median with ties below

Case STAT-HYP-MOOD-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and car
Citation
R Core Team, stats help: oneway.test, chisq.test, binom.test, wilcox.test, aov and kruskal.test; car::Anova and car::leveneTest help.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/hypothesis_methods.R · sha256 e2faa3576983ee8d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance11.00.00E+2passed
/p_valuetail_probability0.317310507862914150.317310507862911153.00E-1514passed
/estimatetolerance6.756.750.00E+0passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 97  grand_median <- stats::median(d$x)
 98  below <- d$x <= grand_median
 99  mood <- stats::chisq.test(table(d$g, below), correct = FALSE)
100  emit("STAT-HYP-MOOD-001", "Mood median with ties below",
101       list(test = "moods_median", response = "x", group = "g"),
102       list("/statistic" = tol(mood$statistic, "median contingency chi-square"),
103            "/p_value" = prob(mood$p.value, "median chi-square tail"),
104            "/estimate" = tol(grand_median, "grand median")),
105       list("Values equal to the pooled median count in the below cell; no Yates correction."))

Whole script generators/r/hypothesis_methods.R


Additional configuration coverage: STAT-HYP-NOMINAL-BASIC-GREATER-001

Case STAT-HYP-NOMINAL-BASIC-GREATER-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats/multcomp
Citation
R stats::pnorm, pt, qt, binom.test, power.t.test; multcomp::glht and mvtnorm documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, multcomp 1.4.32, mvtnorm 1.4.2, jsonlite 2.0.0
Reference script
generators/r/hypothesis_completion.R · sha256 733ecb7b7daec34d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.6251.6250.00E-1passed
/p_valuetail_probability0.0520812794152195610.052081279415219533.44E-1715passed

Convention

Reference code

Whole script generators/r/hypothesis_completion.R


Additional configuration coverage: STAT-HYP-NOMINAL-BASIC-LESS-001

Case STAT-HYP-NOMINAL-BASIC-LESS-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats/multcomp
Citation
R stats::pnorm, pt, qt, binom.test, power.t.test; multcomp::glht and mvtnorm documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, multcomp 1.4.32, mvtnorm 1.4.2, jsonlite 2.0.0
Reference script
generators/r/hypothesis_completion.R · sha256 733ecb7b7daec34d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.6251.6250.00E-1passed
/p_valuetail_probability0.947918720584780440.94791872058478041.28E-1817passed

Convention

Reference code

Whole script generators/r/hypothesis_completion.R


Additional configuration coverage: STAT-HYP-NOMINAL-RANK-RUNS-001

Case STAT-HYP-NOMINAL-RANK-RUNS-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats/multcomp
Citation
R stats::pnorm, pt, qt, binom.test, power.t.test; multcomp::glht and mvtnorm documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, multcomp 1.4.32, mvtnorm 1.4.2, jsonlite 2.0.0
Reference script
generators/r/hypothesis_completion.R · sha256 733ecb7b7daec34d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/runsexact440.00E+2passed
/p_valuetail_probability0.00966062299748495120.009660622997484951.74E-1815passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

58  reference <- stats::median(d$x)
59  sequence <- d$x > reference
60  sequence <- sequence[d$x != reference]
61  runs <- 1L + sum(sequence[-1] != sequence[-length(sequence)])
62  n1 <- sum(sequence)
63  n2 <- length(sequence) - n1
64  expected <- 1 + 2 * n1 * n2 / length(sequence)
65  variance <- 2 * n1 * n2 * (2 * n1 * n2 - length(sequence)) /
66    (length(sequence)^2 * (length(sequence) - 1))
67  p_runs <- 2 * stats::pnorm(-abs((runs - expected) / sqrt(variance)))
68  emit("STAT-HYP-NOMINAL-RANK-RUNS-001", "tabelz.executor.blocked_rank_tests",
69       "blockedRankTestsMenuItem", "HYP_CORE",
70       list(method = "runs", response_column = "x", reference_mode = "median",
71            inference = "asymptotic", alternative = "two-sided", max_exact_n = 20L),
72       list("/runs" = exact(as.integer(runs), "Observed runs in row order."),
73            "/p_value" = prob(p_runs, "no-continuity-correction normal tail")),
74       "nominal", list("Observed median, asymptotic normal runs test without continuity correction."))

Whole script generators/r/hypothesis_completion.R


Additional configuration coverage: STAT-HYP-NUMERIC-EQUIV-001

Case STAT-HYP-NUMERIC-EQUIV-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats/multcomp
Citation
R stats::pnorm, pt, qt, binom.test, power.t.test; multcomp::glht and mvtnorm documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, multcomp 1.4.32, mvtnorm 1.4.2, jsonlite 2.0.0
Reference script
generators/r/hypothesis_completion.R · sha256 733ecb7b7daec34d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/difference_a_minus_btolerance-1.5749999999999993-1.57500000000000028.78E-1615passed
/p_lowertail_probability0.0685392586923868760.068539258692387071.94E-1614passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

77  a <- d$x[d$g == "A"]
78  bb <- d$x[d$g == "B"]
79  effect <- mean(a) - mean(bb)
80  va <- var(a) / length(a)
81  vb <- var(bb) / length(bb)
82  se <- sqrt(va + vb)
83  df <- (va + vb)^2 / (va^2 / (length(a) - 1) + vb^2 / (length(bb) - 1))
84  emit("STAT-HYP-NUMERIC-EQUIV-001", "tabelz.executor.equivalence", "equivalenceMenuItem",
85       "HYP_CORE",
86       list(response_column = "x", group_column = "g", group_a = "A", group_b = "B",
87            lower_bound = -3, upper_bound = 3, max_rows = 100L, max_cells = 1000L,
88            alpha = 0.1),
89       list("/difference_a_minus_b" = tol(effect, "A-B mean"),
90            "/p_lower" = prob(stats::pt((effect + 3) / se, df, lower.tail = FALSE), "Welch lower TOST")),
91       "axis_coverage", list("Non-default alpha and bounded source/cell limits do not alter Welch estimators."))

Whole script generators/r/hypothesis_completion.R


Additional configuration coverage: STAT-HYP-NUMERIC-EQUIVEXT-001

Case STAT-HYP-NUMERIC-EQUIVEXT-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats/multcomp
Citation
R stats::pnorm, pt, qt, binom.test, power.t.test; multcomp::glht and mvtnorm documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, multcomp 1.4.32, mvtnorm 1.4.2, jsonlite 2.0.0
Reference script
generators/r/hypothesis_completion.R · sha256 733ecb7b7daec34d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/estimatetolerance0.812500000000000110.81250000000000021.12E-1615passed
/p_lowertail_probability1.6227244268166216e-051.6227244268166257e-054.08E-2014passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 93  delta <- d$x - 6
 94  one_effect <- mean(delta)
 95  one_se <- sd(delta) / sqrt(length(delta))
 96  emit("STAT-HYP-NUMERIC-EQUIVEXT-001", "tabelz.executor.equivalence_extensions",
 97       "equivalenceMenuItem", "HYP_CORE",
 98       list(mode = "one_sample_difference", response_column = "x", target = 6,
 99            lower_bound = -2, upper_bound = 2, max_rows = 100L, max_cells = 1000L,
100            alpha = 0.1),
101       list("/estimate" = tol(one_effect, "mean-target effect"),
102            "/p_lower" = prob(stats::pt((one_effect + 2) / one_se, length(delta) - 1,
103                                         lower.tail = FALSE), "one-sample lower TOST")),
104       "axis_coverage", list("Non-default alpha and bounded source/cell limits do not alter one-sample estimators."))

Whole script generators/r/hypothesis_completion.R


Additional configuration coverage: STAT-HYP-NUMERIC-POWER-TARGET-001

Case STAT-HYP-NUMERIC-POWER-TARGET-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats/multcomp
Citation
R stats::pnorm, pt, qt, binom.test, power.t.test; multcomp::glht and mvtnorm documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, multcomp 1.4.32, mvtnorm 1.4.2, jsonlite 2.0.0
Reference script
generators/r/hypothesis_completion.R · sha256 733ecb7b7daec34d
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valueexact38380.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

153  target <- 0.85
154  required <- stats::power.t.test(power = target, delta = 0.5, sd = 1,
155                                  sig.level = 0.05, type = "one.sample",
156                                  alternative = "two.sided", strict = TRUE)
157  emit("STAT-HYP-NUMERIC-POWER-TARGET-001",
158       "tabelz.analysis.hypothesis_tests:run_power_analysis", "hypothesisTestsMenuItem",
159       NULL,
160       list(analysis = "one_sample_t", solve_for = "sample_size", effect_size = 0.5,
161            target_power = target),
162       list("/value" = exact(as.integer(ceiling(required$n)), "Round R solved n upward.")),
163       "axis_coverage", list("Target power 0.85 is non-default; strict two-sided noncentral t solution."))

Whole script generators/r/hypothesis_completion.R


One-way classic ANOVA

Case STAT-HYP-ONE-ANOVA-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance3.03871814502898693.03871814502898373.16E-1514passed
/p_valuetail_probability0.103209612875833130.103209612875833228.76E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

124  anova <- stats::aov(x ~ g, data = d)
125  anova_row <- summary(anova)[[1]][1, ]
126  emit("STAT-HYP-ONE-ANOVA-001", "One-way classic ANOVA", "nominal",
127       list(test = "one_way_anova", response = "x", group = "g", method = "classic", posthoc = FALSE),
128       list("/statistic" = tol(anova_row[["F value"]], "one-way ANOVA F"),
129            "/p_value" = prob(anova_row[["Pr(>F)"]], "one-way ANOVA F tail")),
130       list("Classic fixed-effects one-way ANOVA; posthoc disabled."))

Whole script generators/r/hypothesis_tests.R


Welch one-way ANOVA

Case STAT-HYP-ONE-ANOVA-WELCH-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and car
Citation
R Core Team, stats help: oneway.test, chisq.test, binom.test, wilcox.test, aov and kruskal.test; car::Anova and car::leveneTest help.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/hypothesis_methods.R · sha256 e2faa3576983ee8d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance3.03871814502898113.03871814502898372.64E-1515passed
/p_valuetail_probability0.103210460491396040.10321046049139582.37E-1614passed
/df2tolerance13.99945589410847413.9994558941084751.46E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

37  welch_aov <- stats::oneway.test(x ~ g, data = d, var.equal = FALSE)
38  emit("STAT-HYP-ONE-ANOVA-WELCH-001", "Welch one-way ANOVA",
39       list(test = "one_way_anova", response = "x", group = "g", method = "welch", posthoc = FALSE),
40       list("/statistic" = tol(welch_aov$statistic, "Welch ANOVA F"),
41            "/p_value" = prob(welch_aov$p.value, "Welch ANOVA F tail"),
42            "/df2" = tol(welch_aov$parameter[[2]], "Welch denominator df")),
43       list("R oneway.test uses unequal-variance Welch correction."))

Whole script generators/r/hypothesis_methods.R


Exact one-proportion test

Case STAT-HYP-ONE-PROP-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance88.00.00E+2passed
/p_valuetail_probability0.45050403163340780.45050403163340812.75E-1615passed
/estimatetolerance0.50.50.00E+1passed
/ci_lowertolerance0.246510111490575330.24651011149057532.34E-1716passed
/ci_uppertolerance0.753489888509424670.75348988850942474.34E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

104  binomial <- stats::binom.test(sum(d$outcome == "yes"), length(d$outcome), p = 0.4)
105  emit("STAT-HYP-ONE-PROP-001", "Exact one-proportion test", "nominal",
106       list(test = "one_proportion", response = "outcome", success_value = "yes", target = 0.4),
107       list("/statistic" = tol(sum(d$outcome == "yes"), "success count"),
108            "/p_value" = prob(binomial$p.value, "binomial exact"),
109            "/estimate" = tol(unname(binomial$estimate), "success proportion"),
110            "/ci_lower" = tol(binomial$conf.int[[1]], "Clopper-Pearson lower"),
111            "/ci_upper" = tol(binomial$conf.int[[2]], "Clopper-Pearson upper")),
112       list("binom.test exact two-sided p and Clopper-Pearson interval."))

Whole script generators/r/hypothesis_tests.R


One-sample t two-sided

Case STAT-HYP-ONE-T-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.27211950113300441.2721195011330061.58E-1514passed
/p_valuetail_probability0.22269900194912520.22269900194912474.97E-1614passed
/estimatetolerance0.612499999999999820.61250000000000078.91E-1614passed
/dftolerance1515.00.00E+2passed
/ci_lowertolerance-0.41375016399215347-0.413750163992153034.39E-1614passed
/ci_uppertolerance1.63875016399215311.63875016399215451.35E-1515passed

Convention

Reproduce it yourself

  1. Data: data.csv, the rows as analysed.
  2. R: run the script below; it prints every compared value with its output path.
  3. HypoNum: follow the steps below and read the same values from the result window.

R script

# One-sample t test against a target mean. Base R only; tested with R 4.6.1.
report <- function(path, x) cat(path, "=", if (is.numeric(x)) sprintf("%.17g", x) else as.character(x), "\n")

d <- read.csv("data.csv")                       # column x (thickness)
target <- 6.2
fit <- t.test(d$x, mu = target)                 # two-sided, 95 % interval

report("/estimate",  mean(d$x) - target)         # mean difference from target
report("/statistic", fit$statistic)
report("/df",        fit$parameter)
report("/p_value",   fit$p.value)
report("/ci_lower",  fit$conf.int[1] - target)   # interval for the difference
report("/ci_upper",  fit$conf.int[2] - target)

HypoNum

  1. File > Import data, select data.csv, click Import.
  2. Statistics > Compare means & ANOVA > Hypothesis tests.
  3. Test: One-sample t-test. Response: x. Target: 6.2. Click Run test.

Read the mean difference from target, its 95 % interval, t, the degrees of freedom and the p-value from the result window.

Reference code

Whole script generators/r/hypothesis_tests.R


One-sample t less

Case STAT-HYP-ONE-T-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.27211950113300441.2721195011330061.58E-1514passed
/p_valuetail_probability0.888650499025437360.88865049902543773.30E-1615passed
/estimatetolerance0.612499999999999820.61250000000000078.91E-1614passed
/dftolerance1515.00.00E+2passed
/ci_uppertolerance1.45655855102871891.45655855102871966.61E-1615passed

Convention

Reference code

Whole script generators/r/hypothesis_tests.R


One-sample t greater

Case STAT-HYP-ONE-T-003 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.27211950113300441.2721195011330061.58E-1514passed
/p_valuetail_probability0.11134950097456260.111349500974562352.48E-1614passed
/estimatetolerance0.612499999999999820.61250000000000078.91E-1614passed
/dftolerance1515.00.00E+2passed
/ci_lowertolerance-0.23155855102871925-0.231558551028718141.11E-1514passed

Convention

Reference code

Whole script generators/r/hypothesis_tests.R


R sleep group 1 one-sample t-test against zero

Case STAT-HYP-ONESAMPLE-T-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::t.test
Citation
Cushny AR, Peebles AR (1905). The Action of Optical Isomers: II. Hyoscines. The Journal of Physiology 32(5-6), 501-510. doi:10.1113/jphysiol.1905.sp001097
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/smoke.R · sha256 3c019e985f24b0aa
Dataset
reference_data/datasets/SLEEP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1.32571014071382121.32571014071382125.69E-1817passed
/dftolerance99.00.00E+2passed
/p_valuetail_probability0.21759778006844890.217597780068448882.39E-1715passed
/ci_lowertolerance-0.52978041352623217-0.52978041352623183.29E-1615passed
/ci_uppertolerance2.02978041352623212.02978041352623164.81E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 3  data("sleep", package = "datasets")
 4  citation <- paste(
 5    "Cushny AR, Peebles AR (1905). The Action of Optical Isomers: II. Hyoscines.",
 6    "The Journal of Physiology 32(5-6), 501-510. doi:10.1113/jphysiol.1905.sp001097"
 7  )
 8  dataset <- data.frame(extra = as.double(sleep$extra), group = as.character(sleep$group))
 9  export_dataset(
10    dataset, "SLEEP", "R 4.6.1 datasets::sleep", citation,
11    "Published data distributed with R; retain the citation.",
12    c(extra = "Float64", group = "String")
13  )
14  reloaded <- read_dataset("SLEEP")
15  group_one <- reloaded$extra[reloaded$group == "1"]
16  reference <- stats::t.test(group_one, mu = 0, alternative = "two.sided", conf.level = 0.95)
17  case_id <- "STAT-HYP-ONESAMPLE-T-001"

Whole script generators/r/smoke.R


Paired differences

Case STAT-HYP-PAIRED-T-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance13.82164012876388213.8216401287638671.45E-1414passed
/p_valuetail_probability6.1259435838448321e-106.125943583844918e-108.58E-2413passed
/estimatetolerance0.562500000000000220.56250000000000022.04E-1817passed

Convention

Reproduce it yourself

  1. Data: data.csv, the rows as analysed.
  2. R: run the script below; it prints every compared value with its output path.
  3. HypoNum: follow the steps below and read the same values from the result window.

R script

# Paired t test. Base R only; tested with R 4.6.1.
report <- function(path, x) cat(path, "=", if (is.numeric(x)) sprintf("%.17g", x) else as.character(x), "\n")

d <- read.csv("data.csv")                       # columns x and y, one pair per row
fit <- t.test(d$x, d$y, paired = TRUE)          # two-sided

report("/estimate",  mean(d$x - d$y))            # mean of the x - y differences
report("/statistic", fit$statistic)
report("/p_value",   fit$p.value)

HypoNum

  1. File > Import data, select data.csv, click Import.
  2. Statistics > Compare means & ANOVA > Hypothesis tests.
  3. Test: Paired t-test. Response: x. Paired response: y. Click Run test.

Read the mean difference, t and the p-value from the result window.

Reference code

The lines of the reference script that produce this case's expected values.

 96  paired <- stats::t.test(d$x, d$y, paired = TRUE)
 97  emit("STAT-HYP-PAIRED-T-001", "Paired differences", "nominal",
 98       list(test = "paired_t", response = "x", second_response = "y"),
 99       list("/statistic" = tol(paired$statistic, "paired t"),
100            "/p_value" = prob(paired$p.value, "paired t"),
101            "/estimate" = tol(mean(d$x - d$y), "paired mean difference")),
102       list("Pair rows by exported CSV order; difference is x minus y."))

Whole script generators/r/hypothesis_tests.R


Exact independent mean permutation

Case STAT-HYP-PERM-EXACT-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 independent implementation
Citation
R utils::combn documentation; exact finite-label randomization enumeration.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_secondary.R · sha256 13b3dbafe11fd4e1
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-1.5749999999999993-1.57500000000000028.78E-1615passed
/p_valuetail_probability0.10660450660450660.10660450660450667.90E-1917passed
/assignment_countexact12870128700.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

119  pooled <- d$x
120  n_a <- length(a)
121  observed <- diff
122  all_a <- utils::combn(seq_along(pooled), n_a)
123  total_sum <- sum(pooled)
124  assigned_a <- colSums(matrix(pooled[all_a], nrow = n_a))
125  null <- assigned_a / n_a - (total_sum - assigned_a) / (length(pooled) - n_a)
126  p_exact <- mean(abs(null) >= abs(observed) - 1e-12)
127  emit("STAT-HYP-PERM-EXACT-001", "tabelz.executor.permutation_tests", "permutationTestsMenuItem",
128       "Exact independent mean permutation", "nominal", "HYP_CORE",
129       list(mode = "independent", response_column = "x", group_column = "g",
130            group_a = "A", group_b = "B", inference = "exact"),
131       list("/statistic" = tol(observed, "observed A-B mean"),
132            "/p_value" = prob(p_exact, "all choose(16,8) assignments"),
133            "/assignment_count" = exact(ncol(all_a), "All independent assignments enumerated.")),
134       list("combn enumerates every allocation of 8 labels to 16 values; two-sided absolute statistic."),
135       "R utils::combn documentation; exact finite-label randomization enumeration.")

Whole script generators/r/hypothesis_secondary.R


Prospective power axis: STAT-HYP-POWER-ALT-GREATER-001

Case STAT-HYP-POWER-ALT-GREATER-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test and pwr::pwr.chisq.test / pwr.2p2n.test / pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_power_axes.R · sha256 3116c4f725f974d6
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.8482541909291470.84825419092898241.65E-1312passed

Convention

Reference code

Whole script generators/r/hypothesis_power_axes.R


Prospective power axis: STAT-HYP-POWER-ALT-LESS-001

Case STAT-HYP-POWER-ALT-LESS-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test and pwr::pwr.chisq.test / pwr.2p2n.test / pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_power_axes.R · sha256 3116c4f725f974d6
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.8482541909291470.84825419092898241.65E-1312passed

Convention

Reference code

Whole script generators/r/hypothesis_power_axes.R


Prospective power axis: STAT-HYP-POWER-ANOVA-001

Case STAT-HYP-POWER-ANOVA-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test and pwr::pwr.chisq.test / pwr.2p2n.test / pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_power_axes.R · sha256 3116c4f725f974d6
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.635239843591377860.63523984342412891.67E-109passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

85  anova <- pwr::pwr.anova.test(k = 3, n = 10, f = 0.5, sig.level = 0.05)
86  emit("STAT-HYP-POWER-ANOVA-001",
87       list(analysis = "one_way_anova", solve_for = "power", effect_size = 0.5,
88            sample_size = 30L, groups = 3L),
89       tol(anova$power, "three-group ANOVA noncentral F power"),
90       list("Cohen f=0.5, three groups with ten observations each; N=30."))

Whole script generators/r/hypothesis_power_axes.R


Prospective power axis: STAT-HYP-POWER-ONE-PROP-001

Case STAT-HYP-POWER-ONE-PROP-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test and pwr::pwr.chisq.test / pwr.2p2n.test / pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_power_axes.R · sha256 3116c4f725f974d6
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.78190799873197980.78190799873197971.08E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

70  one_prop <- pwr::pwr.chisq.test(w = 0.5, N = 30, df = 1, sig.level = 0.05)
71  emit("STAT-HYP-POWER-ONE-PROP-001",
72       list(analysis = "one_proportion", solve_for = "power", effect_size = 0.5,
73            sample_size = 30L),
74       tol(one_prop$power, "one-degree chi-square power"),
75       list("Cohen h=0.5 enters the one-degree chi-square approximation as w=0.5."))

Whole script generators/r/hypothesis_power_axes.R


One-sample t achieved power

Case STAT-HYP-POWER-ONE-T-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 independent implementation
Citation
R stats::power.t.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_secondary.R · sha256 13b3dbafe11fd4e1
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.753964715743851930.75396471574378996.21E-1413passed
/achieved_powertolerance0.753964715743851930.75396471574378996.21E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

177  power <- stats::power.t.test(n = 30, delta = 0.5, sd = 1, sig.level = 0.05,
178                               type = "one.sample", alternative = "two.sided", strict = TRUE)
179  emit("STAT-HYP-POWER-ONE-T-001", "tabelz.analysis.hypothesis_tests:run_power_analysis",
180       "hypothesisTestsMenuItem", "One-sample t achieved power", "nominal", NULL,
181       list(analysis = "one_sample_t", solve_for = "power", effect_size = 0.5,
182            sample_size = 30L, alpha = 0.05),
183       list("/value" = tol(power$power, "R noncentral-t achieved power"),
184            "/achieved_power" = tol(power$power, "R noncentral-t achieved power")),
185       list("R power.t.test uses standardized delta/sd = 0.5 and n=30.",
186            "strict=TRUE counts both rejection tails. docs/manual/chapters/07-models-and-tests.tex:864-866: 'two-sided alternatives.'"),
187       "R stats::power.t.test documentation.")

Whole script generators/r/hypothesis_secondary.R


Prospective power axis: STAT-HYP-POWER-PAIRED-001

Case STAT-HYP-POWER-PAIRED-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test and pwr::pwr.chisq.test / pwr.2p2n.test / pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_power_axes.R · sha256 3116c4f725f974d6
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.753964715743851930.75396471574378996.21E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

49  pair <- stats::power.t.test(n = 30, delta = 0.5, sd = 1, sig.level = 0.05,
50                              type = "paired", alternative = "two.sided", strict = TRUE)
51  emit("STAT-HYP-POWER-PAIRED-001",
52       list(analysis = "paired_t", solve_for = "power", effect_size = 0.5, sample_size = 30L),
53       tol(pair$power, "paired t power for 30 pairs"),
54       list("Cohen dz = 0.5, n is number of pairs; strict two-sided t rejection."))

Whole script generators/r/hypothesis_power_axes.R


Prospective power axis: STAT-HYP-POWER-SAMPLE-SIZE-001

Case STAT-HYP-POWER-SAMPLE-SIZE-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test and pwr::pwr.chisq.test / pwr.2p2n.test / pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_power_axes.R · sha256 3116c4f725f974d6
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valueexact34340.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

30  sample <- stats::power.t.test(power = 0.8, delta = 0.5, sd = 1, sig.level = 0.05,
31                                type = "one.sample", alternative = "two.sided", strict = TRUE)
32  emit("STAT-HYP-POWER-SAMPLE-SIZE-001",
33       list(analysis = "one_sample_t", solve_for = "sample_size", effect_size = 0.5,
34            target_power = 0.8, alpha = 0.05),
35       exact(ceiling(sample$n), "Smallest integer n meeting target power from R noncentral t."),
36       list("R power.t.test strict=TRUE includes both two-sided rejection tails; round solved n upward."))

Whole script generators/r/hypothesis_power_axes.R


Prospective power axis: STAT-HYP-POWER-TWO-PROP-001

Case STAT-HYP-POWER-TWO-PROP-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test and pwr::pwr.chisq.test / pwr.2p2n.test / pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_power_axes.R · sha256 3116c4f725f974d6
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.490685567669969180.49068556766996892.77E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

77  two_prop <- pwr::pwr.2p2n.test(h = 0.5, n1 = 30, n2 = 30, sig.level = 0.05,
78                                  alternative = "two.sided")
79  emit("STAT-HYP-POWER-TWO-PROP-001",
80       list(analysis = "two_proportion", solve_for = "power", effect_size = 0.5,
81            sample_size = 30L, ratio = 1),
82       tol(two_prop$power, "two-independent-proportion normal power"),
83       list("Cohen h=0.5 and 30 observations per group; two-sided normal approximation."))

Whole script generators/r/hypothesis_power_axes.R


Prospective power axis: STAT-HYP-POWER-TWO-T-001

Case STAT-HYP-POWER-TWO-T-001 · status: passed

Reference

Evidence type
R
Source
R stats and pwr
Citation
R stats::power.t.test and pwr::pwr.chisq.test / pwr.2p2n.test / pwr.anova.test documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, pwr 1.3.0, jsonlite 2.0.0
Reference script
generators/r/hypothesis_power_axes.R · sha256 3116c4f725f974d6
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/valuetolerance0.553086583494863730.55308658349466142.02E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

57  n1 <- 30
58  n2 <- ceiling(1.5 * n1)
59  df <- n1 + n2 - 2
60  ncp <- 0.5 / sqrt(1 / n1 + 1 / n2)
61  critical <- stats::qt(0.975, df)
62  two_t_power <- stats::pt(-critical, df, ncp = ncp) +
63    stats::pt(critical, df, ncp = ncp, lower.tail = FALSE)
64  emit("STAT-HYP-POWER-TWO-T-001",
65       list(analysis = "two_sample_t", solve_for = "power", effect_size = 0.5,
66            sample_size = n1, ratio = 1.5),
67       tol(two_t_power, "unequal-allocation independent t noncentral tail"),
68       list("n1=30, n2=ceil(1.5*n1)=45; df=n1+n2-2; noncentrality d/sqrt(1/n1+1/n2)."))

Whole script generators/r/hypothesis_power_axes.R


Repeated measures one within factor three levels

Case STAT-HYP-REPEATED-MEASURES-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R Core Team, stats package help: aov, anova.mlm, mauchly.test, SSD, t.test; Greenhouse SW, Geisser S (1959), Psychometrika 24, 95-112; Lecoutre B (1991), Journal of Educational Statistics 16, 371-372.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_repeated_measures.R · sha256 8faee62dc29c9cba
Dataset
reference_data/datasets/HYP_RM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sphericity/0/dimensionsexact220.00E+2passed
/sphericity/0/greenhouse_geissertolerance0.6487915036290810.64879150362908261.56E-1514passed
/sphericity/0/huynh_feldttolerance0.721996786816173630.72199678681617622.55E-1514passed
/sphericity/0/huynh_feldt_appliedtolerance0.721996786816173630.72199678681617622.55E-1514passed
/sphericity/0/lower_boundtolerance0.50.50.00E+1passed
/sphericity/0/mauchly_wtolerance0.458672787164445320.45867278716444923.88E-1514passed
/sphericity/0/chi_squaretolerance5.45592743596222365.4559274359621655.86E-1413passed
/sphericity/0/dfexact220.00E+2passed
/sphericity/0/p_valuetail_probability0.0653522298408693160.065352229840871231.92E-1513passed
/within_tests/0/sum_of_squarestolerance17.02740740740741617.0274074074073457.06E-1414passed
/within_tests/0/dfexact220.00E+2passed
/within_tests/0/mean_squaretolerance8.51370370370370828.5137037037036733.55E-1414passed
/within_tests/0/f_valuetolerance41.96622546782285941.9662254678227241.35E-1314passed
/within_tests/0/error_sum_of_squarestolerance3.24592592592593173.24592592592592853.15E-1515passed
/within_tests/0/error_dfexact16160.00E+2passed
/within_tests/0/partial_eta_squaredtolerance0.839891848441667420.8398918484416674.43E-1615passed
/within_tests/0/tests/0/correctionexactSphericity assumedSphericity assumedpassed
/within_tests/0/tests/0/epsilontolerance11.00.00E+2passed
/within_tests/0/tests/0/df1tolerance22.00.00E+2passed
/within_tests/0/tests/0/df2tolerance1616.00.00E+2passed
/within_tests/0/tests/0/p_valuetail_probability4.3182476875850561e-074.3182476875851525e-079.64E-2113passed
/within_tests/0/tests/1/correctionexactGreenhouse-GeisserGreenhouse-Geisserpassed
/within_tests/0/tests/1/epsilontolerance0.6487915036290810.64879150362908261.56E-1514passed
/within_tests/0/tests/1/df1tolerance1.2975830072581621.2975830072581653.11E-1514passed
/within_tests/0/tests/1/df2tolerance10.38066405806529610.380664058065322.49E-1414passed
/within_tests/0/tests/1/p_valuetail_probability3.0901261916248958e-053.090126191624883e-051.29E-1914passed
/within_tests/0/tests/2/correctionexactHuynh-FeldtHuynh-Feldtpassed
/within_tests/0/tests/2/epsilontolerance0.721996786816173630.72199678681617622.55E-1514passed
/within_tests/0/tests/2/df1tolerance1.44399357363234731.44399357363235245.07E-1514passed
/within_tests/0/tests/2/df2tolerance11.55194858905877811.5519485890588194.09E-1414passed
/within_tests/0/tests/2/p_valuetail_probability1.2628833876660175e-051.2628833876659985e-051.90E-1913passed
/multivariate_tests/0/statisticexactPillai's tracePillai's tracepassed
/multivariate_tests/0/valuetolerance0.974250233420545510.97425023342139758.52E-1312passed
/multivariate_tests/0/f_valuetolerance132.42356222726377132.42356223176114.50E-910passed
/multivariate_tests/0/df1tolerance22.00.00E+2passed
/multivariate_tests/0/df2tolerance77.00.00E+2passed
/multivariate_tests/0/p_valuetail_probability2.7397222569971624e-062.7397222566798884e-063.17E-169passed
/multivariate_tests/1/statisticexactWilks' lambdaWilks' lambdapassed
/multivariate_tests/1/valuetolerance0.025749766579458780.0257497665786026668.56E-1310passed
/multivariate_tests/1/f_valuetolerance132.42356222724115132.423562231760264.52E-910passed
/multivariate_tests/1/df1tolerance22.00.00E+2passed
/multivariate_tests/1/df2tolerance77.00.00E+2passed
/multivariate_tests/1/p_valuetail_probability2.7397222569987586e-062.739722256679948e-063.19E-169passed
/multivariate_tests/2/statisticexactHotelling-Lawley traceHotelling-Lawley tracepassed
/multivariate_tests/2/valuetolerance37.83530349349730737.835303494788651.29E-910passed
/multivariate_tests/2/f_valuetolerance132.42356222724058132.423562231760264.52E-910passed
/multivariate_tests/2/df1tolerance22.00.00E+2passed
/multivariate_tests/2/df2tolerance77.00.00E+2passed
/multivariate_tests/2/p_valuetail_probability2.7397222569987976e-062.739722256679948e-063.19E-169passed
/multivariate_tests/3/statisticexactRoy's rootRoy's rootpassed
/multivariate_tests/3/valuetolerance37.835303493497337.835303494788651.29E-910passed
/multivariate_tests/3/f_valuetolerance132.42356222724055132.423562231760264.52E-910passed
/multivariate_tests/3/df1tolerance22.00.00E+2passed
/multivariate_tests/3/df2tolerance77.00.00E+2passed
/multivariate_tests/3/p_valuetail_probability2.7397222569988022e-062.739722256679948e-063.19E-169passed
/within_tests/0/effectexactTimeTimepassed
/between_tests/0/effectexact(Intercept)(Intercept)passed
/between_tests/0/sum_of_squarestolerance4505.97925925925944505.9792592592592.08E-1417passed
/between_tests/0/f_valuetolerance545.12698270454348545.12698270454332.28E-1315passed
/between_tests/0/p_valuetail_probability1.2035017877443359e-081.2035017877443376e-081.69E-2314passed
/between_error_sstolerance66.12740740740738966.127407407407434.31E-1415passed
/design/subjectsexact10100.00E+2passed
/design/usedexact990.00E+2passed
/design/incompleteexact110.00E+2passed
/design/residual_dfexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

46  emit <- function(id, title, kind, config, outputs, notes, excluded = list()) {
47    write_expected("hypothesis", id, list(
48      schema_version = "2.0.0", case_id = id, outputs = outputs, expected_error = NULL,
49      deferred_outputs = list(), generator_run = run))
50    write_case("hypothesis", id, list(
51      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "hypothesis",
52      entry_point = entry, menu_items = list(menu), title = title, case_kind = kind,
53      dataset = list(path = "reference_data/datasets/HYP_RM.json", sha256 = ""),
54      excluded_rows = excluded, configuration = config, "repeat" = 1L,
55      expected = list(path = paste0("reference_data/expected/hypothesis/", id, ".json"), sha256 = ""),
56      oracle = list(oracle_class = "independent_implementation", source = "R 4.6.1 stats",
57                    citation = "R Core Team, stats package help: aov, anova.mlm, mauchly.test, SSD, t.test; Greenhouse SW, Geisser S (1959), Psychometrika 24, 95-112; Lecoutre B (1991), Journal of Educational Statistics 16, 371-372.",
58                    generator = list(path = "generators/r/hypothesis_repeated_measures.R", sha256 = ""),
59                    independent = TRUE,
60                    convention_notes = c(notes, notes_common, "Expected computed from the exported CSV.")),
61      supported = TRUE, unsupported_reason = NULL,
62      requirement_ids = list("REQ-STAT-HYP-REPEATED-MEASURES-001"),
63      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/hypothesis_repeated_measures.R


Repeated measures with a balanced between-subject factor

Case STAT-HYP-REPEATED-MEASURES-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R Core Team, stats package help: aov, anova.mlm, mauchly.test, SSD, t.test; Greenhouse SW, Geisser S (1959), Psychometrika 24, 95-112; Lecoutre B (1991), Journal of Educational Statistics 16, 371-372.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_repeated_measures.R · sha256 8faee62dc29c9cba
Dataset
reference_data/datasets/HYP_RM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sphericity/0/dimensionsexact220.00E+2passed
/sphericity/0/greenhouse_geissertolerance0.580560071159095140.58056007115909492.25E-1615passed
/sphericity/0/huynh_feldttolerance0.633187966683991350.63318796668399085.51E-1615passed
/sphericity/0/huynh_feldt_appliedtolerance0.633187966683991350.63318796668399085.51E-1615passed
/sphericity/0/lower_boundtolerance0.50.50.00E+1passed
/sphericity/0/mauchly_wtolerance0.277525359256126810.277525359256125761.05E-1514passed
/sphericity/0/chi_squaretolerance6.40921482632106276.4092148263210811.86E-1414passed
/sphericity/0/dfexact220.00E+2passed
/sphericity/0/p_valuetail_probability0.0405748276548461020.040574827654845723.82E-1614passed
/within_tests/0/sum_of_squarestolerance16.74333333333332516.7433333333333327.29E-1515passed
/within_tests/0/dfexact220.00E+2passed
/within_tests/0/mean_squaretolerance8.37166666666666268.3716666666666663.55E-1515passed
/within_tests/0/f_valuetolerance41.39835164835157141.398351648351592.18E-1415passed
/within_tests/0/error_sum_of_squarestolerance2.42666666666666992.426666666666674.18E-1716passed
/within_tests/0/error_dfexact12120.00E+2passed
/within_tests/0/partial_eta_squaredtolerance0.873413319422708940.8734133194227094.73E-1817passed
/within_tests/0/tests/0/correctionexactSphericity assumedSphericity assumedpassed
/within_tests/0/tests/0/epsilontolerance11.00.00E+2passed
/within_tests/0/tests/0/df1tolerance22.00.00E+2passed
/within_tests/0/tests/0/df2tolerance1212.00.00E+2passed
/within_tests/0/tests/0/p_valuetail_probability4.1146042495770505e-064.114604249577034e-061.69E-2014passed
/within_tests/0/tests/1/correctionexactGreenhouse-GeisserGreenhouse-Geisserpassed
/within_tests/0/tests/1/epsilontolerance0.580560071159095140.58056007115909492.25E-1615passed
/within_tests/0/tests/1/df1tolerance1.16112014231819031.16112014231818984.70E-1615passed
/within_tests/0/tests/1/df2tolerance6.96672085390914126.9667208539091391.78E-1515passed
/within_tests/0/tests/1/p_valuetail_probability0.000290817112496091890.00029081711249609234.30E-1914passed
/within_tests/0/tests/2/correctionexactHuynh-FeldtHuynh-Feldtpassed
/within_tests/0/tests/2/epsilontolerance0.633187966683991350.63318796668399085.51E-1615passed
/within_tests/0/tests/2/df1tolerance1.26637593336798271.26637593336798161.10E-1515passed
/within_tests/0/tests/2/df2tolerance7.59825560020789587.598255600207896.21E-1515passed
/within_tests/0/tests/2/p_valuetail_probability0.000169600579358681960.000169600579358682565.99E-1914passed
/within_tests/1/sum_of_squarestolerance0.123333333333332370.123333333333333891.52E-1513passed
/within_tests/1/dfexact220.00E+2passed
/within_tests/1/mean_squaretolerance0.0616666666666661830.0616666666666669467.63E-1613passed
/within_tests/1/f_valuetolerance0.304945054945052140.30494505494505593.78E-1513passed
/within_tests/1/error_sum_of_squarestolerance2.42666666666666992.426666666666674.18E-1716passed
/within_tests/1/error_dfexact12120.00E+2passed
/within_tests/1/partial_eta_squaredtolerance0.0483660130718950030.048366013071895575.69E-1613passed
/within_tests/1/tests/0/correctionexactSphericity assumedSphericity assumedpassed
/within_tests/1/tests/0/epsilontolerance11.00.00E+2passed
/within_tests/1/tests/0/df1tolerance22.00.00E+2passed
/within_tests/1/tests/0/df2tolerance1212.00.00E+2passed
/within_tests/1/tests/0/p_valuetail_probability0.742710673067678130.74271067306767542.78E-1514passed
/within_tests/1/tests/1/correctionexactGreenhouse-GeisserGreenhouse-Geisserpassed
/within_tests/1/tests/1/epsilontolerance0.580560071159095140.58056007115909492.25E-1615passed
/within_tests/1/tests/1/df1tolerance1.16112014231819031.16112014231818984.70E-1615passed
/within_tests/1/tests/1/df2tolerance6.96672085390914126.9667208539091391.78E-1515passed
/within_tests/1/tests/1/p_valuetail_probability0.631296266232213580.63129626623221122.33E-1514passed
/within_tests/1/tests/2/correctionexactHuynh-FeldtHuynh-Feldtpassed
/within_tests/1/tests/2/epsilontolerance0.633187966683991350.63318796668399085.51E-1615passed
/within_tests/1/tests/2/df1tolerance1.26637593336798271.26637593336798161.10E-1515passed
/within_tests/1/tests/2/df2tolerance7.59825560020789587.598255600207896.21E-1515passed
/within_tests/1/tests/2/p_valuetail_probability0.649140146175273490.64914014617527082.66E-1514passed
/multivariate_tests/0/statisticexactPillai's tracePillai's tracepassed
/multivariate_tests/0/valuetolerance0.988997168725307190.98899716872563943.32E-1312passed
/multivariate_tests/0/f_valuetolerance224.71424491441152224.71424492127126.86E-910passed
/multivariate_tests/0/df1tolerance22.00.00E+2passed
/multivariate_tests/0/df2tolerance55.00.00E+2passed
/multivariate_tests/0/p_valuetail_probability1.2698754670669953e-051.2698754669711511e-059.58E-1610passed
/multivariate_tests/1/statisticexactWilks' lambdaWilks' lambdapassed
/multivariate_tests/1/valuetolerance0.0110028312746612390.0110028312743606263.01E-1310passed
/multivariate_tests/1/f_valuetolerance224.71424491506357224.714244921271426.21E-910passed
/multivariate_tests/1/df1tolerance22.00.00E+2passed
/multivariate_tests/1/df2tolerance55.00.00E+2passed
/multivariate_tests/1/p_valuetail_probability1.2698754670578864e-051.269875466971148e-058.67E-1610passed
/multivariate_tests/2/statisticexactHotelling-Lawley traceHotelling-Lawley tracepassed
/multivariate_tests/2/valuetolerance89.88569796602831789.885697968508572.48E-910passed
/multivariate_tests/2/f_valuetolerance224.71424491507079224.714244921271426.20E-910passed
/multivariate_tests/2/df1tolerance22.00.00E+2passed
/multivariate_tests/2/df2tolerance55.00.00E+2passed
/multivariate_tests/2/p_valuetail_probability1.2698754670577872e-051.269875466971148e-058.66E-1610passed
/multivariate_tests/3/statisticexactRoy's rootRoy's rootpassed
/multivariate_tests/3/valuetolerance89.88569796602834689.885697968508572.48E-910passed
/multivariate_tests/3/f_valuetolerance224.71424491507088224.714244921271426.20E-910passed
/multivariate_tests/3/df1tolerance22.00.00E+2passed
/multivariate_tests/3/df2tolerance55.00.00E+2passed
/multivariate_tests/3/p_valuetail_probability1.2698754670577826e-051.269875466971148e-058.66E-1610passed
/multivariate_tests/4/statisticexactPillai's tracePillai's tracepassed
/multivariate_tests/4/valuetolerance0.109100159915057280.109100159906964768.09E-1210passed
/multivariate_tests/4/f_valuetolerance0.306151586873826210.30615158684833642.55E-1110passed
/multivariate_tests/4/df1tolerance22.00.00E+2passed
/multivariate_tests/4/df2tolerance55.00.00E+2passed
/multivariate_tests/4/p_valuetail_probability0.749155897038556230.74915589705556871.70E-1110passed
/multivariate_tests/5/statisticexactWilks' lambdaWilks' lambdapassed
/multivariate_tests/5/valuetolerance0.89089984008494250.89089984009303548.09E-1211passed
/multivariate_tests/5/f_valuetolerance0.306151586873826820.30615158684833612.55E-1110passed
/multivariate_tests/5/df1tolerance22.00.00E+2passed
/multivariate_tests/5/df2tolerance55.00.00E+2passed
/multivariate_tests/5/p_valuetail_probability0.749155897038556340.7491558970555691.70E-1110passed
/multivariate_tests/6/statisticexactHotelling-Lawley traceHotelling-Lawley tracepassed
/multivariate_tests/6/valuetolerance0.122460634749530660.122460634739334551.02E-1110passed
/multivariate_tests/6/f_valuetolerance0.306151586873826650.30615158684833642.55E-1110passed
/multivariate_tests/6/df1tolerance22.00.00E+2passed
/multivariate_tests/6/df2tolerance55.00.00E+2passed
/multivariate_tests/6/p_valuetail_probability0.749155897038556010.74915589705556871.70E-1110passed
/multivariate_tests/7/statisticexactRoy's rootRoy's rootpassed
/multivariate_tests/7/valuetolerance0.122460634749531370.122460634739334541.02E-1110passed
/multivariate_tests/7/f_valuetolerance0.306151586873828430.30615158684833632.55E-1110passed
/multivariate_tests/7/df1tolerance22.00.00E+2passed
/multivariate_tests/7/df2tolerance55.00.00E+2passed
/multivariate_tests/7/p_valuetail_probability0.749155897038554790.74915589705556881.70E-1110passed
/within_tests/0/effectexactTimeTimepassed
/within_tests/1/effectexactTime × grpTime × grppassed
/between_tests/1/effectexactgrpgrppassed
/between_tests/1/sum_of_squarestolerance37.75041666666671637.750416666666727.13E-1515passed
/between_tests/1/dfexact110.00E+2passed
/between_tests/1/f_valuetolerance9.69374799386569929.6937479938656963.57E-1515passed
/between_tests/1/p_valuetail_probability0.0207566219241842390.0207566219241842632.42E-1714passed
/between_error_sstolerance23.36583333333333523.3658333333333451.02E-1415passed
/design/usedexact880.00E+2passed
/design/incompleteexact110.00E+2passed
/design/excludedexact110.00E+2passed
/design/residual_dfexact660.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

170  keep <- stats::complete.cases(d[c("t1", "t2", "t3")]) & seq_len(nrow(d)) != 5
171  cc <- d[keep, ]
172  Y <- as.matrix(cc[c("t1", "t2", "t3")])
173  n <- nrow(Y); p <- 3L
174  grp <- factor(cc$grp)
175  fit <- stats::lm(Y ~ grp, contrasts = list(grp = "contr.sum"))
176  nu <- n - 2L
177  long <- data.frame(y = as.vector(Y), subject = factor(rep(seq_len(n), p)), Time = factor(rep(1:p, each = n)), grp = rep(grp, p))
178  strata <- summary(stats::aov(y ~ grp * Time + Error(subject / Time), data = long))
179  within <- strata[["Error: subject:Time"]][[1]]
180  between <- strata[["Error: subject"]][[1]]
181  sph <- sphericity_outputs(0, fit, orthonormal(p), nu)
182  outputs <- c(sph$outputs,
183               univariate_outputs(0, within["Time", "Sum Sq"], within["Time", "Df"], within["Residuals", "Sum Sq"], within["Residuals", "Df"], sph$gg, sph$hf),
184               univariate_outputs(1, within["grp:Time", "Sum Sq"], within["grp:Time", "Df"], within["Residuals", "Sum Sq"], within["Residuals", "Df"], sph$gg, sph$hf),
185               multivariate_outputs(0, fit, ~1, ~Time, idata, "(Intercept)"),
186               multivariate_outputs(4, fit, ~1, ~Time, idata, "grp"))
187  rownames(within) <- trimws(rownames(within)); rownames(between) <- trimws(rownames(between))
188  outputs[["/within_tests/0/effect"]] <- word("Time", "effect name")
189  outputs[["/within_tests/1/effect"]] <- word("Time × grp", "interaction effect name")
190  total <- rowSums(Y) / sqrt(p)
191  bt <- stats::anova(stats::lm(total ~ grp))
192  outputs[["/between_tests/1/effect"]] <- word("grp", "between effect")
193  outputs[["/between_tests/1/sum_of_squares"]] <- tol(bt["grp", "Sum Sq"], "grp SS of the normalized subject total")
194  outputs[["/between_tests/1/df"]] <- exact(bt["grp", "Df"], "grp df")
195  outputs[["/between_tests/1/f_value"]] <- tol(bt["grp", "F value"], "grp F")
196  outputs[["/between_tests/1/p_value"]] <- prob(bt["grp", "Pr(>F)"], "grp F tail")
197  outputs[["/between_error_ss"]] <- tol(bt["Residuals", "Sum Sq"], "residual SS of the normalized subject total")
198  outputs[["/design/used"]] <- exact(n, "complete, non-excluded subjects")
199  outputs[["/design/incomplete"]] <- exact(1L, "subject 10 lacks t2")
200  outputs[["/design/excluded"]] <- exact(1L, "row 5 excluded")
201  outputs[["/design/residual_df"]] <- exact(nu, "N - rank(X) with the intercept and one sum-to-zero column")
202  emit("STAT-HYP-REPEATED-MEASURES-002", "Repeated measures with a balanced between-subject factor", "missing_excluded",
203       list(factor1_name = "Time", factor1_levels = list("t1", "t2", "t3"), cells = list("t1", "t2", "t3"),
204            between = list("grp"), contrast_family = "none", pairwise_procedure = "none"),
205       outputs, c("Within effects: aov(y ~ grp * Time + Error(subject / Time)) rows Time and grp:Time with the subject:Time residual; sphericity from the residual SSD of lm(Y ~ grp) on nu = 6 df.",
206                  "Multivariate tests: anova.mlm rows (Intercept) (= Time) and grp (= Time x grp) with X = ~1, M = ~Time; balanced groups make the sequential tests the Type III tests.",
207                  "Between tests on the normalized subject total: anova(lm(total ~ grp))."),
208       excluded = list(4L))

Whole script generators/r/hypothesis_repeated_measures.R


Repeated measures two occasions equals paired t squared

Case STAT-HYP-REPEATED-MEASURES-003 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R Core Team, stats package help: aov, anova.mlm, mauchly.test, SSD, t.test; Greenhouse SW, Geisser S (1959), Psychometrika 24, 95-112; Lecoutre B (1991), Journal of Educational Statistics 16, 371-372.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_repeated_measures.R · sha256 8faee62dc29c9cba
Dataset
reference_data/datasets/HYP_RM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sphericity/0/dimensionsexact110.00E+2passed
/sphericity/0/greenhouse_geissertolerance11.00.00E+2passed
/sphericity/0/mauchly_wabsentNonepassed
/within_tests/0/effectexactTimeTimepassed
/within_tests/0/sum_of_squarestolerance1.56055555555555461.56055555555555083.79E-1514passed
/within_tests/0/dfexact110.00E+2passed
/within_tests/0/error_sum_of_squarestolerance2.17444444444444682.17444444444444462.18E-1514passed
/within_tests/0/error_dfexact880.00E+2passed
/within_tests/0/f_valuetolerance5.74144098109350055.74144098109349257.95E-1514passed
/within_tests/0/tests/0/p_valuetail_probability0.0434382500910390070.043438250091039141.32E-1614passed
/within_tests/0/tests/1/epsilontolerance11.00.00E+2passed
/within_tests/0/tests/1/p_valuetail_probability0.0434382500910390070.043438250091039141.32E-1614passed
/multivariate_tests/0/f_valuetolerance5.74144098109350055.7414409810934919.73E-1514passed
/multivariate_tests/0/p_valuetail_probability0.0434382500910390070.043438250091039141.32E-1614passed
/multivariate_tests/1/p_valuetail_probability0.0434382500910390070.043438250091039141.32E-1614passed
/design/usedexact990.00E+2passed
/design/residual_dfexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

211  cc <- d[stats::complete.cases(d[c("t1", "t2")]), ]
212  n <- nrow(cc); nu <- n - 1L
213  tt <- stats::t.test(cc$t1, cc$t2, paired = TRUE)
214  score <- (cc$t2 - cc$t1) / sqrt(2)  # orthonormal contrast score
215  outputs <- list(
216    "/sphericity/0/dimensions" = exact(1L, "one contrast"),
217    "/sphericity/0/greenhouse_geisser" = tol(1, "epsilon is 1 with one contrast"),
218    "/sphericity/0/mauchly_w" = absent("no Mauchly test with one contrast"),
219    "/within_tests/0/effect" = word("Time", "effect name"),
220    "/within_tests/0/sum_of_squares" = tol(n * mean(score)^2, "n times the squared mean contrast score"),
221    "/within_tests/0/df" = exact(1L, "one contrast"),
222    "/within_tests/0/error_sum_of_squares" = tol(sum((score - mean(score))^2), "contrast score SS about its mean"),
223    "/within_tests/0/error_df" = exact(nu, "n - 1"),
224    "/within_tests/0/f_value" = tol(unname(tt$statistic)^2, "squared paired t"),
225    "/within_tests/0/tests/0/p_value" = prob(tt$p.value, "paired t two-sided tail"),
226    "/within_tests/0/tests/1/epsilon" = tol(1, "Greenhouse-Geisser epsilon 1"),
227    "/within_tests/0/tests/1/p_value" = prob(tt$p.value, "unchanged by epsilon 1"),
228    "/multivariate_tests/0/f_value" = tol(unname(tt$statistic)^2, "Pillai exact F with one contrast"),
229    "/multivariate_tests/0/p_value" = prob(tt$p.value, "exact F tail with one contrast"),
230    "/multivariate_tests/1/p_value" = prob(tt$p.value, "Wilks exact F tail with one contrast"),
231    "/design/used" = exact(n, "complete subjects"),
232    "/design/residual_df" = exact(nu, "n - 1")
233  )
234  emit("STAT-HYP-REPEATED-MEASURES-003", "Repeated measures two occasions equals paired t squared", "axis_coverage",
235       list(factor1_name = "Time", factor1_levels = list("t1", "t2"), cells = list("t1", "t2"),
236            contrast_family = "none", pairwise_procedure = "none"),
237       outputs, "Two occasions: the orthonormal contrast score (t2 - t1)/sqrt(2) gives F = t^2 of the paired t test and the same p; with one contrast every epsilon is 1, there is no Mauchly test and the multivariate F is exact.")

Whole script generators/r/hypothesis_repeated_measures.R


Repeated measures two within factors 2 x 2

Case STAT-HYP-REPEATED-MEASURES-004 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats
Citation
R Core Team, stats package help: aov, anova.mlm, mauchly.test, SSD, t.test; Greenhouse SW, Geisser S (1959), Psychometrika 24, 95-112; Lecoutre B (1991), Journal of Educational Statistics 16, 371-372.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_repeated_measures.R · sha256 8faee62dc29c9cba
Dataset
reference_data/datasets/HYP_RM.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/design/usedexact990.00E+2passed
/design/residual_dfexact880.00E+2passed
/within_tests/0/effectexactAApassed
/within_tests/0/sum_of_squarestolerance11.90250000000000211.9024999999999871.46E-1414passed
/within_tests/0/dfexact110.00E+2passed
/within_tests/0/error_sum_of_squarestolerance1.42499999999999851.42499999999999859.90E-1817passed
/within_tests/0/error_dfexact880.00E+2passed
/within_tests/0/f_valuetolerance66.82105263157902266.821052631578957.12E-1414passed
/within_tests/0/tests/0/p_valuetail_probability3.7376465619768127e-053.737646561976824e-051.15E-1914passed
/sphericity/0/greenhouse_geissertolerance11.00.00E+2passed
/multivariate_tests/0/p_valuetail_probability3.7376465619768127e-053.737646561976824e-051.15E-1914passed
/within_tests/1/effectexactBBpassed
/within_tests/1/sum_of_squarestolerance0.233611111111111710.233611111111109442.27E-1514passed
/within_tests/1/dfexact110.00E+2passed
/within_tests/1/error_sum_of_squarestolerance1.20388888888889281.20388888888889724.42E-1514passed
/within_tests/1/error_dfexact880.00E+2passed
/within_tests/1/f_valuetolerance1.5523765574526981.55237655745267782.02E-1413passed
/within_tests/1/tests/0/p_valuetail_probability0.248034727249100580.248034727249103472.89E-1513passed
/sphericity/1/greenhouse_geissertolerance11.00.00E+2passed
/multivariate_tests/4/p_valuetail_probability0.248034727249100580.248034727249103663.08E-1513passed
/within_tests/2/effectexactA × BA × Bpassed
/within_tests/2/sum_of_squarestolerance5.06249999999999735.06249999999998858.85E-1514passed
/within_tests/2/dfexact110.00E+2passed
/within_tests/2/error_sum_of_squarestolerance1.40499999999999981.40499999999999544.44E-1514passed
/within_tests/2/error_dfexact880.00E+2passed
/within_tests/2/f_valuetolerance28.82562277580069628.8256227758007424.59E-1414passed
/within_tests/2/tests/0/p_valuetail_probability0.000670445035887317950.00067044503588731473.26E-1814passed
/sphericity/2/greenhouse_geissertolerance11.00.00E+2passed
/multivariate_tests/8/p_valuetail_probability0.000670445035887317950.00067044503588731384.12E-1814passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

258  emit("STAT-HYP-REPEATED-MEASURES-004", "Repeated measures two within factors 2 x 2", "axis_coverage",
259       list(factor1_name = "A", factor1_levels = list("a1", "a2"), factor2_name = "B", factor2_levels = list("b1", "b2"),
260            cells = list("t1", "t2", "t3", "t4"), contrast_family = "none", pairwise_procedure = "none"),
261       outputs, "Cells in first-factor-slowest order: t1 = a1 b1, t2 = a1 b2, t3 = a2 b1, t4 = a2 b2. Each effect is one orthonormal contrast (A: (t1 + t2 - t3 - t4)/2, B: (t1 - t2 + t3 - t4)/2, A x B: (t1 - t2 - t3 + t4)/2) whose one-sample t squared is the F; effects are ordered A, B, A x B with four multivariate rows each.")

Whole script generators/r/hypothesis_repeated_measures.R


Exact paired sign test

Case STAT-HYP-SIGN-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and car
Citation
R Core Team, stats help: oneway.test, chisq.test, binom.test, wilcox.test, aov and kruskal.test; car::Anova and car::leveneTest help.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/hypothesis_methods.R · sha256 e2faa3576983ee8d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance1616.00.00E+2passed
/p_valuetail_probability3.051757812500002e-053.0517578125000014e-056.45E-2115passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

76  paired_diff <- d$x - d$y
77  positive <- sum(paired_diff > 0)
78  nonzero <- sum(paired_diff != 0)
79  sign <- stats::binom.test(positive, nonzero, p = 0.5)
80  emit("STAT-HYP-SIGN-001", "Exact paired sign test",
81       list(test = "sign", response = "x", second_response = "y"),
82       list("/statistic" = tol(positive, "positive paired differences"),
83            "/p_value" = prob(sign$p.value, "exact two-sided binomial tail")),
84       list("Zero paired differences omitted; test positive among nonzero differences."))

Whole script generators/r/hypothesis_methods.R


Decision threshold: STAT-HYP-THRESHOLD-EQUIV-ESTABLISHED-001

Case STAT-HYP-THRESHOLD-EQUIV-ESTABLISHED-001 · status: passed

Reference

Evidence type
R
Source
R Student t tails
Citation
R stats::t.test, pt and qt documentation; TOST analytic one-sided t tests.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_thresholds.R · sha256 e175b2ee1750ea1d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/p_lowertail_probability0.0978995316199993970.097899531619998815.83E-1614passed
/p_uppertail_probability0.0978995316199993550.097899531619999341.35E-1715passed
/equivalentexactTrueTruepassed

Convention

Reference code

Whole script generators/r/hypothesis_thresholds.R


Decision threshold: STAT-HYP-THRESHOLD-EQUIV-NOT-ESTABLISHED-001

Case STAT-HYP-THRESHOLD-EQUIV-NOT-ESTABLISHED-001 · status: passed

Reference

Evidence type
R
Source
R Student t tails
Citation
R stats::t.test, pt and qt documentation; TOST analytic one-sided t tests.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_thresholds.R · sha256 e175b2ee1750ea1d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/p_lowertail_probability0.102136833973959160.102136833973958665.00E-1614passed
/p_uppertail_probability0.102136833973959230.102136833973959121.12E-1614passed
/equivalentexactFalseFalsepassed

Convention

Reference code

Whole script generators/r/hypothesis_thresholds.R


Decision threshold: STAT-HYP-THRESHOLD-EQUIVEXT-ESTABLISHED-001

Case STAT-HYP-THRESHOLD-EQUIVEXT-ESTABLISHED-001 · status: passed

Reference

Evidence type
R
Source
R Student t tails
Citation
R stats::t.test, pt and qt documentation; TOST analytic one-sided t tests.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_thresholds.R · sha256 e175b2ee1750ea1d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/p_lowertail_probability0.097890951812598520.097890951812598552.78E-1715passed
/p_uppertail_probability0.097890951812598520.097890951812597698.33E-1614passed
/equivalentexactTrueTruepassed

Convention

Reference code

Whole script generators/r/hypothesis_thresholds.R


Decision threshold: STAT-HYP-THRESHOLD-EQUIVEXT-NOT-ESTABLISHED-001

Case STAT-HYP-THRESHOLD-EQUIVEXT-NOT-ESTABLISHED-001 · status: passed

Reference

Evidence type
R
Source
R Student t tails
Citation
R stats::t.test, pt and qt documentation; TOST analytic one-sided t tests.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_thresholds.R · sha256 e175b2ee1750ea1d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/p_lowertail_probability0.102145463733694930.102145463733695172.36E-1614passed
/p_uppertail_probability0.102145463733694930.1021454637336956.94E-1715passed
/equivalentexactFalseFalsepassed

Convention

Reference code

Whole script generators/r/hypothesis_thresholds.R


Decision threshold: STAT-HYP-THRESHOLD-TEST-REJECT-001

Case STAT-HYP-THRESHOLD-TEST-REJECT-001 · status: passed

Reference

Evidence type
R
Source
R Student t tails
Citation
R stats::t.test, pt and qt documentation; TOST analytic one-sided t tests.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_thresholds.R · sha256 e175b2ee1750ea1d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/p_valuetail_probability0.0969447507302092230.096944750730207241.98E-1513passed
/conclusioncontains['provide evidence']The data provide evidence that the mean differs from 5.96 (ppassed

Convention

Reference code

Whole script generators/r/hypothesis_thresholds.R


Decision threshold: STAT-HYP-THRESHOLD-TEST-RETAIN-001

Case STAT-HYP-THRESHOLD-TEST-RETAIN-001 · status: passed

Reference

Evidence type
R
Source
R Student t tails
Citation
R stats::t.test, pt and qt documentation; TOST analytic one-sided t tests.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_thresholds.R · sha256 e175b2ee1750ea1d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/p_valuetail_probability0.103139511404112430.103139511404112733.03E-1614passed
/conclusioncontains['do not provide sufficient evidence']The data do not provide sufficient evidence that the mean dipassed

Convention

Reference code

Whole script generators/r/hypothesis_thresholds.R


Simultaneous comparisons: STAT-HYP-TUKEY-001

Case STAT-HYP-TUKEY-001 · status: passed

Reference

Evidence type
R
Source
R stats and multcomp
Citation
R stats::aov and TukeyHSD help; multcomp::glht and mvtnorm::pmvt documentation.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, multcomp 1.4.32, mvtnorm 1.4.2, jsonlite 2.0.0
Reference script
generators/r/hypothesis_simultaneous.R · sha256 3aa116cb96c8afa8
Dataset
reference_data/datasets/HYP_BLOCKS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance8.05284552845528988.0528455284552837.12E-1515passed
/p_valuetail_probability0.00988818268550011040.0098881826855001382.77E-1714passed
/intervals/0/estimatetolerance1.07499999999999751.07500000000000022.68E-1514passed
/intervals/0/significantexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

34  fit <- stats::aov(x ~ treatment, data = d)
35  anova_row <- summary(fit)[[1]][1, ]
36  tukey <- stats::TukeyHSD(fit, "treatment", conf.level = 0.95)$treatment
37  emit("STAT-HYP-TUKEY-001", "tabelz.analysis.hypothesis_tests:run_hypothesis_test",
38       "hypothesisTestsMenuItem",
39       list(test = "one_way_anova", response = "x", group = "treatment", posthoc = TRUE),
40       list("/statistic" = tol(anova_row[["F value"]], "one-way F"),
41            "/p_value" = prob(anova_row[["Pr(>F)"]], "one-way F tail"),
42            "/intervals/0/estimate" = tol(tukey["B-A", "diff"], "Tukey-Kramer B-A contrast"),
43            "/intervals/0/significant" = exact(tukey["B-A", "p adj"] < 0.05,
44                                                "R TukeyHSD adjusted p below alpha.")),
45       list(list(path = "/intervals/0/lower", reason = "Statsmodels rounds the displayed Tukey interval; full-tier numeric comparison is deferred."),
46            list(path = "/intervals/0/upper", reason = "Statsmodels rounds the displayed Tukey interval; full-tier numeric comparison is deferred.")),
47       list("R TukeyHSD is Tukey-Kramer for unequal group sizes; this balanced design uses the same convention.",
48            "The first ordered interval is B minus A; its rejection flag is compared without reducing the numerical tier."))

Whole script generators/r/hypothesis_simultaneous.R


Type II two-way interaction ANOVA

Case STAT-HYP-TWO-ANOVA-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and car
Citation
R Core Team, stats help: oneway.test, chisq.test, binom.test, wilcox.test, aov and kruskal.test; car::Anova and car::leveneTest help.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/hypothesis_methods.R · sha256 e2faa3576983ee8d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance0.00280275604344216090.0028027560434429227.61E-1612passed
/p_valuetail_probability0.95864999381092830.95864999381092265.66E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

45  two_way <- car::Anova(stats::lm(x ~ g * h, data = d), type = 2)
46  interaction <- two_way["g:h", ]
47  emit("STAT-HYP-TWO-ANOVA-001", "Type II two-way interaction ANOVA",
48       list(test = "two_way_anova", response = "x", group = "g", second_group = "h",
49            posthoc = FALSE),
50       list("/statistic" = tol(interaction[["F value"]], "type II interaction F"),
51            "/p_value" = prob(interaction[["Pr(>F)"]], "type II interaction F tail")),
52       list("Type II sums of squares, g:h interaction is the primary statistic."))

Whole script generators/r/hypothesis_methods.R


Sparse two-proportion Fisher test

Case STAT-HYP-TWO-PROP-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/p_valuetail_probability0.619269619269619210.61926961926961942.21E-1615passed
/estimatetolerance0.250.250.00E+0passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

114  tab <- table(d$g, d$outcome == "yes")
115  fisher <- stats::fisher.test(tab)
116  emit("STAT-HYP-TWO-PROP-001", "Sparse two-proportion Fisher test", "nominal",
117       list(test = "two_proportion", response = "outcome", group = "g", success_value = "yes",
118            method = "fisher"),
119       list("/p_value" = prob(fisher$p.value, "Fisher exact two-sided"),
120            "/estimate" = tol(mean(d$outcome[d$g == "A"] == "yes") -
121                                mean(d$outcome[d$g == "B"] == "yes"), "proportion difference")),
122       list("Fisher p-value from 2x2 table; the reported estimate is A minus B proportion."))

Whole script generators/r/hypothesis_tests.R


Pooled two-proportion Z

Case STAT-HYP-TWO-PROP-Z-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and car
Citation
R Core Team, stats help: oneway.test, chisq.test, binom.test, wilcox.test, aov and kruskal.test; car::Anova and car::leveneTest help.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/hypothesis_methods.R · sha256 e2faa3576983ee8d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance11.00.00E+2passed
/p_valuetail_probability0.317310507862914090.317310507862914155.84E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

108  g_a <- d$outcome[d$g == "A"]
109  g_b <- d$outcome[d$g == "B"]
110  p_a <- mean(g_a == "yes")
111  p_b <- mean(g_b == "yes")
112  pooled <- (sum(g_a == "yes") + sum(g_b == "yes")) / (length(g_a) + length(g_b))
113  z <- (p_a - p_b) / sqrt(pooled * (1 - pooled) * (1 / length(g_a) + 1 / length(g_b)))
114  emit("STAT-HYP-TWO-PROP-Z-001", "Pooled two-proportion Z",
115       list(test = "two_proportion", response = "outcome", group = "g",
116            success_value = "yes", method = "z"),
117       list("/statistic" = tol(z, "pooled two-proportion z"),
118            "/p_value" = prob(2 * stats::pnorm(-abs(z)), "uncorrected normal tail")),
119       list("Manual method z uses no continuity correction; pooled null SE shown analytically."))

Whole script generators/r/hypothesis_methods.R


Independent welch t

Case STAT-HYP-TWO-T-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-1.7431919415339725-1.74319194153397348.88E-1615passed
/dftolerance13.99945589410847513.9994558941084686.64E-1515passed
/p_valuetail_probability0.1032104604913960.103210460491395937.22E-1715passed
/estimatetolerance-1.5749999999999993-1.57500000000000028.78E-1615passed

Convention

Reproduce it yourself

  1. Data: data.csv, the rows as analysed.
  2. R: run the script below; it prints every compared value with its output path.
  3. HypoNum: follow the steps below and read the same values from the result window.

R script

# Welch two-sample t test. Base R only; tested with R 4.6.1.
report <- function(path, x) cat(path, "=", if (is.numeric(x)) sprintf("%.17g", x) else as.character(x), "\n")

d <- read.csv("data.csv")                       # columns x (thickness) and g (line A or B)
fit <- t.test(x ~ g, data = d)                  # Welch, two-sided, 95 % interval

report("/estimate",  diff(rev(tapply(d$x, d$g, mean))))   # mean(A) - mean(B)
report("/statistic", fit$statistic)
report("/df",        fit$parameter)
report("/p_value",   fit$p.value)

HypoNum

  1. File > Import data, select data.csv, click Import.
  2. Statistics > Compare means & ANOVA > Hypothesis tests.
  3. Test: Two-sample t-test. Response: x. Group: g. Click Run test.

Welch is the default method. Read the A-minus-B difference, t, the degrees of freedom and the p-value from the result window.

Reference code

Whole script generators/r/hypothesis_tests.R


Independent pooled t

Case STAT-HYP-TWO-T-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-1.7431919415339725-1.74319194153397348.88E-1615passed
/dftolerance1414.00.00E+2passed
/p_valuetail_probability0.103209612875833470.103209612875833222.52E-1614passed
/estimatetolerance-1.5749999999999993-1.57500000000000028.78E-1615passed

Convention

Reference code

Whole script generators/r/hypothesis_tests.R


Bartlett equal variances

Case STAT-HYP-VARIANCE-BARTLETT-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance0.000253930886168518590.00025393088616851862.71E-2116passed
/p_valuetail_probability0.987286080836755510.98728608083675551.45E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

141  bartlett <- stats::bartlett.test(x ~ g, data = d)
142  emit("STAT-HYP-VARIANCE-BARTLETT-001", "Bartlett equal variances", "axis_coverage",
143       list(test = "equal_variance", response = "x", group = "g", method = "bartlett"),
144       list("/statistic" = tol(bartlett$statistic, "Bartlett K2"),
145            "/p_value" = prob(bartlett$p.value, "Bartlett chi-square tail")),
146       list("Bartlett test uses sample variances and chi-square approximation."))

Whole script generators/r/hypothesis_tests.R


Two-group variance ratio

Case STAT-HYP-VARIANCE-F-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats package documentation
Citation
R Core Team, stats package help: t.test, var.test, aov, binom.test, chisq.test, fisher.test, wilcox.test, kruskal.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_tests.R · sha256 d4285bc8becb7ad3
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance0.98760869565217380.98760869565217344.48E-1615passed
/p_valuetail_probability0.987300045323276420.98730004532327595.54E-1615passed
/ci_lowertolerance0.197723052079291390.197723052079291445.48E-1715passed
/ci_uppertolerance4.93301577874006724.9330157787400616.17E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

132  variance <- stats::var.test(x ~ g, data = d)
133  emit("STAT-HYP-VARIANCE-F-001", "Two-group variance ratio", "axis_coverage",
134       list(test = "equal_variance", response = "x", group = "g", method = "f"),
135       list("/statistic" = tol(variance$statistic, "variance ratio F"),
136            "/p_value" = prob(variance$p.value, "variance ratio F tail"),
137            "/ci_lower" = tol(variance$conf.int[[1]], "variance ratio CI lower"),
138            "/ci_upper" = tol(variance$conf.int[[2]], "variance ratio CI upper")),
139       list("var.test compares A/B sample variances in sorted group order."))

Whole script generators/r/hypothesis_tests.R


Brown-Forsythe median-centred variance test

Case STAT-HYP-VARIANCE-LEVENE-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and car
Citation
R Core Team, stats help: oneway.test, chisq.test, binom.test, wilcox.test, aov and kruskal.test; car::Anova and car::leveneTest help.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/hypothesis_methods.R · sha256 e2faa3576983ee8d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance0.0453074433656960390.045307443365696051.36E-1715passed
/p_valuetail_probability0.834507942707458870.83450794270745881.09E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

69  lev <- car::leveneTest(x ~ g, data = d, center = median)
70  emit("STAT-HYP-VARIANCE-LEVENE-001", "Brown-Forsythe median-centred variance test",
71       list(test = "equal_variance", response = "x", group = "g", method = "levene"),
72       list("/statistic" = tol(lev[["F value"]][[1]], "Brown-Forsythe F"),
73            "/p_value" = prob(lev[["Pr(>F)"]][[1]], "Brown-Forsythe F tail")),
74       list("Levene deviations centred on each group's median."))

Whole script generators/r/hypothesis_methods.R


Paired Wilcoxon signed ranks

Case STAT-HYP-WILCOXON-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats and car
Citation
R Core Team, stats help: oneway.test, chisq.test, binom.test, wilcox.test, aov and kruskal.test; car::Anova and car::leveneTest help.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/hypothesis_methods.R · sha256 e2faa3576983ee8d
Dataset
reference_data/datasets/HYP_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance136136.00.00E+2passed
/p_valuetail_probability3.0517578125e-053.0517578125e-050.00E-13passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

88  wilcox <- stats::wilcox.test(paired_diff, mu = 0, exact = TRUE)
89  stopifnot(sum(paired_diff != 0) <= 50L)
90  emit("STAT-HYP-WILCOXON-001", "Paired Wilcoxon signed ranks",
91       list(test = "wilcoxon", response = "x", second_response = "y"),
92       list("/statistic" = tol(wilcox$statistic, "W+, the positive signed-rank sum"),
93            "/p_value" = prob(wilcox$p.value, "exact conditional signed-rank distribution with tied ranks")),
94       list("Issue #12 ruling: W+ and exact conditional inference for m <= 50, including ties.",
95            "Until 2026-09-25 this case used R's explicit normal approximation (exact = FALSE, correct = TRUE)."))

Whole script generators/r/hypothesis_methods.R


Binomial single-sampling plan: AQL 0.01, LTPD 0.06

Case STAT-MSA-ACCEPT-001 · status: passed

Reference

Evidence type
R
Source
R stats::pbinom, stats::qt (noncentral) and survival::survreg
Citation
R stats::Binomial, stats::TDist and survival::survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/measurement_manufacturing.R · sha256 5a42bd20f26f6b22
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sample_sizeexact1101100.00E+2passed
/accept_numberexact330.00E+2passed
/probability_accept_aqltolerance0.97496185363477150.97496185363477141.08E-1615passed
/probability_accept_ltpdtolerance0.0980303807536210960.098030380753620575.27E-1614passed

Convention

Reference code

Whole script generators/r/measurement_manufacturing.R


Binomial single-sampling plan: AQL 0.02, LTPD 0.08

Case STAT-MSA-ACCEPT-002 · status: passed

Reference

Evidence type
R
Source
R stats::pbinom, stats::qt (noncentral) and survival::survreg
Citation
R stats::Binomial, stats::TDist and survival::survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/measurement_manufacturing.R · sha256 5a42bd20f26f6b22
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sample_sizeexact1121120.00E+2passed
/accept_numberexact440.00E+2passed
/probability_accept_aqltolerance0.925097904667959050.92509790466795943.31E-1615passed
/probability_accept_ltpdtolerance0.0494902416702640920.049490241670264392.98E-1614passed

Convention

Reference code

Whole script generators/r/measurement_manufacturing.R


An AQL at or above the LTPD is refused

Case STAT-MSA-ACCEPT-003 · status: passed

Reference

Evidence type
R
Source
R stats::pbinom, stats::qt (noncentral) and survival::survreg
Citation
R stats::Binomial, stats::TDist and survival::survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/measurement_manufacturing.R · sha256 5a42bd20f26f6b22
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

74  emit("STAT-MSA-ACCEPT-003", acc, "An AQL at or above the LTPD is refused", NULL, list(aql = 0.05, ltpd = 0.05), list(),
75       c("A plan needs AQL < LTPD."), kind = "invalid_input", error = list(error_class = "ValueError", message_contains = "LTPD"))

Whole script generators/r/measurement_manufacturing.R


Three appraisers, two trials and a known standard

Case STAT-MSA-AGREE-001 · status: passed

Reference

Evidence type
R
Source
R irr::kappa2, irr::kappam.fleiss and stats::prop.test
Citation
R irr::kappa2, irr::kappam.fleiss and stats::prop.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/measurement_attribute.R · sha256 beefb4724be48657
Dataset
reference_data/datasets/MSA_AGREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/appraisers/[appraiser=Ann]/within/agreement_pcttolerance83.33333333333332983.333333333333341.38E-1415passed
/appraisers/[appraiser=Ann]/within/ci_lowertolerance55.19691377470265855.196913774702668.23E-1717passed
/appraisers/[appraiser=Ann]/within/ci_uppertolerance95.30348578161462695.303485781614633.06E-1617passed
/appraisers/[appraiser=Ann]/within/kappatolerance0.675675675675675880.67567567567567581.11E-1615passed
/appraisers/[appraiser=Ann]/versus_standard/agreement_pcttolerance91.66666666666667191.666666666666661.38E-1415passed
/appraisers/[appraiser=Bob]/within/agreement_pcttolerance7575.00.00E+2passed
/appraisers/[appraiser=Bob]/within/ci_lowertolerance46.7694665066434446.7694665066434251.46E-1415passed
/appraisers/[appraiser=Bob]/within/ci_uppertolerance91.10583316059451491.105833160594531.44E-1415passed
/appraisers/[appraiser=Bob]/within/kappatolerance0.470588235294117640.470588235294117645.28E-1917passed
/appraisers/[appraiser=Bob]/versus_standard/agreement_pcttolerance79.16666666666667179.166666666666661.38E-1415passed
/appraisers/[appraiser=Cy]/within/agreement_pcttolerance58.33333333333333658.3333333333333362.98E-1617passed
/appraisers/[appraiser=Cy]/within/ci_lowertolerance31.95113125495498431.9511312549549731.11E-1415passed
/appraisers/[appraiser=Cy]/within/ci_uppertolerance80.6739686341243580.673968634124351.81E-1617passed
/appraisers/[appraiser=Cy]/within/kappatolerance0.166666666666666740.166666666666666631.10E-1615passed
/appraisers/[appraiser=Cy]/versus_standard/agreement_pcttolerance54.16666666666666454.1666666666666642.98E-1617passed
/between/agreement_pcttolerance41.66666666666666441.666666666666677.40E-1515passed
/between/kappatolerance0.212499999999999830.212499999999999832.08E-1817passed
/n_trialsexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

85  ratings <- sapply(appr, function(ap) c(rating_of(ap, "T1"), rating_of(ap, "T2")))
86  between <- sum(apply(ratings, 1, function(v) length(unique(v)) == 1))
87  out[["/between/agreement_pct"]] <- tol(100 * between / 24, "Item-and-trial pairs where all appraisers agree.")
88  out[["/between/kappa"]] <- tol(irr::kappam.fleiss(ratings)$value, "Fleiss kappa over three appraisers.")
89  out[["/n_trials"]] <- exact(2L, "Two trials.")
90  emit("STAT-MSA-AGREE-001", agree_entry, "attributeAgreementMenuItem", "Three appraisers, two trials and a known standard", "MSA_AGREE",
91       list(rating = "rating", part = "item", appraiser = "appraiser", trial = "trial", standard = "standard"), out,
92       c("Nominal scale (line 288); Fleiss kappa for more than two appraisers (line 302)."))

Whole script generators/r/measurement_attribute.R


Two appraisers use Cohen kappa between them

Case STAT-MSA-AGREE-002 · status: passed

Reference

Evidence type
R
Source
R irr::kappa2, irr::kappam.fleiss and stats::prop.test
Citation
R irr::kappa2, irr::kappam.fleiss and stats::prop.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/measurement_attribute.R · sha256 beefb4724be48657
Dataset
reference_data/datasets/MSA_AGREE_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/between/kappatolerance0.583333333333333260.58333333333333336.82E-1917passed
/between/agreement_pcttolerance79.16666666666665779.166666666666661.93E-1617passed
/overall_standardabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 93  two <- read_dataset("MSA_AGREE"); two <- two[two$appraiser != "Cy", ]
 94  export_dataset(two, "MSA_AGREE_TWO", "MSA_AGREE with appraisers Ann and Bob only", "R stats::runif documentation",
 95                 "Constructed validation data; no third-party license.",
 96                 c(item = "String", appraiser = "String", trial = "String", rating = "String", standard = "String"))
 97  a <- read_dataset("MSA_AGREE_TWO")
 98  ratings2 <- sapply(c("Ann", "Bob"), function(ap) c(rating_of(ap, "T1"), rating_of(ap, "T2")))
 99  emit("STAT-MSA-AGREE-002", agree_entry, "attributeAgreementMenuItem", "Two appraisers use Cohen kappa between them", "MSA_AGREE_TWO",
100       list(rating = "rating", part = "item", appraiser = "appraiser", trial = "trial"),
101       list("/between/kappa" = tol(irr::kappa2(ratings2)$value, "Cohen kappa for two appraisers (line 302)."),
102            "/between/agreement_pct" = tol(100 * mean(ratings2[, 1] == ratings2[, 2]), "Item-and-trial agreement."),
103            "/overall_standard" = absent("No standard was assigned.")),
104       c("Two appraisers."), kind = "axis_coverage")

Whole script generators/r/measurement_attribute.R


Weighted Cohen kappa: ordinal_linear

Case STAT-MSA-AGREE-003 · status: passed

Reference

Evidence type
R
Source
R irr::kappa2, irr::kappam.fleiss and stats::prop.test
Citation
R irr::kappa2, irr::kappam.fleiss and stats::prop.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/measurement_attribute.R · sha256 beefb4724be48657
Dataset
reference_data/datasets/MSA_AGREE_ORDINAL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/between/kappatolerance0.381443298969072030.381443298969072141.12E-1615passed

Convention

Reference code

Whole script generators/r/measurement_attribute.R


Weighted Cohen kappa: ordinal_quadratic

Case STAT-MSA-AGREE-004 · status: passed

Reference

Evidence type
R
Source
R irr::kappa2, irr::kappam.fleiss and stats::prop.test
Citation
R irr::kappa2, irr::kappam.fleiss and stats::prop.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/measurement_attribute.R · sha256 beefb4724be48657
Dataset
reference_data/datasets/MSA_AGREE_ORDINAL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/between/kappatolerance0.444444444444444640.44444444444444415.53E-1614passed

Convention

Reference code

Whole script generators/r/measurement_attribute.R


Repeated columns are refused

Case STAT-MSA-AGREE-005 · status: passed

Reference

Evidence type
R
Source
R irr::kappa2, irr::kappam.fleiss and stats::prop.test
Citation
R irr::kappa2, irr::kappam.fleiss and stats::prop.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, irr 0.85, jsonlite 2.0.0
Reference script
generators/r/measurement_attribute.R · sha256 beefb4724be48657
Dataset
reference_data/datasets/MSA_AGREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

121  emit("STAT-MSA-AGREE-005", agree_entry, "attributeAgreementMenuItem", "Repeated columns are refused", "MSA_AGREE",
122       list(rating = "rating", part = "item", appraiser = "item"), list(),
123       c("Line 293: all assigned columns must be distinct."), kind = "invalid_input",
124       error = list(error_class = "ValueError", message_contains = "must use a different column"))

Whole script generators/r/measurement_attribute.R


AIAG MSA-4 ANOVA Gage R&R on the Figure III-B 15 data (interaction pooled)

Case STAT-MSA-AIAG-ANOVA-001 · status: imprecise_reference

Reference

Evidence type
Published example
Source
MSA-4 worked examples, as recorded in the approved AIAG MSA-4 specification (sections 6 and 7) with its documented printed discrepancies.
Citation
AIAG, Measurement Systems Analysis Reference Manual, 4th edition (2010), Chapter III Section B, Chapter IV Sections E and H, Appendices A and B
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MSA_AIAG_FIG_IIIB15.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/anova/table/[source=Parts]/sstolerance88.361988.361934444444443.44E-56passed
/anova/table/[source=Appraiser]/sstolerance3.16733.1672622222222223.78E-54passed
/anova/table/[source=Appraiser by part]/sstolerance0.35900.358982222222222041.78E-54passed
/anova/table/[source=Equipment]/sstolerance2.75892.75893333333333373.33E-54passed
/anova/decision/ftolerance0.4340.43372103013510294imprecise_reference
/anova/decision/decisionexactpooledpooledpassed
/anova/part_variancetolerance1.0864471.08644660441067843.96E-76passed
/summary/rows/[key=ev]/std_devtolerance0.1999330.199933179720816621.80E-76passed
/summary/rows/[key=av]/std_devtolerance0.2268380.22683752149632234.79E-75passed
/summary/rows/[key=int]/std_devtolerance00.00.00E+215passed
/summary/rows/[key=grr]/variancetolerance0.091430.091428537511870841.46E-64passed
/summary/rows/[key=pv]/std_devtolerance1.0423271.04232749383803474.94E-76passed
/summary/rows/[key=tv]/std_devtolerance1.0851.08529956321863imprecise_reference
/summary/rows/[key=grr]/pct_total_variationtolerance27.927.86065088114229imprecise_reference
/summary/rows/[key=grr]/pct_contributiontolerance7.87.762158675208948imprecise_reference
/summary/rows/[key=pv]/pct_total_variationtolerance96.096.04053379942818imprecise_reference
/summary/rows/[key=grr]/pct_tolerancetolerance22.6822.677864174217852.14E-34passed
/summary/ndc/valueexact440.00E+2passed

Convention


AIAG MSA-4 Average and Range on the Figure III-B 15 raw data

Case STAT-MSA-AIAG-AR-001 · status: imprecise_reference

Reference

Evidence type
Published example
Source
MSA-4 worked examples, as recorded in the approved AIAG MSA-4 specification (sections 6 and 7) with its documented printed discrepancies.
Citation
AIAG, Measurement Systems Analysis Reference Manual, 4th edition (2010), Chapter III Section B, Chapter IV Sections E and H, Appendices A and B
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MSA_AIAG_FIG_IIIB15.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/average_range/r_double_bartolerance0.34170.341666666666666733.33E-54passed
/average_range/x_difftolerance0.44460.44466666666666665imprecise_reference
/average_range/r_ptolerance3.5113.5111111111111111.11E-44passed
/average_range/ucl_rtolerance0.88160.8815000000000002imprecise_reference
/average_range/out_of_control/0/appraiserexactBBpassed
/average_range/out_of_control/0/partexact44passed
/summary/rows/[key=ev]/std_devtolerance0.201880.2018566666666667imprecise_reference
/summary/rows/[key=av]/std_devtolerance0.229630.22966702910320377imprecise_reference
/summary/rows/[key=grr]/std_devtolerance0.305750.305766345654438661.63E-54passed
/summary/rows/[key=pv]/std_devtolerance1.104561.10459555555555563.56E-54passed
/summary/rows/[key=tv]/std_devtolerance1.146101.14613454685213823.45E-54passed
/summary/rows/[key=grr]/pct_total_variationtolerance26.6826.6780498410266771.95E-34passed
/summary/ndc/valueexact550.00E+2passed

Convention


AIAG MSA-4 Appendix B impact of GRR on Cp (corrected tolerance formula)

Case STAT-MSA-AIAG-CP-001 · status: passed

Reference

Evidence type
Published example
Source
MSA-4 worked examples, as recorded in the approved AIAG MSA-4 specification (sections 6 and 7) with its documented printed discrepancies.
Citation
AIAG, Measurement Systems Analysis Reference Manual, 4th edition (2010), Chapter III Section B, Chapter IV Sections E and H, Appendices A and B
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/bases/[basis=process]/actual_cptolerance2.52.50.00E+1passed
/bases/[basis=tolerance]/actual_cptolerance2.52.50.00E+1passed

Convention


AIAG MSA-4 Chapter IV Section H pooled standard deviation study

Case STAT-MSA-AIAG-POOLED-001 · status: imprecise_reference

Reference

Evidence type
Published example
Source
MSA-4 worked examples, as recorded in the approved AIAG MSA-4 specification (sections 6 and 7) with its documented printed discrepancies.
Citation
AIAG, Measurement Systems Analysis Reference Manual, 4th edition (2010), Chapter III Section B, Chapter IV Sections E and H, Appendices A and B
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MSA_AIAG_FIG_IIIB15.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pooled_sd/pooled_repeatabilitytolerance0.214434660.214434657232039422.77E-97passed
/pooled_sd/pooled_reproducibilitytolerance0.208430640.208430643940216513.94E-97passed
/pooled_sd/pooled_grrtolerance0.299041060.29904105998266451.73E-1110passed
/pooled_sd/limits/sigmatolerance0.21530.21532995744686606imprecise_reference
/pooled_sd/limits/s_centertolerance0.19080.1908312061460093imprecise_reference

Convention


AIAG MSA-4 Range method on the Table III-B 6 data

Case STAT-MSA-AIAG-RANGE-001 · status: imprecise_reference

Reference

Evidence type
Published example
Source
MSA-4 worked examples, as recorded in the approved AIAG MSA-4 specification (sections 6 and 7) with its documented printed discrepancies.
Citation
AIAG, Measurement Systems Analysis Reference Manual, 4th edition (2010), Chapter III Section B, Chapter IV Sections E and H, Appendices A and B
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MSA_AIAG_TABLE_IIIB6.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/range_method/average_rangetolerance0.070.076.66E-1816passed
/range_method/d2_startolerance1.191051.191055.79E-1716passed
/range_method/grr_sdtolerance0.05880.05877167205406995imprecise_reference
/range_method/pct_grrtolerance75.675.63921757280559imprecise_reference

Convention


AIAG MSA-4 Chapter IV Section E control-chart worksheet

Case STAT-MSA-AIAG-WORKSHEET-001 · status: imprecise_reference

Reference

Evidence type
Published example
Source
MSA-4 worked examples, as recorded in the approved AIAG MSA-4 specification (sections 6 and 7) with its documented printed discrepancies.
Citation
AIAG, Measurement Systems Analysis Reference Manual, 4th edition (2010), Chapter III Section B, Chapter IV Sections E and H, Appendices A and B
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MSA_AIAG_FIG_IVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/worksheet/average_rangetolerance7.67.63.55E-1616passed
/worksheet/range_upper_limittolerance24.824.836799999999997imprecise_reference
/worksheet/measurement_sdtolerance7.2897.2888512062469751.49E-44passed
/worksheet/signal_to_noisetolerance1.6981.6985623030218318imprecise_reference
/worksheet/ndc/valueexact220.00E+2passed

Convention


Deming regression with lambda 1

Case STAT-MSA-DEMING-001 · status: passed

Reference

Evidence type
R
Source
mcr 1.3.3.1 Deming point estimates and Linnet jackknife intervals; independent R arithmetic for other fields
Citation
mcr 1.3.3.1 mcreg documentation; NCSS Deming Regression chapter 303, Program 1; NIST Dataplot reproduces that NCSS example in case 007
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, mcr 1.3.3.1
Reference script
generators/r/measurement_deming.R · sha256 4f9a1dc380b3c156
Dataset
reference_data/datasets/MSA_DEMING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/slopetolerance1.03096917528822421.03096917528822421.78E-1716passed
/intercepttolerance0.508422174301307090.50842217430129828.88E-1513passed
/sxytolerance1365.40390250600011365.40390250600032.50E-1315passed
/rows/0/xhattolerance8.72047212650564068.720472126505643.10E-1717passed
/rows/0/y_residualtolerance-0.32186013068876917-0.32186013068876563.55E-1513passed
/rows/17/xhattolerance7.07486341143852387.0748634114385241.92E-1717passed
/rows/17/y_residualtolerance-0.20498827086890881-0.204988270868905253.56E-1513passed
/rows/39/xhattolerance6.03841537650399836.0384153765039981.14E-1717passed
/rows/39/y_residualtolerance0.279557704936639340.27955770493664293.55E-1513passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

76  set.seed(20261027)
77  truth <- runif(40, 5, 25)
78  x <- round(truth + rnorm(40, 0, 0.4), 4); y <- round(0.3 + 1.04 * truth + rnorm(40, 0, 0.6), 4)
79  export_dataset(data.frame(method_a = x, method_b = y), "MSA_DEMING", "Constructed method comparison: 40 pairs, X error SD 0.4, Y error SD 0.6, seed 20261027",
80                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.", c(method_a = "Float64", method_b = "Float64"))
81  d <- read_dataset("MSA_DEMING")
82  fit1 <- mcr_fit(d$method_a, d$method_b, 1)
83  out1 <- outputs(deming(d$method_a, d$method_b, 1), d$method_a, d$method_b, c(0L, 17L, 39L))
84  out1[["/slope"]] <- mcr_point(fit1, "Slope"); out1[["/intercept"]] <- mcr_point(fit1, "Intercept")
85  emit("STAT-MSA-DEMING-001", "Deming regression with lambda 1", "MSA_DEMING", list(x_column = "method_a", y_column = "method_b"),
86       out1, c("Orthogonal regression at lambda = 1.", "mcr supplies slope/intercept; closed-form R supplies Sxy and latent row outputs."), oracle = "mcr")

Whole script generators/r/measurement_deming.R


Deming regression with lambda 2.25

Case STAT-MSA-DEMING-002 · status: passed

Reference

Evidence type
R
Source
mcr 1.3.3.1 Deming point estimates and Linnet jackknife intervals; independent R arithmetic for other fields
Citation
mcr 1.3.3.1 mcreg documentation; NCSS Deming Regression chapter 303, Program 1; NIST Dataplot reproduces that NCSS example in case 007
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, mcr 1.3.3.1
Reference script
generators/r/measurement_deming.R · sha256 4f9a1dc380b3c156
Dataset
reference_data/datasets/MSA_DEMING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/slopetolerance1.02852733279456121.02852733279456121.10E-1716passed
/intercepttolerance0.543301580677519880.5433015806775118.88E-1513passed
/sxytolerance1365.40390250600011365.40390250600032.50E-1315passed
/rows/0/xhattolerance8.84187895839854758.841878958398551.82E-1515passed
/rows/0/y_residualtolerance-0.46031576265152729-0.460315762651523743.55E-1514passed
/rows/39/xhattolerance5.92303570179211825.9230357017921184.61E-1717passed
/rows/39/y_residualtolerance0.378094306911276770.378094306911279432.66E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

87  fit2 <- mcr_fit(d$method_a, d$method_b, 2.25)
88  out2 <- outputs(deming(d$method_a, d$method_b, 2.25), d$method_a, d$method_b, c(0L, 39L))
89  out2[["/slope"]] <- mcr_point(fit2, "Slope"); out2[["/intercept"]] <- mcr_point(fit2, "Intercept")
90  emit("STAT-MSA-DEMING-002", "Deming regression with lambda 2.25", "MSA_DEMING",
91       list(x_column = "method_a", y_column = "method_b", lambda_ratio = 2.25, comparable_units = TRUE),
92       out2,
93       c("lambda = 0.6^2 / 0.4^2 = 2.25; mcr error.ratio = 1/2.25.",
94         "mcr supplies slope/intercept; closed-form R supplies Sxy and latent row outputs."),
95       kind = "axis_coverage", oracle = "mcr")

Whole script generators/r/measurement_deming.R


Swapping X and Y with 1/lambda gives the inverse line

Case STAT-MSA-DEMING-003 · status: passed

Reference

Evidence type
R
Source
mcr 1.3.3.1 Deming point estimates and Linnet jackknife intervals; independent R arithmetic for other fields
Citation
mcr 1.3.3.1 mcreg documentation; NCSS Deming Regression chapter 303, Program 1; NIST Dataplot reproduces that NCSS example in case 007
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, mcr 1.3.3.1
Reference script
generators/r/measurement_deming.R · sha256 4f9a1dc380b3c156
Dataset
reference_data/datasets/MSA_DEMING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/slopetolerance0.972263904045164160.97226390404516422.33E-1817passed
/intercepttolerance-0.5282325159034329-0.5282325159034248.88E-1513passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 96  inv <- deming(d$method_b, d$method_a, 1 / 2.25)
 97  fit3 <- mcr_fit(d$method_b, d$method_a, 1 / 2.25)
 98  emit("STAT-MSA-DEMING-003", "Swapping X and Y with 1/lambda gives the inverse line", "MSA_DEMING",
 99       list(x_column = "method_b", y_column = "method_a", lambda_ratio = 1 / 2.25),
100       list("/slope" = mcr_point(fit3, "Slope"), "/intercept" = mcr_point(fit3, "Intercept")),
101       c(sprintf("R: 1 / b(lambda 2.25) = %.15g and b(swapped, 1/2.25) = %.15g.", 1 / deming(d$method_a, d$method_b, 2.25)$b, inv$b)),
102       kind = "axis_coverage", oracle = "mcr")

Whole script generators/r/measurement_deming.R


Missing and excluded pairs keep their rows without fitted values

Case STAT-MSA-DEMING-004 · status: passed

Reference

Evidence type
R
Source
mcr 1.3.3.1 Deming point estimates and Linnet jackknife intervals; independent R arithmetic for other fields
Citation
mcr 1.3.3.1 mcreg documentation; NCSS Deming Regression chapter 303, Program 1; NIST Dataplot reproduces that NCSS example in case 007
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, mcr 1.3.3.1
Reference script
generators/r/measurement_deming.R · sha256 4f9a1dc380b3c156
Dataset
reference_data/datasets/MSA_DEMING_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/slopetolerance1.02189225888616081.02189225888616084.63E-1716passed
/rows/0/xhatabsentNonepassed
/rows/4/xhatabsentNonepassed
/rows/1/xhattolerance23.40927224110617323.4092722411061667.22E-1515passed
/excluded_countexact220.00E+2passed
/missing_countexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

103  m <- read_dataset("MSA_DEMING"); m$method_b[c(5, 30)] <- NA
104  export_dataset(m, "MSA_DEMING_MISSING", "MSA_DEMING with method_b missing in rows 5 and 30 (1-based)", "R stats::rnorm documentation",
105                 "Constructed validation data; no third-party license.", c(method_a = "Float64", method_b = "Float64"))
106  mm <- read_dataset("MSA_DEMING_MISSING")
107  excluded <- c(0L, 20L)
108  keep <- setdiff(which(!is.na(mm$method_b)), excluded + 1L)
109  r <- deming(mm$method_a[keep], mm$method_b[keep], 1)
110  fit4 <- mcr_fit(mm$method_a[keep], mm$method_b[keep], 1)
111  emit("STAT-MSA-DEMING-004", "Missing and excluded pairs keep their rows without fitted values", "MSA_DEMING_MISSING",
112       list(x_column = "method_a", y_column = "method_b"),
113       list("/slope" = mcr_point(fit4, "Slope"),
114            "/rows/0/xhat" = absent("An excluded row keeps its identity with fitted fields unavailable (line 81-82)."),
115            "/rows/4/xhat" = absent("A missing pair has no projection."),
116            "/rows/1/xhat" = tol(r$xhat[1], "The first retained pair."),
117            "/excluded_count" = exact(2L, "Two excluded rows."), "/missing_count" = exact(2L, "Two missing pairs.")),
118       c("Rows 0 and 20 (0-based) are excluded; mcr fits the 36 matched eligible pairs only.",
119         "Row identities, exclusions and latent projection come from the Tabelz policy and closed form, not mcr."),
120       kind = "missing_excluded", excluded = excluded, oracle = "mcr")

Whole script generators/r/measurement_deming.R


Jackknife t intervals on the NCSS ten-pair example

Case STAT-MSA-DEMING-007 · status: passed

Reference

Evidence type
R
Source
mcr 1.3.3.1 Deming point estimates and Linnet jackknife intervals; independent R arithmetic for other fields
Citation
mcr 1.3.3.1 mcreg documentation; NCSS Deming Regression chapter 303, Program 1; NIST Dataplot reproduces that NCSS example in case 007
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, mcr 1.3.3.1
Reference script
generators/r/measurement_deming.R · sha256 4f9a1dc380b3c156
Dataset
reference_data/datasets/MSA_DEMING_NCSS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/slopetolerance1.00119422781949071.00119422781949051.77E-1615passed
/intercepttolerance-0.08974489900704441-0.089744899007042631.78E-1513passed
/interval_methodexactjackknife_tjackknife_tpassed
/jackknife_degrees_of_freedomexact880.00E+2passed
/jackknife_slope_setolerance0.187177052799396370.18717705279939651.34E-1615passed
/jackknife_slope_interval/0tolerance0.569563170048754230.56956317004875384.44E-1615passed
/jackknife_slope_interval/1tolerance1.43282528559022731.43282528559022734.00E-1716passed
/jackknife_intercept_setolerance1.72198741281869831.7219874128187012.64E-1514passed
/jackknife_intercept_interval/0tolerance-4.0606549937364864-4.060654993736493.57E-1515passed
/jackknife_intercept_interval/1tolerance3.88116519572239763.88116519572240477.10E-1514passed
/bootstrap_slope_intervalabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

137  jk_outputs <- function(coef, n) list(
138    "/slope" = mcr_point(coef, "Slope"),
139    "/intercept" = mcr_point(coef, "Intercept"),
140    "/interval_method" = exact("jackknife_t", "The configured interval method."),
141    "/jackknife_degrees_of_freedom" = exact(n - 2L, "Tabelz/CLSI n - 2 interval convention; mcr uses the same degrees of freedom."),
142    "/jackknife_slope_se" = mcr_tol(coef["Slope", "SE"], "Linnet leave-one-out slope standard error."),
143    "/jackknife_slope_interval/0" = mcr_tol(coef["Slope", "LCI"], "Slope lower bound, centred on full-data estimate."),
144    "/jackknife_slope_interval/1" = mcr_tol(coef["Slope", "UCI"], "Slope upper bound."),
145    "/jackknife_intercept_se" = mcr_tol(coef["Intercept", "SE"], "Linnet leave-one-out intercept standard error."),
146    "/jackknife_intercept_interval/0" = mcr_tol(coef["Intercept", "LCI"], "Intercept lower bound."),
147    "/jackknife_intercept_interval/1" = mcr_tol(coef["Intercept", "UCI"], "Intercept upper bound."),
148    "/bootstrap_slope_interval" = absent("Jackknife bounds never appear in the bootstrap fields."))
149  export_dataset(data.frame(method_a = c(7, 8.3, 10.5, 9, 5.1, 8.2, 10.2, 10.3, 7.1, 5.9),
150                            method_b = c(7.9, 8.2, 9.6, 9, 6.5, 7.3, 10.2, 10.6, 6.3, 5.2)),
151                 "MSA_DEMING_NCSS", "NCSS Deming Regression Program 1: ten pairs; reproduced by NIST Dataplot",
152                 "NCSS Deming Regression chapter 303, Program 1; https://www.itl.nist.gov/div898/software/dataplot/refman1/auxillar/demfit.htm",
153                 "Small numeric example transcribed for validation; NCSS source and NIST reproduction credited.", c(method_a = "Float64", method_b = "Float64"))
154  dn <- read_dataset("MSA_DEMING_NCSS")
155  jn <- mcr_fit(dn$method_a, dn$method_b, 0.25, jackknife = TRUE)
156  stopifnot(abs(jn["Slope", "EST"] - 1.00119) < 5.1e-6,
157            abs(jn["Slope", "SE"] - 0.18718) < 5.1e-6,
158            abs(jn["Intercept", "SE"] - 1.72199) < 5.1e-6)
159  emit("STAT-MSA-DEMING-007", "Jackknife t intervals on the NCSS ten-pair example", "MSA_DEMING_NCSS",
160       list(x_column = "method_a", y_column = "method_b", lambda_ratio = 0.25, interval_method = "jackknife_t"),
161       jk_outputs(jn, nrow(dn)),
162       c("NCSS-origin example reproduced by NIST Dataplot; delta = Var(e_x)/Var(e_y) = 4 corresponds to lambda = 0.25 here.",
163         "mcr error.ratio = 4; its slope and Linnet jackknife SEs agree with NIST's reproduced five-decimal values within 5.1e-6.",
164         "Interval method, degrees of freedom and bootstrap-field absence are Tabelz contract checks."), oracle = "mcr")

Whole script generators/r/measurement_deming.R


Jackknife t intervals with lambda 2.25 on 40 pairs

Case STAT-MSA-DEMING-008 · status: passed

Reference

Evidence type
R
Source
mcr 1.3.3.1 Deming point estimates and Linnet jackknife intervals; independent R arithmetic for other fields
Citation
mcr 1.3.3.1 mcreg documentation; NCSS Deming Regression chapter 303, Program 1; NIST Dataplot reproduces that NCSS example in case 007
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, mcr 1.3.3.1
Reference script
generators/r/measurement_deming.R · sha256 4f9a1dc380b3c156
Dataset
reference_data/datasets/MSA_DEMING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/slopetolerance1.02852733279456121.02852733279456121.10E-1716passed
/intercepttolerance0.543301580677519880.5433015806775118.88E-1513passed
/interval_methodexactjackknife_tjackknife_tpassed
/jackknife_degrees_of_freedomexact38380.00E+2passed
/jackknife_slope_setolerance0.0206476495403887450.0206476495403884083.37E-1613passed
/jackknife_slope_interval/0tolerance0.986728351566498630.98672835156649936.65E-1615passed
/jackknife_slope_interval/1tolerance1.07032631402262381.07032631402262316.74E-1615passed
/jackknife_intercept_setolerance0.313803006636052480.313803006636050432.05E-1514passed
/jackknife_intercept_interval/0tolerance-0.091959394574533726-0.09195939457453854.77E-1513passed
/jackknife_intercept_interval/1tolerance1.17856255592957341.17856255592956051.29E-1413passed
/bootstrap_slope_intervalabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

165  jm <- mcr_fit(d$method_a, d$method_b, 2.25, jackknife = TRUE)
166  emit("STAT-MSA-DEMING-008", "Jackknife t intervals with lambda 2.25 on 40 pairs", "MSA_DEMING",
167       list(x_column = "method_a", y_column = "method_b", lambda_ratio = 2.25, interval_method = "jackknife_t"),
168       jk_outputs(jm, nrow(d)), c("mcr error.ratio = 1/2.25 on 40 matched pairs; 38 degrees of freedom.",
169                                 "Interval method, degrees of freedom and bootstrap-field absence are Tabelz contract checks."),
170       oracle = "mcr")

Whole script generators/r/measurement_deming.R


Classic crossed structure reproduces the crossed engine

Case STAT-MSA-EXPANDED-001 · status: passed

Reference

Evidence type
R
Source
R stats::anova and lme4::lmer REML
Citation
R stats::anova.lm and lme4::lmer documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, jsonlite 2.0.0
Reference script
generators/r/measurement_expanded.R · sha256 55edc5d437caddcd
Dataset
reference_data/datasets/MSA_GRR_POOLED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/components/[source=Repeatability]/variancetolerance0.256057924521367660.25605792452136724.41E-1614passed
/components/[source=Total Gage R&R]/variancetolerance0.333610087196582920.33361008719658399.95E-1614passed
/components/[source=Part-To-Part]/variancetolerance2.8671386053741712.86713860537417635.31E-1514passed
/pathexactclassic_crossedclassic_crossedpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

54  g <- read_dataset("MSA_GRR_POOLED")
55  g$part <- factor(g$part); g$operator <- factor(g$operator)
56  full <- anova(lm(y ~ part * operator, data = g)); ms <- full[["Mean Sq"]]; df <- full[["Df"]]; ss <- full[["Sum Sq"]]
57  p_int <- pf(ms[3] / ms[4], df[3], df[4], lower.tail = FALSE); if (p_int < 0.05) stop("expected pooling")
58  ms_err <- (ss[3] + ss[4]) / (df[3] + df[4])
59  v_op <- max(0, (ms[2] - ms_err) / 30); v_part <- max(0, (ms[1] - ms_err) / 9)
60  emit("STAT-MSA-EXPANDED-001", "Classic crossed structure reproduces the crossed engine", "MSA_GRR_POOLED",
61       list(response_column = "y", part_column = "part", operator_column = "operator"),
62       c(component("Repeatability", ms_err, "Pooled error mean square.", "1e-9"),
63         component("Total Gage R&R", ms_err + v_op, "Repeatability + operator.", "1e-9"),
64         # The classic structure reproduces the crossed engine's components,
65         # labels included (tests/unit/analysis/test_gage_expanded.py), so the part
66         # component keeps the crossed spelling; until 2026-09-25 this used "Part-to-part".
67         component("Part-To-Part", v_part, "(MS_part - MS_error) / (o r).", "1e-9"),
68         list("/path" = exact("classic_crossed", "Delegated to the crossed engine (line 177-178)."))),
69       c("Dataset written by measurement_gage.R; 10 parts x 3 operators x 3 trials, interaction pooled."))

Whole script generators/r/measurement_expanded.R


Fully crossed part, operator and site

Case STAT-MSA-EXPANDED-002 · status: passed

Reference

Evidence type
R
Source
R stats::anova and lme4::lmer REML
Citation
R stats::anova.lm and lme4::lmer documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, jsonlite 2.0.0
Reference script
generators/r/measurement_expanded.R · sha256 55edc5d437caddcd
Dataset
reference_data/datasets/MSA_EXPANDED_CROSSED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/components/[source=Repeatability]/variancetolerance0.0806205798811851230.080620580857861689.77E-107passed
/components/[source=Reproducibility]/variancetolerance0.987231634443072160.98723143841981481.96E-76passed
/components/[source=Site (measurement system)]/variancetolerance0.0647252725813222680.064725246785473592.58E-86passed
/components/[source=Total Gage R&R]/variancetolerance1.13257748690557961.132577266063152.21E-76passed
/components/[source=Part-to-part]/variancetolerance3.04932802349250843.04932801034637141.31E-88passed
/pathexactreml_expandedreml_expandedpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

72  set.seed(20261028)
73  rows <- expand.grid(trial = 1:2, site = c("S1", "S2", "S3"), operator = c("A", "B", "C"), part = sprintf("P%d", 1:6), stringsAsFactors = FALSE)
74  pe <- rnorm(6, 0, 1.8); oe <- rnorm(3, 0, 0.5); se <- rnorm(3, 0, 0.4)
75  rows$y <- round(30 + pe[match(rows$part, sprintf("P%d", 1:6))] + oe[match(rows$operator, c("A", "B", "C"))] +
76                    se[match(rows$site, c("S1", "S2", "S3"))] + rnorm(nrow(rows), 0, 0.3), 4)
77  export_dataset(rows[c("part", "operator", "site", "y")], "MSA_EXPANDED_CROSSED",
78                 "Constructed: 6 parts x 3 operators x 3 sites x 2 trials, fully crossed, seed 20261028",
79                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
80                 c(part = "String", operator = "String", site = "String", y = "Float64"))
81  e <- read_dataset("MSA_EXPANDED_CROSSED")
82  vc <- reml(y ~ (1 | part) + (1 | operator) + (1 | site), e)
83  emit("STAT-MSA-EXPANDED-002", "Fully crossed part, operator and site", "MSA_EXPANDED_CROSSED",
84       list(response_column = "y", part_column = "part", operator_column = "operator", site_column = "site", structure = "fully_crossed"),
85       c(component("Repeatability", vc[["Residual"]], "REML residual variance."),
86         component("Reproducibility", vc[["operator"]], "The operator component."),
87         component("Site (measurement system)", vc[["site"]], "A crossed site counts toward Gage R&R (line 190-191)."),
88         component("Total Gage R&R", vc[["Residual"]] + vc[["operator"]] + vc[["site"]], "Repeatability + reproducibility + site."),
89         component("Part-to-part", vc[["part"]], "The part component."),
90         list("/path" = exact("reml_expanded", "REML path."))),
91       c("lmer(y ~ (1|part) + (1|operator) + (1|site), REML) with bobyqa rhoend 1e-12."), kind = "axis_coverage")

Whole script generators/r/measurement_expanded.R


Site nested in part counts toward part-to-part

Case STAT-MSA-EXPANDED-003 · status: passed

Reference

Evidence type
R
Source
R stats::anova and lme4::lmer REML
Citation
R stats::anova.lm and lme4::lmer documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, jsonlite 2.0.0
Reference script
generators/r/measurement_expanded.R · sha256 55edc5d437caddcd
Dataset
reference_data/datasets/MSA_EXPANDED_NESTED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/components/[source=Repeatability]/variancetolerance0.073306484690665280.073306483323036561.37E-97passed
/components/[source=Reproducibility]/variancetolerance0.0713407294288583150.071340710016700421.94E-86passed
/components/[source=Part-to-part]/variancetolerance3.92414326269457273.92414315296449261.10E-77passed
/components/[source=Total Gage R&R]/variancetolerance0.144647214119523580.144647193339736972.08E-86passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 94  set.seed(20261029)
 95  rows <- expand.grid(trial = 1:2, operator = c("A", "B", "C"), site = 1:2, part = sprintf("P%d", 1:8), stringsAsFactors = FALSE)
 96  rows$site_id <- sprintf("%s-s%d", rows$part, rows$site)
 97  pe <- setNames(rnorm(8, 0, 2), sprintf("P%d", 1:8)); sw <- setNames(rnorm(16, 0, 0.7), unique(rows$site_id)); oe <- c(A = 0.3, B = -0.2, C = 0.1)
 98  rows$y <- round(12 + pe[rows$part] + sw[rows$site_id] + oe[rows$operator] + rnorm(nrow(rows), 0, 0.25), 4)
 99  export_dataset(data.frame(part = rows$part, operator = rows$operator, site = rows$site_id, y = rows$y), "MSA_EXPANDED_NESTED",
100                 "Constructed: 8 parts, 2 sites within each part, 3 operators, 2 trials, seed 20261029",
101                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
102                 c(part = "String", operator = "String", site = "String", y = "Float64"))
103  n <- read_dataset("MSA_EXPANDED_NESTED")
104  vn <- reml(y ~ (1 | part) + (1 | part:site) + (1 | operator), n)
105  emit("STAT-MSA-EXPANDED-003", "Site nested in part counts toward part-to-part", "MSA_EXPANDED_NESTED",
106       list(response_column = "y", part_column = "part", operator_column = "operator", site_column = "site", structure = "site_nested_in_part"),
107       c(component("Repeatability", vn[["Residual"]], "REML residual variance."),
108         component("Reproducibility", vn[["operator"]], "The operator component."),
109         component("Part-to-part", vn[["part"]] + vn[["part:site"]], "Part plus site-within-part (line 191-192)."),
110         component("Total Gage R&R", vn[["Residual"]] + vn[["operator"]], "Repeatability + reproducibility.")),
111       c("lmer(y ~ (1|part) + (1|part:site) + (1|operator), REML)."), kind = "axis_coverage")

Whole script generators/r/measurement_expanded.R


A part-by-site interaction is confounded when sites nest in parts

Case STAT-MSA-EXPANDED-004 · status: passed

Reference

Evidence type
R
Source
R stats::anova and lme4::lmer REML
Citation
R stats::anova.lm and lme4::lmer documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, jsonlite 2.0.0
Reference script
generators/r/measurement_expanded.R · sha256 55edc5d437caddcd
Dataset
reference_data/datasets/MSA_EXPANDED_NESTED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

112  emit("STAT-MSA-EXPANDED-004", "A part-by-site interaction is confounded when sites nest in parts", "MSA_EXPANDED_NESTED",
113       list(response_column = "y", part_column = "part", operator_column = "operator", site_column = "site",
114            structure = "site_nested_in_part", interactions = list("part_site")), list(),
115       c("Line 183-185: a declaration confounded with the nesting is rejected."), kind = "invalid_input",
116       error = list(error_class = "ValueError", message_contains = "confounded with the declared nesting"))

Whole script generators/r/measurement_expanded.R


Crossed Gage R&R with a pooled interaction and a tolerance

Case STAT-MSA-GRR-001 · status: passed

Reference

Evidence type
R
Source
R stats::aov, stats::lm and stats::t.test
Citation
R stats::aov, stats::anova.lm, stats::lm and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_GRR_POOLED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/components/[source=Total Gage R&R]/variancetolerance0.333610087196582920.33361008719658399.95E-1614passed
/components/[source=Total Gage R&R]/pct_contributiontolerance10.42287662167679910.422876621676811.11E-1414passed
/components/[source=Total Gage R&R]/pct_study_vartolerance32.28448020593919732.284480205939222.10E-1415passed
/components/[source=Total Gage R&R]/study_vartolerance3.46553937202810983.4655393720281155.33E-1514passed
/components/[source=Total Gage R&R]/pct_tolerancetolerance28.87949476690091528.879494766900964.66E-1414passed
/components/[source=Repeatability]/variancetolerance0.256057924521367660.25605792452136724.41E-1614passed
/components/[source=Repeatability]/pct_contributiontolerance7.9999384242725087.9999384242724783.02E-1414passed
/components/[source=Repeatability]/pct_study_vartolerance28.2841623957162828.2841623957162265.37E-1414passed
/components/[source=Repeatability]/study_vartolerance3.03612998449823213.03612998449822952.64E-1515passed
/components/[source=Repeatability]/pct_tolerancetolerance25.30108320415193625.301083204151912.44E-1415passed
/components/[source=Reproducibility]/variancetolerance0.0775521626752152670.077552162675216721.46E-1513passed
/components/[source=Reproducibility]/pct_contributiontolerance2.42293819740429182.42293819740433274.09E-1413passed
/components/[source=Reproducibility]/pct_study_vartolerance15.56579004549493415.5657900454950651.31E-1314passed
/components/[source=Reproducibility]/study_vartolerance1.67089133587667821.67089133587669371.55E-1414passed
/components/[source=Reproducibility]/pct_tolerancetolerance13.92409446563898513.9240944656391151.30E-1314passed
/components/[source=Operator]/variancetolerance0.0775521626752152670.077552162675216721.46E-1513passed
/components/[source=Operator]/pct_contributiontolerance2.42293819740429182.42293819740433274.09E-1413passed
/components/[source=Operator]/pct_study_vartolerance15.56579004549493415.5657900454950651.31E-1314passed
/components/[source=Operator]/study_vartolerance1.67089133587667821.67089133587669371.55E-1414passed
/components/[source=Operator]/pct_tolerancetolerance13.92409446563898513.9240944656391151.30E-1314passed
/components/[source=Part-To-Part]/variancetolerance2.8671386053741712.86713860537417635.31E-1514passed
/components/[source=Part-To-Part]/pct_contributiontolerance89.57712337832319389.57712337832324.92E-1617passed
/components/[source=Part-To-Part]/pct_study_vartolerance94.64519183684039794.64519183684043.96E-1617passed
/components/[source=Part-To-Part]/study_vartolerance10.15957626052731310.1595762605273241.07E-1414passed
/components/[source=Part-To-Part]/pct_tolerancetolerance84.66313550439427684.663135504394368.54E-1414passed
/components/[source=Total Variation]/variancetolerance3.2007486925707543.20074869257076026.22E-1514passed
/components/[source=Total Variation]/pct_contributiontolerance100100.000000000000011.42E-1415passed
/components/[source=Total Variation]/pct_study_vartolerance100100.00.00E+2passed
/components/[source=Total Variation]/study_vartolerance10.73438181417761710.7343818141776271.05E-1415passed
/components/[source=Total Variation]/pct_tolerancetolerance89.45318178481346689.453181784813571.00E-1314passed
/ndcexact440.00E+2passed
/interaction_droppedexactTrueTruepassed
/anova_with_interaction/[source=Operator*Part]/p_valuetail_probability0.904825348490681190.90482534849067922.00E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 96  interval_deferral <- list(list(path = "/components/[source=Operator]/variance_lower",
 97    reason = "The manual names component-specific F/chi-square intervals without a formula for non-repeatability components."))
 98  export_dataset(crossed_data(20261018, 0.05), "MSA_GRR_POOLED", "Constructed crossed study: 10 parts x 3 operators x 3 trials, negligible interaction, seed 20261018",
 99                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
100                 c(part = "String", operator = "String", y = "Float64"))
101  g1 <- read_dataset("MSA_GRR_POOLED")
102  r1 <- crossed_ref(g1, tolerance = 12)
103  if (!r1$pooled) stop("expected a pooled interaction")
104  emit("STAT-MSA-GRR-001", "tabelz.analysis.gage_rr:compute_gage_rr", "gageRrMenuItem", "Crossed Gage R&R with a pooled interaction and a tolerance",
105       "MSA_GRR_POOLED", list(measurement_column = "y", part_column = "part", operator_column = "operator", tolerance = 12),
106       c(component_outputs(r1),
107         list("/anova_with_interaction/[source=Operator*Part]/p_value" = tail_p(r1$p_int, "Interaction F test against repeatability."))),
108       c(sprintf("Interaction p = %.6g >= 0.05, so its SS and df join repeatability (line 126-127).", r1$p_int)),
109       deferred = interval_deferral)

Whole script generators/r/measurement_gage.R


Crossed Gage R&R with a retained interaction

Case STAT-MSA-GRR-002 · status: passed

Reference

Evidence type
R
Source
R stats::aov, stats::lm and stats::t.test
Citation
R stats::aov, stats::anova.lm, stats::lm and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_GRR_INTERACTION.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/components/[source=Total Gage R&R]/variancetolerance2.10105718885186122.1010571888518556.20E-1514passed
/components/[source=Total Gage R&R]/pct_contributiontolerance21.3068512594154121.3068512594153535.68E-1414passed
/components/[source=Total Gage R&R]/pct_study_vartolerance46.15934494705856446.15934494705856.40E-1414passed
/components/[source=Total Gage R&R]/study_vartolerance8.69701436118550938.6970143611854971.24E-1414passed
/components/[source=Repeatability]/variancetolerance0.261627754666669650.2616277546666672.66E-1513passed
/components/[source=Repeatability]/pct_contributiontolerance2.65317083399513012.6531708339951042.62E-1414passed
/components/[source=Repeatability]/pct_study_vartolerance16.288556823718716.288556823718628.15E-1414passed
/components/[source=Repeatability]/study_vartolerance3.06897363429536623.068973634295351.60E-1414passed
/components/[source=Reproducibility]/variancetolerance1.83942943418519181.8394294341851883.79E-1514passed
/components/[source=Reproducibility]/pct_contributiontolerance18.65368042542028318.653680425420253.15E-1414passed
/components/[source=Reproducibility]/pct_study_vartolerance43.18990672069144743.189906720691413.59E-1415passed
/components/[source=Reproducibility]/study_vartolerance8.13753400181325768.137534001813257.12E-1515passed
/components/[source=Operator]/variancetolerance0.350582982374488820.35058298237448751.34E-1514passed
/components/[source=Operator]/pct_contributiontolerance3.55526708133890823.5552670813388961.24E-1414passed
/components/[source=Operator]/pct_study_vartolerance18.85541588334478318.855415883344753.21E-1414passed
/components/[source=Operator]/study_vartolerance3.55260290005533763.5526029000553316.63E-1514passed
/components/[source=Part-To-Part]/variancetolerance7.75988924227982767.759889242279832.63E-1515passed
/components/[source=Part-To-Part]/pct_contributiontolerance78.6931487405845978.693148740584655.68E-1415passed
/components/[source=Part-To-Part]/pct_study_vartolerance88.70915890740064988.709158907400682.87E-1415passed
/components/[source=Part-To-Part]/study_vartolerance16.71394665308208316.713946653082096.71E-1515passed
/components/[source=Total Variation]/variancetolerance9.86094643113168889.8609464311316853.57E-1515passed
/components/[source=Total Variation]/pct_contributiontolerance100100.00.00E+2passed
/components/[source=Total Variation]/pct_study_vartolerance100100.00.00E+2passed
/components/[source=Total Variation]/study_vartolerance18.84128635525559818.8412863552555943.97E-1515passed
/ndcexact220.00E+2passed
/interaction_droppedexactFalseFalsepassed
/components/[source=Operator*Part]/variancetolerance1.4888464518107031.48884645181070052.48E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

110  export_dataset(crossed_data(20261019, 1.2), "MSA_GRR_INTERACTION", "Constructed crossed study with a strong operator-by-part interaction, seed 20261019",
111                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
112                 c(part = "String", operator = "String", y = "Float64"))
113  g2 <- read_dataset("MSA_GRR_INTERACTION")
114  r2 <- crossed_ref(g2)
115  if (r2$pooled) stop("expected a retained interaction")
116  emit("STAT-MSA-GRR-002", "tabelz.analysis.gage_rr:compute_gage_rr", "gageRrMenuItem", "Crossed Gage R&R with a retained interaction",
117       "MSA_GRR_INTERACTION", list(measurement_column = "y", part_column = "part", operator_column = "operator"),
118       c(component_outputs(r2), list("/components/[source=Operator*Part]/variance" = tol(r2$v[["int"]], "(MS_OP - MS_E) / r."))),
119       c(sprintf("Interaction p = %.6g < 0.05, so it stays in reproducibility (line 127-128).", r2$p_int)),
120       kind = "axis_coverage", deferred = interval_deferral)

Whole script generators/r/measurement_gage.R


Interaction alpha 0.99 and a 5.15-sigma study variation

Case STAT-MSA-GRR-003 · status: passed

Reference

Evidence type
R
Source
R stats::aov, stats::lm and stats::t.test
Citation
R stats::aov, stats::anova.lm, stats::lm and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_GRR_POOLED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/components/[source=Total Gage R&R]/variancetolerance0.364675397604940170.3646753976049418.28E-1614passed
/components/[source=Total Gage R&R]/pct_contributiontolerance11.24788330379230711.247883303792313.27E-1515passed
/components/[source=Total Gage R&R]/pct_study_vartolerance33.53786412965546833.537864129655474.07E-1616passed
/components/[source=Total Gage R&R]/study_vartolerance3.11000051977118293.11000051977118643.51E-1514passed
/components/[source=Repeatability]/variancetolerance0.284016703888889170.284016703888888566.15E-1614passed
/components/[source=Repeatability]/pct_contributiontolerance8.76008297420358598.7600829742035523.37E-1414passed
/components/[source=Repeatability]/pct_study_vartolerance29.59743734549257329.597437345492525.37E-1414passed
/components/[source=Repeatability]/study_vartolerance2.74460070481902062.74460070481901753.08E-1514passed
/components/[source=Reproducibility]/variancetolerance0.0806586937160509860.080658693716052431.44E-1513passed
/components/[source=Reproducibility]/pct_contributiontolerance2.48780032958871862.4878003295887594.04E-1413passed
/components/[source=Reproducibility]/pct_study_vartolerance15.77276237565480515.7727623756549351.30E-1314passed
/components/[source=Reproducibility]/study_vartolerance1.46262442345393741.46262442345395051.31E-1414passed
/components/[source=Operator]/variancetolerance0.0806586937160509860.080658693716052431.44E-1513passed
/components/[source=Operator]/pct_contributiontolerance2.48780032958871862.4878003295887594.04E-1413passed
/components/[source=Operator]/pct_study_vartolerance15.77276237565480515.7727623756549351.30E-1314passed
/components/[source=Operator]/study_vartolerance1.46262442345393741.46262442345395051.31E-1414passed
/components/[source=Part-To-Part]/variancetolerance2.87749370884362372.87749370884362854.84E-1514passed
/components/[source=Part-To-Part]/pct_contributiontolerance88.75211669620769388.75211669620772.80E-1617passed
/components/[source=Part-To-Part]/pct_study_vartolerance94.20834182608655394.208341826086551.72E-1617passed
/components/[source=Part-To-Part]/study_vartolerance8.73603610871687238.736036108716887.11E-1515passed
/components/[source=Total Variation]/variancetolerance3.24216910644856383.24216910644856965.79E-1514passed
/components/[source=Total Variation]/pct_contributiontolerance100100.00.00E+2passed
/components/[source=Total Variation]/pct_study_vartolerance100100.00.00E+2passed
/components/[source=Total Variation]/study_vartolerance9.27310250810277919.2731025081027867.15E-1515passed
/ndcexact330.00E+2passed
/interaction_droppedexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

121  r3 <- crossed_ref(g1, alpha = 0.99, m = 5.15)
122  emit("STAT-MSA-GRR-003", "tabelz.analysis.gage_rr:compute_gage_rr", "gageRrMenuItem", "Interaction alpha 0.99 and a 5.15-sigma study variation",
123       "MSA_GRR_POOLED", list(measurement_column = "y", part_column = "part", operator_column = "operator",
124                              interaction_alpha = 0.99, study_var_multiplier = 5.15),
125       component_outputs(r3), c(sprintf("With alpha 0.99 the interaction p = %.6g decides pooling.", r3$p_int)),
126       kind = "threshold", deferred = interval_deferral)

Whole script generators/r/measurement_gage.R


Simultaneous full-model repeatability interval

Case STAT-MSA-GRR-004 · status: passed

Reference

Evidence type
R
Source
R stats::lm full crossed ANOVA and stats::qchisq
Citation
R stats::lm, stats::anova.lm and stats::qchisq documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_GRR_POOLED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/components/[source=Repeatability]/variance_lowertolerance0.187628206654706920.187628206654706365.56E-1614passed
/components/[source=Repeatability]/variance_uppertolerance0.471560614568669380.471560614568668161.22E-1514passed
/confidence_methodexactSimultaneous Bonferroni intervals from full-model ANOVA (balSimultaneous Bonferroni intervals from full-model ANOVA (balpassed
/interval_joint_confidencetolerance0.949999999999999960.954.41E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

127  # STAT-MSA-GRR-004: the simultaneous full-model repeatability interval (#185 rulings

Whole script generators/r/measurement_gage.R


A missing value unbalances the design and is refused

Case STAT-MSA-GRR-005 · status: passed

Reference

Evidence type
R
Source
R stats::aov, stats::lm and stats::t.test
Citation
R stats::aov, stats::anova.lm, stats::lm and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_GRR_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

153  m <- read_dataset("MSA_GRR_POOLED"); m$y[7] <- NA
154  export_dataset(m, "MSA_GRR_MISSING", "MSA_GRR_POOLED with y missing in row 7 (1-based)", "R stats::rnorm documentation",
155                 "Constructed validation data; no third-party license.", c(part = "String", operator = "String", y = "Float64"))
156  emit("STAT-MSA-GRR-005", "tabelz.analysis.gage_rr:compute_gage_rr", "gageRrMenuItem", "A missing value unbalances the design and is refused", "MSA_GRR_MISSING",
157       list(measurement_column = "y", part_column = "part", operator_column = "operator"), list(),
158       c("Line 99-102: every operator measures every part the same number of times unless unbalanced fitting is allowed."),
159       kind = "invalid_input", error = list(error_class = "ValueError", message_contains = "balanced"))

Whole script generators/r/measurement_gage.R


One trial per cell is refused

Case STAT-MSA-GRR-006 · status: passed

Reference

Evidence type
R
Source
R stats::aov, stats::lm and stats::t.test
Citation
R stats::aov, stats::anova.lm, stats::lm and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_GRR_ONETRIAL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

160  two <- read_dataset("MSA_GRR_POOLED"); two <- two[!duplicated(two[c("part", "operator")]), ]
161  export_dataset(two, "MSA_GRR_ONETRIAL", "MSA_GRR_POOLED with one trial per cell", "Hand-constructed",
162                 "Constructed validation data; no third-party license.", c(part = "String", operator = "String", y = "Float64"))
163  emit("STAT-MSA-GRR-006", "tabelz.analysis.gage_rr:compute_gage_rr", "gageRrMenuItem", "One trial per cell is refused", "MSA_GRR_ONETRIAL",
164       list(measurement_column = "y", part_column = "part", operator_column = "operator"), list(),
165       c("Line 100-101: at least two trials in every cell."), kind = "invalid_input",
166       error = list(error_class = "ValueError", message_contains = "trial"))

Whole script generators/r/measurement_gage.R


Gage linearity and bias

Case STAT-MSA-LIN-001 · status: passed

Reference

Evidence type
R
Source
R stats::aov, stats::lm and stats::t.test
Citation
R stats::aov, stats::anova.lm, stats::lm and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_LINEARITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/metrics/[key=average_bias]/valuetolerance0.139259999999999880.139259999999999632.46E-1614passed
/metrics/[key=bias_p]/valuetail_probability4.6562665468056675e-104.6562665468055666e-101.01E-2313passed
/metrics/[key=slope]/valuetolerance0.0119358333333332880.0119358333333331721.16E-1614passed
/metrics/[key=linearity_p]/valuetail_probability0.023564509354125310.023564509354126411.10E-1513passed
/metrics/[key=linearity]/valuetolerance0.0954866666666663030.095486666666665379.29E-1614passed
/metrics/[key=r_squared]/valuetolerance0.170001471363738440.17000147136373622.24E-1513passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

190  set.seed(20261021)
191  ref <- rep(c(2, 4, 6, 8, 10), each = 6)
192  meas <- round(ref + 0.05 + 0.012 * ref + rnorm(30, 0, 0.08), 4)
193  export_dataset(data.frame(reference = ref, y = meas), "MSA_LINEARITY", "Constructed: five reference levels x six measurements, seed 20261021",
194                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.", c(reference = "Float64", y = "Float64"))
195  dl <- read_dataset("MSA_LINEARITY")
196  bias <- dl$y - dl$reference; lf <- summary(lm(bias ~ dl$reference)); tt <- t.test(bias)
197  emit("STAT-MSA-LIN-001", "tabelz.analysis.gage_studies:compute_linearity_bias", "extendedGageMenuItem", "Gage linearity and bias", "MSA_LINEARITY",
198       list(measurement_column = "y", reference_column = "reference", tolerance = 4),
199       list("/metrics/[key=average_bias]/value" = tol(mean(bias), "Mean of y - r."),
200            "/metrics/[key=bias_p]/value" = tail_p(tt$p.value, "One-sample t test of bias against zero."),
201            "/metrics/[key=slope]/value" = tol(coef(lf)[2, 1], "Slope of bias on reference."),
202            "/metrics/[key=linearity_p]/value" = tail_p(coef(lf)[2, 4], "Slope t test."),
203            "/metrics/[key=linearity]/value" = tol(coef(lf)[2, 1] * diff(range(dl$reference)), "Slope times the reference range (line 230)."),
204            "/metrics/[key=r_squared]/value" = tol(lf$r.squared, "R^2 of the bias regression.")),
205       c("Line 225-231."))

Whole script generators/r/measurement_gage.R


Nested destructive Gage R&R

Case STAT-MSA-NESTED-001 · status: passed

Reference

Evidence type
R
Source
R stats::aov, stats::lm and stats::t.test
Citation
R stats::aov, stats::anova.lm, stats::lm and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_NESTED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/metrics/[key=grr_pct]/valuetolerance21.67542612328188821.6754261232819162.81E-1414passed
/metrics/[key=ndc]/valueexact660.00E+2passed
/tables/1/rows/1/3tolerance8.39560437138889888.3956043713888927.09E-1515passed
/tables/0/rows/0/1tolerance0.13573329305555540.135733293055555622.21E-1614passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

169  set.seed(20261020)
170  rows <- expand.grid(trial = 1:3, part = 1:4, operator = c("A", "B", "C"), stringsAsFactors = FALSE)
171  rows$part_id <- sprintf("%s-%d", rows$operator, rows$part)
172  pe <- setNames(rnorm(12, 0, 1.5), unique(rows$part_id)); oe <- c(A = -0.3, B = 0.2, C = 0.5)
173  rows$y <- round(20 + pe[rows$part_id] + oe[rows$operator] + rnorm(nrow(rows), 0, 0.4), 4)
174  export_dataset(data.frame(operator = rows$operator, part = rows$part_id, y = rows$y), "MSA_NESTED",
175                 "Constructed nested study: 3 operators x 4 own parts x 3 trials, seed 20261020", "R stats::rnorm documentation",
176                 "Constructed validation data; no third-party license.", c(operator = "String", part = "String", y = "Float64"))
177  dn <- read_dataset("MSA_NESTED")
178  fit <- anova(lm(y ~ operator + operator:part, data = transform(dn, operator = factor(operator), part = factor(part))))
179  ms_o <- fit[["Mean Sq"]][1]; ms_p <- fit[["Mean Sq"]][2]; ms_e <- fit[["Mean Sq"]][3]
180  v_o <- max(0, (ms_o - ms_p) / (4 * 3)); v_p <- max(0, (ms_p - ms_e) / 3); v_g <- ms_e + v_o
181  emit("STAT-MSA-NESTED-001", "tabelz.analysis.gage_studies:compute_nested_gage_rr", "extendedGageMenuItem", "Nested destructive Gage R&R", "MSA_NESTED",
182       list(measurement_column = "y", part_column = "part", operator_column = "operator"),
183       list("/metrics/[key=grr_pct]/value" = tol(100 * sqrt(v_g / (v_g + v_p)), "100 SD_GRR / SD_total."),
184            "/metrics/[key=ndc]/value" = exact(max(1L, as.integer(floor(sqrt(2) * sqrt(v_p / v_g)))), "floor(sqrt(2) SD_part / SD_GRR), minimum 1."),
185            "/tables/1/rows/1/3" = tol(ms_p, "Part(Operator) mean square."),
186            "/tables/0/rows/0/1" = tol(v_g, "Total Gage R&R variance = repeatability + operator.")),
187       c("lm(y ~ operator + operator:part): Operator, Part within Operator and Repeatability (line 215-217)."))

Whole script generators/r/measurement_gage.R


NIST (52,3) binomial operating-characteristic example

Case STAT-MSA-OC-NIST-001 · status: passed

Reference

Evidence type
R
Source
R stats::pbinom applied to NIST (52,3) binomial OC example
Citation
NIST/SEMATECH e-Handbook 6.2.3.2, Choosing a Sampling Plan with a given OC Curve, https://www.itl.nist.gov/div898/handbook/pmc/section2/pmc232.htm; R stats::Binomial documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_nist_oc.R · sha256 9f719ee61df305c2
Dataset
reference_data/datasets/MSA_NIST_OC_52_3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/probabilities/0tolerance0.998153065260017240.99815306526001724.03E-1817passed
/probabilities/1tolerance0.929537383868120790.92953738386812091.13E-1615passed
/probabilities/2tolerance0.738317410796620610.73831741079662071.11E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

6  run <- generator_run(c("stats", "jsonlite"))
7  id <- "STAT-MSA-OC-NIST-001"

Whole script generators/r/measurement_nist_oc.R


Right-censored Weibull life model

Case STAT-MSA-REL-001 · status: passed

Reference

Evidence type
R
Source
R stats::pbinom, stats::qt (noncentral) and survival::survreg
Citation
R stats::Binomial, stats::TDist and survival::survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/measurement_manufacturing.R · sha256 5a42bd20f26f6b22
Dataset
reference_data/datasets/MSA_RELIABILITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/shapetolerance2.75033300394348412.75033330433706663.00E-76passed
/scaletolerance949.50608045453725949.50604171781463.87E-57passed
/log_likelihoodtolerance-141.76345898726697-141.763458987267262.86E-1314passed
/median_lifetolerance831.04225782406047831.04223601601132.18E-57passed
/failuresexact19190.00E+2passed
/censoredexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

117  set.seed(20261031)
118  life <- round(rweibull(24, shape = 2.2, scale = 900), 2); cens <- life > 1100; life[cens] <- 1100
119  export_dataset(data.frame(hours = life, status = ifelse(cens, "suspended", "failed")), "MSA_RELIABILITY",
120                 "Constructed: 24 Weibull(2.2, 900) lives, Type I censored at 1100 h, seed 20261031",
121                 "R stats::rweibull documentation", "Constructed validation data; no third-party license.",
122                 c(hours = "Float64", status = "String"))
123  rl <- read_dataset("MSA_RELIABILITY")
124  fit <- survival::survreg(survival::Surv(hours, status == "failed") ~ 1, data = rl, dist = "weibull",
125                           control = survival::survreg.control(rel.tolerance = 1e-12, maxiter = 200))
126  shape <- 1 / fit$scale; scale <- exp(unname(coef(fit)))
127  emit("STAT-MSA-REL-001", "tabelz.analysis.manufacturing_studies:reliability_study", "Right-censored Weibull life model", "MSA_RELIABILITY",
128       list(time_column = "hours", event_column = "status", failure_value = "failed"),
129       list("/shape" = tol(shape, "1 / survreg scale.", "1e-6"), "/scale" = tol(scale, "exp(survreg intercept).", "1e-6"),
130            "/log_likelihood" = tol(fit$loglik[2], "Censored Weibull log likelihood.", "1e-6"),
131            "/median_life" = tol(scale * log(2)^(1 / shape), "scale (ln 2)^(1/shape).", "1e-6"),
132            "/failures" = exact(sum(rl$status == "failed"), "Observed failures."), "/censored" = exact(sum(rl$status != "failed"), "Suspensions.")),
133       c("survreg(Surv(hours, failed) ~ 1, dist = 'weibull'), rel.tolerance 1e-12; shape = 1/scale, scale = exp(intercept)."),
134       deferred = list(list(path = "/shape_lower", reason = "No stated interval convention for the life-model parameters.")))

Whole script generators/r/measurement_manufacturing.R


Probit gage response with Fieller intervals

Case STAT-MSA-RESP-001 · status: passed

Reference

Evidence type
R
Source
R stats::glm probit and twopartm::FiellerRatio 0.1.0 for bounded Fieller outputs
Citation
R stats::glm documentation; twopartm 0.1.0 FiellerRatio; Franz (2007), equations (3)-(5)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, irr 0.85, jsonlite 2.0.0, twopartm 0.1.0
Reference script
generators/r/measurement_attribute.R · sha256 beefb4724be48657
Dataset
reference_data/datasets/MSA_GAGE_RESPONSE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance14.4401718286898114.4401716699731221.59E-77passed
/midpoint/estimatetolerance9.9999842698932229.9999842699114311.82E-1111passed
/midpoint/lowertolerance9.98233688512166149.9823368877488452.63E-99passed
/midpoint/uppertolerance10.01763160521747410.0176316025782382.64E-99passed
/lower_point/estimatetolerance9.88607609439052849.8860760931567351.23E-99passed
/upper_point/estimatetolerance10.11389244539591610.113892446666131.27E-99passed
/biastolerance-1.5730106778022446e-05-1.5730088568588485e-051.82E-115passed
/spreadtolerance0.227816351005390130.227816353509394832.50E-97passed
/spread_lowertolerance0.185612631746040170.185612646948871581.52E-87passed
/probit_scaletolerance0.0692512535074682880.069251254268631667.61E-107passed
/pearson_chi_squaretolerance5.21278250715879745.2127824075102519.96E-87passed
/goodness_of_fit_dfexact990.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

173    "/bias" = list(comparison = "tolerance", value = dec(unname(pr$mid$est - threshold)),
174                   absolute_tolerance = dec(1e-6 * abs(pr$mid$est)), relative_tolerance = "0",
175                   rationale = "Iterative probit fit; owner 1e-6 tier on x_0.5, propagated to x_0.5 minus the exact threshold."),
176    "/spread" = tol(pr$spread, "(qnorm(0.95) - qnorm(0.05)) / |b1|.", "1e-6"),
177    "/spread_lower" = tol(pr$spread_lo, "Inverted Wald slope interval.", "1e-6"),
178    "/probit_scale" = tol(pr$scale, "1 / |b1|.", "1e-6"),
179    "/pearson_chi_square" = tol(pr$pearson, "Pearson chi-square over the pooled levels.", "1e-6"),
180    "/goodness_of_fit_df" = exact(9L, "k - 2 = 11 - 2."))
181  emit("STAT-MSA-RESP-001", resp_entry, "attributeGageResponseMenuItem", "Probit gage response with Fieller intervals", "MSA_GAGE_RESPONSE",
182       list(reference_column = "reference", pass_count_column = "passes", trials_column = "trials", threshold = 10),
183       resp_outputs(pr, 10), c("glm(binomial(link = 'probit')) with epsilon 1e-14; normal quantile in the Fieller equation."))

Whole script generators/r/measurement_attribute.R


Modelling the fail event reverses the curve

Case STAT-MSA-RESP-002 · status: passed

Reference

Evidence type
R
Source
R stats::glm probit and twopartm::FiellerRatio 0.1.0 for bounded Fieller outputs
Citation
R stats::glm documentation; twopartm 0.1.0 FiellerRatio; Franz (2007), equations (3)-(5)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, irr 0.85, jsonlite 2.0.0, twopartm 0.1.0
Reference script
generators/r/measurement_attribute.R · sha256 beefb4724be48657
Dataset
reference_data/datasets/MSA_GAGE_RESPONSE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/midpoint/estimatetolerance9.9999842698932229.9999842699114311.82E-1111passed
/spreadtolerance0.22781635100539210.227816353509394832.50E-97passed
/directionexactdecreasingdecreasingpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

184  prf <- probit_ref(g$reference, g$trials - g$passes, g$trials)
185  emit("STAT-MSA-RESP-002", resp_entry, "attributeGageResponseMenuItem", "Modelling the fail event reverses the curve", "MSA_GAGE_RESPONSE",
186       list(reference_column = "reference", pass_count_column = "passes", trials_column = "trials", threshold = 10, modeled_event = "fail"),
187       list("/midpoint/estimate" = tol(prf$mid$est, "The 50% point is unchanged (line 364-365).", "1e-6"),
188            "/spread" = tol(prf$spread, "The spread is unchanged.", "1e-6"),
189            "/direction" = exact("decreasing", "The fail probability falls with the reference.")),
190       c("The fail count n - y is modelled."), kind = "axis_coverage")

Whole script generators/r/measurement_attribute.R


10/90 transition points at 90% confidence

Case STAT-MSA-RESP-003 · status: passed

Reference

Evidence type
R
Source
R stats::glm probit and twopartm::FiellerRatio 0.1.0 for bounded Fieller outputs
Citation
R stats::glm documentation; twopartm 0.1.0 FiellerRatio; Franz (2007), equations (3)-(5)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, irr 0.85, jsonlite 2.0.0, twopartm 0.1.0
Reference script
generators/r/measurement_attribute.R · sha256 beefb4724be48657
Dataset
reference_data/datasets/MSA_GAGE_RESPONSE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/lower_point/estimatetolerance9.91123521754480139.911235216587549.57E-1010passed
/spreadtolerance0.177498104696844040.177498106647784181.95E-97passed
/midpoint/lowertolerance9.9852920736936239.9852920757927622.10E-99passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

191  pr90 <- probit_ref(g$reference, g$passes, g$trials, level = 0.9, lower = 0.1, upper = 0.9)
192  emit("STAT-MSA-RESP-003", resp_entry, "attributeGageResponseMenuItem", "10/90 transition points at 90% confidence", "MSA_GAGE_RESPONSE",
193       list(reference_column = "reference", pass_count_column = "passes", trials_column = "trials", threshold = 10,
194            lower_percentile = 0.1, upper_percentile = 0.9, confidence_level = 0.9),
195       list("/lower_point/estimate" = tol(pr90$low$est, "x_0.10.", "1e-6"), "/spread" = tol(pr90$spread, "Spread between 10% and 90%.", "1e-6"),
196            "/midpoint/lower" = tol(pr90$mid$lo, "90% Fieller limit.", "1e-6")),
197       c("Non-default transition points and confidence."), kind = "axis_coverage")

Whole script generators/r/measurement_attribute.R


Separated outcomes have no finite estimate

Case STAT-MSA-RESP-004 · status: passed

Reference

Evidence type
R
Source
R stats::glm probit and twopartm::FiellerRatio 0.1.0 for bounded Fieller outputs
Citation
R stats::glm documentation; twopartm 0.1.0 FiellerRatio; Franz (2007), equations (3)-(5)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, irr 0.85, jsonlite 2.0.0, twopartm 0.1.0
Reference script
generators/r/measurement_attribute.R · sha256 beefb4724be48657
Dataset
reference_data/datasets/MSA_GAGE_RESPONSE_SEPARATED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

198  sep <- data.frame(reference = c(1, 2, 3, 4), passes = c(0L, 0L, 10L, 10L), trials = rep(10L, 4))
199  export_dataset(sep, "MSA_GAGE_RESPONSE_SEPARATED", "Constructed: outcomes separated along the reference axis", "Hand-constructed",
200                 "Constructed validation data; no third-party license.", c(reference = "Float64", passes = "Int64", trials = "Int64"))
201  emit("STAT-MSA-RESP-004", resp_entry, "attributeGageResponseMenuItem", "Separated outcomes have no finite estimate", "MSA_GAGE_RESPONSE_SEPARATED",
202       list(reference_column = "reference", pass_count_column = "passes", trials_column = "trials", threshold = 2.5), list(),
203       c("Line 367-371: separated outcomes are an error."), kind = "invalid_input",
204       error = list(error_class = "ValueError", message_contains = "separat"))

Whole script generators/r/measurement_attribute.R


Gage run chart means and trial layout on a hand-checked crossed study

Case STAT-MSA-RUNCHART-001 · status: passed

Reference

Evidence type
Analytical
Source
Hand arithmetic means of the constructed readings
Citation
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MSA_RUN_CHART_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/overall_meantolerance11.062511.06250.00E-2passed
/parts/[part=P1]/meantolerance10.02510.0253.55E-1616passed
/parts/[part=P2]/meantolerance12.112.13.55E-1616passed
/reference_sourceexactoverall_meanoverall_meanpassed
/max_trialsexact220.00E+2passed

Convention


Gage stability sorted by an order column

Case STAT-MSA-STAB-001 · status: passed

Reference

Evidence type
R
Source
R stats::aov, stats::lm and stats::t.test
Citation
R stats::aov, stats::anova.lm, stats::lm and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_STABILITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/metrics/[key=mean]/valuetolerance5.11542999999999995.1154299999999998.67E-1615passed
/metrics/[key=sigma]/valuetolerance0.113122283101671660.113122283101671661.23E-1816passed
/metrics/[key=upper]/valuetolerance5.45479684930501475.4547968493050148.49E-1615passed
/metrics/[key=slope]/valuetolerance0.0140142157953281640.0140142157953281422.21E-1714passed
/metrics/[key=drift_p]/valuetail_probability9.4588733955102822e-079.458873395510365e-078.26E-2114passed
/metrics/[key=bias]/valuetolerance0.0154300000000002770.0154299999999993898.88E-1613passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

235  set.seed(20261023)
236  yst <- round(5 + seq(0, 0.3, length.out = 30) + rnorm(30, 0, 0.1), 4)
237  ord <- sample(30)
238  export_dataset(data.frame(order = ord, y = yst[ord]), "MSA_STABILITY", "Constructed: a drifting check standard stored out of order, seed 20261023",
239                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.", c(order = "Int64", y = "Float64"))
240  ds <- read_dataset("MSA_STABILITY"); ys <- ds$y[order(ds$order)]
241  sig <- mean(abs(diff(ys))) / (2 / sqrt(pi)); lf <- summary(lm(ys ~ seq_along(ys)))
242  emit("STAT-MSA-STAB-001", "tabelz.analysis.gage_studies:compute_gage_stability", "extendedGageMenuItem", "Gage stability sorted by an order column", "MSA_STABILITY",
243       list(measurement_column = "y", order_column = "order", reference = 5.1),
244       list("/metrics/[key=mean]/value" = tol(mean(ys), "Centre."),
245            "/metrics/[key=sigma]/value" = tol(sig, "MR-bar / d2 on the sorted sequence."),
246            "/metrics/[key=upper]/value" = tol(mean(ys) + 3 * sig, "Mean + 3 sigma."),
247            "/metrics/[key=slope]/value" = tol(coef(lf)[2, 1], "Drift per observation."),
248            "/metrics/[key=drift_p]/value" = tail_p(coef(lf)[2, 4], "Slope t test."),
249            "/metrics/[key=bias]/value" = tol(mean(ys) - 5.1, "Mean minus the reference.")),
250       c("The order column actively sorts the sequence (line 257-260)."))

Whole script generators/r/measurement_gage.R


One-sided normal upper tolerance bound

Case STAT-MSA-TOL-001 · status: passed

Reference

Evidence type
R
Source
R stats::pbinom, stats::qt (noncentral) and survival::survreg
Citation
R stats::Binomial, stats::TDist and survival::survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/measurement_manufacturing.R · sha256 5a42bd20f26f6b22
Dataset
reference_data/datasets/MSA_TOLERANCE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/factortolerance3.06390112624039773.06390112623816832.23E-1212passed
/uppertolerance112.72846480060377112.728464800594259.52E-1213passed
/nexact30300.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

78  set.seed(20261030)
79  export_dataset(data.frame(x = round(rnorm(30, 100, 4), 4)), "MSA_TOLERANCE", "Constructed: 30 values N(100, 4^2), seed 20261030",
80                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.", c(x = "Float64"))
81  x <- read_dataset("MSA_TOLERANCE")$x; n <- length(x); m <- mean(x); s <- sd(x)
82  tolx <- "tabelz.analysis.manufacturing_studies:tolerance_interval_study"
83  k1 <- qt(0.95, n - 1, ncp = qnorm(0.99) * sqrt(n)) / sqrt(n)
84  emit("STAT-MSA-TOL-001", tolx, "One-sided normal upper tolerance bound", "MSA_TOLERANCE",
85       list(column = "x", method = "normal", content = 0.99, confidence = 0.95, sided = "one_upper"),
86       list("/factor" = tol(k1, "Exact noncentral-t factor."), "/upper" = tol(m + k1 * s, "Mean + k s."), "/n" = exact(30L, "Thirty values.")),
87       c("99% content, 95% confidence, one-sided upper."))

Whole script generators/r/measurement_manufacturing.R


Two-sided normal tolerance interval (Howe)

Case STAT-MSA-TOL-002 · status: passed

Reference

Evidence type
R
Source
R tolerance::K.factor 3.0.0, explicit HE2
Citation
Howe (1969), equation (2.13) and Table 1; R tolerance 3.0.0 K.factor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/measurement_manufacturing.R · sha256 5a42bd20f26f6b22
Dataset
reference_data/datasets/MSA_TOLERANCE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/factortolerance2.84159082729069112.84159082729069164.68E-1615passed
/lowertolerance87.50191373632949987.50191373632952.39E-1617passed
/uppertolerance111.77881959700383111.778819597003831.63E-1516passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

29  emit <- function(id, entry, title, dataset, config, outputs, notes, kind = "nominal", error = NULL, deferred = list()) {
30    write_expected(family, id, list(
31      schema_version = "2.0.0", case_id = id,
32      outputs = if (is.null(error)) outputs else setNames(list(), character()),
33      expected_error = error, deferred_outputs = deferred, generator_run = run))
34    write_case(family, id, list(
35      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = family, entry_point = entry,
36      menu_items = list(menu), title = title, case_kind = kind,
37      dataset = if (is.null(dataset)) NULL else list(path = paste0("reference_data/datasets/", dataset, ".json"), sha256 = ""),
38      excluded_rows = list(), configuration = config, "repeat" = 1L,
39      expected = list(path = paste0("reference_data/expected/", family, "/", id, ".json"), sha256 = ""),
40      oracle = list(
41        oracle_class = "independent_implementation",
42        source = if (id %in% c("STAT-MSA-TOL-002", "STAT-MSA-TOL-004"))
43          "R tolerance::K.factor 3.0.0, explicit HE2" else
44          "R stats::pbinom, stats::qt (noncentral) and survival::survreg",
45        citation = if (id %in% c("STAT-MSA-TOL-002", "STAT-MSA-TOL-004"))
46          "Howe (1969), equation (2.13) and Table 1; R tolerance 3.0.0 K.factor documentation" else
47          "R stats::Binomial, stats::TDist and survival::survreg documentation",
48        generator = list(path = generator, sha256 = ""), independent = TRUE,
49        convention_notes = as.list(notes)),
50      supported = TRUE, unsupported_reason = NULL, requirement_ids = list("REQ-STAT-MSA-001"),
51      risk_ids = list("RISK-NUMERIC-001"), review_reference = NULL))

Whole script generators/r/measurement_manufacturing.R


Distribution-free intervals are two-sided only

Case STAT-MSA-TOL-003 · status: passed

Reference

Evidence type
R
Source
R stats::pbinom, stats::qt (noncentral) and survival::survreg
Citation
R stats::Binomial, stats::TDist and survival::survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/measurement_manufacturing.R · sha256 5a42bd20f26f6b22
Dataset
reference_data/datasets/MSA_TOLERANCE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

111  emit("STAT-MSA-TOL-003", tolx, "Distribution-free intervals are two-sided only", "MSA_TOLERANCE",
112       list(column = "x", method = "distribution_free", content = 0.9, confidence = 0.9, sided = "one_upper"), list(),
113       c("The distribution-free interval is two-sided."), kind = "invalid_input",
114       error = list(error_class = "ValueError", message_contains = "two-sided"))

Whole script generators/r/measurement_manufacturing.R


Howe Table 1 HE2 factor, N=13

Case STAT-MSA-TOL-004 · status: passed

Reference

Evidence type
R
Source
R tolerance::K.factor 3.0.0, explicit HE2
Citation
Howe (1969), equation (2.13) and Table 1; R tolerance 3.0.0 K.factor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0, tolerance 3.0.0
Reference script
generators/r/measurement_manufacturing.R · sha256 5a42bd20f26f6b22
Dataset
reference_data/datasets/MSA_HOWE_TABLE1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/factortolerance3.08208866133823863.082088661338241.29E-1515passed
/lowertolerance-5.0030108513643761-5.0030108513643815.30E-1514passed
/uppertolerance19.00301085136437619.003010851364383.62E-1515passed
/nexact13130.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

29  emit <- function(id, entry, title, dataset, config, outputs, notes, kind = "nominal", error = NULL, deferred = list()) {
30    write_expected(family, id, list(
31      schema_version = "2.0.0", case_id = id,
32      outputs = if (is.null(error)) outputs else setNames(list(), character()),
33      expected_error = error, deferred_outputs = deferred, generator_run = run))
34    write_case(family, id, list(
35      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = family, entry_point = entry,
36      menu_items = list(menu), title = title, case_kind = kind,
37      dataset = if (is.null(dataset)) NULL else list(path = paste0("reference_data/datasets/", dataset, ".json"), sha256 = ""),
38      excluded_rows = list(), configuration = config, "repeat" = 1L,
39      expected = list(path = paste0("reference_data/expected/", family, "/", id, ".json"), sha256 = ""),
40      oracle = list(
41        oracle_class = "independent_implementation",
42        source = if (id %in% c("STAT-MSA-TOL-002", "STAT-MSA-TOL-004"))
43          "R tolerance::K.factor 3.0.0, explicit HE2" else
44          "R stats::pbinom, stats::qt (noncentral) and survival::survreg",
45        citation = if (id %in% c("STAT-MSA-TOL-002", "STAT-MSA-TOL-004"))
46          "Howe (1969), equation (2.13) and Table 1; R tolerance 3.0.0 K.factor documentation" else
47          "R stats::Binomial, stats::TDist and survival::survreg documentation",
48        generator = list(path = generator, sha256 = ""), independent = TRUE,
49        convention_notes = as.list(notes)),
50      supported = TRUE, unsupported_reason = NULL, requirement_ids = list("REQ-STAT-MSA-001"),
51      risk_ids = list("RISK-NUMERIC-001"), review_reference = NULL))

Whole script generators/r/measurement_manufacturing.R


Type-1 gage Cg and Cgk

Case STAT-MSA-TYPE1-001 · status: passed

Reference

Evidence type
R
Source
Minitab Type 1 Gage Study Cg/Cgk formulas; R stats::sd and stats::t.test for sample statistics and bias test
Citation
Minitab Support, Methods and formulas for Type 1 Gage Study, Cg and CgK formula images; R stats::sd and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_TYPE1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/metrics/[key=bias]/valuetolerance0.0192352000000006740.0192351999999988981.78E-1513passed
/metrics/[key=std]/valuetolerance0.0153854328072582510.0153854328072581311.20E-1614passed
/metrics/[key=cg]/valuetolerance1.29993093145646021.29993093145647041.02E-1414passed
/metrics/[key=cgk]/valuetolerance0.883190407243923840.88319040724396924.54E-1413passed
/metrics/[key=bias_p]/valuetail_probability1.8436166959158541e-061.8436166959163587e-065.05E-1912passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

208  set.seed(20261022)
209  export_dataset(data.frame(y = round(10.02 + rnorm(25, 0, 0.015), 5), run = seq_len(25)), "MSA_TYPE1",
210                 "Constructed: 25 repeated measurements of a 10.00 standard, seed 20261022", "R stats::rnorm documentation",
211                 "Constructed validation data; no third-party license.", c(y = "Float64", run = "Int64"))
212  d1 <- read_dataset("MSA_TYPE1")$y
213  s <- sd(d1); B <- mean(d1) - 10; W <- 0.2 * 0.6
214  emit("STAT-MSA-TYPE1-001", "tabelz.analysis.gage_studies:compute_type_1_gage", "extendedGageMenuItem", "Type-1 gage Cg and Cgk", "MSA_TYPE1",
215       list(measurement_column = "y", reference = 10, tolerance = 0.6),
216       list("/metrics/[key=bias]/value" = tol(B, "Mean minus reference."),
217            "/metrics/[key=std]/value" = tol(s, "Sample standard deviation."),
218            "/metrics/[key=cg]/value" = list(comparison = "tolerance", value = dec(W / (6 * s)),
219              absolute_tolerance = "0", relative_tolerance = "1e-9",
220              rationale = "Minitab Type 1 Cg image formula (K/100*T)/SV with T=0.6, K=20, SV=6 sample SD; owner 1e-9 tier."),
221            "/metrics/[key=cgk]/value" = list(comparison = "tolerance", value = dec((W / 2 - abs(B)) / (3 * s)),
222              absolute_tolerance = "0", relative_tolerance = "1e-9",
223              rationale = "Minitab Type 1 CgK image formula (K/200*T-|mean-reference|)/(SV/2), K=20 and SV=6 sample SD; owner 1e-9 tier."),
224            "/metrics/[key=bias_p]/value" = tail_p(t.test(d1, mu = 10)$p.value, "t test of the bias.")),
225       c("Type 1 Tabelz defaults K=20 percent and SV=6s: Minitab Cg=(K/100*T)/SV and Cgk=(K/200*T-|mean-reference|)/(SV/2). T=0.6 response units.",
226         "The numeric Cg/Cgk values are analytic applications of the Minitab image formulas to constructed measurements, not a published numerical example."),
227       oracle_source = "Minitab Type 1 Gage Study Cg/Cgk formulas; R stats::sd and stats::t.test for sample statistics and bias test",
228       oracle_citation = "Minitab Support, Methods and formulas for Type 1 Gage Study, Cg and CgK formula images; R stats::sd and stats::t.test documentation")

Whole script generators/r/measurement_gage.R


Fewer than ten measurements are refused

Case STAT-MSA-TYPE1-002 · status: passed

Reference

Evidence type
R
Source
R stats::aov, stats::lm and stats::t.test
Citation
R stats::aov, stats::anova.lm, stats::lm and stats::t.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_gage.R · sha256 865663d557479688
Dataset
reference_data/datasets/MSA_TYPE1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

229  emit("STAT-MSA-TYPE1-002", "tabelz.analysis.gage_studies:compute_type_1_gage", "extendedGageMenuItem", "Fewer than ten measurements are refused", "MSA_TYPE1",
230       list(measurement_column = "y", reference = 10, tolerance = 0.6), list(),
231       c("Line 240: at least ten measurements; rows 0-15 (0-based) are excluded, leaving nine."), kind = "invalid_input",
232       excluded = 0:15, error = list(error_class = "ValueError", message_contains = "10"))

Whole script generators/r/measurement_gage.R


Multivariate STAT-MV-ALPHA-001

Case STAT-MV-ALPHA-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_core.R · sha256 90e434c114270562
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cronbach_alphatolerance0.942857142857142840.94285714285714281.89E-1817passed
/standardized_alphatolerance0.942857142857142950.94285714285714272.23E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

80  items <- as.matrix(d[c("i1","i2","i3")])
81  k <- ncol(items)
82  alpha <- k/(k-1) * (1-sum(apply(items,2,stats::var))/stats::var(rowSums(items)))
83  rbar <- mean(stats::cor(items)[upper.tri(stats::cor(items))])
84  standard_alpha <- k*rbar/(1+(k-1)*rbar)
85  emit("STAT-MV-ALPHA-001", "tabelz.executor.item_analysis", "itemAnalysisMenuItem",
86       list(items=list("i1","i2","i3")),
87       list("/cronbach_alpha"=tol(alpha, "sample-variance raw Cronbach alpha"),
88            "/standardized_alpha"=tol(standard_alpha, "mean-correlation standardized alpha")),
89       c("Raw alpha is k/(k-1)*(1-sum sample item variances/sample total variance)."))

Whole script generators/r/multivariate_core.R


Special boundary STAT-MV-ALPHA-CONSTANT-001

Case STAT-MV-ALPHA-CONSTANT-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_boundaries.R · sha256 420706593513ebf0
Dataset
reference_data/datasets/MV_TIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cronbach_alphatolerance00.00.00E+215passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

47  items <- as.matrix(d[c("i1","i2")]); k <- ncol(items)
48  alpha <- k/(k-1)*(1-sum(apply(items,2,stats::var))/stats::var(rowSums(items)))
49  emit("STAT-MV-ALPHA-CONSTANT-001","tabelz.executor.item_analysis","itemAnalysisMenuItem",
50       "MV_TIES",list(items=list("i1","i2")),
51       list("/cronbach_alpha"=tol(alpha,"raw alpha with one zero-variance item")),
52       c("One item is constant; total-score variance remains positive, so raw alpha is defined."))

Whole script generators/r/multivariate_boundaries.R


Hard-numeric STAT-MV-ALPHA-HARD-001

Case STAT-MV-ALPHA-HARD-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::prcomp, factanal, hclust, kmeans, manova and chisq.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_hard.R · sha256 06eca51b7d1d3ab6
Dataset
reference_data/datasets/MV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cronbach_alphatolerance0.942857142857142840.94285714285714281.89E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

82  items <- as.matrix(d[c("i1","i2","i3")]); k <- ncol(items)
83  alpha <- k/(k-1)*(1-sum(apply(items,2,stats::var))/stats::var(rowSums(items)))
84  emit("STAT-MV-ALPHA-HARD-001","tabelz.executor.item_analysis","itemAnalysisMenuItem",
85       "MV_HARD",list(items=list("i1","i2","i3")),
86       list("/cronbach_alpha"=tol(alpha,"analytic raw alpha from sample variances")),
87       c("Large identical item offsets must cancel in sample-variance alpha formula."))

Whole script generators/r/multivariate_hard.R


Invalid multivariate request STAT-MV-ALPHA-INVALID-001

Case STAT-MV-ALPHA-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-ALPHA-INVALID-002

Case STAT-MV-ALPHA-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Minimum-sample STAT-MV-ALPHA-MIN-001

Case STAT-MV-ALPHA-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, factanal, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_minimum.R · sha256 bcd4094568b73b37
Dataset
reference_data/datasets/MV_MIN_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cronbach_alphatolerance11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

44  items <- as.matrix(d[c("i1","i3")]); k <- ncol(items)
45  alpha <- k/(k-1)*(1-sum(apply(items,2,stats::var))/stats::var(rowSums(items)))
46  emit("STAT-MV-ALPHA-MIN-001","tabelz.executor.item_analysis","itemAnalysisMenuItem",
47       "MV_MIN_TWO",list(items=list("i1","i3")),
48       list("/cronbach_alpha"=tol(alpha,"two-row two-item sample-variance alpha")),
49       c("Two rows and two items are the documented minimum for item analysis."))

Whole script generators/r/multivariate_minimum.R


Missing and excluded multivariate population STAT-MV-ALPHA-POP-001

Case STAT-MV-ALPHA-POP-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic identities
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_population.R · sha256 8a6d148e9316e21f
Dataset
reference_data/datasets/MV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/excluded_countexact110.00E+2passed
/missing_countexact110.00E+2passed
/cronbach_alphatolerance0.942467827403482230.94246782740348222.16E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

65  itemd <- eligible[stats::complete.cases(eligible[c("i1","i2","i3")]),c("i1","i2","i3")]
66  itemm <- as.matrix(itemd); k <- ncol(itemm)
67  alpha <- k/(k-1)*(1-sum(apply(itemm,2,stats::var))/stats::var(rowSums(itemm)))
68  emit("STAT-MV-ALPHA-POP-001","tabelz.executor.item_analysis","itemAnalysisMenuItem",
69       list(items=list("i1","i2","i3")),
70       list("/excluded_count"=exact(1L,"One stable row ID excluded"),
71            "/missing_count"=exact(1L,"One item-missing row"),
72            "/cronbach_alpha"=tol(alpha,"Analytic raw alpha on complete selected items")),
73       c("Raw alpha uses complete rows and sample variances after exclusions."))

Whole script generators/r/multivariate_population.R


Multivariate option STAT-MV-ALPHA-REVERSE-001

Case STAT-MV-ALPHA-REVERSE-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cronbach_alphatolerance-1.7865168539325846-1.78651685393258331.30E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

69  items <- as.matrix(d[c("i1","i2","i3")]); items[,3] <- 6-items[,3]
70  k <- ncol(items)
71  alpha <- k/(k-1)*(1-sum(apply(items,2,stats::var))/stats::var(rowSums(items)))
72  emit("STAT-MV-ALPHA-REVERSE-001","tabelz.executor.item_analysis","itemAnalysisMenuItem",
73       list(items=list("i1","i2","i3"),reverse_items=list("i3"),scale_low=1,scale_high=5),
74       list("/cronbach_alpha"=tol(alpha,"Analytic sample-variance alpha after reflecting i3 as 6-i3")),
75       c("Reverse coding uses low+high-value once; raw alpha uses sample variances."))

Whole script generators/r/multivariate_options.R


Retained scoring STAT-MV-ALPHA-SCORE-001

Case STAT-MV-ALPHA-SCORE-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic
Citation
R stats::prcomp, kmeans, hclust and matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring.R · sha256 6aa292c3e1b1894b
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/scale_scoretolerance1.33333333333333331.33333333333333334.07E-1716passed
/1/scale_scoretolerance1.33333333333333331.33333333333333334.07E-1716passed
/2/scale_scoreabsentNonepassed
/3/scale_scoretolerance22.00.00E+2passed
/4/scale_scoretolerance3.33333333333333353.33333333333333351.86E-1717passed
/5/scale_scoretolerance3.33333333333333353.33333333333333351.86E-1717passed
/6/scale_scoretolerance3.33333333333333353.33333333333333351.86E-1717passed
/7/scale_scoretolerance44.00.00E+2passed
/8/scale_scoretolerance4.6666666666666674.6666666666666673.73E-1717passed
/9/scale_scoretolerance4.3333333333333334.3333333333333333.73E-1717passed
/10/scale_scoretolerance4.6666666666666674.6666666666666673.73E-1717passed
/11/scale_scoretolerance4.3333333333333334.3333333333333333.73E-1717passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

86  id <- "STAT-MV-ALPHA-SCORE-001"

Whole script generators/r/multivariate_scoring.R


Retained correction STAT-MV-ALPHA-SCORE-HARD-001

Case STAT-MV-ALPHA-SCORE-HARD-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and analytic row-identity join
Citation
R stats::prcomp, scale and base matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_corrections.R · sha256 d22bd7a20d0e25a3
Dataset
reference_data/datasets/MV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/scale_scoretolerance1000001.51000001.50.00E+1passed
/1/scale_scoretolerance1000001.51000001.50.00E+1passed
/2/scale_scoretolerance10000011000001.00.00E+2passed
/3/scale_scoretolerance10000021000002.00.00E+2passed
/4/scale_scoretolerance1000003.51000003.50.00E+1passed
/5/scale_scoretolerance1000003.51000003.50.00E+1passed
/6/scale_scoretolerance10000031000003.00.00E+2passed
/7/scale_scoretolerance10000041000004.00.00E+2passed
/8/scale_scoretolerance10000051000005.00.00E+2passed
/9/scale_scoretolerance10000041000004.00.00E+2passed
/10/scale_scoretolerance1000004.51000004.50.00E+1passed
/11/scale_scoretolerance1000004.51000004.50.00E+1passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

114  x <- as.matrix(hard[c("x","y","z")])
115  mu <- as.vector(as_serialized(colMeans(x))); sd <- as.vector(as_serialized(apply(x,2,stats::sd)))
116  scores <- as.vector(scale(x,center=mu,scale=sd)%*%as_serialized(weights))
117  model <- list(factor_analysis=list(columns=list("x","y","z"),means=nums(mu),
118    scales=nums(sd),weights=rows(weights),row_ids=ids(id,nrow(hard))))
119  emit(id,"operation:tabelz.factor_scores","factorAnalysisMenuItem",
120       "MV_HARD",list(names=list("factor1"),model=model),scores,"factor1",
121       c("Scoring subtracts 1e8-scale means before applying explicit factor weights."),kind="hard_numeric")
122  id <- "STAT-MV-ALPHA-SCORE-HARD-001"

Whole script generators/r/multivariate_scoring_corrections.R


Invalid multivariate request STAT-MV-ALPHA-SCORE-INVALID-001

Case STAT-MV-ALPHA-SCORE-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-ALPHA-SCORE-INVALID-002

Case STAT-MV-ALPHA-SCORE-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Retained boundary STAT-MV-ALPHA-SCORE-MIN-001

Case STAT-MV-ALPHA-SCORE-MIN-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and retained identity join
Citation
R stats::prcomp and base matrix arithmetic documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_boundaries.R · sha256 4cdf60725804542b
Dataset
reference_data/datasets/MV_MIN_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/scale_scoretolerance1.51.50.00E+1passed
/1/scale_scoretolerance3.53.50.00E+1passed

Convention

Reference code

Whole script generators/r/multivariate_scoring_boundaries.R


Retained correction STAT-MV-ALPHA-SCORE-OMIT-CORRECTED-001

Case STAT-MV-ALPHA-SCORE-OMIT-CORRECTED-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and analytic row-identity join
Citation
R stats::prcomp, scale and base matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_corrections.R · sha256 d22bd7a20d0e25a3
Dataset
reference_data/datasets/MV_SCORE_FOUR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/scale_scoretolerance1.51.50.00E+1passed
/1/scale_scoretolerance3.53.50.00E+1passed
/2/scale_scoretolerance55.00.00E+2passed
/3/scale_scoreabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

84  model <- list(kmeans=list(k=2L,seed=11L,restarts=50L,standardize=TRUE,
85    row_ids=ids(id,3L),labels=as.list(c(1L,2L,2L))))
86  emit(id,"operation:tabelz.kmeans_cluster_labels","kmeansClusteringMenuItem",
87       "MV_SCORE_FOUR",list(name="kcluster",model=model),c(1L,2L,2L,1L),"kcluster",
88       c("Three retained labels, one nonmember; the test concerns replay identity, not refitting k-means."),
89       omitted=4L,numeric=FALSE)
90  id <- "STAT-MV-ALPHA-SCORE-OMIT-CORRECTED-001"

Whole script generators/r/multivariate_scoring_corrections.R


Configuration axis STAT-MV-ALPHA-SCORE-SUM-001

Case STAT-MV-ALPHA-SCORE-SUM-001 · status: passed

Reference

Evidence type
R
Source
R stats and matrix arithmetic
Citation
R stats::hclust and base rowSums documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_axes.R · sha256 8e53089af57e269e
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/scale_sumtolerance44.00.00E+2passed
/11/scale_sumtolerance1313.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

28  id <- "STAT-MV-ALPHA-SCORE-SUM-001"

Whole script generators/r/multivariate_axes.R


Multivariate STAT-MV-CA-001

Case STAT-MV-CA-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_core.R · sha256 90e434c114270562
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chi_squaretolerance10.50000000000000210.5000000000000022.24E-1616passed
/total_inertiatolerance0.875000000000000110.87500000000000011.02E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 97  tab <- table(factor(d$group, levels=c("A","B","C")),
 98               factor(d$cat, levels=c("U","V","W")))
 99  chi <- suppressWarnings(stats::chisq.test(tab, correct=FALSE))$statistic
100  emit("STAT-MV-CA-001", "tabelz.executor.correspondence_analysis",
101       "correspondenceAnalysisMenuItem",
102       list(mode="categorical", row_column="group", column_column="cat"),
103       list("/chi_square"=tol(chi, "Pearson chi-square for contingency table"),
104            "/total_inertia"=tol(chi/sum(tab), "CA total inertia = chi-square / n")),
105       c("First-appearance row/column levels; inertia is chi-square divided by total count."))

Whole script generators/r/multivariate_core.R


CA principal-coordinate sign convention

Case STAT-MV-CA-COORD-001 · status: passed

Reference

Evidence type
R
Source
R ca::ca
Citation
R ca package simple correspondence-analysis documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · ca 0.72, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_ca_coordinates.R · sha256 4b100972a8c30fac
Dataset
reference_data/datasets/MV_CA_COORD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/singular_values/0tolerance0.541708363830499030.54170836383049911.10E-1615passed
/row_categories/0/coordinatessign_invariant_sequence['-0.61106269520984557'][0.6110626952098457]passed
/column_categories/0/coordinatessign_invariant_sequence['-0.63200213675150896'][0.6320021367515091]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 1  source("common.R",local=TRUE)
 2  d <- data.frame(label=c("A","B","C"),U=c(8L,1L,2L),
 3                  V=c(2L,5L,1L),W=c(1L,2L,4L))
 4  export_dataset(d,"MV_CA_COORD","Asymmetric positive three-by-three count table",
 5                 "R ca::ca principal-coordinate map with distinct inertias",
 6                 "Constructed validation data; no third-party license.",
 7                 c(label="String",U="Int64",V="Int64",W="Int64"))
 8  d <- read_dataset("MV_CA_COORD")
 9  tab <- as.matrix(d[c("U","V","W")])
10  fit <- ca::ca(tab)
11  stopifnot(fit$sv[1L]>fit$sv[2L],fit$sv[2L]>0)
12  principal_row <- fit$rowcoord[1L,1L]*fit$sv[1L]
13  principal_col <- fit$colcoord[1L,1L]*fit$sv[1L]
14  sign_one <- function(x,note) list(comparison="sign_invariant_sequence",
15    values=list(dec(unname(x))),absolute_tolerance="0",relative_tolerance="1e-9",
16    rationale=note)
17  tol <- function(x,note) list(comparison="tolerance",value=dec(unname(x)),
18    absolute_tolerance="0",relative_tolerance="1e-9",rationale=note)
19  id <- "STAT-MV-CA-COORD-001"

Whole script generators/r/multivariate_ca_coordinates.R


Multivariate option STAT-MV-CA-COUNTS-001

Case STAT-MV-CA-COUNTS-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CA_COUNTS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chi_squaretolerance10.50000000000000210.5000000000000022.24E-1616passed
/total_inertiatolerance0.875000000000000110.87500000000000011.02E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

86  ca <- data.frame(label=c("A","B","C"),U=c(3L,0L,1L),V=c(1L,3L,0L),W=c(0L,1L,3L))
87  export_dataset(ca,"MV_CA_COUNTS","Constructed three-by-three count table",
88                 "R stats::chisq.test and correspondence-analysis inertia identity",
89                 "Constructed validation data; no third-party license.",
90                 c(label="String",U="Int64",V="Int64",W="Int64"))
91  counts <- read_dataset("MV_CA_COUNTS")
92  table <- as.matrix(counts[c("U","V","W")])
93  chi <- suppressWarnings(stats::chisq.test(table,correct=FALSE))$statistic
94  emit("STAT-MV-CA-COUNTS-001","tabelz.executor.correspondence_analysis",
95       "correspondenceAnalysisMenuItem",
96       list(mode="count_table",row_label_column="label",count_columns=list("U","V","W")),
97       list("/chi_square"=tol(chi,"Pearson chi-square from R count table"),
98            "/total_inertia"=tol(chi/sum(table),"CA total inertia equals chi-square/n")),
99       c("Count-table mode uses literal nonnegative integer cell counts."),dataset="MV_CA_COUNTS")

Whole script generators/r/multivariate_options.R


Hard-numeric STAT-MV-CA-HARD-001

Case STAT-MV-CA-HARD-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::prcomp, factanal, hclust, kmeans, manova and chisq.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_hard.R · sha256 06eca51b7d1d3ab6
Dataset
reference_data/datasets/MV_CA_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chi_squaretolerance1.3999981333358222e-051.3999981333358222e-051.76E-2216passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

88  mat <- as.matrix(ca[c("U","V","W")])
89  chi <- suppressWarnings(stats::chisq.test(mat,correct=FALSE))$statistic
90  emit("STAT-MV-CA-HARD-001","tabelz.executor.correspondence_analysis",
91       "correspondenceAnalysisMenuItem","MV_CA_HARD",
92       list(mode="count_table",row_label_column="label",count_columns=list("U","V","W")),
93       list("/chi_square"=tol(chi,"Pearson chi square with million-count baseline")),
94       c("Large balanced cells with small departures; chi-square uses observed-expected squared/expected."))

Whole script generators/r/multivariate_hard.R


Invalid multivariate request STAT-MV-CA-INVALID-001

Case STAT-MV-CA-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-CA-INVALID-002

Case STAT-MV-CA-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Minimum-sample STAT-MV-CA-MIN-001

Case STAT-MV-CA-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, factanal, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_minimum.R · sha256 bcd4094568b73b37
Dataset
reference_data/datasets/MV_MIN_CA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chi_squaretolerance0.666666666666666630.66666666666666663.41E-1918passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

77  ct <- data.frame(label=c("A","B"),U=c(1L,2L),V=c(2L,1L))
78  export_dataset(ct,"MV_MIN_CA","Two-by-two positive count table",
79                 "R chisq.test correspondence-analysis boundary",
80                 "Constructed validation data; no third-party license.",
81                 c(label="String",U="Int64",V="Int64"))
82  ct <- read_dataset("MV_MIN_CA")
83  chi <- suppressWarnings(stats::chisq.test(as.matrix(ct[c("U","V")]),correct=FALSE))$statistic
84  emit("STAT-MV-CA-MIN-001","tabelz.executor.correspondence_analysis",
85       "correspondenceAnalysisMenuItem","MV_MIN_CA",
86       list(mode="count_table",row_label_column="label",count_columns=list("U","V")),
87       list("/chi_square"=tol(chi,"Pearson chi-square of a two-by-two table")),
88       c("Two rows and two columns give one correspondence-analysis dimension."))

Whole script generators/r/multivariate_minimum.R


Missing and excluded multivariate population STAT-MV-CA-POP-001

Case STAT-MV-CA-POP-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic identities
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_population.R · sha256 8a6d148e9316e21f
Dataset
reference_data/datasets/MV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/excluded_countexact110.00E+2passed
/missing_countexact110.00E+2passed
/chi_squaretolerance11.11111111111111211.1111111111111124.93E-1616passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

75  cad <- eligible[stats::complete.cases(eligible[c("group","cat")]),,drop=FALSE]
76  tab <- table(factor(cad$group,levels=c("A","B","C")),
77               factor(cad$cat,levels=c("U","V","W")))
78  chi <- suppressWarnings(stats::chisq.test(tab,correct=FALSE))$statistic
79  emit("STAT-MV-CA-POP-001","tabelz.executor.correspondence_analysis",
80       "correspondenceAnalysisMenuItem",list(mode="categorical",row_column="group",column_column="cat"),
81       list("/excluded_count"=exact(1L,"One stable row ID excluded"),
82            "/missing_count"=exact(1L,"One category-missing row"),
83            "/chi_square"=tol(chi,"R complete-case contingency chi-square")),
84       c("Event rows with either category missing are omitted after exclusion."))

Whole script generators/r/multivariate_population.R


Special boundary STAT-MV-CA-ZERO-001

Case STAT-MV-CA-ZERO-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_boundaries.R · sha256 420706593513ebf0
Dataset
reference_data/datasets/MV_CA_ZERO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chi_squaretolerance33.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

80  ct <- data.frame(label=c("A","B"),U=c(0L,3L),V=c(2L,1L))
81  export_dataset(ct,"MV_CA_ZERO","Two-by-two table with one zero cell and nonzero margins",
82                 "R chisq.test zero-cell correspondence-analysis boundary",
83                 "Constructed validation data; no third-party license.",
84                 c(label="String",U="Int64",V="Int64"))
85  ct <- read_dataset("MV_CA_ZERO")
86  chi <- suppressWarnings(stats::chisq.test(as.matrix(ct[c("U","V")]),correct=FALSE))$statistic
87  emit("STAT-MV-CA-ZERO-001","tabelz.executor.correspondence_analysis",
88       "correspondenceAnalysisMenuItem","MV_CA_ZERO",
89       list(mode="count_table",row_label_column="label",count_columns=list("U","V")),
90       list("/chi_square"=tol(chi,"Pearson chi-square with one zero observed cell")),
91       c("A zero observed cell is valid because both row and column margins remain positive."))

Whole script generators/r/multivariate_boundaries.R


Canonical correlation of three and two variables

Case STAT-MV-CANCOR-001 · status: passed

Reference

Evidence type
R
Source
R stats::cancor plus closed-form Bartlett test, structure correlations and redundancy
Citation
R Core Team, stats package help: cancor; Bartlett (1941) sequential chi-square; Stewart and Love (1968) redundancy index.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_canonical.R · sha256 32ebf68ddffb5b14
Dataset
reference_data/datasets/MV_CANCOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact19190.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed
/bartlett_multipliertolerance1515.00.00E+2passed
/roots/0/correlationtolerance0.798075595155631870.79807559515563212.19E-1615passed
/roots/0/squaredtolerance0.636924655583016030.63692465558301643.34E-1615passed
/roots/0/wilks_lambdatolerance0.342345663428911480.34234566342891122.73E-1615passed
/roots/0/chi_squaretolerance16.07901510846511416.079015108465131.45E-1415passed
/roots/0/dfexact660.00E+2passed
/roots/0/p_valuetail_probability0.0133360869200753480.0133360869200752687.98E-1714passed
/redundancy/0/first_own_variancetolerance0.363867153963214450.363867153963214561.12E-1615passed
/redundancy/0/first_redundancytolerance0.231755961715992640.231755961715992831.91E-1615passed
/redundancy/0/second_own_variancetolerance0.52326500834953360.52326500834953382.22E-1615passed
/redundancy/0/second_redundancytolerance0.333280385221670690.3332803852216713.37E-1614passed
/roots/1/correlationtolerance0.238944990733285590.23894499073328571.16E-1615passed
/roots/1/squaredtolerance0.0570947085965299380.057094708596529995.50E-1715passed
/roots/1/wilks_lambdatolerance0.942905291403470010.942905291403473.91E-1817passed
/roots/1/chi_squaretolerance0.881841520301189560.88184152030118963.11E-1817passed
/roots/1/dfexact220.00E+2passed
/roots/1/p_valuetail_probability0.643443690934103360.64344369093410331.08E-1615passed
/redundancy/1/first_own_variancetolerance0.308893748764026640.30889374876402651.63E-1615passed
/redundancy/1/first_redundancytolerance0.0176361985729718320.017636198572971846.88E-1815passed
/redundancy/1/second_own_variancetolerance0.476734991650466340.476734991650466572.26E-1615passed
/redundancy/1/second_redundancytolerance0.027219045426052510.0272190454260525483.81E-1714passed
/loadings/0/variableexactx1x1passed
/loadings/0/rawsequence['0.38602700140414303', '0.17398644948956274'][0.38602700140414303, 0.17398644948956274]passed
/loadings/0/standardizedsequence['1.136293706597068', '0.51213958316138908'][1.136293706597068, 0.5121395831613891]passed
/loadings/0/own_structuresequence['0.95223722557777302', '-0.17059829954689215'][0.9522372255777731, -0.17059829954689223]passed
/loadings/0/cross_structuresequence['0.75995729053232886', '-0.040763609104346392'][0.7599572905323292, -0.040763609104346406]passed
/loadings/1/variableexactx2x2passed
/loadings/1/rawsequence['0.19015806504951616', '0.75662662264390179'][0.19015806504951618, 0.7566266226439019]passed
/loadings/1/standardizedsequence['0.29794460641193748', '1.1855022884552877'][0.29794460641193754, 1.185502288455288]passed
/loadings/1/own_structuresequence['-0.40292602124656329', '0.90068286005104004'][-0.40292602124656335, 0.9006828600510398]passed
/loadings/1/cross_structuresequence['-0.32156542421004181', '0.21521365764852493'][-0.32156542421004186, 0.21521365764852488]passed
/loadings/2/variableexactx3x3passed
/loadings/2/rawsequence['0.17423645579878752', '-0.045857031487328781'][0.17423645579878744, -0.04585703148732876]passed
/loadings/2/standardizedsequence['0.25354369011812677', '-0.066729783545341023'][0.25354369011812666, -0.066729783545341]passed
/loadings/2/own_structuresequence['0.14998783122641271', '-0.29385005035573397'][0.14998783122641257, -0.29385005035573397]passed
/loadings/2/cross_structuresequence['0.11970162767212184', '-0.070213997559226332'][0.11970162767212184, -0.07021399755922635]passed
/loadings/3/variableexacty1y1passed
/loadings/3/rawsequence['0.5077080195901108', '-0.11220270973072193'][0.5077080195901109, -0.11220270973072197]passed
/loadings/3/standardizedsequence['0.99869338568718635', '-0.22070973815760844'][0.9986933856871866, -0.22070973815760853]passed
/loadings/3/own_structuresequence['0.99998199054457437', '-0.0060015486760309107'][0.9999819905445746, -0.006001548676030933]passed
/loadings/3/cross_structuresequence['0.79806122224877463', '-0.0014340399927795303'][0.7980612222487747, -0.0014340399927796343]passed
/loadings/4/variableexacty2y2passed
/loadings/4/rawsequence['-0.005015652493060966', '-0.83571131963358025'][-0.005015652493060886, -0.8357113196335805]passed
/loadings/4/standardizedsequence['-0.0061383294950053055', '-1.0227725006295123'][-0.0061383294950052075, -1.0227725006295125]passed
/loadings/4/own_structuresequence['-0.21579164785871111', '-0.97643943217919149'][-0.21579164785871102, -0.9764394321791917]passed
/loadings/4/cross_structuresequence['-0.17221804779445546', '-0.23331531107367159'][-0.1722180477944555, -0.23331531107367162]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

28  emit <- function(id, title, kind, configuration, outputs, notes, excluded = list()) {
29    expected <- list(schema_version = "2.0.0", case_id = id, outputs = outputs,
30                     expected_error = NULL, deferred_outputs = list(), generator_run = run)
31    write_expected("multivariate", id, expected)
32    case <- list(
33      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "multivariate",
34      entry_point = "tabelz.analysis.canonical_correlation:compute_canonical_correlation",
35      menu_items = list("canonicalCorrelationMenuItem"), title = title, case_kind = kind,
36      dataset = list(path = "reference_data/datasets/MV_CANCOR.json", sha256 = ""),
37      excluded_rows = excluded, configuration = configuration, "repeat" = 1L,
38      expected = list(path = paste0("reference_data/expected/multivariate/", id, ".json"),
39                      sha256 = ""),
40      oracle = list(oracle_class = "independent_implementation",
41                    source = "R stats::cancor plus closed-form Bartlett test, structure correlations and redundancy",
42                    citation = "R Core Team, stats package help: cancor; Bartlett (1941) sequential chi-square; Stewart and Love (1968) redundancy index.",
43                    generator = list(path = "generators/r/multivariate_canonical.R", sha256 = ""),
44                    independent = TRUE,
45                    convention_notes = c(notes, "Expected computed from the exported CSV.", manual)),
46      supported = TRUE, unsupported_reason = NULL,
47      requirement_ids = list("REQ-STAT-MV-CANCOR-001"), risk_ids = list("RISK-REFERENCE-001"),
48      review_reference = NULL
49    )
50    write_case("multivariate", id, case)

Whole script generators/r/multivariate_canonical.R


Canonical correlation of one variable against one

Case STAT-MV-CANCOR-SINGLE-001 · status: passed

Reference

Evidence type
R
Source
R stats::cancor plus closed-form Bartlett test, structure correlations and redundancy
Citation
R Core Team, stats package help: cancor; Bartlett (1941) sequential chi-square; Stewart and Love (1968) redundancy index.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_canonical.R · sha256 32ebf68ddffb5b14
Dataset
reference_data/datasets/MV_CANCOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact000.00E+2passed
/bartlett_multipliertolerance17.517.50.00E+1passed
/roots/0/correlationtolerance0.760345861462430550.76034586146243023.33E-1615passed
/roots/0/squaredtolerance0.578125829043045660.57812582904304515.57E-1615passed
/roots/0/wilks_lambdatolerance0.421874170956954340.42187417095695495.57E-1614passed
/roots/0/chi_squaretolerance15.10334319368603515.1033431936860122.31E-1414passed
/roots/0/dfexact110.00E+2passed
/roots/0/p_valuetail_probability0.000101782873608619250.000101782873608620531.28E-1813passed
/redundancy/0/first_own_variancetolerance11.00000000000000044.44E-1615passed
/redundancy/0/first_redundancytolerance0.578125829043045660.57812582904304533.35E-1615passed
/redundancy/0/second_own_variancetolerance10.99999999999999982.22E-1615passed
/redundancy/0/second_redundancytolerance0.578125829043045660.5781258290430456.68E-1614passed
/loadings/0/variableexactx1x1passed
/loadings/0/rawsequence['0.34718947895486757'][0.3471894789548675]passed
/loadings/0/standardizedsequence['1.0000000000000002'][0.9999999999999999]passed
/loadings/0/own_structuresequence['1'][1.0000000000000002]passed
/loadings/0/cross_structuresequence['0.76034586146243044'][0.7603458614624308]passed
/loadings/1/variableexacty1y1passed
/loadings/1/rawsequence['0.52165474526221467'][0.5216547452622147]passed
/loadings/1/standardizedsequence['1.0000000000000002'][1.0]passed
/loadings/1/own_structuresequence['1'][0.9999999999999999]passed
/loadings/1/cross_structuresequence['0.76034586146243033'][0.7603458614624306]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

173  ref <- reference(d, "x1", "y1")
174  stopifnot(ref$N == 20, ref$k == 1, abs(ref$roots[[1]]$correlation - abs(cor(d$x1, d$y1))) < 1e-12)
175  emit("STAT-MV-CANCOR-SINGLE-001", "Canonical correlation of one variable against one", "boundary",
176       list(first = list("x1"), second = list("y1")),
177       outputs_for(ref, 0L, 0L),
178       list("With one variable in each set the canonical correlation is |cor(x1, y1)| (checked in the generator), the structure correlations are +1 and +/-r, and Bartlett's test has one degree of freedom.",
179            sign_note))

Whole script generators/r/multivariate_canonical.R


Canonical correlation of two and two variables with an excluded row

Case STAT-MV-CANCOR-SQUARE-001 · status: passed

Reference

Evidence type
R
Source
R stats::cancor plus closed-form Bartlett test, structure correlations and redundancy
Citation
R Core Team, stats package help: cancor; Bartlett (1941) sequential chi-square; Stewart and Love (1968) redundancy index.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_canonical.R · sha256 32ebf68ddffb5b14
Dataset
reference_data/datasets/MV_CANCOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact19190.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact110.00E+2passed
/bartlett_multipliertolerance15.515.50.00E+1passed
/roots/0/correlationtolerance0.781521012613154430.78152101261315494.45E-1615passed
/roots/0/squaredtolerance0.610775093155890290.6107750931558916.68E-1614passed
/roots/0/wilks_lambdatolerance0.389047659486489170.389047659486488566.08E-1614passed
/roots/0/chi_squaretolerance14.63282808607828814.6328280860783122.44E-1414passed
/roots/0/dfexact440.00E+2passed
/roots/0/p_valuetail_probability0.00552659776274700040.0055265977627469415.90E-1713passed
/redundancy/0/first_own_variancetolerance0.491122117039063940.4911221170390645.90E-1715passed
/redundancy/0/first_redundancytolerance0.299965156785452320.29996515678545273.93E-1614passed
/redundancy/0/second_own_variancetolerance0.507364668799293210.50736466879929331.07E-1615passed
/redundancy/0/second_redundancytolerance0.309885702849895750.309885702849896143.89E-1614passed
/roots/1/correlationtolerance0.0213397620779898840.0213397620779897621.22E-1614passed
/roots/1/squaredtolerance0.000455385445545215130.000455385445545209935.20E-1813passed
/roots/1/wilks_lambdatolerance0.999544614554454760.99954461455445482.18E-1817passed
/roots/1/chi_squaretolerance0.00706008205729324010.00706008205729274055.00E-1613passed
/roots/1/dfexact110.00E+2passed
/roots/1/p_valuetail_probability0.933037115892309310.93303711589231152.22E-1514passed
/redundancy/1/first_own_variancetolerance0.508877882960936060.50887788296093564.48E-1615passed
/redundancy/1/first_redundancytolerance0.00023173558146027170.000231735581460268852.85E-1813passed
/redundancy/1/second_own_variancetolerance0.492635331200706740.49263533120070661.12E-1615passed
/redundancy/1/second_redundancytolerance0.000224338959790148460.000224338959790145862.60E-1813passed
/loadings/0/variableexactx1x1passed
/loadings/0/rawsequence['0.33706444282438358', '-0.037295061810982348'][0.33706444282438347, -0.037295061810982355]passed
/loadings/0/standardizedsequence['0.99593811401855303', '-0.11019724658880013'][0.9959381140185527, -0.11019724658880015]passed
/loadings/0/own_structuresequence['0.984961733639663', '-0.1727726345968868'][0.984961733639663, -0.1727726345968868]passed
/loadings/0/cross_structuresequence['0.76976829145927794', '-0.0036869269158850181'][0.7697682914592783, -0.0036869269158850432]passed
/loadings/1/variableexactx3x3passed
/loadings/1/rawsequence['0.11689898960892811', '0.66643095264815411'][0.11689898960892812, 0.666430952648154]passed
/loadings/1/standardizedsequence['0.17312095249209869', '0.98694746360625163'][0.17312095249209872, 0.9869474636062514]passed
/loadings/1/own_structuresequence['0.10997553065876746', '0.99393429493921914'][0.10997553065876749, 0.9939342949392187]passed
/loadings/1/cross_structuresequence['0.085948188083108987', '0.021210321375157451'][0.08594818808310899, 0.021210321375157503]passed
/loadings/2/variableexacty1y1passed
/loadings/2/rawsequence['0.5180819336156407', '-0.079552636785159392'][0.5180819336156407, -0.07955263678515939]passed
/loadings/2/standardizedsequence['1.0184914743619644', '-0.15639163821684121'][1.0184914743619642, -0.1563916382168412]passed
/loadings/2/own_structuresequence['0.99583843042680975', '0.091136274254918617'][0.9958384304268099, 0.0911362742549186]passed
/loadings/2/cross_structuresequence['0.77826865854625527', '0.0019448264092744717'][0.7782686585462553, 0.0019448264092744487]passed
/loadings/3/variableexacty2y2passed
/loadings/3/rawsequence['0.080711847017508415', '-0.88193158770067281'][0.08071184701750833, -0.8819315877006728]passed
/loadings/3/standardizedsequence['0.093909429157011642', '-1.0261404616170045'][0.09390942915701155, -1.0261404616170045]passed
/loadings/3/own_structuresequence['-0.15177337738765148', '-0.98841531853586007'][-0.15177337738765162, -0.98841531853586]passed
/loadings/3/cross_structuresequence['-0.11861408358371597', '-0.021092547731795766'][-0.11861408358371597, -0.021092547731795724]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

164  ref <- reference(d[-1, ], c("x1", "x3"), c("y1", "y2"))
165  stopifnot(ref$N == 19, ref$k == 2)
166  emit("STAT-MV-CANCOR-SQUARE-001", "Canonical correlation of two and two variables with an excluded row", "axis_coverage",
167       list(first = list("x1", "x3"), second = list("y1", "y2")),
168       outputs_for(ref, 0L, 1L),
169       list("stats::cancor on rows 2-20 (row 1 is excluded by identity; no selected value is missing): p = q = 2, so both roots have a square set of coefficients.",
170            sign_note),
171       excluded = list(0L))

Whole script generators/r/multivariate_canonical.R


Categorical PCA with a binary nominal variable, one dimension

Case STAT-MV-CATPCA-NOMINAL-001 · status: passed

Reference

Evidence type
R
Source
R stats::prcomp / eigen(cor(.)) on the uniquely quantified variables
Citation
Gifi (1990) Nonlinear Multivariate Analysis, ch. 4 (princals single quantification); Linting, Meulman, Groenen and van der Kooij (2007) Nonlinear principal components analysis, Psychological Methods 12(3) (Cronbach's alpha of a CATPCA dimension); R Core Team, stats package help: prcomp.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_catpca.R · sha256 a00047c7f5c09357
Dataset
reference_data/datasets/MV_CATPCA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact18180.00E+2passed
/missing_countexact220.00E+2passed
/convergedexactTrueTruepassed
/losstolerance0.563633093831934720.56363309383193492.19E-1615passed
/dimensions/0/eigenvaluetolerance1.74546762467226091.745467624672268.79E-1615passed
/dimensions/0/percenttolerance43.63669061680652643.63669061680652.88E-1415passed
/dimensions/0/alphatolerance0.569450168451551160.56945016845155074.40E-1615passed
/variables/0/variableexactsitesitepassed
/variables/0/levelexactnominalnominalpassed
/variables/0/loadingssign_invariant_sequence['-0.27487959813455592'][-0.2748795981345563]passed
/variables/1/loadingssign_invariant_sequence['0.88504957226925862'][0.8850495722692578]passed
/variables/2/loadingssign_invariant_sequence['-0.86423325038628296'][-0.8642332503862818]passed
/variables/3/loadingssign_invariant_sequence['0.37376058480584778'][0.37376058480584823]passed
/categories/0/categoryexacteasteastpassed
/categories/0/countexact990.00E+2passed
/artifact/variables/0/quantificationssign_invariant_sequence['-1', '1'][-1.0, 1.0]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

183  outputs[["/categories/0/category"]] <- exact("east", "categories in code-point order")
184  outputs[["/categories/0/count"]] <- exact(as.integer(counts[["east"]]), "rows at east")
185  outputs[["/artifact/variables/0/quantifications"]] <- sign_seq(c(y_east, y_west), "(-sqrt(n_west/n_east), sqrt(n_east/n_west)) up to a common sign: the pair flips together", rel = "1e-8")
186  emit("STAT-MV-CATPCA-NOMINAL-001", "Categorical PCA with a binary nominal variable, one dimension", "axis_coverage",
187       list(nominal = list("site"), numeric = list("x1", "x2", "x3"), dimensions = 1L),
188       outputs,
189       list("A binary nominal variable has one centred normalized quantification up to sign, so the optimum is the first principal component of cor(quantified site, x1, x2, x3) on the 18 rows complete on all four variables.",
190            "Because the nominal sign is free, the site quantifications and every loading (one per variable in one dimension) are compared sign-invariantly; eigenvalue, percent, alpha and loss are sign-free."))

Whole script generators/r/multivariate_catpca.R


Categorical PCA of numeric variables equals correlation PCA

Case STAT-MV-CATPCA-NUMERIC-001 · status: passed

Reference

Evidence type
R
Source
R stats::prcomp / eigen(cor(.)) on the uniquely quantified variables
Citation
Gifi (1990) Nonlinear Multivariate Analysis, ch. 4 (princals single quantification); Linting, Meulman, Groenen and van der Kooij (2007) Nonlinear principal components analysis, Psychological Methods 12(3) (Cronbach's alpha of a CATPCA dimension); R Core Team, stats package help: prcomp.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_catpca.R · sha256 a00047c7f5c09357
Dataset
reference_data/datasets/MV_CATPCA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact19190.00E+2passed
/missing_countexact110.00E+2passed
/convergedexactTrueTruepassed
/losstolerance1.11984938523514991.11984938523515034.19E-1615passed
/dimensions/0/eigenvaluetolerance1.69008384870068771.6900838487006876.36E-1615passed
/dimensions/0/percenttolerance56.3361282900229256.3361282900229061.43E-1415passed
/dimensions/0/alphatolerance0.612470069959440980.61247006995944082.25E-1615passed
/dimensions/1/eigenvaluetolerance0.950367995593862760.95036799559386151.22E-1514passed
/dimensions/1/percenttolerance31.67893318646209331.678933186462054.30E-1414passed
/dimensions/1/alphatolerance-0.078335978225661074-0.078335978225663072.00E-1513passed
/variables/0/variableexactx1x1passed
/variables/0/levelexactnumericnumericpassed
/variables/0/loadingssequence['0.85707623106730413', '-0.31617448641347418'][0.8570762310673037, -0.3161744864134746]passed
/variables/1/variableexactx2x2passed
/variables/1/levelexactnumericnumericpassed
/variables/1/loadingssequence['-0.89822253197929802', '0.092223177017479313'][-0.8982225319792981, 0.09222317701747873]passed
/variables/2/variableexactx3x3passed
/variables/2/levelexactnumericnumericpassed
/variables/2/loadingssequence['0.38561699377082914', '0.91754922230681513'][0.38561699377082925, 0.9175492223068142]passed
/categories/0/variableexactx1x1passed
/categories/0/quantificationtolerance-1.6581154019395148-1.65811540193951481.19E-1717passed
/categories/0/countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

122  sd_pop <- function(v) sqrt(mean((v - mean(v))^2))
123  stopifnot(!anyDuplicated(q[, "x1"]))
124  first_x1 <- sort(unique(q[, "x1"]))[1]
125  outputs[["/categories/0/variable"]] <- exact("x1", "the first category row belongs to x1")
126  outputs[["/categories/0/quantification"]] <- tol((first_x1 - mean(q[, "x1"])) / sd_pop(q[, "x1"]), "numeric quantification: (x - mean)/population SD of the smallest x1 value", rel = "1e-8")
127  outputs[["/categories/0/count"]] <- exact(1L, "distinct x1 values occur once")
128  emit("STAT-MV-CATPCA-NUMERIC-001", "Categorical PCA of numeric variables equals correlation PCA", "nominal",
129       list(numeric = list("x1", "x2", "x3"), dimensions = 2L),
130       outputs,
131       list("With numeric scaling levels only, each quantification is the linear standardization, so the princals optimum is the PCA of the Pearson correlation matrix: eigen(cor(x)) cross-checked with prcomp(scale. = TRUE) on the 19 complete rows.",
132            "Loss at the optimum: dimensions - (lambda_1 + lambda_2)/m. Sign convention replicated from the product: the loadings of each dimension sum to a positive value.",
133            "The eigenvalues are distinct, so the loadings are unique up to the stated sign."))

Whole script generators/r/multivariate_catpca.R


Categorical PCA with a binary ordinal variable

Case STAT-MV-CATPCA-ORDINAL-001 · status: passed

Reference

Evidence type
R
Source
R stats::prcomp / eigen(cor(.)) on the uniquely quantified variables
Citation
Gifi (1990) Nonlinear Multivariate Analysis, ch. 4 (princals single quantification); Linting, Meulman, Groenen and van der Kooij (2007) Nonlinear principal components analysis, Psychological Methods 12(3) (Cronbach's alpha of a CATPCA dimension); R Core Team, stats package help: prcomp.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_catpca.R · sha256 a00047c7f5c09357
Dataset
reference_data/datasets/MV_CATPCA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/missing_countexact000.00E+2passed
/convergedexactTrueTruepassed
/losstolerance1.09078980680955921.09078980680955874.74E-1615passed
/dimensions/0/eigenvaluetolerance2.335993538650522.335993538650524.83E-1716passed
/dimensions/0/percenttolerance77.8664512883506677.866451288350663.20E-1617passed
/dimensions/0/alphatolerance0.857874936218130490.85787493621813054.95E-1918passed
/dimensions/1/eigenvaluetolerance0.391637040920802540.3916370409208034.41E-1614passed
/dimensions/1/percenttolerance13.05456803069341713.0545680306934321.45E-1414passed
/dimensions/1/alphatolerance-2.3300769418368996-2.33007694183689474.87E-1514passed
/variables/0/variableexactratingratingpassed
/variables/0/levelexactordinalordinalpassed
/variables/0/categoriesexact220.00E+2passed
/variables/0/loadingssequence['0.90719547188237959', '0.04388715389750688'][0.9071954718823804, 0.04388715389750762]passed
/variables/1/loadingssequence['0.87414237656941896', '0.41559888847014059'][0.8741423765694183, 0.4155988884701403]passed
/variables/2/loadingssequence['-0.86536987463869763', '0.46582026850059105'][-0.8653698746386972, 0.46582026850059166]passed
/categories/0/categoryexactlowlowpassed
/categories/0/countexact13130.00E+2passed
/categories/0/quantificationtolerance-0.73379938570534275-0.73379938570534284.88E-1817passed
/categories/1/categoryexacthighhighpassed
/categories/1/quantificationtolerance1.36277028773849371.36277028773849373.52E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

155  outputs[["/categories/0/category"]] <- exact("low", "categories in the declared order")
156  outputs[["/categories/0/count"]] <- exact(as.integer(counts[["low"]]), "rows rated low")
157  outputs[["/categories/0/quantification"]] <- tol(y_low, "-sqrt(n_high/n_low): the centred, normalized non-decreasing quantification of a binary ordinal variable", rel = "1e-8")
158  outputs[["/categories/1/category"]] <- exact("high", "categories in the declared order")
159  outputs[["/categories/1/quantification"]] <- tol(y_high, "sqrt(n_low/n_high)", rel = "1e-8")
160  emit("STAT-MV-CATPCA-ORDINAL-001", "Categorical PCA with a binary ordinal variable", "axis_coverage",
161       list(ordinal = list("rating"), orders = list("rating: low < high"), numeric = list("x1", "x2"),
162            dimensions = 2L),
163       outputs,
164       list("A two-category variable has exactly one centred quantification with sum of squared values n (up to sign); the ordinal restriction low < high fixes the sign, so the quantification is -sqrt(n_high/n_low) for low and sqrt(n_low/n_high) for high and the princals optimum is the PCA of cor(quantified rating, x1, x2).",
165            "Sign convention replicated from the product: the loadings of each dimension sum to a positive value."))

Whole script generators/r/multivariate_catpca.R


Multivariate STAT-MV-CORR-001

Case STAT-MV-CORR-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_core.R · sha256 90e434c114270562
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/0/valuetolerance10.99999999999999991.11E-1615passed
/cells/1/valuetolerance0.187210767772165330.187210767772165258.14E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

91  emit("STAT-MV-CORR-001", "tabelz.analysis.correlation:compute_correlation",
92       "correlationMenuItem", list(columns=list("x","y","z"), metric="pearson", max_scatter_columns=0L),
93       list("/cells/0/value"=tol(1, "diagonal correlation"),
94            "/cells/1/value"=tol(stats::cor(d$x,d$y), "Pearson x-y correlation")),
95       c("Pearson product-moment correlation, complete pairwise finite observations."))

Whole script generators/r/multivariate_core.R


Multivariate option STAT-MV-CORR-COVARIANCE-001

Case STAT-MV-CORR-COVARIANCE-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/1/valuetolerance1.43522727272727281.43522727272727244.48E-1615passed

Convention

Reference code

Whole script generators/r/multivariate_options.R


Hard-numeric STAT-MV-CORR-HARD-001

Case STAT-MV-CORR-HARD-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::prcomp, factanal, hclust, kmeans, manova and chisq.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_hard.R · sha256 06eca51b7d1d3ab6
Dataset
reference_data/datasets/MV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/1/valuetolerance1.43522727908284381.43522728166666672.58E-98passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

95  emit("STAT-MV-CORR-HARD-001","tabelz.analysis.correlation:compute_correlation",
96       "correlationMenuItem","MV_HARD",
97       list(columns=xyz,metric="covariance",max_scatter_columns=0L),
98       list("/cells/1/value"=tol(stats::cov(d$x,d$y),"sample covariance after common offsets")),
99       c("R sample covariance uses n-1 and centered products at 1e8 offset."))

Whole script generators/r/multivariate_hard.R


Invalid multivariate request STAT-MV-CORR-INVALID-001

Case STAT-MV-CORR-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-CORR-INVALID-002

Case STAT-MV-CORR-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Minimum-sample STAT-MV-CORR-MIN-001

Case STAT-MV-CORR-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, factanal, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_minimum.R · sha256 bcd4094568b73b37
Dataset
reference_data/datasets/MV_MIN_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/1/valuetolerance10.99999999999999991.11E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

50  emit("STAT-MV-CORR-MIN-001","tabelz.analysis.correlation:compute_correlation",
51       "correlationMenuItem","MV_MIN_TWO",
52       list(columns=list("x","y"),metric="pearson",max_scatter_columns=0L),
53       list("/cells/1/value"=tol(stats::cor(d$x,d$y),"two-row Pearson r is exactly signed one")),
54       c("Two distinct pairs are the smallest nonconstant sample for Pearson r."))

Whole script generators/r/multivariate_minimum.R


Missing and excluded multivariate population STAT-MV-CORR-POP-001

Case STAT-MV-CORR-POP-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic identities
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_population.R · sha256 8a6d148e9316e21f
Dataset
reference_data/datasets/MV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/1/valuetolerance0.137468375295608010.137468375295607965.15E-1715passed
/cells/1/pair_countexact10100.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 96  pair <- eligible[stats::complete.cases(eligible[c("x","y")]),,drop=FALSE]
 97  emit("STAT-MV-CORR-POP-001","tabelz.analysis.correlation:compute_correlation",
 98       "correlationMenuItem",list(columns=list("x","y"),max_scatter_columns=0L),
 99       list("/cells/1/value"=tol(stats::cor(pair$x,pair$y),"R pairwise-complete Pearson x-y"),
100            "/cells/1/pair_count"=exact(nrow(pair),"Pairwise finite rows after exclusion")),
101       c("Correlation uses pairwise finite rows rather than all-variable complete cases."))

Whole script generators/r/multivariate_population.R


Multivariate option STAT-MV-CORR-SPEARMAN-001

Case STAT-MV-CORR-SPEARMAN-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/1/valuetolerance0.433566433566433650.433566433566433651.06E-1817passed

Convention

Reference code

Whole script generators/r/multivariate_options.R


Special boundary STAT-MV-CORR-TIES-001

Case STAT-MV-CORR-TIES-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_boundaries.R · sha256 420706593513ebf0
Dataset
reference_data/datasets/MV_TIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cells/1/valuetolerance0.77777777777777790.77777777777777791.14E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

53  emit("STAT-MV-CORR-TIES-001","tabelz.analysis.correlation:compute_correlation",
54       "correlationMenuItem","MV_TIES",
55       list(columns=list("x","y"),metric="spearman",max_scatter_columns=0L),
56       list("/cells/1/value"=tol(stats::cor(d$x,d$y,method="spearman"),"midrank Spearman correlation with ties")),
57       c("Tied ranks use R average/midrank convention."))

Whole script generators/r/multivariate_boundaries.R


One-factor maximum-likelihood fit

Case STAT-MV-FA-001 · status: passed

Reference

Evidence type
R
Source
R stats::factanal
Citation
R stats::factanal maximum-likelihood factor analysis documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_factor.R · sha256 c9e6d06b9a729e97
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlation_eigenvalues/0tolerance1.99745905548196131.99745905548196162.65E-1615passed
/statusexactheywoodheywoodpassed
/uniquenesses/0tolerance00.00.00E+2passed
/uniquenesses/1tolerance0.964952128430156430.96495213868770411.03E-87passed
/uniquenesses/2tolerance0.0422374449555635860.0422374452811036253.26E-108passed
/source_countexact12120.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

10  r <- stats::cor(d[c("x","y","z")])
11  stopifnot(r[1,2] * r[1,3] / r[2,3] > 1)
12  lambda <- c(1, r[1,2], r[1,3]); psi <- c(0, 1 - r[1,2]^2, 1 - r[1,3]^2)
13  Sigma <- outer(lambda, lambda) + diag(psi); Si <- solve(Sigma); G <- Si - Si %*% r %*% Si
14  stopifnot(max(abs(2 * G %*% lambda)) < 1e-10, max(abs(diag(G)[2:3])) < 1e-10, G[1,1] > 0)
15  eig <- eigen(r, symmetric=TRUE)$values
16  id <- "STAT-MV-FA-001"

Whole script generators/r/multivariate_factor.R


Stable ML factor fit none

Case STAT-MV-FA-002 · status: passed

Reference

Evidence type
R
Source
R stats::factanal
Citation
R stats::factanal maximum-likelihood factor analysis documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_factor_stable.R · sha256 ffea0161e03a21ed
Dataset
reference_data/datasets/MV_FACTOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/uniquenesses/0tolerance0.0225723106810134210.0225723074977176343.18E-96passed
/uniquenesses/2tolerance0.990883786916117560.99088378727358163.57E-109passed
/correlation_eigenvalues/0tolerance3.00768529306745873.0076852930674645.32E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

41  set.seed(1959L)
42  fit1 <- stats::factanal(d,factors=1L,rotation="none",control=list(nstart=20L,opt=list(factr=1,pgtol=0,maxit=10000)))
43  set.seed(1959L)
44  fit2 <- stats::factanal(d,factors=2L,rotation="varimax",control=list(nstart=20L,opt=list(factr=1,pgtol=0,maxit=10000)))
45  emit("STAT-MV-FA-002",1L,"none",fit1)

Whole script generators/r/multivariate_factor_stable.R


Seeded ML factor fit none

Case STAT-MV-FA-003 · status: passed

Reference

Evidence type
R
Source
Seeded R stats::factanal
Citation
R stats::factanal maximum-likelihood factor analysis documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_factor_seeded.R · sha256 d0fd8847879330e0
Dataset
reference_data/datasets/MV_FACTOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/uniquenesses/0tolerance0.0225723106810134210.0225723074977176343.18E-96passed
/uniquenesses/2tolerance0.990883786916117560.99088378727358163.57E-109passed
/correlation_eigenvalues/0tolerance3.00768529306745873.0076852930674645.32E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

29  set.seed(1959)
30  fit1 <- stats::factanal(d,factors=1L,rotation="none",
31                         control=list(nstart=20L,opt=list(factr=1,pgtol=0,maxit=10000)))
32  set.seed(1959)
33  fit2 <- stats::factanal(d,factors=2L,rotation="varimax",
34                         control=list(nstart=20L,opt=list(factr=1,pgtol=0,maxit=10000)))
35  emit("STAT-MV-FA-003",1L,"none",fit1)

Whole script generators/r/multivariate_factor_seeded.R


Hard-numeric STAT-MV-FA-HARD-001

Case STAT-MV-FA-HARD-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::prcomp, factanal, hclust, kmeans, manova and chisq.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_hard.R · sha256 06eca51b7d1d3ab6
Dataset
reference_data/datasets/MV_FACTOR_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/uniquenesses/0tolerance0.0225723087179749390.0225723076494091581.07E-97passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

55  set.seed(1959)
56  fa <- stats::factanal(f,factors=1L,rotation="none",
57                       control=list(nstart=20L,opt=list(factr=1,pgtol=0,maxit=10000)))
58  emit("STAT-MV-FA-HARD-001","tabelz.executor.factor_analysis","factorAnalysisMenuItem",
59       "MV_FACTOR_HARD",list(columns=as.list(names(f)),n_factors=1L,rotation="none"),
60       list("/uniquenesses/0"=tol(fa$uniquenesses[1],"seeded ML uniqueness","1e-6")),
61       c("set.seed(1959), 20 starts, L-BFGS-B factr=1; R ML factor analysis on offset data."))

Whole script generators/r/multivariate_hard.R


Invalid multivariate request STAT-MV-FA-INVALID-001

Case STAT-MV-FA-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-FA-INVALID-002

Case STAT-MV-FA-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Seeded ML factor-loading sign convention

Case STAT-MV-FA-LOADINGS-001 · status: passed

Reference

Evidence type
R
Source
Seeded R stats::factanal
Citation
R stats::factanal maximum-likelihood factor analysis documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_factor_loadings.R · sha256 4c3562f543b3c199
Dataset
reference_data/datasets/MV_FACTOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/loadings/0sign_invariant_sequence['0.98864942686416046'][0.9886494285114245]passed
/loadings/2sign_invariant_sequence['0.09547885970496163'][0.09547885879877761]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

1  source("common.R",local=TRUE)
2  d <- read_dataset("MV_FACTOR")
3  set.seed(1959)
4  fit <- stats::factanal(d,factors=1L,rotation="none",
5                        control=list(nstart=20L,opt=list(factr=1,pgtol=0,maxit=10000)))
6  sign_one <- function(x,note) list(comparison="sign_invariant_sequence",
7    values=list(dec(unname(x))),absolute_tolerance="0",relative_tolerance="1e-6",
8    rationale=note)
9  id <- "STAT-MV-FA-LOADINGS-001"

Whole script generators/r/multivariate_factor_loadings.R


Minimum-sample STAT-MV-FA-MIN-001

Case STAT-MV-FA-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, factanal, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_minimum.R · sha256 bcd4094568b73b37
Dataset
reference_data/datasets/MV_MIN_FACTOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlation_eigenvalues/0tolerance2.51709776129547612.51709776129547574.11E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

55  f <- factor[c(1L,6L,12L,18L),c("a","b","c")]
56  export_dataset(f,"MV_MIN_FACTOR","Four complete observations on three variables",
57                 "R seeded ML factor-analysis minimum-row boundary",
58                 "Constructed validation data; no third-party license.",
59                 setNames(rep("Float64",ncol(f)),names(f)))
60  f <- read_dataset("MV_MIN_FACTOR")
61  eig <- eigen(stats::cor(f),symmetric=TRUE)$values[1]
62  emit("STAT-MV-FA-MIN-001","tabelz.executor.factor_analysis","factorAnalysisMenuItem",
63       "MV_MIN_FACTOR",list(columns=as.list(names(f)),n_factors=1L,rotation="none"),
64       list("/correlation_eigenvalues/0"=tol(eig,"sample-correlation eigenvalue at n=p+1")),
65       c("Four complete rows exceed the three selected columns by one; only the closed-form correlation eigenvalue is compared because a just-identified ML uniqueness fit is nonunique."))

Whole script generators/r/multivariate_minimum.R


Factor preprocessing after missing and exclusion

Case STAT-MV-FA-POP-001 · status: passed

Reference

Evidence type
R
Source
R stats::cor/eigen
Citation
R stats::cor and eigen documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_factor_population.R · sha256 12f01f8bd6ff3de3
Dataset
reference_data/datasets/MV_FACTOR_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact48480.00E+2passed
/excluded_countexact110.00E+2passed
/missing_countexact110.00E+2passed
/correlation_eigenvalues/0tolerance2.89766226316614442.89766226316614975.32E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 1  source("common.R",local=TRUE)
 2  raw <- read_dataset("MV_FACTOR")
 3  raw$a[4] <- NA_real_
 4  export_dataset(raw,"MV_FACTOR_POP","Constructed factor matrix with one missing measurement",
 5                 "R stats::cor/eigen for complete-case factor preprocessing",
 6                 "Constructed validation data; no third-party license.",
 7                 c(a="Float64",b="Float64",c="Float64",d="Float64",e="Float64"))
 8  d <- read_dataset("MV_FACTOR_POP")
 9  selected <- d[-11L,,drop=FALSE]
10  selected <- selected[stats::complete.cases(selected),,drop=FALSE]
11  eig <- eigen(stats::cor(selected),symmetric=TRUE)$values[1]
12  id <- "STAT-MV-FA-POP-001"

Whole script generators/r/multivariate_factor_population.R


Retained scoring STAT-MV-FA-SCORE-001

Case STAT-MV-FA-SCORE-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic
Citation
R stats::prcomp, kmeans, hclust and matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring.R · sha256 6aa292c3e1b1894b
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/factor1tolerance-1.0056800962632124-1.00568009626321186.46E-1615passed
/1/factor1tolerance-1.0385186375316962-1.03851863753169557.12E-1615passed
/2/factor1absentNonepassed
/3/factor1tolerance-0.9344876886373692-0.93448768863736865.59E-1615passed
/4/factor1tolerance0.309643655968443010.30964365596844181.22E-1514passed
/5/factor1tolerance0.248515946880273840.248515946880272811.03E-1514passed
/6/factor1tolerance0.251643239278769970.25164323927876881.17E-1514passed
/7/factor1tolerance0.257756258926815860.25775625892681481.05E-1514passed
/8/factor1tolerance0.722044177518351260.72204417751835039.99E-1614passed
/9/factor1tolerance0.769809125796301450.76980912579630059.97E-1614passed
/10/factor1tolerance0.695233270960889490.69523327096088868.85E-1614passed
/11/factor1tolerance0.78746510524424540.78746510524424431.11E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

50  id <- "STAT-MV-FA-SCORE-001"

Whole script generators/r/multivariate_scoring.R


Retained correction STAT-MV-FA-SCORE-HARD-001

Case STAT-MV-FA-SCORE-HARD-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and analytic row-identity join
Citation
R stats::prcomp, scale and base matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_corrections.R · sha256 d22bd7a20d0e25a3
Dataset
reference_data/datasets/MV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/factor1tolerance-1.0056801436500564-1.00568014365005642.40E-1716passed
/1/factor1tolerance-1.0385186849392949-1.0385186849392951.07E-1716passed
/2/factor1tolerance-1.0634244049368691-1.0634244049368692.79E-1716passed
/3/factor1tolerance-0.93448773542295449-0.93448773542295457.94E-1918passed
/4/factor1tolerance0.309643609548824080.30964360954882413.61E-1816passed
/5/factor1tolerance0.248515900134071490.24851590013407152.06E-1817passed
/6/factor1tolerance0.251643192589435130.251643192589435075.82E-1715passed
/7/factor1tolerance0.257756211864340280.25775621186434031.13E-1817passed
/8/factor1tolerance0.722044131510289460.72204413151028954.72E-1817passed
/9/factor1tolerance0.769809078036807980.76980907803680811.12E-1615passed
/10/factor1tolerance0.695233224558436610.69523322455843663.27E-1817passed
/11/factor1tolerance0.787465058183188350.78746505818318831.09E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

104  pc <- stats::prcomp(hard[c("x","y")],center=TRUE,scale.=TRUE)
105  pc_center <- as.vector(as_serialized(pc$center)); pc_scale <- as.vector(as_serialized(pc$scale))
106  pc_loading <- as.vector(as_serialized(pc$rotation[,1]))
107  scores <- as.vector(scale(as.matrix(hard[c("x","y")]),center=pc_center,scale=pc_scale)%*%pc_loading)
108  model <- list(principal_components=list(columns=list("x","y"),means=nums(pc_center),
109    scales=nums(pc_scale),loadings=rows(matrix(pc_loading,ncol=1)),row_ids=ids(id,nrow(hard))))
110  emit(id,"operation:tabelz.principal_component_scores","principalComponentsMenuItem",
111       "MV_HARD",list(names=list("pc1"),model=model),scores,"pc1",
112       c("Scoring subtracts 1e8-scale means before applying retained PCA loading."),kind="hard_numeric")
113  id <- "STAT-MV-FA-SCORE-HARD-001"

Whole script generators/r/multivariate_scoring_corrections.R


Retained scoring STAT-MV-FA-SCORE-HARD-VALID-001

Case STAT-MV-FA-SCORE-HARD-VALID-001 · status: passed

Reference

Evidence type
R
Source
R stats and centered matrix arithmetic
Citation
R stats::prcomp and scale documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_final.R · sha256 90940efd66e73ecd
Dataset
reference_data/datasets/MV_SCORE_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/factor1tolerance-1.005680096261294-1.00568009630630464.50E-1110passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

62  pc <- stats::prcomp(hard[c("x","y")],center=TRUE,scale.=TRUE)
63  model <- list(principal_components=list(columns=list("x","y"),means=nums(pc$center),
64    scales=nums(pc$scale),loadings=rows(pc$rotation[,1,drop=FALSE]),row_ids=ids(id,nrow(hard))))
65  out <- list("/0/pc1"=tol(pc$x[1,1],"Centered PCA score at 1e5 offset"))
66  emit(id,"operation:tabelz.principal_component_scores","principalComponentsMenuItem",
67       "MV_SCORE_HARD",list(names=list("pc1"),model=model),out,"hard_numeric",
68       c("Measurements have a 1e5 offset with O(1) spread; R centers before projecting."))
69  id <- "STAT-MV-FA-SCORE-HARD-VALID-001"

Whole script generators/r/multivariate_scoring_final.R


Invalid multivariate request STAT-MV-FA-SCORE-INVALID-001

Case STAT-MV-FA-SCORE-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-FA-SCORE-INVALID-002

Case STAT-MV-FA-SCORE-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Retained scoring STAT-MV-FA-SCORE-MIN-VALID-001

Case STAT-MV-FA-SCORE-MIN-VALID-001 · status: passed

Reference

Evidence type
R
Source
R stats and centered matrix arithmetic
Citation
R stats::prcomp and scale documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_final.R · sha256 90940efd66e73ecd
Dataset
reference_data/datasets/MV_SCORE_FOUR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/factor1tolerance-1.0896413684402626-1.08964136844026242.19E-1615passed
/3/factor1tolerance0.501515274163573550.50151527416357314.40E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

39  weights <- matrix(c(0.2,0.3,0.5),ncol=1)
40  id <- "STAT-MV-FA-SCORE-MIN-VALID-001"

Whole script generators/r/multivariate_scoring_final.R


Retained scoring STAT-MV-FA-SCORE-OMIT-VALID-001

Case STAT-MV-FA-SCORE-OMIT-VALID-001 · status: passed

Reference

Evidence type
R
Source
R stats and centered matrix arithmetic
Citation
R stats::prcomp and scale documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_final.R · sha256 90940efd66e73ecd
Dataset
reference_data/datasets/MV_SCORE_FIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/factor1tolerance-1.0896413684402626-1.08964136844026242.19E-1615passed
/3/factor1tolerance0.501515274163573550.50151527416357314.40E-1615passed
/4/factor1absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

41  x <- as.matrix(four[c("x","y","z")]); mu <- colMeans(x); sd <- apply(x,2,stats::sd)
42  scores <- as.vector(scale(x,center=mu,scale=sd)%*%weights)
43  model <- list(factor_analysis=list(columns=list("x","y","z"),means=nums(mu),
44    scales=nums(sd),weights=rows(weights),row_ids=ids(id,4L)))
45  out <- list("/0/factor1"=tol(scores[1],"Four-row retained FA score"),
46              "/3/factor1"=tol(scores[4],"Fourth retained FA score"))
47  emit(id,"operation:tabelz.factor_scores","factorAnalysisMenuItem","MV_SCORE_FOUR",
48       list(names=list("factor1"),model=model),out,"boundary",
49       c("Four retained rows are the smallest population strictly exceeding three scoring columns."))
50  id <- "STAT-MV-FA-SCORE-OMIT-VALID-001"

Whole script generators/r/multivariate_scoring_final.R


Special boundary STAT-MV-FA-SIX-001

Case STAT-MV-FA-SIX-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_boundaries.R · sha256 420706593513ebf0
Dataset
reference_data/datasets/MV_FA_SIX.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlation_eigenvalues/0tolerance2.91875774539336642.918757745393373.56E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

58  f <- read_dataset("MV_FACTOR")[seq_len(6L),]
59  export_dataset(f,"MV_FA_SIX","Six complete observations on five factor variables",
60                 "R sample-correlation eigenvalue at n=p+1",
61                 "Constructed validation data; no third-party license.",
62                 setNames(rep("Float64",ncol(f)),names(f)))
63  f <- read_dataset("MV_FA_SIX")
64  eig <- eigen(stats::cor(f),symmetric=TRUE)$values[1]
65  emit("STAT-MV-FA-SIX-001","tabelz.executor.factor_analysis","factorAnalysisMenuItem",
66       "MV_FA_SIX",list(columns=as.list(names(f)),n_factors=1L,rotation="none"),
67       list("/correlation_eigenvalues/0"=tol(eig,"leading correlation eigenvalue at n=p+1")),
68       c("Five-variable factor analysis on six complete rows is the n=p+1 boundary; ML uniqueness is not compared on this small sample."))

Whole script generators/r/multivariate_boundaries.R


Stable ML factor fit varimax

Case STAT-MV-FA-VARIMAX-001 · status: passed

Reference

Evidence type
R
Source
R stats::factanal
Citation
R stats::factanal maximum-likelihood factor analysis documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_factor_stable.R · sha256 ffea0161e03a21ed
Dataset
reference_data/datasets/MV_FACTOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/uniquenesses/0tolerance0.012980870964739750.0129808704788346934.86E-107passed
/uniquenesses/2tolerance0.0339630470695116840.033963046694790853.75E-107passed
/correlation_eigenvalues/0tolerance3.00768529306745873.0076852930674645.32E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

46  emit("STAT-MV-FA-VARIMAX-001",2L,"varimax",fit2)

Whole script generators/r/multivariate_factor_stable.R


Seeded ML factor fit varimax

Case STAT-MV-FA-VARIMAX-002 · status: passed

Reference

Evidence type
R
Source
Seeded R stats::factanal
Citation
R stats::factanal maximum-likelihood factor analysis documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_factor_seeded.R · sha256 d0fd8847879330e0
Dataset
reference_data/datasets/MV_FACTOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/uniquenesses/0tolerance0.012980870964739750.0129808704788346934.86E-107passed
/uniquenesses/2tolerance0.0339630470695116840.033963046694790853.75E-107passed
/correlation_eigenvalues/0tolerance3.00768529306745873.0076852930674645.32E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

36  emit("STAT-MV-FA-VARIMAX-002",2L,"varimax",fit2)

Whole script generators/r/multivariate_factor_seeded.R


Multivariate STAT-MV-HCLUST-001

Case STAT-MV-HCLUST-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_core.R · sha256 90e434c114270562
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/merge_heightssequence['0.065820535503267441', '0.083096677567938343', '0.10141230[0.06582053550326754, 0.08309667756793819, 0.101412304199116passed
/cluster_labelspartition[1, 1, 3, 3, 1, 1, 2, 2, 2, 2, 3, 3][3, 3, 1, 1, 3, 3, 2, 2, 2, 2, 1, 1]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

61  id_order <- function(id, n) order(paste0(id, ":", seq_len(n) - 1L), method = "radix")
62  emit("STAT-MV-HCLUST-001", "tabelz.executor.hierarchical_clustering",
63       "hierarchicalClusteringMenuItem",
64       list(columns=list("x","y","z"), linkage="ward", cut_k=3L),
65       list("/merge_heights"=seqtol(hc$height, "ward.D2 linkage heights"),
66            "/cluster_labels"=list(comparison="partition", values=as.list(as.integer(stats::cutree(hc,3L))[id_order("STAT-MV-HCLUST-001", nrow(d))]),
67                                    rationale="R Ward.D2 three-way partition")),
68       c("Rows standardized with sample SD; Ward.D2 uses squared Euclidean update."))

Whole script generators/r/multivariate_core.R


Ordered multivariate STAT-MV-HCLUST-002

Case STAT-MV-HCLUST-002 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::hclust/cutree/kmeans documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_ordered.R · sha256 583c4e9cd0e92e8b
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/cluster_labelspartition[1, 1, 2, 2, 1, 1, 3, 3, 3, 3, 2, 2][3, 3, 1, 1, 3, 3, 2, 2, 2, 2, 1, 1]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

29  scaled <- scale(d[order_ids,c("x","y","z")])
30  hc <- stats::hclust(stats::dist(scaled), method="ward.D2")
31  emit("STAT-MV-HCLUST-002", "tabelz.executor.hierarchical_clustering",
32       "hierarchicalClusteringMenuItem",
33       list(columns=list("x","y","z"), linkage="ward", cut_k=3L),
34       list("/cluster_labels"=part(stats::cutree(hc,3L), "R Ward.D2 partition in retained row-ID order")),
35       c("Sample-SD standardization and Euclidean Ward.D2 linkage."))

Whole script generators/r/multivariate_ordered.R


Multivariate option STAT-MV-HCLUST-AVERAGE-001

Case STAT-MV-HCLUST-AVERAGE-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/merge_heights/0tolerance0.0658205355032674410.065820535503267549.72E-1714passed
/cluster_labelspartition[1, 1, 2, 2, 1, 1, 3, 3, 3, 3, 2, 2][2, 2, 1, 1, 2, 2, 3, 3, 3, 3, 1, 1]passed

Convention

Reference code

Whole script generators/r/multivariate_options.R


Multivariate option STAT-MV-HCLUST-COMPLETE-001

Case STAT-MV-HCLUST-COMPLETE-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/merge_heights/0tolerance0.0658205355032674410.065820535503267549.72E-1714passed
/cluster_labelspartition[1, 1, 2, 2, 1, 1, 3, 3, 3, 3, 2, 2][1, 1, 3, 3, 1, 1, 2, 2, 2, 2, 3, 3]passed

Convention

Reference code

Whole script generators/r/multivariate_options.R


Hard-numeric STAT-MV-HCLUST-HARD-001

Case STAT-MV-HCLUST-HARD-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::prcomp, factanal, hclust, kmeans, manova and chisq.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_hard.R · sha256 06eca51b7d1d3ab6
Dataset
reference_data/datasets/MV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/merge_heights/0tolerance0.0658205368802939060.06582053688029391.46E-1917passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

62  ord <- order(as.character(seq_len(nrow(d))-1L))
63  scaled <- scale(d[ord,c("x","y","z")])
64  hc <- stats::hclust(stats::dist(scaled),method="ward.D2")
65  emit("STAT-MV-HCLUST-HARD-001","tabelz.executor.hierarchical_clustering",
66       "hierarchicalClusteringMenuItem","MV_HARD",
67       list(columns=xyz,linkage="ward",cut_k=3L),
68       list("/merge_heights/0"=tol(hc$height[1],"standardized first Ward.D2 merge")),
69       c("Rows are in stable row-ID order and standardized with sample SD."))

Whole script generators/r/multivariate_hard.R


Invalid multivariate request STAT-MV-HCLUST-INVALID-001

Case STAT-MV-HCLUST-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-HCLUST-INVALID-002

Case STAT-MV-HCLUST-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Retained scoring STAT-MV-HCLUST-LABEL-001

Case STAT-MV-HCLUST-LABEL-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic
Citation
R stats::prcomp, kmeans, hclust and matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring.R · sha256 6aa292c3e1b1894b
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/hclusterexact110.00E+2passed
/1/hclusterexact110.00E+2passed
/2/hclusterexact110.00E+2passed
/3/hclusterexact110.00E+2passed
/4/hclusterexact330.00E+2passed
/5/hclusterexact330.00E+2passed
/6/hclusterexact330.00E+2passed
/7/hclusterexact330.00E+2passed
/8/hclusterexact220.00E+2passed
/9/hclusterexact220.00E+2passed
/10/hclusterexact220.00E+2passed
/11/hclusterexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

61  sorted <- order(as.character(seq_len(nrow(d))-1L))
62  scaled <- scale(d[sorted,c("x","y","z")])
63  id <- "STAT-MV-HCLUST-LABEL-001"

Whole script generators/r/multivariate_scoring.R


Invalid multivariate request STAT-MV-HCLUST-LABEL-INVALID-001

Case STAT-MV-HCLUST-LABEL-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-HCLUST-LABEL-INVALID-002

Case STAT-MV-HCLUST-LABEL-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Retained boundary STAT-MV-HCLUST-LABEL-MIN-001

Case STAT-MV-HCLUST-LABEL-MIN-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and retained identity join
Citation
R stats::prcomp and base matrix arithmetic documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_boundaries.R · sha256 4cdf60725804542b
Dataset
reference_data/datasets/MV_MIN_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/hclusterexact110.00E+2passed
/1/hclusterexact220.00E+2passed

Convention

Reference code

Whole script generators/r/multivariate_scoring_boundaries.R


Retained correction STAT-MV-HCLUST-LABEL-OMIT-CORRECTED-001

Case STAT-MV-HCLUST-LABEL-OMIT-CORRECTED-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and analytic row-identity join
Citation
R stats::prcomp, scale and base matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_corrections.R · sha256 d22bd7a20d0e25a3
Dataset
reference_data/datasets/MV_SCORE_FOUR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/hclusterexact110.00E+2passed
/1/hclusterexact220.00E+2passed
/2/hclusterexact220.00E+2passed
/3/hclusterabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

69  x <- as.matrix(ret[c("x","y","z")]); mu <- colMeans(x); sd <- apply(x,2,stats::sd)
70  scores <- as.vector(scale(as.matrix(d[c("x","y","z")]),center=mu,scale=sd)%*%weights)
71  model <- list(factor_analysis=list(columns=list("x","y","z"),means=nums(mu),
72    scales=nums(sd),weights=rows(weights),row_ids=ids(id,3L)))
73  emit(id,"operation:tabelz.factor_scores","factorAnalysisMenuItem",
74       "MV_SCORE_FOUR",list(names=list("factor1"),model=model),scores,"factor1",
75       c("Three retained FA score rows plus one nonmember."),omitted=4L)
76  id <- "STAT-MV-HCLUST-LABEL-OMIT-CORRECTED-001"

Whole script generators/r/multivariate_scoring_corrections.R


Minimum-sample STAT-MV-HCLUST-MIN-001

Case STAT-MV-HCLUST-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, factanal, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_minimum.R · sha256 bcd4094568b73b37
Dataset
reference_data/datasets/MV_MIN_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/merge_heights/0tolerance1.99999999999999981.99999999999999982.20E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

34  hc <- stats::hclust(stats::dist(scale(d[c("x","y")])),method="ward.D2")
35  emit("STAT-MV-HCLUST-MIN-001","tabelz.executor.hierarchical_clustering",
36       "hierarchicalClusteringMenuItem","MV_MIN_TWO",
37       list(columns=list("x","y"),linkage="ward",cut_k=2L),
38       list("/merge_heights/0"=tol(hc$height[1],"only Ward.D2 merge height")),
39       c("Two distinct rows give one unambiguous merge."))

Whole script generators/r/multivariate_minimum.R


Missing and excluded multivariate population STAT-MV-HCLUST-POP-001

Case STAT-MV-HCLUST-POP-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic identities
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_population.R · sha256 8a6d148e9316e21f
Dataset
reference_data/datasets/MV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/excluded_countexact110.00E+2passed
/missing_countexact110.00E+2passed
/cluster_labelspartition[1, 1, 2, 2, 1, 3, 3, 3, 3, 2][2, 2, 1, 1, 2, 3, 3, 3, 3, 1]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

44  ord <- which(seq_len(nrow(d)) != 10L & stats::complete.cases(d[c("x","y","z")]))
45  ord <- ord[order(as.character(ord-1L))]
46  scaled <- scale(d[ord,c("x","y","z")])
47  hc <- stats::hclust(stats::dist(scaled),method="ward.D2")
48  emit("STAT-MV-HCLUST-POP-001","tabelz.executor.hierarchical_clustering",
49       "hierarchicalClusteringMenuItem",list(columns=list("x","y","z"),cut_k=3L),
50       list("/excluded_count"=exact(1L,"One stable row ID excluded"),
51            "/missing_count"=exact(1L,"One selected-column missing row"),
52            "/cluster_labels"=part(stats::cutree(hc,3L),"R Ward.D2 partition in sorted row-ID order")),
53       c("R uses sample-SD standardization and Ward.D2 on the same sorted complete rows."))

Whole script generators/r/multivariate_population.R


Configuration axis STAT-MV-HCLUST-RAW-001

Case STAT-MV-HCLUST-RAW-001 · status: passed

Reference

Evidence type
R
Source
R stats and matrix arithmetic
Citation
R stats::hclust and base rowSums documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_axes.R · sha256 8e53089af57e269e
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/merge_heights/0tolerance0.173205080756887370.173205080756887374.03E-1917passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

21  ord <- order(as.character(seq_len(nrow(d))-1L))
22  hc <- stats::hclust(stats::dist(d[ord,c("x","y","z")]),method="ward.D2")
23  emit("STAT-MV-HCLUST-RAW-001","tabelz.executor.hierarchical_clustering",
24       "hierarchicalClusteringMenuItem",
25       list(columns=list("x","y","z"),mode="rows",linkage="ward",standardize=FALSE,cut_k=3L),
26       list("/merge_heights/0"=tol(hc$height[1],"R Ward.D2 first merge on original measurement scale")),
27       c("Row-mode standardize=false uses Euclidean distances in original units."))

Whole script generators/r/multivariate_axes.R


Special boundary STAT-MV-HCLUST-TIES-001

Case STAT-MV-HCLUST-TIES-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_boundaries.R · sha256 420706593513ebf0
Dataset
reference_data/datasets/MV_TIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/merge_heights/0tolerance00.00.00E+215passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

34  ord <- order(as.character(seq_len(nrow(d))-1L))
35  hc <- stats::hclust(stats::dist(scale(d[ord,c("x","y")])),method="ward.D2")
36  emit("STAT-MV-HCLUST-TIES-001","tabelz.executor.hierarchical_clustering",
37       "hierarchicalClusteringMenuItem","MV_TIES",
38       list(columns=list("x","y"),linkage="ward",cut_k=2L),
39       list("/merge_heights/0"=tol(hc$height[1],"zero-height duplicate-row merge")),
40       c("Exact duplicate measurements force first Ward.D2 merge height zero."))

Whole script generators/r/multivariate_boundaries.R


Multivariate option STAT-MV-HCLUST-VARIABLES-001

Case STAT-MV-HCLUST-VARIABLES-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/merge_heights/0tolerance0.0213465602960175580.0213465602960175035.53E-1714passed
/cluster_labelspartition[1, 2, 1][2, 1, 2]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

41  var_dist <- as.dist(1-stats::cor(d[c("x","y","z")]))
42  var_hc <- stats::hclust(var_dist,method="average")
43  emit("STAT-MV-HCLUST-VARIABLES-001","tabelz.executor.hierarchical_clustering",
44       "hierarchicalClusteringMenuItem",
45       list(columns=cols,mode="variables",linkage="average",standardize=TRUE,cut_k=2L),
46       list("/merge_heights/0"=tol(var_hc$height[1],"R average linkage on 1-Pearson correlation"),
47            "/cluster_labels"=part(stats::cutree(var_hc,2L),"R variable partition")),
48       c("Variable mode distance is 1-r on complete cases; scaling is immaterial to correlation.",
49         "Variable mode requires standardize=true (strict request contract); until 2026-09-25 this case sent false."))

Whole script generators/r/multivariate_options.R


Variable-mode average linkage

Case STAT-MV-HCLUST-VARIABLES-002 · status: passed

Reference

Evidence type
R
Source
R stats::hclust/cutree
Citation
R stats::hclust and cutree documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_variables.R · sha256 ab1b4850d11e43e0
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/merge_heights/0tolerance0.0213465602960175580.0213465602960175035.53E-1714passed
/cluster_labelspartition[1, 2, 1][2, 1, 2]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

1  source("common.R",local=TRUE)
2  d <- read_dataset("MV_CORE")
3  distance <- as.dist(1-stats::cor(d[c("x","y","z")]))
4  fit <- stats::hclust(distance,method="average")
5  id <- "STAT-MV-HCLUST-VARIABLES-002"

Whole script generators/r/multivariate_variables.R


Kendall tau-b tie-corrected normal

Case STAT-MV-KENDALL-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::cor.test
Citation
R Core Team, stats package help: cor.test (method = 'kendall'); Kendall (1970), Rank Correlation Methods, for S and the tie-corrected variance.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_correlation_extensions.R · sha256 0d15a50f16440286
Dataset
reference_data/datasets/MV_CORREXT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pairs/0/nexact16160.00E+2passed
/pairs/0/tau_btolerance0.833333333333333260.83333333333333336.82E-1918passed
/pairs/0/sexact1001000.00E+2passed
/pairs/0/ztolerance4.50225168890748154.50225168890748158.29E-1817passed
/pairs/0/p_valuetail_probability6.7237289325878942e-066.723728932587893e-067.99E-2215passed
/pairs/1/nexact16160.00E+2passed
/pairs/1/tau_btolerance0.79567831159028590.79567831159028592.71E-1817passed
/pairs/1/sexact91910.00E+2passed
/pairs/1/ztolerance4.14924284854673434.1492428485467348.00E-1817passed
/pairs/1/p_valuetail_probability3.3357678071950611e-053.335767807195062e-057.03E-2115passed
/pairs/2/nexact16160.00E+2passed
/pairs/2/tau_btolerance0.883115488688119620.88311548868811951.10E-1615passed
/pairs/2/sexact1011010.00E+2passed
/pairs/2/ztolerance4.60520360113428764.6052036011342881.65E-1717passed
/pairs/2/p_valuetail_probability4.1206250634638877e-064.120625063463884e-063.38E-2115passed
/rows_analysedexact16160.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

102  outputs <- c(kendall_pair(d$x, d$y, "/pairs/0"), kendall_pair(d$x, d$z, "/pairs/1"),
103               kendall_pair(d$y, d$z, "/pairs/2"))
104  outputs[["/rows_analysed"]] <- exact(nrow(d), "all rows are analysed; deletion is per pair")
105  emit("STAT-MV-KENDALL-001", kendall_entry, kendall_menu, "Kendall tau-b tie-corrected normal",
106       "nominal", list(columns = list("x", "y", "z"), deletion = "pairwise", inference = "asymptotic"),
107       outputs, kendall_notes, kendall_cite, "R 4.6.1 stats::cor.test")

Whole script generators/r/multivariate_correlation_extensions.R


Kendall tau-b pairwise deletion

Case STAT-MV-KENDALL-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::cor.test
Citation
R Core Team, stats package help: cor.test (method = 'kendall'); Kendall (1970), Rank Correlation Methods, for S and the tie-corrected variance.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_correlation_extensions.R · sha256 0d15a50f16440286
Dataset
reference_data/datasets/MV_CORREXT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pairs/0/nexact16160.00E+2passed
/pairs/0/tau_btolerance0.833333333333333260.83333333333333336.82E-1918passed
/pairs/0/sexact1001000.00E+2passed
/pairs/0/ztolerance4.50225168890748154.50225168890748158.29E-1817passed
/pairs/0/p_valuetail_probability6.7237289325878942e-066.723728932587893e-067.99E-2215passed
/pairs/1/nexact15150.00E+2passed
/pairs/1/tau_btolerance0.695238095238095320.69523809523809531.63E-1817passed
/pairs/1/sexact73730.00E+2passed
/pairs/1/ztolerance3.61256311292937323.61256311292937274.89E-1615passed
/pairs/1/p_valuetail_probability0.000303185281695832820.000303185281695832651.66E-1915passed
/pairs/2/nexact15150.00E+2passed
/pairs/2/tau_btolerance0.828571428571428740.82857142857142861.15E-1615passed
/pairs/2/sexact87870.00E+2passed
/pairs/2/ztolerance4.30538343595692394.3053834359569244.57E-1716passed
/pairs/2/p_valuetail_probability1.6669674796231492e-051.666967479623147e-052.05E-2014passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

110  outputs <- c(kendall_pair(d$x, d$y, "/pairs/0"), kendall_pair(d$x, d$w, "/pairs/1"),
111               kendall_pair(d$y, d$w, "/pairs/2"))
112  emit("STAT-MV-KENDALL-002", kendall_entry, kendall_menu, "Kendall tau-b pairwise deletion",
113       "missing_excluded", list(columns = list("x", "y", "w"), deletion = "pairwise", inference = "asymptotic"),
114       outputs, c(kendall_notes, "Pairwise deletion: each pair drops only its own incomplete rows."),
115       kendall_cite, "R 4.6.1 stats::cor.test")

Whole script generators/r/multivariate_correlation_extensions.R


Kendall tau-b listwise deletion

Case STAT-MV-KENDALL-003 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::cor.test
Citation
R Core Team, stats package help: cor.test (method = 'kendall'); Kendall (1970), Rank Correlation Methods, for S and the tie-corrected variance.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_correlation_extensions.R · sha256 0d15a50f16440286
Dataset
reference_data/datasets/MV_CORREXT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pairs/0/nexact15150.00E+2passed
/pairs/0/tau_btolerance0.828571428571428740.82857142857142861.15E-1615passed
/pairs/0/sexact87870.00E+2passed
/pairs/0/ztolerance4.30538343595692394.3053834359569244.57E-1716passed
/pairs/0/p_valuetail_probability1.6669674796231492e-051.666967479623147e-052.05E-2014passed
/pairs/1/nexact15150.00E+2passed
/pairs/1/tau_btolerance0.695238095238095320.69523809523809531.63E-1817passed
/pairs/1/sexact73730.00E+2passed
/pairs/1/ztolerance3.61256311292937323.61256311292937274.89E-1615passed
/pairs/1/p_valuetail_probability0.000303185281695832820.000303185281695832651.66E-1915passed
/pairs/2/nexact15150.00E+2passed
/pairs/2/tau_btolerance0.828571428571428740.82857142857142861.15E-1615passed
/pairs/2/sexact87870.00E+2passed
/pairs/2/ztolerance4.30538343595692394.3053834359569244.57E-1716passed
/pairs/2/p_valuetail_probability1.6669674796231492e-051.666967479623147e-052.05E-2014passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

118  complete <- is.finite(d$x) & is.finite(d$y) & is.finite(d$w)
119  outputs <- c(kendall_pair(d$x[complete], d$y[complete], "/pairs/0"),
120               kendall_pair(d$x[complete], d$w[complete], "/pairs/1"),
121               kendall_pair(d$y[complete], d$w[complete], "/pairs/2"))
122  emit("STAT-MV-KENDALL-003", kendall_entry, kendall_menu, "Kendall tau-b listwise deletion",
123       "axis_coverage", list(columns = list("x", "y", "w"), deletion = "listwise", inference = "asymptotic"),
124       outputs, c(kendall_notes, "Listwise deletion: rows incomplete on any chosen column are dropped from every pair."),
125       kendall_cite, "R 4.6.1 stats::cor.test")

Whole script generators/r/multivariate_correlation_extensions.R


Kendall tau-b exact inference

Case STAT-MV-KENDALL-004 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::cor.test
Citation
R Core Team, stats package help: cor.test (method = 'kendall'); Kendall (1970), Rank Correlation Methods, for S and the tie-corrected variance.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_correlation_extensions.R · sha256 0d15a50f16440286
Dataset
reference_data/datasets/MV_CORREXT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/pairs/0/nexact16160.00E+2passed
/pairs/0/tau_btolerance0.833333333333333260.83333333333333336.82E-1918passed
/pairs/0/sexact1001000.00E+2passed
/pairs/0/p_valuetail_probability2.4419372501860437e-072.4419372499316283e-072.54E-179passed
/pairs/1/nexact16160.00E+2passed
/pairs/1/tau_btolerance0.79567831159028590.79567831159028592.71E-1817passed
/pairs/1/sexact91910.00E+2passed
/pairs/1/p_valueabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

128  outputs <- c(kendall_pair(d$x, d$y, "/pairs/0", exact_p = TRUE),
129               kendall_pair(d$x, d$z, "/pairs/1", exact_p = TRUE))
130  emit("STAT-MV-KENDALL-004", kendall_entry, kendall_menu, "Kendall tau-b exact inference",
131       "axis_coverage", list(columns = list("x", "y", "z"), deletion = "pairwise", inference = "exact"),
132       outputs, c(kendall_notes, "cor.test(exact = TRUE) enumerates the permutation distribution of S for n < 50 without ties; with ties the p-value is withheld."),
133       kendall_cite, "R 4.6.1 stats::cor.test")

Whole script generators/r/multivariate_correlation_extensions.R


Multivariate STAT-MV-KMEANS-001

Case STAT-MV-KMEANS-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_core.R · sha256 90e434c114270562
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/inertiatolerance0.116962942715426880.11696294271542691.86E-1715passed
/cluster_labelspartition[2, 2, 1, 1, 2, 2, 3, 3, 3, 3, 1, 1][1, 1, 3, 3, 1, 1, 2, 2, 2, 2, 3, 3]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

70  set.seed(11)
71  km <- stats::kmeans(scaled, centers=3L, nstart=50L, algorithm="Hartigan-Wong")
72  emit("STAT-MV-KMEANS-001", "tabelz.executor.clustering", "kmeansClusteringMenuItem",
73       list(columns=list("x","y","z"), k=3L, seed=11L, restarts=50L),
74       list("/inertia"=tol(km$tot.withinss, "unique three-cluster optimum"),
75            "/cluster_labels"=list(comparison="partition", values=as.list(as.integer(km$cluster)[id_order("STAT-MV-KMEANS-001", nrow(d))]),
76                                    rationale="Well-separated cluster partition independent of numeric labels")),
77       c("R Hartigan-Wong and Tabelz Lloyd agree on this unique separated optimum.") ,
78       kind="determinism", repeat_count=2L)

Whole script generators/r/multivariate_core.R


Ordered multivariate STAT-MV-KMEANS-002

Case STAT-MV-KMEANS-002 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::hclust/cutree/kmeans documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_ordered.R · sha256 583c4e9cd0e92e8b
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/inertiatolerance0.116962942715426880.11696294271542691.86E-1715passed
/cluster_labelspartition[2, 2, 3, 3, 2, 2, 1, 1, 1, 1, 3, 3][1, 1, 3, 3, 1, 1, 2, 2, 2, 2, 3, 3]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

37  set.seed(11)
38  km <- stats::kmeans(scaled, centers=3L, nstart=50L, algorithm="Hartigan-Wong")
39  emit("STAT-MV-KMEANS-002", "tabelz.executor.clustering", "kmeansClusteringMenuItem",
40       list(columns=list("x","y","z"), k=3L, seed=11L, restarts=50L),
41       list("/inertia"=tol(km$tot.withinss, "R unique optimum total within-cluster SS"),
42            "/cluster_labels"=part(km$cluster, "R unique optimum partition in retained row-ID order")),
43       c("Well-separated optimum is shared by R Hartigan-Wong and scikit-learn Lloyd."),
44       kind="determinism", repeat_count=2L)

Whole script generators/r/multivariate_ordered.R


Hard-numeric STAT-MV-KMEANS-HARD-001

Case STAT-MV-KMEANS-HARD-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::prcomp, factanal, hclust, kmeans, manova and chisq.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_hard.R · sha256 06eca51b7d1d3ab6
Dataset
reference_data/datasets/MV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/inertiatolerance0.116962943170034220.116962943170034211.13E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

70  set.seed(11)
71  km <- stats::kmeans(scaled,3L,nstart=50L,algorithm="Hartigan-Wong")
72  emit("STAT-MV-KMEANS-HARD-001","tabelz.executor.clustering","kmeansClusteringMenuItem",
73       "MV_HARD",list(columns=xyz,k=3L,seed=11L,restarts=50L),
74       list("/inertia"=tol(km$tot.withinss,"unique separated three-cluster within SS")),
75       c("set.seed(11), nstart=50; unique separated optimum, independent of label names."))

Whole script generators/r/multivariate_hard.R


Invalid multivariate request STAT-MV-KMEANS-INVALID-001

Case STAT-MV-KMEANS-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-KMEANS-INVALID-002

Case STAT-MV-KMEANS-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Retained scoring STAT-MV-KMEANS-LABEL-001

Case STAT-MV-KMEANS-LABEL-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic
Citation
R stats::prcomp, kmeans, hclust and matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring.R · sha256 6aa292c3e1b1894b
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/kclusterexact220.00E+2passed
/1/kclusterexact220.00E+2passed
/2/kclusterexact220.00E+2passed
/3/kclusterexact220.00E+2passed
/4/kclusterexact110.00E+2passed
/5/kclusterexact110.00E+2passed
/6/kclusterexact110.00E+2passed
/7/kclusterexact110.00E+2passed
/8/kclusterexact330.00E+2passed
/9/kclusterexact330.00E+2passed
/10/kclusterexact330.00E+2passed
/11/kclusterexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

74  id <- "STAT-MV-KMEANS-LABEL-001"

Whole script generators/r/multivariate_scoring.R


Invalid multivariate request STAT-MV-KMEANS-LABEL-INVALID-001

Case STAT-MV-KMEANS-LABEL-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-KMEANS-LABEL-INVALID-002

Case STAT-MV-KMEANS-LABEL-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Retained boundary STAT-MV-KMEANS-LABEL-MIN-001

Case STAT-MV-KMEANS-LABEL-MIN-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and retained identity join
Citation
R stats::prcomp and base matrix arithmetic documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_boundaries.R · sha256 4cdf60725804542b
Dataset
reference_data/datasets/MV_MIN_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/kclusterexact110.00E+2passed
/1/kclusterexact220.00E+2passed

Convention

Reference code

Whole script generators/r/multivariate_scoring_boundaries.R


Retained correction STAT-MV-KMEANS-LABEL-OMIT-CORRECTED-001

Case STAT-MV-KMEANS-LABEL-OMIT-CORRECTED-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and analytic row-identity join
Citation
R stats::prcomp, scale and base matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_corrections.R · sha256 d22bd7a20d0e25a3
Dataset
reference_data/datasets/MV_SCORE_FOUR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/kclusterexact110.00E+2passed
/1/kclusterexact220.00E+2passed
/2/kclusterexact220.00E+2passed
/3/kclusterabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

77  model <- list(hierarchical_clustering=list(mode="rows",linkage="ward",cut_k=2L,
78    row_ids=ids(id,3L),labels=as.list(c(1L,2L,2L))))
79  emit(id,"operation:tabelz.hierarchical_cluster_labels","hierarchicalClusteringMenuItem",
80       "MV_SCORE_FOUR",list(name="hcluster",model=model),c(1L,2L,2L,1L),"hcluster",
81       c("Three retained labels, one nonmember; the test concerns replay identity, not refitting linkage."),
82       omitted=4L,numeric=FALSE)
83  id <- "STAT-MV-KMEANS-LABEL-OMIT-CORRECTED-001"

Whole script generators/r/multivariate_scoring_corrections.R


Minimum-sample STAT-MV-KMEANS-MIN-001

Case STAT-MV-KMEANS-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, factanal, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_minimum.R · sha256 bcd4094568b73b37
Dataset
reference_data/datasets/MV_MIN_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/inertiatolerance00.00.00E+215passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

40  emit("STAT-MV-KMEANS-MIN-001","tabelz.executor.clustering","kmeansClusteringMenuItem",
41       "MV_MIN_TWO",list(columns=list("x","y"),k=2L,seed=11L,restarts=50L),
42       list("/inertia"=tol(0,"two singleton clusters have zero within SS")),
43       c("With k=n=2 every observation forms a singleton, independent of initialization."))

Whole script generators/r/multivariate_minimum.R


Missing and excluded multivariate population STAT-MV-KMEANS-POP-001

Case STAT-MV-KMEANS-POP-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic identities
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_population.R · sha256 8a6d148e9316e21f
Dataset
reference_data/datasets/MV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/excluded_countexact110.00E+2passed
/missing_countexact110.00E+2passed
/inertiatolerance0.087048011179124390.087048011179124362.73E-1715passed
/cluster_labelspartition[3, 3, 1, 1, 3, 2, 2, 2, 2, 1][1, 1, 3, 3, 1, 2, 2, 2, 2, 3]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

55  set.seed(37)
56  km <- stats::kmeans(scaled,3L,nstart=50L,algorithm="Hartigan-Wong")
57  emit("STAT-MV-KMEANS-POP-001","tabelz.executor.clustering","kmeansClusteringMenuItem",
58       list(columns=list("x","y","z"),k=3L,seed=37L,restarts=50L),
59       list("/excluded_count"=exact(1L,"One stable row ID excluded"),
60            "/missing_count"=exact(1L,"One selected-column missing row"),
61            "/inertia"=tol(km$tot.withinss,"R unique optimum complete-case within SS"),
62            "/cluster_labels"=part(km$cluster,"R complete-case partition in sorted row-ID order")),
63       c("Well-separated optimum shared by R Hartigan-Wong and scikit-learn Lloyd."))

Whole script generators/r/multivariate_population.R


Multivariate option STAT-MV-KMEANS-RAW-001

Case STAT-MV-KMEANS-RAW-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/inertiatolerance0.817499999999999010.81749999999999933.28E-1615passed
/cluster_labelspartition[1, 1, 2, 2, 1, 1, 3, 3, 3, 3, 2, 2][1, 1, 3, 3, 1, 1, 2, 2, 2, 2, 3, 3]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

51  set.seed(29)
52  km <- stats::kmeans(d[ord,c("x","y","z")],3L,nstart=50L,algorithm="Hartigan-Wong")
53  emit("STAT-MV-KMEANS-RAW-001","tabelz.executor.clustering","kmeansClusteringMenuItem",
54       list(columns=cols,k=3L,standardize=FALSE,seed=29L,restarts=50L),
55       list("/inertia"=tol(km$tot.withinss,"R unique raw-scale minimum within SS"),
56            "/cluster_labels"=part(km$cluster,"R raw-scale three-way partition")),
57       c("Well-separated optimum shared by R Hartigan-Wong and scikit-learn Lloyd; no standardization."),
58       kind="determinism",repeat_count=2L)

Whole script generators/r/multivariate_options.R


Special boundary STAT-MV-KMEANS-TIES-001

Case STAT-MV-KMEANS-TIES-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_boundaries.R · sha256 420706593513ebf0
Dataset
reference_data/datasets/MV_TIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/inertiatolerance2.01433433862514872.0143343386251484.87E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

41  set.seed(11)
42  km <- stats::kmeans(scale(d[ord,c("x","y")]),centers=2L,nstart=50L,algorithm="Hartigan-Wong")
43  emit("STAT-MV-KMEANS-TIES-001","tabelz.executor.clustering","kmeansClusteringMenuItem",
44       "MV_TIES",list(columns=list("x","y"),k=2L,seed=11L,restarts=50L),
45       list("/inertia"=tol(km$tot.withinss,"within SS with exact duplicate observations")),
46       c("set.seed(11), nstart=50; duplicate points remain together in the unique two-way optimum."))

Whole script generators/r/multivariate_boundaries.R


One-class latent class fit in closed form

Case STAT-MV-LATENT-ONE-001 · status: passed

Reference

Evidence type
R
Source
closed-form one-class multinomial maximum likelihood (base R)
Citation
Lazarsfeld and Henry (1968) Latent Structure Analysis; McCutcheon (1987) Latent Class Analysis, Sage.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · mclust 6.1.3, cluster 2.1.8.2, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_model_clustering.R · sha256 b03bbb2bfb4448e9
Dataset
reference_data/datasets/MV_LATENT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact23230.00E+2passed
/missing_countexact110.00E+2passed
/chosen_kexact110.00E+2passed
/fits/[k=1]/parametersexact550.00E+2passed
/fits/[k=1]/log_likelihoodtolerance-60.39735569922253-60.397355699222532.21E-1617passed
/fits/[k=1]/bictolerance136.47218247809082136.472182478090824.96E-1516passed
/profiles/0/sizeexact23230.00E+2passed
/profiles/0/probabilities/0/0/0exactglossglosspassed
/profiles/0/probabilities/0/0/1tolerance0.304347826086956540.304347826086956541.05E-1817passed
/profiles/0/probabilities/0/1/1tolerance0.347826086956521730.347826086956521735.24E-1917passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

182  finish_probs <- table(factor(complete$finish, levels = sort(unique(complete$finish), method = "radix"))) / nl
183  outputs <- list(
184    "/n" = exact(nl, "complete rows"),
185    "/missing_count" = exact(1L, "one row has a missing line"),
186    "/chosen_k" = exact(1L, "the only K in the range"),
187    "/fits/[k=1]/parameters" = exact(as.integer(m), "sum over variables of (categories - 1)"),
188    "/fits/[k=1]/log_likelihood" = tol(loglik, "sum_j sum_c n_jc log(n_jc / n): the multinomial log-likelihood at the observed proportions", rel = "1e-9"),
189    "/fits/[k=1]/bic" = tol(bic_of(loglik, m, nl), "-2 log L + m log n", rel = "1e-9"),
190    "/profiles/0/size" = exact(nl, "every row in the single class"),
191    "/profiles/0/probabilities/0/0/0" = exact(names(finish_probs)[1], "first finish category in code-point order"),
192    "/profiles/0/probabilities/0/0/1" = tol(finish_probs[[1]], "observed proportion of that category", rel = "1e-9"),
193    "/profiles/0/probabilities/0/1/1" = tol(finish_probs[[2]], "observed proportion of the second category", rel = "1e-9")
194  )
195  emit("STAT-MV-LATENT-ONE-001", "One-class latent class fit in closed form", "boundary",
196       "MV_LATENT",
197       list(variables = list("finish", "defect", "line"), kind = "latent_class", k_min = 1L,
198            k_max = 1L),
199       outputs,
200       list("With one class the conditional-independence likelihood factorizes into one multinomial per variable, maximized at the observed category proportions; complete rows only (missing_items complete).",
201            "Categories are listed in code-point order within each variable."),
202       "closed-form one-class multinomial maximum likelihood (base R)",
203       "Lazarsfeld and Henry (1968) Latent Structure Analysis; McCutcheon (1987) Latent Class Analysis, Sage.")

Whole script generators/r/multivariate_model_clustering.R


Multivariate STAT-MV-MANOVA-001

Case STAT-MV-MANOVA-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_core.R · sha256 90e434c114270562
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/joint_tests/0/valuetolerance1.99335936558775151.99335936558775086.58E-1615passed
/joint_tests/1/valuetolerance9.0403124637035046e-069.040312463706415e-062.91E-1812passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

107  mv <- stats::manova(cbind(x,y) ~ group, data=d)
108  pillai <- summary(mv, test="Pillai")$stats[1,]
109  wilks <- summary(mv, test="Wilks")$stats[1,]
110  emit("STAT-MV-MANOVA-001", "tabelz.executor.manova", "manovaMenuItem",
111       list(response_columns=list("x","y"), factor_columns=list("group")),
112       list("/joint_tests/0/value"=tol(unname(pillai["Pillai"]), "Pillai trace"),
113            "/joint_tests/1/value"=tol(unname(wilks["Wilks"]), "Wilks lambda")),
114       c("R manova one-way factor; Tabelz emits Pillai and Wilks only; no Hotelling-Lawley or Roy."))

Whole script generators/r/multivariate_core.R


Hard-numeric STAT-MV-MANOVA-HARD-001

Case STAT-MV-MANOVA-HARD-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::prcomp, factanal, hclust, kmeans, manova and chisq.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_hard.R · sha256 06eca51b7d1d3ab6
Dataset
reference_data/datasets/MV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/joint_tests/0/valuetolerance1.99335936561824621.99335936554493097.33E-1110passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

76  mv <- stats::manova(cbind(x,y)~group,data=d)
77  pillai <- summary(mv,test="Pillai")$stats[1,"Pillai"]
78  emit("STAT-MV-MANOVA-HARD-001","tabelz.executor.manova","manovaMenuItem",
79       "MV_HARD",list(response_columns=list("x","y"),factor_columns=list("group")),
80       list("/joint_tests/0/value"=tol(pillai,"Pillai trace after cancelling common offset")),
81       c("One-way MANOVA; Pillai trace computed from R hypothesis/error SSCP."))

Whole script generators/r/multivariate_hard.R


Multivariate option STAT-MV-MANOVA-II-001

Case STAT-MV-MANOVA-II-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/joint_tests/0/valuetolerance1.99335936558775151.99335936558775086.58E-1615passed
/joint_tests/1/valuetolerance9.0403124637035046e-069.040312463706415e-062.91E-1812passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

77  mv <- stats::manova(cbind(x,y)~group,data=d)
78  pillai <- summary(mv,test="Pillai")$stats[1,"Pillai"]
79  wilks <- summary(mv,test="Wilks")$stats[1,"Wilks"]
80  emit("STAT-MV-MANOVA-II-001","tabelz.executor.manova","manovaMenuItem",
81       list(response_columns=list("x","y"),factor_columns=list("group"),test_type="II"),
82       list("/joint_tests/0/value"=tol(pillai,"R one-way Pillai trace; Type II equals Type III"),
83            "/joint_tests/1/value"=tol(wilks,"R one-way Wilks lambda; Type II equals Type III")),
84       c("One-factor design has identical Type-II and Type-III hypotheses."))

Whole script generators/r/multivariate_options.R


Invalid multivariate request STAT-MV-MANOVA-INVALID-001

Case STAT-MV-MANOVA-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-MANOVA-INVALID-002

Case STAT-MV-MANOVA-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Minimum-sample STAT-MV-MANOVA-MIN-001

Case STAT-MV-MANOVA-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, factanal, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_minimum.R · sha256 bcd4094568b73b37
Dataset
reference_data/datasets/MV_MIN_MANOVA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/joint_tests/0/valuetolerance0.997823490225805740.99782349022580624.47E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

66  m <- core[c(1:3,5:7),]
67  export_dataset(m,"MV_MIN_MANOVA","Two groups with three observations each",
68                 "R manova minimum-estimable one-way boundary",
69                 "Constructed validation data; no third-party license.",types)
70  m <- read_dataset("MV_MIN_MANOVA")
71  mv <- stats::manova(cbind(x,y)~group,data=m)
72  pillai <- summary(mv,test="Pillai")$stats[1,"Pillai"]
73  emit("STAT-MV-MANOVA-MIN-001","tabelz.executor.manova","manovaMenuItem",
74       "MV_MIN_MANOVA",list(response_columns=list("x","y"),factor_columns=list("group")),
75       list("/joint_tests/0/value"=tol(pillai,"two-group Pillai trace")),
76       c("Each group has three observations for two responses; residual SSCP is estimable."))

Whole script generators/r/multivariate_minimum.R


Missing and excluded multivariate population STAT-MV-MANOVA-POP-001

Case STAT-MV-MANOVA-POP-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic identities
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_population.R · sha256 8a6d148e9316e21f
Dataset
reference_data/datasets/MV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/excluded_countexact110.00E+2passed
/missing_countexact110.00E+2passed
/joint_tests/0/valuetolerance1.99556585717565671.99556585717565679.90E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

86  md <- eligible[stats::complete.cases(eligible[c("x","y","group")]),,drop=FALSE]
87  mv <- stats::manova(cbind(x,y)~group,data=md)
88  pillai <- summary(mv,test="Pillai")$stats[1,"Pillai"]
89  emit("STAT-MV-MANOVA-POP-001","tabelz.executor.manova","manovaMenuItem",
90       list(response_columns=list("x","y"),factor_columns=list("group")),
91       list("/excluded_count"=exact(1L,"One stable row ID excluded"),
92            "/missing_count"=exact(1L,"One response/factor-missing row"),
93            "/joint_tests/0/value"=tol(pillai,"R complete-case one-way Pillai trace")),
94       c("One-way factor Type-III Pillai trace on complete response/factor rows."))

Whole script generators/r/multivariate_population.R


Special boundary STAT-MV-MANOVA-SMALL-001

Case STAT-MV-MANOVA-SMALL-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_boundaries.R · sha256 420706593513ebf0
Dataset
reference_data/datasets/MV_MANOVA_FOUR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/joint_tests/0/valuetolerance0.999452354873814230.99945235487404192.28E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

69  m <- base[c(1L,2L,5L,6L),]
70  export_dataset(m,"MV_MANOVA_FOUR","Two groups with two observations each",
71                 "R manova small residual-SSCP boundary",
72                 "Constructed validation data; no third-party license.",types)
73  m <- read_dataset("MV_MANOVA_FOUR")
74  mv <- stats::manova(cbind(x,y)~group,data=m)
75  pillai <- summary(mv,test="Pillai")$stats[1,"Pillai"]
76  emit("STAT-MV-MANOVA-SMALL-001","tabelz.executor.manova","manovaMenuItem",
77       "MV_MANOVA_FOUR",list(response_columns=list("x","y"),factor_columns=list("group")),
78       list("/joint_tests/0/value"=tol(pillai,"Pillai trace with two residual degrees of freedom")),
79       c("Two groups of two observations leave residual df=2 for two responses."))

Whole script generators/r/multivariate_boundaries.R


Indicator MCA of three variables, complete rows

Case STAT-MV-MCA-001 · status: passed

Reference

Evidence type
R
Source
R ca::mjca (lambda = 'indicator' and 'adjusted') with the textbook supplementary projection
Citation
Nenadic and Greenacre (2007), Correspondence analysis in R, with two- and three-dimensional graphics: the ca package, Journal of Statistical Software 20(3); Greenacre (2007) Correspondence Analysis in Practice, 2nd ed., chapters 18-19.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · ca 0.72, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_mca.R · sha256 6edebdab7f958ce0
Dataset
reference_data/datasets/MV_MCA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_analysedexact23230.00E+2passed
/variablesexact330.00E+2passed
/categoriesexact880.00E+2passed
/total_inertiatolerance1.66666666666666671.66666666666666674.07E-1716passed
/greenacre_denominatortolerance0.357042648709314290.357042648709315961.67E-1514passed
/dimensions/0/inertiatolerance0.691479734948668880.69147973494866923.31E-1615passed
/dimensions/0/raw_percenttolerance41.48878409692013541.488784096920151.41E-1415passed
/dimensions/0/cumulative_percenttolerance41.48878409692012841.488784096920152.11E-1415passed
/dimensions/0/adjustedtolerance0.28860490122752980.28860490122753034.97E-1614passed
/dimensions/0/benzecri_percenttolerance92.46443382881456392.464433828814532.81E-1415passed
/dimensions/0/greenacre_percenttolerance80.8320524931174380.83205249311722.27E-1314passed
/dimensions/1/inertiatolerance0.435575767039333870.435575767039334263.91E-1615passed
/dimensions/1/raw_percenttolerance26.13454602236003126.1345460223600552.44E-1415passed
/dimensions/1/cumulative_percenttolerance67.62333011928015567.623330119280215.66E-1415passed
/dimensions/1/adjustedtolerance0.0235204093127833160.0235204093127834961.80E-1614passed
/dimensions/1/benzecri_percenttolerance7.53556617118543237.5355661711854754.26E-1414passed
/dimensions/1/greenacre_percenttolerance6.58756296980432146.5875629698043422.05E-1414passed
/dimensions/2/inertiatolerance0.305199490161439770.30519949016144024.42E-1614passed
/dimensions/2/raw_percenttolerance18.31196940968638418.3119694096864122.83E-1414passed
/dimensions/2/cumulative_percenttolerance85.93529952896655385.935299528966627.09E-1415passed
/dimensions/2/adjustedtolerance00.00.00E+2passed
/dimensions/2/benzecri_percenttolerance00.00.00E+2passed
/dimensions/2/greenacre_percenttolerance00.00.00E+2passed
/dimensions/3/inertiatolerance0.167665573593928170.16766557359392831.38E-1615passed
/dimensions/3/raw_percenttolerance10.0599344156356910.0599344156356988.46E-1515passed
/dimensions/3/cumulative_percenttolerance95.99523394460223495.995233944602328.48E-1415passed
/dimensions/3/adjustedtolerance00.00.00E+2passed
/dimensions/3/benzecri_percenttolerance00.00.00E+2passed
/dimensions/3/greenacre_percenttolerance00.00.00E+2passed
/dimensions/4/inertiatolerance0.0667461009232949150.066746100923294838.29E-1714passed
/dimensions/4/raw_percenttolerance4.00476605539769454.004766055397694.46E-1514passed
/dimensions/4/cumulative_percenttolerance99.999999999999929100.07.10E-1415passed
/dimensions/4/adjustedtolerance00.00.00E+2passed
/dimensions/4/benzecri_percenttolerance00.00.00E+2passed
/dimensions/4/greenacre_percenttolerance00.00.00E+2passed
/categories_table/0/variableexactregionregionpassed
/categories_table/0/categoryexactnorthnorthpassed
/categories_table/0/countexact12120.00E+2passed
/categories_table/0/masstolerance0.173913043478260860.173913043478260864.74E-1816passed
/categories_table/0/contributions/0tolerance0.113602703457617940.11360270345761793.88E-1715passed
/categories_table/0/cosine_squared/0tolerance0.492747612915135690.49274761291513542.77E-1615passed
/categories_table/0/contributions/1tolerance0.00686424749336785480.00686424749336793558.07E-1713passed
/categories_table/0/cosine_squared/1tolerance0.0187548264389032460.0187548264389034682.22E-1613passed
/categories_table/1/variableexactregionregionpassed
/categories_table/1/categoryexactsouthsouthpassed
/categories_table/1/countexact660.00E+2passed
/categories_table/1/masstolerance0.0869565217391304320.086956521739130433.69E-1917passed
/categories_table/1/contributions/0tolerance0.00988699194907927990.0098869919490794331.53E-1613passed
/categories_table/1/cosine_squared/0tolerance0.027748774440872880.0277487744408732133.33E-1613passed
/categories_table/1/contributions/1tolerance0.252142814952255880.25214281495225721.33E-1514passed
/categories_table/1/cosine_squared/1tolerance0.445769629518454170.445769629518455068.87E-1614passed
/categories_table/2/variableexactregionregionpassed
/categories_table/2/categoryexactwestwestpassed
/categories_table/2/countexact550.00E+2passed
/categories_table/2/masstolerance0.0724637681159420320.072463768115942034.00E-1917passed
/categories_table/2/contributions/0tolerance0.170760549201174640.170760549201174372.74E-1614passed
/categories_table/2/cosine_squared/0tolerance0.452630260655049740.45263026065504916.63E-1614passed
/categories_table/2/contributions/1tolerance0.17784189528543060.17784189528543064.34E-1816passed
/categories_table/2/cosine_squared/1tolerance0.296943876477607780.296943876477607951.69E-1615passed
/categories_table/3/variableexactsizesizepassed
/categories_table/3/categoryexactlargelargepassed
/categories_table/3/countexact11110.00E+2passed
/categories_table/3/masstolerance0.159420289855072480.159420289855072452.99E-1715passed
/categories_table/3/contributions/0tolerance0.14774288127464230.147742881274642272.63E-1715passed
/categories_table/3/cosine_squared/0tolerance0.587426948209967990.5874269482099681.77E-1817passed
/categories_table/3/contributions/1tolerance0.00891363251896840240.008913632518968251.53E-1613passed
/categories_table/3/cosine_squared/1tolerance0.0223247333489493550.0223247333489489843.71E-1613passed
/categories_table/4/variableexactsizesizepassed
/categories_table/4/categoryexactsmallsmallpassed
/categories_table/4/countexact12120.00E+2passed
/categories_table/4/masstolerance0.173913043478260860.173913043478260864.74E-1816passed
/categories_table/4/contributions/0tolerance0.135430974501755260.13543097450175541.37E-1614passed
/categories_table/4/cosine_squared/0tolerance0.587426948209967880.5874269482099681.12E-1615passed
/categories_table/4/contributions/1tolerance0.00817082980905439040.0081708298090542271.63E-1613passed
/categories_table/4/cosine_squared/1tolerance0.0223247333489494420.022324733348948994.51E-1613passed
/categories_table/5/variableexactgradegradepassed
/categories_table/5/categoryexactAApassed
/categories_table/5/countexact990.00E+2passed
/categories_table/5/masstolerance0.130434782608695650.130434782608695651.45E-1816passed
/categories_table/5/contributions/0tolerance0.233839551395222140.23383955139522231.66E-1615passed
/categories_table/5/cosine_squared/0tolerance0.796926890023612940.79692689002361344.48E-1615passed
/categories_table/5/contributions/1tolerance0.0417303511019291720.041730351101929116.22E-1714passed
/categories_table/5/cosine_squared/1tolerance0.0895853106152144220.08958531061521431.24E-1614passed
/categories_table/6/variableexactgradegradepassed
/categories_table/6/categoryexactBBpassed
/categories_table/6/countexact440.00E+2passed
/categories_table/6/masstolerance0.0579710144927536240.0579710144927536241.08E-1917passed
/categories_table/6/contributions/0tolerance0.00173096148607174630.0017309614860717651.86E-1713passed
/categories_table/6/cosine_squared/0tolerance0.00434672686747747390.0043467268674775174.34E-1714passed
/categories_table/6/contributions/1tolerance0.45100313552042370.45100313552042324.96E-1614passed
/categories_table/6/cosine_squared/1tolerance0.713409291142646620.71340929114264551.11E-1514passed
/categories_table/7/variableexactgradegradepassed
/categories_table/7/categoryexactCCpassed
/categories_table/7/countexact10100.00E+2passed
/categories_table/7/masstolerance0.144927536231884060.144927536231884063.20E-1816passed
/categories_table/7/contributions/0tolerance0.187005386734437120.18700538673443728.12E-1715passed
/categories_table/7/cosine_squared/0tolerance0.686340002497234040.68634000249723373.37E-1615passed
/categories_table/7/contributions/1tolerance0.0533330933185703420.053333093318570471.25E-1614passed
/categories_table/7/cosine_squared/1tolerance0.12330089300971930.123300893009719512.12E-1614passed
/incomplete_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

28  emit <- function(id, title, kind, configuration, outputs, notes) {
29    expected <- list(schema_version = "2.0.0", case_id = id, outputs = outputs,
30                     expected_error = NULL, deferred_outputs = list(), generator_run = run)
31    write_expected("multivariate", id, expected)
32    case <- list(
33      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "multivariate",
34      entry_point = "tabelz.analysis.multiple_correspondence_analysis:compute_mca",
35      menu_items = list("multipleCorrespondenceMenuItem"), title = title, case_kind = kind,
36      dataset = list(path = "reference_data/datasets/MV_MCA.json", sha256 = ""),
37      excluded_rows = list(), configuration = configuration, "repeat" = 1L,
38      expected = list(path = paste0("reference_data/expected/multivariate/", id, ".json"),
39                      sha256 = ""),
40      oracle = list(oracle_class = "independent_implementation",
41                    source = "R ca::mjca (lambda = 'indicator' and 'adjusted') with the textbook supplementary projection",
42                    citation = "Nenadic and Greenacre (2007), Correspondence analysis in R, with two- and three-dimensional graphics: the ca package, Journal of Statistical Software 20(3); Greenacre (2007) Correspondence Analysis in Practice, 2nd ed., chapters 18-19.",
43                    generator = list(path = "generators/r/multivariate_mca.R", sha256 = ""),
44                    independent = TRUE,
45                    convention_notes = c(notes, "Expected computed from the exported CSV.", manual)),
46      supported = TRUE, unsupported_reason = NULL,
47      requirement_ids = list("REQ-STAT-MV-MCA-001"), risk_ids = list("RISK-REFERENCE-001"),
48      review_reference = NULL
49    )
50    write_case("multivariate", id, case)

Whole script generators/r/multivariate_mca.R


Indicator MCA reporting one dimension

Case STAT-MV-MCA-DIM1-001 · status: passed

Reference

Evidence type
R
Source
R ca::mjca (lambda = 'indicator' and 'adjusted') with the textbook supplementary projection
Citation
Nenadic and Greenacre (2007), Correspondence analysis in R, with two- and three-dimensional graphics: the ca package, Journal of Statistical Software 20(3); Greenacre (2007) Correspondence Analysis in Practice, 2nd ed., chapters 18-19.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · ca 0.72, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_mca.R · sha256 6edebdab7f958ce0
Dataset
reference_data/datasets/MV_MCA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_analysedexact23230.00E+2passed
/variablesexact330.00E+2passed
/categoriesexact880.00E+2passed
/total_inertiatolerance1.66666666666666671.66666666666666674.07E-1716passed
/greenacre_denominatortolerance0.357042648709314290.357042648709315961.67E-1514passed
/dimensions/0/inertiatolerance0.691479734948668880.69147973494866923.31E-1615passed
/dimensions/0/raw_percenttolerance41.48878409692013541.488784096920151.41E-1415passed
/dimensions/0/cumulative_percenttolerance41.48878409692012841.488784096920152.11E-1415passed
/dimensions/0/adjustedtolerance0.28860490122752980.28860490122753034.97E-1614passed
/dimensions/0/benzecri_percenttolerance92.46443382881456392.464433828814532.81E-1415passed
/dimensions/0/greenacre_percenttolerance80.8320524931174380.83205249311722.27E-1314passed
/dimensions/1/inertiatolerance0.435575767039333870.435575767039334263.91E-1615passed
/dimensions/1/raw_percenttolerance26.13454602236003126.1345460223600552.44E-1415passed
/dimensions/1/cumulative_percenttolerance67.62333011928015567.623330119280215.66E-1415passed
/dimensions/1/adjustedtolerance0.0235204093127833160.0235204093127834961.80E-1614passed
/dimensions/1/benzecri_percenttolerance7.53556617118543237.5355661711854754.26E-1414passed
/dimensions/1/greenacre_percenttolerance6.58756296980432146.5875629698043422.05E-1414passed
/dimensions/2/inertiatolerance0.305199490161439770.30519949016144024.42E-1614passed
/dimensions/2/raw_percenttolerance18.31196940968638418.3119694096864122.83E-1414passed
/dimensions/2/cumulative_percenttolerance85.93529952896655385.935299528966627.09E-1415passed
/dimensions/2/adjustedtolerance00.00.00E+2passed
/dimensions/2/benzecri_percenttolerance00.00.00E+2passed
/dimensions/2/greenacre_percenttolerance00.00.00E+2passed
/dimensions/3/inertiatolerance0.167665573593928170.16766557359392831.38E-1615passed
/dimensions/3/raw_percenttolerance10.0599344156356910.0599344156356988.46E-1515passed
/dimensions/3/cumulative_percenttolerance95.99523394460223495.995233944602328.48E-1415passed
/dimensions/3/adjustedtolerance00.00.00E+2passed
/dimensions/3/benzecri_percenttolerance00.00.00E+2passed
/dimensions/3/greenacre_percenttolerance00.00.00E+2passed
/dimensions/4/inertiatolerance0.0667461009232949150.066746100923294838.29E-1714passed
/dimensions/4/raw_percenttolerance4.00476605539769454.004766055397694.46E-1514passed
/dimensions/4/cumulative_percenttolerance99.999999999999929100.07.10E-1415passed
/dimensions/4/adjustedtolerance00.00.00E+2passed
/dimensions/4/benzecri_percenttolerance00.00.00E+2passed
/dimensions/4/greenacre_percenttolerance00.00.00E+2passed
/categories_table/0/variableexactregionregionpassed
/categories_table/0/categoryexactnorthnorthpassed
/categories_table/0/countexact12120.00E+2passed
/categories_table/0/masstolerance0.173913043478260860.173913043478260864.74E-1816passed
/categories_table/0/coordinatessign_invariant_sequence['-0.6720753766050902'][-0.6720753766050903]passed
/categories_table/0/contributions/0tolerance0.113602703457617940.11360270345761793.88E-1715passed
/categories_table/0/cosine_squared/0tolerance0.492747612915135690.49274761291513542.77E-1615passed
/categories_table/1/variableexactregionregionpassed
/categories_table/1/categoryexactsouthsouthpassed
/categories_table/1/countexact660.00E+2passed
/categories_table/1/masstolerance0.0869565217391304320.086956521739130433.69E-1917passed
/categories_table/1/coordinatessign_invariant_sequence['0.2803953059208964'][0.2803953059208986]passed
/categories_table/1/contributions/0tolerance0.00988699194907927990.0098869919490794331.53E-1613passed
/categories_table/1/cosine_squared/0tolerance0.027748774440872880.0277487744408732133.33E-1613passed
/categories_table/2/variableexactregionregionpassed
/categories_table/2/categoryexactwestwestpassed
/categories_table/2/countexact550.00E+2passed
/categories_table/2/masstolerance0.0724637681159420320.072463768115942034.00E-1917passed
/categories_table/2/coordinatessign_invariant_sequence['1.2765065367471409'][1.2765065367471402]passed
/categories_table/2/contributions/0tolerance0.170760549201174640.170760549201174372.74E-1614passed
/categories_table/2/cosine_squared/0tolerance0.452630260655049740.45263026065504916.63E-1614passed
/categories_table/3/variableexactsizesizepassed
/categories_table/3/categoryexactlargelargepassed
/categories_table/3/countexact11110.00E+2passed
/categories_table/3/masstolerance0.159420289855072480.159420289855072452.99E-1715passed
/categories_table/3/coordinatessign_invariant_sequence['-0.80051820594364875'][-0.800518205943649]passed
/categories_table/3/contributions/0tolerance0.14774288127464230.147742881274642272.63E-1715passed
/categories_table/3/cosine_squared/0tolerance0.587426948209967990.5874269482099681.77E-1817passed
/categories_table/4/variableexactsizesizepassed
/categories_table/4/categoryexactsmallsmallpassed
/categories_table/4/countexact12120.00E+2passed
/categories_table/4/masstolerance0.173913043478260860.173913043478260864.74E-1816passed
/categories_table/4/coordinatessign_invariant_sequence['0.73380835544834422'][0.7338083554483448]passed
/categories_table/4/contributions/0tolerance0.135430974501755260.13543097450175541.37E-1614passed
/categories_table/4/cosine_squared/0tolerance0.587426948209967880.5874269482099681.12E-1615passed
/categories_table/5/variableexactgradegradepassed
/categories_table/5/categoryexactAApassed
/categories_table/5/countexact990.00E+2passed
/categories_table/5/masstolerance0.130434782608695650.130434782608695651.45E-1816passed
/categories_table/5/coordinatessign_invariant_sequence['-1.1134020168599672'][-1.1134020168599679]passed
/categories_table/5/contributions/0tolerance0.233839551395222140.23383955139522231.66E-1615passed
/categories_table/5/cosine_squared/0tolerance0.796926890023612940.79692689002361344.48E-1615passed
/categories_table/6/variableexactgradegradepassed
/categories_table/6/categoryexactBBpassed
/categories_table/6/countexact440.00E+2passed
/categories_table/6/masstolerance0.0579710144927536240.0579710144927536241.08E-1917passed
/categories_table/6/coordinatessign_invariant_sequence['0.14369047505147298'][0.14369047505147378]passed
/categories_table/6/contributions/0tolerance0.00173096148607174630.0017309614860717651.86E-1713passed
/categories_table/6/cosine_squared/0tolerance0.00434672686747747390.0043467268674775174.34E-1714passed
/categories_table/7/variableexactgradegradepassed
/categories_table/7/categoryexactCCpassed
/categories_table/7/countexact10100.00E+2passed
/categories_table/7/masstolerance0.144927536231884060.144927536231884063.20E-1816passed
/categories_table/7/coordinatessign_invariant_sequence['0.94458562515338074'][0.9445856251533812]passed
/categories_table/7/contributions/0tolerance0.187005386734437120.18700538673443728.12E-1715passed
/categories_table/7/cosine_squared/0tolerance0.686340002497234040.68634000249723373.37E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

187            "The first three inertias are distinct, so each axis is unique up to sign; with two axes the signs are independent, so the sign-free contributions and squared cosines are compared here and the signed first-axis coordinates in STAT-MV-MCA-DIM1-001."))

Whole script generators/r/multivariate_mca.R


Indicator MCA with missing values as a category

Case STAT-MV-MCA-MISSING-001 · status: passed

Reference

Evidence type
R
Source
R ca::mjca (lambda = 'indicator' and 'adjusted') with the textbook supplementary projection
Citation
Nenadic and Greenacre (2007), Correspondence analysis in R, with two- and three-dimensional graphics: the ca package, Journal of Statistical Software 20(3); Greenacre (2007) Correspondence Analysis in Practice, 2nd ed., chapters 18-19.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · ca 0.72, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_mca.R · sha256 6edebdab7f958ce0
Dataset
reference_data/datasets/MV_MCA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_analysedexact24240.00E+2passed
/variablesexact330.00E+2passed
/categoriesexact990.00E+2passed
/total_inertiatolerance22.00.00E+2passed
/greenacre_denominatortolerance0.390740740740738610.39074074074074011.50E-1514passed
/dimensions/0/inertiatolerance0.711299136787406530.71129913678740737.75E-1614passed
/dimensions/0/raw_percenttolerance35.56495683937032535.564956839370374.24E-1414passed
/dimensions/0/cumulative_percenttolerance35.56495683937032535.564956839370374.24E-1414passed
/dimensions/0/adjustedtolerance0.321430834306537020.32143083430653831.27E-1514passed
/dimensions/0/benzecri_percenttolerance92.04548862100267692.04548862100272.81E-1415passed
/dimensions/0/greenacre_percenttolerance82.26191968034642382.261919680346441.46E-1415passed
/dimensions/1/inertiatolerance0.443909438786017460.443909438786017462.86E-1817passed
/dimensions/1/raw_percenttolerance22.19547193930087322.1954719393008733.22E-1717passed
/dimensions/1/cumulative_percenttolerance57.76042877867120557.760428778671232.84E-1415passed
/dimensions/1/adjustedtolerance0.0275109189684370310.027510918968437032.54E-1917passed
/dimensions/1/benzecri_percenttolerance7.87807424986389387.87807424986386452.93E-1414passed
/dimensions/1/greenacre_percenttolerance7.04070911988439727.040709119884372.76E-1414passed
/dimensions/2/inertiatolerance0.344225228147079180.34422522814707892.81E-1615passed
/dimensions/2/raw_percenttolerance17.21126140735395917.2112614073539451.43E-1415passed
/dimensions/2/cumulative_percenttolerance74.97169018602517174.971690186025191.42E-1415passed
/dimensions/2/adjustedtolerance0.000266925088425833980.00026692508842582041.36E-1713passed
/dimensions/2/benzecri_percenttolerance0.0764371291334465520.076437129133442374.18E-1513passed
/dimensions/2/greenacre_percenttolerance0.0683125818720147630.0683125818720113.76E-1513passed
/dimensions/3/inertiatolerance0.273680010131871880.2736800101318721.07E-1615passed
/dimensions/3/raw_percenttolerance13.68400050659359313.68400050659367.12E-1515passed
/dimensions/3/cumulative_percenttolerance88.65569069261874888.655690692618782.84E-1415passed
/dimensions/3/adjustedtolerance00.00.00E+2passed
/dimensions/3/benzecri_percenttolerance00.00.00E+2passed
/dimensions/3/greenacre_percenttolerance00.00.00E+2passed
/dimensions/4/inertiatolerance0.163582975494571040.163582975494571042.36E-1816passed
/dimensions/4/raw_percenttolerance8.1791487747285538.1791487747285534.92E-1717passed
/dimensions/4/cumulative_percenttolerance96.83483946734729496.834839467347344.25E-1415passed
/dimensions/4/adjustedtolerance00.00.00E+2passed
/dimensions/4/benzecri_percenttolerance00.00.00E+2passed
/dimensions/4/greenacre_percenttolerance00.00.00E+2passed
/dimensions/5/inertiatolerance0.0633032106530528330.06330321065305282.81E-1715passed
/dimensions/5/raw_percenttolerance3.16516053265264173.16516053265264041.29E-1515passed
/dimensions/5/cumulative_percenttolerance99.99999999999994399.999999999999972.86E-1415passed
/dimensions/5/adjustedtolerance00.00.00E+2passed
/dimensions/5/benzecri_percenttolerance00.00.00E+2passed
/dimensions/5/greenacre_percenttolerance00.00.00E+2passed
/categories_table/0/variableexactregionregionpassed
/categories_table/0/categoryexactnorthnorthpassed
/categories_table/0/countexact12120.00E+2passed
/categories_table/0/masstolerance0.166666666666666660.166666666666666662.59E-1816passed
/categories_table/0/contributions/0tolerance0.0747422622428528250.074742262242852951.25E-1614passed
/categories_table/0/cosine_squared/0tolerance0.318984639689274870.318984639689276041.17E-1514passed
/categories_table/0/contributions/1tolerance0.00113820184530750510.00113820184530750123.88E-1814passed
/categories_table/0/cosine_squared/1tolerance0.00303155125425398280.0030315512542539757.77E-1814passed
/categories_table/1/variableexactregionregionpassed
/categories_table/1/categoryexactsouthsouthpassed
/categories_table/1/countexact660.00E+2passed
/categories_table/1/masstolerance0.0833333333333333290.083333333333333332.93E-1917passed
/categories_table/1/contributions/0tolerance0.0138142709387470160.0138142709387469645.21E-1714passed
/categories_table/1/cosine_squared/0tolerance0.0393043159763124470.039304315976312331.18E-1614passed
/categories_table/1/contributions/1tolerance0.204054778360136770.204054778360137022.54E-1614passed
/categories_table/1/cosine_squared/1tolerance0.362327368573814090.362327368573814372.78E-1615passed
/categories_table/2/variableexactregionregionpassed
/categories_table/2/categoryexactwestwestpassed
/categories_table/2/countexact550.00E+2passed
/categories_table/2/masstolerance0.0694444444444444480.069444444444444454.72E-1917passed
/categories_table/2/contributions/0tolerance0.157657843576483050.157657843576483161.09E-1615passed
/categories_table/2/cosine_squared/0tolerance0.424958733639346890.42495873363934693.84E-1817passed
/categories_table/2/contributions/1tolerance0.197125655062521390.197125655062522228.30E-1614passed
/categories_table/2/cosine_squared/1tolerance0.33160145270828240.33160145270828327.73E-1614passed
/categories_table/3/variableexactregionregionpassed
/categories_table/3/categoryexact(missing value)(missing value)passed
/categories_table/3/countexact110.00E+2passed
/categories_table/3/masstolerance0.0138888888888888880.0138888888888888881.18E-1917passed
/categories_table/3/contributions/0tolerance0.0523052219766615210.0523052219766616058.37E-1714passed
/categories_table/3/cosine_squared/0tolerance0.116466759364611250.116466759364611613.58E-1614passed
/categories_table/3/contributions/1tolerance0.0531648231530314560.053164823153031044.17E-1614passed
/categories_table/3/cosine_squared/1tolerance0.0738794091412802170.073879409141279665.55E-1614passed
/categories_table/4/variableexactsizesizepassed
/categories_table/4/categoryexactlargelargepassed
/categories_table/4/countexact12120.00E+2passed
/categories_table/4/masstolerance0.166666666666666660.166666666666666662.59E-1816passed
/categories_table/4/contributions/0tolerance0.145352532213612050.1453525322136125.43E-1715passed
/categories_table/4/cosine_squared/0tolerance0.620334784160436190.62033478416043622.30E-1817passed
/categories_table/4/contributions/1tolerance0.00763604695706814830.0076360469570682931.45E-1613passed
/categories_table/4/cosine_squared/1tolerance0.0203382799155348050.0203382799155351763.71E-1613passed
/categories_table/5/variableexactsizesizepassed
/categories_table/5/categoryexactsmallsmallpassed
/categories_table/5/countexact12120.00E+2passed
/categories_table/5/masstolerance0.166666666666666660.166666666666666662.59E-1816passed
/categories_table/5/contributions/0tolerance0.145352532213612080.1453525322136128.43E-1715passed
/categories_table/5/cosine_squared/0tolerance0.62033478416043630.62033478416043612.19E-1615passed
/categories_table/5/contributions/1tolerance0.00763604695706809450.0076360469570682381.43E-1613passed
/categories_table/5/cosine_squared/1tolerance0.0203382799155346630.0203382799155350243.61E-1613passed
/categories_table/6/variableexactgradegradepassed
/categories_table/6/categoryexactAApassed
/categories_table/6/countexact10100.00E+2passed
/categories_table/6/masstolerance0.13888888888888890.13888888888888894.94E-1816passed
/categories_table/6/contributions/0tolerance0.222810472717245160.222810472717245271.13E-1615passed
/categories_table/6/cosine_squared/0tolerance0.815065184113563210.81506518411356342.25E-1615passed
/categories_table/6/contributions/1tolerance0.0262784649852881270.0262784649852881583.10E-1714passed
/categories_table/6/cosine_squared/1tolerance0.0599927587394259860.059992758739425987.02E-1815passed
/categories_table/7/variableexactgradegradepassed
/categories_table/7/categoryexactBBpassed
/categories_table/7/countexact440.00E+2passed
/categories_table/7/masstolerance0.0555555555555555520.055555555555555554.72E-1917passed
/categories_table/7/contributions/0tolerance0.0048688203760994990.0048688203760994623.73E-1714passed
/categories_table/7/cosine_squared/0tolerance0.0124674758304930240.0124674758304929329.22E-1714passed
/categories_table/7/contributions/1tolerance0.437351733924604250.437351733924604866.07E-1614passed
/categories_table/7/cosine_squared/1tolerance0.698920425930825150.69892042593082564.45E-1615passed
/categories_table/8/variableexactgradegradepassed
/categories_table/8/categoryexactCCpassed
/categories_table/8/countexact10100.00E+2passed
/categories_table/8/masstolerance0.13888888888888890.13888888888888894.94E-1816passed
/categories_table/8/contributions/0tolerance0.183096043744686850.183096043744686883.16E-1715passed
/categories_table/8/cosine_squared/0tolerance0.669785440447466350.66978544044746684.46E-1615passed
/categories_table/8/contributions/1tolerance0.0656142487549739440.065614248754973944.89E-1917passed
/categories_table/8/cosine_squared/1tolerance0.14979489089753140.14979489089753138.69E-1715passed
/incomplete_countexact000.00E+2passed
/categories_table/3/is_missingexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

195  d_missing <- d
196  d_missing$region[is.na(d_missing$region)] <- "(missing value)"
197  ref <- mca_reference(d_missing, active)
198  stopifnot(ref$N == 24, any(grepl("(missing value)", ref$names, fixed = TRUE)))
199  missing_index <- which(ref$names == "region:(missing value)")
200  outputs <- outputs_for(ref, 2L)
201  outputs[["/incomplete_count"]] <- exact(0L, "the missing region is a category, so no row is incomplete")
202  outputs[[sprintf("/categories_table/%d/is_missing", missing_index - 1L)]] <- exact(TRUE, "the typed missing category")
203  emit("STAT-MV-MCA-MISSING-001", "Indicator MCA with missing values as a category", "missing_excluded",
204       list(columns = list("region", "size", "grade"), missing = "category", dimensions = 2L),
205       outputs,
206       list("The missing region is recoded as the category '(missing value)' before ca::mjca, so all 24 rows are active and region has four categories (J = 9).",
207            "The product lists the missing category last within its variable; the recoded label sorts after the text categories in R as well (asserted by the index used)."))

Whole script generators/r/multivariate_mca.R


Indicator MCA with a supplementary variable

Case STAT-MV-MCA-SUPP-001 · status: passed

Reference

Evidence type
R
Source
R ca::mjca (lambda = 'indicator' and 'adjusted') with the textbook supplementary projection
Citation
Nenadic and Greenacre (2007), Correspondence analysis in R, with two- and three-dimensional graphics: the ca package, Journal of Statistical Software 20(3); Greenacre (2007) Correspondence Analysis in Practice, 2nd ed., chapters 18-19.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · ca 0.72, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_mca.R · sha256 6edebdab7f958ce0
Dataset
reference_data/datasets/MV_MCA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows_analysedexact23230.00E+2passed
/variablesexact330.00E+2passed
/categoriesexact880.00E+2passed
/total_inertiatolerance1.66666666666666671.66666666666666674.07E-1716passed
/greenacre_denominatortolerance0.357042648709314290.357042648709315961.67E-1514passed
/dimensions/0/inertiatolerance0.691479734948668880.69147973494866923.31E-1615passed
/dimensions/0/raw_percenttolerance41.48878409692013541.488784096920151.41E-1415passed
/dimensions/0/cumulative_percenttolerance41.48878409692012841.488784096920152.11E-1415passed
/dimensions/0/adjustedtolerance0.28860490122752980.28860490122753034.97E-1614passed
/dimensions/0/benzecri_percenttolerance92.46443382881456392.464433828814532.81E-1415passed
/dimensions/0/greenacre_percenttolerance80.8320524931174380.83205249311722.27E-1314passed
/dimensions/1/inertiatolerance0.435575767039333870.435575767039334263.91E-1615passed
/dimensions/1/raw_percenttolerance26.13454602236003126.1345460223600552.44E-1415passed
/dimensions/1/cumulative_percenttolerance67.62333011928015567.623330119280215.66E-1415passed
/dimensions/1/adjustedtolerance0.0235204093127833160.0235204093127834961.80E-1614passed
/dimensions/1/benzecri_percenttolerance7.53556617118543237.5355661711854754.26E-1414passed
/dimensions/1/greenacre_percenttolerance6.58756296980432146.5875629698043422.05E-1414passed
/dimensions/2/inertiatolerance0.305199490161439770.30519949016144024.42E-1614passed
/dimensions/2/raw_percenttolerance18.31196940968638418.3119694096864122.83E-1414passed
/dimensions/2/cumulative_percenttolerance85.93529952896655385.935299528966627.09E-1415passed
/dimensions/2/adjustedtolerance00.00.00E+2passed
/dimensions/2/benzecri_percenttolerance00.00.00E+2passed
/dimensions/2/greenacre_percenttolerance00.00.00E+2passed
/dimensions/3/inertiatolerance0.167665573593928170.16766557359392831.38E-1615passed
/dimensions/3/raw_percenttolerance10.0599344156356910.0599344156356988.46E-1515passed
/dimensions/3/cumulative_percenttolerance95.99523394460223495.995233944602328.48E-1415passed
/dimensions/3/adjustedtolerance00.00.00E+2passed
/dimensions/3/benzecri_percenttolerance00.00.00E+2passed
/dimensions/3/greenacre_percenttolerance00.00.00E+2passed
/dimensions/4/inertiatolerance0.0667461009232949150.066746100923294838.29E-1714passed
/dimensions/4/raw_percenttolerance4.00476605539769454.004766055397694.46E-1514passed
/dimensions/4/cumulative_percenttolerance99.999999999999929100.07.10E-1415passed
/dimensions/4/adjustedtolerance00.00.00E+2passed
/dimensions/4/benzecri_percenttolerance00.00.00E+2passed
/dimensions/4/greenacre_percenttolerance00.00.00E+2passed
/categories_table/0/variableexactregionregionpassed
/categories_table/0/categoryexactnorthnorthpassed
/categories_table/0/countexact12120.00E+2passed
/categories_table/0/masstolerance0.173913043478260860.173913043478260864.74E-1816passed
/categories_table/0/coordinatessign_invariant_sequence['-0.6720753766050902'][-0.6720753766050903]passed
/categories_table/0/contributions/0tolerance0.113602703457617940.11360270345761793.88E-1715passed
/categories_table/0/cosine_squared/0tolerance0.492747612915135690.49274761291513542.77E-1615passed
/categories_table/1/variableexactregionregionpassed
/categories_table/1/categoryexactsouthsouthpassed
/categories_table/1/countexact660.00E+2passed
/categories_table/1/masstolerance0.0869565217391304320.086956521739130433.69E-1917passed
/categories_table/1/coordinatessign_invariant_sequence['0.2803953059208964'][0.2803953059208986]passed
/categories_table/1/contributions/0tolerance0.00988699194907927990.0098869919490794331.53E-1613passed
/categories_table/1/cosine_squared/0tolerance0.027748774440872880.0277487744408732133.33E-1613passed
/categories_table/2/variableexactregionregionpassed
/categories_table/2/categoryexactwestwestpassed
/categories_table/2/countexact550.00E+2passed
/categories_table/2/masstolerance0.0724637681159420320.072463768115942034.00E-1917passed
/categories_table/2/coordinatessign_invariant_sequence['1.2765065367471409'][1.2765065367471402]passed
/categories_table/2/contributions/0tolerance0.170760549201174640.170760549201174372.74E-1614passed
/categories_table/2/cosine_squared/0tolerance0.452630260655049740.45263026065504916.63E-1614passed
/categories_table/8/variableexactsectorsectorpassed
/categories_table/8/categoryexactmixedmixedpassed
/categories_table/8/countexact770.00E+2passed
/categories_table/8/supplementaryexactTrueTruepassed
/categories_table/8/coordinatessign_invariant_sequence['-0.086904046842107105'][-0.08690404684210698]passed
/categories_table/9/variableexactsectorsectorpassed
/categories_table/9/categoryexactprivateprivatepassed
/categories_table/9/countexact990.00E+2passed
/categories_table/9/supplementaryexactTrueTruepassed
/categories_table/9/coordinatessign_invariant_sequence['-0.14174653089154446'][-0.14174653089154457]passed
/categories_table/10/variableexactsectorsectorpassed
/categories_table/10/categoryexactpublicpublicpassed
/categories_table/10/countexact770.00E+2passed
/categories_table/10/supplementaryexactTrueTruepassed
/categories_table/10/coordinatessign_invariant_sequence['0.26914958655980725'][0.269149586559807]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

223  emit("STAT-MV-MCA-SUPP-001", "Indicator MCA with a supplementary variable", "axis_coverage",
224       list(columns = list("region", "size", "grade"), supplementary = list("sector"),
225            missing = "complete", dimensions = 1L),
226       outputs,
227       list("The active fit is that of STAT-MV-MCA-001; each sector category is projected as the average of the standard row coordinates (D_r^-1/2 U) of the complete rows holding it, which is the simple-CA supplementary-column principal coordinate.",
228            "Supplementary categories follow the active ones in the categories table, variable by variable, categories sorted; one dimension is requested so the single coordinate is compared up to the axis sign."))

Whole script generators/r/multivariate_mca.R


Gaussian mixture with diagonal covariance, K chosen by BIC

Case STAT-MV-MCLUST-001 · status: passed

Reference

Evidence type
R
Source
R mclust::Mclust (VVI/VVV) with emControl(tol = 1e-12); cluster::silhouette
Citation
Scrucca, Fop, Murphy and Raftery (2016) mclust 5: clustering, classification and density estimation using Gaussian finite mixture models, The R Journal 8(1); Rousseeuw (1987) Silhouettes, J. Comput. Appl. Math. 20.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · mclust 6.1.3, cluster 2.1.8.2, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_model_clustering.R · sha256 b03bbb2bfb4448e9
Dataset
reference_data/datasets/MV_MIXTURE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact29290.00E+2passed
/missing_countexact110.00E+2passed
/chosen_kexact220.00E+2passed
/fits/[k=1]/statusexactfittedfittedpassed
/fits/[k=1]/parametersexact440.00E+2passed
/fits/[k=1]/log_likelihoodtolerance-141.10473027092343-141.104730270938941.55E-1112passed
/fits/[k=1]/bictolerance295.67864386179275295.67864386182383.10E-1112passed
/fits/[k=2]/statusexactfittedfittedpassed
/fits/[k=2]/parametersexact990.00E+2passed
/fits/[k=2]/log_likelihoodtolerance-85.931249845434138-85.931249850762925.33E-910passed
/fits/[k=2]/bictolerance202.16816216074653202.16816217140411.07E-810passed
/silhouettetolerance0.797082271530502480.79708227153050254.15E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

23  emit <- function(id, title, kind, dataset, configuration, outputs, notes, source, citation) {
24    expected <- list(schema_version = "2.0.0", case_id = id, outputs = outputs,
25                     expected_error = NULL, deferred_outputs = list(), generator_run = run)
26    write_expected("multivariate", id, expected)
27    case <- list(
28      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "multivariate",
29      entry_point = "tabelz.analysis.model_based_clustering:compute_model_clustering",
30      menu_items = list("modelClusteringMenuItem"), title = title, case_kind = kind,
31      dataset = list(path = paste0("reference_data/datasets/", dataset, ".json"), sha256 = ""),
32      excluded_rows = list(), configuration = configuration, "repeat" = 1L,
33      expected = list(path = paste0("reference_data/expected/multivariate/", id, ".json"),
34                      sha256 = ""),
35      oracle = list(oracle_class = "independent_implementation", source = source,
36                    citation = citation,
37                    generator = list(path = "generators/r/multivariate_model_clustering.R",
38                                     sha256 = ""),
39                    independent = TRUE,
40                    convention_notes = c(notes, "Expected computed from the exported CSV.", manual)),
41      supported = TRUE, unsupported_reason = NULL,
42      requirement_ids = list("REQ-STAT-MV-MCLUST-001"), risk_ids = list("RISK-REFERENCE-001"),
43      review_reference = NULL
44    )
45    write_case("multivariate", id, case)

Whole script generators/r/multivariate_model_clustering.R


Gaussian mixture with full covariance, two components

Case STAT-MV-MCLUST-FULL-001 · status: passed

Reference

Evidence type
R
Source
R mclust::Mclust (VVI/VVV) with emControl(tol = 1e-12); cluster::silhouette
Citation
Scrucca, Fop, Murphy and Raftery (2016) mclust 5: clustering, classification and density estimation using Gaussian finite mixture models, The R Journal 8(1); Rousseeuw (1987) Silhouettes, J. Comput. Appl. Math. 20.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · mclust 6.1.3, cluster 2.1.8.2, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_model_clustering.R · sha256 b03bbb2bfb4448e9
Dataset
reference_data/datasets/MV_MIXTURE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact29290.00E+2passed
/chosen_kexact220.00E+2passed
/fits/[k=2]/parametersexact11110.00E+2passed
/fits/[k=2]/log_likelihoodtolerance-83.1922537990433-83.192253812506311.35E-89passed
/fits/[k=2]/bictolerance203.42476172793781203.424761754863822.69E-89passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

119  outputs <- list(
120    "/n" = exact(n, "complete rows"),
121    "/chosen_k" = exact(2L, "the only K in the range"),
122    "/fits/[k=2]/parameters" = exact(m2f, "K - 1 + K p means + K p(p+1)/2 covariances"),
123    "/fits/[k=2]/log_likelihood" = tol(fit2_full$loglik, "mclust VVV (full, varying) log-likelihood", rel = "1e-7"),
124    "/fits/[k=2]/bic" = tol(bic_of(fit2_full$loglik, m2f, n), "-2 log L + m log n", rel = "1e-7")
125  )
126  emit("STAT-MV-MCLUST-FULL-001", "Gaussian mixture with full covariance, two components", "axis_coverage",
127       "MV_MIXTURE",
128       list(variables = list("length", "width"), kind = "gaussian", covariance = "full",
129            k_min = 2L, k_max = 2L),
130       outputs,
131       list("mclust::Mclust(G = 2, modelNames = 'VVV') on the 29 complete rows; VVV is the unrestricted per-component covariance model.", em_note),
132       mclust_source, mclust_citation)

Whole script generators/r/multivariate_model_clustering.R


One-component Gaussian fit in closed form

Case STAT-MV-MCLUST-ONE-001 · status: passed

Reference

Evidence type
R
Source
closed-form Gaussian maximum likelihood (base R dnorm)
Citation
Scrucca, Fop, Murphy and Raftery (2016) mclust 5: clustering, classification and density estimation using Gaussian finite mixture models, The R Journal 8(1); Rousseeuw (1987) Silhouettes, J. Comput. Appl. Math. 20.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · mclust 6.1.3, cluster 2.1.8.2, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_model_clustering.R · sha256 b03bbb2bfb4448e9
Dataset
reference_data/datasets/MV_MIXTURE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact29290.00E+2passed
/chosen_kexact110.00E+2passed
/fits/[k=1]/parametersexact440.00E+2passed
/fits/[k=1]/log_likelihoodtolerance-141.10473027092343-141.104730270938941.55E-1112passed
/fits/[k=1]/bictolerance295.67864386179275295.67864386182383.10E-1112passed
/profiles/0/sizeexact29290.00E+2passed
/profiles/0/weighttolerance11.00.00E+2passed
/profiles/0/meanssequence['15.077999999999999', '3.4384482758620689'][15.078, 3.4384482758620685]passed
/profiles/0/variancessequence['21.299738620689656', '2.7099001093935793'][21.299760681133222, 2.7099029160758374]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

134  outputs <- list(
135    "/n" = exact(n, "complete rows"),
136    "/chosen_k" = exact(1L, "the only K in the range"),
137    "/fits/[k=1]/parameters" = exact(m1, "p means and p variances"),
138    "/fits/[k=1]/log_likelihood" = tol(fit1, "closed-form Gaussian log-likelihood with MLE variances", rel = "1e-9"),
139    "/fits/[k=1]/bic" = tol(bic_of(fit1, m1, n), "-2 log L + m log n", rel = "1e-9"),
140    "/profiles/0/size" = exact(n, "every row in the single cluster"),
141    "/profiles/0/weight" = tol(1, "single component weight"),
142    "/profiles/0/means" = list(comparison = "sequence", values = as.list(dec(colMeans(x))),
143                               absolute_tolerance = "0", relative_tolerance = "1e-9",
144                               rationale = "sample means"),
145    "/profiles/0/variances" = list(comparison = "sequence",
146                                   values = as.list(dec(apply(x, 2, function(v) mean((v - mean(v))^2)))),
147                                   absolute_tolerance = "0", relative_tolerance = "1e-5",
148                                   rationale = "MLE variances; the product adds 1e-6 times the sample variance as a floor, hence 1e-5 relative")
149  )
150  emit("STAT-MV-MCLUST-ONE-001", "One-component Gaussian fit in closed form", "boundary",
151       "MV_MIXTURE",
152       list(variables = list("length", "width"), kind = "gaussian", covariance = "diagonal",
153            k_min = 1L, k_max = 1L),
154       outputs,
155       list("With one component EM is the Gaussian MLE: means are the sample means and variances the population (1/n) variances; log L = sum_j sum_i log N(x_ij; mean_j, var_j).",
156            "The product's variance floor (1e-6 times the sample variance) perturbs the variances at 1e-6 relative and the log-likelihood only to second order."),
157       "closed-form Gaussian maximum likelihood (base R dnorm)", mclust_citation)

Whole script generators/r/multivariate_model_clustering.R


Classical scaling of unstandardized city-block distances

Case STAT-MV-MDS-CITYBLOCK-001 · status: passed

Reference

Evidence type
R
Source
Retained R reference with independently reviewed analytical corrections
Citation
R Core Team, stats package help: cmdscale (classical multidimensional scaling, Torgerson 1958; Mardia 1978 goodness-of-fit ratios).
Reference software
not recorded
Reference script
generators/analytic/reviewed_reference_corrections.py · sha256 49e3645836bb3ac3
Dataset
reference_data/datasets/MV_MDS_VARS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/dimensionsexact220.00E+2passed
/gof_absolutetolerance0.776486018189265770.77648601818926582.16E-1918passed
/gof_positivetolerance0.806224988550626190.80622498855062622.94E-1817passed
/eigenvalues/0/eigenvaluetolerance672.92730338881984672.92730338882141.59E-1214passed
/eigenvalues/0/retainedexactTrueTruepassed
/eigenvalues/1/eigenvaluetolerance220.23034351850265220.230343518502568.52E-1415passed
/eigenvalues/1/retainedexactTrueTruepassed
/eigenvalues/2/eigenvaluetolerance103.26657434760966103.266574347609973.15E-1314passed
/eigenvalues/2/retainedexactFalseFalsepassed
/eigenvalues/3/eigenvaluetolerance47.97806152899610147.9780615289960445.73E-1414passed
/eigenvalues/3/retainedexactFalseFalsepassed
/eigenvalues/4/eigenvaluetolerance40.04225435297606640.042254352976191.21E-1314passed
/eigenvalues/4/retainedexactFalseFalsepassed
/eigenvalues/5/eigenvaluetolerance11.72487707681434411.7248770768143483.99E-1515passed
/eigenvalues/5/retainedexactFalseFalsepassed
/eigenvalues/6/eigenvaluetolerance8.50503981186959388.5050398118695375.69E-1414passed
/eigenvalues/6/retainedexactFalseFalsepassed
/eigenvalues/7/eigenvaluetolerance3.15234319757605163.1523431975760123.95E-1413passed
/eigenvalues/7/retainedexactFalseFalsepassed
/eigenvalues/8/eigenvaluetolerance04.3955782049271525e-144.40E-14passed
/eigenvalues/8/retainedexactFalseFalsepassed
/eigenvalues/9/eigenvaluetolerance-5.6757617106128926-5.6757617106128484.44E-1414passed
/eigenvalues/9/retainedexactFalseFalsepassed
/eigenvalues/10/eigenvaluetolerance-36.753370967098888-36.753370967098945.01E-1414passed
/eigenvalues/10/retainedexactFalseFalsepassed
/objects/0/coordinatessequence['-2.4812798330800709', '0.96146365911720877'][-2.481279833080072, 0.9614636591172209]passed
/objects/1/coordinatessequence['-2.3398777387581386', '3.2911253455489216'][-2.339877738758127, 3.2911253455489122]passed
/objects/2/coordinatessequence['-7.5094623984637865', '-6.6449895970314827'][-7.509462398463808, -6.644989597031502]passed
/objects/3/coordinatessequence['-3.3602813813003882', '6.5050248763842085'][-3.3602813813003864, 6.505024876384214]passed
/objects/4/coordinatessequence['15.65097061757503', '-4.0860336047804982'][15.650970617575078, -4.086033604780502]passed
/objects/5/coordinatessequence['13.681419646746303', '-0.99803533480064366'][13.681419646746315, -0.9980353348006439]passed
/objects/6/coordinatessequence['-5.0457743355063336', '1.7188593431388968'][-5.045774335506331, 1.7188593431389114]passed
/objects/7/coordinatessequence['-5.8608029962304009', '-3.7384024046778119'][-5.860802996230433, -3.7384024046778315]passed
/objects/8/coordinatessequence['5.4844362976911167', '0.25928400704953908'][5.484436297691114, 0.25928400704950566]passed
/objects/9/coordinatessequence['-8.4577282722636529', '-5.0987820796134891'][-8.457728272263664, -5.098782079613489]passed
/objects/10/coordinatessequence['0.238380393590328', '7.8304857896652207'][0.2383803935903263, 7.830485789665198]passed
/rows_usedexact11110.00E+2passed

Convention

Reference code

Whole script generators/analytic/reviewed_reference_corrections.py


Classical scaling of correlation distances between variables

Case STAT-MV-MDS-CORRELATION-001 · status: passed

Reference

Evidence type
R
Source
Retained R reference with independently reviewed analytical corrections
Citation
R Core Team, stats package help: cmdscale (classical multidimensional scaling, Torgerson 1958; Mardia 1978 goodness-of-fit ratios).
Reference software
not recorded
Reference script
generators/analytic/reviewed_reference_corrections.py · sha256 49e3645836bb3ac3
Dataset
reference_data/datasets/MV_MDS_VARS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/dimensionsexact220.00E+2passed
/gof_absolutetolerance0.785426480440145290.78542648044014631.00E-1514passed
/gof_positivetolerance0.785426480440145290.78542648044014631.00E-1514passed
/eigenvalues/0/eigenvaluetolerance1.54047432149240171.54047432149240224.54E-1615passed
/eigenvalues/0/retainedexactTrueTruepassed
/eigenvalues/1/eigenvaluetolerance0.99344612496590590.99344612496590655.52E-1615passed
/eigenvalues/1/retainedexactTrueTruepassed
/eigenvalues/2/eigenvaluetolerance0.692250951580528630.69225095158052642.22E-1514passed
/eigenvalues/2/retainedexactFalseFalsepassed
/eigenvalues/3/eigenvaluetolerance06.705959843412296e-186.71E-18passed
/eigenvalues/3/retainedexactFalseFalsepassed
/objects/0/coordinatessequence['0.79570933992228288', '0.11896890050703234'][0.7957093399222828, 0.11896890050703247]passed
/objects/1/coordinatessequence['0.39360135137049856', '-0.36196788595957713'][0.3936013513704991, -0.3619678859595771]passed
/objects/2/coordinatessequence['-0.44437396982006383', '0.76132271899526494'][-0.4443739698200644, 0.7613227189952652]passed
/objects/3/coordinatessequence['-0.7449367214727175', '-0.51832373354272065'][-0.7449367214727174, -0.518323733542721]passed
/rows_usedexact11110.00E+2passed
/objects/0/labelexactv1v1passed
/objects/3/labelexactv4v4passed

Convention

Reference code

Whole script generators/analytic/reviewed_reference_corrections.py


Classical scaling of a one-dimensional supplied matrix

Case STAT-MV-MDS-MATRIX-001 · status: passed

Reference

Evidence type
R
Source
Retained R reference with independently reviewed analytical corrections
Citation
R Core Team, stats package help: cmdscale (classical multidimensional scaling, Torgerson 1958; Mardia 1978 goodness-of-fit ratios).
Reference software
not recorded
Reference script
generators/analytic/reviewed_reference_corrections.py · sha256 49e3645836bb3ac3
Dataset
reference_data/datasets/MV_MDS_CONVEYOR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/dimensionsexact110.00E+2passed
/gof_absolutetolerance0.999999999999999220.99999999999999985.58E-1615passed
/gof_positivetolerance0.999999999999999220.99999999999999985.58E-1615passed
/eigenvalues/0/eigenvaluetolerance4.66666666666666794.6666666666666684.91E-1716passed
/eigenvalues/0/retainedexactTrueTruepassed
/eigenvalues/1/eigenvaluetolerance01.2606523053526741e-151.26E-15passed
/eigenvalues/1/retainedexactFalseFalsepassed
/eigenvalues/2/eigenvaluetolerance0-3.724738856525493e-163.72E-16passed
/eigenvalues/2/retainedexactFalseFalsepassed
/objects/0/coordinatessequence['-1.333333333333333'][-1.3333333333333333]passed
/objects/1/coordinatessequence['-0.33333333333333365'][-0.3333333333333339]passed
/objects/2/coordinatessequence['1.6666666666666674'][1.666666666666667]passed
/objects/0/labelexactAApassed
/rows_usedexact330.00E+2passed
/missing_countexact000.00E+2passed

Convention

Reference code

Whole script generators/analytic/reviewed_reference_corrections.py


Classical scaling of a non-Euclidean matrix with labels

Case STAT-MV-MDS-MATRIX-002 · status: passed

Reference

Evidence type
R
Source
Retained R reference with independently reviewed analytical corrections
Citation
R Core Team, stats package help: cmdscale (classical multidimensional scaling, Torgerson 1958; Mardia 1978 goodness-of-fit ratios).
Reference software
not recorded
Reference script
generators/analytic/reviewed_reference_corrections.py · sha256 49e3645836bb3ac3
Dataset
reference_data/datasets/MV_MDS_SITES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/dimensionsexact220.00E+2passed
/gof_absolutetolerance0.815488679709379110.81548867970937943.28E-1615passed
/gof_positivetolerance11.00.00E+2passed
/eigenvalues/0/eigenvaluetolerance17.41392210806465417.4139221080646822.80E-1414passed
/eigenvalues/0/retainedexactTrueTruepassed
/eigenvalues/1/eigenvaluetolerance7.65905420432692277.6590542043269472.40E-1414passed
/eigenvalues/1/retainedexactTrueTruepassed
/eigenvalues/2/eigenvaluetolerance0-1.4140410896003506e-151.41E-15passed
/eigenvalues/2/retainedexactFalseFalsepassed
/eigenvalues/3/eigenvaluetolerance-1.5764533572795938-1.57645335727959251.34E-1515passed
/eigenvalues/3/retainedexactFalseFalsepassed
/eigenvalues/4/eigenvaluetolerance-4.0965229551120341-4.0965229551120361.80E-1515passed
/eigenvalues/4/retainedexactFalseFalsepassed
/objects/0/coordinatessequence['-2.3828463202857404', '0.4986714014739243'][-2.382846320285741, 0.498671401473925]passed
/objects/1/coordinatessequence['-1.5513796507817434', '-1.3059822417755038'][-1.5513796507817463, -1.3059822417755047]passed
/objects/2/coordinatessequence['-0.18359339527408747', '1.7264422699623765'][-0.18359339527408805, 1.7264422699623794]passed
/objects/3/coordinatessequence['1.4196671629139705', '-1.5323628107059395'][1.4196671629139717, -1.5323628107059435]passed
/objects/4/coordinatessequence['2.6981522034275995', '0.6132313810451453'][2.6981522034276026, 0.6132313810451464]passed
/objects/0/labelexactnorthnorthpassed
/objects/4/labelexactcentrecentrepassed

Convention

Reference code

Whole script generators/analytic/reviewed_reference_corrections.py


Ordinal SMACOF on an exactly Euclidean matrix, secondary ties

Case STAT-MV-MDS-ORDINAL-001 · status: passed

Reference

Evidence type
R
Source
R stats::cmdscale and analytic perfect-fit stress
Citation
R Core Team, stats package help: cmdscale (classical multidimensional scaling, Torgerson 1958; Mardia 1978 goodness-of-fit ratios).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_mds.R · sha256 92a560135cfde9af
Dataset
reference_data/datasets/MV_MDS_PLANE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/dimensionsexact220.00E+2passed
/convergedexactTrueTruepassed
/iterationsexact110.00E+2passed
/raw_stresstolerance03.944304526105059e-313.94E-31passed
/normalized_stresstolerance03.975119405215255e-323.98E-32passed
/smacof_stresstolerance01.993770148541515e-161.99E-16passed
/stress_1tolerance01.986027322597818e-161.99E-16passed
/starts/0/startexactclassicalclassicalpassed
/starts/0/normalized_stresstolerance03.975119405215255e-323.98E-32passed
/eigenvalues/0/eigenvaluetolerance17.20856444288742917.2085644428874367.03E-1515passed
/eigenvalues/1/eigenvaluetolerance6.69143555711258396.69143555711257051.34E-1414passed
/objects/0/coordinatessequence['-0.39326786792865198', '-0.4163818642282579'][-0.39326786792865204, -0.4163818642282582]passed
/objects/1/coordinatessequence['0.75924710724200373', '-0.23878156020080102'][0.7592471072420044, -0.2387815602008013]passed
/objects/2/coordinatessequence['0.449448991918459', '0.47586498768943902'][0.44944899191845955, 0.475864987689438]passed
/objects/3/coordinatessequence['-0.70306598325219694', '0.2982646836619815'][-0.7030659832521968, 0.298264683661981]passed
/objects/4/coordinatessequence['-0.1123622479796148', '-0.1189662469223597'][-0.11236224797961505, -0.11896624692235955]passed

Convention

Reference code

Whole script generators/r/multivariate_mds.R


Ordinal SMACOF on an exactly Euclidean matrix, primary ties

Case STAT-MV-MDS-ORDINAL-002 · status: passed

Reference

Evidence type
R
Source
R stats::cmdscale and analytic perfect-fit stress
Citation
R Core Team, stats package help: cmdscale (classical multidimensional scaling, Torgerson 1958; Mardia 1978 goodness-of-fit ratios).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_mds.R · sha256 92a560135cfde9af
Dataset
reference_data/datasets/MV_MDS_PLANE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/dimensionsexact220.00E+2passed
/convergedexactTrueTruepassed
/iterationsexact110.00E+2passed
/raw_stresstolerance05.577493118945435e-315.58E-31passed
/normalized_stresstolerance05.669937756276022e-325.67E-32passed
/smacof_stresstolerance02.3811631099687444e-162.38E-16passed
/stress_1tolerance02.361671678905735e-162.36E-16passed
/starts/0/startexactclassicalclassicalpassed
/starts/0/normalized_stresstolerance05.669937756276022e-325.67E-32passed
/eigenvalues/0/eigenvaluetolerance17.20856444288742917.2085644428874367.03E-1515passed
/eigenvalues/1/eigenvaluetolerance6.69143555711258396.69143555711257051.34E-1414passed
/objects/0/coordinatessequence['-0.39326786792865198', '-0.4163818642282579'][-0.39326786792865204, -0.4163818642282582]passed
/objects/1/coordinatessequence['0.75924710724200373', '-0.23878156020080102'][0.7592471072420044, -0.2387815602008013]passed
/objects/2/coordinatessequence['0.449448991918459', '0.47586498768943902'][0.44944899191845955, 0.475864987689438]passed
/objects/3/coordinatessequence['-0.70306598325219694', '0.2982646836619815'][-0.7030659832521968, 0.298264683661981]passed
/objects/4/coordinatessequence['-0.1123622479796148', '-0.1189662469223597'][-0.11236224797961505, -0.11896624692235955]passed

Convention

Reference code

Whole script generators/r/multivariate_mds.R


Classical scaling of standardized Euclidean distances between rows

Case STAT-MV-MDS-VARIABLES-001 · status: passed

Reference

Evidence type
R
Source
Retained R reference with independently reviewed analytical corrections
Citation
R Core Team, stats package help: cmdscale (classical multidimensional scaling, Torgerson 1958; Mardia 1978 goodness-of-fit ratios).
Reference software
not recorded
Reference script
generators/analytic/reviewed_reference_corrections.py · sha256 49e3645836bb3ac3
Dataset
reference_data/datasets/MV_MDS_VARS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/dimensionsexact220.00E+2passed
/gof_absolutetolerance0.635436312430741390.6354363124307425.55E-1615passed
/gof_positivetolerance0.63543631243074150.6354363124307424.45E-1615passed
/eigenvalues/0/eigenvaluetolerance15.46842105021515215.4684210502151913.93E-1414passed
/eigenvalues/0/retainedexactTrueTruepassed
/eigenvalues/1/eigenvaluetolerance9.94903144701447499.9490314470144848.93E-1515passed
/eigenvalues/1/retainedexactTrueTruepassed
/eigenvalues/2/eigenvaluetolerance9.20687154540668479.2068715454066831.74E-1515passed
/eigenvalues/2/retainedexactFalseFalsepassed
/eigenvalues/3/eigenvaluetolerance5.37567595736362685.3756759573636325.33E-1515passed
/eigenvalues/3/retainedexactFalseFalsepassed
/eigenvalues/4/eigenvaluetolerance02.421380353515736e-152.42E-15passed
/eigenvalues/4/retainedexactFalseFalsepassed
/eigenvalues/5/eigenvaluetolerance01.60321581260916e-151.60E-15passed
/eigenvalues/5/retainedexactFalseFalsepassed
/eigenvalues/6/eigenvaluetolerance01.3310908222176163e-151.33E-15passed
/eigenvalues/6/retainedexactFalseFalsepassed
/eigenvalues/7/eigenvaluetolerance02.0144967829558408e-162.01E-16passed
/eigenvalues/7/retainedexactFalseFalsepassed
/eigenvalues/8/eigenvaluetolerance0-2.955351189134437e-162.96E-16passed
/eigenvalues/8/retainedexactFalseFalsepassed
/eigenvalues/9/eigenvaluetolerance0-1.344988412040238e-151.34E-15passed
/eigenvalues/9/retainedexactFalseFalsepassed
/eigenvalues/10/eigenvaluetolerance0-2.7543484068296546e-152.75E-15passed
/eigenvalues/10/retainedexactFalseFalsepassed
/objects/0/coordinatessequence['-0.61929236698565027', '1.2338419410113577'][-0.6192923669856465, 1.2338419410113661]passed
/objects/1/coordinatessequence['-1.0248572828376896', '-0.57511246726732757'][-1.0248572828376927, -0.5751124672673401]passed
/objects/2/coordinatessequence['0.057899676400370775', '-0.21087958035127785'][0.05789967640037349, -0.21087958035126017]passed
/objects/3/coordinatessequence['-1.3311317304529344', '0.17563355255255569'][-1.3311317304529375, 0.1756335525525604]passed
/objects/4/coordinatessequence['1.4225792689703474', '1.6989763899421912'][1.422579268970352, 1.6989763899421808]passed
/objects/5/coordinatessequence['1.0603123441629405', '0.69753661588814486'][1.0603123441629403, 0.6975366158881289]passed
/objects/6/coordinatessequence['-1.3112234241313061', '0.78775682605887754'][-1.3112234241313039, 0.7877568260588819]passed
/objects/7/coordinatessequence['-0.021012150067988383', '-1.2188486026714596'][-0.021012150067989833, -1.2188486026714607]passed
/objects/8/coordinatessequence['2.4840933496214963', '-1.0981951962197751'][2.484093349621497, -1.098195196219777]passed
/objects/9/coordinatessequence['0.33540949381985807', '-0.41052424669205795'][0.33540949381986035, -0.4105242466920375]passed
/objects/10/coordinatessequence['-1.0527771784994453', '-1.080185232251228'][-1.0527771784994528, -1.0801852322512422]passed
/rows_usedexact11110.00E+2passed
/missing_countexact110.00E+2passed
/objects/0/labelexactu01u01passed
/objects/6/labelexactu08u08passed

Convention

Reference code

Whole script generators/analytic/reviewed_reference_corrections.py


Partial correlation one control

Case STAT-MV-PARTIAL-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm, stats::cor, stats::pt
Citation
R Core Team, stats package help: lm, residuals, cor, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_correlation_extensions.R · sha256 0d15a50f16440286
Dataset
reference_data/datasets/MV_CORREXT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact16160.00E+2passed
/incomplete_countexact000.00E+2passed
/control_rankexact110.00E+2passed
/dfexact13130.00E+2passed
/zero_order_rtolerance0.956828008006929690.95682800800692981.08E-1615passed
/partial_rtolerance0.379753836171511150.379753836171510268.88E-1614passed
/t_valuetolerance1.48009979095836061.48009979095835663.96E-1514passed
/p_valuetail_probability0.162669811121302740.16266981112130381.06E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

174  partial_case("STAT-MV-PARTIAL-001", "Partial correlation one control", "nominal", "c1")

Whole script generators/r/multivariate_correlation_extensions.R


Partial correlation two controls with a missing control value

Case STAT-MV-PARTIAL-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm, stats::cor, stats::pt
Citation
R Core Team, stats package help: lm, residuals, cor, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_correlation_extensions.R · sha256 0d15a50f16440286
Dataset
reference_data/datasets/MV_CORREXT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact15150.00E+2passed
/incomplete_countexact110.00E+2passed
/control_rankexact220.00E+2passed
/dfexact11110.00E+2passed
/zero_order_rtolerance0.959193232274462180.95919323227446211.09E-1615passed
/partial_rtolerance0.27709670103323830.277096701033236471.83E-1514passed
/t_valuetolerance0.956479580902620750.9564795809026146.77E-1514passed
/p_valuetail_probability0.359379179608914630.35937917960891783.17E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

175  partial_case("STAT-MV-PARTIAL-002", "Partial correlation two controls with a missing control value",
176               "missing_excluded", c("c1", "c3"),
177               "c3 is missing on one row, so n = 15 complete cases and incomplete_count = 1.")

Whole script generators/r/multivariate_correlation_extensions.R


Partial correlation redundant control

Case STAT-MV-PARTIAL-003 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm, stats::cor, stats::pt
Citation
R Core Team, stats package help: lm, residuals, cor, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_correlation_extensions.R · sha256 0d15a50f16440286
Dataset
reference_data/datasets/MV_CORREXT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact16160.00E+2passed
/incomplete_countexact000.00E+2passed
/control_rankexact110.00E+2passed
/dfexact13130.00E+2passed
/zero_order_rtolerance0.956828008006929690.95682800800692981.08E-1615passed
/partial_rtolerance0.379753836171511150.379753836171509541.61E-1514passed
/t_valuetolerance1.48009979095836061.48009979095835357.07E-1514passed
/p_valuetail_probability0.162669811121302740.162669811121304631.89E-1513passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

178  partial_case("STAT-MV-PARTIAL-003", "Partial correlation redundant control", "axis_coverage",
179               c("c1", "c2"),
180               "c2 = 2 c1 + 3 is collinear with c1 and the intercept: the controls have rank 1, df = n - 3 and the residuals equal those on c1 alone.")

Whole script generators/r/multivariate_correlation_extensions.R


Multivariate STAT-MV-PCA-001

Case STAT-MV-PCA-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_core.R · sha256 90e434c114270562
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/eigenvalues/0tolerance1.99745905548196091.99745905548196166.65E-1615passed
/meanssequence['5.0083333333333337', '4.0250000000000004', '5.691666666666[5.008333333333333, 4.025, 5.691666666666666]passed
/loadings/0sign_invariant_sequence['0.70672573267086569', '-0.0052863510808498241'][0.7067257326708654, -0.0052863510808498285]passed
/source_countexact12120.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

46  pc <- stats::prcomp(d[c("x","y","z")], center=TRUE, scale.=TRUE)
47  emit("STAT-MV-PCA-001", "tabelz.executor.principal_components", "principalComponentsMenuItem",
48       list(columns=list("x","y","z"), scaling="correlation", n_components=2L),
49       list("/eigenvalues/0"=tol(pc$sdev[1]^2, "sample-correlation first eigenvalue"),
50            "/means"=seqtol(pc$center, "column means"),
51            "/loadings/0"=seqtol(pc$rotation[1,1:2], "first variable loadings", TRUE),
52            "/source_count"=exact(nrow(d), "all source rows")),
53       c("prcomp uses sample SD (n-1); each component may flip sign."))

Whole script generators/r/multivariate_core.R


Multivariate option STAT-MV-PCA-COV-001

Case STAT-MV-PCA-COV-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic formulae
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_options.R · sha256 5d6896c30e5e2257
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/eigenvalues/0tolerance18.29015486765841918.290154867658421.36E-1617passed
/scales/0tolerance11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

22  cols <- list("x","y","z")
23  pc <- stats::prcomp(d[c("x","y","z")],center=TRUE,scale.=FALSE)
24  emit("STAT-MV-PCA-COV-001","tabelz.executor.principal_components","principalComponentsMenuItem",
25       list(columns=cols,scaling="covariance",n_components=2L),
26       list("/eigenvalues/0"=tol(pc$sdev[1]^2,"R prcomp sample-covariance first eigenvalue"),
27            "/scales/0"=tol(1,"Covariance PCA retains raw measurement scale")),
28       c("prcomp sample covariance divides by n-1 and does not standardize."))

Whole script generators/r/multivariate_options.R


Hard-numeric STAT-MV-PCA-HARD-001

Case STAT-MV-PCA-HARD-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::prcomp, factanal, hclust, kmeans, manova and chisq.test documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_hard.R · sha256 06eca51b7d1d3ab6
Dataset
reference_data/datasets/MV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/eigenvalues/0tolerance1.99745905564030961.99745905564030932.56E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

49  xyz <- list("x","y","z")
50  pc <- stats::prcomp(d[c("x","y","z")],center=TRUE,scale.=TRUE)
51  emit("STAT-MV-PCA-HARD-001","tabelz.executor.principal_components","principalComponentsMenuItem",
52       "MV_HARD",list(columns=xyz,scaling="correlation",n_components=2L),
53       list("/eigenvalues/0"=tol(pc$sdev[1]^2,"sample-correlation leading eigenvalue")),
54       c("Large common offset must cancel before computing covariance; prcomp uses n-1."))

Whole script generators/r/multivariate_hard.R


Invalid multivariate request STAT-MV-PCA-INVALID-001

Case STAT-MV-PCA-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-PCA-INVALID-002

Case STAT-MV-PCA-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Minimum-sample STAT-MV-PCA-MIN-001

Case STAT-MV-PCA-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, factanal, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_minimum.R · sha256 bcd4094568b73b37
Dataset
reference_data/datasets/MV_MIN_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/eigenvalues/0tolerance1.99999999999999961.99999999999999964.41E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

24  two <- core[c(1L,5L),]
25  export_dataset(two,"MV_MIN_TWO","Two distinct observations",
26                 "R closed-form two-row multivariate boundary",
27                 "Constructed validation data; no third-party license.",types)
28  d <- read_dataset("MV_MIN_TWO")
29  pc <- stats::prcomp(d[c("x","y")],center=TRUE,scale.=TRUE)
30  emit("STAT-MV-PCA-MIN-001","tabelz.executor.principal_components","principalComponentsMenuItem",
31       "MV_MIN_TWO",list(columns=list("x","y"),scaling="correlation",n_components=1L),
32       list("/eigenvalues/0"=tol(pc$sdev[1]^2,"two-point sample-correlation eigenvalue")),
33       c("Two complete rows are the minimum accepted for PCA."))

Whole script generators/r/multivariate_minimum.R


Missing and excluded multivariate population STAT-MV-PCA-POP-001

Case STAT-MV-PCA-POP-001 · status: passed

Reference

Evidence type
R
Source
R stats and analytic identities
Citation
R stats::prcomp/hclust/kmeans/manova/cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_population.R · sha256 8a6d148e9316e21f
Dataset
reference_data/datasets/MV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact12120.00E+2passed
/excluded_countexact110.00E+2passed
/missing_countexact110.00E+2passed
/eigenvalues/0tolerance1.98115518675034191.98115518675034162.65E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

33  eligible <- d[-10L,,drop=FALSE]
34  xyzd <- eligible[stats::complete.cases(eligible[c("x","y","z")]),,drop=FALSE]
35  pc <- stats::prcomp(xyzd[c("x","y","z")],center=TRUE,scale.=TRUE)
36  emit("STAT-MV-PCA-POP-001","tabelz.executor.principal_components","principalComponentsMenuItem",
37       list(columns=list("x","y","z"),n_components=2L),
38       list("/source_count"=exact(nrow(d),"CSV source population"),
39            "/excluded_count"=exact(1L,"One stable row ID excluded"),
40            "/missing_count"=exact(1L,"One selected-column missing row"),
41            "/eigenvalues/0"=tol(pc$sdev[1]^2,"R complete-case sample-correlation eigenvalue")),
42       c("Complete-case standardization divides by n-1."))

Whole script generators/r/multivariate_population.R


Retained scoring STAT-MV-PCA-SCORE-001

Case STAT-MV-PCA-SCORE-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic
Citation
R stats::prcomp, kmeans, hclust and matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring.R · sha256 6aa292c3e1b1894b
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/pc1tolerance-1.647388504772167-1.6473885047721689.16E-1615passed
/1/pc1tolerance-1.6803191885164597-1.68031918851646014.16E-1615passed
/2/pc1absentNonepassed
/3/pc1tolerance-1.5609201506993629-1.56092015069936356.18E-1615passed
/4/pc1tolerance0.0328840355261049710.032884035526104871.04E-1614passed
/5/pc1tolerance0.0231156487555106150.0231156487555106087.01E-1815passed
/6/pc1tolerance-0.047544843428177191-0.0475448434281771856.49E-1815passed
/7/pc1tolerance-0.026782108802030635-0.0267821088020307531.18E-1614passed
/8/pc1tolerance1.64435422870033631.64435422870033631.78E-1716passed
/9/pc1tolerance1.64803864091185771.64803864091185774.22E-1918passed
/10/pc1tolerance1.63216402986191511.63216402986191564.55E-1615passed
/11/pc1tolerance1.7101488470280951.7101488470280951.72E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

39  id <- "STAT-MV-PCA-SCORE-001"

Whole script generators/r/multivariate_scoring.R


Retained correction STAT-MV-PCA-SCORE-HARD-001

Case STAT-MV-PCA-SCORE-HARD-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and analytic row-identity join
Citation
R stats::prcomp, scale and base matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_corrections.R · sha256 d22bd7a20d0e25a3
Dataset
reference_data/datasets/MV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/pc1tolerance-1.4678119640661258-1.46781196406612583.92E-1817passed
/1/pc1tolerance-1.3949013476239474-1.39490134762394744.81E-1716passed
/2/pc1tolerance-1.5722189126001063-1.57221891260010633.35E-1716passed
/3/pc1tolerance-1.3841121615967962-1.38411216159679622.45E-1716passed
/4/pc1tolerance0.946079452675325670.94607945267532571.68E-1817passed
/5/pc1tolerance0.913711897809299510.91371189780929951.41E-1817passed
/6/pc1tolerance0.998282927746172890.99828292774617292.15E-1817passed
/7/pc1tolerance0.862379651813709680.86237965181370974.16E-2019passed
/8/pc1tolerance0.524430008734014570.52443000873401464.40E-1817passed
/9/pc1tolerance0.525301236201558110.52530123620155811.74E-1817passed
/10/pc1tolerance0.513640827400182240.51364082740018222.71E-1817passed
/11/pc1tolerance0.535219193153452230.53521919315345222.89E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

102  as_serialized <- function(x) array(as.double(sprintf("%.15g", x)), dim = dim(as.array(x)))
103  id <- "STAT-MV-PCA-SCORE-HARD-001"

Whole script generators/r/multivariate_scoring_corrections.R


Retained scoring STAT-MV-PCA-SCORE-HARD-VALID-001

Case STAT-MV-PCA-SCORE-HARD-VALID-001 · status: passed

Reference

Evidence type
R
Source
R stats and centered matrix arithmetic
Citation
R stats::prcomp and scale documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_final.R · sha256 90940efd66e73ecd
Dataset
reference_data/datasets/MV_SCORE_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/pc1tolerance-1.4678120318312429-1.46781203176193796.93E-1110passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

51  x <- as.matrix(four[c("x","y","z")]); mu <- colMeans(x); sd <- apply(x,2,stats::sd)
52  scores <- as.vector(scale(as.matrix(five[c("x","y","z")]),center=mu,scale=sd)%*%weights)
53  model <- list(factor_analysis=list(columns=list("x","y","z"),means=nums(mu),
54    scales=nums(sd),weights=rows(weights),row_ids=ids(id,4L)))
55  out <- list("/0/factor1"=tol(scores[1],"First retained FA score"),
56              "/3/factor1"=tol(scores[4],"Fourth retained FA score"),
57              "/4/factor1"=absent("Fifth source row is outside retained model population"))
58  emit(id,"operation:tabelz.factor_scores","factorAnalysisMenuItem","MV_SCORE_FIVE",
59       list(names=list("factor1"),model=model),out,"boundary",
60       c("Four retained rows satisfy the model schema; fifth source row must remain missing."))
61  id <- "STAT-MV-PCA-SCORE-HARD-VALID-001"

Whole script generators/r/multivariate_scoring_final.R


Invalid multivariate request STAT-MV-PCA-SCORE-INVALID-001

Case STAT-MV-PCA-SCORE-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid multivariate request STAT-MV-PCA-SCORE-INVALID-002

Case STAT-MV-PCA-SCORE-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict configuration contract
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/MV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Retained boundary STAT-MV-PCA-SCORE-MIN-001

Case STAT-MV-PCA-SCORE-MIN-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and retained identity join
Citation
R stats::prcomp and base matrix arithmetic documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_boundaries.R · sha256 4cdf60725804542b
Dataset
reference_data/datasets/MV_MIN_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/pc1tolerance-0.99999999999999989-0.99999999999999935.56E-1615passed
/1/pc1tolerance0.999999999999999890.99999999999999935.56E-1615passed

Convention

Reference code

Whole script generators/r/multivariate_scoring_boundaries.R


Retained correction STAT-MV-PCA-SCORE-OMIT-CORRECTED-001

Case STAT-MV-PCA-SCORE-OMIT-CORRECTED-001 · status: passed

Reference

Evidence type
R
Source
R matrix arithmetic and analytic row-identity join
Citation
R stats::prcomp, scale and base matrix multiplication documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_scoring_corrections.R · sha256 d22bd7a20d0e25a3
Dataset
reference_data/datasets/MV_SCORE_FOUR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/pc1tolerance-1.2302436583607195-1.23024365836071771.80E-1514passed
/1/pc1tolerance0.78494473606746240.78494473606746262.23E-1615passed
/2/pc1tolerance0.445298922293257020.445298922293257747.17E-1614passed
/3/pc1absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

57  d <- four
58  ret <- three
59  id <- "STAT-MV-PCA-SCORE-OMIT-CORRECTED-001"

Whole script generators/r/multivariate_scoring_corrections.R


Special boundary STAT-MV-PCA-TIES-001

Case STAT-MV-PCA-TIES-001 · status: passed

Reference

Evidence type
R
Source
R stats and closed-form formulae
Citation
R stats::prcomp, hclust, kmeans, manova, chisq.test and cor documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/multivariate_boundaries.R · sha256 420706593513ebf0
Dataset
reference_data/datasets/MV_TIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/eigenvalues/1tolerance0.656038981934367160.65603898193436722.43E-1918passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

23  d <- base[c(1L,1L,5L,9L),]
24  d$i2 <- c(2L,2L,2L,2L)
25  export_dataset(d,"MV_TIES","Duplicate measurements and a constant item",
26                 "R multivariate tie and zero-item-variance references",
27                 "Constructed validation data; no third-party license.",types)
28  d <- read_dataset("MV_TIES")
29  pc <- stats::prcomp(d[c("x","y")],center=TRUE,scale.=TRUE)
30  emit("STAT-MV-PCA-TIES-001","tabelz.executor.principal_components","principalComponentsMenuItem",
31       "MV_TIES",list(columns=list("x","y"),scaling="correlation",n_components=2L),
32       list("/eigenvalues/1"=tol(pc$sdev[2]^2,"second eigenvalue with one repeated row")),
33       c("One exact duplicate row tests covariance computation without assuming a sign for loadings."))

Whole script generators/r/multivariate_boundaries.R


NIST AtmWtAg: fit

Case STAT-NIST-ATMWTAG-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/AtmWtAg.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_ATMWTAG.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre1.59467335677930E+0115.9467335677929752.46E-1414passed
/responses/0/anova/[source=Model]/dftolerance11.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance4646.00.00E+215passed
/responses/0/observationsexact48480.00E+2passed
/responses/0/anova/[source=Model]/sslre3.63834187500000E-093.638341875000001e-091.08E-2415passed
/responses/0/anova/[source=Error]/sslre1.04951729166667E-081.0495172916666666e-083.35E-2314passed

Convention


NIST AtmWtAg: general linear

Case STAT-NIST-ATMWTAG-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/AtmWtAg.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_ATMWTAG.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre1.59467335677930E+0115.9467335677929653.53E-1414passed
/term_tests/[term=group]/numerator_dfexact110.00E+2passed
/residual_dfexact46460.00E+2passed
/observationsexact48480.00E+2passed
/term_tests/[term=group]/sum_sqlre3.63834187500000E-093.6383418749999986e-091.40E-2415passed
/term_tests/[term=group]/mean_sqlre3.63834187500000E-093.6383418749999986e-091.40E-2415passed
/residual_variancelre2.28155932971014E-102.281559329710145e-104.94E-2514passed

Convention


NIST AtmWtAg: linear effects

Case STAT-NIST-ATMWTAG-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/AtmWtAg.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_ATMWTAG.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre1.59467335677930E+0115.9467335677929722.82E-1414passed
/term_tests/[term=group]/numerator_dfexact110.00E+2passed
/residual_dfexact46460.00E+2passed
/observationsexact48480.00E+2passed
/term_tests/[term=group]/sum_sqlre3.63834187500000E-093.6383418750000003e-092.57E-2516passed
/term_tests/[term=group]/mean_sqlre3.63834187500000E-093.6383418750000003e-092.57E-2516passed
/residual_variancelre2.28155932971014E-102.281559329710145e-104.94E-2514passed

Convention


NIST Bennett5: nonlinear regression Start 1

Case STAT-NIST-BENNETT5-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Bennett5.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_BENNETT5.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact1511510.00E+2passed
/sselre5.2404744073E-040.00052404744072656713.43E-1511passed
/artifact/residual_variancelre0.000003470512852658977507843.470512852493822e-061.65E-1610passed
/artifact/estimates/0lre-2.5235058043E+03-2523.50580430455064.55E-911passed
/artifact/covariance/0/0lre88299.164970649901892188299.164970052185.98E-711passed
/artifact/estimates/1lre4.6736564644E+0146.736564643575724.24E-1011passed
/artifact/covariance/1/1lre1.549744104871088847361.54974410474982931.21E-1010passed
/artifact/estimates/2lre9.3218483193E-010.93218483193294512.95E-1211passed
/artifact/covariance/2/2lre0.0004109661220712591868840.00041096612207948978.23E-1510passed

Notes

Convention


NIST Bennett5: nonlinear regression Start 2

Case STAT-NIST-BENNETT5-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Bennett5.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_BENNETT5.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact1511510.00E+2passed
/sselre5.2404744073E-040.00052404744072624983.75E-1511passed
/artifact/residual_variancelre0.000003470512852658977507843.4705128524917205e-061.67E-1610passed
/artifact/estimates/0lre-2.5235058043E+03-2523.50580430465934.66E-911passed
/artifact/covariance/0/0lre88299.164970649901892188299.164968959361.69E-610passed
/artifact/estimates/1lre4.6736564644E+0146.7365646435761954.24E-1011passed
/artifact/covariance/1/1lre1.549744104871088847361.54974410473041321.41E-1010passed
/artifact/estimates/2lre9.3218483193E-010.93218483193293782.94E-1211passed
/artifact/covariance/2/2lre0.0004109661220712591868840.00041096612207437233.11E-1511passed

Notes

Convention


NIST BoxBOD: nonlinear regression Start 1

Case STAT-NIST-BOXBOD-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/BoxBOD.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_BOXBOD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact440.00E+2passed
/sselre1.1680088766E+031168.00887655555184.44E-810passed
/artifact/residual_variancelre292.002219133693090929292.002219138887945.19E-910passed
/artifact/estimates/0lre2.1380940889E+02213.80940889039793.98E-1011passed
/artifact/covariance/0/0lre152.634045234014310976152.634045245775641.18E-810passed
/artifact/estimates/1lre5.4723748542E-010.54723748541919948.01E-1311passed
/artifact/covariance/1/1lre0.01093277945721897381690.0109327794572342441.53E-1411passed

Convention


NIST BoxBOD: nonlinear regression Start 2

Case STAT-NIST-BOXBOD-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/BoxBOD.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_BOXBOD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact440.00E+2passed
/sselre1.1680088766E+031168.00887655555184.44E-810passed
/artifact/residual_variancelre292.002219133693090929292.002219138887945.19E-910passed
/artifact/estimates/0lre2.1380940889E+02213.80940889039793.98E-1011passed
/artifact/covariance/0/0lre152.634045234014310976152.634045245775641.18E-810passed
/artifact/estimates/1lre5.4723748542E-010.54723748541919948.01E-1311passed
/artifact/covariance/1/1lre0.01093277945721897381690.0109327794572342441.53E-1411passed

Notes

Convention


NIST Chwirut1: nonlinear regression Start 1

Case STAT-NIST-CHWIRUT1-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Chwirut1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_CHWIRUT1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2112110.00E+2passed
/sselre2.3844771393E+032384.47713930935069.35E-911passed
/artifact/residual_variancelre11.3008395230654254240011.3008395227931312.72E-1010passed
/artifact/estimates/0lre1.9027818370E-010.190278183696138643.86E-1210passed
/artifact/covariance/0/0lre0.0004813002847779479912250.00048130028476652731.14E-1410passed
/artifact/estimates/1lre6.1314004477E-030.0061314004477082678.27E-1511passed
/artifact/covariance/1/1lre1.190251728519627552601E-71.1902517284875868e-073.20E-1810passed
/artifact/estimates/2lre1.0530908399E-020.0105309083989447065.53E-1411passed
/artifact/covariance/2/2lre6.285611382337052671504E-76.285611382315404e-072.16E-1811passed

Notes

Convention


NIST Chwirut1: nonlinear regression Start 2

Case STAT-NIST-CHWIRUT1-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Chwirut1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_CHWIRUT1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2112110.00E+2passed
/sselre2.3844771393E+032384.47713930934929.35E-911passed
/artifact/residual_variancelre11.3008395230654254240011.3008395227931242.72E-1010passed
/artifact/estimates/0lre1.9027818370E-010.190278183696358133.64E-1210passed
/artifact/covariance/0/0lre0.0004813002847779479912250.00048130028476703351.09E-1410passed
/artifact/estimates/1lre6.1314004477E-030.0061314004477110271.10E-1411passed
/artifact/covariance/1/1lre1.190251728519627552601E-71.1902517284871835e-073.24E-1810passed
/artifact/estimates/2lre1.0530908399E-020.0105309083989373446.27E-1411passed
/artifact/covariance/2/2lre6.285611382337052671504E-76.2856113823153e-072.18E-1811passed

Notes

Convention


NIST Chwirut2: nonlinear regression Start 1

Case STAT-NIST-CHWIRUT2-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Chwirut2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_CHWIRUT2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact51510.00E+2passed
/sselre5.1304802941E+02513.04802940686533.13E-911passed
/artifact/residual_variancelre10.0597652827705964160010.0597652824875562.83E-1010passed
/artifact/estimates/0lre1.6657666537E-010.166576665368040171.96E-1210passed
/artifact/covariance/0/0lre0.0014671417804491199761000.0014671417804745182.54E-1410passed
/artifact/estimates/1lre5.1653291286E-030.0051653291285725832.74E-1411passed
/artifact/covariance/1/1lre4.438438269564669475876E-74.4384382696228935e-075.82E-1810passed
/artifact/estimates/2lre1.2150007096E-020.012150007095829641.70E-1310passed
/artifact/covariance/2/2lre0.000002342196018034515442892.3421960179905598e-064.40E-1710passed

Notes

Convention


NIST Chwirut2: nonlinear regression Start 2

Case STAT-NIST-CHWIRUT2-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Chwirut2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_CHWIRUT2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact51510.00E+2passed
/sselre5.1304802941E+02513.04802940686573.13E-911passed
/artifact/residual_variancelre10.0597652827705964160010.0597652824875632.83E-1010passed
/artifact/estimates/0lre1.6657666537E-010.166576665368039561.96E-1210passed
/artifact/covariance/0/0lre0.0014671417804491199761000.00146714178047451222.54E-1410passed
/artifact/estimates/1lre5.1653291286E-030.0051653291285725752.74E-1411passed
/artifact/covariance/1/1lre4.438438269564669475876E-74.438438269622887e-075.82E-1810passed
/artifact/estimates/2lre1.2150007096E-020.012150007095829661.70E-1310passed
/artifact/covariance/2/2lre0.000002342196018034515442892.342196017990556e-064.40E-1710passed

Notes

Convention


NIST DanWood: nonlinear regression Start 1

Case STAT-NIST-DANWOOD-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/DanWood.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_DANWOOD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact440.00E+2passed
/sselre4.3173084083E-030.0043173084082910948.91E-1511passed
/artifact/residual_variancelre0.0010793271020595388935210.00107932710207277361.32E-1410passed
/artifact/estimates/0lre7.6886226176E-010.76886226176498264.98E-1211passed
/artifact/covariance/0/0lre0.0003342305682177232996000.00033423056822290045.18E-1510passed
/artifact/estimates/1lre3.8604055871E+003.86040558707605762.39E-1111passed
/artifact/covariance/1/1lre0.0026756422765448906935690.0026756422765311931.37E-1411passed

Notes

Convention


NIST DanWood: nonlinear regression Start 2

Case STAT-NIST-DANWOOD-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/DanWood.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_DANWOOD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact440.00E+2passed
/sselre4.3173084083E-030.0043173084082910838.92E-1511passed
/artifact/residual_variancelre0.0010793271020595388935210.00107932710207277071.32E-1410passed
/artifact/estimates/0lre7.6886226176E-010.76886226176498254.98E-1211passed
/artifact/covariance/0/0lre0.0003342305682177232996000.00033423056822289935.18E-1510passed
/artifact/estimates/1lre3.8604055871E+003.8604055870760582.39E-1111passed
/artifact/covariance/1/1lre0.0026756422765448906935690.0026756422765311861.37E-1411passed

Convention


NIST Eckerle4: nonlinear regression Start 1

Case STAT-NIST-ECKERLE4-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Eckerle4.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_ECKERLE4.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact32320.00E+2passed
/sselre1.4635887487E-030.00146358874872747162.75E-1410passed
/artifact/residual_variancelre0.000045737148397305702298094.5737148397733486e-054.28E-1611passed
/artifact/estimates/0lre1.5543827178E+001.55438271775261354.74E-1110passed
/artifact/covariance/0/0lre0.0002374080406416256525690.000237408040656850781.52E-1410passed
/artifact/estimates/1lre4.0888321754E+004.0888321754230072.30E-1111passed
/artifact/covariance/1/1lre0.0021905227516057486870090.00219052275163537452.96E-1410passed
/artifact/estimates/2lre4.5154121844E+02451.541218436144963.86E-911passed
/artifact/covariance/2/2lre0.0021902885614467700418560.0021902885614885654.18E-1410passed

Convention


NIST Eckerle4: nonlinear regression Start 2

Case STAT-NIST-ECKERLE4-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Eckerle4.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_ECKERLE4.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact32320.00E+2passed
/sselre1.4635887487E-030.00146358874872747112.75E-1410passed
/artifact/residual_variancelre0.000045737148397305702298094.573714839773347e-054.28E-1611passed
/artifact/estimates/0lre1.5543827178E+001.55438271775261374.74E-1110passed
/artifact/covariance/0/0lre0.0002374080406416256525690.0002374080406568511.52E-1410passed
/artifact/estimates/1lre4.0888321754E+004.0888321754230082.30E-1111passed
/artifact/covariance/1/1lre0.0021905227516057486870090.00219052275163537752.96E-1410passed
/artifact/estimates/2lre4.5154121844E+02451.541218436144963.86E-911passed
/artifact/covariance/2/2lre0.0021902885614467700418560.00219028856148856474.18E-1410passed

Notes

Convention


NIST ENSO: nonlinear regression Start 1

Case STAT-NIST-ENSO-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/ENSO.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_ENSO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact1591590.00E+2passed
/sselre7.8853978668E+02788.5397866829122.91E-911passed
/artifact/residual_variancelre4.959369727574956144094.9593697275654849.47E-1211passed
/artifact/estimates/0lre1.0510749193E+0110.5107491927713532.29E-1010passed
/artifact/covariance/0/0lre0.03058592610587933060890.0305859261061214272.42E-1311passed
/artifact/estimates/1lre3.0762128085E+003.07621280853853.85E-1110passed
/artifact/covariance/1/1lre0.05909786350008984753210.059097863500447753.58E-1311passed
/artifact/estimates/2lre5.3280138227E-010.53280138226595544.04E-1211passed
/artifact/covariance/2/2lre0.05931507602609882779240.0593150760263828352.84E-1311passed
/artifact/estimates/3lre4.4311088700E+0144.311088700220462.20E-1011passed
/artifact/covariance/3/3lre0.89128753686850907525760.89128753687674118.23E-1211passed
/artifact/estimates/4lre-1.6231428586E+00-1.62314285862666452.67E-1110passed
/artifact/covariance/4/4lre0.07883948400119282913210.078839483998904662.29E-1210passed
/artifact/estimates/5lre5.2554493756E-010.52554493755817981.82E-1211passed
/artifact/covariance/5/5lre0.23110807392789418521610.23110807392515852.74E-1210passed
/artifact/estimates/6lre2.6887614440E+0126.8876144400245922.46E-1112passed
/artifact/covariance/6/6lre0.17316367030370851569000.173163670299549324.16E-1210passed
/artifact/estimates/7lre2.1232288488E-010.212322884881517321.52E-1211passed
/artifact/covariance/7/7lre0.26481339580006449139210.264813395801250241.19E-1211passed
/artifact/estimates/8lre1.4966870418E+001.49668704183160473.16E-1110passed
/artifact/covariance/8/8lre0.06469122078689846454490.064691220787901321.00E-1210passed

Convention


NIST ENSO: nonlinear regression Start 2

Case STAT-NIST-ENSO-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/ENSO.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_ENSO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact1591590.00E+2passed
/sselre7.8853978668E+02788.5397866829122.91E-911passed
/artifact/residual_variancelre4.959369727574956144094.9593697275654849.47E-1211passed
/artifact/estimates/0lre1.0510749193E+0110.5107491927713532.29E-1010passed
/artifact/covariance/0/0lre0.03058592610587933060890.0305859261061214482.42E-1311passed
/artifact/estimates/1lre3.0762128085E+003.07621280853853.85E-1110passed
/artifact/covariance/1/1lre0.05909786350008984753210.0590978635004477053.58E-1311passed
/artifact/estimates/2lre5.3280138227E-010.53280138226595544.04E-1211passed
/artifact/covariance/2/2lre0.05931507602609882779240.059315076026382842.84E-1311passed
/artifact/estimates/3lre4.4311088700E+0144.311088700220462.20E-1011passed
/artifact/covariance/3/3lre0.89128753686850907525760.89128753687673938.23E-1211passed
/artifact/estimates/4lre-1.6231428586E+00-1.62314285862666472.67E-1110passed
/artifact/covariance/4/4lre0.07883948400119282913210.078839483998904672.29E-1210passed
/artifact/estimates/5lre5.2554493756E-010.52554493755817971.82E-1211passed
/artifact/covariance/5/5lre0.23110807392789418521610.231108073925158472.74E-1210passed
/artifact/estimates/6lre2.6887614440E+0126.8876144400245922.46E-1112passed
/artifact/covariance/6/6lre0.17316367030370851569000.173163670299549074.16E-1210passed
/artifact/estimates/7lre2.1232288488E-010.21232288488151721.52E-1211passed
/artifact/covariance/7/7lre0.26481339580006449139210.264813395801250241.19E-1211passed
/artifact/estimates/8lre1.4966870418E+001.4966870418316053.16E-1110passed
/artifact/covariance/8/8lre0.06469122078689846454490.064691220787901341.00E-1210passed

Convention


NIST Filip: fit

Case STAT-NIST-FILIP-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Filip.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_FILIP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/observationsexact82820.00E+2passed
/responses/0/parameters/[term=Intercept]/estimatelre-1467.48961422980-1467.48961422977822.18E-1113passed
/responses/0/parameters/[term=Intercept]/std_errorlre298.084530995537298.08453099553363.39E-1213passed
/responses/0/parameters/[term=x]/estimatelre-2772.17959193342-2772.17959193339272.73E-1114passed
/responses/0/parameters/[term=x]/std_errorlre559.779865474950559.77986547494386.21E-1213passed
/responses/0/parameters/[term=x^2]/estimatelre-2316.37108160893-2316.3710816089052.49E-1113passed
/responses/0/parameters/[term=x^2]/std_errorlre466.477572127796466.47757212779134.71E-1213passed
/responses/0/parameters/[term=x^3]/estimatelre-1127.97394098372-1127.97394098370341.66E-1113passed
/responses/0/parameters/[term=x^3]/std_errorlre227.204274477751227.20427447774882.19E-1214passed
/responses/0/parameters/[term=x^4]/estimatelre-354.478233703349-354.47823370334494.07E-1213passed
/responses/0/parameters/[term=x^4]/std_errorlre71.647866087592771.647866087591947.64E-1313passed
/responses/0/parameters/[term=x^5]/estimatelre-75.1242017393757-75.12420173937498.03E-1313passed
/responses/0/parameters/[term=x^5]/std_errorlre15.289717874740015.289717874739831.70E-1313passed
/responses/0/parameters/[term=x^6]/estimatelre-10.8753180355343-10.8753180355341341.66E-1313passed
/responses/0/parameters/[term=x^6]/std_errorlre2.236911598160332.23691159816030672.33E-1413passed
/responses/0/parameters/[term=x^7]/estimatelre-1.06221498588947-1.06221498588945631.37E-1413passed
/responses/0/parameters/[term=x^7]/std_errorlre0.2216243219342270.22162432193422482.21E-1514passed
/responses/0/parameters/[term=x^8]/estimatelre-0.670191154593408E-01-0.067019115459340136.74E-1613passed
/responses/0/parameters/[term=x^8]/std_errorlre0.142363763154724E-010.0142363763154722281.72E-1613passed
/responses/0/parameters/[term=x^9]/estimatelre-0.246781078275479E-02-0.00246781078275476032.97E-1713passed
/responses/0/parameters/[term=x^9]/std_errorlre0.535617408889821E-030.00053561740888981446.57E-1813passed
/responses/0/parameters/[term=x^10]/estimatelre-0.402962525080404E-04-4.0296252508039944e-054.56E-1913passed
/responses/0/parameters/[term=x^10]/std_errorlre0.896632837373868E-058.966328373738576e-061.04E-1913passed
/responses/0/anova/[source=Error]/sslre0.795851382172941E-030.00079585138217292831.27E-1713passed
/responses/0/anova/[source=Model]/dftolerance1010.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance7171.00.00E+215passed
/responses/0/anova/[source=Error]/mslre0.000011209174396801994569681680791.1209174396801807e-051.88E-1913passed
/responses/0/metrics/r_squaredlre0.9967274161856200.99672741618562022.19E-1615passed
/responses/0/anova/[source=Model]/f_statlre2162.439545114892162.4395451149223.21E-1113passed

Convention


NIST Gauss1: nonlinear regression Start 1

Case STAT-NIST-GAUSS1-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Gauss1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_GAUSS1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2422420.00E+2passed
/sselre1.3158222432E+031315.82224320337763.38E-911passed
/artifact/residual_variancelre5.437281996748728324005.43728199670817254.06E-1111passed
/artifact/estimates/0lre9.8778210871E+0198.778210871030723.07E-1112passed
/artifact/covariance/0/0lre0.33093917099352200529000.33093917099557232.05E-1211passed
/artifact/estimates/1lre1.0497276517E-020.0104972765171685471.69E-1310passed
/artifact/covariance/1/1lre1.30103429139334826289E-81.3010342914530905e-085.97E-1910passed
/artifact/estimates/2lre1.0048990633E+02100.489906329564934.35E-1011passed
/artifact/covariance/2/2lre0.34611778381336151655040.34611778381510481.74E-1211passed
/artifact/estimates/3lre6.7481111276E+0167.481111276319473.19E-1011passed
/artifact/covariance/3/3lre0.01094240145311778017440.0109424014532130659.53E-1411passed
/artifact/estimates/4lre2.3129773360E+0123.1297733596585543.41E-1010passed
/artifact/covariance/4/4lre0.03041518959748667133160.030415189598245427.59E-1310passed
/artifact/estimates/5lre7.1994503004E+0171.994503003733632.66E-1011passed
/artifact/covariance/5/5lre0.39216143174700698515690.392161431746864351.43E-1312passed
/artifact/estimates/6lre1.7899805021E+02178.998050207263872.74E-910passed
/artifact/covariance/6/6lre0.01546786759097968390890.0154678675910974171.18E-1311passed
/artifact/estimates/7lre1.8389389025E+0118.3893890254870344.87E-1010passed
/artifact/covariance/7/7lre0.04053905532168398602240.0405390553223917367.08E-1310passed

Notes

Convention


NIST Gauss1: nonlinear regression Start 2

Case STAT-NIST-GAUSS1-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Gauss1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_GAUSS1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2422420.00E+2passed
/sselre1.3158222432E+031315.82224320337793.38E-911passed
/artifact/residual_variancelre5.437281996748728324005.4372819967081734.06E-1111passed
/artifact/estimates/0lre9.8778210871E+0198.778210871030723.07E-1112passed
/artifact/covariance/0/0lre0.33093917099352200529000.3309391709955732.05E-1211passed
/artifact/estimates/1lre1.0497276517E-020.0104972765171685471.69E-1310passed
/artifact/covariance/1/1lre1.30103429139334826289E-81.3010342914530912e-085.97E-1910passed
/artifact/estimates/2lre1.0048990633E+02100.489906329564924.35E-1011passed
/artifact/covariance/2/2lre0.34611778381336151655040.346117783815105161.74E-1211passed
/artifact/estimates/3lre6.7481111276E+0167.481111276319473.19E-1011passed
/artifact/covariance/3/3lre0.01094240145311778017440.0109424014532130919.53E-1411passed
/artifact/estimates/4lre2.3129773360E+0123.1297733596585573.41E-1010passed
/artifact/covariance/4/4lre0.03041518959748667133160.0304151895982454387.59E-1310passed
/artifact/estimates/5lre7.1994503004E+0171.994503003733632.66E-1011passed
/artifact/covariance/5/5lre0.39216143174700698515690.39216143174686421.43E-1312passed
/artifact/estimates/6lre1.7899805021E+02178.998050207263872.74E-910passed
/artifact/covariance/6/6lre0.01546786759097968390890.015467867591097451.18E-1311passed
/artifact/estimates/7lre1.8389389025E+0118.3893890254870344.87E-1010passed
/artifact/covariance/7/7lre0.04053905532168398602240.040539055322391767.08E-1310passed

Convention


NIST Gauss2: nonlinear regression Start 1

Case STAT-NIST-GAUSS2-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Gauss2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_GAUSS2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2422420.00E+2passed
/sselre1.2475282092E+031247.52820923099833.10E-810passed
/artifact/residual_variancelre5.155075244543921715245.1550752447561912.12E-1010passed
/artifact/estimates/0lre9.9018328406E+0199.018328405936426.36E-1112passed
/artifact/covariance/0/0lre0.28889299410130874006410.28889299409827913.03E-1210passed
/artifact/estimates/1lre1.0994945399E-020.0109949453986392773.61E-1310passed
/artifact/covariance/1/1lre1.77830406543325378756E-81.7783040653063657e-081.27E-1810passed
/artifact/estimates/2lre1.0188022528E+02101.880225275250334.75E-910passed
/artifact/covariance/2/2lre0.35066904872407599559840.350669048728896654.82E-1210passed
/artifact/estimates/3lre1.0703095519E+02107.03095518596374.04E-910passed
/artifact/covariance/3/3lre0.02252039956971004358560.022520399570064013.54E-1310passed
/artifact/estimates/4lre2.3578584029E+0123.578584028568994.31E-1010passed
/artifact/covariance/4/4lre0.05150900354452347344890.051509003544085754.38E-1311passed
/artifact/estimates/5lre7.2045589471E+0172.045589470637783.62E-1011passed
/artifact/covariance/5/5lre0.38096010725856599014560.380960107257508341.06E-1211passed
/artifact/estimates/6lre1.5327010194E+02153.27010194127321.27E-911passed
/artifact/covariance/6/6lre0.03789514098985477842810.037895140988939759.15E-1310passed
/artifact/estimates/7lre1.9525972636E+0119.5259726363356173.36E-1010passed
/artifact/covariance/7/7lre0.06978340818328086684490.069783408184000567.20E-1310passed

Convention


NIST Gauss2: nonlinear regression Start 2

Case STAT-NIST-GAUSS2-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Gauss2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_GAUSS2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2422420.00E+2passed
/sselre1.2475282092E+031247.52820923099853.10E-810passed
/artifact/residual_variancelre5.155075244543921715245.1550752447561922.12E-1010passed
/artifact/estimates/0lre9.9018328406E+0199.018328405936426.36E-1112passed
/artifact/covariance/0/0lre0.28889299410130874006410.288892994098281953.03E-1210passed
/artifact/estimates/1lre1.0994945399E-020.010994945398639373.61E-1310passed
/artifact/covariance/1/1lre1.77830406543325378756E-81.7783040653064282e-081.27E-1810passed
/artifact/estimates/2lre1.0188022528E+02101.880225275249914.75E-910passed
/artifact/covariance/2/2lre0.35066904872407599559840.35066904872889274.82E-1210passed
/artifact/estimates/3lre1.0703095519E+02107.030955185964034.04E-910passed
/artifact/covariance/3/3lre0.02252039956971004358560.022520399570064273.54E-1310passed
/artifact/estimates/4lre2.3578584029E+0123.5785840285697684.30E-1010passed
/artifact/covariance/4/4lre0.05150900354452347344890.051509003544088274.35E-1311passed
/artifact/estimates/5lre7.2045589471E+0172.045589470637673.62E-1011passed
/artifact/covariance/5/5lre0.38096010725856599014560.380960107257518171.05E-1211passed
/artifact/estimates/6lre1.5327010194E+02153.270101941273771.27E-911passed
/artifact/covariance/6/6lre0.03789514098985477842810.037895140988937819.17E-1310passed
/artifact/estimates/7lre1.9525972636E+0119.5259726363350733.35E-1010passed
/artifact/covariance/7/7lre0.06978340818328086684490.06978340818399567.15E-1310passed

Convention


NIST Gauss3: nonlinear regression Start 1

Case STAT-NIST-GAUSS3-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Gauss3.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_GAUSS3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2422420.00E+2passed
/sselre1.2444846360E+031244.48463601315711.32E-810passed
/artifact/residual_variancelre5.142498496102756406255.1424984959221371.81E-1010passed
/artifact/estimates/0lre9.8940368970E+0198.940368969869741.30E-1011passed
/artifact/covariance/0/0lre0.28095504672635145658890.280955046723253673.10E-1210passed
/artifact/estimates/1lre1.0945879335E-020.0109458793348869191.13E-1310passed
/artifact/covariance/1/1lre1.57604395140858505921E-81.5760439513043066e-081.04E-1810passed
/artifact/estimates/2lre1.0069553078E+02100.695530778158021.84E-910passed
/artifact/covariance/2/2lre0.66026329824052452584890.66026329824624465.72E-1211passed
/artifact/estimates/3lre1.1163619459E+02111.636194587253242.75E-910passed
/artifact/covariance/3/3lre0.12473512178151200990490.124735121781210483.02E-1311passed
/artifact/estimates/4lre2.3300500029E+0123.3005000291696551.70E-1011passed
/artifact/covariance/4/4lre0.13384463488399890185290.133844634886090442.09E-1210passed
/artifact/estimates/5lre7.3705031418E+0173.705031417603883.96E-1011passed
/artifact/covariance/5/5lre1.461980625380842027241.46198062544066295.98E-1110passed
/artifact/estimates/6lre1.4776164251E+02147.761642505105474.89E-910passed
/artifact/covariance/6/6lre0.16392929910641935892010.163929299102950573.47E-1210passed
/artifact/estimates/7lre1.9668221230E+0119.6682212295977834.02E-1010passed
/artifact/covariance/7/7lre0.14293415998160141608960.142934159979765771.84E-1210passed

Convention


NIST Gauss3: nonlinear regression Start 2

Case STAT-NIST-GAUSS3-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Gauss3.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_GAUSS3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2422420.00E+2passed
/sselre1.2444846360E+031244.4846360131571.32E-810passed
/artifact/residual_variancelre5.142498496102756406255.1424984959221361.81E-1010passed
/artifact/estimates/0lre9.8940368970E+0198.940368969869721.30E-1011passed
/artifact/covariance/0/0lre0.28095504672635145658890.280955046723253443.10E-1210passed
/artifact/estimates/1lre1.0945879335E-020.0109458793348869021.13E-1310passed
/artifact/covariance/1/1lre1.57604395140858505921E-81.5760439513042984e-081.04E-1810passed
/artifact/estimates/2lre1.0069553078E+02100.695530778157941.84E-910passed
/artifact/covariance/2/2lre0.66026329824052452584890.66026329824625415.73E-1211passed
/artifact/estimates/3lre1.1163619459E+02111.636194587253142.75E-910passed
/artifact/covariance/3/3lre0.12473512178151200990490.12473512178120953.03E-1311passed
/artifact/estimates/4lre2.3300500029E+0123.3005000291694951.69E-1011passed
/artifact/covariance/4/4lre0.13384463488399890185290.133844634886088582.09E-1210passed
/artifact/estimates/5lre7.3705031418E+0173.705031417604293.96E-1011passed
/artifact/covariance/5/5lre1.461980625380842027241.46198062544061855.98E-1110passed
/artifact/estimates/6lre1.4776164251E+02147.761642505105364.89E-910passed
/artifact/covariance/6/6lre0.16392929910641935892010.163929299102951043.47E-1210passed
/artifact/estimates/7lre1.9668221230E+0119.668221229597854.02E-1010passed
/artifact/covariance/7/7lre0.14293415998160141608960.14293415997976661.83E-1210passed

Notes

Convention


NIST Hahn1: nonlinear regression Start 1

Case STAT-NIST-HAHN1-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Hahn1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_HAHN1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2292290.00E+2passed
/sselre1.5324382854E+001.53243828536061383.94E-1110passed
/artifact/residual_variancelre0.0066918702417945761310490.0066918702417494054.52E-1411passed
/artifact/estimates/0lre1.0776351733E+001.07763517330989019.89E-1211passed
/artifact/covariance/0/0lre0.02913901829158250865640.0291390182321721245.94E-118passed
/artifact/estimates/1lre-1.2269296921E-01-0.122692969210204982.05E-1311passed
/artifact/covariance/1/1lre0.0001440069406196302777210.000144006940549676087.00E-149passed
/artifact/estimates/2lre4.0863750610E-030.0040863750610466854.67E-1410passed
/artifact/covariance/2/2lre5.06624241303177313969E-85.0662424122480604e-087.84E-189passed
/artifact/estimates/3lre-1.4262662514E-06-1.42626625140785e-067.85E-1811passed
/artifact/covariance/3/3lre7.60548161507314727556E-147.605481614902566e-141.71E-2410passed
/artifact/estimates/4lre-5.7609940901E-03-0.0057609940900923577.64E-1511passed
/artifact/covariance/4/4lre6.10726844124788991961E-86.107268441530212e-082.82E-1810passed
/artifact/estimates/5lre2.4053735503E-040.000240537355031765331.77E-1511passed
/artifact/covariance/5/5lre1.09189412143366517824E-101.0918941212565823e-101.77E-209passed
/artifact/estimates/6lre-1.2314450199E-07-1.231445019919573e-071.96E-1810passed
/artifact/covariance/6/6lre1.69711465722102196929E-161.6971146566852321e-165.36E-269passed

Convention


NIST Hahn1: nonlinear regression Start 2

Case STAT-NIST-HAHN1-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Hahn1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_HAHN1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact2292290.00E+2passed
/sselre1.5324382854E+001.53243828536062843.94E-1110passed
/artifact/residual_variancelre0.0066918702417945761310490.0066918702417494694.51E-1411passed
/artifact/estimates/0lre1.0776351733E+001.0776351733115521.16E-1110passed
/artifact/covariance/0/0lre0.02913901829158250865640.0291390182712247362.04E-119passed
/artifact/estimates/1lre-1.2269296921E-01-0.122692969210330163.30E-1311passed
/artifact/covariance/1/1lre0.0001440069406196302777210.000144006940602781341.68E-149passed
/artifact/estimates/2lre4.0863750610E-030.0040863750610491124.91E-1410passed
/artifact/covariance/2/2lre5.06624241303177313969E-85.0662424119312555e-081.10E-179passed
/artifact/estimates/3lre-1.4262662514E-06-1.4262662514110709e-061.11E-1711passed
/artifact/covariance/3/3lre7.60548161507314727556E-147.60548161375172e-141.32E-239passed
/artifact/estimates/4lre-5.7609940901E-03-0.0057609940900918388.16E-1511passed
/artifact/covariance/4/4lre6.10726844124788991961E-86.107268441323152e-087.53E-1910passed
/artifact/estimates/5lre2.4053735503E-040.000240537355031872881.87E-1511passed
/artifact/covariance/5/5lre1.09189412143366517824E-101.0918941213019717e-101.32E-209passed
/artifact/estimates/6lre-1.2314450199E-07-1.2314450199211132e-072.11E-1810passed
/artifact/covariance/6/6lre1.69711465722102196929E-161.6971146573557257e-161.35E-2610passed

Convention


NIST Kirby2: nonlinear regression Start 1

Case STAT-NIST-KIRBY2-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Kirby2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_KIRBY2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact1461460.00E+2passed
/sselre3.9050739624E+003.90507396239089439.11E-1211passed
/artifact/residual_variancelre0.02674708193511131871610.0267470819341842089.27E-1310passed
/artifact/estimates/0lre1.6745063063E+001.67450630632180642.18E-1110passed
/artifact/covariance/0/0lre0.0077421757509349486982440.00774217575077980861.55E-1310passed
/artifact/estimates/1lre-1.3927397867E-01-0.139273978674737564.74E-1210passed
/artifact/covariance/1/1lre0.000016959605327182168009961.6959605327157046e-052.51E-1711passed
/artifact/estimates/2lre2.5961181191E-030.00259611811909347056.53E-1511passed
/artifact/covariance/2/2lre1.751968304850972529764E-91.7519683048531834e-092.21E-2111passed
/artifact/estimates/3lre-1.7241811870E-03-0.00172418118698271961.73E-1410passed
/artifact/covariance/3/3lre3.472968525860255996025E-93.4729685259017387e-094.15E-2010passed
/artifact/estimates/4lre2.1664802578E-052.1664802577673612e-053.26E-1610passed
/artifact/covariance/4/4lre4.05207314915622562561E-144.052073149317941e-141.62E-2410passed

Notes

Convention


NIST Kirby2: nonlinear regression Start 2

Case STAT-NIST-KIRBY2-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Kirby2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_KIRBY2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact1461460.00E+2passed
/sselre3.9050739624E+003.90507396239088379.12E-1211passed
/artifact/residual_variancelre0.02674708193511131871610.0267470819341841359.27E-1310passed
/artifact/estimates/0lre1.6745063063E+001.67450630632232622.23E-1110passed
/artifact/covariance/0/0lre0.0077421757509349486982440.0077421757510395721.05E-1310passed
/artifact/estimates/1lre-1.3927397867E-01-0.139273978674768044.77E-1210passed
/artifact/covariance/1/1lre0.000016959605327182168009961.695960532726947e-058.73E-1711passed
/artifact/estimates/2lre2.5961181191E-030.00259611811909379536.20E-1511passed
/artifact/covariance/2/2lre1.751968304850972529764E-91.7519683048634274e-091.25E-2011passed
/artifact/estimates/3lre-1.7241811870E-03-0.00172418118698229181.77E-1410passed
/artifact/covariance/3/3lre3.472968525860255996025E-93.4729685259190483e-095.88E-2010passed
/artifact/estimates/4lre2.1664802578E-052.166480257767509e-053.25E-1610passed
/artifact/covariance/4/4lre4.05207314915622562561E-144.052073149325224e-141.69E-2410passed

Notes

Convention


NIST Lanczos1: nonlinear regression Start 1

Case STAT-NIST-LANCZOS1-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Lanczos1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LANCZOS1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact18180.00E+2passed
/sselre1.4307867721E-251.4307870632247253e-252.91E-326passed
/artifact/residual_variancelre7.948815400495619163801E-277.94881701791514e-271.62E-336passed
/artifact/estimates/0lre9.5100000027E-020.095100000027431464.31E-1311passed
/artifact/covariance/0/0lre2.845934869034954326756E-212.8459354481681904e-215.79E-286passed
/artifact/estimates/1lre1.0000000001E+001.0000000001277192.77E-1110passed
/artifact/covariance/1/1lre7.54767826784791636041E-207.547679803850871e-201.54E-266passed
/artifact/estimates/2lre8.6070000013E-010.86070000013445794.46E-1211passed
/artifact/covariance/2/2lre1.84309465537071950625E-201.8430950303783614e-203.75E-276passed
/artifact/estimates/3lre3.0000000002E+003.00000000023462433.46E-1110passed
/artifact/covariance/3/3lre1.109439722508547918761E-191.1094399482683673e-192.26E-266passed
/artifact/estimates/4lre1.5575999998E+001.55759999983817133.82E-1110passed
/artifact/covariance/4/4lre3.54031749923256176704E-203.540318219491003e-207.20E-276passed
/artifact/estimates/5lre5.0000000001E+005.0000000001115491.15E-1111passed
/artifact/covariance/5/5lre1.22268318147870289444E-201.222683430339539e-202.49E-276passed

Convention


NIST Lanczos1: nonlinear regression Start 2

Case STAT-NIST-LANCZOS1-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Lanczos1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LANCZOS1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact18180.00E+2passed
/sselre1.4307867721E-251.4307870632247253e-252.91E-326passed
/artifact/residual_variancelre7.948815400495619163801E-277.94881701791514e-271.62E-336passed
/artifact/estimates/0lre9.5100000027E-020.095100000027431464.31E-1311passed
/artifact/covariance/0/0lre2.845934869034954326756E-212.8459354481681904e-215.79E-286passed
/artifact/estimates/1lre1.0000000001E+001.0000000001277192.77E-1110passed
/artifact/covariance/1/1lre7.54767826784791636041E-207.547679803850871e-201.54E-266passed
/artifact/estimates/2lre8.6070000013E-010.86070000013445794.46E-1211passed
/artifact/covariance/2/2lre1.84309465537071950625E-201.8430950303783614e-203.75E-276passed
/artifact/estimates/3lre3.0000000002E+003.00000000023462433.46E-1110passed
/artifact/covariance/3/3lre1.109439722508547918761E-191.1094399482683673e-192.26E-266passed
/artifact/estimates/4lre1.5575999998E+001.55759999983817133.82E-1110passed
/artifact/covariance/4/4lre3.54031749923256176704E-203.540318219491003e-207.20E-276passed
/artifact/estimates/5lre5.0000000001E+005.0000000001115491.15E-1111passed
/artifact/covariance/5/5lre1.22268318147870289444E-201.222683430339539e-202.49E-276passed

Notes

Convention


NIST Lanczos2: nonlinear regression Start 1

Case STAT-NIST-LANCZOS2-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Lanczos2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LANCZOS2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact18180.00E+2passed
/sselre2.2299428125E-112.2299428122902724e-112.10E-2110passed
/artifact/residual_variancelre1.23885711799958014201E-121.2388571179390403e-126.05E-2310passed
/artifact/estimates/0lre9.6251029939E-020.096251030236514232.98E-108passed
/artifact/covariance/0/0lre4.458309749529753777529E-74.4583097580268464e-078.50E-168passed
/artifact/estimates/1lre1.0057332849E+001.00573328637292651.47E-98passed
/artifact/covariance/1/1lre0.000011552960471696714229761.1552960423758281e-054.79E-148passed
/artifact/estimates/2lre8.6424689056E-010.86424689130691157.47E-109passed
/artifact/covariance/2/2lre0.000002953533262803254892252.953533282756303e-062.00E-148passed
/artifact/estimates/3lre3.0078283915E+003.00782839334493351.84E-99passed
/artifact/covariance/3/3lre0.000017394743374724182927361.7394743409401197e-053.47E-148passed
/artifact/estimates/4lre1.5529016879E+001.55290168682311981.08E-99passed
/artifact/covariance/4/4lre0.000005637956488759749278895.63795651925078e-063.05E-148passed
/artifact/estimates/5lre5.0028798100E+005.0028798105964865.96E-109passed
/artifact/covariance/5/5lre0.000001948477424399600966561.948477435005764e-061.06E-148passed

Convention


NIST Lanczos2: nonlinear regression Start 2

Case STAT-NIST-LANCZOS2-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Lanczos2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LANCZOS2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact18180.00E+2passed
/sselre2.2299428125E-112.2299428123953906e-111.05E-2110passed
/artifact/residual_variancelre1.23885711799958014201E-121.2388571179974393e-122.14E-2411passed
/artifact/estimates/0lre9.6251029939E-020.096251029881802455.72E-119passed
/artifact/covariance/0/0lre4.458309749529753777529E-74.4583097476102824e-071.92E-169passed
/artifact/estimates/1lre1.0057332849E+001.00573328457025733.30E-109passed
/artifact/covariance/1/1lre0.000011552960471696714229761.1552960480350489e-058.65E-159passed
/artifact/estimates/2lre8.6424689056E-010.86424689040819491.52E-109passed
/artifact/covariance/2/2lre0.000002953533262803254892252.953533258737468e-064.07E-158passed
/artifact/estimates/3lre3.0078283915E+003.007828391131173.69E-109passed
/artifact/covariance/3/3lre0.000017394743374724182927361.73947433673672e-057.36E-159passed
/artifact/estimates/4lre1.5529016879E+001.55290168807640821.76E-109passed
/artifact/covariance/4/4lre0.000005637956488759749278895.637956482362313e-066.40E-158passed
/artifact/estimates/5lre5.0028798100E+005.0028798098645311.35E-1010passed
/artifact/covariance/5/5lre0.000001948477424399600966561.9484774223167206e-062.08E-158passed

Convention


NIST Lanczos3: nonlinear regression Start 1

Case STAT-NIST-LANCZOS3-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Lanczos3.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LANCZOS3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact18180.00E+2passed
/sselre1.6117193594E-081.6117193593564184e-084.36E-1910passed
/artifact/residual_variancelre8.95399644080031805584E-108.953996440868991e-106.87E-2111passed
/artifact/estimates/0lre8.6816414977E-020.086816415046614126.96E-119passed
/artifact/covariance/0/0lre0.0002957680691224706818810.000295768069316433851.94E-139passed
/artifact/estimates/1lre9.5498101505E-010.95498101543669993.87E-109passed
/artifact/covariance/1/1lre0.0094170768807469190256250.0094170768717616858.99E-129passed
/artifact/estimates/2lre8.4400777463E-010.84400777479030221.60E-109passed
/artifact/covariance/2/2lre0.0017213091809271750115240.00172130918372640722.80E-128passed
/artifact/estimates/3lre2.9515951832E+002.9515951836245564.25E-109passed
/artifact/covariance/3/3lre0.01159134849768520000360.0115913485044202196.74E-129passed
/artifact/estimates/4lre1.5825685901E+001.58256858991523571.85E-109passed
/artifact/covariance/4/4lre0.0034072409175286017909610.00340724092204322454.51E-128passed
/artifact/estimates/5lre4.9863565084E+004.9863565085365651.37E-1010passed
/artifact/covariance/5/5lre0.0011858658539889572112250.00118586585553330321.54E-128passed

Notes

Convention


NIST Lanczos3: nonlinear regression Start 2

Case STAT-NIST-LANCZOS3-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Lanczos3.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LANCZOS3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact18180.00E+2passed
/sselre1.6117193594E-081.6117193593562484e-084.38E-1910passed
/artifact/residual_variancelre8.95399644080031805584E-108.953996440868046e-106.77E-2111passed
/artifact/estimates/0lre8.6816414977E-020.086816414977452884.53E-1311passed
/artifact/covariance/0/0lre0.0002957680691224706818810.000295768069126713054.24E-1510passed
/artifact/estimates/1lre9.5498101505E-010.95498101504727052.73E-1211passed
/artifact/covariance/1/1lre0.0094170768807469190256250.0094170768806748547.21E-1411passed
/artifact/estimates/2lre8.4400777463E-010.84400777462569474.31E-1211passed
/artifact/covariance/2/2lre0.0017213091809271750115240.0017213091809459961.88E-1410passed
/artifact/estimates/3lre2.9515951832E+002.95159518319136678.63E-1211passed
/artifact/covariance/3/3lre0.01159134849768520000360.0115913484977730838.79E-1411passed
/artifact/estimates/4lre1.5825685901E+001.58256859014897524.90E-1110passed
/artifact/covariance/4/4lre0.0034072409175286017909610.00340724091749058333.80E-1410passed
/artifact/estimates/5lre4.9863565084E+004.9863565083995094.91E-1313passed
/artifact/covariance/5/5lre0.0011858658539889572112250.00118586585399322874.27E-1511passed

Notes

Convention


NIST Lew: acf

Case STAT-NIST-LEW-ACF-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/Lew.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LEW.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlations/acf/1lre-0.307304800605679-0.307304800605679444.40E-1614passed
/correlations/sample_countexact2002000.00E+2passed

Convention


NIST Lew: tabulation

Case STAT-NIST-LEW-TABULATION-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/Lew.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LEW.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/meanlre-177.435000000000-177.4352.27E-1516passed
/numeric_summaries/0/standard_deviationlre277.332168044316277.33216804431611.20E-1315passed
/numeric_summaries/0/valid_nexact2002000.00E+2passed

Convention


NIST Longley: fit

Case STAT-NIST-LONGLEY-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Longley.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LONGLEY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/observationsexact16160.00E+2passed
/responses/0/parameters/[term=Intercept]/estimatelre-3482258.63459582-3482258.6345958211.15E-915passed
/responses/0/parameters/[term=Intercept]/std_errorlre890420.383607373890420.3836073721.06E-914passed
/responses/0/parameters/[term=x1]/estimatelre15.061872271373315.0618722713739226.22E-1313passed
/responses/0/parameters/[term=x1]/std_errorlre84.914925774766984.9149257747679.92E-1414passed
/responses/0/parameters/[term=x2]/estimatelre-0.358191792925910E-01-0.035819179292591222.22E-1614passed
/responses/0/parameters/[term=x2]/std_errorlre0.334910077722432E-010.0334910077722431752.46E-1715passed
/responses/0/parameters/[term=x3]/estimatelre-2.02022980381683-2.0202298038168282.18E-1514passed
/responses/0/parameters/[term=x3]/std_errorlre0.4883996816516990.48839968165169955.02E-1614passed
/responses/0/parameters/[term=x4]/estimatelre-1.03322686717359-1.0332268671735932.91E-1514passed
/responses/0/parameters/[term=x4]/std_errorlre0.2142741631616750.214274163161675161.63E-1615passed
/responses/0/parameters/[term=x5]/estimatelre-0.511041056535807E-01-0.051104105653579481.22E-1513passed
/responses/0/parameters/[term=x5]/std_errorlre0.2260732000693700.226073200069370484.84E-1614passed
/responses/0/parameters/[term=x6]/estimatelre1829.151464613551829.15146461355333.34E-1214passed
/responses/0/parameters/[term=x6]/std_errorlre455.478499142212455.478499142211663.40E-1315passed
/responses/0/anova/[source=Error]/sslre836424.055505915836424.05550591455.03E-1015passed
/responses/0/metrics/r_squaredlre0.9954790045772960.99547900457729421.79E-1514passed
/responses/0/anova/[source=Model]/f_statlre330.285339234588330.285339234588434.34E-1314passed

Notes

Convention


NIST Lottery: acf

Case STAT-NIST-LOTTERY-ACF-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/Lottery.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LOTTERY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlations/acf/1lre-0.120948622967393-0.120948622967392881.20E-1615passed
/correlations/sample_countexact2182180.00E+2passed

Convention


NIST Lottery: tabulation

Case STAT-NIST-LOTTERY-TABULATION-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/Lottery.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_LOTTERY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/meanlre518.958715596330518.95871559633033.08E-1315passed
/numeric_summaries/0/standard_deviationlre291.699727470969291.699727470969043.73E-1415passed
/numeric_summaries/0/valid_nexact2182180.00E+2passed

Convention


NIST Mavro: acf

Case STAT-NIST-MAVRO-ACF-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/Mavro.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MAVRO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlations/acf/1lre0.9379891834382480.93798918343826481.68E-1413passed
/correlations/sample_countexact50500.00E+2passed

Convention


NIST Mavro: tabulation

Case STAT-NIST-MAVRO-TABULATION-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/Mavro.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MAVRO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/meanlre2.001856000000002.0018567.98E-1716passed
/numeric_summaries/0/standard_deviationlre0.0004291234540030530.00042912345400308543.24E-1713passed
/numeric_summaries/0/valid_nexact50500.00E+2passed

Convention


NIST MGH09: nonlinear regression Start 1

Case STAT-NIST-MGH09-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/MGH09.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MGH09.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact770.00E+2passed
/sselre3.0750560385E-040.00030750560384923867.61E-1611passed
/artifact/residual_variancelre0.000043929371977834143756014.392937197846265e-056.29E-1610passed
/artifact/estimates/0lre1.9280693458E-010.192806934579037369.63E-1311passed
/artifact/covariance/0/0lre0.0001307663657289756995290.000130766365718785571.02E-1410passed
/artifact/estimates/1lre1.9128232873E-010.19128232873437834.38E-1210passed
/artifact/covariance/1/1lre0.03854633633401276699210.038546336334080286.75E-1411passed
/artifact/estimates/2lre1.2305650693E-010.123056506926322793.68E-1210passed
/artifact/covariance/2/2lre0.0065354340137156634378240.0065354340136760343.96E-1411passed
/artifact/estimates/3lre1.3606233068E-010.136062330683800083.80E-1210passed
/artifact/covariance/3/3lre0.0081045982678494979668640.0081045982678928934.34E-1411passed

Notes

Convention


NIST MGH09: nonlinear regression Start 2

Case STAT-NIST-MGH09-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/MGH09.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MGH09.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact770.00E+2passed
/sselre3.0750560385E-040.00030750560384923787.62E-1611passed
/artifact/residual_variancelre0.000043929371977834143756014.392937197846254e-056.28E-1610passed
/artifact/estimates/0lre1.9280693458E-010.192806934579038039.62E-1311passed
/artifact/covariance/0/0lre0.0001307663657289756995290.000130766365718780331.02E-1410passed
/artifact/estimates/1lre1.9128232873E-010.19128232873436394.36E-1210passed
/artifact/covariance/1/1lre0.03854633633401276699210.038546336334075696.29E-1411passed
/artifact/estimates/2lre1.2305650693E-010.123056506926320363.68E-1210passed
/artifact/covariance/2/2lre0.0065354340137156634378240.0065354340136754714.02E-1411passed
/artifact/estimates/3lre1.3606233068E-010.136062330683793423.79E-1210passed
/artifact/covariance/3/3lre0.0081045982678494979668640.0081045982678920884.26E-1411passed

Notes

Convention


NIST MGH10: nonlinear regression Start 1

Case STAT-NIST-MGH10-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/MGH10.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MGH10.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact13130.00E+2passed
/sselre8.7945855171E+0187.94585517077312.27E-1011passed
/artifact/residual_variancelre6.765065782488662042256.7650657823671611.22E-1010passed
/artifact/estimates/0lre5.6096364710E-030.0056096364710287182.87E-1411passed
/artifact/covariance/0/0lre2.46109970181658485841E-82.4610997017799544e-083.66E-1910passed
/artifact/estimates/1lre6.1813463463E+036181.346346286271.37E-811passed
/artifact/covariance/1/1lre543.310464966571505449543.31046496050686.06E-910passed
/artifact/estimates/2lre3.4522363462E+02345.223634624132964.13E-910passed
/artifact/covariance/2/2lre0.61600684438684899060640.61600684438080916.04E-1211passed

Convention


NIST MGH10: nonlinear regression Start 2

Case STAT-NIST-MGH10-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/MGH10.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MGH10.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact13130.00E+2passed
/sselre8.7945855171E+0187.945855170627063.73E-1011passed
/artifact/residual_variancelre6.765065782488662042256.76506578235592751.33E-1010passed
/artifact/estimates/0lre5.6096364710E-030.005609636481446911.04E-118passed
/artifact/covariance/0/0lre2.46109970181658485841E-82.4610997096973467e-087.88E-178passed
/artifact/estimates/1lre6.1813463463E+036181.3463447378921.56E-69passed
/artifact/covariance/1/1lre543.310464966571505449543.31046455409364.12E-79passed
/artifact/estimates/2lre3.4522363462E+02345.223634572009644.80E-89passed
/artifact/covariance/2/2lre0.61600684438684899060640.61600684407387433.13E-109passed

Convention


NIST MGH17: nonlinear regression Start 1

Case STAT-NIST-MGH17-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/MGH17.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MGH17.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact28280.00E+2passed
/sselre5.4648946975E-055.4648946974828085e-051.72E-1611passed
/artifact/residual_variancelre0.000001951748106239105539561.95174810624386e-064.75E-1811passed
/artifact/estimates/0lre3.7541005211E-010.375410052106952163.05E-1211passed
/artifact/covariance/0/0lre0.000004294490930983239096014.29449093117976e-061.97E-1610passed
/artifact/estimates/1lre1.9358469127E+001.9358469127123811.24E-1111passed
/artifact/covariance/1/1lre0.04853944487076220852840.048539444872630931.87E-1210passed
/artifact/estimates/2lre-1.4646871366E+00-1.46468713661343661.34E-1111passed
/artifact/covariance/2/2lre0.04917620137255444921210.049176201371419031.14E-1210passed
/artifact/estimates/3lre1.2867534640E-020.0128675346400573145.73E-1411passed
/artifact/covariance/3/3lre2.012541451832553636996E-72.012541451852016e-071.95E-1811passed
/artifact/estimates/4lre2.2122699662E-020.022122699661672553.27E-1310passed
/artifact/covariance/4/4lre8.005238171116811730625E-78.005238171028975e-078.78E-1810passed

Convention


NIST MGH17: nonlinear regression Start 2

Case STAT-NIST-MGH17-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/MGH17.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MGH17.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact28280.00E+2passed
/sselre5.4648946975E-055.464894697482894e-051.71E-1611passed
/artifact/residual_variancelre0.000001951748106239105539561.9517481062438905e-064.78E-1811passed
/artifact/estimates/0lre3.7541005211E-010.375410051961940441.48E-109passed
/artifact/covariance/0/0lre0.000004294490930983239096014.294490933258638e-062.28E-159passed
/artifact/estimates/1lre1.9358469127E+001.93584689577584231.69E-88passed
/artifact/covariance/1/1lre0.04853944487076220852840.0485394417292252243.14E-97passed
/artifact/estimates/2lre-1.4646871366E+00-1.46468711956742941.70E-87passed
/artifact/covariance/2/2lre0.04917620137255444921210.049176198207761143.16E-97passed
/artifact/estimates/3lre1.2867534640E-020.0128675346057242243.43E-118passed
/artifact/covariance/3/3lre2.012541451832553636996E-72.0125413966144034e-075.52E-157passed
/artifact/estimates/4lre2.2122699662E-020.0221226997302272426.82E-118passed
/artifact/covariance/4/4lre8.005238171116811730625E-78.005238068507154e-071.03E-147passed

Notes

Convention


NIST Michelso: acf

Case STAT-NIST-MICHELSO-ACF-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/Michelso.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MICHELSO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlations/acf/1lre0.5351996686212830.53519966862126331.97E-1413passed
/correlations/sample_countexact1001000.00E+2passed

Convention


NIST Michelso: tabulation

Case STAT-NIST-MICHELSO-TABULATION-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/Michelso.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MICHELSO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/meanlre299.852400000000299.85241.13E-1416passed
/numeric_summaries/0/standard_deviationlre0.07901054781905180.079010547819050661.14E-1513passed
/numeric_summaries/0/valid_nexact1001000.00E+2passed

Convention


NIST Misra1a: nonlinear regression Start 1

Case STAT-NIST-MISRA1A-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Misra1a.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MISRA1A.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact12120.00E+2passed
/sselre1.2455138894E-010.12455138894441314.41E-1210passed
/artifact/residual_variancelre0.01037928241153742689000.0103792824120344264.97E-1310passed
/artifact/estimates/0lre2.3894212918E+02238.942129178861561.14E-911passed
/artifact/covariance/0/0lre7.327889735534012080817.3278897357480082.14E-1010passed
/artifact/estimates/1lre5.5015643181E-040.00055015643180591394.09E-1511passed
/artifact/covariance/1/1lre5.280738279008440126096E-115.280738279036431e-112.80E-2211passed

Notes

Convention


NIST Misra1a: nonlinear regression Start 2

Case STAT-NIST-MISRA1A-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Misra1a.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MISRA1A.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact12120.00E+2passed
/sselre1.2455138894E-010.12455138894440214.40E-1210passed
/artifact/residual_variancelre0.01037928241153742689000.0103792824120335084.96E-1310passed
/artifact/estimates/0lre2.3894212918E+02238.94212917886531.13E-911passed
/artifact/covariance/0/0lre7.327889735534012080817.3278897357478512.14E-1010passed
/artifact/estimates/1lre5.5015643181E-040.0005501564318059044.10E-1511passed
/artifact/covariance/1/1lre5.280738279008440126096E-115.2807382790359324e-112.75E-2211passed

Notes

Convention


NIST Misra1b: nonlinear regression Start 1

Case STAT-NIST-MISRA1B-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Misra1b.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MISRA1B.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact12120.00E+2passed
/sselre7.5464681533E-020.075464681533381763.82E-1311passed
/artifact/residual_variancelre0.0062887234610495781120040.00628872346111514656.56E-1410passed
/artifact/estimates/0lre3.3799746163E+02337.99746163368813.69E-910passed
/artifact/covariance/0/0lre10.0133958470309534284910.0133958467337762.97E-1010passed
/artifact/estimates/1lre3.9039091287E-040.00039039091286569944.30E-1510passed
/artifact/covariance/1/1lre1.810274595245973123556E-111.810274595251836e-115.86E-2311passed

Notes

Convention


NIST Misra1b: nonlinear regression Start 2

Case STAT-NIST-MISRA1B-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Misra1b.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MISRA1B.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact12120.00E+2passed
/sselre7.5464681533E-020.075464681533381763.82E-1311passed
/artifact/residual_variancelre0.0062887234610495781120040.00628872346111514656.56E-1410passed
/artifact/estimates/0lre3.3799746163E+02337.99746163368813.69E-910passed
/artifact/covariance/0/0lre10.0133958470309534284910.0133958467337762.97E-1010passed
/artifact/estimates/1lre3.9039091287E-040.00039039091286569944.30E-1510passed
/artifact/covariance/1/1lre1.810274595245973123556E-111.810274595251836e-115.86E-2311passed

Notes

Convention


NIST Misra1c: nonlinear regression Start 1

Case STAT-NIST-MISRA1C-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Misra1c.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MISRA1C.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact12120.00E+2passed
/sselre4.0966836971E-020.040966836970675113.25E-1311passed
/artifact/residual_variancelre0.0034139030808505331760490.00341390308088959263.91E-1410passed
/artifact/estimates/0lre6.3642725809E+02636.42725809300733.01E-911passed
/artifact/covariance/0/0lre21.7513350543652127118421.7513350547137533.49E-1010passed
/artifact/estimates/1lre2.0813627256E-040.000208136272563285323.29E-1510passed
/artifact/covariance/1/1lre3.14296987562740154025E-123.142969875782169e-121.55E-2210passed

Notes

Convention


NIST Misra1c: nonlinear regression Start 2

Case STAT-NIST-MISRA1C-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Misra1c.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MISRA1C.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact12120.00E+2passed
/sselre4.0966836971E-020.040966836970678173.22E-1311passed
/artifact/residual_variancelre0.0034139030808505331760490.00341390308088984763.93E-1410passed
/artifact/estimates/0lre6.3642725809E+02636.42725809300333.00E-911passed
/artifact/covariance/0/0lre21.7513350543652127118421.7513350547153053.50E-1010passed
/artifact/estimates/1lre2.0813627256E-040.000208136272563286813.29E-1510passed
/artifact/covariance/1/1lre3.14296987562740154025E-123.142969875782486e-121.55E-2210passed

Notes

Convention


NIST Misra1d: nonlinear regression Start 1

Case STAT-NIST-MISRA1D-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Misra1d.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MISRA1D.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact12120.00E+2passed
/sselre5.6419295283E-020.0564192952826489263.51E-1311passed
/artifact/residual_variancelre0.0047016079402881403963210.0047016079402207446.74E-1410passed
/artifact/estimates/0lre4.3736970754E+02437.36970753989361.06E-1012passed
/artifact/covariance/0/0lre13.3145984437980617902513.3145984439782361.80E-1010passed
/artifact/estimates/1lre3.0227324449E-040.00030227324449004314.31E-1712passed
/artifact/covariance/1/1lre8.60504352699627361441E-128.605043527146268e-121.50E-2210passed

Notes

Convention


NIST Misra1d: nonlinear regression Start 2

Case STAT-NIST-MISRA1D-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Misra1d.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_MISRA1D.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact12120.00E+2passed
/sselre5.6419295283E-020.0564192952826460953.54E-1311passed
/artifact/residual_variancelre0.0047016079402881403963210.0047016079402205086.76E-1410passed
/artifact/estimates/0lre4.3736970754E+02437.36970753989331.07E-1012passed
/artifact/covariance/0/0lre13.3145984437980617902513.3145984439775291.79E-1010passed
/artifact/estimates/1lre3.0227324449E-040.00030227324449004344.34E-1712passed
/artifact/covariance/1/1lre8.60504352699627361441E-128.605043527145838e-121.50E-2210passed

Notes

Convention


NIST Nelson: nonlinear regression Start 1

Case STAT-NIST-NELSON-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Nelson.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NELSON_LOG_RESPONSE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact1251250.00E+2passed
/sselre3.7976833176E+003.7976833176451494.51E-1110passed
/artifact/residual_variancelre0.03038146654102728169000.030381466541161191.34E-1311passed
/artifact/estimates/0lre2.5906836021E+002.5906836021362463.62E-1110passed
/artifact/covariance/0/0lre0.0003667223626179128665690.00036672236259940941.85E-1410passed
/artifact/estimates/1lre5.6177717026E-095.6177717026494515e-094.95E-2011passed
/artifact/covariance/1/1lre3.736155177831239971600E-173.7361551778531725e-172.19E-2811passed
/artifact/estimates/2lre-5.7701013174E-02-0.057701013174447754.48E-1311passed
/artifact/covariance/2/2lre0.000015659721938135457708491.565972193848099e-053.46E-1610passed

Convention


NIST Nelson: nonlinear regression Start 2

Case STAT-NIST-NELSON-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Nelson.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NELSON_LOG_RESPONSE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact1251250.00E+2passed
/sselre3.7976833176E+003.79768331764514774.51E-1110passed
/artifact/residual_variancelre0.03038146654102728169000.030381466541161181.34E-1311passed
/artifact/estimates/0lre2.5906836021E+002.59068360213620073.62E-1110passed
/artifact/covariance/0/0lre0.0003667223626179128665690.000366722362599384161.85E-1410passed
/artifact/estimates/1lre5.6177717026E-095.617771702616804e-091.68E-2011passed
/artifact/covariance/1/1lre3.736155177831239971600E-173.7361551778141565e-171.71E-2811passed
/artifact/estimates/2lre-5.7701013174E-02-0.057701013174468894.69E-1311passed
/artifact/covariance/2/2lre0.000015659721938135457708491.565972193849939e-053.64E-1610passed

Convention


NIST NoInt1: fit

Case STAT-NIST-NOINT1-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/NoInt1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NOINT1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/eligible_countexact11110.00E+2passed
/artifact/degrees_of_freedomexact10100.00E+2passed
/artifact/estimates/0lre2.074380165289262.0743801652892563.83E-1514passed
/validation_standard_errors/0lre0.165289256198347E-010.0165289256198347044.26E-1815passed
/sselre127.272727272727127.272727272727151.53E-1314passed

Convention


NIST NoInt2: fit

Case STAT-NIST-NOINT2-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/NoInt2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NOINT2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/eligible_countexact330.00E+2passed
/artifact/degrees_of_freedomexact220.00E+2passed
/artifact/estimates/0lre0.7272727272727270.72727272727272732.93E-1615passed
/validation_standard_errors/0lre0.420827318078432E-010.0420827318078432141.42E-1715passed
/sselre0.2727272727272730.27272727272727227.93E-1614passed

Convention


NIST Norris: fit

Case STAT-NIST-NORRIS-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Norris.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NORRIS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/[term=Intercept]/estimatelre-0.262323073774029-0.262323073774054172.52E-1413passed
/responses/0/parameters/[term=x]/estimatelre1.002116818020451.00211681802045454.49E-1514passed
/responses/0/parameters/[term=Intercept]/std_errorlre0.2328182343011520.23281823430115472.70E-1513passed
/responses/0/parameters/[term=x]/std_errorlre0.429796848199937E-030.00042979684819994093.92E-1814passed
/responses/0/metrics/r_squaredlre0.9999937458837120.99999374588371172.96E-1615passed
/responses/0/anova/[source=Error]/sslre26.617398529422426.6173985294228684.68E-1313passed
/responses/0/observationsexact36360.00E+2passed
/responses/0/anova/[source=Model]/f_statlre5436385.540797855436385.5407977431.07E-713passed

Convention


NIST Norris: general linear

Case STAT-NIST-NORRIS-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Norris.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NORRIS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/Interceptlre-0.262323073774029-0.262323073774192841.64E-1312passed
/coefficients/xlre1.002116818020451.00211681802045474.71E-1514passed
/observationsexact36360.00E+2passed
/residual_dfexact34340.00E+2passed
/residual_variancelre0.7828646626300690.78286466263008641.74E-1413passed
/term_tests/[term=x]/f_statisticlre5436385.540797855436385.5407977291.21E-713passed

Convention


NIST Norris: linear effects

Case STAT-NIST-NORRIS-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Norris.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NORRIS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/Interceptlre-0.262323073774029-0.26232307377387091.58E-1312passed
/coefficients/xlre1.002116818020451.0021168180204544.05E-1514passed
/observationsexact36360.00E+2passed
/residual_dfexact34340.00E+2passed
/residual_variancelre0.7828646626300690.7828646626300791.00E-1413passed
/term_tests/[term=x]/f_statisticlre5436385.540797855436385.5407977787.16E-813passed

Convention


NIST NumAcc1: tabulation

Case STAT-NIST-NUMACC1-TABULATION-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/NumAcc1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NUMACC1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/meanlre1000000210000002.00.00E+2passed
/numeric_summaries/0/standard_deviationlre11.00.00E+2passed
/numeric_summaries/0/valid_nexact330.00E+2passed

Convention


NIST NumAcc2: acf

Case STAT-NIST-NUMACC2-ACF-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/NumAcc2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NUMACC2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlations/acf/1lre-0.999-0.99899999999999991.12E-1615passed
/correlations/sample_countexact100110010.00E+2passed

Convention


NIST NumAcc2: tabulation

Case STAT-NIST-NUMACC2-TABULATION-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/NumAcc2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NUMACC2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/meanlre1.21.24.44E-1716passed
/numeric_summaries/0/standard_deviationlre0.10.099999999999999982.22E-1715passed
/numeric_summaries/0/valid_nexact100110010.00E+2passed

Convention


NIST NumAcc3: acf

Case STAT-NIST-NUMACC3-ACF-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/NumAcc3.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NUMACC3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlations/acf/1lre-0.999-0.9998.88E-1918passed
/correlations/sample_countexact100110010.00E+2passed

Convention


NIST NumAcc3: tabulation

Case STAT-NIST-NUMACC3-TABULATION-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/NumAcc3.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NUMACC3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/meanlre1000000.21000000.20000000016.98E-1116passed
/numeric_summaries/0/standard_deviationlre0.10.10000000003492463.49E-119passed
/numeric_summaries/0/valid_nexact100110010.00E+2passed

Convention


NIST NumAcc4: acf

Case STAT-NIST-NUMACC4-ACF-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/NumAcc4.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NUMACC4.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlations/acf/1lre-0.999-0.99899999999999982.23E-1615passed
/correlations/sample_countexact100110010.00E+2passed

Convention


NIST NumAcc4: tabulation

Case STAT-NIST-NUMACC4-TABULATION-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/NumAcc4.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_NUMACC4.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/meanlre10000000.210000000.2000000011.12E-915passed
/numeric_summaries/0/standard_deviationlre0.10.100000000558793545.59E-108passed
/numeric_summaries/0/valid_nexact100110010.00E+2passed

Convention


NIST PiDigits: acf

Case STAT-NIST-PIDIGITS-ACF-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/PiDigits.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_PIDIGITS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/correlations/acf/1lre-0.00355099287237972-0.00355099287237972161.62E-1815passed
/correlations/sample_countexact500050000.00E+2passed

Convention


NIST PiDigits: tabulation

Case STAT-NIST-PIDIGITS-TABULATION-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/univ/data/PiDigits.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_PIDIGITS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/numeric_summaries/0/meanlre4.534800000000004.53482.80E-1616passed
/numeric_summaries/0/standard_deviationlre2.867339060288712.8673390602887081.89E-1515passed
/numeric_summaries/0/valid_nexact500050000.00E+2passed

Convention


NIST Pontius: fit

Case STAT-NIST-PONTIUS-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Pontius.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_PONTIUS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/observationsexact40400.00E+2passed
/responses/0/parameters/[term=Intercept]/estimatelre0.673565789473684E-030.00067356578947364084.32E-1713passed
/responses/0/parameters/[term=Intercept]/std_errorlre0.107938612033077E-030.000107938612033077242.36E-1914passed
/responses/0/parameters/[term=x]/estimatelre0.732059160401003E-067.320591604010026e-074.22E-2215passed
/responses/0/parameters/[term=x]/std_errorlre0.157817399981659E-091.578173999816591e-101.08E-2515passed
/responses/0/parameters/[term=x^2]/estimatelre-0.316081871345029E-14-3.160818713450321e-153.08E-29passed
/responses/0/parameters/[term=x^2]/std_errorlre0.486652849992036E-164.866528499920374e-171.39E-3114passed
/responses/0/anova/[source=Error]/sslre0.155761768796992E-051.5576176879699351e-061.51E-2014passed
/responses/0/metrics/r_squaredlre0.9999999001785370.99999990017853711.27E-1615passed
/responses/0/anova/[source=Model]/f_statlre185330865.995752185330865.995750371.63E-614passed

Notes

Convention


NIST Rat42: nonlinear regression Start 1

Case STAT-NIST-RAT42-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Rat42.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_RAT42.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact660.00E+2passed
/sselre8.0565229338E+008.0565229338112841.13E-1111passed
/artifact/residual_variancelre1.342753822401747990011.34275382230188079.99E-1110passed
/artifact/estimates/0lre7.2462237576E+0172.462237575569184.31E-1011passed
/artifact/covariance/0/0lre3.006854284269961268013.00685428411123821.59E-1010passed
/artifact/estimates/1lre2.6180768402E+002.6180768402084978.50E-1211passed
/artifact/covariance/1/1lre0.0077960454397295619112960.00779604543981059048.10E-1410passed
/artifact/estimates/2lre6.7359200066E-020.067359200066105031.05E-1311passed
/artifact/covariance/2/2lre0.000011878819520166790441291.1878819520044612e-051.22E-1610passed

Notes

Convention


NIST Rat42: nonlinear regression Start 2

Case STAT-NIST-RAT42-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Rat42.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_RAT42.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact660.00E+2passed
/sselre8.0565229338E+008.0565229338112941.13E-1111passed
/artifact/residual_variancelre1.342753822401747990011.34275382230188249.99E-1110passed
/artifact/estimates/0lre7.2462237576E+0172.462237575569194.31E-1011passed
/artifact/covariance/0/0lre3.006854284269961268013.00685428411125161.59E-1010passed
/artifact/estimates/1lre2.6180768402E+002.6180768402084978.50E-1211passed
/artifact/covariance/1/1lre0.0077960454397295619112960.0077960454398106398.11E-1410passed
/artifact/estimates/2lre6.7359200066E-020.067359200066105021.05E-1311passed
/artifact/covariance/2/2lre0.000011878819520166790441291.1878819520044698e-051.22E-1610passed

Convention


NIST Rat43: nonlinear regression Start 1

Case STAT-NIST-RAT43-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Rat43.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_RAT43.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact11110.00E+2passed
/sselre8.7864049080E+038786.4049079631513.68E-811passed
/artifact/residual_variancelre798.764082526832574244798.76408254210471.53E-810passed
/artifact/estimates/0lre6.9964151270E+02699.64151269644563.55E-911passed
/artifact/covariance/0/0lre265.764914114162965489265.76491410428089.88E-910passed
/artifact/estimates/1lre5.2771253025E+005.2771253024500485.00E-1111passed
/artifact/covariance/1/1lre4.338362362342683172414.33836236253487151.92E-1010passed
/artifact/estimates/2lre7.5962938329E-010.75962938329413494.13E-1211passed
/artifact/covariance/2/2lre0.03828331868997721494010.0382833186895376464.40E-1310passed
/artifact/estimates/3lre1.2792483859E+001.27924838590971469.71E-1211passed
/artifact/covariance/3/3lre0.47282038954147868682250.47282038954110553.73E-1312passed

Notes

Convention


NIST Rat43: nonlinear regression Start 2

Case STAT-NIST-RAT43-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Rat43.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_RAT43.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact11110.00E+2passed
/sselre8.7864049080E+038786.404907963153.69E-811passed
/artifact/residual_variancelre798.764082526832574244798.76408254210451.53E-810passed
/artifact/estimates/0lre6.9964151270E+02699.64151269644563.55E-911passed
/artifact/covariance/0/0lre265.764914114162965489265.764914104281169.88E-910passed
/artifact/estimates/1lre5.2771253025E+005.2771253024500355.00E-1111passed
/artifact/covariance/1/1lre4.338362362342683172414.33836236253485551.92E-1010passed
/artifact/estimates/2lre7.5962938329E-010.75962938329413384.13E-1211passed
/artifact/covariance/2/2lre0.03828331868997721494010.03828331868953754.40E-1310passed
/artifact/estimates/3lre1.2792483859E+001.27924838590971019.71E-1211passed
/artifact/covariance/3/3lre0.47282038954147868682250.47282038954110183.77E-1312passed

Notes

Convention


NIST Roszman1: nonlinear regression Start 1

Case STAT-NIST-ROSZMAN1-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Roszman1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_ROSZMAN1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact21210.00E+2passed
/sselre4.9484847331E-040.00049484847330968813.12E-1612passed
/artifact/residual_variancelre0.000023564213014494140836002.3564213014747053e-052.53E-1610passed
/artifact/estimates/0lre2.0196866396E-010.201968663960292732.93E-1311passed
/artifact/covariance/0/0lre0.0003675911224294986365290.0003675911224242885.21E-1510passed
/artifact/estimates/1lre-6.1953516256E-06-6.1953516254825735e-061.17E-1610passed
/artifact/covariance/1/1lre1.027775101168204119481E-111.0277751011881212e-111.99E-2210passed
/artifact/estimates/2lre1.2044556708E+031204.45567078616771.38E-810passed
/artifact/covariance/2/2lre5483.5480917081010652495483.54809170339654.70E-912passed
/artifact/estimates/3lre-1.8134269537E+02-181.342695370381483.81E-1011passed
/artifact/covariance/3/3lre2457.5332753369947948012457.53327540354356.65E-810passed

Convention


NIST Roszman1: nonlinear regression Start 2

Case STAT-NIST-ROSZMAN1-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Roszman1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_ROSZMAN1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact21210.00E+2passed
/sselre4.9484847331E-040.00049484847330969233.08E-1612passed
/artifact/residual_variancelre0.000023564213014494140836002.3564213014747253e-052.53E-1610passed
/artifact/estimates/0lre2.0196866396E-010.201968663877778238.22E-119passed
/artifact/covariance/0/0lre0.0003675911224294986365290.000367591122486581065.71E-149passed
/artifact/estimates/1lre-6.1953516256E-06-6.195351612987561e-061.26E-148passed
/artifact/covariance/1/1lre1.027775101168204119481E-111.0277751013804922e-112.12E-219passed
/artifact/estimates/2lre1.2044556708E+031204.45567112872623.29E-79passed
/artifact/covariance/2/2lre5483.5480917081010652495483.5480905814261.13E-69passed
/artifact/estimates/3lre-1.8134269537E+02-181.342695548628341.79E-79passed
/artifact/covariance/3/3lre2457.5332753369947948012457.53327854561353.21E-68passed

Convention


NIST SiRstv: fit

Case STAT-NIST-SIRSTV-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SiRstv.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SIRSTV.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre1.18046237440255E+001.18046237440244631.04E-1313passed
/responses/0/observationsexact25250.00E+2passed
/responses/0/anova/[source=Model]/dftolerance44.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance2020.00.00E+215passed
/responses/0/anova/[source=Model]/sslre5.11462616000000E-020.051146261599999495.07E-1614passed
/responses/0/anova/[source=Error]/sslre2.16636560000000E-010.21663656000001651.65E-1413passed

Convention


NIST SiRstv: general linear

Case STAT-NIST-SIRSTV-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SiRstv.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SIRSTV.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre1.18046237440255E+001.18046237440244741.03E-1313passed
/term_tests/[term=group]/numerator_dfexact440.00E+2passed
/residual_dfexact20200.00E+2passed
/observationsexact25250.00E+2passed
/term_tests/[term=group]/sum_sqlre5.11462616000000E-020.051146261599999554.52E-1614passed
/term_tests/[term=group]/mean_sqlre1.27865654000000E-020.0127865653999998871.13E-1614passed
/residual_variancelre1.08318280000000E-020.0108318280000008268.26E-1613passed

Convention


NIST SiRstv: linear effects

Case STAT-NIST-SIRSTV-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SiRstv.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SIRSTV.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre1.18046237440255E+001.18046237440244671.03E-1313passed
/term_tests/[term=group]/numerator_dfexact440.00E+2passed
/residual_dfexact20200.00E+2passed
/observationsexact25250.00E+2passed
/term_tests/[term=group]/sum_sqlre5.11462616000000E-020.051146261599999524.79E-1614passed
/term_tests/[term=group]/mean_sqlre1.27865654000000E-020.012786565399999881.20E-1614passed
/residual_variancelre1.08318280000000E-020.0108318280000008268.26E-1613passed

Convention


NIST SmLs01: fit

Case STAT-NIST-SMLS01-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs01.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre2.10000000000000E+0121.0000000000000043.55E-1515passed
/responses/0/anova/[source=Model]/dftolerance88.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance180180.00.00E+215passed
/responses/0/observationsexact1891890.00E+2passed
/responses/0/anova/[source=Model]/sslre1.68000000000000E+001.68000000000000131.27E-1515passed
/responses/0/anova/[source=Error]/sslre1.80000000000000E+001.8000000000000019.33E-1615passed

Convention


NIST SmLs01: general linear

Case STAT-NIST-SMLS01-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs01.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.10000000000000E+0120.9999999999999861.42E-1415passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact1801800.00E+2passed
/observationsexact1891890.00E+2passed
/term_tests/[term=group]/sum_sqlre1.68000000000000E+001.67999999999999955.06E-1615passed
/term_tests/[term=group]/mean_sqlre2.10000000000000E-010.209999999999999946.33E-1715passed
/residual_variancelre1.00000000000000E-020.0100000000000000043.68E-1815passed

Convention


NIST SmLs01: linear effects

Case STAT-NIST-SMLS01-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs01.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.10000000000000E+0121.00.00E-11passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact1801800.00E+2passed
/observationsexact1891890.00E+2passed
/term_tests/[term=group]/sum_sqlre1.68000000000000E+001.68000000000000066.04E-1615passed
/term_tests/[term=group]/mean_sqlre2.10000000000000E-010.210000000000000087.55E-1715passed
/residual_variancelre1.00000000000000E-020.0100000000000000043.68E-1815passed

Convention


NIST SmLs02: fit

Case STAT-NIST-SMLS02-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs02.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS02.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre2.01000000000000E+02201.000000000000065.68E-1415passed
/responses/0/anova/[source=Model]/dftolerance88.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance18001800.00.00E+215passed
/responses/0/observationsexact180918090.00E+2passed
/responses/0/anova/[source=Model]/sslre1.60800000000000E+0116.0800000000000131.25E-1415passed
/responses/0/anova/[source=Error]/sslre1.80000000000000E+0118.000000000000011.07E-1415passed

Convention


NIST SmLs02: general linear

Case STAT-NIST-SMLS02-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs02.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS02.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.01000000000000E+02201.000000000000373.69E-1314passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact180018000.00E+2passed
/observationsexact180918090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60800000000000E+0116.0800000000000161.61E-1415passed
/term_tests/[term=group]/mean_sqlre2.01000000000000E+002.0100000000000022.01E-1515passed
/residual_variancelre1.00000000000000E-020.0099999999999999928.47E-1815passed

Convention


NIST SmLs02: linear effects

Case STAT-NIST-SMLS02-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs02.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS02.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.01000000000000E+02201.000000000000262.56E-1314passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact180018000.00E+2passed
/observationsexact180918090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60800000000000E+0116.080000000000033.03E-1414passed
/term_tests/[term=group]/mean_sqlre2.01000000000000E+002.0100000000000043.78E-1514passed
/residual_variancelre1.00000000000000E-020.0100000000000000055.41E-1815passed

Convention


NIST SmLs03: fit

Case STAT-NIST-SMLS03-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs03.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS03.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre2.00100000000000E+032001.0000000000019.09E-1315passed
/responses/0/anova/[source=Model]/dftolerance88.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance1800018000.00.00E+215passed
/responses/0/observationsexact18009180090.00E+2passed
/responses/0/anova/[source=Model]/sslre1.60080000000000E+02160.080000000000151.55E-1315passed
/responses/0/anova/[source=Error]/sslre1.80000000000000E+02180.000000000000098.53E-1415passed

Convention


NIST SmLs03: general linear

Case STAT-NIST-SMLS03-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs03.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS03.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.00100000000000E+032001.00000000002612.61E-1113passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact18000180000.00E+2passed
/observationsexact18009180090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60080000000000E+02160.080000000002142.14E-1213passed
/term_tests/[term=group]/mean_sqlre2.00100000000000E+0120.0100000000002682.68E-1313passed
/residual_variancelre1.00000000000000E-020.0100000000000000043.68E-1815passed

Convention


NIST SmLs03: linear effects

Case STAT-NIST-SMLS03-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs03.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS03.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.00100000000000E+032000.99999999997952.05E-1113passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact18000180000.00E+2passed
/observationsexact18009180090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60080000000000E+02160.07999999999851.49E-1214passed
/term_tests/[term=group]/mean_sqlre2.00100000000000E+0120.0099999999998131.87E-1314passed
/residual_variancelre1.00000000000000E-020.0100000000000000098.88E-1815passed

Convention


NIST SmLs04: fit

Case STAT-NIST-SMLS04-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs04.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS04.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre2.10000000000000E+0120.9999999999999963.55E-1515passed
/responses/0/anova/[source=Model]/dftolerance88.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance180180.00.00E+215passed
/responses/0/observationsexact1891890.00E+2passed
/responses/0/anova/[source=Model]/sslre1.68000000000000E+001.68000000000000021.60E-1616passed
/responses/0/anova/[source=Error]/sslre1.80000000000000E+001.80000000000000032.66E-1615passed

Convention


NIST SmLs04: general linear

Case STAT-NIST-SMLS04-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs04.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS04.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.10000000000000E+0120.999999999999972.84E-1414passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact1801800.00E+2passed
/observationsexact1891890.00E+2passed
/term_tests/[term=group]/sum_sqlre1.68000000000000E+001.67999999999999821.84E-1514passed
/term_tests/[term=group]/mean_sqlre2.10000000000000E-010.209999999999999772.30E-1614passed
/residual_variancelre1.00000000000000E-020.0100000000000000021.94E-1815passed

Convention


NIST SmLs04: linear effects

Case STAT-NIST-SMLS04-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs04.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS04.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.10000000000000E+0121.00.00E-11passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact1801800.00E+2passed
/observationsexact1891890.00E+2passed
/term_tests/[term=group]/sum_sqlre1.68000000000000E+001.686.22E-1716passed
/term_tests/[term=group]/mean_sqlre2.10000000000000E-010.217.77E-1816passed
/residual_variancelre1.00000000000000E-020.012.08E-1916passed

Convention


NIST SmLs05: fit

Case STAT-NIST-SMLS05-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs05.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS05.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre2.01000000000000E+02200.999999999999945.68E-1415passed
/responses/0/anova/[source=Model]/dftolerance88.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance18001800.00.00E+215passed
/responses/0/observationsexact180918090.00E+2passed
/responses/0/anova/[source=Model]/sslre1.60800000000000E+0116.081.71E-1515passed
/responses/0/anova/[source=Error]/sslre1.80000000000000E+0118.0000000000000043.55E-1515passed

Convention


NIST SmLs05: general linear

Case STAT-NIST-SMLS05-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs05.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS05.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.01000000000000E+02200.999999999999663.41E-1314passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact180018000.00E+2passed
/observationsexact180918090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60800000000000E+0116.0799999999999841.59E-1415passed
/term_tests/[term=group]/mean_sqlre2.01000000000000E+002.0099999999999981.99E-1515passed
/residual_variancelre1.00000000000000E-020.0100000000000000077.15E-1815passed

Convention


NIST SmLs05: linear effects

Case STAT-NIST-SMLS05-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs05.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS05.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.01000000000000E+02201.000000000000098.53E-1415passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact180018000.00E+2passed
/observationsexact180918090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60800000000000E+0116.0800000000000131.25E-1415passed
/term_tests/[term=group]/mean_sqlre2.01000000000000E+002.01000000000000161.56E-1515passed
/residual_variancelre1.00000000000000E-020.0100000000000000043.68E-1815passed

Convention


NIST SmLs06: fit

Case STAT-NIST-SMLS06-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs06.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS06.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre2.00100000000000E+032001.00000000000054.55E-1315passed
/responses/0/anova/[source=Model]/dftolerance88.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance1800018000.00.00E+215passed
/responses/0/observationsexact18009180090.00E+2passed
/responses/0/anova/[source=Model]/sslre1.60080000000000E+02160.080000000000076.93E-1415passed
/responses/0/anova/[source=Error]/sslre1.80000000000000E+02180.000000000000032.84E-1415passed

Convention


NIST SmLs06: general linear

Case STAT-NIST-SMLS06-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs06.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS06.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.00100000000000E+032001.00000000002462.46E-1113passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact18000180000.00E+2passed
/observationsexact18009180090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60080000000000E+02160.08000000000212.09E-1213passed
/term_tests/[term=group]/mean_sqlre2.00100000000000E+0120.010000000000262.61E-1313passed
/residual_variancelre1.00000000000000E-020.0100000000000000077.15E-1815passed

Convention


NIST SmLs06: linear effects

Case STAT-NIST-SMLS06-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs06.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS06.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.00100000000000E+032000.99999999998481.52E-1114passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact18000180000.00E+2passed
/observationsexact18009180090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60080000000000E+02160.079999999998731.27E-1214passed
/term_tests/[term=group]/mean_sqlre2.00100000000000E+0120.009999999999841.58E-1314passed
/residual_variancelre1.00000000000000E-020.0099999999999999973.26E-1815passed

Convention


NIST SmLs07: fit

Case STAT-NIST-SMLS07-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs07.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS07.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre2.10000000000000E+0120.9999999999999963.55E-1515passed
/responses/0/anova/[source=Model]/dftolerance88.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance180180.00.00E+215passed
/responses/0/observationsexact1891890.00E+2passed
/responses/0/anova/[source=Model]/sslre1.68000000000000E+001.68000000000000021.60E-1616passed
/responses/0/anova/[source=Error]/sslre1.80000000000000E+001.80000000000000032.66E-1615passed

Convention


NIST SmLs07: general linear

Case STAT-NIST-SMLS07-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs07.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS07.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.10000000000000E+0120.999999999999972.84E-1414passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact1801800.00E+2passed
/observationsexact1891890.00E+2passed
/term_tests/[term=group]/sum_sqlre1.68000000000000E+001.67999999999999821.84E-1514passed
/term_tests/[term=group]/mean_sqlre2.10000000000000E-010.209999999999999772.30E-1614passed
/residual_variancelre1.00000000000000E-020.0100000000000000021.94E-1815passed

Convention


NIST SmLs07: linear effects

Case STAT-NIST-SMLS07-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs07.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS07.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.10000000000000E+0121.00.00E-11passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact1801800.00E+2passed
/observationsexact1891890.00E+2passed
/term_tests/[term=group]/sum_sqlre1.68000000000000E+001.686.22E-1716passed
/term_tests/[term=group]/mean_sqlre2.10000000000000E-010.217.77E-1816passed
/residual_variancelre1.00000000000000E-020.012.08E-1916passed

Convention


NIST SmLs08: fit

Case STAT-NIST-SMLS08-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs08.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS08.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre2.01000000000000E+02200.999999999999945.68E-1415passed
/responses/0/anova/[source=Model]/dftolerance88.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance18001800.00.00E+215passed
/responses/0/observationsexact180918090.00E+2passed
/responses/0/anova/[source=Model]/sslre1.60800000000000E+0116.081.71E-1515passed
/responses/0/anova/[source=Error]/sslre1.80000000000000E+0118.0000000000000043.55E-1515passed

Convention


NIST SmLs08: general linear

Case STAT-NIST-SMLS08-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs08.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS08.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.01000000000000E+02200.999999999999663.41E-1314passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact180018000.00E+2passed
/observationsexact180918090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60800000000000E+0116.0799999999999841.59E-1415passed
/term_tests/[term=group]/mean_sqlre2.01000000000000E+002.0099999999999981.99E-1515passed
/residual_variancelre1.00000000000000E-020.0100000000000000077.15E-1815passed

Convention


NIST SmLs08: linear effects

Case STAT-NIST-SMLS08-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs08.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS08.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.01000000000000E+02201.000000000000098.53E-1415passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact180018000.00E+2passed
/observationsexact180918090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60800000000000E+0116.0800000000000131.25E-1415passed
/term_tests/[term=group]/mean_sqlre2.01000000000000E+002.01000000000000161.56E-1515passed
/residual_variancelre1.00000000000000E-020.0100000000000000043.68E-1815passed

Convention


NIST SmLs09: fit

Case STAT-NIST-SMLS09-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs09.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS09.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/anova/[source=Model]/f_statlre2.00100000000000E+032001.00000000000054.55E-1315passed
/responses/0/anova/[source=Model]/dftolerance88.00.00E+215passed
/responses/0/anova/[source=Error]/dftolerance1800018000.00.00E+215passed
/responses/0/observationsexact18009180090.00E+2passed
/responses/0/anova/[source=Model]/sslre1.60080000000000E+02160.080000000000076.93E-1415passed
/responses/0/anova/[source=Error]/sslre1.80000000000000E+02180.000000000000032.84E-1415passed

Convention


NIST SmLs09: general linear

Case STAT-NIST-SMLS09-GENERAL-LINEAR-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs09.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS09.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.00100000000000E+032001.00000000002462.46E-1113passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact18000180000.00E+2passed
/observationsexact18009180090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60080000000000E+02160.08000000000212.09E-1213passed
/term_tests/[term=group]/mean_sqlre2.00100000000000E+0120.010000000000262.61E-1313passed
/residual_variancelre1.00000000000000E-020.0100000000000000077.15E-1815passed

Convention


NIST SmLs09: linear effects

Case STAT-NIST-SMLS09-LINEAR-EFFECTS-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/anova/SmLs09.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_SMLS09.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/[term=group]/f_statisticlre2.00100000000000E+032000.99999999998481.52E-1114passed
/term_tests/[term=group]/numerator_dfexact880.00E+2passed
/residual_dfexact18000180000.00E+2passed
/observationsexact18009180090.00E+2passed
/term_tests/[term=group]/sum_sqlre1.60080000000000E+02160.079999999998731.27E-1214passed
/term_tests/[term=group]/mean_sqlre2.00100000000000E+0120.009999999999841.58E-1314passed
/residual_variancelre1.00000000000000E-020.0099999999999999973.26E-1815passed

Convention


NIST Thurber: nonlinear regression Start 1

Case STAT-NIST-THURBER-START1-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Thurber.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_THURBER.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact30300.00E+2passed
/sselre5.6427082397E+035642.70823966700853.30E-811passed
/artifact/residual_variancelre188.090274664493414656188.090274655566968.93E-910passed
/artifact/estimates/0lre1.2881396800E+031288.13968009108189.11E-810passed
/artifact/covariance/0/0lre21.7603248414316766233621.760324850748539.32E-99passed
/artifact/estimates/1lre1.4910792535E+031491.07924698680186.51E-68passed
/artifact/covariance/1/1lre1565.8763939825748393961565.87661575314772.22E-46passed
/artifact/estimates/2lre5.8323836877E+02583.23836389352144.88E-68passed
/artifact/covariance/2/2lre823.615157954949994404823.6152748304161.17E-46passed
/artifact/estimates/3lre7.5416644291E+0175.416643356139569.35E-77passed
/artifact/covariance/3/3lre30.9974685470159987290030.997473049182884.50E-66passed
/artifact/estimates/4lre9.6629502864E-010.96629502312706155.51E-98passed
/artifact/covariance/4/4lre0.0009817782386076096319690.00098177835961036861.21E-106passed
/artifact/estimates/5lre3.9797285797E-010.39797285541859262.55E-98passed
/artifact/covariance/5/5lre0.0002245480730992155272040.0002245481011925992.81E-116passed
/artifact/estimates/6lre4.9727297349E-020.0497272966705376166.78E-107passed
/artifact/covariance/6/6lre0.000043352143454538970121294.335214948657674e-056.03E-126passed

Notes

Convention


NIST Thurber: nonlinear regression Start 2

Case STAT-NIST-THURBER-START2-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/nls/data/LINKS/DATA/Thurber.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_THURBER.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/degrees_of_freedomexact30300.00E+2passed
/sselre5.6427082397E+035642.7082396670733.29E-811passed
/artifact/residual_variancelre188.090274664493414656188.09027465556918.92E-910passed
/artifact/estimates/0lre1.2881396800E+031288.1396799537594.62E-810passed
/artifact/covariance/0/0lre21.7603248414316766233621.7603248410342153.97E-1010passed
/artifact/estimates/1lre1.4910792535E+031491.07925354221564.22E-810passed
/artifact/covariance/1/1lre1565.8763939825748393961565.87639396210532.05E-810passed
/artifact/estimates/2lre5.8323836877E+02583.2383687666693.33E-911passed
/artifact/covariance/2/2lre823.615157954949994404823.61515792552382.94E-810passed
/artifact/estimates/3lre7.5416644291E+0175.416644291011211.12E-1112passed
/artifact/covariance/3/3lre30.9974685470159987290030.997468547165171.49E-1011passed
/artifact/estimates/4lre9.6629502864E-010.96629502864461664.62E-1211passed
/artifact/covariance/4/4lre0.0009817782386076096319690.0009817782386071754.35E-1612passed
/artifact/estimates/5lre3.9797285797E-010.39797285797390743.91E-1211passed
/artifact/covariance/5/5lre0.0002245480730992155272040.000224548073089930629.28E-1510passed
/artifact/estimates/6lre4.9727297349E-020.049727297348850591.49E-1311passed
/artifact/covariance/6/6lre0.000043352143454538970121294.335214345468464e-051.46E-1611passed

Notes

Convention


NIST Wampler1: fit

Case STAT-NIST-WAMPLER1-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Wampler1.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_WAMPLER1.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/observationsexact21210.00E+2passed
/responses/0/parameters/[term=Intercept]/estimatelre1.000000000000000.99999999998476851.52E-1110passed
/responses/0/parameters/[term=x]/estimatelre1.000000000000001.0000000000027852.79E-1211passed
/responses/0/parameters/[term=x^2]/estimatelre1.000000000000000.99999999999885041.15E-1211passed
/responses/0/parameters/[term=x^3]/estimatelre1.000000000000001.00000000000013521.35E-1312passed
/responses/0/parameters/[term=x^4]/estimatelre1.000000000000000.99999999999999396.11E-1514passed
/responses/0/parameters/[term=x^5]/estimatelre1.000000000000001.00000000000000022.22E-1615passed
/responses/0/parameterszero_residual_standard_errors[{'term': 'Intercept', 'estimate': 0.9999999999847685, 'std_passed
/responses/0/anova/[source=Error]/sslre0.0000000000000002.4067300062000036e-192.41E-19passed
/responses/0/metrics/r_squaredlre1.000000000000001.00.00E-12passed

Convention


NIST Wampler2: fit

Case STAT-NIST-WAMPLER2-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Wampler2.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_WAMPLER2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/observationsexact21210.00E+2passed
/responses/0/parameters/[term=Intercept]/estimatelre1.000000000000000.99999999999999982.22E-1615passed
/responses/0/parameters/[term=x]/estimatelre0.1000000000000000.100000000000000646.44E-1614passed
/responses/0/parameters/[term=x^2]/estimatelre0.100000000000000E-010.0099999999999996813.19E-1613passed
/responses/0/parameters/[term=x^3]/estimatelre0.100000000000000E-020.00100000000000005425.42E-1713passed
/responses/0/parameters/[term=x^4]/estimatelre0.100000000000000E-039.999999999999639e-053.61E-1813passed
/responses/0/parameters/[term=x^5]/estimatelre0.100000000000000E-041.000000000000008e-058.04E-2014passed
/responses/0/parameterszero_residual_standard_errors[{'term': 'Intercept', 'estimate': 0.9999999999999998, 'std_passed
/responses/0/anova/[source=Error]/sslre0.0000000000000001.9669194870171214e-291.97E-29passed
/responses/0/metrics/r_squaredlre1.000000000000001.00.00E-12passed

Convention


NIST Wampler3: fit

Case STAT-NIST-WAMPLER3-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Wampler3.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_WAMPLER3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/observationsexact21210.00E+2passed
/responses/0/parameters/[term=Intercept]/estimatelre1.000000000000000.9999999999957354.27E-1211passed
/responses/0/parameters/[term=Intercept]/std_errorlre2152.326246781702152.3262467811125.88E-1012passed
/responses/0/parameters/[term=x]/estimatelre1.000000000000001.00000000002508662.51E-1110passed
/responses/0/parameters/[term=x]/std_errorlre2363.551734696812363.5517346968584.81E-1113passed
/responses/0/parameters/[term=x^2]/estimatelre1.000000000000000.99999999999010469.90E-1211passed
/responses/0/parameters/[term=x^2]/std_errorlre779.343524331583779.34352433160061.76E-1113passed
/responses/0/parameters/[term=x^3]/estimatelre1.000000000000001.00000000000139641.40E-1211passed
/responses/0/parameters/[term=x^3]/std_errorlre101.475507550350101.475507550351581.58E-1213passed
/responses/0/parameters/[term=x^4]/estimatelre1.000000000000000.99999999999991918.09E-1413passed
/responses/0/parameters/[term=x^4]/std_errorlre5.645665121707525.6456651217076088.77E-1413passed
/responses/0/parameters/[term=x^5]/estimatelre1.000000000000001.00000000000000161.55E-1514passed
/responses/0/parameters/[term=x^5]/std_errorlre0.1123248546793120.112324854679313571.57E-1513passed
/responses/0/anova/[source=Error]/sslre83554268.000000083554268.000000282.83E-714passed
/responses/0/metrics/r_squaredlre0.9999955590258200.99999555902581996.44E-1716passed
/responses/0/anova/[source=Model]/f_statlre675524.458240122675524.45824012021.79E-914passed

Convention


NIST Wampler4: fit

Case STAT-NIST-WAMPLER4-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Wampler4.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_WAMPLER4.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/observationsexact21210.00E+2passed
/responses/0/parameters/[term=Intercept]/estimatelre1.000000000000000.99999999988488141.15E-109passed
/responses/0/parameters/[term=Intercept]/std_errorlre215232.624678170215232.624678110825.92E-812passed
/responses/0/parameters/[term=x]/estimatelre1.000000000000000.99999999922698887.73E-109passed
/responses/0/parameters/[term=x]/std_errorlre236355.173469681236355.173469685434.43E-913passed
/responses/0/parameters/[term=x^2]/estimatelre1.000000000000001.00000000042921384.29E-109passed
/responses/0/parameters/[term=x^2]/std_errorlre77934.352433158377934.352433159951.65E-913passed
/responses/0/parameters/[term=x^3]/estimatelre1.000000000000000.99999999993132536.87E-1110passed
/responses/0/parameters/[term=x^3]/std_errorlre10147.550755035010147.5507550351411.41E-1013passed
/responses/0/parameters/[term=x^4]/estimatelre1.000000000000001.0000000000042334.23E-1211passed
/responses/0/parameters/[term=x^4]/std_errorlre564.566512170752564.56651217075987.78E-1213passed
/responses/0/parameters/[term=x^5]/estimatelre1.000000000000000.9999999999999118.90E-1413passed
/responses/0/parameters/[term=x^5]/std_errorlre11.232485467931211.2324854679313391.39E-1313passed
/responses/0/anova/[source=Error]/sslre835542680000.000835542680000.00011.22E-415passed
/responses/0/metrics/r_squaredlre0.9574784408256620.95747844082566211.36E-1615passed
/responses/0/anova/[source=Model]/f_statlre67.552445824012267.552445824012222.43E-1415passed

Convention


NIST Wampler5: fit

Case STAT-NIST-WAMPLER5-FIT-001 · status: passed

Reference

Evidence type
NIST
Source
NIST Statistical Reference Datasets
Citation
https://itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Wampler5.dat
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/NIST_WAMPLER5.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/observationsexact21210.00E+2passed
/responses/0/parameters/[term=Intercept]/estimatelre1.000000000000000.99999999031348359.69E-98passed
/responses/0/parameters/[term=Intercept]/std_errorlre21523262.467817021523262.467811085.92E-612passed
/responses/0/parameters/[term=x]/estimatelre1.000000000000000.99999992433404687.57E-87passed
/responses/0/parameters/[term=x]/std_errorlre23635517.346968123635517.3469685434.43E-713passed
/responses/0/parameters/[term=x^2]/estimatelre1.000000000000001.00000004212297674.21E-87passed
/responses/0/parameters/[term=x^2]/std_errorlre7793435.243315837793435.2433159941.64E-713passed
/responses/0/parameters/[term=x^3]/estimatelre1.000000000000000.99999999326027146.74E-98passed
/responses/0/parameters/[term=x^3]/std_errorlre1014755.075503501014755.0755035141.40E-813passed
/responses/0/parameters/[term=x^4]/estimatelre1.000000000000001.00000000041503764.15E-109passed
/responses/0/parameters/[term=x^4]/std_errorlre56456.651217075256456.6512170759747.74E-1013passed
/responses/0/parameters/[term=x^5]/estimatelre1.000000000000000.9999999999912858.72E-1211passed
/responses/0/parameters/[term=x^5]/std_errorlre1123.248546793121123.24854679313381.38E-1113passed
/responses/0/anova/[source=Error]/sslre0.835542680000000E+168355426800000000.00.00E+2passed
/responses/0/metrics/r_squaredlre0.224668921574940E-020.00224668921574955331.53E-1613passed
/responses/0/anova/[source=Model]/f_statlre6.7552445824012241E-030.006755244582401224.31E-1815passed

Convention


Fixed approved limits replace the calculated ones

Case STAT-QUAL-APPROVED-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded I-MR charts and rules in R
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_stages.R · sha256 046de2272a987401
Dataset
reference_data/datasets/QUAL_SPC_STAGES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/upper_control_limittolerance24.60000000000000124.64.21E-1616passed
/primary/lower_control_limittolerance18.39999999999999918.44.21E-1616passed
/primary/center_linetolerance21.521.50.00E+1passed
/violation_keysset['primary:rule_1:11', 'primary:rule_1:17', 'primary:rule_1:2['primary:rule_1:11', 'primary:rule_1:17', 'primary:rule_1:2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 96  whole <- stage(d$x)
 97  approved <- list(list(chart = "individuals", center = 21.5, lower = 18.4, upper = 24.6),
 98                   list(chart = "moving_range", center = 1.0, lower = 0, upper = 3.3))
 99  keys <- c(rules_for(d$x, rep(21.5, 60), 1, rep(24.6, 60), rep(18.4, 60), 0L),
100            rules_for(whole$mr, rep(1, 59), 1, rep(3.3, 59), rep(0, 59), 0L, "secondary", FALSE))
101  emit("STAT-QUAL-APPROVED-001", approved_entry, "Fixed approved limits replace the calculated ones",
102       list(y_column = "x", x_column = "label", approved = approved),
103       list("/primary/upper_control_limit" = list(comparison = "tolerance", value = dec(24.6), absolute_tolerance = "0",
104              relative_tolerance = "1e-9", rationale = "The approved upper limit applies to every point."),
105            "/primary/lower_control_limit" = list(comparison = "tolerance", value = dec(18.4), absolute_tolerance = "0",
106              relative_tolerance = "1e-9", rationale = "The approved lower limit applies to every point."),
107            "/primary/center_line" = list(comparison = "tolerance", value = dec(21.5), absolute_tolerance = "0",
108              relative_tolerance = "1e-9", rationale = "The approved centre."),
109            "/violation_keys" = set_of(keys, "Rule 1 against the approved limits, rules 2-4 against the approved centre; rule 1 on the MR chart.")),
110       c("Approved values (individuals 18.4/21.5/24.6, moving range 0/1.0/3.3) are hand-chosen round numbers.",
111         "Rules 2-4 need only the approved centre, so the fixed basis needs no process sigma."),
112       kind = "axis_coverage")

Whole script generators/r/quality_spc_stages.R


Approved limits for the wrong chart are refused

Case STAT-QUAL-APPROVED-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded I-MR charts and rules in R
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_stages.R · sha256 046de2272a987401
Dataset
reference_data/datasets/QUAL_SPC_STAGES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

114  emit("STAT-QUAL-APPROVED-002", approved_entry, "Approved limits for the wrong chart are refused",
115       list(y_column = "x", x_column = "label", approved = list(list(chart = "xbar", center = 21, lower = 18, upper = 24))),
116       list(), c("A variables result needs approvals for the charts it plots."), kind = "invalid_input",
117       error = list(error_class = "ValueError", message_contains = "No approved limits are available"))

Whole script generators/r/quality_spc_stages.R


p chart with varying inspected counts

Case STAT-QUAL-ATTR-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.3.1-2, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc331.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc332.htm; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance0.0594451783355350040.0594451783355350041.45E-1917passed
/chart/point_upper_limitssequence['0.1290044393164344', '0.13184475260149978', '0.12834509494[0.1290044393164344, 0.13184475260149978, 0.1283450949478652passed
/chart/point_lower_limitssequence['0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0',[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,passed
/chart/points/0/valuetolerance0.0961538461538461590.096153846153846161.84E-1917passed
/chart/points/5/valuetolerance0.0588235294117647050.0588235294117647056.60E-2017passed
/violation_keysset['primary:rule_1:16']['primary:rule_1:16']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

121  laney_deferred <- list(list(path = "/dispersion_ratio",
122    reason = "The manual does not define the reported dispersion ratio; sigma_z is checked through the adjusted limits."))
123  emit_chart("STAT-QUAL-ATTR-001", "p chart with varying inspected counts",
124             list(count_column = "defectives", chart = "p", sample_size_column = "inspected"),
125             p_chart(a$defectives, a$inspected), c("Line 139-141; stepped limits for varying n_i (line 160-161)."))

Whole script generators/r/quality_spc_attributes.R


Laney p' chart

Case STAT-QUAL-ATTR-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Laney, Improved Control Charts for Attributes, Quality Engineering 14(4), 2002, pp. 531-537; NIST/SEMATECH e-Handbook 6.3.3
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance0.0594451783355350040.0594451783355350041.45E-1917passed
/chart/point_upper_limitssequence['0.15614894767927182', '0.16009765266707113', '0.1552323035[0.15614894767927182, 0.16009765266707113, 0.155232303585630passed
/chart/point_lower_limitssequence['0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0',[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,passed
/chart/points/0/valuetolerance0.0961538461538461590.096153846153846161.84E-1917passed
/chart/points/5/valuetolerance0.0588235294117647050.0588235294117647056.60E-2017passed
/violation_keysset['primary:rule_1:16']['primary:rule_1:16']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

126  ch <- p_chart(a$defectives, a$inspected, laney = TRUE)
127  emit_chart("STAT-QUAL-ATTR-002", "Laney p' chart", list(count_column = "defectives", chart = "laney_p", sample_size_column = "inspected"),
128             ch, c("sigma_z = mean |z_i - z_(i-1)| / d2(2) over standardized p values."), kind = "axis_coverage",
129             deferred = laney_deferred)

Whole script generators/r/quality_spc_attributes.R


c chart

Case STAT-QUAL-ATTR-003 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.3.1-2, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc331.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc332.htm; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance7.8333333333333337.8333333333333333.73E-1717passed
/chart/upper_control_limittolerance16.22976114520666616.2297611452066661.29E-1617passed
/chart/lower_control_limittolerance00.00.00E+2passed
/chart/points/8/valuetolerance2525.00.00E+2passed
/violation_keysset['primary:rule_1:8']['primary:rule_1:8']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

130  cbar <- mean(a$defects)
131  emit("STAT-QUAL-ATTR-003", attr_entry, "c chart", "QUAL_ATTR", list(count_column = "defects", chart = "c"),
132       list("/chart/center_line" = tol(cbar, "cbar."),
133            "/chart/upper_control_limit" = tol(cbar + 3 * sqrt(cbar), "cbar + 3 sqrt(cbar)."),
134            "/chart/lower_control_limit" = tol(max(0, cbar - 3 * sqrt(cbar)), "Clipped at 0."),
135            "/chart/points/8/value" = tol(a$defects[9], "The ninth count."),
136            "/violation_keys" = set_of(window_keys(a$defects, rep(cbar, k), rep(sqrt(cbar), k), rep(cbar + 3 * sqrt(cbar), k),
137                                                   rep(max(0, cbar - 3 * sqrt(cbar)), k)), "Rules 1-4 in R.")),
138       c("Line 144-145: cbar +/- 3 sqrt(cbar), clipped at 0."), kind = "axis_coverage")

Whole script generators/r/quality_spc_attributes.R


u chart with an area column

Case STAT-QUAL-ATTR-004 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.3.1-2, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc331.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc332.htm; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance4.25016277219127494.2501627721912771.79E-1515passed
/chart/point_upper_limitssequence['8.564540080336922', '9.2437352347215551', '8.2608436838305[8.564540080336926, 9.243735234721559, 8.260843683830597, 8.passed
/chart/point_lower_limitssequence['0', '0', '0.23948186055195553', '0', '0', '0.2875157000385[0.0, 0.0, 0.23948186055195642, 0.0, 0.0, 0.2875157000385373passed
/chart/points/0/valuetolerance5.35279805352798025.352798053527981.54E-1717passed
/chart/points/5/valuetolerance4.92610837438423674.9261083743842374.55E-1717passed
/violation_keysset['primary:rule_1:8']['primary:rule_1:8']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

139  emit_chart("STAT-QUAL-ATTR-004", "u chart with an area column", list(count_column = "defects", chart = "u", sample_size_column = "area"),
140             u_chart(a$defects, a$area), c("Line 146-148."), kind = "axis_coverage")

Whole script generators/r/quality_spc_attributes.R


Laney u' chart

Case STAT-QUAL-ATTR-005 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Laney, Improved Control Charts for Attributes, Quality Engineering 14(4), 2002, pp. 531-537; NIST/SEMATECH e-Handbook 6.3.3
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance4.25016277219127494.2501627721912771.79E-1515passed
/chart/point_upper_limitssequence['10.00147955174851', '10.906886336785874', '9.5966345286161[10.001479551748513, 10.906886336785877, 9.596634528616164, passed
/chart/point_lower_limitssequence['0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0',[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,passed
/chart/points/0/valuetolerance5.35279805352798025.352798053527981.54E-1717passed
/chart/points/5/valuetolerance4.92610837438423674.9261083743842374.55E-1717passed
/violation_keysset['primary:rule_1:8']['primary:rule_1:8']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

141  ch <- u_chart(a$defects, a$area, laney = TRUE)
142  emit_chart("STAT-QUAL-ATTR-005", "Laney u' chart", list(count_column = "defects", chart = "laney_u", sample_size_column = "area"),
143             ch, c("sigma_z from standardized u values."), kind = "axis_coverage",
144             deferred = laney_deferred)

Whole script generators/r/quality_spc_attributes.R


np chart with a constant size of 100

Case STAT-QUAL-ATTR-006 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.3.1-2, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc331.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc332.htm; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR_NP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance88.00.00E+2passed
/chart/upper_control_limittolerance16.1387959797503216.138795979750321.34E-1718passed
/chart/lower_control_limittolerance00.00.00E+2passed
/violation_keysset[][]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

147  set.seed(20261007)
148  dn <- rbinom(k, 100, 0.08)
149  export_dataset(data.frame(lot = seq_len(k), defectives = as.integer(dn)), "QUAL_ATTR_NP",
150                 "Constructed: rbinom(30, 100, 0.08), seed 20261007", "R stats::rbinom documentation",
151                 "Constructed validation data; no third-party license.", c(lot = "Int64", defectives = "Int64"))
152  an <- read_dataset("QUAL_ATTR_NP")
153  pbar <- sum(an$defectives) / (100 * k); s <- sqrt(100 * pbar * (1 - pbar))
154  emit("STAT-QUAL-ATTR-006", attr_entry, "np chart with a constant size of 100", "QUAL_ATTR_NP",
155       list(count_column = "defectives", chart = "np", sample_size = 100),
156       list("/chart/center_line" = tol(100 * pbar, "n pbar."),
157            "/chart/upper_control_limit" = tol(100 * pbar + 3 * s, "n pbar + 3 sqrt(n pbar (1 - pbar))."),
158            "/chart/lower_control_limit" = tol(max(0, 100 * pbar - 3 * s), "Clipped at 0."),
159            "/violation_keys" = set_of(window_keys(an$defectives, rep(100 * pbar, k), rep(s, k), rep(100 * pbar + 3 * s, k),
160                                                   rep(max(0, 100 * pbar - 3 * s), k)), "Rules 1-4 in R.")),
161       c("Line 142-143."), kind = "axis_coverage")

Whole script generators/r/quality_spc_attributes.R


p chart with a Phase I baseline of 20 lots

Case STAT-QUAL-ATTR-007 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.3.1-2, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc331.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc332.htm; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance0.066463723997970570.066463723997970573.76E-1917passed
/chart/point_upper_limitssequence['0.13973985326469524', '0.14273193743148355', '0.1390452771[0.13973985326469524, 0.14273193743148355, 0.139045277121975passed
/chart/point_lower_limitssequence['0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0',[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,passed
/chart/points/0/valuetolerance0.0961538461538461590.096153846153846161.84E-1917passed
/chart/points/5/valuetolerance0.0588235294117647050.0588235294117647056.60E-2017passed
/violation_keysset['primary:rule_1:16', 'primary:rule_2:25', 'primary:rule_2:2['primary:rule_1:16', 'primary:rule_2:25', 'primary:rule_2:2passed
/baseline_countexact20200.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

163  bp <- p_chart(a$defectives, a$inspected, base = a$phase <= 1)
164  emit("STAT-QUAL-ATTR-007", attr_entry, "p chart with a Phase I baseline of 20 lots", "QUAL_ATTR",
165       list(count_column = "defectives", chart = "p", sample_size_column = "inspected",
166            baseline = list(column = "phase", op = "le", value = 1)),
167       c(chart_outputs(bp$values, bp$center, bp$ucl, bp$lcl, bp$zone, "From the baseline pbar."),
168         list("/baseline_count" = exact(20L, "Twenty baseline lots."))),
169       c("pbar from the first 20 lots; all 30 are evaluated (line 167-168)."), kind = "axis_coverage")

Whole script generators/r/quality_spc_attributes.R


Missing counts are omitted and excluded lots removed

Case STAT-QUAL-ATTR-008 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.3.1-2, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc331.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc332.htm; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance0.0566473988439306350.0566473988439306352.93E-1917passed
/chart/point_upper_limitssequence['0.12742773976401811', '0.12400635276914609', '0.1207617624[0.1274277397640181, 0.12400635276914608, 0.1207617624381226passed
/chart/point_lower_limitssequence['0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0',[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,passed
/chart/points/0/valuetolerance0.031250.031250.00E-3passed
/chart/points/5/valuetolerance0.0833333333333333290.083333333333333332.93E-1917passed
/violation_keysset['primary:rule_1:13']['primary:rule_1:13']passed
/subgroup_countexact26260.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

171  m <- read_dataset("QUAL_ATTR"); m$defectives[c(5, 23)] <- NA
172  export_dataset(m, "QUAL_ATTR_MISSING", "QUAL_ATTR with defectives missing in rows 5 and 23 (1-based)",
173                 "R stats::rbinom documentation", "Constructed validation data; no third-party license.",
174                 c(lot = "Int64", defectives = "Int64", inspected = "Int64", defects = "Int64", area = "Float64", phase = "Int64"))
175  mm <- read_dataset("QUAL_ATTR_MISSING")
176  excluded <- c(0L, 11L)
177  keep <- setdiff(which(!is.na(mm$defectives)), excluded + 1L)
178  mp <- p_chart(mm$defectives[keep], mm$inspected[keep])
179  emit("STAT-QUAL-ATTR-008", attr_entry, "Missing counts are omitted and excluded lots removed", "QUAL_ATTR_MISSING",
180       list(count_column = "defectives", chart = "p", sample_size_column = "inspected"),
181       c(chart_outputs(mp$values, mp$center, mp$ucl, mp$lcl, mp$zone, ""),
182         list("/subgroup_count" = exact(length(keep), "Retained lots."))),
183       c("Line 164-165: rows with null counts are omitted and counted; excluded lots 0 and 11 (0-based) are removed first."),
184       kind = "missing_excluded", excluded = excluded)

Whole script generators/r/quality_spc_attributes.R


g chart plots the counts between events

Case STAT-QUAL-ATTR-009 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Benneyan, Number-between g-type statistical quality control charts for monitoring adverse events, Health Care Management Science 4(4), 2001, pp. 305-318; NIST/SEMATECH e-Handbook 6.3.3
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/points/0/valuetolerance1111.00.00E+2passed
/chart/points/8/valuetolerance2525.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

186  emit("STAT-QUAL-ATTR-009", attr_entry, "g chart plots the counts between events", "QUAL_ATTR",
187       list(count_column = "defects", chart = "g"),
188       list("/chart/points/0/value" = tol(a$defects[1], "The plotted statistic is the count itself."),
189            "/chart/points/8/value" = tol(a$defects[9], "The plotted statistic is the count itself.")),
190       c("Line 152-153 say only that g limits are probability based under a geometric model."), kind = "axis_coverage",
191       deferred = list(list(path = "/chart/center_line", reason = "The manual does not define the g chart's centre or probability limits."),
192                       list(path = "/chart/upper_control_limit", reason = "The manual does not define the g chart's probability limits.")))

Whole script generators/r/quality_spc_attributes.R


Defectives above the sample size are refused

Case STAT-QUAL-ATTR-010 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.3.1-2, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc331.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc332.htm; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR_OVER.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

194  big <- read_dataset("QUAL_ATTR"); big$defectives[3] <- big$inspected[3] + 1L
195  export_dataset(big, "QUAL_ATTR_OVER", "QUAL_ATTR with defectives exceeding the inspected count in row 3 (1-based)",
196                 "Hand-constructed", "Constructed validation data; no third-party license.",
197                 c(lot = "Int64", defectives = "Int64", inspected = "Int64", defects = "Int64", area = "Float64", phase = "Int64"))
198  emit("STAT-QUAL-ATTR-010", attr_entry, "Defectives above the sample size are refused", "QUAL_ATTR_OVER",
199       list(count_column = "defectives", chart = "p", sample_size_column = "inspected"), list(),
200       c("Line 166: binomial counts cannot exceed sample size."), kind = "invalid_input",
201       error = list(error_class = "ValueError", message_contains = "exceed"))

Whole script generators/r/quality_spc_attributes.R


np with a varying size column is refused

Case STAT-QUAL-ATTR-011 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.3.1-2, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc331.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc332.htm; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_ATTR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

202  emit("STAT-QUAL-ATTR-011", attr_entry, "np with a varying size column is refused", "QUAL_ATTR",
203       list(count_column = "defectives", chart = "np", sample_size_column = "inspected"), list(),
204       c("Line 142 and 167: np requires a constant size."), kind = "invalid_input",
205       error = list(error_class = "ValueError", message_contains = "constant"))

Whole script generators/r/quality_spc_attributes.R


Binomial defectives with a Wilson interval

Case STAT-QUAL-ATTRCAP-001 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_ATTR_CAP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/ratetolerance0.0119562976705833850.0119562976705833851.07E-1917passed
/ppmtolerance11956.29767058338411956.2976705833843.88E-1316passed
/yield_percenttolerance98.80437023294166598.804370232941663.47E-1617passed
/ci_lowertolerance0.00926081400174614060.0092608140017461421.72E-1815passed
/ci_uppertolerance0.0154241236707201410.0154241236707201392.20E-1815passed
/sigma_leveltolerance2.2585306593877832.2585306593877833.16E-1716passed
/total_eventsexact58580.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

76  emit("STAT-QUAL-ATTRCAP-001", attr_entry, "capabilityMenuItem", "Binomial defectives with a Wilson interval", "QUAL_ATTR_CAP",
77       list(family = "binomial", event_column = "defectives", denominator_column = "inspected"),
78       binomial_outputs(a$defectives, a$inspected, 0.95),
79       c("The sigma level is taken as the normal quantile of the yield without a 1.5 shift; the manual says only 'where defined'."))

Whole script generators/r/quality_capability_counts_batches.R


Binomial at a 90% confidence level

Case STAT-QUAL-ATTRCAP-002 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_ATTR_CAP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/ci_lowertolerance0.00964784201096863420.0096478420109686341.04E-2017passed
/ci_uppertolerance0.0148088421433032590.0148088421433032592.85E-2017passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

80  emit("STAT-QUAL-ATTRCAP-002", attr_entry, "capabilityMenuItem", "Binomial at a 90% confidence level", "QUAL_ATTR_CAP",
81       list(family = "binomial", event_column = "defectives", denominator_column = "inspected", confidence_level = 0.9),
82       binomial_outputs(a$defectives, a$inspected, 0.9)[c("/ci_lower", "/ci_upper")],
83       c("A non-default confidence level (the engine API accepts it; the dialog fixes 95%)."), kind = "axis_coverage")

Whole script generators/r/quality_capability_counts_batches.R


Poisson defects with an exposure column

Case STAT-QUAL-ATTRCAP-003 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_ATTR_CAP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/ratetolerance2.1978021978021982.1978021978021981.24E-1717passed
/yield_percenttolerance11.10469495916230411.1046949591623043.53E-1616passed
/ci_lowertolerance1.86846597363482061.86846597363482061.28E-1717passed
/ci_uppertolerance2.56847000860024942.56847000860025039.21E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

92  emit("STAT-QUAL-ATTRCAP-003", attr_entry, "capabilityMenuItem", "Poisson defects with an exposure column", "QUAL_ATTR_CAP",
93       list(family = "poisson", event_column = "defects", denominator_column = "exposure"),
94       poisson_outputs(a$defects, a$exposure, 0.95), c("Line 173-176."), kind = "axis_coverage")

Whole script generators/r/quality_capability_counts_batches.R


Poisson defects with unit exposure

Case STAT-QUAL-ATTRCAP-004 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_ATTR_CAP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/ratetolerance6.32000000000000036.321.58E-1717passed
/ci_lowertolerance5.37296075378429015.372960753784294.45E-1717passed
/ci_uppertolerance7.3858923567308777.385892356730882.63E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

95  emit("STAT-QUAL-ATTRCAP-004", attr_entry, "capabilityMenuItem", "Poisson defects with unit exposure", "QUAL_ATTR_CAP",
96       list(family = "poisson", event_column = "defects"),
97       poisson_outputs(a$defects, rep(1, k), 0.95)[c("/rate", "/ci_lower", "/ci_upper")],
98       c("With the exposure column off, each row has unit exposure (line 161-162)."), kind = "axis_coverage")

Whole script generators/r/quality_capability_counts_batches.R


Missing and excluded rows are omitted

Case STAT-QUAL-ATTRCAP-005 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_ATTR_CAP_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/ratetolerance0.0127015144113336590.0127015144113336593.39E-1916passed
/ci_lowertolerance0.00969938268557152090.0096993826855715231.71E-1815passed
/ci_uppertolerance0.0166172671664043260.0166172671664043261.16E-1917passed
/dropped_rowsexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 99  m <- read_dataset("QUAL_ATTR_CAP"); m$defectives[c(4, 19)] <- NA
100  export_dataset(m, "QUAL_ATTR_CAP_MISSING", "QUAL_ATTR_CAP with defectives missing in rows 4 and 19 (1-based)",
101                 "R stats::rbinom documentation", "Constructed validation data; no third-party license.",
102                 c(defectives = "Int64", inspected = "Int64", defects = "Int64", exposure = "Float64"))
103  mm <- read_dataset("QUAL_ATTR_CAP_MISSING")
104  excluded <- c(0L, 10L)
105  keep <- setdiff(which(!is.na(mm$defectives)), excluded + 1L)
106  emit("STAT-QUAL-ATTRCAP-005", attr_entry, "capabilityMenuItem", "Missing and excluded rows are omitted", "QUAL_ATTR_CAP_MISSING",
107       list(family = "binomial", event_column = "defectives", denominator_column = "inspected"),
108       c(binomial_outputs(mm$defectives[keep], mm$inspected[keep], 0.95)[c("/rate", "/ci_lower", "/ci_upper")],
109         list("/dropped_rows" = exact(2L, "Two rows with a missing count are disclosed and omitted."))),
110       c("Rows 0 and 10 (0-based) are excluded before the call."), kind = "missing_excluded", excluded = excluded)

Whole script generators/r/quality_capability_counts_batches.R


Defectives above inspected are refused

Case STAT-QUAL-ATTRCAP-006 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_ATTR_CAP_OVER.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

111  over <- read_dataset("QUAL_ATTR_CAP"); over$defectives[2] <- over$inspected[2] + 1L
112  export_dataset(over, "QUAL_ATTR_CAP_OVER", "QUAL_ATTR_CAP with more defectives than inspected in row 2 (1-based)",
113                 "Hand-constructed", "Constructed validation data; no third-party license.",
114                 c(defectives = "Int64", inspected = "Int64", defects = "Int64", exposure = "Float64"))
115  emit("STAT-QUAL-ATTRCAP-006", attr_entry, "capabilityMenuItem", "Defectives above inspected are refused", "QUAL_ATTR_CAP_OVER",
116       list(family = "binomial", event_column = "defectives", denominator_column = "inspected"), list(),
117       c("Line 159-160."), kind = "invalid_input",
118       error = list(error_class = "ValueError", message_contains = "cannot exceed inspected counts"))

Whole script generators/r/quality_capability_counts_batches.R


Binomial without an inspected column is refused

Case STAT-QUAL-ATTRCAP-007 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_ATTR_CAP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

119  emit("STAT-QUAL-ATTRCAP-007", attr_entry, "capabilityMenuItem", "Binomial without an inspected column is refused", "QUAL_ATTR_CAP",
120       list(family = "binomial", event_column = "defectives"), list(),
121       c("Line 159: inspected counts are required for Binomial."), kind = "invalid_input",
122       error = list(error_class = "ValueError", message_contains = "inspected-count column"))

Whole script generators/r/quality_capability_counts_batches.R


Within and overall capability of balanced batches

Case STAT-QUAL-BATCH-001 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_BATCH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/within_capability/sigmatolerance0.840848891497948610.84084889149795152.89E-1514passed
/within_capability/cpktolerance1.51473411320196491.5147341132019683.15E-1514passed
/within_capability/cpk_intervalinterval[0.68580159139771, 2.01382598571698]passed
/within_capability/cptolerance1.58569910338828831.58569910338828285.51E-1514passed
/within_capability/cp_intervalinterval[1.1696129051716113, 2.01382598571698]passed
/overall_capability/sigmatolerance1.03535952348784771.03535952348784836.21E-1615passed
/overall_capability/cpktolerance1.2301644705110491.23016447051105536.26E-1514passed
/overall_capability/cpk_intervalinterval[0.38505677331520866, 2.01382598571698]passed
/overall_capability/cptolerance1.28779742986445121.28779742986445066.40E-1615passed
/overall_capability/cp_intervalinterval[0.6567021379686924, 2.01382598571698]passed
/overall_meantolerance50.17901249999999950.179012499999982.14E-1415passed
/total_nexact48480.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

171  set.seed(20261016)
172  G <- 8L; nb <- 6L
173  batch <- rep(sprintf("B%d", seq_len(G)), each = nb)
174  yv <- round(50 + rep(rnorm(G, 0, 0.6), each = nb) + rnorm(G * nb, 0, 0.9), 4)
175  export_dataset(data.frame(batch = batch, y = yv), "QUAL_BATCH", "Constructed: 8 balanced batches of 6, between SD 0.6, within SD 0.9, seed 20261016",
176                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.", c(batch = "String", y = "Float64"))
177  b <- read_dataset("QUAL_BATCH")
178  rb <- batch_ref(b$y, b$batch, 46, 54)
179  if (rb$s2b <= 0) stop("expected a positive between-batch component")
180  emit("STAT-QUAL-BATCH-001", batch_entry, "batchCapabilityMenuItem", "Within and overall capability of balanced batches", "QUAL_BATCH",
181       list(response_column = "y", batch_column = "batch", lower_limit = 46, upper_limit = 54),
182       c(batch_outputs(rb, "within"), batch_outputs(rb, "overall"),
183         list("/overall_mean" = tol(rb$mu, "Analyzed grand mean."), "/total_n" = exact(48L, "48 analyzed rows."))),
184       c("Balanced batches: REML equals the ANOVA estimator, s2_e = MSW and s2_b = (MSB - MSW) / n."))

Whole script generators/r/quality_capability_counts_batches.R


Upper limit only at a 90% level

Case STAT-QUAL-BATCH-002 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_BATCH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/within_capability/sigmatolerance0.840848891497948610.84084889149795152.89E-1514passed
/within_capability/cpktolerance1.51473411320196491.5147341132019683.15E-1514passed
/within_capability/cpk_intervalinterval[0.8029206734865707, 2.4502471616626513]passed
/within_capability/cpabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

185  emit("STAT-QUAL-BATCH-002", batch_entry, "batchCapabilityMenuItem", "Upper limit only at a 90% level", "QUAL_BATCH",
186       list(response_column = "y", batch_column = "batch", upper_limit = 54, confidence_level = 0.9),
187       c(batch_outputs(batch_ref(b$y, b$batch, NULL, 54, 0.9), "within", FALSE),
188         list("/within_capability/cp" = list(comparison = "absent", rationale = "Cp needs both limits."))),
189       c("One-sided upper specification; Cpk = CPU."), kind = "axis_coverage")

Whole script generators/r/quality_capability_counts_batches.R


A zero between-batch component makes within and overall coincide

Case STAT-QUAL-BATCH-003 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_BATCH_FLAT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/within_capability/sigmatolerance0.899793052968703380.89979305296870331.14E-1615passed
/overall_capability/sigmatolerance0.899793052968703380.89979305296870331.14E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

190  set.seed(20261017)
191  flat <- round(50 + rnorm(G * nb, 0, 0.9), 4)
192  flat <- flat - ave(flat, batch) + mean(flat) + rep(seq(-0.02, 0.02, length.out = G), each = nb)
193  export_dataset(data.frame(batch = batch, y = round(flat, 6)), "QUAL_BATCH_FLAT",
194                 "Constructed: batch means forced nearly equal, seed 20261017", "R stats::rnorm documentation",
195                 "Constructed validation data; no third-party license.", c(batch = "String", y = "Float64"))
196  bf <- read_dataset("QUAL_BATCH_FLAT")
197  rf <- batch_ref(bf$y, bf$batch, 46, 54)
198  if (rf$s2b != 0) stop("expected MSB < MSW")
199  emit("STAT-QUAL-BATCH-003", batch_entry, "batchCapabilityMenuItem", "A zero between-batch component makes within and overall coincide", "QUAL_BATCH_FLAT",
200       list(response_column = "y", batch_column = "batch", lower_limit = 46, upper_limit = 54),
201       list("/within_capability/sigma" = tol(rf$within$sigma, "Constrained REML at the zero boundary: sqrt(SST / (N - 1)) = sd(y)."),
202            "/overall_capability/sigma" = tol(rf$overall$sigma, "Chapter 43: at a zero optimum total and within capability coincide.")),
203       c("MSB < MSW, so the constrained REML optimum has zero between-batch variance (issue #2 ruling).",
204         "Until 2026-09-24 this case expected sqrt(MSW), the truncated-ANOVA convention; the statistician ruled for REML."),
205       kind = "boundary")

Whole script generators/r/quality_capability_counts_batches.R


No specification limit is refused

Case STAT-QUAL-BATCH-004 · status: passed

Reference

Evidence type
R
Source
R stats::prop.test, stats::poisson.test and one-way ANOVA mean squares
Citation
R stats::prop.test (correct = FALSE gives the Wilson interval), stats::poisson.test and stats::aov documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_capability_counts_batches.R · sha256 67f13e823884ff37
Dataset
reference_data/datasets/QUAL_BATCH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

206  emit("STAT-QUAL-BATCH-004", batch_entry, "batchCapabilityMenuItem", "No specification limit is refused", "QUAL_BATCH",
207       list(response_column = "y", batch_column = "batch"), list(),
208       c("Line 6: at least one specification limit."), kind = "invalid_input",
209       error = list(error_class = "ValueError", message_contains = "At least one specification limit"))

Whole script generators/r/quality_capability_counts_batches.R


Between/within I-MR-R chart by subgroup column

Case STAT-QUAL-BETWEEN-WITHIN-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas with stats::integrate constants; nlme::lme REML components
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 (I-MR, R and S chart constants); Pinheiro & Bates (2000) nlme; R stats::integrate
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_BW_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stages/0/subgroupsexact660.00E+2passed
/stages/0/observationsexact30300.00E+2passed
/stages/0/grand_meantolerance10.71110.7112.98E-1616passed
/stages/0/subgroup_mean_centertolerance10.71110.7112.98E-1616passed
/stages/0/mean_moving_rangetolerance2.04280000000000022.04280000000000151.30E-1515passed
/stages/0/i_chart_scaletolerance1.81038436331489421.81038436331489551.30E-1515passed
/stages/0/within_sigmatolerance0.537419684062672380.5374196840626736.62E-1614passed
/individuals/points/0/centertolerance10.71110.7112.98E-1616passed
/individuals/points/0/uppertolerance16.14215308994468516.1421530899446857.14E-1818passed
/individuals/points/0/lowertolerance5.27984691005531745.2798469100553143.59E-1515passed
/individuals/points/0/valuetolerance10.11399999999999910.1139999999999998.97E-1818passed
/within/points/0/centertolerance1.25000000000000041.254.00E-1615passed
/within/points/0/uppertolerance2.64312393136984942.64312393136986051.11E-1414passed
/within/points/0/lowertolerance00.00.00E+2passed
/stages/0/components/df_betweenexact550.00E+2passed
/stages/0/components/df_withinexact24240.00E+2passed
/stages/0/components/ms_betweentolerance8.04947800000000068.0494780000000111.06E-1414passed
/stages/0/components/ms_withintolerance0.275095000000000140.275095000000000261.16E-1615passed
/stages/0/components/n0tolerance55.00.00E+2passed
/stages/0/components/between_sigmatolerance1.24694691146014721.24694691146014724.48E-1716passed
/stages/0/components/within_component_sigmatolerance0.524494995209678020.52449499520967811.08E-1615passed
/stages/0/components/total_sigmatolerance1.35276442886409451.3527644288640954.81E-1615passed
/moving_range/points/0/valuetolerance1.76800000000000071.76800000000000251.76E-1515passed
/moving_range/points/0/centertolerance2.04280000000000022.04280000000000151.30E-1515passed
/moving_range/points/0/uppertolerance6.67287140472275556.672871404722764.47E-1515passed
/moving_range/points/0/subgroupexact110.00E+2passed
/subgroups/2/meantolerance9.47199999999999959.4723.10E-1717passed
/subgroups/2/rangetolerance0.780000000000001140.78000000000000113.13E-1817passed
/subgroups/2/standard_deviationtolerance0.33877721292908730.33877721292908731.33E-1817passed
/subgroups/2/sizeexact550.00E+2passed
/source_countexact30300.00E+2passed
/excluded_countexact000.00E+2passed
/missing_countexact000.00E+2passed
/dropped_subgroup_countexact000.00E+2passed
/individuals/points/4/beyond_limitsexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 96  set.seed(20261010)
 97  batch_means <- c(10.4, 11.9, 9.1, 10.8, 12.6, 9.7)
 98  bw_value <- round(rep(batch_means, each = 5) + rnorm(30, 0, 0.5), 2)
 99  lot_sizes <- c(4, 5, 6, 5, 4, 6)
100  bw <- data.frame(value = bw_value, batch = rep(1:6, each = 5), stage = rep(c("A", "B"), each = 15),
101                   lot = rep(sprintf("L%d", 1:6), times = lot_sizes), stringsAsFactors = FALSE)
102  export_dataset(bw, "QUAL_BW_01",
103                 "Simulated (R 4.6.1, set.seed 20261010): six batches of five coating measurements with batch-to-batch shifts; lot gives unequal subgroup sizes 4-6; 30 rows",
104                 "Simulated with stats::rnorm", "Constructed validation data; no third-party license.",
105                 c(value = "Float64", batch = "Int64", stage = "String", lot = "String"))
106  d <- read_dataset("QUAL_BW_01")
107  bw_ep <- "tabelz.analysis.spc_between_within:compute_between_within"
108  bw_menu <- "betweenWithinChartMenuItem"
109  bw_source <- "Hand-coded chart formulas with stats::integrate constants; nlme::lme REML components"
110  bw_cite <- "Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 (I-MR, R and S chart constants); Pinheiro & Bates (2000) nlme; R stats::integrate"
111  bw_req <- "REQ-STAT-QUAL-BETWEEN-WITHIN-001"

Whole script generators/r/quality_subgroup_capability.R


Between/within I-MR-S chart

Case STAT-QUAL-BETWEEN-WITHIN-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas with stats::integrate constants; nlme::lme REML components
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 (I-MR, R and S chart constants); Pinheiro & Bates (2000) nlme; R stats::integrate
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_BW_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stages/0/subgroupsexact660.00E+2passed
/stages/0/observationsexact30300.00E+2passed
/stages/0/grand_meantolerance10.71110.7112.98E-1616passed
/stages/0/subgroup_mean_centertolerance10.71110.7112.98E-1616passed
/stages/0/mean_moving_rangetolerance2.04280000000000022.04280000000000151.30E-1515passed
/stages/0/i_chart_scaletolerance1.81038436331489421.81038436331489551.30E-1515passed
/stages/0/within_sigmatolerance0.535531373643903570.53553137364390364.33E-1817passed
/individuals/points/0/centertolerance10.71110.7112.98E-1616passed
/individuals/points/0/uppertolerance16.14215308994468516.1421530899446857.14E-1818passed
/individuals/points/0/lowertolerance5.27984691005531745.2798469100553143.59E-1515passed
/individuals/points/0/valuetolerance10.11399999999999910.1139999999999998.97E-1818passed
/within/points/0/centertolerance0.503391781172920380.50339178117292051.14E-1615passed
/within/points/0/uppertolerance1.0515843579562331.05158435795623254.93E-1615passed
/within/points/0/lowertolerance00.00.00E+2passed
/stages/0/components/df_betweenexact550.00E+2passed
/stages/0/components/df_withinexact24240.00E+2passed
/stages/0/components/ms_betweentolerance8.04947800000000068.0494780000000111.06E-1414passed
/stages/0/components/ms_withintolerance0.275095000000000140.275095000000000261.16E-1615passed
/stages/0/components/n0tolerance55.00.00E+2passed
/stages/0/components/between_sigmatolerance1.24694691146014721.24694691146014724.48E-1716passed
/stages/0/components/within_component_sigmatolerance0.524494995209678020.52449499520967811.08E-1615passed
/stages/0/components/total_sigmatolerance1.35276442886409451.3527644288640954.81E-1615passed
/within/points/2/valuetolerance0.33877721292908730.33877721292908731.33E-1817passed
/moving_range/points/0/uppertolerance6.67287140472275556.672871404722764.47E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

196  s2 <- bw_stage(groups, "stddev")
197  out <- c(stage_outputs(s2, 0L, "stddev", 0L, comp, "1e-6"),
198           list("/within/points/2/value" = num(sd(groups[[3]]), "Sample SD of batch 3 on the S chart."),
199                "/moving_range/points/0/upper" = num(s2$mr_upper, "D4(2) MR-bar.")))
200  emit("STAT-QUAL-BETWEEN-WITHIN-002", bw_ep, bw_menu, "Between/within I-MR-S chart", "QUAL_BW_01",
201       list(response = "value", subgroup_mode = "column", subgroup_column = "batch", dispersion = "stddev"), out,
202       c(bw_notes, "S chart: within sigma is the mean of s_i/c4(n_i); limits (c4 +/- 3 sqrt(1 - c4^2)) sigma_w; the I and MR charts are unchanged."),
203       bw_source, bw_cite, bw_req, kind = "axis_coverage")

Whole script generators/r/quality_subgroup_capability.R


Between/within chart with consecutive subgroups of five and two stages

Case STAT-QUAL-BETWEEN-WITHIN-003 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas with stats::integrate constants; nlme::lme REML components
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 (I-MR, R and S chart constants); Pinheiro & Bates (2000) nlme; R stats::integrate
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_BW_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stages/1/labelexactBBpassed
/stages/1/first_subgroupexact330.00E+2passed
/subgroups/3/labelexact44passed
/moving_range/points/2/subgroupexact440.00E+2passed
/stages/0/subgroupsexact330.00E+2passed
/stages/0/observationsexact15150.00E+2passed
/stages/0/grand_meantolerance10.48933333333333310.4893333333333312.04E-1515passed
/stages/0/subgroup_mean_centertolerance10.48933333333333310.4893333333333351.51E-1515passed
/stages/0/mean_moving_rangetolerance2.08900000000000042.0890000000000021.79E-1515passed
/stages/0/i_chart_scaletolerance1.85132804727081181.85132804727081331.55E-1515passed
/stages/0/within_sigmatolerance0.560349590582679860.56034959058268067.74E-1614passed
/individuals/points/0/centertolerance10.48933333333333310.4893333333333351.51E-1515passed
/individuals/points/0/uppertolerance16.0433174751457716.0433174751457777.00E-1515passed
/individuals/points/0/lowertolerance4.93534919152089734.9353491915208943.50E-1515passed
/individuals/points/0/valuetolerance10.11399999999999910.1139999999999998.97E-1818passed
/within/points/0/centertolerance1.30333333333333411.3033333333333342.01E-1615passed
/within/points/0/uppertolerance2.75589721910829692.7558972191083091.20E-1414passed
/within/points/0/lowertolerance00.00.00E+2passed
/stages/0/components/df_betweenexact220.00E+2passed
/stages/0/components/df_withinexact12120.00E+2passed
/stages/0/components/ms_betweentolerance7.78840666666666967.7884066666666821.24E-1414passed
/stages/0/components/ms_withintolerance0.325823333333333630.325823333333333631.23E-1817passed
/stages/0/components/n0tolerance55.00.00E+2passed
/stages/0/components/between_sigmatolerance1.22168599348059461.22168599348059442.32E-1615passed
/stages/0/components/within_component_sigmatolerance0.570809366893478390.57080936689347842.92E-1817passed
/stages/0/components/total_sigmatolerance1.34845837903881961.34845837903881942.23E-1615passed
/moving_range/points/0/uppertolerance6.82378517939388916.8237851793938956.18E-1515passed
/stages/1/subgroupsexact330.00E+2passed
/stages/1/observationsexact15150.00E+2passed
/stages/1/grand_meantolerance10.93266666666666610.9326666666666681.87E-1515passed
/stages/1/subgroup_mean_centertolerance10.93266666666666610.9326666666666681.87E-1515passed
/stages/1/mean_moving_rangetolerance2.29399999999999962.29400000000000141.77E-1515passed
/stages/1/i_chart_scaletolerance2.03300456698862672.0330045669886281.38E-1515passed
/stages/1/within_sigmatolerance0.514489777542664780.51448977754266524.48E-1615passed
/individuals/points/3/centertolerance10.93266666666666610.9326666666666681.87E-1515passed
/individuals/points/3/uppertolerance17.03168036763254617.0316803676325537.44E-1515passed
/individuals/points/3/lowertolerance4.83365296570078674.8336529657007833.52E-1515passed
/individuals/points/3/valuetolerance10.9210.927.11E-1717passed
/within/points/3/centertolerance1.19666666666666651.19666666666666578.43E-1615passed
/within/points/3/uppertolerance2.53035064363140142.53035064363141079.31E-1514passed
/within/points/3/lowertolerance00.00.00E+2passed
/stages/1/components/df_betweenexact220.00E+2passed
/stages/1/components/df_withinexact12120.00E+2passed
/stages/1/components/ms_betweentolerance11.59824666666666311.5982466666666751.23E-1414passed
/stages/1/components/ms_withintolerance0.224366666666666740.224366666666666838.52E-1715passed
/stages/1/components/n0tolerance55.00.00E+2passed
/stages/1/components/between_sigmatolerance1.50823605579498031.5082360557949816.90E-1615passed
/stages/1/components/within_component_sigmatolerance0.47367358662550180.473673586625501855.23E-1715passed
/stages/1/components/total_sigmatolerance1.58086769423208411.58086769423208454.08E-1615passed
/moving_range/points/2/uppertolerance7.49342422284805037.4934242228480565.35E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

215  emit("STAT-QUAL-BETWEEN-WITHIN-003", bw_ep, bw_menu, "Between/within chart with consecutive subgroups of five and two stages", "QUAL_BW_01",
216       list(response = "value", subgroup_mode = "size", subgroup_size = 5L, stage_column = "stage", dispersion = "range"), out,
217       c(bw_notes, "Consecutive rows of five form subgroups 1-6; stage A holds subgroups 1-3 and stage B 4-6; every stage estimates its own centre, scales and components (chapter 84 line 35: stage boundaries restart the calculations).",
218         "Both stages are balanced with positive between components (checked), so the ANOVA closed form is the REML solution."),
219       bw_source, bw_cite, bw_req, kind = "axis_coverage")

Whole script generators/r/quality_subgroup_capability.R


Between/within chart with unequal lot sizes and an excluded row

Case STAT-QUAL-BETWEEN-WITHIN-004 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas with stats::integrate constants; nlme::lme REML components
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 (I-MR, R and S chart constants); Pinheiro & Bates (2000) nlme; R stats::integrate
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_BW_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stages/0/subgroupsexact660.00E+2passed
/stages/0/observationsexact29290.00E+2passed
/stages/0/grand_meantolerance10.74827586206896510.7482758620689651.25E-1616passed
/stages/0/subgroup_mean_centertolerance10.82122222222222210.8212222222222231.41E-1515passed
/stages/0/mean_moving_rangetolerance1.80623333333333361.80623333333333426.46E-1615passed
/stages/0/i_chart_scaletolerance1.60073261365028691.60073261365028756.17E-1615passed
/stages/0/within_sigmatolerance0.794438752507558020.79443875250755921.22E-1514passed
/individuals/points/0/centertolerance10.82122222222222210.8212222222222231.41E-1515passed
/individuals/points/0/uppertolerance15.62342006317308215.6234200631730863.74E-1515passed
/individuals/points/0/lowertolerance6.01902438127136116.0190243812713612.06E-1717passed
/individuals/points/0/valuetolerance10.12249999999999910.1224999999999992.79E-1616passed
/within/points/0/centertolerance1.63555137438254291.63555137438254344.88E-1615passed
/within/points/0/uppertolerance3.73241256767711123.73241256767712451.33E-1414passed
/within/points/0/lowertolerance00.00.00E+2passed
/stages/0/components/df_betweenexact550.00E+2passed
/stages/0/components/df_withinexact23230.00E+2passed
/stages/0/components/ms_betweentolerance5.94064209195402445.9406420919540316.19E-1514passed
/stages/0/components/ms_withintolerance0.692939275362319250.69293927536231985.56E-1615passed
/stages/0/components/n0tolerance4.81379310344827624.8137931034482764.39E-1717passed
/stages/0/components/between_sigmatolerance1.06796498069927371.06796498077624447.70E-1110passed
/stages/0/components/within_component_sigmatolerance0.833012244807961990.83301222991912331.49E-87passed
/stages/0/components/total_sigmatolerance1.35442186928593271.35442186018951829.10E-98passed
/within/points/5/uppertolerance4.03458245153586334.0345824515358781.51E-1414passed
/within/points/5/centertolerance2.01343568058764032.01343568058764033.07E-1716passed
/within/points/5/lowertolerance00.00.00E+2passed
/subgroups/2/sizeexact550.00E+2passed
/excluded_countexact110.00E+2passed
/source_countexact30300.00E+2passed
/stages/0/subgroup_size_minexact440.00E+2passed
/stages/0/subgroup_size_maxexact660.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

223  keep <- rep(TRUE, nrow(d)); keep[11] <- FALSE  # 0-based row 10 (lot L3) excluded
224  groups4 <- groups_of(d$value[keep], d$lot[keep])
225  s4 <- bw_stage(groups4, "range")
226  lme4 <- nlme::lme(value ~ 1, random = ~ 1 | lot, data = data.frame(value = d$value[keep], lot = factor(d$lot[keep], levels = unique(d$lot))),
227                    method = "REML", control = nlme::lmeControl(tolerance = 1e-10, msTol = 1e-12, niterEM = 100, msMaxIter = 500, opt = "optim"))
228  vc4 <- nlme::VarCorr(lme4)
229  comp4 <- anova_components(groups4)
230  comp4$between <- as.numeric(vc4[1, 1]); comp4$within <- as.numeric(vc4[2, 1])
231  out <- stage_outputs(s4, 0L, "range", 0L, comp4, "1e-5")
232  wl6 <- w_limits(6, s4$within, "range")
233  out <- c(out, list("/within/points/5/upper" = num(wl6[["upper"]], "(d2(6) + 3 d3(6)) sigma_w for the six-value lot."),
234                     "/within/points/5/center" = num(wl6[["center"]], "d2(6) sigma_w."),
235                     "/within/points/5/lower" = num(wl6[["lower"]], "max(0, d2(6) - 3 d3(6)) sigma_w = 0 for n = 6."),
236                     "/subgroups/2/size" = exact(5L, "Lot L3 keeps five values after the exclusion."),
237                     "/excluded_count" = exact(1L, "One excluded row."), "/source_count" = exact(30L, "Rows in the table."),
238                     "/stages/0/subgroup_size_min" = exact(4L, "Smallest lot."), "/stages/0/subgroup_size_max" = exact(6L, "Largest lot.")))
239  emit("STAT-QUAL-BETWEEN-WITHIN-004", bw_ep, bw_menu, "Between/within chart with unequal lot sizes and an excluded row", "QUAL_BW_01",
240       list(response = "value", subgroup_mode = "column", subgroup_column = "lot", dispersion = "range"), out,
241       c(bw_notes, "Unequal sizes 4-6: within sigma averages R_i/d2(n_i) and the R-chart limits change with n_i (chapter 84 lines 25-26, 31-33).",
242         "Unbalanced components come from nlme::lme(value ~ 1, random = ~1 | lot, method = 'REML'); tolerance 1e-5 on the component SDs because both programs stop at a numerical REML optimum. MS between/within and n0 are the descriptive ANOVA quantities.",
243         "Row index 10 (0-based) is excluded, so lot L3 has five values; the I chart is centred at the unweighted mean of the six lot means."),
244       bw_source, bw_cite, bw_req, kind = "axis_coverage", excluded = 10L)

Whole script generators/r/quality_subgroup_capability.R


Bissell Table 3/4 capability: overall sigma

Case STAT-QUAL-BISSELL-001 · status: passed

Reference

Evidence type
R
Source
R recomputation from Bissell (1990), Table 3
Citation
Bissell (1990), How Reliable is Your Capability Index?, Applied Statistics 39(3), 331-340, Tables 3-4, pp. 338-339
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_bissell_capability.R · sha256 6c6439542e744990
Dataset
reference_data/datasets/QUAL_BISSELL_TABLE3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows/0/nexact75750.00E+2passed
/rows/0/meantolerance5.61359999999999995.6136000000000019.11E-1615passed
/rows/0/stdtolerance0.0483131675856817160.0483131675856817164.57E-1917passed
/rows/0/ppktolerance0.852769587647780320.85276958764778646.11E-1514passed
/rows/0/ppk_ci_lowertolerance0.696032468763810420.69603246876381565.22E-1514passed
/rows/0/ppk_ci_uppertolerance1.00950670653175021.00950670653175716.91E-1514passed
/rows/0/uncertainty_methodcontains['Ppk nu=N-1']two-sided 0.95 large-sample normal-index approximation, SE =passed

Convention

Reference code

Whole script generators/r/quality_bissell_capability.R


Bissell Table 3/4 capability: range sigma

Case STAT-QUAL-BISSELL-002 · status: passed

Reference

Evidence type
R
Source
R recomputation from Bissell (1990), Table 3
Citation
Bissell (1990), How Reliable is Your Capability Index?, Applied Statistics 39(3), 331-340, Tables 3-4, pp. 338-339
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_bissell_capability.R · sha256 6c6439542e744990
Dataset
reference_data/datasets/QUAL_BISSELL_TABLE3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows/0/nexact75750.00E+2passed
/rows/0/meantolerance5.61359999999999995.6136000000000019.11E-1615passed
/rows/0/stdtolerance0.0483131675856817160.0483131675856817164.57E-1917passed
/rows/0/ppktolerance0.852769587647780320.85276958764778646.11E-1514passed
/rows/0/ppk_ci_lowertolerance0.696032468763810420.69603246876381565.22E-1514passed
/rows/0/ppk_ci_uppertolerance1.00950670653175021.00950670653175716.91E-1514passed
/rows/0/uncertainty_methodcontains['Cpk nu=54.34308553', 'Ppk nu=N-1']two-sided 0.95 large-sample normal-index approximation, SE =passed
/rows/0/within_sigmatolerance0.0495859228495159480.0495859228495160045.57E-1714passed
/rows/0/cpktolerance0.830880976543166170.83088097654317125.00E-1514passed
/rows/0/cpk_ci_lowertolerance0.657411685397982890.65741168539798683.89E-1514passed
/rows/0/cpk_ci_uppertolerance1.00435026768834961.00435026768835565.96E-1514passed

Convention

Reference code

Whole script generators/r/quality_bissell_capability.R


Bissell Table 3/4 capability: stddev sigma

Case STAT-QUAL-BISSELL-003 · status: passed

Reference

Evidence type
R
Source
R recomputation from Bissell (1990), Table 3
Citation
Bissell (1990), How Reliable is Your Capability Index?, Applied Statistics 39(3), 331-340, Tables 3-4, pp. 338-339
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_bissell_capability.R · sha256 6c6439542e744990
Dataset
reference_data/datasets/QUAL_BISSELL_TABLE3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows/0/nexact75750.00E+2passed
/rows/0/meantolerance5.61359999999999995.6136000000000019.11E-1615passed
/rows/0/stdtolerance0.0483131675856817160.0483131675856817164.57E-1917passed
/rows/0/ppktolerance0.852769587647780320.85276958764778646.11E-1514passed
/rows/0/ppk_ci_lowertolerance0.696032468763810420.69603246876381565.22E-1514passed
/rows/0/ppk_ci_uppertolerance1.00950670653175021.00950670653175716.91E-1514passed
/rows/0/uncertainty_methodcontains['Cpk nu=56.91801074', 'Ppk nu=N-1']two-sided 0.95 large-sample normal-index approximation, SE =passed
/rows/0/within_sigmatolerance0.049558997876299610.04955899787629967.21E-1815passed
/rows/0/cpktolerance0.831332386963030250.83133238696303656.21E-1514passed
/rows/0/cpk_ci_lowertolerance0.661000198241762150.66100019824176755.33E-1514passed
/rows/0/cpk_ci_uppertolerance1.00166457568429831.00166457568430547.15E-1514passed

Convention

Reference code

Whole script generators/r/quality_bissell_capability.R


Bissell Table 3/4 capability: pooled sigma

Case STAT-QUAL-BISSELL-004 · status: passed

Reference

Evidence type
R
Source
R recomputation from Bissell (1990), Table 3
Citation
Bissell (1990), How Reliable is Your Capability Index?, Applied Statistics 39(3), 331-340, Tables 3-4, pp. 338-339
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_bissell_capability.R · sha256 6c6439542e744990
Dataset
reference_data/datasets/QUAL_BISSELL_TABLE3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows/0/nexact75750.00E+2passed
/rows/0/meantolerance5.61359999999999995.6136000000000019.11E-1615passed
/rows/0/stdtolerance0.0483131675856817160.0483131675856817164.57E-1917passed
/rows/0/ppktolerance0.852769587647780320.85276958764778646.11E-1514passed
/rows/0/ppk_ci_lowertolerance0.696032468763810420.69603246876381565.22E-1514passed
/rows/0/ppk_ci_uppertolerance1.00950670653175021.00950670653175716.91E-1514passed
/rows/0/uncertainty_methodcontains['Cpk nu=60', 'Ppk nu=N-1']two-sided 0.95 large-sample normal-index approximation, SE =passed
/rows/0/within_sigmatolerance0.0483390801181266820.048339080118126684.04E-1917passed
/rows/0/cpktolerance0.852312454008621210.85231245400862736.11E-1514passed
/rows/0/cpk_ci_lowertolerance0.682177654826066520.68217765482607165.10E-1514passed
/rows/0/cpk_ci_uppertolerance1.02244725319117591.0224472531911837.11E-1514passed

Convention

Reference code

Whole script generators/r/quality_bissell_capability.R


Normal capability with individual observations

Case STAT-QUAL-CAPABILITY-001 · status: passed

Reference

Evidence type
R
Source
R stats and manual formulas
Citation
R stats::mean and stats::sd documentation; Tabelz manual chapter 08, Capability continuous data
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_normal_capability.R · sha256 eb3f29c277d1830a
Dataset
reference_data/datasets/QUAL_CAPABILITY_NORMAL_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows/0/nexact30300.00E+2passed
/rows/0/meantolerance9.9888127359128349.9888127359128361.81E-1515passed
/rows/0/stdtolerance0.346648162022056120.34664816202205613.38E-1817passed
/rows/0/within_sigmatolerance0.203833908696091330.203833908696091331.56E-1817passed
/rows/0/cptolerance2.12591386833197232.12591386833197233.26E-1716passed
/rows/0/cpktolerance2.10761912996920442.1076191299692072.68E-1514passed
/rows/0/pptolerance1.25006672703997191.2500667270399723.32E-1817passed
/rows/0/ppktolerance1.23930916811537541.23930916811537721.77E-1514passed
/rows/0/oosexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

11  index <- seq_len(30L)
12  raw <- data.frame(value = 10 + 0.41 * sin(index * 0.61) +
13                              0.24 * cos(index * 1.19) + 0.07 * sin(index * 2.31))
14  export_dataset(raw, "QUAL_CAPABILITY_NORMAL_01", "Constructed ordered continuous measurements",
15                 "R stats::mean and stats::sd documentation; manual chapter 08 capability formulas",
16                 "Constructed validation data; no third-party license.", c(value = "Float64"))
17  d <- read_dataset("QUAL_CAPABILITY_NORMAL_01")
18  x <- d$value
19  lsl <- 8.7
20  usl <- 11.3
21  d2 <- 2 / sqrt(pi)
22  mu <- mean(x)
23  overall <- stats::sd(x)
24  mr_bar <- mean(abs(diff(x)))
25  within <- mr_bar / d2
26  cp <- (usl - lsl) / (6 * within)
27  cpk <- min((mu - lsl) / (3 * within), (usl - mu) / (3 * within))
28  pp <- (usl - lsl) / (6 * overall)
29  ppk <- min((mu - lsl) / (3 * overall), (usl - mu) / (3 * overall))
30  id <- "STAT-QUAL-CAPABILITY-001"

Whole script generators/r/quality_normal_capability.R


Compare capability of two production lines

Case STAT-QUAL-CAPABILITY-COMPARISON-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded manual capability formulas with stats::pnorm, qnorm, shapiro.test and stats::integrate constants
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 8 (Cp, Cpk, Pp, Ppk); NIST Dataplot Cpm; Royston (1995) shapiro.test; R stats documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_SIXPACK_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/populations/0/populationexactAApassed
/populations/0/rowsexact20200.00E+2passed
/populations/0/excludedexact000.00E+2passed
/populations/0/capability/nexact20200.00E+2passed
/populations/0/capability/meantolerance9.91934999999999979.9193499999999981.81E-1515passed
/populations/0/capability/stdtolerance0.233352291836320060.233352291836319842.24E-1615passed
/populations/0/capability/within_sigmatolerance0.247537173335673020.24753717333567291.15E-1615passed
/populations/0/capability/cptolerance1.61591891274277181.61591891274277246.40E-1615passed
/populations/0/capability/cpktolerance1.50731569581551741.50731569581551581.57E-1514passed
/populations/0/capability/pptolerance1.71414643864124171.71414643864124331.58E-1515passed
/populations/0/capability/ppktolerance1.59894151341089441.59894151341089331.10E-1515passed
/populations/0/capability/cpmtolerance1.62011429190550691.62011429190550452.41E-1514passed
/populations/0/capability/cpk_ci_lowertolerance0.878988566211608590.87898856621160778.87E-1614passed
/populations/0/capability/cpk_ci_uppertolerance2.1356428254194262.1356428254194241.82E-1515passed
/populations/0/capability/ppk_ci_lowertolerance1.06998768246406641.06998768246406576.74E-1615passed
/populations/0/capability/ppk_ci_uppertolerance2.12789534435772242.12789534435772071.75E-1515passed
/populations/0/capability/dpmotolerance0.826322712513305580.82632271242166779.16E-119passed
/populations/0/capability/within_ppmtolerance3.1785026478394623.1785026478395327.02E-1413passed
/populations/0/capability/normality_ptail_probability0.441465036018377050.44146503601838598.83E-1513passed
/populations/0/capability/oosexact000.00E+2passed
/populations/1/populationexactBBpassed
/populations/1/rowsexact20200.00E+2passed
/populations/1/excludedexact000.00E+2passed
/populations/1/capability/nexact20200.00E+2passed
/populations/1/capability/meantolerance10.24060000000000110.24064.09E-1616passed
/populations/1/capability/stdtolerance0.234364852762794590.234364852762794535.56E-1715passed
/populations/1/capability/within_sigmatolerance0.270065994651130070.270065994651130071.92E-1918passed
/populations/1/capability/cptolerance1.48111945939998061.48111945939998062.42E-1716passed
/populations/1/capability/cpktolerance1.18415500779028361.18415500779028361.99E-1817passed
/populations/1/capability/pptolerance1.70674055979224781.7067405597922482.35E-1615passed
/populations/1/capability/ppktolerance1.3645390775539011.36453907755390122.47E-1615passed
/populations/1/capability/cpmtolerance1.19090219959849161.19090219959849192.62E-1615passed
/populations/1/capability/cpk_ci_lowertolerance0.6823305255596690.6823305255596691.12E-1817passed
/populations/1/capability/cpk_ci_uppertolerance1.68597949002089821.68597949002089822.86E-1817passed
/populations/1/capability/ppk_ci_lowertolerance0.906751382869183110.90675138286918321.10E-1615passed
/populations/1/capability/ppk_ci_uppertolerance1.82232677223861891.82232677223861922.64E-1615passed
/populations/1/capability/dpmotolerance21.23515487211592821.2351548721789746.30E-1111passed
/populations/1/capability/within_ppmtolerance190.86784552668396190.867845526683682.79E-1314passed
/populations/1/capability/normality_ptail_probability0.417048967290293670.41704896729029935.61E-1513passed
/populations/1/capability/oosexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

495  cmp_ep <- "tabelz.analysis.capability_sixpack:compute_capability_comparison"
496  cmp_req <- "REQ-STAT-QUAL-CAPABILITY-COMPARISON-001"

Whole script generators/r/quality_subgroup_capability.R


Compare capability before and after a split point with subgroups of five

Case STAT-QUAL-CAPABILITY-COMPARISON-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded manual capability formulas with stats::pnorm, qnorm, shapiro.test and stats::integrate constants
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 8 (Cp, Cpk, Pp, Ppk); NIST Dataplot Cpm; Royston (1995) shapiro.test; R stats documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_SIXPACK_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/populations/0/populationexactbefore 21before 21passed
/populations/0/rowsexact20200.00E+2passed
/populations/0/capability/nexact20200.00E+2passed
/populations/0/capability/meantolerance9.91934999999999979.9193499999999981.81E-1515passed
/populations/0/capability/stdtolerance0.233352291836320060.233352291836319842.24E-1615passed
/populations/0/capability/within_sigmatolerance0.239259243344701560.239259243344701721.62E-1615passed
/populations/0/capability/cptolerance1.67182673659014511.67182673659014381.30E-1515passed
/populations/0/capability/cpktolerance1.55946604800181521.55946604800181173.54E-1514passed
/populations/0/capability/pptolerance1.71414643864124171.71414643864124331.58E-1515passed
/populations/0/capability/ppktolerance1.59894151341089441.59894151341089331.10E-1515passed
/populations/0/capability/cpmabsentNonepassed
/populations/0/capability/cpk_ci_lowertolerance0.973228370332438250.97322837033243136.99E-1514passed
/populations/0/capability/cpk_ci_uppertolerance2.14570372567119222.1457037256711923.64E-1716passed
/populations/0/capability/ppk_ci_lowertolerance1.06998768246406641.06998768246406576.74E-1615passed
/populations/0/capability/ppk_ci_uppertolerance2.12789534435772242.12789534435772071.75E-1515passed
/populations/0/capability/dpmotolerance0.826322712513305580.82632271242166779.16E-119passed
/populations/0/capability/within_ppmtolerance1.48898237389866091.48898237389873117.02E-1413passed
/populations/0/capability/normality_ptail_probability0.441465036018377050.44146503601838598.83E-1513passed
/populations/0/capability/oosexact000.00E+2passed
/populations/0/capability/subgroup_countexact440.00E+2passed
/populations/1/populationexactfrom 21from 21passed
/populations/1/rowsexact20200.00E+2passed
/populations/1/capability/nexact20200.00E+2passed
/populations/1/capability/meantolerance10.24060000000000110.24064.09E-1616passed
/populations/1/capability/stdtolerance0.234364852762794590.234364852762794535.56E-1715passed
/populations/1/capability/within_sigmatolerance0.239366727281514420.239366727281514985.59E-1614passed
/populations/1/capability/cptolerance1.6710760285808971.6710760285808934.01E-1514passed
/populations/1/capability/cpktolerance1.33602528485042621.3360252848504233.16E-1514passed
/populations/1/capability/pptolerance1.70674055979224781.7067405597922482.35E-1615passed
/populations/1/capability/ppktolerance1.3645390775539011.36453907755390122.47E-1615passed
/populations/1/capability/cpmabsentNonepassed
/populations/1/capability/cpk_ci_lowertolerance0.8281632142650730.8281632142650676.00E-1514passed
/populations/1/capability/cpk_ci_uppertolerance1.84388735543577931.8438873554357792.15E-1615passed
/populations/1/capability/ppk_ci_lowertolerance0.906751382869183110.90675138286918321.10E-1615passed
/populations/1/capability/ppk_ci_uppertolerance1.82232677223861891.82232677223861922.64E-1615passed
/populations/1/capability/dpmotolerance21.23515487211592821.2351548721789746.30E-1111passed
/populations/1/capability/within_ppmtolerance30.60861091630757630.6086109163088341.26E-1213passed
/populations/1/capability/normality_ptail_probability0.417048967290293670.41704896729029935.61E-1513passed
/populations/1/capability/oosexact000.00E+2passed
/populations/1/capability/subgroup_countexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

524  emit("STAT-QUAL-CAPABILITY-COMPARISON-002", cmp_ep, "capabilityComparisonMenuItem", "Compare capability before and after a split point with subgroups of five", "QUAL_SIXPACK_01",
525       list(column = "value", lsl = lsl, usl = usl, split_kind = "point", split_column = "index", split_value = 21, method = "normal", subgroup_mode = "constant", subgroup_size = 5L, confidence_level = 0.95),
526       out, c(cap_notes, "docs/manual/chapters/85-capability-views.tex:34-37: populations before and from the declared split point 21 on the index column; no target so Cpm is unavailable.",
527              "Within each population consecutive rows of five form subgroups: s_w = mean(R_i/d2(5))."),
528       cap_source, cap_cite, cmp_req, kind = "axis_coverage")

Whole script generators/r/quality_subgroup_capability.R


Capability of two characteristics with their own limits

Case STAT-QUAL-CAPABILITY-MULTIPLE-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded manual capability formulas with stats::pnorm, qnorm, shapiro.test and stats::integrate constants
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 8 (Cp, Cpk, Pp, Ppk); NIST Dataplot Cpm; Royston (1995) shapiro.test; R stats documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_SIXPACK_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows/0/nexact40400.00E+2passed
/rows/0/meantolerance10.07997499999999910.0799752.41E-1616passed
/rows/0/stdtolerance0.282399924411626480.282399924411626371.12E-1615passed
/rows/0/within_sigmatolerance0.256278647827082210.25627864782708223.41E-1816passed
/rows/0/cptolerance1.5608011178125541.5608011178125543.97E-1716passed
/rows/0/cpktolerance1.45678022664833891.4567802266483394.95E-1716passed
/rows/0/pptolerance1.41643097402872981.41643097402873046.16E-1615passed
/rows/0/ppktolerance1.32203175140544091.32203175140544167.02E-1615passed
/rows/0/cpmtolerance1.36283457785801691.36283457785801756.37E-1615passed
/rows/0/cpk_ci_lowertolerance1.03009902926282871.03009902926282872.61E-1817passed
/rows/0/cpk_ci_uppertolerance1.88346142403384921.88346142403384923.61E-1817passed
/rows/0/ppk_ci_lowertolerance1.01098936224844781.01098936224844834.64E-1615passed
/rows/0/ppk_ci_uppertolerance1.63307414056243411.6330741405624358.39E-1615passed
/rows/0/dpmotolerance39.44445087871426439.4444508786673254.69E-1111passed
/rows/0/within_ppmtolerance6.49763582539419826.497635825394198.04E-1514passed
/rows/0/normality_ptail_probability0.820858017758414050.8208580177584216.99E-1514passed
/rows/0/oosexact000.00E+2passed
/rows/1/nexact40400.00E+2passed
/rows/1/meantolerance2.01719999999999992.01721.80E-1717passed
/rows/1/stdtolerance0.0829343886485745810.082934388648574582.95E-1917passed
/rows/1/within_sigmatolerance0.0703982311551960140.070398231155196013.75E-1917passed
/rows/1/cptolerance1.18374186347979831.18374186347979833.81E-1716passed
/rows/1/cpktolerance1.10230042327238881.10230042327238884.81E-1716passed
/rows/1/pptolerance1.00481036505193561.00481036505193564.99E-1716passed
/rows/1/ppktolerance0.935679411936362950.9356794119363631.95E-1817passed
/rows/1/cpmabsentNonepassed
/rows/1/cpk_ci_lowertolerance0.772455769684860180.77245576968486022.88E-1817passed
/rows/1/cpk_ci_uppertolerance1.43214507685991731.43214507685991732.09E-1716passed
/rows/1/ppk_ci_lowertolerance0.703756077918781480.70375607791878154.77E-1817passed
/rows/1/ppk_ci_uppertolerance1.16760274595394441.16760274595394442.09E-1716passed
/rows/1/dpmotolerance3136.85219730665633136.8521973065837.32E-1113passed
/rows/1/within_ppmtolerance545.32894552608229545.32894552608166.79E-1314passed
/rows/1/normality_ptail_probability0.153434836840451060.153434836840459948.88E-1513passed
/rows/1/oosexact110.00E+2passed
/rows/0/columnexactvaluevaluepassed
/rows/1/columnexactthicknessthicknesspassed
/rows/1/cpm_noteexactCpm needs both specification limits and a target.Cpm needs both specification limits and a target.passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

531  mul_ep <- "tabelz.analysis.capability_sixpack:compute_capability_multiple"
532  mul_req <- "REQ-STAT-QUAL-CAPABILITY-MULTIPLE-001"

Whole script generators/r/quality_subgroup_capability.R


Capability of two characteristics by lot with one-sided limits at 99% confidence

Case STAT-QUAL-CAPABILITY-MULTIPLE-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded manual capability formulas with stats::pnorm, qnorm, shapiro.test and stats::integrate constants
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 8 (Cp, Cpk, Pp, Ppk); NIST Dataplot Cpm; Royston (1995) shapiro.test; R stats documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_SIXPACK_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/rows/0/nexact40400.00E+2passed
/rows/0/meantolerance10.07997499999999910.0799752.41E-1616passed
/rows/0/stdtolerance0.282399924411626480.282399924411626371.12E-1615passed
/rows/0/within_sigmatolerance0.239312985313107990.239312985313108353.61E-1614passed
/rows/0/cpabsentNonepassed
/rows/0/cpktolerance1.56005603364230591.56005603364230372.18E-1514passed
/rows/0/ppabsentNonepassed
/rows/0/ppktolerance1.32203175140544091.32203175140544167.02E-1615passed
/rows/0/cpmabsentNonepassed
/rows/0/cpk_ci_lowertolerance1.01507449072544991.0150744907254445.81E-1514passed
/rows/0/cpk_ci_uppertolerance2.1050375765591622.10503757655916331.35E-1515passed
/rows/0/ppk_ci_lowertolerance0.91325276007899570.91325276007899614.40E-1615passed
/rows/0/ppk_ci_uppertolerance1.73081074273188621.7308107427318878.64E-1615passed
/rows/0/dpmotolerance36.52983719433764536.5298371942657247.19E-1111passed
/rows/0/within_ppmtolerance1.43319897347570561.4331989734757464.04E-1413passed
/rows/0/normality_ptail_probability0.820858017758414050.8208580177584216.99E-1514passed
/rows/0/oosexact000.00E+2passed
/rows/1/nexact40400.00E+2passed
/rows/1/meantolerance2.01719999999999992.01721.80E-1717passed
/rows/1/stdtolerance0.0829343886485745810.082934388648574582.95E-1917passed
/rows/1/within_sigmatolerance0.0787857256835877150.078785725683587841.25E-1614passed
/rows/1/cpabsentNonepassed
/rows/1/cpktolerance1.13049243240289911.1304924324028972.00E-1514passed
/rows/1/ppabsentNonepassed
/rows/1/ppktolerance1.07394131816750841.07394131816750842.58E-1716passed
/rows/1/cpmabsentNonepassed
/rows/1/cpk_ci_lowertolerance0.724642527868169850.72464252786816544.44E-1514passed
/rows/1/cpk_ci_uppertolerance1.53634233693762831.53634233693762884.83E-1615passed
/rows/1/ppk_ci_lowertolerance0.732565545335876810.73256554533587689.95E-1917passed
/rows/1/ppk_ci_uppertolerance1.41531709099913991.415317090999143.73E-1716passed
/rows/1/dpmotolerance636.88686861114172636.88686861113071.10E-1113passed
/rows/1/within_ppmtolerance347.58461076648001347.584610766486666.65E-1213passed
/rows/1/normality_ptail_probability0.153434836840451060.153434836840459948.88E-1513passed
/rows/1/oosexact000.00E+2passed
/rows/0/subgroup_countexact880.00E+2passed
/rows/1/subgroup_countexact880.00E+2passed
/rows/0/confidence_leveltolerance0.989999999999999990.991.12E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

546  g_value <- groups_of(cp_d$value, cp_d$lot); g_thick <- groups_of(cp_d$thickness, cp_d$lot)
547  r1 <- capability(cp_d$value, NULL, usl, NULL, level = 0.99, groups = g_value)
548  r2 <- capability(cp_d$thickness, 1.75, NULL, NULL, level = 0.99, groups = g_thick)
549  out <- c(cap_outputs(r1, "/rows/0"), cap_outputs(r2, "/rows/1"),
550           list("/rows/0/subgroup_count" = exact(8L, "Eight lots."), "/rows/1/subgroup_count" = exact(8L, "Eight lots."),
551                "/rows/0/confidence_level" = num(0.99, "Configured level.")))
552  emit("STAT-QUAL-CAPABILITY-MULTIPLE-002", mul_ep, "capabilityMultipleMenuItem", "Capability of two characteristics by lot with one-sided limits at 99% confidence", "QUAL_SIXPACK_01",
553       list(columns = list("value", "thickness"), lsls = list("", "1.75"), usls = list("11.2", ""),
554            method = "normal", subgroup_mode = "by_column", group_column = "lot", confidence_level = 0.99),
555       out, c(cap_notes, "One-sided limits per column: value has only USL (Cpk = CPU) and thickness only LSL (Cpk = CPL); Cp, Pp and Cpm are unavailable; lots of five give s_w = mean(R_i/d2(5)); 99% intervals with z = qnorm(0.995)."),
556       cap_source, cap_cite, mul_req, kind = "axis_coverage")

Whole script generators/r/quality_subgroup_capability.R


Capability sixpack of ordered individuals with a target

Case STAT-QUAL-CAPABILITY-SIXPACK-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded manual capability formulas with stats::pnorm, qnorm, shapiro.test and stats::integrate constants
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 8 (Cp, Cpk, Pp, Ppk); NIST Dataplot Cpm; Royston (1995) shapiro.test; R stats documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_SIXPACK_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/capability/nexact40400.00E+2passed
/capability/meantolerance10.07997499999999910.0799752.41E-1616passed
/capability/stdtolerance0.282399924411626480.282399924411626371.12E-1615passed
/capability/within_sigmatolerance0.256278647827082210.25627864782708223.41E-1816passed
/capability/cptolerance1.5608011178125541.5608011178125543.97E-1716passed
/capability/cpktolerance1.45678022664833891.4567802266483394.95E-1716passed
/capability/pptolerance1.41643097402872981.41643097402873046.16E-1615passed
/capability/ppktolerance1.32203175140544091.32203175140544167.02E-1615passed
/capability/cpmtolerance1.36283457785801691.36283457785801756.37E-1615passed
/capability/cpk_ci_lowertolerance1.03009902926282871.03009902926282872.61E-1817passed
/capability/cpk_ci_uppertolerance1.88346142403384921.88346142403384923.61E-1817passed
/capability/ppk_ci_lowertolerance1.01098936224844781.01098936224844834.64E-1615passed
/capability/ppk_ci_uppertolerance1.63307414056243411.6330741405624358.39E-1615passed
/capability/dpmotolerance39.44445087871426439.4444508786673254.69E-1111passed
/capability/within_ppmtolerance6.49763582539419826.497635825394198.04E-1514passed
/capability/normality_ptail_probability0.820858017758414050.8208580177584216.99E-1514passed
/capability/oosexact000.00E+2passed
/within_spread/0tolerance9.31113905651875319.3111390565187533.45E-1717passed
/within_spread/1tolerance10.84881094348124510.8488109434812454.17E-1616passed
/overall_spread/0tolerance9.23277522676512069.232775226765123.53E-1717passed
/overall_spread/1tolerance10.92717477323487810.9271747732348781.53E-1616passed
/overall_spread_labelexactmean +/- 3 overall SDmean +/- 3 overall SDpassed
/chart_familyexactI-MRI-MRpassed
/primary/0/centertolerance10.07997499999999910.0799752.41E-1616passed
/primary/0/uppertolerance10.84881094348124510.8488109434812454.17E-1616passed
/primary/0/lowertolerance9.31113905651875319.3111390565187533.45E-1717passed
/secondary/0/centertolerance0.289179487179487220.28917948717948724.29E-1816passed
/secondary/0/uppertolerance0.94461402527530380.94461402527530383.57E-1817passed
/run_plot_subgroupsexact25250.00E+2passed
/excluded_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

432  six_ep <- "tabelz.analysis.capability_sixpack:compute_capability_sixpack"
433  six_req <- "REQ-STAT-QUAL-CAPABILITY-SIXPACK-001"

Whole script generators/r/quality_subgroup_capability.R


Capability sixpack with consecutive subgroups of five at 90% confidence

Case STAT-QUAL-CAPABILITY-SIXPACK-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded manual capability formulas with stats::pnorm, qnorm, shapiro.test and stats::integrate constants
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 8 (Cp, Cpk, Pp, Ppk); NIST Dataplot Cpm; Royston (1995) shapiro.test; R stats documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_SIXPACK_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/capability/nexact40400.00E+2passed
/capability/meantolerance10.07997499999999910.0799752.41E-1616passed
/capability/stdtolerance0.282399924411626480.282399924411626371.12E-1615passed
/capability/within_sigmatolerance0.239312985313107990.239312985313108353.61E-1614passed
/capability/cptolerance1.67145129829313271.67145129829313042.26E-1514passed
/capability/cpktolerance1.56005603364230591.56005603364230372.18E-1514passed
/capability/pptolerance1.41643097402872981.41643097402873046.16E-1615passed
/capability/ppktolerance1.32203175140544091.32203175140544167.02E-1615passed
/capability/cpmtolerance1.36283457785801691.36283457785801756.37E-1615passed
/capability/cpk_ci_lowertolerance1.21204583504205471.212045835042054.71E-1514passed
/capability/cpk_ci_uppertolerance1.90806623224255721.90806623224255742.47E-1615passed
/capability/ppk_ci_lowertolerance1.06099674975432981.06099674975433046.48E-1615passed
/capability/ppk_ci_uppertolerance1.58306675305655211.58306675305655286.56E-1615passed
/capability/dpmotolerance39.44445087871426439.4444508786673254.69E-1111passed
/capability/within_ppmtolerance1.47753234035395751.47753234035399974.22E-1413passed
/capability/normality_ptail_probability0.820858017758414050.8208580177584216.99E-1514passed
/capability/oosexact000.00E+2passed
/primary/0/centertolerance10.07997500000000110.0799750000000011.75E-1717passed
/primary/0/uppertolerance10.40104706183511110.4010470618351132.18E-1515passed
/primary/0/lowertolerance9.75890293816489069.7589029381648891.75E-1515passed
/primary/0/ytolerance10.081610.0815999999999981.88E-1515passed
/secondary/0/centertolerance0.556625000000000150.55662500000000012.56E-1817passed
/secondary/0/uppertolerance1.17698308663899391.1769830866389995.08E-1514passed
/secondary/0/ytolerance0.330000000000000070.32999999999999831.78E-1514passed
/run_plot_subgroupsexact880.00E+2passed
/capability/subgroup_countexact880.00E+2passed
/capability/confidence_leveltolerance0.900000000000000020.92.20E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

455  groups5 <- split(x, rep(1:8, each = 5))
456  r <- capability(x, lsl, usl, target, level = 0.90, groups = groups5)
457  means5 <- vapply(groups5, mean, 0); ranges5 <- vapply(groups5, function(g) max(g) - min(g), 0)
458  out <- c(cap_outputs(r, "/capability"),
459           list("/primary/0/center" = num(mean(means5), "X-bar centre: mean of the eight subgroup means."),
460                "/primary/0/upper" = num(mean(means5) + 3 * r$within / sqrt(5), "centre + 3 s_w/sqrt(5)."),
461                "/primary/0/lower" = num(mean(means5) - 3 * r$within / sqrt(5), "centre - 3 s_w/sqrt(5)."),
462                "/primary/0/y" = num(means5[1], "First subgroup mean."),
463                "/secondary/0/center" = num(d2(5) * r$within, "R chart centre d2(5) s_w."),
464                "/secondary/0/upper" = num((d2(5) + 3 * d3(5)) * r$within, "(d2(5) + 3 d3(5)) s_w."),
465                "/secondary/0/y" = num(ranges5[1], "First subgroup range."),
466                "/run_plot_subgroups" = exact(8L, "Eight subgroups, fewer than 25."),
467                "/capability/subgroup_count" = exact(8L, "Eight complete subgroups."),
468                "/capability/confidence_level" = num(0.90, "Configured level.")))
469  emit("STAT-QUAL-CAPABILITY-SIXPACK-002", six_ep, "capabilitySixpackMenuItem", "Capability sixpack with consecutive subgroups of five at 90% confidence", "QUAL_SIXPACK_01",
470       list(column = "value", lsl = lsl, usl = usl, target = target, method = "normal", subgroup_mode = "constant", subgroup_size = 5L, confidence_level = 0.90),
471       out, c(cap_notes, "Subgroups of five: s_w = mean(R_i/d2(5)) (the default range estimator for sizes up to ten); Cpk nu = 1/(2 v) with v = g^-2 sum (d3(5)/d2(5))^2; 90% intervals with z = qnorm(0.95)."),
472       cap_source, cap_cite, six_req, kind = "axis_coverage")

Whole script generators/r/quality_subgroup_capability.R


Capability sixpack by lot column with an upper limit only and excluded rows

Case STAT-QUAL-CAPABILITY-SIXPACK-003 · status: passed

Reference

Evidence type
R
Source
Hand-coded manual capability formulas with stats::pnorm, qnorm, shapiro.test and stats::integrate constants
Citation
Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 8 (Cp, Cpk, Pp, Ppk); NIST Dataplot Cpm; Royston (1995) shapiro.test; R stats documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_SIXPACK_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/capability/nexact38380.00E+2passed
/capability/meantolerance2.01744736842105252.0174473684210523.99E-1615passed
/capability/stdtolerance0.0851382650873084710.085138265087308473.53E-1917passed
/capability/within_sigmatolerance0.0811221710403820380.081122171040382151.11E-1614passed
/capability/cpabsentNonepassed
/capability/cpktolerance0.955565449808554510.95556544980855515.57E-1615passed
/capability/ppabsentNonepassed
/capability/ppktolerance0.910490057322117850.91049005732211961.78E-1514passed
/capability/cpmabsentNonepassed
/capability/cpk_ci_lowertolerance0.678734660156534280.67873466015653291.34E-1514passed
/capability/cpk_ci_uppertolerance1.23239623946057471.23239623946057722.49E-1514passed
/capability/ppk_ci_lowertolerance0.677538025664515490.67753802566451691.44E-1514passed
/capability/ppk_ci_uppertolerance1.14344208897972031.14344208897972232.03E-1514passed
/capability/dpmotolerance3152.62241159791673152.62241159786075.60E-1113passed
/capability/within_ppmtolerance2073.90385979726312073.903859797251.32E-1114passed
/capability/normality_ptail_probability0.206839348604739790.206839348604735024.77E-1513passed
/capability/oosexact110.00E+2passed
/capability/cpm_noteexactCpm needs both specification limits and a target.Cpm needs both specification limits and a target.passed
/within_spread/1tolerance2.26081388154219852.26081388154219853.35E-1716passed
/excluded_countexact220.00E+2passed
/capability/subgroup_countexact880.00E+2passed
/primary/0/centertolerance2.01744736842105252.01744736842105254.48E-1716passed
/primary/0/uppertolerance2.13913062498162582.13913062498162583.88E-1716passed
/primary/1/uppertolerance2.12628418177824812.1262841817782482.76E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

485  out[["/primary/0/center"]] <- num(mean(xk), "X-bar centre: size-weighted mean of the 38 retained observations.")
486  out[["/primary/0/upper"]] <- num(mean(xk) + 3 * r$within / sqrt(4), "Size-weighted centre + 3 s_w/sqrt(4): lot L01 has four values after the exclusion.")
487  out[["/primary/1/upper"]] <- num(mean(xk) + 3 * r$within / sqrt(5), "Size-weighted centre + 3 s_w/sqrt(5) for the complete lot L02.")
488  emit("STAT-QUAL-CAPABILITY-SIXPACK-003", six_ep, "capabilitySixpackMenuItem", "Capability sixpack by lot column with an upper limit only and excluded rows", "QUAL_SIXPACK_01",
489       list(column = "thickness", usl = 2.25, method = "normal", subgroup_mode = "by_column", group_column = "lot", confidence_level = 0.95),
490       out, c(cap_notes, "Only USL: Cp, Pp and Cpm are unavailable; Cpk = CPU and Ppk = PPU. Lots L01 and L05 lose one row each to the exclusions, so s_w averages R_i/d2(n_i) over sizes 4 and 5 and the X-bar limits change with n_i.",
491              "The X-bar centre is the size-weighted mean of all retained observations."),
492       cap_source, cap_cite, six_req, kind = "axis_coverage", excluded = c(2L, 21L))

Whole script generators/r/quality_subgroup_capability.R


Slifker-Shapiro supplied-parameter Johnson SU transform

Case STAT-QUAL-JOHNSON-PARAM-001 · status: passed

Reference

Evidence type
R
Source
R recomputation from Slifker and Shapiro (1980), p. 243
Citation
Slifker and Shapiro (1980), The Johnson System: Selection and Parameter Estimation, Technometrics 22(2), 239-246, p. 243
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_johnson_supplied_parameters.R · sha256 7eda3f986a1a3fa6
Dataset
reference_data/datasets/QUAL_JOHNSON_SU_PUBLISHED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/scores/0tolerance3.00052188124300743.00052188124300744.93E-1716passed
/scores/1tolerance1.00067495872033191.00067495872033162.69E-1615passed
/scores/2tolerance-0.99920282692934959-0.99920282692934964.32E-1817passed
/scores/3tolerance-2.9990399696617649-2.9990399696617657.19E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 6  run <- generator_run(c("stats", "jsonlite"))
 7  examples <- list(
 8    list(id = "STAT-QUAL-JOHNSON-PARAM-001", dataset = "QUAL_JOHNSON_SU_PUBLISHED",
 9         family = "SU", x = c(16.689, 15.242, 13.581, 10.409),
10         a = 1.402, b = 2.333, loc = 15.516, scale = 1.585,
11         printed = c(3, 1, -1, -3)),
12    list(id = "STAT-QUAL-JOHNSON-PARAM-002", dataset = "QUAL_JOHNSON_SB_PUBLISHED",
13         family = "SB", x = c(0.786, 0.635, 0.516, 0.432),
14         a = 2.373, b = 1.959, loc = 0.295, scale = 1.203,
15         printed = c(1.645, 0.5483, -0.5483, -1.645)))

Whole script generators/r/quality_johnson_supplied_parameters.R


Slifker-Shapiro supplied-parameter Johnson SB transform

Case STAT-QUAL-JOHNSON-PARAM-002 · status: passed

Reference

Evidence type
R
Source
R recomputation from Slifker and Shapiro (1980), p. 243
Citation
Slifker and Shapiro (1980), The Johnson System: Selection and Parameter Estimation, Technometrics 22(2), 239-246, p. 243
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_johnson_supplied_parameters.R · sha256 7eda3f986a1a3fa6
Dataset
reference_data/datasets/QUAL_JOHNSON_SB_PUBLISHED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/scores/0tolerance1.64496941789340581.64496941789340581.69E-1716passed
/scores/1tolerance0.548252085065960680.54825208506596071.48E-1817passed
/scores/2tolerance-0.54870864079312076-0.54870864079312081.00E-1817passed
/scores/3tolerance-1.6462561629605261-1.64625616296052612.60E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

12    list(id = "STAT-QUAL-JOHNSON-PARAM-002", dataset = "QUAL_JOHNSON_SB_PUBLISHED",
13         family = "SB", x = c(0.786, 0.635, 0.516, 0.432),
14         a = 2.373, b = 1.959, loc = 0.295, scale = 1.203,
15         printed = c(1.645, 0.5483, -0.5483, -1.645)))

Whole script generators/r/quality_johnson_supplied_parameters.R


Width 3 with a Phase I marker and the moving-range sigma

Case STAT-QUAL-MA-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Example 9.3, pp. 428-429; NIST/SEMATECH e-Handbook 6.3.2.1 for d2 moving-range factor
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_MA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/0/moving_averagetolerance11.908511.90858.53E-1717passed
/points/0/window_sizeexact110.00E+2passed
/points/0/ucltolerance14.60846247283869914.60846247283871.42E-1617passed
/points/0/lcltolerance9.26513752716130069.26513752716133.20E-1717passed
/points/1/moving_averagetolerance11.827811.82781.31E-1616passed
/points/1/window_sizeexact220.00E+2passed
/points/1/ucltolerance13.82595065158586313.8259506515858632.27E-1616passed
/points/1/lcltolerance10.04764934841413610.0476493484141364.83E-1616passed
/points/2/moving_averagetolerance11.33393333333333311.3339333333333341.30E-1515passed
/points/2/window_sizeexact330.00E+2passed
/points/2/ucltolerance13.47928504787724413.4792850478772443.52E-1616passed
/points/2/lcltolerance10.39431495212275510.3943149521227553.58E-1616passed
/points/10/moving_averagetolerance12.561912.56193.78E-1616passed
/points/10/window_sizeexact330.00E+2passed
/points/10/ucltolerance13.47928504787724413.4792850478772443.52E-1616passed
/points/10/lcltolerance10.39431495212275510.3943149521227553.58E-1616passed
/points/29/moving_averagetolerance11.95066666666666711.9506666666666691.55E-1515passed
/points/29/window_sizeexact330.00E+2passed
/points/29/ucltolerance13.47928504787724413.4792850478772443.52E-1616passed
/points/29/lcltolerance10.39431495212275510.3943149521227553.58E-1616passed
/baseline_meantolerance11.936811.93681.45E-1616passed
/baseline_sigmatolerance0.890554157612899580.89055415761289962.68E-1817passed
/baseline_countexact30300.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

82  set.seed(20261010)
83  x <- round(c(rnorm(30, 12, 0.8), rnorm(20, 12.9, 0.8)), 4)
84  export_dataset(data.frame(x = x, phase1 = c(rep(TRUE, 30), rep(FALSE, 20)),
85                            stage = c(rep("S1", 25), rep("S2", 25))),
86                 "QUAL_MA", "Constructed: 30 points N(12, 0.8^2) then 20 points N(12.9, 0.8^2), seed 20261010",
87                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
88                 c(x = "Float64", phase1 = "Boolean", stage = "String"))
89  d <- read_dataset("QUAL_MA")
90  base <- d$x[d$phase1]
91  s_mr <- mean(abs(diff(base))) / d2
92  r <- moving(d$x, 3, mean(base), s_mr)
93  emit("STAT-QUAL-MA-001", ma_entry, ma_menu, "Width 3 with a Phase I marker and the moving-range sigma", "QUAL_MA",
94       list(value_column = "x", baseline_column = "phase1"),
95       c(ma_outputs(r), list("/baseline_mean" = tol(mean(base), "Mean of the 30 Phase I rows."),
96                             "/baseline_sigma" = tol(s_mr, "MR-bar / d2 over the Phase I rows."),
97                             "/baseline_count" = exact(30L, "Thirty Phase I rows."))),
98       c("Expanding startup: k = 1, 2, 3 (line 28-29)."))

Whole script generators/r/quality_ma_shortrun.R


Width 5 with the sample standard deviation

Case STAT-QUAL-MA-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Example 9.3, pp. 428-429; NIST/SEMATECH e-Handbook 6.3.2.1 for d2 moving-range factor
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_MA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/0/moving_averagetolerance11.908511.90858.53E-1717passed
/points/0/window_sizeexact110.00E+2passed
/points/0/ucltolerance14.49238786138167814.4923878613816781.06E-1617passed
/points/0/lcltolerance9.38121213861832189.3812121386183221.64E-1717passed
/points/1/moving_averagetolerance11.827811.82781.31E-1616passed
/points/1/window_sizeexact220.00E+2passed
/points/1/ucltolerance13.74387350670101113.743873506701012.88E-1616passed
/points/1/lcltolerance10.12972649329898910.1297264932989891.50E-1818passed
/points/2/moving_averagetolerance11.33393333333333311.3339333333333341.30E-1515passed
/points/2/window_sizeexact330.00E+2passed
/points/2/ucltolerance13.41226933970645213.4122693397064523.68E-1616passed
/points/2/lcltolerance10.46133066029354810.4613306602935487.82E-1717passed
/points/10/moving_averagetolerance12.38941999999999912.3894199999999981.34E-1515passed
/points/10/window_sizeexact550.00E+2passed
/points/10/ucltolerance13.07969363610454713.0796936361045474.94E-1616passed
/points/10/lcltolerance10.79390636389545210.7939063638954522.16E-1616passed
/points/29/moving_averagetolerance11.54516000000000111.545161.80E-1515passed
/points/29/window_sizeexact550.00E+2passed
/points/29/ucltolerance13.07969363610454713.0796936361045474.94E-1616passed
/points/29/lcltolerance10.79390636389545210.7939063638954522.16E-1616passed
/baseline_sigmatolerance0.851862620460559270.85186262046055932.22E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 99  s_sd <- sd(base)
100  r <- moving(d$x, 5, mean(base), s_sd)
101  emit("STAT-QUAL-MA-002", ma_entry, ma_menu, "Width 5 with the sample standard deviation", "QUAL_MA",
102       list(value_column = "x", baseline_column = "phase1", window_width = 5L, sigma_estimator = "stddev"),
103       c(ma_outputs(r), list("/baseline_sigma" = tol(s_sd, "Sample standard deviation of the Phase I rows."))),
104       c("sigma_estimator = stddev (line 29-31)."), kind = "axis_coverage")

Whole script generators/r/quality_ma_shortrun.R


Full-window startup with historical parameters

Case STAT-QUAL-MA-003 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Example 9.3, pp. 428-429; NIST/SEMATECH e-Handbook 6.3.2.1 for d2 moving-range factor
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_MA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/0/limits_availableexactFalseFalsepassed
/points/2/limits_availableexactFalseFalsepassed
/points/3/limits_availableexactTrueTruepassed
/points/3/ucltolerance13.19999999999999913.22.89E-1616passed
/points/0/moving_averagetolerance11.908511.90858.53E-1717passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

105  r <- moving(d$x, 4, 12, 0.8, full = TRUE)
106  emit("STAT-QUAL-MA-003", ma_entry, ma_menu, "Full-window startup with historical parameters", "QUAL_MA",
107       list(value_column = "x", window_width = 4L, partial_window_policy = "full_window", historical_mean = 12, historical_sigma = 0.8),
108       list("/points/0/limits_available" = exact(FALSE, "The first w - 1 rows have no limits (line 34-36)."),
109            "/points/2/limits_available" = exact(FALSE, "Still a partial window."),
110            "/points/3/limits_available" = exact(TRUE, "The first complete window starts monitoring."),
111            "/points/3/ucl" = tol(12 + 3 * 0.8 / 2, "Historical centre + 3 sigma / sqrt(4)."),
112            "/points/0/moving_average" = tol(r$m[1], "The expanding partial statistic is still reported.")),
113       c("Historical mean 12 and sigma 0.8 are used exactly (line 12-13)."), kind = "axis_coverage")

Whole script generators/r/quality_ma_shortrun.R


A stage column resets the window

Case STAT-QUAL-MA-004 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Example 9.3, pp. 428-429; NIST/SEMATECH e-Handbook 6.3.2.1 for d2 moving-range factor
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_MA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/25/window_sizeexact110.00E+2passed
/points/26/moving_averagetolerance10.93690000000000110.9369000000000013.98E-1616passed
/points/24/window_sizeexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

114  r <- moving(d$x, 3, 12, 0.8, stage = d$stage)
115  emit("STAT-QUAL-MA-004", ma_entry, ma_menu, "A stage column resets the window", "QUAL_MA",
116       list(value_column = "x", window_width = 3L, stage_column = "stage", historical_mean = 12, historical_sigma = 0.8),
117       list("/points/25/window_size" = exact(1L, "The first row of stage S2 starts a new window (line 39-47)."),
118            "/points/26/moving_average" = tol(r$m[27], "Mean of the first two S2 rows."),
119            "/points/24/window_size" = exact(3L, "A full window at the end of S1.")),
120       c("Stages S1 (rows 1-25) and S2 (rows 26-50)."), kind = "axis_coverage")

Whole script generators/r/quality_ma_shortrun.R


Width one is the individuals specification

Case STAT-QUAL-MA-005 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Example 9.3, pp. 428-429; NIST/SEMATECH e-Handbook 6.3.2.1 for d2 moving-range factor
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_MA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/40/moving_averagetolerance11.926711.92673.01E-1616passed
/points/40/ucltolerance14.60846247283869914.60846247283871.42E-1617passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

121  r <- moving(d$x, 1, mean(base), s_mr)
122  emit("STAT-QUAL-MA-005", ma_entry, ma_menu, "Width one is the individuals specification", "QUAL_MA",
123       list(value_column = "x", window_width = 1L, baseline_column = "phase1"),
124       list("/points/40/moving_average" = tol(d$x[41], "Each average is the observation itself (line 6-8)."),
125            "/points/40/ucl" = tol(mean(base) + 3 * s_mr, "Centre + 3 sigma.")),
126       c("Width 1."), kind = "boundary")

Whole script generators/r/quality_ma_shortrun.R


Missing and excluded rows are omitted before the windows

Case STAT-QUAL-MA-006 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Example 9.3, pp. 428-429; NIST/SEMATECH e-Handbook 6.3.2.1 for d2 moving-range factor
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_MA_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/0/moving_averagetolerance11.908511.90858.53E-1717passed
/points/0/window_sizeexact110.00E+2passed
/points/0/ucltolerance14.36361941578520214.3636194157852022.62E-1616passed
/points/0/lcltolerance9.66508058421479899.6650805842148023.53E-1515passed
/points/5/moving_averagetolerance11.78336666666666611.7833666666666661.22E-1616passed
/points/5/window_sizeexact330.00E+2passed
/points/5/ucltolerance13.37070132960254113.3707013296025431.98E-1515passed
/points/5/lcltolerance10.65799867039745910.6579986703974612.18E-1515passed
/points/10/moving_averagetolerance12.30313333333333312.3031333333333352.03E-1515passed
/points/10/window_sizeexact330.00E+2passed
/points/10/ucltolerance13.37070132960254113.3707013296025431.98E-1515passed
/points/10/lcltolerance10.65799867039745910.6579986703974612.18E-1515passed
/points/30/moving_averagetolerance12.912912.91294.89E-1616passed
/points/30/window_sizeexact330.00E+2passed
/points/30/ucltolerance13.37070132960254113.3707013296025431.98E-1515passed
/points/30/lcltolerance10.65799867039745910.6579986703974612.18E-1515passed
/excluded_countexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

128  m <- read_dataset("QUAL_MA"); m$x[c(6, 33)] <- NA
129  export_dataset(m, "QUAL_MA_MISSING", "QUAL_MA with x missing in rows 6 and 33 (1-based)", "R stats::rnorm documentation",
130                 "Constructed validation data; no third-party license.", c(x = "Float64", phase1 = "Boolean", stage = "String"))
131  mm <- read_dataset("QUAL_MA_MISSING")
132  excluded <- c(9L, 40L)
133  keep <- setdiff(which(!is.na(mm$x)), excluded + 1L)
134  xk <- mm$x[keep]; bk <- mm$phase1[keep]
135  r <- moving(xk, 3, mean(xk[bk]), mean(abs(diff(xk[bk]))) / d2)
136  emit("STAT-QUAL-MA-006", ma_entry, ma_menu, "Missing and excluded rows are omitted before the windows", "QUAL_MA_MISSING",
137       list(value_column = "x", baseline_column = "phase1"),
138       c(ma_outputs(r, c(0L, 5L, 10L, 30L)), list("/excluded_count" = exact(2L, "Two excluded rows."))),
139       c("Line 41-43: the retained rows become adjacent; excluded rows 9 and 40 (0-based) and two missing values are omitted."),
140       kind = "missing_excluded", excluded = excluded)

Whole script generators/r/quality_ma_shortrun.R


A constant baseline is rejected

Case STAT-QUAL-MA-007 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Example 9.3, pp. 428-429; NIST/SEMATECH e-Handbook 6.3.2.1 for d2 moving-range factor
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_MA_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

141  export_dataset(data.frame(x = rep(5, 8), phase1 = rep(TRUE, 8)), "QUAL_MA_CONSTANT", "Constructed: eight equal values",
142                 "Hand-constructed", "Constructed validation data; no third-party license.", c(x = "Float64", phase1 = "Boolean"))
143  emit("STAT-QUAL-MA-007", ma_entry, ma_menu, "A constant baseline is rejected", "QUAL_MA_CONSTANT",
144       list(value_column = "x", baseline_column = "phase1"), list(),
145       c("Line 32: constant estimated baseline data are rejected; the fragment is the product's wording."), kind = "invalid_input",
146       error = list(error_class = "SpcMovingAverageExecutionError", message_contains = "no varying measurements"))

Whole script generators/r/quality_ma_shortrun.R


A historical mean without a sigma is refused

Case STAT-QUAL-MA-008 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Example 9.3, pp. 428-429; NIST/SEMATECH e-Handbook 6.3.2.1 for d2 moving-range factor
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_MA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

147  emit("STAT-QUAL-MA-008", ma_entry, ma_menu, "A historical mean without a sigma is refused", "QUAL_MA",
148       list(value_column = "x", historical_mean = 12), list(),
149       c("Line 12-13: the two historical values must be supplied together."), kind = "invalid_input",
150       error = list(error_class = "ValueError", message_contains = "supplied together"))

Whole script generators/r/quality_ma_shortrun.R


MEWMA with lambda 0.2 and a trial limit of 11

Case STAT-QUAL-MEWMA-001 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/0/statistictolerance2.31565264258923122.31565264258924631.51E-1414passed
/points/9/statistictolerance6.31990970071343666.3199097007133379.95E-1413passed
/points/19/statistictolerance113.85192367474005113.85192367473946.54E-1314passed
/signal_countexact10100.00E+2passed
/asymptotic_factortolerance0.111111111111111120.111111111111111121.18E-1816passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

131  mw <- mewma(M0, 0.2, xbar, S)
132  emit("STAT-QUAL-MEWMA-001", mewma_entry, "mewmaMenuItem", "MEWMA with lambda 0.2 and a trial limit of 11", "QUAL_T2",
133       list(variables = as.list(vars), reference_column = "role", reference_value = "ref", decision_limit = 11),
134       list("/points/0/statistic" = tol(mw[1], "Exact finite-time covariance at step 1."),
135            "/points/9/statistic" = tol(mw[10], "Step 10."),
136            "/points/19/statistic" = tol(mw[20], "Step 20."),
137            "/signal_count" = exact(sum(mw > 11), "Points with T^2 > h."),
138            "/asymptotic_factor" = tol(0.2 / 1.8, "Steady-state lambda / (2 - lambda).")),
139       c("Z_0 = xbar; no reset after a signal (line 77-78)."))

Whole script generators/r/quality_multivariate_spc.R


Lambda 1 reproduces the individual Hotelling statistic

Case STAT-QUAL-MEWMA-002 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/0/statistictolerance2.31565264258925342.3156526425892458.43E-1514passed
/points/15/statistictolerance34.55081776270690634.550817762706819.95E-1414passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

140  m1 <- mewma(M0, 1, xbar, S)
141  emit("STAT-QUAL-MEWMA-002", mewma_entry, "mewmaMenuItem", "Lambda 1 reproduces the individual Hotelling statistic", "QUAL_T2",
142       list(variables = as.list(vars), reference_column = "role", reference_value = "ref", smoothing = 1, decision_limit = 11),
143       list("/points/0/statistic" = tol(t2[1], "Line 63-65: with lambda = 1 the statistic equals the individual T^2."),
144            "/points/15/statistic" = tol(t2[16], "Same identity at a shifted row.")),
145       c(sprintf("R check: max |MEWMA(lambda = 1) - T^2| = %.3g.", max(abs(m1 - t2)))), kind = "boundary")

Whole script generators/r/quality_multivariate_spc.R


A stage column restarts the recursion

Case STAT-QUAL-MEWMA-003 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/10/stepexact110.00E+2passed
/points/10/statistictolerance14.91945426834562514.919454268345548.54E-1414passed
/segment_countexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

146  ms <- mewma(M0, 0.3, xbar, S, stage = d$stage[d$role == "mon"])
147  emit("STAT-QUAL-MEWMA-003", mewma_entry, "mewmaMenuItem", "A stage column restarts the recursion", "QUAL_T2",
148       list(variables = as.list(vars), reference_column = "role", reference_value = "ref", smoothing = 0.3,
149            decision_limit = 12, stage_column = "stage"),
150       list("/points/10/step" = exact(1L, "The first S2 row restarts at step 1 (line 69-71)."),
151            "/points/10/statistic" = tol(ms[11], "Restarted from Z = xbar."),
152            "/segment_count" = exact(2L, "Two stages.")),
153       c("Monitoring stages S1 (rows 41-50) and S2 (rows 51-60), 1-based."), kind = "axis_coverage")

Whole script generators/r/quality_multivariate_spc.R


A user-calibrated limit needs a source note

Case STAT-QUAL-MEWMA-004 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

154  emit("STAT-QUAL-MEWMA-004", mewma_entry, "mewmaMenuItem", "A user-calibrated limit needs a source note", "QUAL_T2",
155       list(variables = as.list(vars), reference_column = "role", reference_value = "ref", decision_limit = 11,
156            limit_provenance = "user_calibrated"), list(),
157       c("Line 25-28: a user-calibrated limit requires a source note."), kind = "invalid_input",
158       error = list(error_class = "ValueError", message_contains = "source"))

Whole script generators/r/quality_multivariate_spc.R


A smoothing weight above 1 is refused

Case STAT-QUAL-MEWMA-005 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

159  emit("STAT-QUAL-MEWMA-005", mewma_entry, "mewmaMenuItem", "A smoothing weight above 1 is refused", "QUAL_T2",
160       list(variables = as.list(vars), reference_column = "role", reference_value = "ref", smoothing = 1.2, decision_limit = 11), list(),
161       c("Line 19: 0 < lambda <= 1."), kind = "invalid_input",
162       error = list(error_class = "ValueError", message_contains = "less than or equal to 1"))

Whole script generators/r/quality_multivariate_spc.R


NIST R-chart D4(5): printed table versus full precision

Case STAT-QUAL-PUBCHART-001 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_NIST_R5.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/secondary/center_linetolerance44.00.00E+2passed
/secondary/upper_control_limittolerance8.45799658038355428.4579965803835521.80E-1515passed

Convention

Reference code

Whole script generators/chart_published_reference.py


NIST wafer c chart with strict on-limit decision

Case STAT-QUAL-PUBCHART-002 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc331.htm
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_NIST_WAFERS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance1616.00.00E+2passed
/chart/lower_control_limittolerance44.00.00E+2passed
/chart/upper_control_limittolerance2828.00.00E+2passed
/chart/points/2/valuetolerance2828.00.00E+2passed
/chart/points/23/valuetolerance3131.00.00E+2passed
/violation_keysset['primary:rule_1:23']['primary:rule_1:23']passed

Convention

Reference code

Whole script generators/chart_published_reference.py


NIST tabular CUSUM strict H boundary

Case STAT-QUAL-PUBCHART-003 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc323.htm
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_NIST_CUSUM_DERIVED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/points/0/valuetolerance0.50.50.00E+1passed
/primary/points/1/valuetolerance11.00.00E+2passed
/primary/points/2/valuetolerance1.51.50.00E+1passed
/primary/points/3/valuetolerance00.00.00E+2passed
/violation_keysset['primary:rule_1:2']['primary:rule_1:2']passed

Convention

Reference code

Whole script generators/chart_published_reference.py


Montgomery Example 9.2 finite-time EWMA

Case STAT-QUAL-PUBCHART-004 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Examples 9.2-9.3 and Tables 9.10/9.14, pp. 421-429
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_MONTGOMERY_9_2_9_3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/points/0/valuetolerance9.94500000000000039.9451.58E-1717passed
/primary/points/1/valuetolerance9.74949999999999949.74951.11E-1717passed
/primary/points/2/valuetolerance9.70354999999999999.703552.32E-1818passed
/primary/points/3/valuetolerance9.89919500000000069.8991953.27E-1717passed
/primary/points/4/valuetolerance10.12527550000000110.12527552.39E-1616passed
/primary/points/5/valuetolerance10.13074795000000210.1307479500000022.77E-1616passed
/primary/points/6/valuetolerance9.92167315500000249.9216731550000024.68E-1717passed
/primary/points/7/valuetolerance10.07550583950000310.0755058395000034.47E-1616passed
/primary/points/8/valuetolerance9.98795525555000379.9879552555500044.14E-1717passed
/primary/points/9/valuetolerance10.02315972999500510.0231597299950054.03E-1616passed
/primary/points/10/valuetolerance9.92384375699550439.9238437569955044.95E-1717passed
/primary/points/11/valuetolerance10.07845938129595410.0784593812959544.66E-1616passed
/primary/points/12/valuetolerance10.12161344316635910.1216134431663593.08E-1616passed
/primary/points/13/valuetolerance10.04945209884972210.0494520988497222.07E-1616passed
/primary/points/14/valuetolerance10.0525068889647510.052506888964752.41E-1717passed
/primary/points/15/valuetolerance9.98425620006827519.9842562000682752.22E-1717passed
/primary/points/16/valuetolerance10.04783058006144710.0478305800614471.37E-1616passed
/primary/points/17/valuetolerance10.07404752205530410.0740475220553047.14E-1717passed
/primary/points/18/valuetolerance9.91864276984977349.9186427698497734.40E-1717passed
/primary/points/19/valuetolerance10.01077849286479610.0107784928647962.63E-1616passed
/primary/points/20/valuetolerance10.09970064357831610.0997006435783163.17E-1616passed
/primary/points/21/valuetolerance10.02273057922048410.0227305792204841.59E-1616passed
/primary/points/22/valuetolerance10.24945752129843510.2494575212984351.25E-1616passed
/primary/points/23/valuetolerance10.37451176916859210.3745117691685923.19E-1717passed
/primary/points/24/valuetolerance10.39706059225173410.3970605922517343.76E-1616passed
/primary/points/25/valuetolerance10.46535453302656110.4653545330265611.57E-1616passed
/primary/points/26/valuetolerance10.45681907972390510.4568190797239051.19E-1616passed
/primary/points/27/valuetolerance10.57313717175151510.5731371717515152.40E-1616passed
/primary/points/28/valuetolerance10.64682345457636410.6468234545763644.12E-1616passed
/primary/points/29/valuetolerance10.63414110911872810.6341411091187281.03E-1617passed
/primary/point_upper_limitssequence['10.27', '10.363247849270991', '10.424003172157946', '10.46[10.27, 10.36324784927099, 10.424003172157946, 10.4674620186passed
/primary/point_lower_limitssequence['9.7300000000000004', '9.6367521507290093', '9.575996827842[9.73, 9.63675215072901, 9.575996827842054, 9.53253798132896passed
/violation_keysset['primary:rule_1:28', 'primary:rule_1:29']['primary:rule_1:28', 'primary:rule_1:29']passed

Convention

Reference code

Whole script generators/chart_published_reference.py


Montgomery Example 9.3 all 30 moving averages

Case STAT-QUAL-PUBCHART-005 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Examples 9.2-9.3 and Tables 9.10/9.14, pp. 421-429
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_MONTGOMERY_9_2_9_3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/points/0/moving_averagetolerance9.44999999999999939.451.05E-1717passed
/points/1/moving_averagetolerance8.72000000000000068.7199999999999991.74E-1515passed
/points/2/moving_averagetolerance8.91000000000000018.9099999999999981.73E-1515passed
/points/3/moving_averagetolerance9.59750000000000019.59754.21E-1717passed
/points/4/moving_averagetolerance10.10999999999999910.114.32E-1616passed
/points/5/moving_averagetolerance10.25610.2562.27E-1616passed
/points/6/moving_averagetolerance10.26610.2661.42E-1717passed
/points/7/moving_averagetolerance10.69999999999999910.72.89E-1616passed
/points/8/moving_averagetolerance10.20810.2080000000000021.96E-1515passed
/points/9/moving_averagetolerance9.84399999999999949.8441.74E-1717passed
/points/10/moving_averagetolerance9.61400000000000089.6143.26E-1717passed
/points/11/moving_averagetolerance10.30000000000000110.32.89E-1616passed
/points/12/moving_averagetolerance10.10999999999999910.114.32E-1616passed
/points/13/moving_averagetolerance10.1510.1499999999999991.42E-1515passed
/points/14/moving_averagetolerance10.09800000000000110.0979999999999992.02E-1515passed
/points/15/moving_averagetolerance10.16610.1663.69E-1616passed
/points/16/moving_averagetolerance9.99600000000000049.9959999999999991.74E-1515passed
/points/17/moving_averagetolerance9.95599999999999959.9561.68E-1717passed
/points/18/moving_averagetolerance9.77999999999999949.7800000000000011.74E-1515passed
/points/19/moving_averagetolerance9.93200000000000049.9319999999999991.79E-1515passed
/points/20/moving_averagetolerance10.23810.2384.55E-1616passed
/points/21/moving_averagetolerance9.98000000000000049.982.63E-1717passed
/points/22/moving_averagetolerance10.37599999999999910.3764.46E-1616passed
/points/23/moving_averagetolerance10.97210.9724.69E-1616passed
/points/24/moving_averagetolerance10.92399999999999910.9244.88E-1616passed
/points/25/moving_averagetolerance10.96000000000000110.9599999999999991.92E-1515passed
/points/26/moving_averagetolerance11.1711.177.11E-1717passed
/points/27/moving_averagetolerance11.03611.0364.12E-1616passed
/points/28/moving_averagetolerance10.99799999999999910.9980000000000012.11E-1515passed
/points/29/moving_averagetolerance10.98199999999999910.9823.18E-1616passed
/points/0/window_sizeexact110.00E+2passed
/points/4/window_sizeexact550.00E+2passed
/points/4/ucltolerance11.34164078649987311.3416407864998731.72E-1616passed

Convention

Reference code

Whole script generators/chart_published_reference.py


Lucas-Saccucci Table 1 EWMA finite-time decisions

Case STAT-QUAL-PUBCHART-006 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
Lucas and Saccucci, Technometrics 32(1), 1990, Table 1
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_LUCAS_EWMA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/points/0/valuetolerance0.250.250.00E+0passed
/primary/points/1/valuetolerance0.06250.06250.00E-2passed
/primary/points/2/valuetolerance0.0468750.0468750.00E-4passed
/primary/points/3/valuetolerance-0.16484375000000001-0.164843751.10E-1817passed
/primary/points/4/valuetolerance-0.32363281250000003-0.323632812500000033.31E-1816passed
/primary/points/5/valuetolerance-0.542724609375-0.5427246093750.00E-10passed
/primary/points/6/valuetolerance-0.03204345703125-0.032043457031250.00E-12passed
/primary/points/7/valuetolerance-0.17403259277343749-0.17403259277343754.45E-1816passed
/primary/points/8/valuetolerance0.119475555419921890.119475555419921893.90E-1816passed
/primary/points/9/valuetolerance-0.1353933334350586-0.13539333343505864.85E-1816passed
/primary/points/10/valuetolerance0.198454999923706020.198454999923706021.38E-1817passed
/primary/points/11/valuetolerance0.273841249942779530.273841249942779538.66E-2018passed
/primary/points/12/valuetolerance0.855380937457084660.85538093745708474.24E-1817passed
/primary/points/13/valuetolerance0.816535703092813540.81653570309281353.77E-1817passed
/primary/points/14/valuetolerance0.887401777319610230.88740177731961021.12E-1918passed
/primary/points/15/valuetolerance1.16555133298970761.16555133298970761.09E-1717passed
/primary/points/16/valuetolerance1.22416349974228081.22416349974228081.93E-1716passed
/primary/points/17/valuetolerance1.39312262480671061.39312262480671063.67E-1716passed
/primary/points/18/valuetolerance1.24484196860503291.2448419686050333.36E-1716passed
/violation_keysset['primary:rule_1:15', 'primary:rule_1:16', 'primary:rule_1:1['primary:rule_1:15', 'primary:rule_1:16', 'primary:rule_1:1passed

Convention

Reference code

Whole script generators/chart_published_reference.py


Minitab Z-MR printed factors versus full precision

Case STAT-QUAL-PUBCHART-007 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
https://support.minitab.com/en-us/minitab/help-and-how-to/quality-and-process-improvement/control-charts/how-to/variables-charts-for-individuals/z-mr-chart/methods-and-formulas/methods-and-formulas/
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_ZMR_DERIVED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/mr_chart/upper_control_limittolerance3.68588656637777763.68588656637777763.77E-1716passed

Convention

Reference code

Whole script generators/chart_published_reference.py


Benneyan geometric support translated to Tabelz g chart

Case STAT-QUAL-PUBCHART-008 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
Benneyan, Health Care Management Science 4(4), 2001, Example 1, p. 311
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_BENNEYAN_G_TRANSLATED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/chart/center_linetolerance42.20000000000000342.21.58E-1617passed
/chart/lower_control_limittolerance11.00.00E+2passed
/chart/upper_control_limittolerance276276.00.00E+2passed
/violation_keysset['primary:rule_1:6']['primary:rule_1:6']passed

Convention

Reference code

Whole script generators/chart_published_reference.py


Lowry Table 4 MEWMA from printed inputs

Case STAT-QUAL-PUBCHART-009 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
Lowry, Woodall, Champ and Rigdon, Technometrics 34(1), 1992, Table 4
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_LOWRY_TABLE4.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/baseline/mean/0tolerance00.00.00E+2passed
/baseline/covariance/0/1tolerance0.50.499999999999999945.55E-1715passed
/points/0/statistictolerance3.28840000000000023.28840000000000151.34E-1515passed
/points/1/statistictolerance3.17719558011049723.17719558011049851.32E-1515passed
/points/2/statistictolerance7.36809149669518787.36809149669519052.68E-1515passed
/points/3/statistictolerance5.25907221610424545.2590722161042468.45E-1615passed
/points/4/statistictolerance1.09203332713644061.09203332713644041.87E-1615passed
/points/5/statistictolerance1.27995496390891921.27995496390891874.81E-1615passed
/points/6/statistictolerance5.66105151090627255.6610515109062738.43E-1615passed
/points/7/statistictolerance8.32442330895979418.3244233089597941.12E-1818passed
/points/8/statistictolerance9.64498269049053479.6449826904905356.41E-1818passed
/points/9/statistictolerance17.20744472781586217.2074447278158721.02E-1415passed
/signal_countexact220.00E+2passed

Convention

Reference code

Whole script generators/chart_published_reference.py


Montgomery Example 11.1 subgroup Hotelling full-precision limit

Case STAT-QUAL-PUBCHART-010 · status: passed

Reference

Evidence type
Analytical
Source
Published inputs and independent analytic calculation
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., Example 11.1
Reference software
not recorded
Reference script
generators/chart_published_reference.py · sha256 18ce58af14ce93a9
Dataset
reference_data/datasets/PUB_MONTGOMERY_11_1_DERIVED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/baseline/t_squared_ucltolerance15.16503018300000115.165030183009919.91E-1212passed
/baseline/reference_mean_generalized_variancetolerance10.99999999999999973.33E-1615passed
/baseline/sigma_determinanttolerance1.1251.12499999999999982.22E-1615passed

Convention

Reference code

Whole script generators/chart_published_reference.py


Standardized Z-MR across three products

Case STAT-QUAL-SHORTRUN-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Minitab Z-MR Chart Methods and Formulas, https://support.minitab.com/en-us/minitab/help-and-how-to/quality-and-process-improvement/control-charts/how-to/variables-charts-for-individuals/z-mr-chart/methods-and-formulas/methods-and-formulas/
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_SHORTRUN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observations/0/ztolerance0.0638000000000005230.063800000000000524.05E-2018passed
/observations/0/moving_rangeabsentNonepassed
/observations/1/ztolerance-0.20499999999999829-0.20499999999999834.70E-1816passed
/observations/1/moving_rangetolerance0.268799999999998820.26879999999999882.34E-1817passed
/observations/2/ztolerance-1.0705000000000009-1.0705000000000014.72E-1817passed
/observations/2/moving_rangetolerance0.86550000000000260.86550000000000265.86E-1918passed
/observations/13/ztolerance3.39999999999999863.3999999999999953.57E-1514passed
/observations/13/moving_rangetolerance4.01079999999999974.0107999999999963.55E-1515passed
/observations/14/ztolerance0.983833333333332010.9838333333333324.79E-1817passed
/observations/14/moving_rangetolerance2.41616666666666642.4161666666666633.60E-1514passed
/observations/35/ztolerance1.46700000000000051.46700000000000052.58E-1716passed
/observations/35/moving_rangetolerance0.868333333333334620.86833333333333462.67E-1817passed
/observations/13/signal_reasonscontains['Z']['Z above UCL', 'MR above UCL']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

186  all_valid <- rep(TRUE, 36)
187  r <- short_run(s$x, s$product, all_valid)
188  emit("STAT-QUAL-SHORTRUN-001", sr_entry, sr_menu, "Standardized Z-MR across three products", "QUAL_SHORTRUN",
189       list(measurement_column = "x", product_column = "product", products = products),
190       c(sr_outputs(r, c(0L, 1L, 2L, 13L, 14L, 35L)),
191         list("/observations/13/signal_reasons" = list(comparison = "contains", values = list("Z"),
192                rationale = "Row 14 (1-based) lies 3.4 sigma above its product centre: z > 3."))),
193       c("Reset at product change is off, so cross-product pairs contribute moving ranges (line 28-31).",
194         sprintf("MR upper limit d2 + 3 d3 = %.15g.", d2 + 3 * d3)))

Whole script generators/r/quality_ma_shortrun.R


Reset the moving range at each product change

Case STAT-QUAL-SHORTRUN-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Minitab Z-MR Chart Methods and Formulas, https://support.minitab.com/en-us/minitab/help-and-how-to/quality-and-process-improvement/control-charts/how-to/variables-charts-for-individuals/z-mr-chart/methods-and-formulas/methods-and-formulas/
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_SHORTRUN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observations/1/ztolerance-0.20499999999999829-0.20499999999999834.70E-1816passed
/observations/1/moving_rangetolerance0.268799999999998820.26879999999999882.34E-1817passed
/observations/2/ztolerance-1.0705000000000009-1.0705000000000014.72E-1817passed
/observations/2/moving_rangeabsentNonepassed
/observations/3/ztolerance-1.9280000000000008-1.92800000000000082.42E-1716passed
/observations/3/moving_rangetolerance0.857499999999999930.85749999999999991.05E-1817passed
/observations/5/ztolerance1.15600000000000151.15600000000000152.92E-1716passed
/observations/5/moving_rangeabsentNonepassed
/observations/6/ztolerance-0.56360000000000099-0.5636000000000019.25E-1917passed
/observations/6/moving_rangeabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

195  r <- short_run(s$x, s$product, all_valid, reset = TRUE)
196  emit("STAT-QUAL-SHORTRUN-002", sr_entry, sr_menu, "Reset the moving range at each product change", "QUAL_SHORTRUN",
197       list(measurement_column = "x", product_column = "product", products = products, reset_on_product_change = TRUE),
198       sr_outputs(r, c(1L, 2L, 3L, 5L, 6L)), c("A product transition starts a new moving-range sequence (line 30-31)."),
199       kind = "axis_coverage")

Whole script generators/r/quality_ma_shortrun.R


A stage column breaks adjacency

Case STAT-QUAL-SHORTRUN-003 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Minitab Z-MR Chart Methods and Formulas, https://support.minitab.com/en-us/minitab/help-and-how-to/quality-and-process-improvement/control-charts/how-to/variables-charts-for-individuals/z-mr-chart/methods-and-formulas/methods-and-formulas/
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_SHORTRUN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observations/19/ztolerance0.257400000000000520.25740000000000052.72E-1816passed
/observations/19/moving_rangetolerance2.23782.23781.14E-1717passed
/observations/20/ztolerance-0.85183333333333466-0.85183333333333473.53E-1817passed
/observations/20/moving_rangeabsentNonepassed
/observations/21/ztolerance0.906083333333332570.90608333333333264.53E-1817passed
/observations/21/moving_rangetolerance1.75791666666666721.75791666666666723.81E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

200  r <- short_run(s$x, s$product, all_valid, stage = s$stage)
201  emit("STAT-QUAL-SHORTRUN-003", sr_entry, sr_menu, "A stage column breaks adjacency", "QUAL_SHORTRUN",
202       list(measurement_column = "x", product_column = "product", stage_column = "stage", products = products),
203       sr_outputs(r, c(19L, 20L, 21L)), c("Stage I is rows 1-20 and stage II rows 21-36 (1-based)."), kind = "axis_coverage")

Whole script generators/r/quality_ma_shortrun.R


Missing and excluded rows break adjacency but stay in the table

Case STAT-QUAL-SHORTRUN-004 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Minitab Z-MR Chart Methods and Formulas, https://support.minitab.com/en-us/minitab/help-and-how-to/quality-and-process-improvement/control-charts/how-to/variables-charts-for-individuals/z-mr-chart/methods-and-formulas/methods-and-formulas/
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_SHORTRUN_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observations/3/zabsentNonepassed
/observations/3/moving_rangeabsentNonepassed
/observations/4/ztolerance-0.83158333333333445-0.83158333333333451.25E-1817passed
/observations/4/moving_rangeabsentNonepassed
/observations/7/zabsentNonepassed
/observations/7/moving_rangeabsentNonepassed
/observations/8/ztolerance-0.75916666666666788-0.75916666666666792.45E-1817passed
/observations/8/moving_rangeabsentNonepassed
/observations/25/zabsentNonepassed
/observations/25/moving_rangeabsentNonepassed
/observations/26/ztolerance0.283583333333332630.283583333333332632.26E-1817passed
/observations/26/moving_rangeabsentNonepassed
/observations/3/statusexactexcludedexcludedpassed
/observations/7/statusexactmissingmissingpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

204  m <- read_dataset("QUAL_SHORTRUN"); m$x[8] <- NA
205  export_dataset(m, "QUAL_SHORTRUN_MISSING", "QUAL_SHORTRUN with x missing in row 8 (1-based)", "R stats::rnorm documentation",
206                 "Constructed validation data; no third-party license.", c(x = "Float64", product = "String", stage = "String"))
207  mm <- read_dataset("QUAL_SHORTRUN_MISSING")
208  valid <- !is.na(mm$x); valid[c(3, 25) + 1] <- FALSE
209  r <- short_run(mm$x, mm$product, valid)
210  emit("STAT-QUAL-SHORTRUN-004", sr_entry, sr_menu, "Missing and excluded rows break adjacency but stay in the table", "QUAL_SHORTRUN_MISSING",
211       list(measurement_column = "x", product_column = "product", products = products),
212       c(sr_outputs(r, c(3L, 4L, 7L, 8L, 25L, 26L)),
213         list("/observations/3/status" = exact("excluded", "Excluded rows are retained in the detail table (line 47-48)."),
214              "/observations/7/status" = exact("missing", "A missing measurement."))),
215       c("Rows 3 and 25 (0-based) are excluded; row 7 (0-based) is missing."), kind = "missing_excluded", excluded = c(3L, 25L))

Whole script generators/r/quality_ma_shortrun.R


A product without parameters is an input error

Case STAT-QUAL-SHORTRUN-005 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Minitab Z-MR Chart Methods and Formulas, https://support.minitab.com/en-us/minitab/help-and-how-to/quality-and-process-improvement/control-charts/how-to/variables-charts-for-individuals/z-mr-chart/methods-and-formulas/methods-and-formulas/
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_SHORTRUN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

216  emit("STAT-QUAL-SHORTRUN-005", sr_entry, sr_menu, "A product without parameters is an input error", "QUAL_SHORTRUN",
217       list(measurement_column = "x", product_column = "product", products = products[1:2]), list(),
218       c("Line 13-14: every active product must have a record."), kind = "invalid_input",
219       error = list(error_class = "ShortRunExecutionError", message_contains = "parameters were supplied for product"))

Whole script generators/r/quality_ma_shortrun.R


A zero sigma is refused

Case STAT-QUAL-SHORTRUN-006 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
Minitab Z-MR Chart Methods and Formulas, https://support.minitab.com/en-us/minitab/help-and-how-to/quality-and-process-improvement/control-charts/how-to/variables-charts-for-individuals/z-mr-chart/methods-and-formulas/methods-and-formulas/
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_ma_shortrun.R · sha256 ad9724ed63c7b629
Dataset
reference_data/datasets/QUAL_SHORTRUN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

220  emit("STAT-QUAL-SHORTRUN-006", sr_entry, sr_menu, "A zero sigma is refused", "QUAL_SHORTRUN",
221       list(measurement_column = "x", product_column = "product",
222            products = list(list(product = "A", center = 10, sigma = 0), products[[2]], products[[3]])), list(),
223       c("Line 14: sigma must be strictly positive."), kind = "invalid_input",
224       error = list(error_class = "ValueError", message_contains = "must be finite and positive"))

Whole script generators/r/quality_ma_shortrun.R


Individuals and moving range with the default rules

Case STAT-QUAL-SPC-001 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_IMR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/center_linetolerance50.63048285714285950.630482857142863.23E-1617passed
/primary/upper_control_limittolerance57.4261287189227957.426128718922792.85E-1718passed
/primary/lower_control_limittolerance43.83483699536292743.834836995362933.25E-1617passed
/within_sigmatolerance2.26521528725997672.2652152872599774.17E-1615passed
/secondary/center_linetolerance2.55602173913043452.5560217391304354.55E-1615passed
/secondary/upper_control_limittolerance8.34932659726512768.3493265972651283.01E-1818passed
/secondary/lower_control_limittolerance00.00.00E+2passed
/primary/points/0/valuetolerance48.63329999999999848.63334.20E-1617passed
/secondary/points/0/valuetolerance2.39809999999999952.39809999999999954.57E-1716passed
/violation_keysset['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

114  set.seed(20261001)
115  x <- round(rnorm(70, 50, 2), 4)
116  x[20] <- 62                                   # beyond the limits
117  x[30:38] <- round(51 + abs(rnorm(9, 1.2, 0.3)), 4)   # nine above the centre
118  x[45:50] <- round(seq(46, 52, length.out = 6), 4)    # six rising
119  x[55:57] <- c(55.5, 49.8, 55.9)                # two of three beyond 2 sigma
120  export_dataset(data.frame(t = seq_along(x), x = x, phase = ifelse(seq_along(x) <= 40, 1L, 2L)),
121                 "QUAL_SPC_IMR", "Constructed individuals series, seed 20261001, with an outlier, a shift, a trend and a two-of-three pattern",
122                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
123                 c(t = "Int64", x = "Float64", phase = "Int64"))
124  d <- read_dataset("QUAL_SPC_IMR")
125  default_rules <- c("rule_2", "rule_3", "rule_4")
126  all_rules <- c(default_rules, "rule_5", "rule_6", "rule_7", "rule_8")
127  r <- imr(d$x, default_rules)
128  emit("STAT-QUAL-SPC-001", "Individuals and moving range with the default rules", "QUAL_SPC_IMR",
129       list(y_column = "x"), imr_outputs(r, d$x, "Rules 1-4 on the I chart and rule 1 on the MR chart."),
130       c("Rules 2-4 are on by default (line 216-217)."))

Whole script generators/r/quality_spc.R


Individuals chart with all eight rules

Case STAT-QUAL-SPC-002 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_IMR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/violation_keysset['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

131  r <- imr(d$x, all_rules)
132  emit("STAT-QUAL-SPC-002", "Individuals chart with all eight rules", "QUAL_SPC_IMR",
133       list(y_column = "x", run_rules = as.list(all_rules)), list("/violation_keys" = set_of(r$keys, "Rules 1-8 evaluated independently in R.")),
134       c("Zones use the I chart's sigma = MR-bar/d2 (line 238-240)."), kind = "axis_coverage")

Whole script generators/r/quality_spc.R


Phase I baseline freezes the limits from the first 40 points

Case STAT-QUAL-SPC-003 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_IMR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/center_linetolerance50.89240749999999750.89240753.56E-1617passed
/primary/upper_control_limittolerance57.18909070812301857.189090708123022.03E-1617passed
/primary/lower_control_limittolerance44.59572429187697744.595724291876984.92E-1616passed
/within_sigmatolerance2.0988944027076732.0988944027076732.37E-1716passed
/secondary/center_linetolerance2.36834871794871752.36834871794871751.44E-1717passed
/secondary/upper_control_limittolerance7.7362866831857227.7362866831857229.30E-1817passed
/secondary/lower_control_limittolerance00.00.00E+2passed
/primary/points/0/valuetolerance48.63329999999999848.63334.20E-1617passed
/secondary/points/0/valuetolerance2.39809999999999952.39809999999999954.57E-1716passed
/violation_keysset['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4passed
/baseline_countexact40400.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

135  rb <- imr(d$x, default_rules, baseline = d$phase <= 1)
136  emit("STAT-QUAL-SPC-003", "Phase I baseline freezes the limits from the first 40 points", "QUAL_SPC_IMR",
137       list(y_column = "x", baseline = list(column = "phase", op = "le", value = 1)),
138       c(imr_outputs(rb, d$x, "All 70 points are evaluated against the baseline limits."),
139         list("/baseline_count" = exact(40L, "Forty baseline points."))),
140       c("Line 25-27: the baseline estimates the limits and every point is still plotted and evaluated."), kind = "axis_coverage")

Whole script generators/r/quality_spc.R


Missing and excluded rows are omitted before the moving ranges

Case STAT-QUAL-SPC-004 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_IMR_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/center_linetolerance50.58226060606060850.58226060606067.41E-1515passed
/primary/upper_control_limittolerance57.60371791835740457.60371791835747.15E-1515passed
/primary/lower_control_limittolerance43.56080329376381143.5608032937638046.67E-1515passed
/within_sigmatolerance2.34048577076559772.3404857707655984.61E-1615passed
/secondary/center_linetolerance2.64095538461538392.64095538461538444.71E-1615passed
/secondary/upper_control_limittolerance8.62676506126325658.6267650612632581.76E-1515passed
/secondary/lower_control_limittolerance00.00.00E+2passed
/primary/points/0/valuetolerance48.63329999999999848.63334.20E-1617passed
/secondary/points/0/valuetolerance2.39809999999999952.39809999999999954.57E-1716passed
/violation_keysset['primary:rule_1:17', 'primary:rule_2:35', 'primary:rule_3:4['primary:rule_1:17', 'primary:rule_2:35', 'primary:rule_3:4passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

143  m <- read_dataset("QUAL_SPC_IMR")
144  m$x[c(8, 61)] <- NA
145  export_dataset(m, "QUAL_SPC_IMR_MISSING", "QUAL_SPC_IMR with x missing in rows 8 and 61 (1-based)",
146                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
147                 c(t = "Int64", x = "Float64", phase = "Int64"))
148  mm <- read_dataset("QUAL_SPC_IMR_MISSING")
149  excluded <- c(2L, 40L)
150  xm <- mm$x[-(excluded + 1L)]; xm <- xm[!is.na(xm)]
151  emit("STAT-QUAL-SPC-004", "Missing and excluded rows are omitted before the moving ranges", "QUAL_SPC_IMR_MISSING",
152       list(y_column = "x"), imr_outputs(imr(xm, default_rules), xm, "Rules on the retained sequence."),
153       c("Line 75-77: missing observations are omitted and the adjacent retained values become successive.",
154         "Excluded rows (0-based 2 and 40) are removed from the plotted source."),
155       kind = "missing_excluded", excluded = excluded)

Whole script generators/r/quality_spc.R


Individuals chart at a 1e7 offset

Case STAT-QUAL-SPC-005 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_IMR_OFFSET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/center_linetolerance10000050.63048285810000050.6304828584.69E-1117passed
/primary/upper_control_limittolerance10000057.42612871910000057.4261287192.01E-1218passed
/primary/lower_control_limittolerance10000043.83483699710000043.8348369951.77E-915passed
/within_sigmatolerance2.26521528727586172.26521528725997671.59E-1111passed
/secondary/center_linetolerance2.55602173914835932.55602173913043451.79E-1111passed
/secondary/upper_control_limittolerance8.34932659732367818.3493265972651285.86E-1111passed
/secondary/lower_control_limittolerance00.00.00E+2passed
/primary/points/0/valuetolerance10000048.63330000110000048.63331.98E-1016passed
/secondary/points/0/valuetolerance2.39809999987483022.39811.25E-1010passed
/violation_keysset['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

157  h <- read_dataset("QUAL_SPC_IMR"); h$x <- h$x + 1e7
158  export_dataset(h, "QUAL_SPC_IMR_OFFSET", "QUAL_SPC_IMR shifted by 1e7", "R stats::rnorm documentation",
159                 "Constructed validation data; no third-party license.", c(t = "Int64", x = "Float64", phase = "Int64"))
160  ho <- read_dataset("QUAL_SPC_IMR_OFFSET")
161  ro <- imr(ho$x, default_rules)
162  emit("STAT-QUAL-SPC-005", "Individuals chart at a 1e7 offset", "QUAL_SPC_IMR_OFFSET", list(y_column = "x"),
163       imr_outputs(ro, ho$x, "The offset changes no signal."), c("A 1e7 shift moves the centre and limits only."), kind = "hard_numeric")

Whole script generators/r/quality_spc.R


X-bar and R with constant subgroups of 5

Case STAT-QUAL-SPC-006 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_SUB5.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/center_linetolerance19.97251999999999919.972523.85E-1616passed
/within_sigmatolerance1.4305204347615251.4305204347615272.02E-1514passed
/primary/points/0/valuetolerance19.8797819.879782.29E-1616passed
/secondary/points/0/valuetolerance2.0148000000000012.0148000000000013.46E-1716passed
/violation_keysset[][]passed
/subgroup_countexact990.00E+2passed
/primary/upper_control_limittolerance21.89176456119759421.8917645611975973.06E-1515passed
/primary/lower_control_limittolerance18.05327543880240518.05327543880243.29E-1515passed
/secondary/center_linetolerance3.3272888888888893.32728888888888854.82E-1615passed
/secondary/upper_control_limittolerance7.03554951104257237.0355495110426022.93E-1414passed
/secondary/lower_control_limittolerance00.00.00E+2passed
/dropped_subgroupsexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

225  set.seed(20261002)
226  xs <- round(rnorm(52, 20, 1.5), 4)
227  xs[13] <- NA                                  # makes subgroup 3 incomplete
228  export_dataset(data.frame(row = seq_along(xs), x = xs), "QUAL_SPC_SUB5",
229                 "Constructed: 52 rows, seed 20261002, x missing at row 13 (1-based)", "R stats::rnorm documentation",
230                 "Constructed validation data; no third-party license.", c(row = "Int64", x = "Float64"))
231  s5 <- read_dataset("QUAL_SPC_SUB5")$x
232  blocks <- split(s5[1:50], rep(1:10, each = 5))
233  kept <- Filter(function(g) !anyNA(g), blocks)
234  ch <- xbar_chart(unname(kept), "R", default_rules)
235  emit("STAT-QUAL-SPC-006", "X-bar and R with constant subgroups of 5", "QUAL_SPC_SUB5",
236       list(y_column = "x", subgroup_mode = "constant", subgroup_size = 5L),
237       c(xbar_outputs(ch), list("/dropped_subgroups" = exact(2L, "Subgroup 3 has a missing value and rows 51-52 are an incomplete final subgroup."))),
238       c("Line 44-48: subgroups are formed from consecutive row positions before missing values are assessed; incomplete ones are dropped and counted.",
239         sprintf("A2(5) = %.15g, D4(5) = %.15g from the integrals.", 3 / (d2(5) * sqrt(5)), 1 + 3 * d3(5) / d2(5))))

Whole script generators/r/quality_spc.R


X-bar and S with constant subgroups of 12

Case STAT-QUAL-SPC-007 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_SUB12.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/center_linetolerance7.00161833333333317.0016183333333334.48E-1717passed
/within_sigmatolerance0.371193556587707420.37119355658770722.24E-1615passed
/primary/points/0/valuetolerance6.90872166666666666.9087216666666668.38E-1615passed
/secondary/points/0/valuetolerance0.412597140514334990.412597140514334945.31E-1715passed
/violation_keysset[][]passed
/subgroup_countexact880.00E+2passed
/primary/upper_control_limittolerance7.32308138305938447.3230813830593841.39E-1717passed
/primary/lower_control_limittolerance6.68015528360728176.6801552836072822.43E-1717passed
/secondary/center_linetolerance0.362863732590546770.36286373259054675.30E-1715passed
/secondary/upper_control_limittolerance0.597450842646042530.597450842646042.55E-1514passed
/secondary/lower_control_limittolerance0.128276622535051010.128276622535053432.42E-1513passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

241  set.seed(20261003)
242  x12 <- round(rnorm(96, 7, 0.4), 5)
243  export_dataset(data.frame(x = x12, row = seq_along(x12)), "QUAL_SPC_SUB12", "Constructed: 96 rows, seed 20261003",
244                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.", c(x = "Float64", row = "Int64"))
245  s12 <- read_dataset("QUAL_SPC_SUB12")$x
246  ch <- xbar_chart(unname(split(s12, rep(1:8, each = 12))), "S", default_rules)
247  emit("STAT-QUAL-SPC-007", "X-bar and S with constant subgroups of 12", "QUAL_SPC_SUB12",
248       list(y_column = "x", subgroup_mode = "constant", subgroup_size = 12L), xbar_outputs(ch),
249       c("Sizes 11-25 use X-bar/S (line 46).", sprintf("c4(12) = %.15g.", c4(12))), kind = "axis_coverage")

Whole script generators/r/quality_spc.R


X-bar and R with the smallest subgroups of 2

Case STAT-QUAL-SPC-008 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_SUB12.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/center_linetolerance7.00161833333333317.0016183333333334.48E-1717passed
/within_sigmatolerance0.35574921241524610.35574921241524613.39E-1817passed
/primary/upper_control_limittolerance7.75627637483511567.7562763748351163.87E-1717passed
/secondary/upper_control_limittolerance1.31125124304083011.311251243040832.77E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

251  x2 <- read_dataset("QUAL_SPC_SUB12")$x
252  ch <- xbar_chart(unname(split(x2, rep(1:48, each = 2))), "R", default_rules)
253  emit("STAT-QUAL-SPC-008", "X-bar and R with the smallest subgroups of 2", "QUAL_SPC_SUB12",
254       list(y_column = "x", subgroup_mode = "constant", subgroup_size = 2L),
255       c(xbar_outputs(ch)[c("/primary/center_line", "/within_sigma")],
256         list("/primary/upper_control_limit" = tol(ch$ucl[1], "Grand mean + A2(2) R-bar."),
257              "/secondary/upper_control_limit" = tol(ch$s_ucl[1], "D4(2) R-bar."))),
258       c("All 96 rows form 48 subgroups of 2, the smallest subgroup size (line 44)."),
259       kind = "boundary")

Whole script generators/r/quality_spc.R


Grouping column with varying sizes (X-bar/R, point-specific limits)

Case STAT-QUAL-SPC-009 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_BYLOT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/center_linetolerance100.43124600000002100.431245999999993.26E-1415passed
/within_sigmatolerance2.33542142258375092.33542142258375574.84E-1514passed
/primary/points/0/valuetolerance100.6999100.69994.77E-1617passed
/secondary/points/0/valuetolerance6.2030000000000036.2030000000000034.41E-1717passed
/violation_keysset[][]passed
/subgroup_countexact10100.00E+2passed
/primary/point_upper_limitssequence['103.93437813387564', '103.29154140984751', '103.5645426342[103.93437813387563, 103.2915414098475, 103.56454263420392, passed
/primary/point_lower_limitssequence['96.928113866124392', '97.57095059015252', '97.297949365796[96.92811386612435, 97.57095059015248, 97.29794936579606, 96passed
/secondary/point_upper_limitssequence['10.97221936991426', '11.860511913292751', '11.486010719222[10.972219369914306, 11.8605119132928, 11.486010719222381, 1passed
/secondary/point_center_linessequence['4.8080505959871935', '5.9189217628128414', '5.432024290887[4.808050595987198, 5.918921762812844, 5.4320242908878145, 3passed
/dropped_subgroupsexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

262  set.seed(20261004)
263  sizes <- c(4, 6, 5, 3, 7, 5, 4, 6, 5, 1, 5)
264  lot <- rep(sprintf("L%02d", seq_along(sizes)), sizes)
265  vals <- round(rnorm(length(lot), 100, 3), 4)
266  export_dataset(data.frame(lot = lot, x = vals), "QUAL_SPC_BYLOT", "Constructed: lots of sizes 4,6,5,3,7,5,4,6,5,1,5, seed 20261004",
267                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.", c(lot = "String", x = "Float64"))
268  bl <- read_dataset("QUAL_SPC_BYLOT")
269  groups <- split(bl$x, factor(bl$lot, levels = unique(bl$lot)))
270  groups <- groups[vapply(groups, length, 1L) >= 2]
271  ch <- xbar_chart(unname(groups), "R", default_rules)
272  emit("STAT-QUAL-SPC-009", "Grouping column with varying sizes (X-bar/R, point-specific limits)", "QUAL_SPC_BYLOT",
273       list(y_column = "x", subgroup_mode = "by_column", group_column = "lot"),
274       c(xbar_outputs(ch, varying = TRUE), list("/dropped_subgroups" = exact(1L, "Lot L10 has one measurement."))),
275       c("Line 49-52 and 93-97: first-appearance order; sigma averages R_i/d2(n_i); the median size is 5, so X-bar/R."),
276       kind = "axis_coverage")

Whole script generators/r/quality_spc.R


Grouping column with a median size above 10 uses X-bar/S

Case STAT-QUAL-SPC-010 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_BYLOT_S.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/center_linetolerance3.00437181818181823.0043718181818181.08E-1818passed
/within_sigmatolerance0.192511430763598140.19251143076359812.74E-1715passed
/primary/points/0/valuetolerance3.03723916666666673.03723916666666634.32E-1615passed
/secondary/points/0/valuetolerance0.164162862839334230.164162862839334328.70E-1715passed
/violation_keysset[][]passed
/subgroup_countexact660.00E+2passed
/primary/point_upper_limitssequence['3.1710916077419835', '3.158724350386469', '3.1785049591901[3.171091607741983, 3.158724350386469, 3.1785049591901826, 3passed
/primary/point_lower_limitssequence['2.8376520286216529', '2.8500192859771674', '2.830238677173[2.8376520286216533, 2.8500192859771674, 2.830238677173454, passed
/secondary/point_upper_limitssequence['0.30985483041791551', '0.30097771397180728', '0.3152065766[0.3098548304179143, 0.30097771397180784, 0.3152065766456035passed
/secondary/point_center_linessequence['0.18819134948189775', '0.18884821482802699', '0.1877660388[0.1881913494818978, 0.18884821482802694, 0.1877660388466138passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

279  set.seed(20261005)
280  sizes <- c(12, 14, 11, 13, 12, 15)
281  lot <- rep(sprintf("B%d", seq_along(sizes)), sizes)
282  export_dataset(data.frame(lot = lot, x = round(rnorm(length(lot), 3, 0.2), 5)), "QUAL_SPC_BYLOT_S",
283                 "Constructed: batches of sizes 12,14,11,13,12,15, seed 20261005", "R stats::rnorm documentation",
284                 "Constructed validation data; no third-party license.", c(lot = "String", x = "Float64"))
285  bs <- read_dataset("QUAL_SPC_BYLOT_S")
286  ch <- xbar_chart(unname(split(bs$x, factor(bs$lot, levels = unique(bs$lot)))), "S", default_rules)
287  emit("STAT-QUAL-SPC-010", "Grouping column with a median size above 10 uses X-bar/S", "QUAL_SPC_BYLOT_S",
288       list(y_column = "x", subgroup_mode = "by_column", group_column = "lot"), xbar_outputs(ch, varying = TRUE),
289       c("Median usable size 12.5 > 10 selects X-bar/S; sigma averages s_i/c4(n_i) (line 119-120)."), kind = "axis_coverage")

Whole script generators/r/quality_spc.R


Spec limits are overlays only

Case STAT-QUAL-SPC-011 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_IMR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/lower_spec_limittolerance40400.00E+2passed
/upper_spec_limittolerance60600.00E+2passed
/violation_keysset['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4['primary:rule_1:19', 'primary:rule_2:37', 'primary:rule_3:4passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

291  emit("STAT-QUAL-SPC-011", "Spec limits are overlays only", "QUAL_SPC_IMR",
292       list(y_column = "x", lower_spec = 40, upper_spec = 60),
293       list("/lower_spec_limit" = tol(40, "Passed through as an overlay."), "/upper_spec_limit" = tol(60, "Passed through as an overlay."),
294            "/violation_keys" = set_of(imr(d$x, default_rules)$keys, "Spec limits never create signals (line 30-31).")),
295       c("Line 30-31: applied specifications never replace limits or create signals."), kind = "axis_coverage")

Whole script generators/r/quality_spc.R


A subgroup size of 26 is refused

Case STAT-QUAL-SPC-012 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_SUB12.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

297  emit("STAT-QUAL-SPC-012", "A subgroup size of 26 is refused", "QUAL_SPC_SUB12",
298       list(y_column = "x", subgroup_mode = "constant", subgroup_size = 26L), list(),
299       c("Line 44: sizes 2-25."), kind = "invalid_input", error = list(error_class = "ValueError", message_contains = "25"))

Whole script generators/r/quality_spc.R


An unknown run rule is refused

Case STAT-QUAL-SPC-013 · status: passed

Reference

Evidence type
R
Source
R stats::integrate constants and hand-coded chart formulas and rules
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R stats::integrate and base::lgamma
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc.R · sha256 933d3416ca032dbb
Dataset
reference_data/datasets/QUAL_SPC_IMR.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

300  emit("STAT-QUAL-SPC-013", "An unknown run rule is refused", "QUAL_SPC_IMR",
301       list(y_column = "x", run_rules = list("rule_9")), list(),
302       c("Only rules 2-8 are optional (line 219-236)."), kind = "invalid_input",
303       error = list(error_class = "ValueError", message_contains = "Unknown run rules"))

Whole script generators/r/quality_spc.R


Default normalization matches through whitespace and case

Case STAT-QUAL-SPECMATCH-001 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation from the constructed spec sheet
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/quality_spec_import.R · sha256 3e2833e638a2586d
Dataset
reference_data/datasets/QUAL_SPEC_SHEET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/matches/0/variableexact Diameter A Diameter Apassed
/matches/0/columnexactDiameterADiameterApassed
/matches/0/lowertolerance9.90000000000000049.94.47E-1717passed
/matches/0/uppertolerance10.110.13.55E-1616passed
/matches/0/targettolerance1010.00.00E+2passed
/matches/0/unitsexactmmmmpassed
/matches/1/columnexactLengthLengthpassed
/matches/1/lowerabsentNonepassed
/matches/1/uppertolerance50.550.50.00E+1passed
/matches/2absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

44  sheet <- data.frame(
45    variable = c(" Diameter A", "length", "Width", "Height", "Mass", "Unknown"),
46    LSL = c("9.9", "", "3", "abc", "", "1"),
47    USL = c("10.1", "50.5", "2", "7", "", "2"),
48    Target = c("10", "", "", "", "", ""),
49    Units = c("mm", "mm", "mm", "mm", "g", "mm"),
50    stringsAsFactors = FALSE)
51  export_dataset(sheet, "QUAL_SPEC_SHEET", "Constructed specification sheet with whitespace, case, one-sided, inverted, malformed, empty and unmatched rows",
52                 "Hand-constructed", "Constructed validation data; no third-party license.",
53                 c(variable = "String", LSL = "String", USL = "String", Target = "String", Units = "String"))
54  targets <- list("DiameterA", "Length", "Width", "Height", "Mass", "Other")
55  base <- list(target_columns = targets, name_column = "variable", lsl_column = "LSL", usl_column = "USL",
56               target_column = "Target", units_column = "Units")
57  emit("STAT-QUAL-SPECMATCH-001", match_entry, "Default normalization matches through whitespace and case", "QUAL_SPEC_SHEET", base,
58       list("/matches/0/variable" = exact(" Diameter A", "The first resolved row."),
59            "/matches/0/column" = exact("DiameterA", "' Diameter A' without whitespace and case equals 'DiameterA'."),
60            "/matches/0/lower" = num(9.9, "LSL 9.9."), "/matches/0/upper" = num(10.1, "USL 10.1."),
61            "/matches/0/target" = num(10, "Target 10."), "/matches/0/units" = exact("mm", "Units retained."),
62            "/matches/1/column" = exact("Length", "'length' matches 'Length' case-insensitively."),
63            "/matches/1/lower" = absent("A blank LSL is an absent side (line 47-48)."),
64            "/matches/1/upper" = num(50.5, "USL 50.5, a one-sided specification."),
65            "/matches/2" = absent("Only two rows resolve.")),
66       c("Width has LSL 3 > USL 2 (inverted); Height has the malformed LSL 'abc' and is unresolved even though USL 7 parses (line 48-50);",
67         "Mass has neither limit; Unknown names no target column. None of these cancels the two valid rows (line 51-52)."))

Whole script generators/r/quality_spec_import.R


Case-sensitive matching leaves 'length' unmatched

Case STAT-QUAL-SPECMATCH-002 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation from the constructed spec sheet
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/quality_spec_import.R · sha256 3e2833e638a2586d
Dataset
reference_data/datasets/QUAL_SPEC_SHEET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/matches/0/columnexactDiameterADiameterApassed
/matches/1absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

68  emit("STAT-QUAL-SPECMATCH-002", match_entry, "Case-sensitive matching leaves 'length' unmatched", "QUAL_SPEC_SHEET",
69       modifyList(base, list(case_insensitive = FALSE)),
70       list("/matches/0/column" = exact("DiameterA", "Whitespace is still removed and the letters match exactly."),
71            "/matches/1" = absent("'length' does not match 'Length' when case matters.")),
72       c("case_insensitive = false (line 26)."), kind = "axis_coverage")

Whole script generators/r/quality_spec_import.R


Keeping whitespace leaves ' Diameter A' unmatched

Case STAT-QUAL-SPECMATCH-003 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation from the constructed spec sheet
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/quality_spec_import.R · sha256 3e2833e638a2586d
Dataset
reference_data/datasets/QUAL_SPEC_SHEET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/matches/0/columnexactLengthLengthpassed
/matches/1absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

73  emit("STAT-QUAL-SPECMATCH-003", match_entry, "Keeping whitespace leaves ' Diameter A' unmatched", "QUAL_SPEC_SHEET",
74       modifyList(base, list(ignore_whitespace = FALSE)),
75       list("/matches/0/column" = exact("Length", "Only 'length' still matches."),
76            "/matches/1" = absent("' Diameter A' keeps its spaces and matches nothing.")),
77       c("ignore_whitespace = false (line 24-25)."), kind = "axis_coverage")

Whole script generators/r/quality_spec_import.R


Neither limit role assigned is refused

Case STAT-QUAL-SPECMATCH-004 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation from the constructed spec sheet
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/quality_spec_import.R · sha256 3e2833e638a2586d
Dataset
reference_data/datasets/QUAL_SPEC_SHEET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

78  emit("STAT-QUAL-SPECMATCH-004", match_entry, "Neither limit role assigned is refused", "QUAL_SPEC_SHEET",
79       list(target_columns = targets, name_column = "variable"), list(),
80       c("Line 32-33: at least one of LSL or USL must be assigned."), kind = "invalid_input",
81       error = list(error_class = "ValueError", message_contains = "lsl_column or usl_column"))

Whole script generators/r/quality_spec_import.R


Matching a transposed sheet

Case STAT-QUAL-SPECMATCH-005 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation from the constructed spec sheet
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/quality_spec_import.R · sha256 3e2833e638a2586d
Dataset
reference_data/datasets/QUAL_SPEC_WIDE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/matches/0/columnexactalphaalphapassed
/matches/0/lowertolerance11.00.00E+2passed
/matches/1/lowerabsentNonepassed
/matches/1/uppertolerance99.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 97  emit("STAT-QUAL-SPECMATCH-005", match_entry, "Matching a transposed sheet", "QUAL_SPEC_WIDE",
 98       list(transposed = TRUE, target_columns = list("alpha", "beta"), name_column = "variable", lsl_column = "LSL", usl_column = "USL"),
 99       list("/matches/0/column" = exact("alpha", "'Alpha' matches 'alpha' case-insensitively."),
100            "/matches/0/lower" = num(1, "LSL 1."), "/matches/1/lower" = absent("Beta has a blank LSL."),
101            "/matches/1/upper" = num(9, "USL 9.")),
102       c("Prepare then match, as the desktop does."), kind = "axis_coverage")

Whole script generators/r/quality_spec_import.R


Transposing a sheet laid out across columns

Case STAT-QUAL-SPECPREP-001 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation from the constructed spec sheet
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/quality_spec_import.R · sha256 3e2833e638a2586d
Dataset
reference_data/datasets/QUAL_SPEC_WIDE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/variableexactAlphaAlphapassed
/0/LSLexact11passed
/1/variableexactBetaBetapassed
/1/USLexact99passed
/2absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

83  wide <- data.frame(field = c("variable", "LSL", "USL"), c1 = c("Alpha", "1", "2"), c2 = c("Beta", "", "9"),
84                     stringsAsFactors = FALSE)
85  export_dataset(wide, "QUAL_SPEC_WIDE", "Constructed specification sheet laid out across columns", "Hand-constructed",
86                 "Constructed validation data; no third-party license.", c(field = "String", c1 = "String", c2 = "String"))
87  emit("STAT-QUAL-SPECPREP-001", prep_entry, "Transposing a sheet laid out across columns", "QUAL_SPEC_WIDE",
88       list(transposed = TRUE),
89       list("/0/variable" = exact("Alpha", "The first characteristic becomes the first row."),
90            "/0/LSL" = exact("1", "Its LSL."), "/1/variable" = exact("Beta", "The second characteristic."),
91            "/1/USL" = exact("9", "Its USL."), "/2" = absent("Two characteristics become two rows.")),
92       c("The first source column (variable, LSL, USL) supplies the new headings (line 20-22)."))

Whole script generators/r/quality_spec_import.R


Without transposing the sheet is unchanged

Case STAT-QUAL-SPECPREP-002 · status: passed

Reference

Evidence type
Analytical
Source
Hand derivation from the constructed spec sheet
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
generators/r/quality_spec_import.R · sha256 3e2833e638a2586d
Dataset
reference_data/datasets/QUAL_SPEC_WIDE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/fieldexactvariablevariablepassed
/2/c2exact99passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

93  emit("STAT-QUAL-SPECPREP-002", prep_entry, "Without transposing the sheet is unchanged", "QUAL_SPEC_WIDE",
94       list(transposed = FALSE),
95       list("/0/field" = exact("variable", "The original first row."), "/2/c2" = exact("9", "The original cell.")),
96       c("transposed = false is the default (line 20)."), kind = "axis_coverage")

Whole script generators/r/quality_spec_import.R


Two process stages recalculated independently

Case STAT-QUAL-STAGES-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded I-MR charts and rules in R
Citation
NIST/SEMATECH e-Handbook 6.3.2.1, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm; Nelson, Journal of Quality Technology 16(4), 1984, pp. 237-239; R base arithmetic
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_stages.R · sha256 046de2272a987401
Dataset
reference_data/datasets/QUAL_SPC_STAGES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/point_center_linessequence['19.869526666666665', '19.869526666666665', '19.86952666666[19.86952666666667, 19.86952666666667, 19.86952666666667, 19passed
/primary/point_upper_limitssequence['23.599142395471546', '23.599142395471546', '23.59914239547[23.599142395471546, 23.599142395471546, 23.599142395471546,passed
/primary/point_lower_limitssequence['16.139910937861785', '16.139910937861785', '16.13991093786[16.139910937861792, 16.139910937861792, 16.139910937861792,passed
/violation_keysset['primary:rule_1:11', 'primary:rule_1:46', 'secondary:rule_1['primary:rule_1:11', 'primary:rule_1:46', 'secondary:rule_1passed
/stage_countexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

79  a <- stage(d$x[1:30]); b <- stage(d$x[31:60])
80  keys <- c(rules_for(a$values, a$center, a$sigma, a$ucl, a$lcl, 0L),
81            rules_for(b$values, b$center, b$sigma, b$ucl, b$lcl, 30L),
82            rules_for(a$mr, rep(0, 29), 1, a$mr_ucl, rep(0, 29), 0L, "secondary", FALSE),
83            rules_for(b$mr, rep(0, 29), 1, b$mr_ucl, rep(0, 29), 29L, "secondary", FALSE))
84  emit("STAT-QUAL-STAGES-001", stage_entry, "Two process stages recalculated independently",
85       list(y_column = "x", x_column = "label",
86            stages = list(list(name = "Before", start_label = "P01", end_label = "P30"),
87                          list(name = "After", start_label = "P31"))),
88       list("/primary/point_center_lines" = seqtol(c(a$center, b$center), "Each stage's own mean."),
89            "/primary/point_upper_limits" = seqtol(c(a$ucl, b$ucl), "Each stage's mean + 3 MR-bar/d2."),
90            "/primary/point_lower_limits" = seqtol(c(a$lcl, b$lcl), "Each stage's mean - 3 MR-bar/d2."),
91            "/violation_keys" = set_of(keys, "Rules restart at the stage boundary; moving ranges stay within a stage."),
92            "/stage_count" = exact(2L, "Two calculated stages.")),
93       c("Stage 'After' starts at label P31 with no end label, so it runs to the last point.",
94         "Moving ranges are formed within each stage: 29 + 29 points on the MR chart."))

Whole script generators/r/quality_spc_stages.R


Individual Phase II T-squared at alpha 0.0027

Case STAT-QUAL-T2-001 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/baseline/meansequence['9.8712924999999991', '20.095361749999999', '4.778065749999[9.871292500000004, 20.095361750000002, 4.778065749999999]passed
/baseline/covariance/0sequence['0.44568973674743589', '0.24653224772884599', '0.2109089705[0.4456897367474356, 0.24653224772884605, 0.2109089705878206passed
/baseline/upper_control_limittolerance18.3463626692310718.3463626692310521.80E-1415passed
/monitoring_points/0/t_squaredtolerance2.31565264258925342.31565264258924547.99E-1514passed
/monitoring_points/15/t_squaredtolerance34.55081776270690634.550817762706819.95E-1414passed
/breach_countexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

63  set.seed(20261012)
64  L <- matrix(c(1, 0, 0, 0.6, 0.8, 0, 0.3, -0.2, 0.9), 3, byrow = TRUE)
65  ref <- matrix(rnorm(40 * 3), 40) %*% t(L)
66  mon <- matrix(rnorm(20 * 3), 20) %*% t(L); mon[11:20, 1] <- mon[11:20, 1] + 2.2
67  X <- round(rbind(ref, mon) + matrix(rep(c(10, 20, 5), each = 60), 60), 5)
68  export_dataset(data.frame(a = X[, 1], b = X[, 2], c = X[, 3], role = c(rep("ref", 40), rep("mon", 20)),
69                            stage = c(rep("ref", 40), rep("S1", 10), rep("S2", 10))),
70                 "QUAL_T2", "Constructed: 40 reference and 20 monitoring rows of three correlated variables, seed 20261012; a +2.2 shift in a from monitoring row 11",
71                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
72                 c(a = "Float64", b = "Float64", c = "Float64", role = "String", stage = "String"))
73  d <- read_dataset("QUAL_T2")
74  vars <- c("a", "b", "c")
75  R0 <- as.matrix(d[d$role == "ref", vars]); M0 <- as.matrix(d[d$role == "mon", vars])
76  m <- nrow(R0); p <- 3L; xbar <- colMeans(R0); S <- cov(R0); Si <- solve(S)
77  t2 <- apply(M0, 1, function(x) quad(x - xbar, Si))
78  ucl_ind <- function(alpha) p * (m + 1) * (m - 1) / (m * (m - p)) * qf(1 - alpha, p, m - p)
79  for (alpha in c(0.0027, 0.01)) {
80    id <- if (alpha == 0.0027) "STAT-QUAL-T2-001" else "STAT-QUAL-T2-002"
81    emit(id, t2_entry, "hotellingMenuItem", sprintf("Individual Phase II T-squared at alpha %g", alpha), "QUAL_T2",
82         list(variables = as.list(vars), reference_column = "role", reference_value = "ref", alpha = alpha),
83         list("/baseline/mean" = seqtol(xbar, "Reference mean vector."),
84              "/baseline/covariance/0" = seqtol(S[1, ], "First row of S, denominator m - 1."),
85              "/baseline/upper_control_limit" = tol(ucl_ind(alpha), "p(m+1)(m-1)/(m(m-p)) F_(1-alpha)(p, m-p)."),
86              "/monitoring_points/0/t_squared" = tol(t2[1], "First monitoring T^2."),
87              "/monitoring_points/15/t_squared" = tol(t2[16], "A shifted monitoring row."),
88              "/breach_count" = exact(sum(t2 > ucl_ind(alpha)), "Rows with T^2 above the limit.")),
89         c("The standard Phase-II individual-observation limit is used (Montgomery, 6th ed., section 11.3)."),
90         kind = if (alpha == 0.0027) "nominal" else "threshold")

Whole script generators/r/quality_multivariate_spc.R


Individual Phase II T-squared at alpha 0.01

Case STAT-QUAL-T2-002 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/baseline/meansequence['9.8712924999999991', '20.095361749999999', '4.778065749999[9.871292500000004, 20.095361750000002, 4.778065749999999]passed
/baseline/covariance/0sequence['0.44568973674743589', '0.24653224772884599', '0.2109089705[0.4456897367474356, 0.24653224772884605, 0.2109089705878206passed
/baseline/upper_control_limittolerance14.13021104913324614.130211049133245.34E-1515passed
/monitoring_points/0/t_squaredtolerance2.31565264258925342.31565264258924547.99E-1514passed
/monitoring_points/15/t_squaredtolerance34.55081776270690634.550817762706819.95E-1414passed
/breach_countexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

63  set.seed(20261012)
64  L <- matrix(c(1, 0, 0, 0.6, 0.8, 0, 0.3, -0.2, 0.9), 3, byrow = TRUE)
65  ref <- matrix(rnorm(40 * 3), 40) %*% t(L)
66  mon <- matrix(rnorm(20 * 3), 20) %*% t(L); mon[11:20, 1] <- mon[11:20, 1] + 2.2
67  X <- round(rbind(ref, mon) + matrix(rep(c(10, 20, 5), each = 60), 60), 5)
68  export_dataset(data.frame(a = X[, 1], b = X[, 2], c = X[, 3], role = c(rep("ref", 40), rep("mon", 20)),
69                            stage = c(rep("ref", 40), rep("S1", 10), rep("S2", 10))),
70                 "QUAL_T2", "Constructed: 40 reference and 20 monitoring rows of three correlated variables, seed 20261012; a +2.2 shift in a from monitoring row 11",
71                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
72                 c(a = "Float64", b = "Float64", c = "Float64", role = "String", stage = "String"))
73  d <- read_dataset("QUAL_T2")
74  vars <- c("a", "b", "c")
75  R0 <- as.matrix(d[d$role == "ref", vars]); M0 <- as.matrix(d[d$role == "mon", vars])
76  m <- nrow(R0); p <- 3L; xbar <- colMeans(R0); S <- cov(R0); Si <- solve(S)
77  t2 <- apply(M0, 1, function(x) quad(x - xbar, Si))
78  ucl_ind <- function(alpha) p * (m + 1) * (m - 1) / (m * (m - p)) * qf(1 - alpha, p, m - p)
79  for (alpha in c(0.0027, 0.01)) {
80    id <- if (alpha == 0.0027) "STAT-QUAL-T2-001" else "STAT-QUAL-T2-002"
81    emit(id, t2_entry, "hotellingMenuItem", sprintf("Individual Phase II T-squared at alpha %g", alpha), "QUAL_T2",
82         list(variables = as.list(vars), reference_column = "role", reference_value = "ref", alpha = alpha),
83         list("/baseline/mean" = seqtol(xbar, "Reference mean vector."),
84              "/baseline/covariance/0" = seqtol(S[1, ], "First row of S, denominator m - 1."),
85              "/baseline/upper_control_limit" = tol(ucl_ind(alpha), "p(m+1)(m-1)/(m(m-p)) F_(1-alpha)(p, m-p)."),
86              "/monitoring_points/0/t_squared" = tol(t2[1], "First monitoring T^2."),
87              "/monitoring_points/15/t_squared" = tol(t2[16], "A shifted monitoring row."),
88              "/breach_count" = exact(sum(t2 > ucl_ind(alpha)), "Rows with T^2 above the limit.")),
89         c("The standard Phase-II individual-observation limit is used (Montgomery, 6th ed., section 11.3)."),
90         kind = if (alpha == 0.0027) "nominal" else "threshold")

Whole script generators/r/quality_multivariate_spc.R


Excluded and incomplete rows leave the baseline and the monitoring

Case STAT-QUAL-T2-003 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/baseline/meansequence['9.8533569230769231', '20.062712564102565', '4.766203076923[9.853356923076927, 20.062712564102565, 4.766203076923077]passed
/baseline/upper_control_limittolerance18.47930087053000718.4793008705300074.72E-1616passed
/monitoring_points/0/t_squaredtolerance2.23233554652365212.2323355465236391.29E-1414passed
/incomplete_countexact110.00E+2passed
/excluded_countexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 92  mm <- read_dataset("QUAL_T2"); mm$b[45] <- NA
 93  export_dataset(mm, "QUAL_T2_MISSING", "QUAL_T2 with b missing in row 45 (1-based)", "R stats::rnorm documentation",
 94                 "Constructed validation data; no third-party license.",
 95                 c(a = "Float64", b = "Float64", c = "Float64", role = "String", stage = "String"))
 96  dm <- read_dataset("QUAL_T2_MISSING")
 97  excluded <- c(3L, 52L)
 98  keep <- setdiff(which(stats::complete.cases(dm[vars])), excluded + 1L)
 99  Rk <- as.matrix(dm[intersect(keep, which(dm$role == "ref")), vars]); Mk <- as.matrix(dm[intersect(keep, which(dm$role == "mon")), vars])
100  xk <- colMeans(Rk); Sk <- solve(cov(Rk)); mk <- nrow(Rk)
101  emit("STAT-QUAL-T2-003", t2_entry, "hotellingMenuItem", "Excluded and incomplete rows leave the baseline and the monitoring", "QUAL_T2_MISSING",
102       list(variables = as.list(vars), reference_column = "role", reference_value = "ref", alpha = 0.0027),
103       list("/baseline/mean" = seqtol(xk, "Mean of the 39 remaining reference rows."),
104            "/baseline/upper_control_limit" = tol(p * (mk + 1) * (mk - 1) / (mk * (mk - p)) * qf(1 - 0.0027, p, mk - p), "Limit at m = 39."),
105            "/monitoring_points/0/t_squared" = tol(quad(Mk[1, ] - xk, Sk), "First remaining monitoring row."),
106            "/incomplete_count" = exact(1L, "Row 45 (1-based) has a missing value."),
107            "/excluded_count" = exact(2L, "Two excluded rows.")),
108       c("Reference row 3 and monitoring row 52 (0-based) are excluded; monitoring row 44 (0-based) is incomplete."),
109       kind = "missing_excluded", excluded = excluded)

Whole script generators/r/quality_multivariate_spc.R


Three reference rows for three variables are too few

Case STAT-QUAL-T2-004 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2_FEWREF.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

110  small <- read_dataset("QUAL_T2"); small$role[4:40] <- "mon"
111  export_dataset(small, "QUAL_T2_FEWREF", "QUAL_T2 with only three reference rows", "Hand-constructed",
112                 "Constructed validation data; no third-party license.",
113                 c(a = "Float64", b = "Float64", c = "Float64", role = "String", stage = "String"))
114  emit("STAT-QUAL-T2-004", t2_entry, "hotellingMenuItem", "Three reference rows for three variables are too few", "QUAL_T2_FEWREF",
115       list(variables = as.list(vars), reference_column = "role", reference_value = "ref"), list(),
116       c("The baseline needs more reference rows than variables (m > p)."), kind = "invalid_input",
117       error = list(error_class = "ValueError", message_contains = "reference rows"))

Whole script generators/r/quality_multivariate_spc.R


Subgroup T-squared and generalized variance

Case STAT-QUAL-T2SUB-001 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2_SUB.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/baseline/grand_meansequence['5.0256395999999999', '5.2890446666666664'][5.0256396, 5.289044666666668]passed
/baseline/pooled_covariance/0sequence['1.0006864177346666', '0.25431099362000004'][1.0006864177346666, 0.25431099362000004]passed
/baseline/t_squared_ucltolerance14.20821769389992514.2082176938999131.20E-1415passed
/baseline/f_denominator_dfexact59590.00E+2passed
/monitoring_points/0/t_squaredtolerance4.98191841491019544.98191841491021051.51E-1414passed
/monitoring_points/7/t_squaredtolerance10.57307595925280810.5730759592528157.18E-1515passed
/monitoring_points/0/generalized_variancetolerance0.616927191255437050.6169271912554373.72E-1817passed
/baseline/generalized_variance_centertolerance1.03102442715327511.03102442715327513.59E-1716passed
/baseline/generalized_variance_ucltolerance4.81925006545954914.8192500654595494.83E-1716passed
/t_squared_breach_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

207  export_dataset(subgroup_data(20261013, 2, 5, 15, 8, 1.6), "QUAL_T2_SUB", "Constructed: p = 2, n = 5, 15 reference and 8 monitoring subgroups, seed 20261013; +1.6 shift in v1 in the last four",
208                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
209                 c(v1 = "Float64", v2 = "Float64", lot = "String", role = "String"))
210  ds <- read_dataset("QUAL_T2_SUB")
211  sc <- sub_chart(ds, c("v1", "v2"), 5)
212  sub_config <- list(variables = list("v1", "v2"), subgroup_column = "lot", subgroup_size = 5L,
213                     reference_column = "role", reference_value = "ref")
214  emit("STAT-QUAL-T2SUB-001", sub_entry, "hotellingSubgroupsMenuItem", "Subgroup T-squared and generalized variance", "QUAL_T2_SUB",
215       sub_config,
216       list("/baseline/grand_mean" = seqtol(sc$gm, "Mean of the reference subgroup means."),
217            "/baseline/pooled_covariance/0" = seqtol(sc$Sbar[1, ], "Mean of the reference subgroup covariances."),
218            "/baseline/t_squared_ucl" = tol(sc$ucl, "p(m+1)(n-1)/(mn-m-p+1) F_(1-alpha)(p, mn-m-p+1)."),
219            "/baseline/f_denominator_df" = exact(as.integer(sc$df2), "mn - m - p + 1."),
220            "/monitoring_points/0/t_squared" = tol(sc$t2[[1]], "n (xbar_j - xbarbar)' Sbar^-1 (xbar_j - xbarbar)."),
221            "/monitoring_points/7/t_squared" = tol(sc$t2[[8]], "A shifted subgroup."),
222            "/monitoring_points/0/generalized_variance" = tol(sc$gv[[1]], "|S_j|."),
223            "/baseline/generalized_variance_center" = tol(sc$center, "Mean reference |S_i|."),
224            "/baseline/generalized_variance_ucl" = tol(sc$gv_ucl, "|Sigma| (b1 + 3 sqrt(b2))."),
225            "/t_squared_breach_count" = exact(sum(sc$t2 > sc$ucl), "Subgroups above the limit.")),
226       c("alpha defaults to 0.0027 (line 24-25).", "Alt's b1 and b2 for n = 5, p = 2 (line 70-73)."))

Whole script generators/r/quality_multivariate_spc.R


n <= p keeps the location chart and withholds the spread chart

Case STAT-QUAL-T2SUB-002 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2_SUB_SMALLN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/baseline/t_squared_ucltolerance22.82553880185447822.8255388018544856.92E-1515passed
/monitoring_points/0/t_squaredtolerance1.23574119895847081.23574119895847062.34E-1615passed
/baseline/generalized_variance_availableexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

227  export_dataset(subgroup_data(20261014, 3, 3, 12, 6, 0), "QUAL_T2_SUB_SMALLN", "Constructed: p = 3, n = 3, seed 20261014",
228                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
229                 c(v1 = "Float64", v2 = "Float64", v3 = "Float64", lot = "String", role = "String"))
230  dn <- read_dataset("QUAL_T2_SUB_SMALLN")
231  sn <- sub_chart(dn, c("v1", "v2", "v3"), 3)
232  emit("STAT-QUAL-T2SUB-002", sub_entry, "hotellingSubgroupsMenuItem", "n <= p keeps the location chart and withholds the spread chart", "QUAL_T2_SUB_SMALLN",
233       list(variables = list("v1", "v2", "v3"), subgroup_column = "lot", subgroup_size = 3L, reference_column = "role", reference_value = "ref"),
234       list("/baseline/t_squared_ucl" = tol(sn$ucl, "Location limit still computed."),
235            "/monitoring_points/0/t_squared" = tol(sn$t2[[1]], "Location statistic."),
236            "/baseline/generalized_variance_available" = exact(FALSE, "Every subgroup covariance is singular when n <= p (line 83-85).")),
237       c("m(n - 1) = 24 >= p = 3, so the location chart is defined."), kind = "boundary")

Whole script generators/r/quality_multivariate_spc.R


A subgroup size of 1 is refused

Case STAT-QUAL-T2SUB-003 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2_SUB.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

238  emit("STAT-QUAL-T2SUB-003", sub_entry, "hotellingSubgroupsMenuItem", "A subgroup size of 1 is refused", "QUAL_T2_SUB",
239       modifyList(sub_config, list(subgroup_size = 1L)), list(),
240       c("Line 22-24: a size of 1 points to the individual-observation chart."), kind = "invalid_input",
241       error = list(error_class = "ValueError", message_contains = "individual"))

Whole script generators/r/quality_multivariate_spc.R


Subgroups larger than n stop the run

Case STAT-QUAL-T2SUB-004 · status: passed

Reference

Evidence type
R
Source
R base linear algebra (solve, det, cov) and stats::qf
Citation
Montgomery, Introduction to Statistical Quality Control, 6th ed., sec. 11.3 and Example 11.1; Alt, Multivariate Quality Control, Encyclopedia of Statistical Sciences 6, 1985, pp. 110-122; Lowry et al., Technometrics 34(1), 1992, pp. 46-53; R base::solve, base::det, stats::cov and stats::qf
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_multivariate_spc.R · sha256 07605e0a04ab57ce
Dataset
reference_data/datasets/QUAL_T2_SUB.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

242  emit("STAT-QUAL-T2SUB-004", sub_entry, "hotellingSubgroupsMenuItem", "Subgroups larger than n stop the run", "QUAL_T2_SUB",
243       modifyList(sub_config, list(subgroup_size = 4L)), list(),
244       c("Line 36-38: a subgroup with more than n rows stops the run."), kind = "invalid_input",
245       error = list(error_class = "ValueError", message_contains = "G01"))

Whole script generators/r/quality_multivariate_spc.R


CUSUM with target and sigma from the data

Case STAT-QUAL-TW-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.2.3-4, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc323.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc324.htm; Montgomery, Introduction to Statistical Quality Control, 6th ed., ch. 9
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_TW.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/targettolerance10.28536410.2853644.93E-1616passed
/sigmatolerance0.770838371132430590.77083837113243051.12E-1615passed
/primary/points/49/valuetolerance3.8310849587716363.83108495877163342.65E-1515passed
/secondary/points/49/valuetolerance00.00.00E+2passed
/primary/upper_control_limittolerance3.85419185566215283.85419185566215244.12E-1615passed
/violation_keysset[][]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

226  t0 <- mean(tw); s0 <- mean(abs(diff(tw))) / d2_2
227  cu <- cusum(tw, t0, s0, 0.5, 5)
228  emit("STAT-QUAL-TW-001", tw_entry, "CUSUM with target and sigma from the data", "QUAL_TW", list(y_column = "y"),
229       list("/target" = tol(t0, "Baseline mean."), "/sigma" = tol(s0, "MR-bar / d2(2)."),
230            "/primary/points/49/value" = tol(cu$up[50], "Final upper accumulator C+."),
231            "/secondary/points/49/value" = tol(cu$lo[50], "Final lower accumulator C-."),
232            "/primary/upper_control_limit" = tol(cu$H, "Decision interval H = h sigma."),
233            "/violation_keys" = set_of(cu$keys, "Points where an accumulator exceeds H.")),
234       c("Defaults k = 0.5 and h = 5 (line 180-182); accumulators start at zero."))

Whole script generators/r/quality_spc_attributes.R


CUSUM with a supplied target and sigma

Case STAT-QUAL-TW-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.2.3-4, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc323.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc324.htm; Montgomery, Introduction to Statistical Quality Control, 6th ed., ch. 9
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_TW.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/points/30/valuetolerance2.72470000000000122.7247000000000029.20E-1615passed
/violation_keysset['primary:rule_1:39', 'primary:rule_1:40', 'primary:rule_1:4['primary:rule_1:39', 'primary:rule_1:40', 'primary:rule_1:4passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

235  cu2 <- cusum(tw, 10, 1.1, 0.25, 4)
236  emit("STAT-QUAL-TW-002", tw_entry, "CUSUM with a supplied target and sigma", "QUAL_TW",
237       list(y_column = "y", target = 10, sigma = 1.1, cusum_k = 0.25, cusum_h = 4),
238       list("/primary/points/30/value" = tol(cu2$up[31], "C+ at point 31."),
239            "/violation_keys" = set_of(cu2$keys, "Points where an accumulator exceeds H.")),
240       c("Supplied target 10 and sigma 1.1 override the baseline estimates (line 173-176)."), kind = "axis_coverage")

Whole script generators/r/quality_spc_attributes.R


EWMA with finite-time limits

Case STAT-QUAL-TW-003 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.2.3-4, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc323.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc324.htm; Montgomery, Introduction to Statistical Quality Control, 6th ed., ch. 9
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_TW.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/points/0/valuetolerance9.98375120000000049.98375121.91E-1717passed
/primary/points/49/valuetolerance11.0081200577166711.008120057716674.59E-1616passed
/primary/point_upper_limitssequence['10.747867022679458', '10.877656862880958', '10.94750294361[10.747867022679458, 10.877656862880958, 10.947502943618526,passed
/primary/point_lower_limitssequence['9.8228609773205413', '9.6930711371190412', '9.623225056381[9.822860977320541, 9.693071137119041, 9.623225056381473, 9.passed
/violation_keysset['primary:rule_1:48']['primary:rule_1:48']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

241  ew <- ewma(tw, t0, s0, 0.2, 3)
242  emit("STAT-QUAL-TW-003", tw_entry, "EWMA with finite-time limits", "QUAL_TW", list(y_column = "y", method = "ewma"),
243       list("/primary/points/0/value" = tol(ew$z[1], "z_1 = lambda x_1 + (1 - lambda) T."),
244            "/primary/points/49/value" = tol(ew$z[50], "z_50."),
245            "/primary/point_upper_limits" = seqtol(ew$ucl, "T + L sigma sqrt(lambda/(2-lambda) (1 - (1-lambda)^(2i)))."),
246            "/primary/point_lower_limits" = seqtol(ew$lcl, "Mirror of the upper limits."),
247            "/violation_keys" = set_of(ew$keys, "Points beyond their limits.")),
248       c("Defaults lambda = 0.2 and L = 3 (line 195-196)."), kind = "axis_coverage")

Whole script generators/r/quality_spc_attributes.R


EWMA with lambda 0.1, L 2.7 and a supplied target

Case STAT-QUAL-TW-004 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.2.3-4, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc323.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc324.htm; Montgomery, Introduction to Statistical Quality Control, 6th ed., ch. 9
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_TW.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/primary/points/49/valuetolerance10.77923653252324410.7792365325232442.07E-1616passed
/primary/point_upper_limitssequence['10.208126360205757', '10.280005380449408', '10.32683791458[10.208126360205757, 10.280005380449408, 10.326837914581214,passed
/violation_keysset['primary:rule_1:40', 'primary:rule_1:41', 'primary:rule_1:4['primary:rule_1:40', 'primary:rule_1:41', 'primary:rule_1:4passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

249  ew2 <- ewma(tw, 10, s0, 0.1, 2.7)
250  emit("STAT-QUAL-TW-004", tw_entry, "EWMA with lambda 0.1, L 2.7 and a supplied target", "QUAL_TW",
251       list(y_column = "y", method = "ewma", target = 10, ewma_lambda = 0.1, ewma_l = 2.7),
252       list("/primary/points/49/value" = tol(ew2$z[50], "z_50."),
253            "/primary/point_upper_limits" = seqtol(ew2$ucl, "Finite-time limits."),
254            "/violation_keys" = set_of(ew2$keys, "Points beyond their limits.")),
255       c("Sigma still comes from the data (MR-bar / d2)."), kind = "axis_coverage")

Whole script generators/r/quality_spc_attributes.R


A non-positive sigma is refused

Case STAT-QUAL-TW-005 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.2.3-4, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc323.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc324.htm; Montgomery, Introduction to Statistical Quality Control, 6th ed., ch. 9
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_TW.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

256  emit("STAT-QUAL-TW-005", tw_entry, "A non-positive sigma is refused", "QUAL_TW",
257       list(y_column = "y", sigma = 0), list(), c("Line 176: a supplied sigma must be positive."), kind = "invalid_input",
258       error = list(error_class = "ValueError", message_contains = "igma must be positive"))

Whole script generators/r/quality_spc_attributes.R


An EWMA lambda above 1 is refused

Case STAT-QUAL-TW-006 · status: passed

Reference

Evidence type
R
Source
Hand-coded chart formulas in R
Citation
NIST/SEMATECH e-Handbook 6.3.2.3-4, https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc323.htm and https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc324.htm; Montgomery, Introduction to Statistical Quality Control, 6th ed., ch. 9
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_spc_attributes.R · sha256 6e47cf20f469f98b
Dataset
reference_data/datasets/QUAL_TW.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

259  emit("STAT-QUAL-TW-006", tw_entry, "An EWMA lambda above 1 is refused", "QUAL_TW",
260       list(y_column = "y", method = "ewma", ewma_lambda = 1.5), list(), c("Line 195: 0 < lambda <= 1."), kind = "invalid_input",
261       error = list(error_class = "ValueError", message_contains = "ambda"))

Whole script generators/r/quality_spc_attributes.R


known sigma one-sided variables plan

Case STAT-QUAL-VARACC-001 · status: passed

Reference

Evidence type
R
Source
R stats analytic normal/noncentral-t formulas
Citation
R stats::pnorm, stats::qnorm, stats::pt and stats::uniroot documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_variables_acceptance.R · sha256 a00adcb842fc579c
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sample_sizeexact15150.00E+2passed
/acceptance_constanttolerance1.62071.62072.95E-1716passed
/probability_accept_aqltolerance0.953247452202224510.95324745220222472.22E-1615passed
/probability_accept_ltpdtolerance0.094504526370646130.094504526370645941.94E-1614passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

71  for (status in c("known", "unknown")) {
72    config <- list(spec_side = if (status == "known") "upper" else "lower",
73                   spec_limit = if (status == "known") 87.5 else 12.5,
74                   sigma_status = status, aql = 0.02, ltpd = 0.10,
75                   producer_risk = 0.05, consumer_risk = 0.10,
76                   max_sample_size = 500L, oc_points = 11L)
77    if (status == "known") config$sigma <- 2.25
78    result <- plan(config)
79    id <- if (status == "known") "STAT-QUAL-VARACC-001" else "STAT-QUAL-VARACC-002"
80    outputs <- list("/sample_size" = exact(result$n,
81      "First n with a feasible k interval; one fewer n fails a risk constraint."))
82    deferred <- list()
83    if (status == "known") {
84      outputs[["/acceptance_constant"]] <- tol(result$k,
85        "Midpoint k quantized down to four decimals; owner 1e-9 tier.")
86      outputs[["/probability_accept_aql"]] <- tol(result$aql,
87        "Normal AQL acceptance probability; owner 1e-9 tier.")
88      outputs[["/probability_accept_ltpd"]] <- tol(result$ltpd,
89        "Normal LTPD acceptance probability; owner 1e-9 tier.")
90    } else {
91      deferred <- lapply(c("/acceptance_constant", "/probability_accept_aql",
92                           "/probability_accept_ltpd"), function(path)
93        list(path = path, reason = "reference precision below tier"))
94    }
95    emit(id, paste(status, "sigma one-sided variables plan"), config,
96         outputs,
97         c("Manual chapter 64: Pa(p)=Phi(sqrt(n)*(z_p-k)) for known sigma, with z_p=qnorm(1-p).",
98           "For unknown sigma Pa(p)=P(T>=sqrt(n)*k), T noncentral-t(df=n-1,ncp=sqrt(n)*z_p).",
99           "R scans n from 2 to 500, solves the two k bounds, takes their midpoint and quantizes down to four decimals.",

Whole script generators/r/quality_variables_acceptance.R


unknown sigma one-sided variables plan

Case STAT-QUAL-VARACC-002 · status: passed

Reference

Evidence type
R
Source
R stats analytic normal/noncentral-t formulas
Citation
R stats::pnorm, stats::qnorm, stats::pt and stats::uniroot documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/quality_variables_acceptance.R · sha256 a00adcb842fc579c
Dataset
None

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sample_sizeexact35350.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

71  for (status in c("known", "unknown")) {
72    config <- list(spec_side = if (status == "known") "upper" else "lower",
73                   spec_limit = if (status == "known") 87.5 else 12.5,
74                   sigma_status = status, aql = 0.02, ltpd = 0.10,
75                   producer_risk = 0.05, consumer_risk = 0.10,
76                   max_sample_size = 500L, oc_points = 11L)
77    if (status == "known") config$sigma <- 2.25
78    result <- plan(config)
79    id <- if (status == "known") "STAT-QUAL-VARACC-001" else "STAT-QUAL-VARACC-002"
80    outputs <- list("/sample_size" = exact(result$n,
81      "First n with a feasible k interval; one fewer n fails a risk constraint."))
82    deferred <- list()
83    if (status == "known") {
84      outputs[["/acceptance_constant"]] <- tol(result$k,
85        "Midpoint k quantized down to four decimals; owner 1e-9 tier.")
86      outputs[["/probability_accept_aql"]] <- tol(result$aql,
87        "Normal AQL acceptance probability; owner 1e-9 tier.")
88      outputs[["/probability_accept_ltpd"]] <- tol(result$ltpd,
89        "Normal LTPD acceptance probability; owner 1e-9 tier.")
90    } else {
91      deferred <- lapply(c("/acceptance_constant", "/probability_accept_aql",
92                           "/probability_accept_ltpd"), function(path)
93        list(path = path, reason = "reference precision below tier"))
94    }
95    emit(id, paste(status, "sigma one-sided variables plan"), config,
96         outputs,
97         c("Manual chapter 64: Pa(p)=Phi(sqrt(n)*(z_p-k)) for known sigma, with z_p=qnorm(1-p).",
98           "For unknown sigma Pa(p)=P(T>=sqrt(n)*k), T noncentral-t(df=n-1,ncp=sqrt(n)*z_p).",
99           "R scans n from 2 to 500, solves the two k bounds, takes their midpoint and quantizes down to four decimals.",

Whole script generators/r/quality_variables_acceptance.R


Zone chart of subgroup means with R-bar within sigma

Case STAT-QUAL-ZONE-001 · status: passed

Reference

Evidence type
R
Source
Hand-coded zone scoring rule (Jaehn 1987) with stats::integrate chart constants
Citation
Jaehn, A. H. (1987) Zone control charts: SPC made easy, Quality 26(10), 51-53; Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 constants
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_ZONE_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stages/0/centertolerance50.95725000000000250.957256.73E-1717passed
/stages/0/sigmatolerance2.07164466522727692.07164466522727952.62E-1514passed
/points/0/scaletolerance1.03582233261363841.03582233261363981.36E-1514passed
/stages/0/signalsexact660.00E+2passed
/signal_countexact660.00E+2passed
/customized_signalexactFalseFalsepassed
/source_countexact40400.00E+2passed
/dropped_subgroup_countexact000.00E+2passed
/points/0/ztolerance-2.4398489204448017-2.43984892044480574.02E-1514passed
/points/0/zoneexact330.00E+2passed
/points/0/sideexact-1-10.00E+2passed
/points/0/scoretolerance44.00.00E+2passed
/points/0/signalexactFalseFalsepassed
/points/1/ztolerance-2.2419385322001895-2.24193853220018643.15E-1514passed
/points/1/zoneexact330.00E+2passed
/points/1/sideexact-1-10.00E+2passed
/points/1/scoretolerance88.00.00E+2passed
/points/1/signalexactTrueTruepassed
/points/2/ztolerance-0.834843942607436-0.83484394260743491.11E-1514passed
/points/2/zoneexact110.00E+2passed
/points/2/sideexact-1-10.00E+2passed
/points/2/scoretolerance88.00.00E+2passed
/points/2/signalexactTrueTruepassed
/points/3/ztolerance-1.3223792892587882-1.32237928925878651.75E-1514passed
/points/3/zoneexact220.00E+2passed
/points/3/sideexact-1-10.00E+2passed
/points/3/scoretolerance1010.00.00E+2passed
/points/3/signalexactTrueTruepassed
/points/4/ztolerance-1.6602750740666559-1.66027507406664689.14E-1514passed
/points/4/zoneexact220.00E+2passed
/points/4/sideexact-1-10.00E+2passed
/points/4/scoretolerance1212.00.00E+2passed
/points/4/signalexactTrueTruepassed
/points/5/ztolerance0.642243345266951820.64224334526695098.93E-1614passed
/points/5/zoneexact110.00E+2passed
/points/5/sideexact110.00E+2passed
/points/5/scoretolerance00.00.00E+2passed
/points/5/signalexactFalseFalsepassed
/points/6/ztolerance0.352618386860205190.352618386860204754.43E-1614passed
/points/6/zoneexact110.00E+2passed
/points/6/sideexact110.00E+2passed
/points/6/scoretolerance00.00.00E+2passed
/points/6/signalexactFalseFalsepassed
/points/7/ztolerance2.64065555827346812.6406555582734653.11E-1514passed
/points/7/zoneexact330.00E+2passed
/points/7/sideexact110.00E+2passed
/points/7/scoretolerance44.00.00E+2passed
/points/7/signalexactFalseFalsepassed
/points/8/ztolerance3.38402628485077723.38402628485076561.16E-1414passed
/points/8/zoneexact440.00E+2passed
/points/8/sideexact110.00E+2passed
/points/8/scoretolerance1212.00.00E+2passed
/points/8/signalexactTrueTruepassed
/points/9/ztolerance1.47974218332644861.47974218332644662.02E-1514passed
/points/9/zoneexact220.00E+2passed
/points/9/sideexact110.00E+2passed
/points/9/scoretolerance1414.00.00E+2passed
/points/9/signalexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

249  set.seed(20261011)
250  zone_means <- c(50.0, 49.5, 50.4, 49.8, 50.2, 51.6, 52.0, 51.8, 52.4, 52.1)
251  zone_value <- round(rep(zone_means, each = 4) + rnorm(40, 0, 2), 2)
252  zone_gap <- zone_value; zone_gap[15] <- NA
253  zd <- data.frame(value = zone_value, sample = rep(1:10, each = 4), stage = rep(c("P", "Q"), each = 20),
254                   value_gap = zone_gap, stringsAsFactors = FALSE)
255  export_dataset(zd, "QUAL_ZONE_01",
256                 "Simulated (R 4.6.1, set.seed 20261011): ten samples of four fill weights whose mean shifts upward in the second half; value_gap is value with row 15 missing; 40 rows",
257                 "Simulated with stats::rnorm", "Constructed validation data; no third-party license.",
258                 c(value = "Float64", sample = "Int64", stage = "String", value_gap = "Float64"))
259  z <- read_dataset("QUAL_ZONE_01")
260  zone_ep <- "tabelz.analysis.spc_zone:compute_zone_chart"
261  zone_menu <- "zoneChartMenuItem"
262  zone_source <- "Hand-coded zone scoring rule (Jaehn 1987) with stats::integrate chart constants"
263  zone_cite <- "Jaehn, A. H. (1987) Zone control charts: SPC made easy, Quality 26(10), 51-53; Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 constants"
264  zone_req <- "REQ-STAT-QUAL-ZONE-001"

Whole script generators/r/quality_subgroup_capability.R


Zone chart of subgroup means with S-bar within sigma

Case STAT-QUAL-ZONE-002 · status: passed

Reference

Evidence type
R
Source
Hand-coded zone scoring rule (Jaehn 1987) with stats::integrate chart constants
Citation
Jaehn, A. H. (1987) Zone control charts: SPC made easy, Quality 26(10), 51-53; Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 constants
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_ZONE_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stages/0/sigmatolerance2.08320995466561822.08320995466561864.46E-1615passed
/signal_countexact660.00E+2passed
/points/0/ztolerance-2.426303689976038-2.4263036899760446.18E-1514passed
/points/0/zoneexact330.00E+2passed
/points/0/sideexact-1-10.00E+2passed
/points/0/scoretolerance44.00.00E+2passed
/points/0/signalexactFalseFalsepassed
/points/1/ztolerance-2.2294920344433047-2.22949203444330424.62E-1615passed
/points/1/zoneexact330.00E+2passed
/points/1/sideexact-1-10.00E+2passed
/points/1/scoretolerance88.00.00E+2passed
/points/1/signalexactTrueTruepassed
/points/2/ztolerance-0.8302091664484238-0.83020916644842362.26E-1615passed
/points/2/zoneexact110.00E+2passed
/points/2/sideexact-1-10.00E+2passed
/points/2/scoretolerance88.00.00E+2passed
/points/2/signalexactTrueTruepassed
/points/3/ztolerance-1.3150378788583177-1.31503787885831752.34E-1615passed
/points/3/zoneexact220.00E+2passed
/points/3/sideexact-1-10.00E+2passed
/points/3/scoretolerance1010.00.00E+2passed
/points/3/signalexactTrueTruepassed
/points/4/ztolerance-1.6510577785483429-1.65105777854833567.32E-1514passed
/points/4/zoneexact220.00E+2passed
/points/4/sideexact-1-10.00E+2passed
/points/4/scoretolerance1212.00.00E+2passed
/points/4/signalexactTrueTruepassed
/points/5/ztolerance0.638677823625109830.63867782362510962.21E-1615passed
/points/5/zoneexact110.00E+2passed
/points/5/sideexact110.00E+2passed
/points/5/scoretolerance00.00.00E+2passed
/points/5/signalexactFalseFalsepassed
/points/6/ztolerance0.35066076674794250.35066076674794241.13E-1615passed
/points/6/zoneexact110.00E+2passed
/points/6/sideexact110.00E+2passed
/points/6/scoretolerance00.00.00E+2passed
/points/6/signalexactFalseFalsepassed
/points/7/ztolerance2.62599551607752972.6259955160775299.09E-1615passed
/points/7/zoneexact330.00E+2passed
/points/7/sideexact110.00E+2passed
/points/7/scoretolerance44.00.00E+2passed
/points/7/signalexactFalseFalsepassed
/points/8/ztolerance3.36523929539558483.3652392953955777.60E-1514passed
/points/8/zoneexact440.00E+2passed
/points/8/sideexact110.00E+2passed
/points/8/scoretolerance1212.00.00E+2passed
/points/8/signalexactTrueTruepassed
/points/9/ztolerance1.47152714642823981.47152714642823962.37E-1615passed
/points/9/zoneexact220.00E+2passed
/points/9/sideexact110.00E+2passed
/points/9/scoretolerance1414.00.00E+2passed
/points/9/signalexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

292  zz <- (zmeans - zcenter) / (sigma_s / sqrt(zsizes))
293  steps <- zone_scores(zz)
294  out <- c(list("/stages/0/sigma" = num(sigma_s, "Mean of s_i/c4(4)."), "/signal_count" = exact(as.integer(sum(steps$signal)), "Signalling points.")),
295           zone_outputs(steps, zz))
296  emit("STAT-QUAL-ZONE-002", zone_ep, zone_menu, "Zone chart of subgroup means with S-bar within sigma", "QUAL_ZONE_01",
297       list(response = "value", statistic = "subgroup_mean", subgroup_mode = "column", subgroup_column = "sample", within_sigma = "sbar"),
298       out, c(zone_notes, "within_sigma = sbar: the scale is S-bar/c4(n) divided by sqrt(n_i)."), zone_source, zone_cite, zone_req, kind = "axis_coverage")

Whole script generators/r/quality_subgroup_capability.R


Zone chart as an individuals chart of subgroup means

Case STAT-QUAL-ZONE-003 · status: passed

Reference

Evidence type
R
Source
Hand-coded zone scoring rule (Jaehn 1987) with stats::integrate chart constants
Citation
Jaehn, A. H. (1987) Zone control charts: SPC made easy, Quality 26(10), 51-53; Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 constants
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_ZONE_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stages/0/sigmatolerance1.01571452622724561.01571452622724332.26E-1514passed
/points/0/scaletolerance1.01571452622724561.01571452622724332.26E-1514passed
/signal_countexact660.00E+2passed
/points/0/ztolerance-2.4881499030905672-2.48814990309057961.24E-1414passed
/points/0/zoneexact330.00E+2passed
/points/0/sideexact-1-10.00E+2passed
/points/0/scoretolerance44.00.00E+2passed
/points/0/signalexactFalseFalsepassed
/points/1/ztolerance-2.2863215401927217-2.28632154019272664.90E-1514passed
/points/1/zoneexact330.00E+2passed
/points/1/sideexact-1-10.00E+2passed
/points/1/scoretolerance88.00.00E+2passed
/points/1/signalexactTrueTruepassed
/points/2/ztolerance-0.85137110641905944-0.85137110641906131.89E-1514passed
/points/2/zoneexact110.00E+2passed
/points/2/sideexact-1-10.00E+2passed
/points/2/scoretolerance88.00.00E+2passed
/points/2/signalexactTrueTruepassed
/points/3/ztolerance-1.3485580491673992-1.3485580491674022.87E-1514passed
/points/3/zoneexact220.00E+2passed
/points/3/sideexact-1-10.00E+2passed
/points/3/scoretolerance1010.00.00E+2passed
/points/3/signalexactTrueTruepassed
/points/4/ztolerance-1.6931430589929808-1.69314305899297753.34E-1514passed
/points/4/zoneexact220.00E+2passed
/points/4/sideexact-1-10.00E+2passed
/points/4/scoretolerance1212.00.00E+2passed
/points/4/signalexactTrueTruepassed
/points/5/ztolerance0.654957650818478230.65495765081847971.45E-1514passed
/points/5/zoneexact110.00E+2passed
/points/5/sideexact110.00E+2passed
/points/5/scoretolerance00.00.00E+2passed
/points/5/signalexactFalseFalsepassed
/points/6/ztolerance0.359599070967976660.359599070967977447.81E-1614passed
/points/6/zoneexact110.00E+2passed
/points/6/sideexact110.00E+2passed
/points/6/scoretolerance00.00.00E+2passed
/points/6/signalexactFalseFalsepassed
/points/7/ztolerance2.69293185178690342.6929318517869095.79E-1514passed
/points/7/zoneexact330.00E+2passed
/points/7/sideexact110.00E+2passed
/points/7/scoretolerance44.00.00E+2passed
/points/7/signalexactFalseFalsepassed
/points/8/ztolerance3.4510188734031833.45101887340318354.73E-1615passed
/points/8/zoneexact440.00E+2passed
/points/8/sideexact110.00E+2passed
/points/8/scoretolerance1212.00.00E+2passed
/points/8/signalexactTrueTruepassed
/points/9/ztolerance1.50903621088616611.50903621088616953.36E-1514passed
/points/9/zoneexact220.00E+2passed
/points/9/sideexact110.00E+2passed
/points/9/scoretolerance1414.00.00E+2passed
/points/9/signalexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

301  sigma_i <- mean(abs(diff(zmeans))) / d2_2
302  zz <- (zmeans - zcenter) / sigma_i
303  steps <- zone_scores(zz)
304  out <- c(list("/stages/0/sigma" = num(sigma_i, "MR-bar of the subgroup means / d2(2)."),
305                "/points/0/scale" = num(sigma_i, "The MR scale is not divided by sqrt(n)."),
306                "/signal_count" = exact(as.integer(sum(steps$signal)), "Signalling points.")),
307           zone_outputs(steps, zz))
308  emit("STAT-QUAL-ZONE-003", zone_ep, zone_menu, "Zone chart as an individuals chart of subgroup means", "QUAL_ZONE_01",
309       list(response = "value", statistic = "individuals_of_means", subgroup_mode = "size", subgroup_size = 4L),
310       out, c(zone_notes, "statistic = individuals_of_means: the scale is MR-bar/d2(2) of the consecutive subgroup means (chapter 84 line 106); the centre stays the size-weighted mean, which equals the mean of the means for equal sizes."),
311       zone_source, zone_cite, zone_req, kind = "axis_coverage")

Whole script generators/r/quality_subgroup_capability.R


Zone chart with custom weights, reset after signal and two stages

Case STAT-QUAL-ZONE-004 · status: passed

Reference

Evidence type
R
Source
Hand-coded zone scoring rule (Jaehn 1987) with stats::integrate chart constants
Citation
Jaehn, A. H. (1987) Zone control charts: SPC made easy, Quality 26(10), 51-53; Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 constants
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_ZONE_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/customized_signalexactFalseFalsepassed
/weights/1tolerance11.00.00E+2passed
/reset_after_signalexactTrueTruepassed
/stages/0/centertolerance49.196549.19653.41E-1617passed
/stages/0/sigmatolerance2.1226464682399062.1226464682399082.21E-1514passed
/stages/0/signalsexact000.00E+2passed
/points/0/ztolerance-0.72221164613962385-0.72221164613962975.89E-1514passed
/points/0/zoneexact110.00E+2passed
/points/0/sideexact-1-10.00E+2passed
/points/0/scoretolerance00.00.00E+2passed
/points/0/signalexactFalseFalsepassed
/points/1/ztolerance-0.52905654182308548-0.52905654182308495.55E-1614passed
/points/1/zoneexact110.00E+2passed
/points/1/sideexact-1-10.00E+2passed
/points/1/scoretolerance00.00.00E+2passed
/points/1/signalexactFalseFalsepassed
/points/2/ztolerance0.84422913886640960.84422913886640878.87E-1614passed
/points/2/zoneexact110.00E+2passed
/points/2/sideexact110.00E+2passed
/points/2/scoretolerance00.00.00E+2passed
/points/2/signalexactFalseFalsepassed
/points/3/ztolerance0.368408028232995950.368408028232995573.84E-1614passed
/points/3/zoneexact110.00E+2passed
/points/3/sideexact110.00E+2passed
/points/3/scoretolerance00.00.00E+2passed
/points/3/signalexactFalseFalsepassed
/points/4/ztolerance0.038631020863303660.0386310208633103146.65E-1512passed
/points/4/zoneexact110.00E+2passed
/points/4/sideexact110.00E+2passed
/points/4/scoretolerance00.00.00E+2passed
/points/4/signalexactFalseFalsepassed
/stages/1/centertolerance52.71799999999999652.7179999999999964.19E-1617passed
/stages/1/sigmatolerance2.02064286221464772.020642862214652.25E-1514passed
/stages/1/signalsexact000.00E+2passed
/points/5/ztolerance-1.0843083856979194-1.08430838569791831.08E-1515passed
/points/5/zoneexact220.00E+2passed
/points/5/sideexact-1-10.00E+2passed
/points/5/scoretolerance11.00.00E+2passed
/points/5/signalexactFalseFalsepassed
/points/6/ztolerance-1.3812435894490671-1.38124358944906561.54E-1514passed
/points/6/zoneexact220.00E+2passed
/points/6/sideexact-1-10.00E+2passed
/points/6/scoretolerance22.00.00E+2passed
/points/6/signalexactFalseFalsepassed
/points/7/ztolerance0.964544520184963170.96454452018496229.97E-1614passed
/points/7/zoneexact110.00E+2passed
/points/7/sideexact110.00E+2passed
/points/7/scoretolerance00.00.00E+2passed
/points/7/signalexactFalseFalsepassed
/points/8/ztolerance1.7266782098129011.7266782098128928.89E-1514passed
/points/8/zoneexact220.00E+2passed
/points/8/sideexact110.00E+2passed
/points/8/scoretolerance11.00.00E+2passed
/points/8/signalexactFalseFalsepassed
/points/9/ztolerance-0.22567075485086346-0.225670754850863212.46E-1614passed
/points/9/zoneexact110.00E+2passed
/points/9/sideexact-1-10.00E+2passed
/points/9/scoretolerance00.00.00E+2passed
/points/9/signalexactFalseFalsepassed
/signal_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

328  out[["/signal_count"]] <- exact(as.integer(total_signals), "Signals over both stages.")
329  emit("STAT-QUAL-ZONE-004", zone_ep, zone_menu, "Zone chart with custom weights, reset after signal and two stages", "QUAL_ZONE_01",
330       list(response = "value", statistic = "subgroup_mean", subgroup_mode = "column", subgroup_column = "sample", stage_column = "stage",
331            within_sigma = "rbar", weights = list(0, 1, 3, 6), signal_score = 6, reset_after_signal = TRUE),
332       out, c(zone_notes, "Weights 0, 1, 3, 6 with signal 6: a signalling point keeps its score and the next point starts from zero (chapter 84 lines 118-120); each stage (P: samples 1-5, Q: samples 6-10) estimates its own centre and sigma and starts a new accumulation."),
333       zone_source, zone_cite, zone_req, kind = "axis_coverage")

Whole script generators/r/quality_subgroup_capability.R


Raw-individual zone chart with a missing observation

Case STAT-QUAL-ZONE-005 · status: passed

Reference

Evidence type
R
Source
Hand-coded zone scoring rule (Jaehn 1987) with stats::integrate chart constants
Citation
Jaehn, A. H. (1987) Zone control charts: SPC made easy, Quality 26(10), 51-53; Montgomery, Introduction to Statistical Quality Control, 7th ed., ch. 6 constants
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, nlme 3.1.169
Reference script
generators/r/quality_subgroup_capability.R · sha256 301362213a445f16
Dataset
reference_data/datasets/QUAL_ZONE_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stages/0/centertolerance50.93641025641025550.9364102564102552.15E-1617passed
/stages/0/sigmatolerance2.25029782827801622.2502978282780163.44E-1716passed
/stages/0/mean_moving_rangetolerance2.53918918918918872.53918918918918872.21E-1717passed
/stages/0/moving_rangesexact37370.00E+2passed
/stages/0/subgroupsexact39390.00E+2passed
/gap_countexact110.00E+2passed
/missing_countexact110.00E+2passed
/unitsexactggpassed
/signal_countexact880.00E+2passed
/points/14/eventexactafter_gapafter_gappassed
/points/0/ztolerance-0.35835712334457426-0.358357123344574262.35E-1817passed
/points/0/zoneexact110.00E+2passed
/points/0/sideexact-1-10.00E+2passed
/points/0/scoretolerance00.00.00E+2passed
/points/0/signalexactFalseFalsepassed
/points/1/ztolerance-0.59832535919947916-0.59832535919947923.52E-1817passed
/points/1/zoneexact110.00E+2passed
/points/1/sideexact-1-10.00E+2passed
/points/1/scoretolerance00.00.00E+2passed
/points/1/signalexactFalseFalsepassed
/points/2/ztolerance-1.3671124851790846-1.36711248517908463.11E-1716passed
/points/2/zoneexact220.00E+2passed
/points/2/sideexact-1-10.00E+2passed
/points/2/scoretolerance22.00.00E+2passed
/points/2/signalexactFalseFalsepassed
/points/3/ztolerance-2.1314557549391528-2.1314557549391534.68E-1716passed
/points/3/zoneexact330.00E+2passed
/points/3/sideexact-1-10.00E+2passed
/points/3/scoretolerance66.00.00E+2passed
/points/3/signalexactFalseFalsepassed
/points/4/ztolerance-1.3760001976181535-1.37600019761815358.83E-1817passed
/points/4/zoneexact220.00E+2passed
/points/4/sideexact-1-10.00E+2passed
/points/4/scoretolerance88.00.00E+2passed
/points/4/signalexactTrueTruepassed
/points/5/ztolerance0.388210721537352980.3882107215373533.73E-1817passed
/points/5/zoneexact110.00E+2passed
/points/5/sideexact110.00E+2passed
/points/5/scoretolerance00.00.00E+2passed
/points/5/signalexactFalseFalsepassed
/points/6/ztolerance-1.1404758179827861-1.1404758179827862.24E-1716passed
/points/6/zoneexact220.00E+2passed
/points/6/sideexact-1-10.00E+2passed
/points/6/scoretolerance22.00.00E+2passed
/points/6/signalexactFalseFalsepassed
/points/7/ztolerance-1.9625892185968108-1.96258921859681088.31E-1817passed
/points/7/zoneexact220.00E+2passed
/points/7/sideexact-1-10.00E+2passed
/points/7/scoretolerance44.00.00E+2passed
/points/7/signalexactFalseFalsepassed
/points/8/ztolerance-0.91383915078648448-0.91383915078648452.39E-1817passed
/points/8/zoneexact110.00E+2passed
/points/8/sideexact-1-10.00E+2passed
/points/8/scoretolerance44.00.00E+2passed
/points/8/signalexactFalseFalsepassed
/points/9/ztolerance-0.81607431395670882-0.81607431395670881.09E-1817passed
/points/9/zoneexact110.00E+2passed
/points/9/sideexact-1-10.00E+2passed
/points/9/scoretolerance44.00.00E+2passed
/points/9/signalexactFalseFalsepassed
/points/10/ztolerance-0.26947999895387054-0.269479998953870543.52E-1816passed
/points/10/zoneexact110.00E+2passed
/points/10/sideexact-1-10.00E+2passed
/points/10/scoretolerance44.00.00E+2passed
/points/10/signalexactFalseFalsepassed
/points/11/ztolerance0.499307127025735040.499307127025735042.47E-1817passed
/points/11/zoneexact110.00E+2passed
/points/11/sideexact110.00E+2passed
/points/11/scoretolerance00.00.00E+2passed
/points/11/signalexactFalseFalsepassed
/points/12/ztolerance-1.0471548373725448-1.04715483737254482.22E-1716passed
/points/12/zoneexact220.00E+2passed
/points/12/sideexact-1-10.00E+2passed
/points/12/scoretolerance22.00.00E+2passed
/points/12/signalexactFalseFalsepassed
/points/13/ztolerance0.0949162110480227110.094916211048022711.39E-1917passed
/points/13/zoneexact110.00E+2passed
/points/13/sideexact110.00E+2passed
/points/13/scoretolerance00.00.00E+2passed
/points/13/signalexactFalseFalsepassed
/points/14/ztolerance-1.8159419633521472-1.81594196335214727.38E-1817passed
/points/14/zoneexact220.00E+2passed
/points/14/sideexact-1-10.00E+2passed
/points/14/scoretolerance22.00.00E+2passed
/points/14/signalexactFalseFalsepassed
/points/15/ztolerance-0.14505202480688212-0.145052024806882121.70E-1816passed
/points/15/zoneexact110.00E+2passed
/points/15/sideexact-1-10.00E+2passed
/points/15/scoretolerance22.00.00E+2passed
/points/15/signalexactFalseFalsepassed
/points/16/ztolerance-0.68720248359018599-0.6872024835901863.73E-1817passed
/points/16/zoneexact110.00E+2passed
/points/16/sideexact-1-10.00E+2passed
/points/16/scoretolerance22.00.00E+2passed
/points/16/signalexactFalseFalsepassed
/points/17/ztolerance-2.087017192743799-2.0870171927437994.33E-1716passed
/points/17/zoneexact330.00E+2passed
/points/17/sideexact-1-10.00E+2passed
/points/17/scoretolerance66.00.00E+2passed
/points/17/signalexactFalseFalsepassed
/points/18/ztolerance-0.10061346261152868-0.100613462611528684.35E-1816passed
/points/18/zoneexact110.00E+2passed
/points/18/sideexact-1-10.00E+2passed
/points/18/scoretolerance66.00.00E+2passed
/points/18/signalexactFalseFalsepassed
/points/19/ztolerance-0.46056581639388755-0.460565816393887554.51E-1918passed
/points/19/zoneexact110.00E+2passed
/points/19/sideexact-1-10.00E+2passed
/points/19/scoretolerance66.00.00E+2passed
/points/19/signalexactFalseFalsepassed
/points/20/ztolerance1.09922771666299561.09922771666299563.13E-1716passed
/points/20/zoneexact220.00E+2passed
/points/20/sideexact110.00E+2passed
/points/20/scoretolerance22.00.00E+2passed
/points/20/signalexactFalseFalsepassed
/points/21/ztolerance-0.35835712334457426-0.358357123344574262.35E-1817passed
/points/21/zoneexact110.00E+2passed
/points/21/sideexact-1-10.00E+2passed
/points/21/scoretolerance00.00.00E+2passed
/points/21/signalexactFalseFalsepassed
/points/22/ztolerance0.939248892759725670.93924889275972573.15E-1817passed
/points/22/zoneexact110.00E+2passed
/points/22/sideexact110.00E+2passed
/points/22/scoretolerance00.00.00E+2passed
/points/22/signalexactFalseFalsepassed
/points/23/ztolerance-0.60721307163855109-0.60721307163855114.01E-1817passed
/points/23/zoneexact110.00E+2passed
/points/23/sideexact-1-10.00E+2passed
/points/23/scoretolerance00.00.00E+2passed
/points/23/signalexactFalseFalsepassed
/points/24/ztolerance1.65915360032444341.65915360032444343.28E-1716passed
/points/24/zoneexact220.00E+2passed
/points/24/sideexact110.00E+2passed
/points/24/scoretolerance22.00.00E+2passed
/points/24/signalexactFalseFalsepassed
/points/25/ztolerance0.725943794222033480.72594379422203351.71E-1817passed
/points/25/zoneexact110.00E+2passed
/points/25/sideexact110.00E+2passed
/points/25/scoretolerance22.00.00E+2passed
/points/25/signalexactFalseFalsepassed
/points/26/ztolerance-1.0915933995678984-1.09159339956789844.08E-1716passed
/points/26/zoneexact220.00E+2passed
/points/26/sideexact-1-10.00E+2passed
/points/26/scoretolerance22.00.00E+2passed
/points/26/signalexactFalseFalsepassed
/points/27/ztolerance2.56570026910964042.56570026910964044.12E-1716passed
/points/27/zoneexact330.00E+2passed
/points/27/sideexact110.00E+2passed
/points/27/scoretolerance44.00.00E+2passed
/points/27/signalexactFalseFalsepassed
/points/28/ztolerance1.07256457934578611.0725645793457861.56E-1716passed
/points/28/zoneexact220.00E+2passed
/points/28/sideexact110.00E+2passed
/points/28/scoretolerance66.00.00E+2passed
/points/28/signalexactFalseFalsepassed
/points/29/ztolerance0.210456472755939150.210456472755939151.54E-1918passed
/points/29/zoneexact110.00E+2passed
/points/29/sideexact110.00E+2passed
/points/29/scoretolerance66.00.00E+2passed
/points/29/signalexactFalseFalsepassed
/points/30/ztolerance1.05034529824810761.05034529824810762.83E-1716passed
/points/30/zoneexact220.00E+2passed
/points/30/sideexact110.00E+2passed
/points/30/scoretolerance88.00.00E+2passed
/points/30/signalexactTrueTruepassed
/points/31/ztolerance1.57472033215327081.57472033215327081.46E-1717passed
/points/31/zoneexact220.00E+2passed
/points/31/sideexact110.00E+2passed
/points/31/scoretolerance1010.00.00E+2passed
/points/31/signalexactTrueTruepassed
/points/32/ztolerance1.556944907275131.556944907275131.89E-1716passed
/points/32/zoneexact220.00E+2passed
/points/32/sideexact110.00E+2passed
/points/32/scoretolerance1212.00.00E+2passed
/points/32/signalexactTrueTruepassed
/points/33/ztolerance1.18366098483416821.18366098483416824.94E-1716passed
/points/33/zoneexact220.00E+2passed
/points/33/sideexact110.00E+2passed
/points/33/scoretolerance1414.00.00E+2passed
/points/33/signalexactTrueTruepassed
/points/34/ztolerance1.95244811081377031.95244811081377034.33E-1716passed
/points/34/zoneexact220.00E+2passed
/points/34/sideexact110.00E+2passed
/points/34/scoretolerance1616.00.00E+2passed
/points/34/signalexactTrueTruepassed
/points/35/ztolerance0.725943794222033480.72594379422203351.71E-1817passed
/points/35/zoneexact110.00E+2passed
/points/35/sideexact110.00E+2passed
/points/35/scoretolerance1616.00.00E+2passed
/points/35/signalexactTrueTruepassed
/points/36/ztolerance1.69026059386119051.69026059386119054.12E-1716passed
/points/36/zoneexact220.00E+2passed
/points/36/sideexact110.00E+2passed
/points/36/scoretolerance1818.00.00E+2passed
/points/36/signalexactTrueTruepassed
/points/37/ztolerance-0.11838888748966943-0.118388887489669431.79E-1816passed
/points/37/zoneexact110.00E+2passed
/points/37/sideexact-1-10.00E+2passed
/points/37/scoretolerance00.00.00E+2passed
/points/37/signalexactFalseFalsepassed
/points/38/ztolerance0.463756277269450350.463756277269450354.54E-1817passed
/points/38/zoneexact110.00E+2passed
/points/38/sideexact110.00E+2passed
/points/38/scoretolerance00.00.00E+2passed
/points/38/signalexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

336  x <- z$value_gap
337  present <- which(!is.na(x)); xv <- x[present]
338  centre <- mean(xv)
339  adjacent <- diff(present) == 1
340  mr <- abs(diff(xv))[adjacent]
341  sigma_raw <- mean(mr) / d2_2
342  zz <- (xv - centre) / sigma_raw
343  runs <- cumsum(c(TRUE, !adjacent))
344  steps <- do.call(rbind, lapply(split(seq_along(zz), runs), function(idx) zone_scores(zz[idx])))
345  out <- c(list("/stages/0/center" = num(centre, "Mean of the 39 finite observations."),
346                "/stages/0/sigma" = num(sigma_raw, "MR-bar of adjacent observations / d2(2); no sqrt(n)."),
347                "/stages/0/mean_moving_range" = num(mean(mr), "Mean of the 37 adjacent moving ranges; none spans the missing row."),
348                "/stages/0/moving_ranges" = exact(length(mr), "Adjacent pairs."),
349                "/stages/0/subgroups" = exact(length(xv), "Observations in the stage."),
350                "/gap_count" = exact(1L, "The missing row breaks adjacency once."),
351                "/missing_count" = exact(1L, "One missing observation."),
352                "/units" = exact("g", "Units recorded with the raw chart."),
353                "/signal_count" = exact(as.integer(sum(steps$signal)), "Signals over both runs."),
354                "/points/14/event" = exact("after_gap", "The observation after the missing row starts a new run.")),
355           zone_outputs(steps, zz))
356  emit("STAT-QUAL-ZONE-005", zone_ep, zone_menu, "Raw-individual zone chart with a missing observation", "QUAL_ZONE_01",
357       list(response = "value_gap", statistic = "raw_individuals", subgroup_mode = "individual", units = "g"),
358       out, c("docs/manual/chapters/103-raw-zone.tex:9-20: centre is the mean of the finite observations; sigma = MR-bar/(2/sqrt(pi)) over consecutive eligible pairs; z = (x - xbar)/sigma with no sqrt(n); missing rows break adjacency and restart score accumulation.",
359              "docs/manual/chapters/103-raw-zone.tex:22-24: same zones, weights 0, 2, 4, 8 and signal 8; the score restarts after the gap, so the two runs are scored separately.",
360              zone_notes[2]),
361       zone_source, zone_cite, zone_req, kind = "axis_coverage")

Whole script generators/r/quality_subgroup_capability.R


Regression II remaining axis STAT-REG-AXIS-ADV-DA-MEAN-LDA-001

Case STAT-REG-AXIS-ADV-DA-MEAN-LDA-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_DISCRIMINANT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact24240.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_predictor_countexact000.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II remaining axis STAT-REG-AXIS-ADV-DA-MEAN-QDA-001

Case STAT-REG-AXIS-ADV-DA-MEAN-QDA-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_DISCRIMINANT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact24240.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_predictor_countexact000.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II option axis STAT-REG-AXIS-ADV-DA-PRIOR-CUSTOM-001

Case STAT-REG-AXIS-ADV-DA-PRIOR-CUSTOM-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact24240.00E+2passed
/artifact/priors/0tolerance0.699999999999999960.74.41E-1817passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-DA-PRIOR-OBSERVED-001

Case STAT-REG-AXIS-ADV-DA-PRIOR-OBSERVED-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact24240.00E+2passed
/artifact/priors/0tolerance0.50.50.00E+1passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II remaining axis STAT-REG-AXIS-ADV-DA-SCORE-MEAN-001

Case STAT-REG-AXIS-ADV-DA-SCORE-MEAN-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_DISCRIMINANT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/1/probabilities/0tolerance0.50.50.00E+1passed
/1/row_idexactSTAT-REG-AXIS-ADV-DA-SCORE-MEAN-001:1STAT-REG-AXIS-ADV-DA-SCORE-MEAN-001:1passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

123  lda_score$artifact$imputation <- "mean"
124  lda_score$artifact$imputation_values <- list(0)
125  filled_x <- da$x
126  filled_x[is.na(filled_x)] <- 0
127  lda_ref <- MASS::lda(class ~ x, data = read_dataset("REG_DISCRIMINANT"),
128                       prior = c(0.5, 0.5))
129  lda_pred <- stats::predict(lda_ref, newdata = data.frame(x = filled_x))
130  emit("STAT-REG-AXIS-ADV-DA-SCORE-MEAN-001", "STAT-REG-DISCRIMINANT-SCORE-LDA-001",
131       "REG_DISCRIMINANT_MISSING", lda_score,
132       list("/1/probabilities/0" = tol(lda_pred$posterior[2L, 1L],
133                                        "R LDA posterior after mean filling", "1e-9"),
134            "/1/row_id" = exact("STAT-REG-AXIS-ADV-DA-SCORE-MEAN-001:1",
135                                "missing-x source row keeps identity")),
136       "Stored full-population LDA artifact uses zero training mean; the 50/50 class tie has no unique label.")

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II option axis STAT-REG-AXIS-ADV-MIXED-CATEGORICAL-001

Case STAT-REG-AXIS-ADV-MIXED-CATEGORICAL-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.660185185185185540.66018518518518514.43E-1615passed
/analyzed_countexact1081080.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

65  mixed <- nlme::lme(distance ~ age + sex, random = ~1 | subject, data = orth,
66                     method = "REML", control = nlme::lmeControl(opt = "optim"))
67  emit("STAT-REG-AXIS-ADV-MIXED-CATEGORICAL-001", "STAT-REG-MIXED-REML-001",
68       "REG_ORTHODONT", list(response_column = "distance", group_column = "subject",
69                             fixed_numeric_columns = list("age"),
70                             fixed_categorical_columns = list("sex"), estimation = "reml"),
71       list("/coefficients/1/estimate" = tol(nlme::fixef(mixed)[["age"]], "adjusted age slope"),
72            "/analyzed_count" = exact(as.integer(nrow(orth)), "complete observations")),
73       list("Female/Male factor levels are alphabetic; nlme random intercept is REML."))

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-NONLINEAR-4PL-001

Case STAT-REG-AXIS-ADV-NONLINEAR-4PL-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_4PL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/estimates/2tolerance3.95625453747962033.95625453632357661.16E-99passed
/artifact/estimates/3tolerance2.02113893132951892.02113893265865171.33E-99passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

125  logistic_fit <- stats::nls(y ~ lo + (hi - lo) / (1 + (x / mid)^slope), data = logistic,
126                             start = list(lo = 1, hi = 5, mid = 4, slope = 2),
127                             algorithm = "port",
128                             lower = c(lo = 0, hi = 2, mid = 0.1, slope = 0.1),
129                             upper = c(lo = 3, hi = 8, mid = 10, slope = 5),
130                             control = stats::nls.control(maxiter = 1000, tol = 1e-12))
131  emit("STAT-REG-AXIS-ADV-NONLINEAR-4PL-001", "STAT-REG-NONLINEAR-DECAY-001",
132       "REG_4PL", list(x_column = "x", response = "y", curve = "four_parameter_logistic",
133                       parameters = list(list(name = "lo", start = 1, lower = 0, upper = 3),
134                                         list(name = "hi", start = 5, lower = 2, upper = 8),
135                                         list(name = "mid", start = 4, lower = 0.1, upper = 10),
136                                         list(name = "slope", start = 2, lower = 0.1, upper = 5))),
137       list("/artifact/estimates/2" = tol(coef(logistic_fit)[["mid"]], "four-parameter midpoint"),
138            "/artifact/estimates/3" = tol(coef(logistic_fit)[["slope"]], "four-parameter slope")),
139       list("R nls port curve matches manual 4PL direction and positive midpoint/slope."),
140       kind = "nominal")

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-NONLINEAR-GROWTH-001

Case STAT-REG-AXIS-ADV-NONLINEAR-GROWTH-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_GROWTH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/estimates/0tolerance3.98124972072586043.98124972122117934.95E-109passed
/artifact/estimates/1tolerance0.792853487036724650.79285348726870042.32E-109passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

110  growth_fit <- stats::nls(growth ~ c + a * (1 - exp(-k * x)), data = curve,
111                           start = list(a = 4, k = 0.8, c = 1), algorithm = "port",
112                           lower = c(a = 0, k = 0.01, c = 0),
113                           upper = c(a = 10, k = 3, c = 5),
114                           control = stats::nls.control(maxiter = 1000, tol = 1e-12))
115  emit("STAT-REG-AXIS-ADV-NONLINEAR-GROWTH-001", "STAT-REG-NONLINEAR-DECAY-001",
116       "REG_GROWTH",
117       list(x_column = "x", response = "growth", curve = "asymptotic_growth",
118            parameters = list(list(name = "a", start = 4, lower = 0, upper = 10),
119                              list(name = "k", start = 0.8, lower = 0.01, upper = 3),
120                              list(name = "c", start = 1, lower = 0, upper = 5))),
121       list("/artifact/estimates/0" = tol(coef(growth_fit)[["a"]], "growth amplitude"),
122            "/artifact/estimates/1" = tol(coef(growth_fit)[["k"]], "growth rate")),
123       list("NLS port bounded fit on original response scale; same parameter order as manual."),
124       kind = "nominal")

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-PANEL-90-001

Case STAT-REG-AXIS-ADV-PANEL-90-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/estimatetolerance0.66018518518518510.66018518518518512.81E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

102  panel <- plm::plm(distance ~ age, data = orth, index = c("subject", "time"),
103                    model = "within", effect = "individual")
104  emit("STAT-REG-AXIS-ADV-PANEL-90-001", "STAT-REG-PANEL-001", "REG_ORTHODONT",
105       list(response_column = "distance", predictor_columns = list("age"),
106            entity_column = "subject", time_column = "time", confidence_level = 0.9),
107       list("/coefficients/0/estimate" = tol(coef(panel)[["age"]], "within age slope", "1e-9")),
108       list("90% intervals change quantile only; plm within slope is invariant."))

Whole script generators/r/regression_advanced_axes.R


Regression II remaining axis STAT-REG-AXIS-ADV-PLS-MEAN-001

Case STAT-REG-AXIS-ADV-PLS-MEAN-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_PLS_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact30300.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_predictor_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 97  emit("STAT-REG-AXIS-ADV-PLS-MEAN-001", "STAT-REG-PLS-FIT-001",
 98       "REG_PLS_MISSING", list(response = "y", predictors = list("x"),
 99                               imputation = "mean", scale = FALSE, cv_folds = 3L),
100       list("/source_count" = exact(as.integer(nrow(pls_data)), "source rows"),
101            "/missing_response_count" = exact(1L, "one missing y"),
102            "/incomplete_predictor_count" = exact(0L, "mean-filled x retained")),
103       "PLS mean filling uses training-only mean; random train split prevents R coefficient oracle.",
104       kind = "determinism", repeat_count = 2L,
105       deferred = list(list(path = "/artifact/coefficients",
106         reason = "NumPy training split/CV folds are not reproducible with R RNG.")))

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II option axis STAT-REG-AXIS-ADV-PLS-SCALE-001

Case STAT-REG-AXIS-ADV-PLS-SCALE-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact30300.00E+2passed
/training_metrics/row_countexact23230.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

171  emit("STAT-REG-AXIS-ADV-PLS-SCALE-001", "STAT-REG-PLS-FIT-001", "REG_PLS",
172       list(response = "y", predictors = list("x"), scale = TRUE, cv_folds = 3L,
173            seed = 3L),
174       list("/source_count" = exact(as.integer(nrow(pls_data)), "complete PLS source"),
175            "/training_metrics/row_count" = exact(23L, "floor(0.25*30) held out")),
176       list("R pls::plsr scale=TRUE; split-dependent coefficients deferred."),
177       kind = "determinism", repeat_count = 2L,
178       deferred = list(list(path = "/artifact/coefficients",
179         reason = "NumPy-selected train rows and CV folds differ from R RNG.")))

Whole script generators/r/regression_advanced_axes.R


Regression II remaining axis STAT-REG-AXIS-ADV-PLS-SCORE-MEAN-001

Case STAT-REG-AXIS-ADV-PLS-SCORE-MEAN-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_PLS_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/1/predictedtolerance34.00046675059380234.00046675059383.65E-1616passed
/1/row_idexactSTAT-REG-AXIS-ADV-PLS-SCORE-MEAN-001:1STAT-REG-AXIS-ADV-PLS-SCORE-MEAN-001:1passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

139  pls_score$artifact$imputation <- "mean"
140  pls_score$artifact$imputation_values <- list(pls_score$artifact$x_means[[1L]])
141  expected_mean <- as.double(pls_score$artifact$intercept)
142  emit("STAT-REG-AXIS-ADV-PLS-SCORE-MEAN-001", "STAT-REG-PLS-SCORE-001",
143       "REG_PLS_MISSING", pls_score,
144       list("/1/predicted" = tol(expected_mean,
145                                  "centered PLS prediction at training mean", "1e-9"),
146            "/1/row_id" = exact("STAT-REG-AXIS-ADV-PLS-SCORE-MEAN-001:1",
147                                 "missing-x source row keeps identity")),
148       "Mean imputation inserts stored training x_mean; centered PLS prediction equals intercept.")

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II option axis STAT-REG-AXIS-ADV-SELECTION-AIC-BACKWARD-001

Case STAT-REG-AXIS-ADV-SELECTION-AIC-BACKWARD-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/selectedset['age_sq']['age_sq']passed
/nexact1081080.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-SELECTION-AIC-BOTH-001

Case STAT-REG-AXIS-ADV-SELECTION-AIC-BOTH-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/selectedset['age_sq']['age_sq']passed
/nexact1081080.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-SELECTION-AIC-FORWARD-001

Case STAT-REG-AXIS-ADV-SELECTION-AIC-FORWARD-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/selectedset['age_sq']['age_sq']passed
/nexact1081080.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-SELECTION-BEST-AIC-001

Case STAT-REG-AXIS-ADV-SELECTION-BEST-AIC-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/selectedset['age_sq']['age_sq']passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-SELECTION-BEST-BIC-001

Case STAT-REG-AXIS-ADV-SELECTION-BEST-BIC-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/selectedset['age_sq']['age_sq']passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-SELECTION-BIC-BACKWARD-001

Case STAT-REG-AXIS-ADV-SELECTION-BIC-BACKWARD-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/selectedset['age_sq']['age_sq']passed
/nexact1081080.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-SELECTION-BIC-BOTH-001

Case STAT-REG-AXIS-ADV-SELECTION-BIC-BOTH-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/selectedset['age_sq']['age_sq']passed
/nexact1081080.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-SELECTION-BIC-FORWARD-001

Case STAT-REG-AXIS-ADV-SELECTION-BIC-FORWARD-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/selectedset['age_sq']['age_sq']passed
/nexact1081080.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-STRUCT-CROSSED-001

Case STAT-REG-AXIS-ADV-STRUCT-CROSSED-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_CROSSED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.79896711100613460.7989671109813142.48E-1110passed
/analyzed_countexact64640.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

75  crossed_fit <- lme4::lmer(y ~ x + (1 | a) + (1 | b), data = crossed, REML = TRUE,
76                            control = lme4::lmerControl(optimizer = "bobyqa",
77                                                          optCtrl = list(maxfun = 100000L)))
78  crossed_config <- list(response_column = "y", fixed_numeric_columns = list("x"),
79                         model_mode = "variance_components",
80                         components = list(list(name = "a", level_column = "a", relation = "crossed"),
81                                           list(name = "b", level_column = "b", relation = "crossed")),
82                         estimation = "reml", max_rows = 1000L)
83  emit("STAT-REG-AXIS-ADV-STRUCT-CROSSED-001", "STAT-REG-MIXED-STRUCT-SLOPE-001",
84       "REG_CROSSED", crossed_config,
85       list("/coefficients/1/estimate" = tol(lme4::fixef(crossed_fit)[["x"]],
86                                                "crossed random-intercept fixed slope"),
87            "/analyzed_count" = exact(as.integer(nrow(crossed)), "balanced crossed cells")),
88       list("Two independent crossed grouping factors; lmer REML; bobyqa maxfun 100000."),
89       kind = "nominal")

Whole script generators/r/regression_advanced_axes.R


Regression II option axis STAT-REG-AXIS-ADV-STRUCT-CROSSED-ML-001

Case STAT-REG-AXIS-ADV-STRUCT-CROSSED-ML-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_CROSSED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.798969412405336480.79896941240504182.95E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 90  crossed_fit_ml <- lme4::lmer(y ~ x + (1 | a) + (1 | b), data = crossed, REML = FALSE,
 91                               control = lme4::lmerControl(optimizer = "bobyqa",
 92                                                             optCtrl = list(maxfun = 100000L)))
 93  crossed_config_ml <- crossed_config
 94  crossed_config_ml$estimation <- "ml"
 95  emit("STAT-REG-AXIS-ADV-STRUCT-CROSSED-ML-001", "STAT-REG-MIXED-STRUCT-SLOPE-001",
 96       "REG_CROSSED", crossed_config_ml,
 97       list("/coefficients/1/estimate" = tol(lme4::fixef(crossed_fit_ml)[["x"]],
 98                                                "ML crossed-model fixed slope")),
 99       list("Variance-component mode with ML explicitly; do not compare REML likelihoods."),
100       kind = "axis_coverage")

Whole script generators/r/regression_advanced_axes.R


Regression II remaining axis STAT-REG-AXIS-ADV-STRUCT-NESTED-001

Case STAT-REG-AXIS-ADV-STRUCT-NESTED-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_ADV_NESTED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.798956173466570660.79895617347135254.78E-1211passed
/analyzed_countexact64640.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

45  nested_fit <- lme4::lmer(y ~ x + (1 | a) + (1 | a:ab), data = nested, REML = TRUE,
46                           control = lme4::lmerControl(optimizer = "bobyqa",
47                                                         optCtrl = list(maxfun = 100000L)))
48  emit("STAT-REG-AXIS-ADV-STRUCT-NESTED-001", "STAT-REG-MIXED-STRUCT-SLOPE-001",
49       "REG_ADV_NESTED", list(response_column = "y", fixed_numeric_columns = list("x"),
50                          model_mode = "variance_components", estimation = "reml",
51                          max_rows = 1000L,
52                          components = list(list(name = "a", level_column = "a",
53                                                 relation = "crossed"),
54                                            list(name = "ab", level_column = "ab",
55                                                 relation = "nested", parent_column = "a"))),
56       list("/coefficients/1/estimate" = tol(lme4::fixef(nested_fit)[["x"]],
57                                                "nested fixed slope"),
58            "/analyzed_count" = exact(as.integer(nrow(nested)), "all nested observations")),
59       "R lmer a and a:ab random intercepts; ab labels are unique within parent a.")

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II remaining axis STAT-REG-AXIS-ADV-TREE-BALANCED-001

Case STAT-REG-AXIS-ADV-TREE-BALANCED-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/threshold/0tolerance00.00.00E+2passed
/source_countexact40400.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

62  class_cfg$class_weight <- "balanced"
63  class_cfg$split_assignment <- NULL
64  emit("STAT-REG-AXIS-ADV-TREE-BALANCED-001", "STAT-REG-TREE-CLASS-001",
65       "REG_TREE", class_cfg,
66       list("/artifact/threshold/0" = tol(0, "unambiguous class split", "1e-12"),
67            "/source_count" = exact(as.integer(nrow(tree)), "complete source rows")),
68       "Balanced weights on an exactly balanced two-class table preserve the pure split.",
69       kind = "determinism", repeat_count = 2L)

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II remaining axis STAT-REG-AXIS-ADV-TREE-REG-SEED-001

Case STAT-REG-AXIS-ADV-TREE-REG-SEED-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/threshold/0tolerance00.00.00E+2passed
/source_countexact40400.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

72  reg_cfg$seed <- 7L
73  emit("STAT-REG-AXIS-ADV-TREE-REG-SEED-001", "STAT-REG-TREE-REG-001",
74       "REG_TREE", reg_cfg,
75       list("/artifact/threshold/0" = tol(0, "unambiguous numeric split", "1e-12"),
76            "/source_count" = exact(as.integer(nrow(tree)), "complete source rows")),
77       "Seed changes the holdout, not the pure two-leaf split; repeat verifies replay.",
78       kind = "determinism", repeat_count = 2L)

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II remaining axis STAT-REG-AXIS-ADV-VALIDATED-KNN-001

Case STAT-REG-AXIS-ADV-VALIDATED-KNN-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_PLS_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact30300.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_numeric_countexact000.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II option axis STAT-REG-AXIS-ADV-VALIDATED-LASSO-001

Case STAT-REG-AXIS-ADV-VALIDATED-LASSO-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact1081080.00E+2passed
/training_metrics/row_countexact81810.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Regression II remaining axis STAT-REG-AXIS-ADV-VALIDATED-MEAN-001

Case STAT-REG-AXIS-ADV-VALIDATED-MEAN-001 · status: passed

Reference

Evidence type
R
Source
R lme4/rpart/MASS/pls
Citation
R lme4::lmer, rpart::rpart, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, lme4 2.0.6, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_remaining_axes.R · sha256 b1f30831b6d1324e
Dataset
reference_data/datasets/REG_PLS_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact30300.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_numeric_countexact000.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_remaining_axes.R


Regression II option axis STAT-REG-AXIS-ADV-VALIDATED-RIDGE-001

Case STAT-REG-AXIS-ADV-VALIDATED-RIDGE-001 · status: passed

Reference

Evidence type
R
Source
R statistical packages
Citation
R nlme::lme, lme4::lmer, plm::plm, stats::nls/step, MASS::lda, pls::plsr
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, lme4 2.0.6, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_axes.R · sha256 603ee3cda6c6de34
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact1081080.00E+2passed
/training_metrics/row_countexact81810.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_axes.R


Fit axis STAT-REG-AXIS-FIT-CLUSTER-001

Case STAT-REG-AXIS-FIT-CLUSTER-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/std_errortolerance0.00566832299337894420.0056683229933789321.22E-1714passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

116  X <- stats::model.matrix(default_fit)
117  e <- stats::residuals(default_fit)
118  score <- rowsum(X * as.vector(e), group = d$cluster)
119  B <- solve(crossprod(X))
120  G <- nrow(score)
121  N <- nrow(X)
122  K <- ncol(X)
123  cluster_cov <- (G / (G - 1)) * ((N - 1) / (N - K)) * B %*% crossprod(score) %*% B
124  emit("STAT-REG-AXIS-FIT-CLUSTER-001", fit_entry, "REG_CORE",
125       list(y_columns = list("y"), x_columns = list("x", "z"),
126            covariance = "cluster", cluster_column = "cluster"),
127       list("/responses/0/parameters/1/std_error" = tol(sqrt(diag(cluster_cov))[["x"]],
128                                                        "CR1 x SE")),
129       list("Manual CR1 factor is G/(G-1)*(N-1)/(N-K); cluster scores are row sums of x_i*e_i.",
130            "No random split is applied; G=12 clusters."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-EXPONENTIAL-001

Case STAT-REG-AXIS-FIT-EXPONENTIAL-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/estimatetolerance0.109382149534880030.109382149534880052.49E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

57  exp_fit <- stats::lm(log(y) ~ x, data = p)
58  emit("STAT-REG-AXIS-FIT-EXPONENTIAL-001", fit_entry, "REG_POSITIVE",
59       list(y_columns = list("y"), x_columns = list("x"), model_type = "exponential"),
60       list("/responses/0/parameters/1/estimate" = tol(coef(exp_fit)[["x"]], "log-response slope")),
61       list("Exponential fits log(y) against untransformed x; the reported coefficient is on log scale."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-HC3-001

Case STAT-REG-AXIS-FIT-HC3-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/std_errortolerance0.0325200952986305360.0325200952986305361.71E-1917passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

110  hc3 <- sandwich::vcovHC(default_fit, type = "HC3")
111  emit("STAT-REG-AXIS-FIT-HC3-001", fit_entry, "REG_CORE",
112       list(y_columns = list("y"), x_columns = list("x", "z"), covariance = "hc3"),
113       list("/responses/0/parameters/1/std_error" = tol(sqrt(diag(hc3))[["x"]], "HC3 x SE")),
114       list("HC3 sandwich covariance, t inference with N-K residual degrees of freedom."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-INTERACTIONS-001

Case STAT-REG-AXIS-FIT-INTERACTIONS-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/3/estimatetolerance0.000561809457073371480.00056180945707347191.00E-1612passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

79  interaction <- stats::lm(y ~ x * z, data = d)
80  emit("STAT-REG-AXIS-FIT-INTERACTIONS-001", fit_entry, "REG_CORE",
81       list(y_columns = list("y"), x_columns = list("x", "z"), interactions = TRUE),
82       list("/responses/0/parameters/3/estimate" = tol(coef(interaction)[["x:z"]], "x:z slope")),
83       list("Pairwise interaction is the product x*z, added after main effects."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-LASSO-001

Case STAT-REG-AXIS-FIT-LASSO-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CENTERED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/0/estimatetolerance55.00.00E+2passed
/responses/0/observationsexact72720.00E+2passed

Convention

Reference code

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-LOGARITHMIC-001

Case STAT-REG-AXIS-FIT-LOGARITHMIC-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/estimatetolerance4.7965418406563444.7965418406563443.57E-1717passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

62  log_fit <- stats::lm(y ~ log(x), data = p)
63  emit("STAT-REG-AXIS-FIT-LOGARITHMIC-001", fit_entry, "REG_POSITIVE",
64       list(y_columns = list("y"), x_columns = list("x"), model_type = "logarithmic"),
65       list("/responses/0/parameters/1/estimate" = tol(coef(log_fit)[[2]], "log-x slope")),
66       list("Logarithmic model fits untransformed y against log(x)."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-MISSING-KNN-001

Case STAT-REG-AXIS-FIT-MISSING-KNN-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/estimatetolerance1.3987626254070821.39876262540708221.83E-1615passed

Convention

Reference code

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-MISSING-NEIGHBOR-AVERAGE-001

Case STAT-REG-AXIS-FIT-MISSING-NEIGHBOR-AVERAGE-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/estimatetolerance1.3987626254070821.39876262540708221.83E-1615passed

Convention

Reference code

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-POLYNOMIAL-001

Case STAT-REG-AXIS-FIT-POLYNOMIAL-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/2/estimatetolerance0.0136333660630931010.0136333660630928692.32E-1613passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

52  poly <- stats::lm(y ~ x + I(x^2), data = p)
53  emit("STAT-REG-AXIS-FIT-POLYNOMIAL-001", fit_entry, "REG_POSITIVE",
54       list(y_columns = list("y"), x_columns = list("x"), model_type = "polynomial"),
55       list("/responses/0/parameters/2/estimate" = tol(coef(poly)[[3]], "quadratic coefficient")),
56       list("Polynomial degree two adds the square of numeric x."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-POWER-001

Case STAT-REG-AXIS-FIT-POWER-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/estimatetolerance0.382169341609478630.38216934160947865.55E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

67  power_fit <- stats::lm(log(y) ~ log(x), data = p)
68  emit("STAT-REG-AXIS-FIT-POWER-001", fit_entry, "REG_POSITIVE",
69       list(y_columns = list("y"), x_columns = list("x"), model_type = "power"),
70       list("/responses/0/parameters/1/estimate" = tol(coef(power_fit)[[2]], "log-log slope")),
71       list("Power model fits log(y) against log(x)."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-FIT-RIDGE-001

Case STAT-REG-AXIS-FIT-RIDGE-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CENTERED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/0/estimatetolerance55.00.00E+2passed
/responses/0/observationsexact72720.00E+2passed

Convention

Reference code

Whole script generators/r/regression_fit_axes.R


Regression setting STAT-REG-AXIS-GENERAL-INTERACTION-001

Case STAT-REG-AXIS-GENERAL-INTERACTION-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/residual_variancetolerance0.231111927969271880.23111192796927192.75E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

258  config_interaction <- list(response_column = "y", covariate_columns = list("x", "z"),
259                             terms = list(list(kind = "interaction", columns = list("x", "z"))))
260  interaction_fit <- stats::lm(y ~ x * z, data = d)
261  emit("STAT-REG-AXIS-GENERAL-INTERACTION-001", general_entry, "generalLinearMenuItem",
262       "nominal", config_interaction,
263       list("/residual_variance" = tol(summary(interaction_fit)$sigma^2,
264                                        "numeric interaction model MSE", "1e-9")),
265       list("The explicit x:z term adds one product column to the already included main effects."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-AXIS-GENERAL-MAIN-001

Case STAT-REG-AXIS-GENERAL-MAIN-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/xtolerance1.39591901807085271.39591901807085251.98E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

250  config_main <- list(response_column = "y", covariate_columns = list("x"),
251                      terms = list(list(kind = "main", columns = list("x"))))
252  main_fit <- stats::lm(y ~ x, data = d)
253  emit("STAT-REG-AXIS-GENERAL-MAIN-001", general_entry, "generalLinearMenuItem",
254       "nominal", config_main,
255       list("/coefficients/x" = tol(coef(main_fit)[["x"]], "explicit main-term slope", "1e-9")),
256       list("An explicit main term does not duplicate the automatically included covariate main effect."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-AXIS-GENERAL-NESTED-001

Case STAT-REG-AXIS-GENERAL-NESTED-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_NESTED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/residual_variancetolerance0.239654767295622540.239654767295622522.28E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

275  nested_data <- d
276  nested_data$child <- rep(c("L", "R"), length.out = nrow(nested_data))
277  export_dataset(nested_data, "REG_NESTED", "REG_CORE with balanced within-parent child factor",
278                 "R stats::lm factor nesting and Tabelz manual chapter 08",
279                 "Constructed test data; no third-party dataset license.",
280                 c(x = "Float64", z = "Float64", group = "String", binary = "String",
281                   ordinal = "String", nominal = "String", exposure = "Float64",
282                   cluster = "String", weight = "Float64", y = "Float64",
283                   count = "Int64", nb_count = "Int64", child = "String"))
284  nested_data <- read_dataset("REG_NESTED")
285  nested_fit <- stats::lm(y ~ x + group / child, data = nested_data,
286                          contrasts = list(group = stats::contr.sum(3),
287                                           child = stats::contr.sum(2)))
288  config_nested <- list(response_column = "y", factor_columns = list("group", "child"),
289                        covariate_columns = list("x"),
290                        terms = list(list(kind = "nested", columns = list("group", "child"))))
291  emit("STAT-REG-AXIS-GENERAL-NESTED-001", general_entry, "generalLinearMenuItem",
292       "nominal", config_nested,
293       list("/residual_variance" = tol(summary(nested_fit)$sigma^2,
294                                        "nested-factor residual variance", "1e-9")),
295       list("R group/child spans group and child-within-group contrasts; the child has no crossed main effect."),
296       dataset = "REG_NESTED")

Whole script generators/r/regression_options.R


Regression setting STAT-REG-AXIS-GENERAL-QUADRATIC-001

Case STAT-REG-AXIS-GENERAL-QUADRATIC-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/x^2tolerance0.0136333660630929050.0136333660630930421.37E-1613passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

267  config_quadratic <- list(response_column = "y", covariate_columns = list("x"),
268                           terms = list(list(kind = "quadratic", columns = list("x"))))
269  quadratic_fit <- stats::lm(y ~ x + I(x^2), data = d)
270  emit("STAT-REG-AXIS-GENERAL-QUADRATIC-001", general_entry, "generalLinearMenuItem",
271       "nominal", config_quadratic,
272       list("/coefficients/x^2" = tol(coef(quadratic_fit)[[3]], "quadratic x coefficient", "1e-9")),
273       list("A quadratic term is x squared and remains an adjustment term in Type II inference."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-AXIS-GENERAL-TYPEII-001

Case STAT-REG-AXIS-GENERAL-TYPEII-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/0/f_statistictolerance28.05837698495935228.0583769849593381.39E-1415passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

225  general_entry <- "tabelz.executor.general_linear"
226  general_config <- list(response_column = "y", factor_columns = list("group"),
227                         covariate_columns = list("x"))
228  general_fit <- stats::lm(y ~ group + x, data = d,
229                           contrasts = list(group = stats::contr.sum(3)))
230  type2 <- car::Anova(general_fit, type = 2)
231  config_ii <- general_config
232  config_ii$test_type <- "II"
233  emit("STAT-REG-AXIS-GENERAL-TYPEII-001", general_entry, "generalLinearMenuItem",
234       "nominal", config_ii,
235       list("/term_tests/0/f_statistic" = tol(type2["group", "F value"],
236                                               "Type II whole-factor F", "1e-9")),
237       list("car::Anova type=2 tests group conditional on x while preserving marginality."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-AXIS-GENERAL-WEIGHTED-001

Case STAT-REG-AXIS-GENERAL-WEIGHTED-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/residual_variancetolerance0.338891539962506430.338891539962506375.82E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

239  weighted_fit <- stats::lm(y ~ group + x, data = d, weights = weight,
240                            contrasts = list(group = stats::contr.sum(3)))
241  weighted_config <- general_config
242  weighted_config$weight_column <- "weight"
243  weighted_config$weight_kind <- "precision"
244  emit("STAT-REG-AXIS-GENERAL-WEIGHTED-001", general_entry, "generalLinearMenuItem",
245       "axis_coverage", weighted_config,
246       list("/residual_variance" = tol(summary(weighted_fit)$sigma^2,
247                                        "precision-weighted residual variance", "1e-9")),
248       list("Manual precision weights are inverse relative variances; R lm weights use the same WLS SSE/df."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-AXIS-POISSON-NOOFFSET-001

Case STAT-REG-AXIS-POISSON-NOOFFSET-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance0.180878934091863120.18087893408932912.53E-1210passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

105  no_offset <- stats::glm(count ~ x, data = d, family = stats::poisson())
106  emit("STAT-REG-AXIS-POISSON-NOOFFSET-001",
107       "tabelz.analysis.poisson_regression:compute_poisson_regression",
108       "poissonRegressionMenuItem", "nominal",
109       list(count_column = "count", predictors = list("x")),
110       list("/coefficients/1/coefficient" = tol(coef(no_offset)[["x"]], "no-offset log-count slope")),
111       list("Without exposure, the model is count intensity per row, not per unit exposure."))

Whole script generators/r/regression_options.R


Fit axis STAT-REG-AXIS-VIF-EXPONENTIAL-001

Case STAT-REG-AXIS-VIF-EXPONENTIAL-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance1.00010902110332341.00010902110332344.78E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

151  vif_config <- function(model, missing = "drop", interactions = FALSE, degree = 2L)
152    list(y_columns = list("y"), x_columns = list("x", "z"), model_type = model,
153         missing = missing, degree = degree, interactions = interactions,
154         split_percent = 100L)
155  vif_r <- function(x, z) 1 / (1 - summary(stats::lm(x ~ z))$r.squared)
156  emit("STAT-REG-AXIS-VIF-EXPONENTIAL-001", vif_entry, "REG_POSITIVE",
157       vif_config("exponential"), list("/0" = tol(vif_r(p$x, p$z), "exponential X VIF")),
158       list("Exponential response transformation leaves the X design untransformed."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-VIF-INTERACTIONS-001

Case STAT-REG-AXIS-VIF-INTERACTIONS-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance1.11944685576133881.1194468557613392.66E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

174  interaction_design <- data.frame(x = d$x, z = d$z, xz = d$x * d$z)
175  interaction_vif <- 1 / (1 - summary(stats::lm(x ~ z + xz, data = interaction_design))$r.squared)
176  emit("STAT-REG-AXIS-VIF-INTERACTIONS-001", vif_entry, "REG_CORE",
177       vif_config("linear", interactions = TRUE),
178       list("/0" = tol(interaction_vif, "x VIF with x:z interaction")),
179       list("The base X interaction x*z is an additional VIF design term."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-VIF-LOGARITHMIC-001

Case STAT-REG-AXIS-VIF-LOGARITHMIC-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance1.00020187540320561.00020187540320542.24E-1615passed

Convention

Reference code

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-VIF-MISSING-KNN-001

Case STAT-REG-AXIS-VIF-MISSING-KNN-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance1.00010902110332341.00010902110332372.70E-1615passed

Convention

Reference code

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-VIF-MISSING-NEIGHBOR-AVERAGE-001

Case STAT-REG-AXIS-VIF-MISSING-NEIGHBOR-AVERAGE-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance1.00010902110332341.00010902110332372.70E-1615passed

Convention

Reference code

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-VIF-POLYNOMIAL-001

Case STAT-REG-AXIS-VIF-POLYNOMIAL-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance31.87892959292359131.8789295929233951.96E-1314passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

164  poly_design <- data.frame(x = p$x, z = p$z, x2 = p$x^2, z2 = p$z^2)
165  poly_vif <- 1 / (1 - summary(stats::lm(x ~ z + x2 + z2, data = poly_design))$r.squared)
166  emit("STAT-REG-AXIS-VIF-POLYNOMIAL-001", vif_entry, "REG_POSITIVE",
167       vif_config("polynomial"), list("/0" = tol(poly_vif, "quadratic-design x VIF")),
168       list("Quadratic design includes x, z, x-squared and z-squared; R auxiliary regression conditions on the other three."))

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-AXIS-VIF-POWER-001

Case STAT-REG-AXIS-VIF-POWER-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance1.00020187540320561.00020187540320542.24E-1615passed

Convention

Reference code

Whole script generators/r/regression_fit_axes.R


All models of x1 and x2, Zellner-Siow medium scale

Case STAT-REG-BAYES-REGRESSION-001 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, pt, solve, combn
Citation
Liang, Paulo, Molina, Clyde and Berger (2008), Mixtures of g priors for Bayesian variable selection, Journal of the American Statistical Association 103(481), 410-423, eqs. 5 and 9, section 3.2; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/[numerator=x1]/bf10tolerance159.54961679681637159.549616796816082.88E-1314passed
/bayes_factors/[numerator=x1]/bf01tolerance0.00626764275638143610.00626764275638144651.04E-1714passed
/bayes_factors/[numerator=x1]/log_bf10tolerance5.07235495094647875.0723549509464771.77E-1515passed
/bayes_factors/[numerator=x1]/posterior_probabilitytolerance0.699376343625691680.69937634362569124.48E-1615passed
/bayes_factors/[numerator=x1]/prior_probabilitytolerance0.250.250.00E+0passed
/bayes_factors/[numerator=x2]/bf10tolerance0.405168486108855810.405168486108855532.80E-1615passed
/bayes_factors/[numerator=x2]/bf01tolerance2.46810903188391612.4681090318839181.74E-1515passed
/bayes_factors/[numerator=x2]/log_bf10tolerance-0.903452283302286-0.90345228330228676.63E-1615passed
/bayes_factors/[numerator=x2]/posterior_probabilitytolerance0.00177603218394456760.0017760321839445683.99E-1915passed
/bayes_factors/[numerator=x2]/prior_probabilitytolerance0.250.250.00E+0passed
/bayes_factors/[numerator=x1 + x2]/bf10tolerance67.1764894606308967.176489460630896.19E-1819passed
/bayes_factors/[numerator=x1 + x2]/bf01tolerance0.0148861604413855970.0148861604413855971.86E-1916passed
/bayes_factors/[numerator=x1 + x2]/log_bf10tolerance4.20732332709728764.2073233270972882.41E-1717passed
/bayes_factors/[numerator=x1 + x2]/posterior_probabilitytolerance0.294464183116253720.2944641831162542.77E-1615passed
/bayes_factors/[numerator=x1 + x2]/prior_probabilitytolerance0.250.250.00E+0passed
/posteriors/0/meantolerance0.930669504254175610.93066950425417593.29E-1615passed
/posteriors/0/sdtolerance0.242005325119139330.24200532511913984.72E-1614passed
/posteriors/0/mediantolerance0.936208055426372420.93620805542637724.77E-1514passed
/posteriors/0/lowertolerance0.434804710089630290.4348047100896282.28E-1514passed
/posteriors/0/uppertolerance1.39510182360089451.39510182360089472.42E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

572    emit("regression", "bayesRegressionMenuItem",
573         "tabelz.analysis.bayesian.linear:compute_bayes_regression", id, title, kind,
574         configuration, outputs,
575         c(sprintf("Candidate models: %s, plus intercept only; n = %d; the most probable model (null included) is %s, whose coefficient posteriors are compared by index.",
576                   paste(vapply(models, paste, "", collapse = " + "), collapse = "; "), n, selected),
577           "BF10(g) = (1 + g)^((n-1-p)/2) (1 + g (1 - R^2))^(-(n-1)/2) against the intercept-only model (Liang, Paulo, Molina, Clyde and Berger 2008, eq. 5), averaged over g by R integrate on u = log g.",
578           if (prior == "hyper_g") sprintf("Hyper-g prior pi(g) = (a - 2)/2 (1 + g)^(-a/2), a = %g (Liang et al. 2008, eq. 9).", hyper_a)
579           else sprintf("Zellner-Siow prior g ~ Inverse-Gamma(1/2, n r^2/2), r = %s (Liang et al. 2008, section 3.2).", format(r, digits = 17)),
580           "Posterior model probabilities proportional to prior probability times BF10 over the candidate set including the intercept-only model (BF 1).",
581           "Coefficient posteriors given g: Student-t on n - 1 df with centre g/(1+g) beta-hat and scale sqrt(g/(1+g) s^2 (X'X)^-1_jj), s^2 = SSY (1 - g/(1+g) R^2)/(n - 1), mixed over the posterior of g by R integrate; quantiles by uniroot.",
582           "Tolerance 1e-6: both sides integrate numerically.",
583           paste0(manual, ":33-34: 'Linear regression & Zellner--Siow or hyper-g prior; declare candidate models and their prior probabilities, including the intercept-only model.'"),
584           paste0(manual, ":93-96: 'Regression compares up to 64 declared models ... Coefficient summaries for the most probable model are conditional on that model'")),
585         "R stats: integrate, uniroot, pt, solve, combn",
586         "Liang, Paulo, Molina, Clyde and Berger (2008), Mixtures of g priors for Bayesian variable selection, Journal of the American Statistical Association 103(481), 410-423, eqs. 5 and 9, section 3.2; R Core Team, stats package help: integrate, uniroot.",
587         "REQ-STAT-REG-BAYES-REGRESSION-001")

Whole script generators/r/hypothesis_bayesian.R


Hyper-g prior, a = 3

Case STAT-REG-BAYES-REGRESSION-002 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, pt, solve, combn
Citation
Liang, Paulo, Molina, Clyde and Berger (2008), Mixtures of g priors for Bayesian variable selection, Journal of the American Statistical Association 103(481), 410-423, eqs. 5 and 9, section 3.2; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/[numerator=x1]/bf10tolerance126.09221116684985126.092211166849642.13E-1314passed
/bayes_factors/[numerator=x1]/bf01tolerance0.00793070397248219620.007930703972482211.39E-1714passed
/bayes_factors/[numerator=x1]/log_bf10tolerance4.83701347394861444.8370134739486131.81E-1515passed
/bayes_factors/[numerator=x1]/posterior_probabilitytolerance0.696294369534593780.69629436953459334.42E-1615passed
/bayes_factors/[numerator=x1]/prior_probabilitytolerance0.250.250.00E+0passed
/bayes_factors/[numerator=x2]/bf10tolerance0.507269356089077750.50726935608907761.11E-1615passed
/bayes_factors/[numerator=x2]/bf01tolerance1.97133926580890791.97133926580890867.14E-1615passed
/bayes_factors/[numerator=x2]/log_bf10tolerance-0.67871314212976863-0.6787131421297693.35E-1615passed
/bayes_factors/[numerator=x2]/posterior_probabilitytolerance0.00280119440537754480.0028011944053775483.04E-1814passed
/bayes_factors/[numerator=x2]/prior_probabilitytolerance0.250.250.00E+0passed
/bayes_factors/[numerator=x1 + x2]/bf10tolerance53.49089832233369953.49089832233371.00E-1617passed
/bayes_factors/[numerator=x1 + x2]/bf01tolerance0.0186947692292256140.0186947692292256141.54E-1917passed
/bayes_factors/[numerator=x1 + x2]/log_bf10tolerance3.97951151461290033.97951151461290031.36E-1717passed
/bayes_factors/[numerator=x1 + x2]/posterior_probabilitytolerance0.295382331537543620.2953823315375443.87E-1614passed
/bayes_factors/[numerator=x1 + x2]/prior_probabilitytolerance0.250.250.00E+0passed
/posteriors/0/meantolerance0.926143106837962080.92614310683796221.13E-1615passed
/posteriors/0/sdtolerance0.247307035867695170.24730703586769651.34E-1514passed
/posteriors/0/mediantolerance0.933987324999279610.93398731510157759.90E-97passed
/posteriors/0/lowertolerance0.409874787325764710.409874605168549341.82E-76passed
/posteriors/0/uppertolerance1.39444877787921921.39444877424304183.64E-98passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

591  regression_case("STAT-REG-BAYES-REGRESSION-002", "Hyper-g prior, a = 3", "axis_coverage",
592                  c("x1", "x2"), NULL, prior = "hyper_g", hyper_a = 3)

Whole script generators/r/hypothesis_bayesian.R


Zellner-Siow unit scale, 90% interval

Case STAT-REG-BAYES-REGRESSION-003 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, pt, solve, combn
Citation
Liang, Paulo, Molina, Clyde and Berger (2008), Mixtures of g priors for Bayesian variable selection, Journal of the American Statistical Association 103(481), 410-423, eqs. 5 and 9, section 3.2; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/[numerator=x1]/bf10tolerance235.46934106784553235.46934106784496.22E-1314passed
/bayes_factors/[numerator=x1]/bf01tolerance0.00424683738216208390.0042468373821620951.13E-1714passed
/bayes_factors/[numerator=x1]/log_bf10tolerance5.46158071843304165.4615807184330392.66E-1515passed
/bayes_factors/[numerator=x1]/posterior_probabilitytolerance0.757915851430850830.75791585143085035.56E-1615passed
/bayes_factors/[numerator=x1]/prior_probabilitytolerance0.250.250.00E+0passed
/bayes_factors/[numerator=x2]/bf10tolerance0.175152720126847320.17515272012684732.56E-1715passed
/bayes_factors/[numerator=x2]/bf01tolerance5.70930328273400535.7093032827340071.81E-1515passed
/bayes_factors/[numerator=x2]/log_bf10tolerance-1.7420969991889101-1.74209699918891042.67E-1615passed
/bayes_factors/[numerator=x2]/posterior_probabilitytolerance0.000563772007019459810.00056377200701946091.09E-1814passed
/bayes_factors/[numerator=x2]/prior_probabilitytolerance0.250.250.00E+0passed
/bayes_factors/[numerator=x1 + x2]/bf10tolerance74.03556472156891774.035564721568924.35E-1617passed
/bayes_factors/[numerator=x1 + x2]/bf01tolerance0.0135070219800547850.0135070219800547851.46E-1916passed
/bayes_factors/[numerator=x1 + x2]/log_bf10tolerance4.30454558209641744.3045455820964174.46E-1716passed
/bayes_factors/[numerator=x1 + x2]/posterior_probabilitytolerance0.238301631191740030.23830163119174054.68E-1614passed
/bayes_factors/[numerator=x1 + x2]/prior_probabilitytolerance0.250.250.00E+0passed
/posteriors/0/meantolerance0.982971450756496950.98297145075649755.50E-1615passed
/posteriors/0/sdtolerance0.22436279813745430.22436279813745431.18E-1918passed
/posteriors/0/mediantolerance0.98369525723316920.98369525723316996.62E-1615passed
/posteriors/0/lowertolerance0.614597278083640170.6145972780836241.62E-1413passed
/posteriors/0/uppertolerance1.34882706911150851.34882706911153722.87E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

593  regression_case("STAT-REG-BAYES-REGRESSION-003", "Zellner-Siow unit scale, 90% interval",
594                  "axis_coverage", c("x1", "x2"), NULL, r = 1, level = 0.9)

Whole script generators/r/hypothesis_bayesian.R


Declared models with custom prior weights

Case STAT-REG-BAYES-REGRESSION-004 · status: passed

Reference

Evidence type
R
Source
R stats: integrate, uniroot, pt, solve, combn
Citation
Liang, Paulo, Molina, Clyde and Berger (2008), Mixtures of g priors for Bayesian variable selection, Journal of the American Statistical Association 103(481), 410-423, eqs. 5 and 9, section 3.2; R Core Team, stats package help: integrate, uniroot.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/hypothesis_bayesian.R · sha256 5f55d2b476c40d20
Dataset
reference_data/datasets/HYP_BAYES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact20200.00E+2passed
/bayes_factors/[numerator=x1]/bf10tolerance159.54961679681637159.549616796816082.88E-1314passed
/bayes_factors/[numerator=x1]/bf01tolerance0.00626764275638143610.00626764275638144651.04E-1714passed
/bayes_factors/[numerator=x1]/log_bf10tolerance5.07235495094647875.0723549509464771.77E-1515passed
/bayes_factors/[numerator=x1]/posterior_probabilitytolerance0.823958783362523710.82395878336252352.26E-1615passed
/bayes_factors/[numerator=x1]/prior_probabilitytolerance0.50.50.00E+1passed
/bayes_factors/[numerator=x1 + x2]/bf10tolerance67.1764894606308967.176489460630896.19E-1819passed
/bayes_factors/[numerator=x1 + x2]/bf01tolerance0.0148861604413855970.0148861604413855971.86E-1916passed
/bayes_factors/[numerator=x1 + x2]/log_bf10tolerance4.20732332709728764.2073233270972882.41E-1717passed
/bayes_factors/[numerator=x1 + x2]/posterior_probabilitytolerance0.173459076987426740.173459076987427053.07E-1614passed
/bayes_factors/[numerator=x1 + x2]/prior_probabilitytolerance0.250.250.00E+0passed
/posteriors/0/meantolerance0.930669504254175610.93066950425417593.29E-1615passed
/posteriors/0/sdtolerance0.242005325119139330.24200532511913984.72E-1614passed
/posteriors/0/mediantolerance0.936208055426372420.93620805542637724.77E-1514passed
/posteriors/0/lowertolerance0.434804710089630290.4348047100896282.28E-1514passed
/posteriors/0/uppertolerance1.39510182360089451.39510182360089472.42E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

595  regression_case("STAT-REG-BAYES-REGRESSION-004", "Declared models with custom prior weights",
596                  "axis_coverage", c("x1", "x2"), list("x1", c("x1", "x2")),
597                  model_prior = c(0.5, 0.25), null_prior = 0.25)

Whole script generators/r/hypothesis_bayesian.R


STAT REG BINARY 001

Case STAT-REG-BINARY-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS and nnet
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_core.R · sha256 4b30ab4a6f513e54
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance0.0722220614813967740.072222061481396725.58E-1715passed
/log_likelihoodtolerance-18.127056845888642-18.1270568458886423.00E-1616passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

87  binary_entry <- "tabelz.analysis.logistic_regression:compute_binary_logistic_regression"
88  binary_fit <- stats::glm(I(binary == "event") ~ x, family = stats::binomial(), data = d)
89  emit("STAT-REG-BINARY-001", binary_entry, "logisticRegressionMenuItem", "nominal",
90       list(outcome_column = "binary", event_value = "event", predictors = list("x")),
91       list("/coefficients/1/coefficient" = tol(coef(binary_fit)[["x"]], "binary logit slope"),
92            "/log_likelihood" = tol(as.numeric(stats::logLik(binary_fit)), "binomial log likelihood")),
93       list("R glm binomial logit; event explicitly coded TRUE; Wald inference per manual."))

Whole script generators/r/regression_core.R


Regression setting STAT-REG-BOUNDARY-BINARY-CORRECTED-001

Case STAT-REG-BOUNDARY-BINARY-CORRECTED-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_BIN_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact330.00E+2passed
/coefficients/1/coefficienttolerance08.66101954420646e-168.66E-16passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

179  binary_min <- d[1:3, , drop = FALSE]
180  binary_min$x <- c(-1, 0, 1)
181  binary_min$binary <- c("event", "other", "event")
182  export_dataset(binary_min, "REG_BIN_MIN", "Three deliberately non-separated binary rows",
183                 "R stats::glm binomial manual; three rows exceed intercept-plus-slope count",
184                 "Constructed test data; no third-party dataset license.",
185                 c(x = "Float64", z = "Float64", group = "String", binary = "String",
186                   ordinal = "String", nominal = "String", exposure = "Float64",
187                   cluster = "String", weight = "Float64", y = "Float64",
188                   count = "Int64", nb_count = "Int64"))
189  binary_min <- read_dataset("REG_BIN_MIN")
190  binary_min_fit <- stats::glm(I(binary == "event") ~ x, data = binary_min,
191                               family = stats::binomial())
192  emit("STAT-REG-BOUNDARY-BINARY-CORRECTED-001",
193       "tabelz.analysis.logistic_regression:compute_binary_logistic_regression",
194       "logisticRegressionMenuItem", "boundary",
195       list(outcome_column = "binary", event_value = "event", predictors = list("x")),
196       list("/n" = exact(3, "three complete rows"),
197            "/coefficients/1/coefficient" = list(comparison = "tolerance", value = "0", absolute_tolerance = "1e-12", relative_tolerance = "0",
198                  rationale = paste("Exactly zero by the design's symmetry; any implementation returns rounding",
199                                    "noise of order 1e-15, so an absolute 1e-12 replaces a relative tier that",
200                                    "cannot compare with zero. R gives", format(coef(binary_min_fit)[["x"]], digits = 17)))),
201       list("Events occur at x=-1 and +1 and non-event at x=0: both outcomes, finite MLE and residual df one.",
202            "This corrects coverage lost by the separately retained failed REG_MIN3 draft."),
203       dataset = "REG_BIN_MIN")

Whole script generators/r/regression_options.R


Design boundary STAT-REG-BOUNDARY-BINARY-MIN-001

Case STAT-REG-BOUNDARY-BINARY-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_MIN3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

80  bfit <- stats::glm(I(binary == "event") ~ x, data = d3, family = stats::binomial())
81  emit("STAT-REG-BOUNDARY-BINARY-MIN-001",
82       "tabelz.analysis.logistic_regression:compute_binary_logistic_regression",
83       "logisticRegressionMenuItem", "REG_MIN3",
84       list(outcome_column = "binary", event_value = "event", predictors = list("x")),
85       list(),
86       list("REG_MIN3's three rows all have the same outcome, so a binary logit is undefined and must be refused.",
87            "Until 2026-09-25 this case wrongly asserted both outcomes occur; STAT-REG-BOUNDARY-BINARY-CORRECTED-001 covers the valid minimum design."),
88       error = list(error_class = "ValueError", message_contains = "exactly two outcome values"))

Whole script generators/r/regression_boundaries.R


Regression completion STAT-REG-BOUNDARY-BINARY-MIN-PASS-001

Case STAT-REG-BOUNDARY-BINARY-MIN-PASS-001 · status: passed

Reference

Evidence type
R
Source
R regression fit
Citation
R stats::glm/lm and car::Anova help pages; Tabelz manual chapters 07 and 08
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_lastmile.R · sha256 512832af1db86a5f
Dataset
reference_data/datasets/REG_BIN_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact330.00E+2passed
/coefficients/0/coefficienttolerance0.693147180559742780.69314718055974243.32E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

28  binary_entry <- "tabelz.analysis.logistic_regression:compute_binary_logistic_regression"
29  binary_menu <- "logisticRegressionMenuItem"
30  small_fit <- stats::glm(I(binary == "event") ~ x, data = small,
31                          family = stats::binomial())
32  emit("STAT-REG-BOUNDARY-BINARY-MIN-PASS-001", binary_entry, binary_menu,
33       "boundary", "REG_BIN_MIN",
34       list(outcome_column = "binary", event_value = "event", predictors = list("x")),
35       list("/n" = exact(3, "three-row accepted minimum"),
36            "/coefficients/0/coefficient" = tol(coef(small_fit)[[1]], "nonzero intercept")),
37       list("Symmetric events at x=-1 and +1 make the slope mathematically zero; compare the nonzero intercept and n instead."))

Whole script generators/r/regression_lastmile.R


Design boundary STAT-REG-BOUNDARY-BINARY-ONELEVEL-001

Case STAT-REG-BOUNDARY-BINARY-ONELEVEL-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_DEGENERATE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

89  emit("STAT-REG-BOUNDARY-BINARY-ONELEVEL-001",
90       "tabelz.analysis.logistic_regression:compute_binary_logistic_regression",
91       "logisticRegressionMenuItem", "REG_DEGENERATE",
92       list(outcome_column = "binary", event_value = "event", predictors = list("x")),
93       list(), list("Only one observed binary level is present."),
94       list(error_class = "ValueError", message_contains = "exactly two"))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-EFFECTS-MIN-001

Case STAT-REG-BOUNDARY-EFFECTS-MIN-001 · status: passed

Reference

Evidence type
R
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_MIN3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/xtolerance-0.090029661103937703-0.0900296611039382.91E-1614passed
/residual_dfexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

193  emit("STAT-REG-BOUNDARY-EFFECTS-MIN-001", "tabelz.executor.linear_effects",
194       "linearEffectsMenuItem", "REG_MIN3",
195       list(response_column = "y", covariate_columns = list("x")),
196       list("/coefficients/x" = tol(coef(gfit)[["x"]], "n=3 effects slope", "1e-9"),
197            "/residual_df" = exact(1, "one residual degree of freedom")),
198       list("Three rows, intercept and one covariate give residual df one."))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-EFFECTS-ONELEVEL-001

Case STAT-REG-BOUNDARY-EFFECTS-ONELEVEL-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_DEGENERATE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

199  emit("STAT-REG-BOUNDARY-EFFECTS-ONELEVEL-001", "tabelz.executor.linear_effects",
200       "linearEffectsMenuItem", "REG_DEGENERATE",
201       list(response_column = "y", factor_columns = list("group")),
202       list(), list("A one-level factor has no estimable contrast."),
203       list(error_class = "LinearEffectsExecutionError", message_contains = "two"))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-GENERAL-MIN-001

Case STAT-REG-BOUNDARY-GENERAL-MIN-001 · status: passed

Reference

Evidence type
R
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_MIN3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/xtolerance-0.090029661103937703-0.090029661103938721.01E-1513passed
/residual_dfexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

180  gfit <- stats::lm(y ~ x, data = d3)
181  emit("STAT-REG-BOUNDARY-GENERAL-MIN-001", "tabelz.executor.general_linear",
182       "generalLinearMenuItem", "REG_MIN3",
183       list(response_column = "y", covariate_columns = list("x")),
184       list("/coefficients/x" = tol(coef(gfit)[["x"]], "n=3 general-linear slope", "1e-9"),
185            "/residual_df" = exact(1, "one residual degree of freedom")),
186       list("Three rows, intercept and one covariate give residual df one."))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-GENERAL-ONELEVEL-001

Case STAT-REG-BOUNDARY-GENERAL-ONELEVEL-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_DEGENERATE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

187  emit("STAT-REG-BOUNDARY-GENERAL-ONELEVEL-001", "tabelz.executor.general_linear",
188       "generalLinearMenuItem", "REG_DEGENERATE",
189       list(response_column = "y", factor_columns = list("group")),
190       list(), list("A one-level factor has no estimable contrast."),
191       list(error_class = "GeneralLinearExecutionError", message_contains = "two"))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-NB-MIN-001

Case STAT-REG-BOUNDARY-NB-MIN-001 · status: passed

Reference

Evidence type
R
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_MIN8.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance-1.1143482340505979-1.11434823405019334.05E-1312passed
/nexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

164  nbfit <- MASS::glm.nb(nb_count ~ x, data = d8,
165                        control = stats::glm.control(maxit = 1000, epsilon = 1e-12))
166  emit("STAT-REG-BOUNDARY-NB-MIN-001",
167       "tabelz.analysis.negative_binomial:compute_negative_binomial",
168       "nbRegressionMenuItem", "REG_MIN8",
169       list(count = "nb_count", predictors = list("x")),
170       list("/coefficients/1/coefficient" = tol(coef(nbfit)[["x"]], "small-design NB2 slope"),
171            "/n" = exact(8, "eight observations")),
172       list("Eight-row design is the smallest constructed subset on which the R NB2 fit has finite dispersion."))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-NB-ZEROEXPOSURE-001

Case STAT-REG-BOUNDARY-NB-ZEROEXPOSURE-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_DEGENERATE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

173  emit("STAT-REG-BOUNDARY-NB-ZEROEXPOSURE-001",
174       "tabelz.analysis.negative_binomial:compute_negative_binomial",
175       "nbRegressionMenuItem", "REG_DEGENERATE",
176       list(count = "nb_count", predictors = list("x"), exposure = "exposure"),
177       list(), list("One exposure is zero; manual requires positive exposure."),
178       list(error_class = "ValueError", message_contains = "exposure"))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-NOMINAL-MIN-001

Case STAT-REG-BOUNDARY-NOMINAL-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_MIN5.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

133  stopifnot(max(d5$x[d5$nominal == "C"]) < min(d5$x[d5$nominal != "C"]))
134  emit("STAT-REG-BOUNDARY-NOMINAL-MIN-001",
135       "tabelz.analysis.nominal_logistic:compute_nominal_logistic_regression",
136       "nominalRegressionMenuItem", "REG_MIN5",
137       list(outcome_column = "nominal", reference_value = "A", predictors = list("x")),
138       list(),
139       list("Issue #11 ruling: complete separation of C establishes that no finite MLE exists; the fit is refused.",
140            "Until 2026-09-25 this case compared nnet::multinom's arbitrary stopping point."),
141       list(error_class = "ValueError", message_contains = "No finite maximum-likelihood estimate"))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-NOMINAL-ONELEVEL-001

Case STAT-REG-BOUNDARY-NOMINAL-ONELEVEL-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_DEGENERATE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

142  emit("STAT-REG-BOUNDARY-NOMINAL-ONELEVEL-001",
143       "tabelz.analysis.nominal_logistic:compute_nominal_logistic_regression",
144       "nominalRegressionMenuItem", "REG_DEGENERATE",
145       list(outcome_column = "nominal", predictors = list("x")),
146       list(), list("Only one nominal category is present."),
147       list(error_class = "ValueError", message_contains = "three"))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-OLS-DOMAIN-001

Case STAT-REG-BOUNDARY-OLS-DOMAIN-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

60  emit("STAT-REG-BOUNDARY-OLS-DOMAIN-001", "tabelz.analysis.fit_y_by_x:compute_fits",
61       "fitYByXMenuItem", "REG_CORE",
62       list(y_columns = list("y"), x_columns = list("x"), model_type = "logarithmic"),
63       list(), list("Logarithmic regression requires strictly positive x; REG_CORE contains negative x."),
64       list(error_class = "ValueError", message_contains = "positive"))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-OLS-MIN-001

Case STAT-REG-BOUNDARY-OLS-MIN-001 · status: passed

Reference

Evidence type
R
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_MIN3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/estimatetolerance-0.090029661103937703-0.090029661103937881.80E-1614passed
/responses/0/observationsexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

49  d3 <- read_dataset("REG_MIN3")
50  d4 <- read_dataset("REG_MIN4")
51  d5 <- read_dataset("REG_MIN5")
52  d8 <- read_dataset("REG_MIN8")
53  fit <- stats::lm(y ~ x, data = d3)
54  emit("STAT-REG-BOUNDARY-OLS-MIN-001", "tabelz.analysis.fit_y_by_x:compute_fits",
55       "fitYByXMenuItem", "REG_MIN3",
56       list(y_columns = list("y"), x_columns = list("x")),
57       list("/responses/0/parameters/1/estimate" = tol(coef(fit)[["x"]], "n=3 slope", "1e-9"),
58            "/responses/0/observations" = exact(3, "three observations")),
59       list("Three rows give one residual degree of freedom for intercept plus one slope."))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-ORDINAL-MIN-001

Case STAT-REG-BOUNDARY-ORDINAL-MIN-001 · status: passed

Reference

Evidence type
R
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_MIN4.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/coefficienttolerance-10.437130316938166-10.4371300700424782.47E-77passed
/nexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

108  ord_start <- c(coef(ofit)[["x"]], ofit$zeta[[1L]], log(ofit$zeta[[2L]] - ofit$zeta[[1L]]))
109  ord_refined <- stats::optim(ord_start, function(p) -ord_loglik(p), method = "Nelder-Mead",
110                              control = list(reltol = 1e-16, maxit = 100000L))
111  stopifnot(-ord_refined$value >= ord_loglik(ord_start))
112  ord_min_slope <- ord_refined$par[[1L]]
113  emit("STAT-REG-BOUNDARY-ORDINAL-MIN-001",
114       "tabelz.analysis.ordinal_logistic:compute_ordinal_logistic_regression",
115       "ordinalRegressionMenuItem", "REG_MIN4",
116       list(outcome_column = "ordinal", predictors = list("x"),
117            level_order = list("low", "middle", "high")),
118       list("/coefficients/0/coefficient" = tol(ord_min_slope, "n=4 ordered slope, polr refined by Nelder-Mead"),
119            "/n" = exact(4, "four observations")),
120       list("Four rows exceed one slope plus two thresholds."))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-ORDINAL-ONELEVEL-001

Case STAT-REG-BOUNDARY-ORDINAL-ONELEVEL-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_DEGENERATE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

121  emit("STAT-REG-BOUNDARY-ORDINAL-ONELEVEL-001",
122       "tabelz.analysis.ordinal_logistic:compute_ordinal_logistic_regression",
123       "ordinalRegressionMenuItem", "REG_DEGENERATE",
124       list(outcome_column = "ordinal", predictors = list("x")),
125       list(), list("Only one ordered level is present."),
126       list(error_class = "ValueError", message_contains = "three"))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-POISSON-MIN-001

Case STAT-REG-BOUNDARY-POISSON-MIN-001 · status: passed

Reference

Evidence type
R
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_MIN3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance-13.866489280647015-13.8664892888881348.24E-99passed
/nexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

149  pfit <- stats::glm(count ~ x, data = d3, family = stats::poisson())
150  emit("STAT-REG-BOUNDARY-POISSON-MIN-001",
151       "tabelz.analysis.poisson_regression:compute_poisson_regression",
152       "poissonRegressionMenuItem", "REG_MIN3",
153       list(count_column = "count", predictors = list("x")),
154       list("/coefficients/1/coefficient" = tol(coef(pfit)[["x"]], "n=3 Poisson slope"),
155            "/n" = exact(3, "three observations")),
156       list("Three rows exceed intercept and slope count."))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-POISSON-ZEROEXPOSURE-001

Case STAT-REG-BOUNDARY-POISSON-ZEROEXPOSURE-001 · status: passed

Reference

Evidence type
Analytical
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_DEGENERATE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

157  emit("STAT-REG-BOUNDARY-POISSON-ZEROEXPOSURE-001",
158       "tabelz.analysis.poisson_regression:compute_poisson_regression",
159       "poissonRegressionMenuItem", "REG_DEGENERATE",
160       list(count_column = "count", predictors = list("x"), exposure_column = "exposure"),
161       list(), list("One exposure is zero; manual requires positive exposure."),
162       list(error_class = "ValueError", message_contains = "exposure"))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-VIF-COLLINEAR-001

Case STAT-REG-BOUNDARY-VIF-COLLINEAR-001 · status: passed

Reference

Evidence type
R
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_DEGENERATE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

73  emit("STAT-REG-BOUNDARY-VIF-COLLINEAR-001", "tabelz.analysis.fit_y_by_x:compute_fit_vif",
74       "fitYByXMenuItem", "REG_DEGENERATE",
75       list(y_columns = list("y"), x_columns = list("x", "z"), missing = "drop",
76            model_type = "linear", degree = 2L, interactions = FALSE, split_percent = 100L),
77       list("/0" = list(comparison = "absent", rationale = "Perfect collinearity makes VIF undefined.")),
78       list("z equals x exactly; R auxiliary regression has R-squared one, so VIF is undefined."))

Whole script generators/r/regression_boundaries.R


Design boundary STAT-REG-BOUNDARY-VIF-MIN-001

Case STAT-REG-BOUNDARY-VIF-MIN-001 · status: passed

Reference

Evidence type
R
Source
R minimal-design fit and manual design conditions
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_boundaries.R · sha256 d598c8abc1c5a2a7
Dataset
reference_data/datasets/REG_MIN3.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance35.79267242514120535.792672425140289.23E-1313passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

66  vif_fit <- stats::lm(x ~ z, data = d3)
67  emit("STAT-REG-BOUNDARY-VIF-MIN-001", "tabelz.analysis.fit_y_by_x:compute_fit_vif",
68       "fitYByXMenuItem", "REG_MIN3",
69       list(y_columns = list("y"), x_columns = list("x", "z"), missing = "drop",
70            model_type = "linear", degree = 2L, interactions = FALSE, split_percent = 100L),
71       list("/0" = tol(1 / (1 - summary(vif_fit)$r.squared), "n=3 auxiliary VIF", "1e-9")),
72       list("Three rows are the smallest design for an auxiliary intercept-plus-slope fit."))

Whole script generators/r/regression_boundaries.R


CHAID root split on a binary nominal predictor

Case STAT-REG-CHAID-001 · status: passed

Reference

Evidence type
Analytical
Source
perfect-association chi-square identity (c - 1) N, checked with stats::chisq.test; Kass (1980) Bonferroni multiplier
Citation
Kass (1980) An exploratory technique for investigating large quantities of categorical data, Applied Statistics 29(2), 119-127; R Core Team, stats package help: chisq.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_chaid.R · sha256 eb0c2c6fc3193963
Dataset
reference_data/datasets/REG_CHAID.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact30300.00E+2passed
/test_rowsexact12120.00E+2passed
/missing_targetexact000.00E+2passed
/excluded_countexact000.00E+2passed
/classessequence['no', 'yes']['no', 'yes']passed
/risk_nameexactmisclassification ratemisclassification ratepassed
/training_risktolerance00.00.00E+2passed
/test_risktolerance00.00.00E+2passed
/nodes/0/nexact30300.00E+2passed
/nodes/0/split/predictorexactshiftshiftpassed
/nodes/0/split/statistictolerance3030.00.00E+2passed
/nodes/0/split/df/0exact110.00E+2passed
/nodes/0/split/p_valuetail_probability4.3204630578274968e-084.3204630578274955e-081.30E-2315passed
/nodes/0/split/multipliertolerance11.00.00E+2passed
/nodes/0/split/adjusted_ptail_probability4.3204630578274968e-084.3204630578274955e-081.30E-2315passed
/nodes/0/split/eligibleexactTrueTruepassed
/nodes/0/split/groups/0sequence['day']['day']passed
/nodes/1/stop_reasonexactpure nodepure nodepassed
/nodes/1/parentexact000.00E+2passed
/nodes/0/split/groups/1sequence['night']['night']passed
/nodes/2/stop_reasonexactpure nodepure nodepassed
/nodes/2/parentexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

30  emit <- function(id, title, kind, configuration, outputs, notes) {
31    expected <- list(schema_version = "2.0.0", case_id = id, outputs = outputs,
32                     expected_error = NULL, deferred_outputs = list(), generator_run = run)
33    write_expected("regression", id, expected)
34    case <- list(
35      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
36      entry_point = "tabelz.analysis.chaid:compute_chaid",
37      menu_items = list("chaidMenuItem"), title = title, case_kind = kind,
38      dataset = list(path = "reference_data/datasets/REG_CHAID.json", sha256 = ""),
39      excluded_rows = list(), configuration = configuration, "repeat" = 1L,
40      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"),
41                      sha256 = ""),
42      oracle = list(oracle_class = "analytic",
43                    source = "perfect-association chi-square identity (c - 1) N, checked with stats::chisq.test; Kass (1980) Bonferroni multiplier",
44                    citation = "Kass (1980) An exploratory technique for investigating large quantities of categorical data, Applied Statistics 29(2), 119-127; R Core Team, stats package help: chisq.test.",
45                    generator = list(path = "generators/r/regression_chaid.R", sha256 = ""),
46                    independent = TRUE,
47                    convention_notes = c(notes, "Expected computed from the exported CSV.", manual)),
48      supported = TRUE, unsupported_reason = NULL,
49      requirement_ids = list("REQ-STAT-REG-CHAID-001"), risk_ids = list("RISK-REFERENCE-001"),
50      review_reference = NULL
51    )
52    write_case("regression", id, case)

Whole script generators/r/regression_chaid.R


CHAID root split on an ordered three-category predictor

Case STAT-REG-CHAID-ORDINAL-001 · status: passed

Reference

Evidence type
Analytical
Source
perfect-association chi-square identity (c - 1) N, checked with stats::chisq.test; Kass (1980) Bonferroni multiplier
Citation
Kass (1980) An exploratory technique for investigating large quantities of categorical data, Applied Statistics 29(2), 119-127; R Core Team, stats package help: chisq.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_chaid.R · sha256 eb0c2c6fc3193963
Dataset
reference_data/datasets/REG_CHAID.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact30300.00E+2passed
/test_rowsexact12120.00E+2passed
/missing_targetexact000.00E+2passed
/excluded_countexact000.00E+2passed
/classessequence['A', 'B', 'C']['A', 'B', 'C']passed
/risk_nameexactmisclassification ratemisclassification ratepassed
/training_risktolerance00.00.00E+2passed
/test_risktolerance00.00.00E+2passed
/nodes/0/nexact30300.00E+2passed
/nodes/0/split/predictorexactmachinemachinepassed
/nodes/0/split/statistictolerance6060.00.00E+2passed
/nodes/0/split/df/0exact440.00E+2passed
/nodes/0/split/p_valuetail_probability2.9008631203404549e-122.9008631203404573e-122.39E-2715passed
/nodes/0/split/multipliertolerance11.00.00E+2passed
/nodes/0/split/adjusted_ptail_probability2.9008631203404549e-122.9008631203404573e-122.39E-2715passed
/nodes/0/split/eligibleexactTrueTruepassed
/nodes/0/split/groups/0sequence['m1']['m1']passed
/nodes/1/stop_reasonexactpure nodepure nodepassed
/nodes/1/parentexact000.00E+2passed
/nodes/0/split/groups/1sequence['m2']['m2']passed
/nodes/2/stop_reasonexactpure nodepure nodepassed
/nodes/2/parentexact000.00E+2passed
/nodes/0/split/groups/2sequence['m3']['m3']passed
/nodes/3/stop_reasonexactpure nodepure nodepassed
/nodes/3/parentexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

135  emit("STAT-REG-CHAID-ORDINAL-001", "CHAID root split on an ordered three-category predictor", "axis_coverage",
136       list(target = "grade", target_kind = "categorical", predictors = list("machine"),
137            ordinal = list("machine"), min_parent = 2L, min_child = 1L, max_depth = 3L),
138       root_outputs(3L, c("A", "B", "C"), "machine", list("m1", "m2", "m3")),
139       list(invariance, "Declared ordinal: only adjacent categories may merge, and none does; the ordinal multiplier C(c - 1, r - 1) for r = c is 1.", sizes))

Whole script generators/r/regression_chaid.R


CHAID root split with Kass merge/re-split

Case STAT-REG-CHAID-RESPLIT-001 · status: passed

Reference

Evidence type
Analytical
Source
perfect-association chi-square identity (c - 1) N, checked with stats::chisq.test; Kass (1980) Bonferroni multiplier
Citation
Kass (1980) An exploratory technique for investigating large quantities of categorical data, Applied Statistics 29(2), 119-127; R Core Team, stats package help: chisq.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_chaid.R · sha256 eb0c2c6fc3193963
Dataset
reference_data/datasets/REG_CHAID.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact30300.00E+2passed
/test_rowsexact12120.00E+2passed
/missing_targetexact000.00E+2passed
/excluded_countexact000.00E+2passed
/classessequence['A', 'B', 'C']['A', 'B', 'C']passed
/risk_nameexactmisclassification ratemisclassification ratepassed
/training_risktolerance00.00.00E+2passed
/test_risktolerance00.00.00E+2passed
/nodes/0/nexact30300.00E+2passed
/nodes/0/split/predictorexactmachinemachinepassed
/nodes/0/split/statistictolerance6060.00.00E+2passed
/nodes/0/split/df/0exact440.00E+2passed
/nodes/0/split/p_valuetail_probability2.9008631203404549e-122.9008631203404573e-122.39E-2715passed
/nodes/0/split/multipliertolerance11.00.00E+2passed
/nodes/0/split/adjusted_ptail_probability2.9008631203404549e-122.9008631203404573e-122.39E-2715passed
/nodes/0/split/eligibleexactTrueTruepassed
/nodes/0/split/groups/0sequence['m1']['m1']passed
/nodes/1/stop_reasonexactpure nodepure nodepassed
/nodes/1/parentexact000.00E+2passed
/nodes/0/split/groups/1sequence['m2']['m2']passed
/nodes/2/stop_reasonexactpure nodepure nodepassed
/nodes/2/parentexact000.00E+2passed
/nodes/0/split/groups/2sequence['m3']['m3']passed
/nodes/3/stop_reasonexactpure nodepure nodepassed
/nodes/3/parentexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

141  emit("STAT-REG-CHAID-RESPLIT-001", "CHAID root split with Kass merge/re-split", "axis_coverage",
142       list(target = "grade", target_kind = "categorical", predictors = list("machine"),
143            algorithm = "kass_resplit", resplit_alpha = 0.01, min_parent = 2L, min_child = 1L,
144            max_depth = 3L),
145       root_outputs(3L, c("A", "B", "C"), "machine", list("m1", "m2", "m3")),
146       list(invariance, "Kass merge/re-split mode: no compound forms, so no re-split is attempted and the partition is the three single categories.", sizes))

Whole script generators/r/regression_chaid.R


CHAID root split on a three-category nominal predictor

Case STAT-REG-CHAID-THREE-001 · status: passed

Reference

Evidence type
Analytical
Source
perfect-association chi-square identity (c - 1) N, checked with stats::chisq.test; Kass (1980) Bonferroni multiplier
Citation
Kass (1980) An exploratory technique for investigating large quantities of categorical data, Applied Statistics 29(2), 119-127; R Core Team, stats package help: chisq.test.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_chaid.R · sha256 eb0c2c6fc3193963
Dataset
reference_data/datasets/REG_CHAID.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact30300.00E+2passed
/test_rowsexact12120.00E+2passed
/missing_targetexact000.00E+2passed
/excluded_countexact000.00E+2passed
/classessequence['A', 'B', 'C']['A', 'B', 'C']passed
/risk_nameexactmisclassification ratemisclassification ratepassed
/training_risktolerance00.00.00E+2passed
/test_risktolerance00.00.00E+2passed
/nodes/0/nexact30300.00E+2passed
/nodes/0/split/predictorexactmachinemachinepassed
/nodes/0/split/statistictolerance6060.00.00E+2passed
/nodes/0/split/df/0exact440.00E+2passed
/nodes/0/split/p_valuetail_probability2.9008631203404549e-122.9008631203404573e-122.39E-2715passed
/nodes/0/split/multipliertolerance11.00.00E+2passed
/nodes/0/split/adjusted_ptail_probability2.9008631203404549e-122.9008631203404573e-122.39E-2715passed
/nodes/0/split/eligibleexactTrueTruepassed
/nodes/0/split/groups/0sequence['m1']['m1']passed
/nodes/1/stop_reasonexactpure nodepure nodepassed
/nodes/1/parentexact000.00E+2passed
/nodes/0/split/groups/1sequence['m2']['m2']passed
/nodes/2/stop_reasonexactpure nodepure nodepassed
/nodes/2/parentexact000.00E+2passed
/nodes/0/split/groups/2sequence['m3']['m3']passed
/nodes/3/stop_reasonexactpure nodepure nodepassed
/nodes/3/parentexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

129  emit("STAT-REG-CHAID-THREE-001", "CHAID root split on a three-category nominal predictor", "axis_coverage",
130       list(target = "grade", target_kind = "categorical", predictors = list("machine"),
131            min_parent = 2L, min_child = 1L, max_depth = 3L),
132       root_outputs(3L, c("A", "B", "C"), "machine", list("m1", "m2", "m3")),
133       list(invariance, "Three categories and three classes: chi-square 2N on 4 df; the nominal Kass multiplier for c = r = 3 is 1.", sizes))

Whole script generators/r/regression_chaid.R


Curve estimation original-scale curves

Case STAT-REG-CURVE-ESTIMATION-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm
Citation
R Core Team, stats package help: lm, summary.lm, confint.lm, pf, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_curve_estimation.R · sha256 f830ec43141611cc
Dataset
reference_data/datasets/REG_CURVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[curve=linear]/r_squaredtolerance0.989424314851511720.98942431485151181.12E-1615passed
/fits/[curve=linear]/adjusted_r_squaredtolerance0.988366746336662880.9883667463366631.13E-1615passed
/fits/[curve=linear]/standard_errortolerance1.58379321011328391.5837932101132812.90E-1514passed
/fits/[curve=linear]/f_valuetolerance935.56521488629039935.56521488629383.41E-1214passed
/fits/[curve=linear]/df1exact110.00E+2passed
/fits/[curve=linear]/df2exact10100.00E+2passed
/fits/[curve=linear]/p_valuetail_probability3.270135381274213e-113.270135381274155e-115.82E-2513passed
/fits/[curve=linear]/coefficients/0/estimatetolerance-2.9151515151515044-2.9151515151515292.44E-1414passed
/fits/[curve=linear]/coefficients/0/standard_errortolerance0.97475733817893850.97475733817893731.22E-1514passed
/fits/[curve=linear]/coefficients/0/t_valuetolerance-2.9906433129271432-2.9906433129271722.89E-1414passed
/fits/[curve=linear]/coefficients/0/p_valuetail_probability0.0135587406273825670.01355874062738196.67E-1613passed
/fits/[curve=linear]/coefficients/0/lowertolerance-5.0870462116067205-5.0870462116067432.23E-1414passed
/fits/[curve=linear]/coefficients/0/uppertolerance-0.74325681869628824-0.74325681869631492.66E-1413passed
/fits/[curve=linear]/coefficients/0/parametertolerance-2.9151515151515044-2.9151515151515292.44E-1414passed
/fits/[curve=linear]/coefficients/0/parameter_lowertolerance-5.0870462116067205-5.0870462116067432.23E-1414passed
/fits/[curve=linear]/coefficients/0/parameter_uppertolerance-0.74325681869628824-0.74325681869631492.66E-1413passed
/fits/[curve=linear]/coefficients/1/estimatetolerance4.05104895104895054.0510489510489521.74E-1515passed
/fits/[curve=linear]/coefficients/1/standard_errortolerance0.132443441746071510.13244344174607131.99E-1614passed
/fits/[curve=linear]/coefficients/1/t_valuetolerance30.58701055818123430.5870105581812946.03E-1414passed
/fits/[curve=linear]/coefficients/1/p_valuetail_probability3.270135381274213e-113.27013538127415e-116.27E-2513passed
/fits/[curve=linear]/coefficients/1/lowertolerance3.75594657280374783.755946572803752.19E-1515passed
/fits/[curve=linear]/coefficients/1/uppertolerance4.34615132929415364.34615132929415458.86E-1615passed
/fits/[curve=linear]/coefficients/1/parametertolerance4.05104895104895054.0510489510489521.74E-1515passed
/fits/[curve=linear]/coefficients/1/parameter_lowertolerance3.75594657280374783.755946572803752.19E-1515passed
/fits/[curve=linear]/coefficients/1/parameter_uppertolerance4.34615132929415364.34615132929415458.86E-1615passed
/fits/[curve=quadratic]/r_squaredtolerance0.995799945160291580.99579994516029162.53E-1817passed
/fits/[curve=quadratic]/adjusted_r_squaredtolerance0.99486659964035640.99486659964035642.86E-1817passed
/fits/[curve=quadratic]/standard_errortolerance1.05208407310485371.05208407310485298.43E-1615passed
/fits/[curve=quadratic]/f_valuetolerance1066.91458188968981066.91458188969161.79E-1214passed
/fits/[curve=quadratic]/df1exact220.00E+2passed
/fits/[curve=quadratic]/df2exact990.00E+2passed
/fits/[curve=quadratic]/p_valuetail_probability2.016727983116802e-112.016727983116789e-111.29E-2514passed
/fits/[curve=quadratic]/coefficients/0/estimatetolerance0.313636363636381920.31363636363635522.67E-1413passed
/fits/[curve=quadratic]/coefficients/0/standard_errortolerance1.08735920879025061.0873592087902521.34E-1514passed
/fits/[curve=quadratic]/coefficients/0/t_valuetolerance0.288438596096795230.28843859609677032.49E-1413passed
/fits/[curve=quadratic]/coefficients/0/p_valuetail_probability0.779543790353187220.77954379035320551.83E-1413passed
/fits/[curve=quadratic]/coefficients/0/lowertolerance-2.1461410590630727-2.1461410590631022.93E-1413passed
/fits/[curve=quadratic]/coefficients/0/uppertolerance2.77341378633583612.7734137863358122.40E-1414passed
/fits/[curve=quadratic]/coefficients/0/parametertolerance0.313636363636381920.31363636363635522.67E-1413passed
/fits/[curve=quadratic]/coefficients/0/parameter_lowertolerance-2.1461410590630727-2.1461410590631022.93E-1413passed
/fits/[curve=quadratic]/coefficients/0/parameter_uppertolerance2.77341378633583612.7734137863358122.40E-1414passed
/fits/[curve=quadratic]/coefficients/1/estimatetolerance2.66728271728271342.6672827172827195.75E-1514passed
/fits/[curve=quadratic]/coefficients/1/standard_errortolerance0.38457433941344160.384574339413441665.71E-1715passed
/fits/[curve=quadratic]/coefficients/1/t_valuetolerance6.93567522302941966.9356752230294341.42E-1414passed
/fits/[curve=quadratic]/coefficients/1/p_valuetail_probability6.7917799226063782e-056.791779922606277e-051.02E-1813passed
/fits/[curve=quadratic]/coefficients/1/lowertolerance1.79731512075020851.7973151207502145.55E-1514passed
/fits/[curve=quadratic]/coefficients/1/uppertolerance3.53725031381521853.53725031381522435.76E-1514passed
/fits/[curve=quadratic]/coefficients/1/parametertolerance2.66728271728271342.6672827172827195.75E-1514passed
/fits/[curve=quadratic]/coefficients/1/parameter_lowertolerance1.79731512075020851.7973151207502145.55E-1514passed
/fits/[curve=quadratic]/coefficients/1/parameter_uppertolerance3.53725031381521853.53725031381522435.76E-1514passed
/fits/[curve=quadratic]/coefficients/2/estimatetolerance0.106443556443556670.10644355644355642.75E-1614passed
/fits/[curve=quadratic]/coefficients/2/standard_errortolerance0.0287981135027861670.0287981135027861643.42E-1815passed
/fits/[curve=quadratic]/coefficients/2/t_valuetolerance3.69619893446348293.69619893446347369.32E-1514passed
/fits/[curve=quadratic]/coefficients/2/p_valuetail_probability0.00494961532105654430.0049496153210566157.11E-1713passed
/fits/[curve=quadratic]/coefficients/2/lowertolerance0.0412976977081532390.0412976977081529752.64E-1614passed
/fits/[curve=quadratic]/coefficients/2/uppertolerance0.171589415178960090.171589415178959812.76E-1614passed
/fits/[curve=quadratic]/coefficients/2/parametertolerance0.106443556443556670.10644355644355642.75E-1614passed
/fits/[curve=quadratic]/coefficients/2/parameter_lowertolerance0.0412976977081532390.0412976977081529752.64E-1614passed
/fits/[curve=quadratic]/coefficients/2/parameter_uppertolerance0.171589415178960090.171589415178959812.76E-1614passed
/fits/[curve=cubic]/r_squaredtolerance0.999487714762808730.99948771476280861.09E-1615passed
/fits/[curve=cubic]/adjusted_r_squaredtolerance0.999295607798861950.99929560779886181.10E-1615passed
/fits/[curve=cubic]/standard_errortolerance0.389722201553096340.38972220155309932.94E-1514passed
/fits/[curve=cubic]/f_valuetolerance5202.76669949294235202.7666994928637.91E-1113passed
/fits/[curve=cubic]/df1exact330.00E+2passed
/fits/[curve=cubic]/df2exact880.00E+2passed
/fits/[curve=cubic]/p_valuetail_probability1.6945677443311567e-131.6945677443312526e-139.59E-2713passed
/fits/[curve=cubic]/coefficients/0/estimatetolerance4.06464646464650414.0646464646464544.97E-1413passed
/fits/[curve=cubic]/coefficients/0/standard_errortolerance0.637617879864753380.63761787986475754.11E-1514passed
/fits/[curve=cubic]/coefficients/0/t_valuetolerance6.3747372729081346.3747372729080151.19E-1313passed
/fits/[curve=cubic]/coefficients/0/p_valuetail_probability0.000214847191680327970.000214847191680355072.71E-1712passed
/fits/[curve=cubic]/coefficients/0/lowertolerance2.59429699699826972.59429699699821065.91E-1413passed
/fits/[curve=cubic]/coefficients/0/uppertolerance5.53499593229473865.5349959322946993.99E-1414passed
/fits/[curve=cubic]/coefficients/0/parametertolerance4.06464646464650414.0646464646464544.97E-1413passed
/fits/[curve=cubic]/coefficients/0/parameter_lowertolerance2.59429699699826972.59429699699821065.91E-1413passed
/fits/[curve=cubic]/coefficients/0/parameter_uppertolerance5.53499593229473865.5349959322946993.99E-1414passed
/fits/[curve=cubic]/coefficients/1/estimatetolerance-0.23184963184964852-0.231849631849628951.96E-1413passed
/fits/[curve=cubic]/coefficients/1/standard_errortolerance0.407726340594443180.40772634059444562.44E-1514passed
/fits/[curve=cubic]/coefficients/1/t_valuetolerance-0.56864030788794306-0.56864030788789175.14E-1413passed
/fits/[curve=cubic]/coefficients/1/p_valuetail_probability0.585205008383340890.58520500838337433.34E-1413passed
/fits/[curve=cubic]/coefficients/1/lowertolerance-1.1720682592920968-1.17206825929208281.40E-1413passed
/fits/[curve=cubic]/coefficients/1/uppertolerance0.708368995592799870.70836899559282482.50E-1413passed
/fits/[curve=cubic]/coefficients/1/parametertolerance-0.23184963184964852-0.231849631849628951.96E-1413passed
/fits/[curve=cubic]/coefficients/1/parameter_lowertolerance-1.1720682592920968-1.17206825929208281.40E-1413passed
/fits/[curve=cubic]/coefficients/1/parameter_uppertolerance0.708368995592799870.70836899559282482.50E-1413passed
/fits/[curve=cubic]/coefficients/2/estimatetolerance0.642302142302144550.64230214230214232.22E-1514passed
/fits/[curve=cubic]/coefficients/2/standard_errortolerance0.0714133725206851090.071413372520685574.58E-1614passed
/fits/[curve=cubic]/coefficients/2/t_valuetolerance8.99414380851569118.9941438085156028.89E-1414passed
/fits/[curve=cubic]/coefficients/2/p_valuetail_probability1.8620068854698013e-051.8620068854699338e-051.33E-1813passed
/fits/[curve=cubic]/coefficients/2/lowertolerance0.477622609960569160.47762260996056593.28E-1514passed
/fits/[curve=cubic]/coefficients/2/uppertolerance0.806981674643719950.80698167464371881.11E-1514passed
/fits/[curve=cubic]/coefficients/2/parametertolerance0.642302142302144550.64230214230214232.22E-1514passed
/fits/[curve=cubic]/coefficients/2/parameter_lowertolerance0.477622609960569160.47762260996056593.28E-1514passed
/fits/[curve=cubic]/coefficients/2/parameter_uppertolerance0.806981674643719950.80698167464371881.11E-1514passed
/fits/[curve=cubic]/coefficients/3/estimatetolerance-0.027479927479927571-0.0274799274799274888.31E-1714passed
/fits/[curve=cubic]/coefficients/3/standard_errortolerance0.00362113419079055720.0036211341907905832.56E-1714passed
/fits/[curve=cubic]/coefficients/3/t_valuetolerance-7.5887625346268157-7.5887625346267387.73E-1413passed
/fits/[curve=cubic]/coefficients/3/p_valuetail_probability6.3717449265666336e-056.371744926567096e-054.62E-1813passed
/fits/[curve=cubic]/coefficients/3/lowertolerance-0.035830277898019791-0.035830277898019762.76E-1715passed
/fits/[curve=cubic]/coefficients/3/uppertolerance-0.019129577061835355-0.0191295770618352121.43E-1614passed
/fits/[curve=cubic]/coefficients/3/parametertolerance-0.027479927479927571-0.0274799274799274888.31E-1714passed
/fits/[curve=cubic]/coefficients/3/parameter_lowertolerance-0.035830277898019791-0.035830277898019762.76E-1715passed
/fits/[curve=cubic]/coefficients/3/parameter_uppertolerance-0.019129577061835355-0.0191295770618352121.43E-1614passed
/fits/[curve=logarithmic]/r_squaredtolerance0.829796489718234830.8297964897182352.23E-1615passed
/fits/[curve=logarithmic]/adjusted_r_squaredtolerance0.812776138690058270.81277613869005852.22E-1615passed
/fits/[curve=logarithmic]/standard_errortolerance6.35372591911134696.3537259191113424.46E-1515passed
/fits/[curve=logarithmic]/f_valuetolerance48.75319482803497148.753194828035024.97E-1414passed
/fits/[curve=logarithmic]/df1exact110.00E+2passed
/fits/[curve=logarithmic]/df2exact10100.00E+2passed
/fits/[curve=logarithmic]/p_valuetail_probability3.7956131600958105e-053.7956131600957935e-051.70E-1914passed
/fits/[curve=logarithmic]/coefficients/0/estimatetolerance-6.0598937775669279-6.0598937775669368.02E-1514passed
/fits/[curve=logarithmic]/coefficients/0/standard_errortolerance4.60281547050979524.6028154705097932.64E-1515passed
/fits/[curve=logarithmic]/coefficients/0/t_valuetolerance-1.3165623989040236-1.3165623989040262.44E-1514passed
/fits/[curve=logarithmic]/coefficients/0/p_valuetail_probability0.217360280618592920.217360280618592286.42E-1614passed
/fits/[curve=logarithmic]/coefficients/0/lowertolerance-16.315605755933284-16.315605755933296.77E-1515passed
/fits/[curve=logarithmic]/coefficients/0/uppertolerance4.19581820079942874.1958182007994171.16E-1414passed
/fits/[curve=logarithmic]/coefficients/0/parametertolerance-6.0598937775669279-6.0598937775669368.02E-1514passed
/fits/[curve=logarithmic]/coefficients/0/parameter_lowertolerance-16.315605755933284-16.315605755933296.77E-1515passed
/fits/[curve=logarithmic]/coefficients/0/parameter_uppertolerance4.19581820079942874.1958182007994171.16E-1414passed
/fits/[curve=logarithmic]/coefficients/1/estimatetolerance17.69724967966776317.6972496796677673.74E-1515passed
/fits/[curve=logarithmic]/coefficients/1/standard_errortolerance2.53456969497168012.53456969497167871.38E-1515passed
/fits/[curve=logarithmic]/coefficients/1/t_valuetolerance6.98234880452380056.9823488045238054.41E-1515passed
/fits/[curve=logarithmic]/coefficients/1/p_valuetail_probability3.7956131600958179e-053.7956131600957955e-052.24E-1914passed
/fits/[curve=logarithmic]/coefficients/1/lowertolerance12.04987646923436312.049876469234376.84E-1515passed
/fits/[curve=logarithmic]/coefficients/1/uppertolerance23.34462289010116423.3446228901011643.52E-1616passed
/fits/[curve=logarithmic]/coefficients/1/parametertolerance17.69724967966776317.6972496796677673.74E-1515passed
/fits/[curve=logarithmic]/coefficients/1/parameter_lowertolerance12.04987646923436312.049876469234376.84E-1515passed
/fits/[curve=logarithmic]/coefficients/1/parameter_uppertolerance23.34462289010116423.3446228901011643.52E-1616passed
/fits/[curve=inverse]/r_squaredtolerance0.521089677415634060.52108967741563421.07E-1615passed
/fits/[curve=inverse]/adjusted_r_squaredtolerance0.473198645157197450.473198645157197561.12E-1615passed
/fits/[curve=inverse]/standard_errortolerance10.65789210565209510.6578921056520923.11E-1515passed
/fits/[curve=inverse]/f_valuetolerance10.88073597168784610.880735971687853.78E-1515passed
/fits/[curve=inverse]/df1exact110.00E+2passed
/fits/[curve=inverse]/df2exact10100.00E+2passed
/fits/[curve=inverse]/p_valuetail_probability0.00803252279346353990.0080325227934635336.90E-1815passed
/fits/[curve=inverse]/coefficients/0/estimatetolerance33.82821256628533733.8282125662853457.56E-1515passed
/fits/[curve=inverse]/coefficients/0/standard_errortolerance4.40777455533594154.40777455533594153.54E-1717passed
/fits/[curve=inverse]/coefficients/0/t_valuetolerance7.67466941459919927.6746694145992011.73E-1515passed
/fits/[curve=inverse]/coefficients/0/p_valuetail_probability1.6906972774617145e-051.6906972774617098e-054.70E-2014passed
/fits/[curve=inverse]/coefficients/0/lowertolerance24.00707882874480224.007078828744817.58E-1515passed
/fits/[curve=inverse]/coefficients/0/uppertolerance43.64934630382587243.649346303825887.55E-1515passed
/fits/[curve=inverse]/coefficients/0/parametertolerance33.82821256628533733.8282125662853457.56E-1515passed
/fits/[curve=inverse]/coefficients/0/parameter_lowertolerance24.00707882874480224.007078828744817.58E-1515passed
/fits/[curve=inverse]/coefficients/0/parameter_uppertolerance43.64934630382587243.649346303825887.55E-1515passed
/fits/[curve=inverse]/coefficients/1/estimatetolerance-40.26105983479794-40.2610598347979551.47E-1415passed
/fits/[curve=inverse]/coefficients/1/standard_errortolerance12.20551382887811812.2055138288781143.58E-1515passed
/fits/[curve=inverse]/coefficients/1/t_valuetolerance-3.298596060703376-3.2985960607033782.19E-1515passed
/fits/[curve=inverse]/coefficients/1/p_valuetail_probability0.00803252279346357980.0080325227934635423.81E-1714passed
/fits/[curve=inverse]/coefficients/1/lowertolerance-67.456639405377018-67.456639405377031.44E-1415passed
/fits/[curve=inverse]/coefficients/1/uppertolerance-13.065480264218859-13.0654802642188771.80E-1414passed
/fits/[curve=inverse]/coefficients/1/parametertolerance-40.26105983479794-40.2610598347979551.47E-1415passed
/fits/[curve=inverse]/coefficients/1/parameter_lowertolerance-67.456639405377018-67.456639405377031.44E-1415passed
/fits/[curve=inverse]/coefficients/1/parameter_uppertolerance-13.065480264218859-13.0654802642188771.80E-1414passed
/nexact12120.00E+2passed
/missing_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

37  emit <- function(id, title, kind, config, outputs, notes) {
38    write_expected("regression", id, list(
39      schema_version = "2.0.0", case_id = id, outputs = outputs, expected_error = NULL,
40      deferred_outputs = list(), generator_run = run))
41    write_case("regression", id, list(
42      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
43      entry_point = entry, menu_items = list(menu), title = title, case_kind = kind,
44      dataset = list(path = "reference_data/datasets/REG_CURVE.json", sha256 = ""),
45      excluded_rows = list(), configuration = config, "repeat" = 1L,
46      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"), sha256 = ""),
47      oracle = list(oracle_class = "independent_implementation", source = "R 4.6.1 stats::lm",
48                    citation = "R Core Team, stats package help: lm, summary.lm, confint.lm, pf, pt.",
49                    generator = list(path = "generators/r/regression_curve_estimation.R", sha256 = ""),
50                    independent = TRUE,
51                    convention_notes = c(notes, notes_common, "Expected computed from the exported CSV.")),
52      supported = TRUE, unsupported_reason = NULL,
53      requirement_ids = list("REQ-STAT-REG-CURVE-ESTIMATION-001"),
54      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/regression_curve_estimation.R


Curve estimation log-scale curves 90 percent

Case STAT-REG-CURVE-ESTIMATION-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm
Citation
R Core Team, stats package help: lm, summary.lm, confint.lm, pf, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_curve_estimation.R · sha256 f830ec43141611cc
Dataset
reference_data/datasets/REG_CURVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[curve=power]/r_squaredtolerance0.975896437131221650.97589643713122181.11E-1615passed
/fits/[curve=power]/adjusted_r_squaredtolerance0.973486080844343850.9734860808443441.09E-1615passed
/fits/[curve=power]/standard_errortolerance0.129605968732160430.12960596873216022.21E-1614passed
/fits/[curve=power]/f_valuetolerance404.87642530030951404.8764253003111.47E-1214passed
/fits/[curve=power]/df1exact110.00E+2passed
/fits/[curve=power]/df2exact10100.00E+2passed
/fits/[curve=power]/p_valuetail_probability2.022622486831102e-092.0226224868310693e-093.27E-2313passed
/fits/[curve=power]/coefficients/0/estimatetolerance1.17751462804494871.1775146280449471.59E-1514passed
/fits/[curve=power]/coefficients/0/standard_errortolerance0.0938901623308033150.093890162330803131.81E-1614passed
/fits/[curve=power]/coefficients/0/t_valuetolerance12.54140581732311912.5414058173231266.76E-1515passed
/fits/[curve=power]/coefficients/0/p_valuetail_probability1.9270809113056189e-071.9270809113056083e-071.06E-2114passed
/fits/[curve=power]/coefficients/0/lowertolerance1.00734235900589031.00734235900588921.10E-1514passed
/fits/[curve=power]/coefficients/0/uppertolerance1.3476868970840071.3476868970840051.98E-1514passed
/fits/[curve=power]/coefficients/0/parametertolerance3.24629591544744093.2462959154474364.89E-1514passed
/fits/[curve=power]/coefficients/0/parameter_lowertolerance2.73831388078838512.7383138807883823.10E-1514passed
/fits/[curve=power]/coefficients/0/parameter_uppertolerance3.84851321997339023.84851321997338277.51E-1514passed
/fits/[curve=power]/coefficients/1/estimatetolerance1.04030816851574741.0403081685157486.33E-1615passed
/fits/[curve=power]/coefficients/1/standard_errortolerance0.0517012167062323380.051701216706232259.05E-1714passed
/fits/[curve=power]/coefficients/1/t_valuetolerance20.12154132516465720.1215413251647064.94E-1414passed
/fits/[curve=power]/coefficients/1/p_valuetail_probability2.0226224868311165e-092.0226224868310693e-094.72E-2313passed
/fits/[curve=power]/coefficients/1/lowertolerance0.946601723233639670.94660172323364068.87E-1615passed
/fits/[curve=power]/coefficients/1/uppertolerance1.13401461379785511.13401461379785554.09E-1615passed
/fits/[curve=power]/coefficients/1/parametertolerance1.04030816851574741.0403081685157486.33E-1615passed
/fits/[curve=power]/coefficients/1/parameter_lowertolerance0.946601723233639670.94660172323364068.87E-1615passed
/fits/[curve=power]/coefficients/1/parameter_uppertolerance1.13401461379785511.13401461379785554.09E-1615passed
/fits/[curve=compound]/r_squaredtolerance0.956279218980532250.95627921898053211.08E-1615passed
/fits/[curve=compound]/adjusted_r_squaredtolerance0.951907140878585520.95190714087858532.19E-1615passed
/fits/[curve=compound]/standard_errortolerance0.174553409823995030.174553409823995282.51E-1614passed
/fits/[curve=compound]/f_valuetolerance218.72418485724816218.72418485724756.52E-1314passed
/fits/[curve=compound]/df1exact110.00E+2passed
/fits/[curve=compound]/df2exact10100.00E+2passed
/fits/[curve=compound]/p_valuetail_probability4.00503223571717e-084.005032235717225e-085.49E-2213passed
/fits/[curve=compound]/coefficients/0/estimatetolerance1.50704508555536741.50704508555536631.09E-1515passed
/fits/[curve=compound]/coefficients/0/standard_errortolerance0.107430197353810260.107430197353810472.12E-1614passed
/fits/[curve=compound]/coefficients/0/t_valuetolerance14.02813289630354114.0281328963035043.73E-1414passed
/fits/[curve=compound]/coefficients/0/p_valuetail_probability6.6436794777500471e-086.643679477750222e-081.75E-2113passed
/fits/[curve=compound]/coefficients/0/lowertolerance1.31233202943560051.31233202943559911.38E-1514passed
/fits/[curve=compound]/coefficients/0/uppertolerance1.70175814167513441.70175814167513358.93E-1615passed
/fits/[curve=compound]/coefficients/0/parametertolerance4.51337443527074374.5133744352707385.36E-1514passed
/fits/[curve=compound]/coefficients/0/parameter_lowertolerance3.71482670399478783.7148267039947834.93E-1514passed
/fits/[curve=compound]/coefficients/0/parameter_uppertolerance5.48357983188011575.4835798318801114.48E-1515passed
/fits/[curve=compound]/coefficients/1/estimatetolerance0.215878166866517810.21587816686651798.54E-1715passed
/fits/[curve=compound]/coefficients/1/standard_errortolerance0.0145968894284809070.0145968894284809312.43E-1714passed
/fits/[curve=compound]/coefficients/1/t_valuetolerance14.78932672089057314.7893267208905551.78E-1414passed
/fits/[curve=compound]/coefficients/1/p_valuetail_probability4.0050322357171839e-084.005032235717234e-085.02E-2213passed
/fits/[curve=compound]/coefficients/1/lowertolerance0.189421872263415420.189421872263415485.50E-1715passed
/fits/[curve=compound]/coefficients/1/uppertolerance0.24233446146962020.242334461469620321.16E-1615passed
/fits/[curve=compound]/coefficients/1/parametertolerance1.2409511808195271.2409511808195274.44E-1716passed
/fits/[curve=compound]/coefficients/1/parameter_lowertolerance1.20855069897021621.20855069897021641.86E-1615passed
/fits/[curve=compound]/coefficients/1/parameter_uppertolerance1.27422029914802071.27422029914802077.49E-1817passed
/fits/[curve=s]/r_squaredtolerance0.762764659782079190.76276465978207942.23E-1615passed
/fits/[curve=s]/adjusted_r_squaredtolerance0.739041125760287130.73904112576028742.24E-1615passed
/fits/[curve=s]/standard_errortolerance0.406606272276772820.406606272276772542.80E-1615passed
/fits/[curve=s]/f_valuetolerance32.15223579595751632.152235795957553.55E-1414passed
/fits/[curve=s]/df1exact110.00E+2passed
/fits/[curve=s]/df2exact10100.00E+2passed
/fits/[curve=s]/p_valuetail_probability0.000206585707491088350.000206585707491087627.29E-1914passed
/fits/[curve=s]/coefficients/0/estimatetolerance3.59305460575844913.59305460575844964.93E-1615passed
/fits/[curve=s]/coefficients/0/standard_errortolerance0.168159778989609210.168159778989609138.34E-1715passed
/fits/[curve=s]/coefficients/0/t_valuetolerance21.36690846852545221.3669084685254621.04E-1415passed
/fits/[curve=s]/coefficients/0/p_valuetail_probability1.1234381231613971e-091.1234381231613934e-093.68E-2414passed
/fits/[curve=s]/coefficients/0/lowertolerance3.28827154391917853.28827154391917949.18E-1615passed
/fits/[curve=s]/coefficients/0/uppertolerance3.89783766759771933.89783766759771984.69E-1615passed
/fits/[curve=s]/coefficients/0/parametertolerance3.59305460575844913.59305460575844964.93E-1615passed
/fits/[curve=s]/coefficients/0/parameter_lowertolerance3.28827154391917853.28827154391917949.18E-1615passed
/fits/[curve=s]/coefficients/0/parameter_uppertolerance3.89783766759771933.89783766759771984.69E-1615passed
/fits/[curve=s]/coefficients/1/estimatetolerance-2.6403676954260353-2.6403676954260369.21E-1615passed
/fits/[curve=s]/coefficients/1/standard_errortolerance0.465649157449326940.465649157449326672.73E-1615passed
/fits/[curve=s]/coefficients/1/t_valuetolerance-5.6702941542707856-5.6702941542707915.38E-1515passed
/fits/[curve=s]/coefficients/1/p_valuetail_probability0.000206585707491090170.00020658570749108881.38E-1814passed
/fits/[curve=s]/coefficients/1/lowertolerance-3.4843386901729536-3.48433869017295361.20E-1717passed
/fits/[curve=s]/coefficients/1/uppertolerance-1.7963967006791175-1.79639670067911881.33E-1515passed
/fits/[curve=s]/coefficients/1/parametertolerance-2.6403676954260353-2.6403676954260369.21E-1615passed
/fits/[curve=s]/coefficients/1/parameter_lowertolerance-3.4843386901729536-3.48433869017295361.20E-1717passed
/fits/[curve=s]/coefficients/1/parameter_uppertolerance-1.7963967006791175-1.79639670067911881.33E-1515passed
/fits/[curve=growth]/r_squaredtolerance0.956279218980532250.95627921898053211.08E-1615passed
/fits/[curve=growth]/adjusted_r_squaredtolerance0.951907140878585520.95190714087858532.19E-1615passed
/fits/[curve=growth]/standard_errortolerance0.174553409823995030.174553409823995282.51E-1614passed
/fits/[curve=growth]/f_valuetolerance218.72418485724816218.72418485724756.52E-1314passed
/fits/[curve=growth]/df1exact110.00E+2passed
/fits/[curve=growth]/df2exact10100.00E+2passed
/fits/[curve=growth]/p_valuetail_probability4.00503223571717e-084.005032235717225e-085.49E-2213passed
/fits/[curve=growth]/coefficients/0/estimatetolerance1.50704508555536741.50704508555536631.09E-1515passed
/fits/[curve=growth]/coefficients/0/standard_errortolerance0.107430197353810260.107430197353810472.12E-1614passed
/fits/[curve=growth]/coefficients/0/t_valuetolerance14.02813289630354114.0281328963035043.73E-1414passed
/fits/[curve=growth]/coefficients/0/p_valuetail_probability6.6436794777500471e-086.643679477750222e-081.75E-2113passed
/fits/[curve=growth]/coefficients/0/lowertolerance1.31233202943560051.31233202943559911.38E-1514passed
/fits/[curve=growth]/coefficients/0/uppertolerance1.70175814167513441.70175814167513358.93E-1615passed
/fits/[curve=growth]/coefficients/0/parametertolerance1.50704508555536741.50704508555536631.09E-1515passed
/fits/[curve=growth]/coefficients/0/parameter_lowertolerance1.31233202943560051.31233202943559911.38E-1514passed
/fits/[curve=growth]/coefficients/0/parameter_uppertolerance1.70175814167513441.70175814167513358.93E-1615passed
/fits/[curve=growth]/coefficients/1/estimatetolerance0.215878166866517810.21587816686651798.54E-1715passed
/fits/[curve=growth]/coefficients/1/standard_errortolerance0.0145968894284809070.0145968894284809312.43E-1714passed
/fits/[curve=growth]/coefficients/1/t_valuetolerance14.78932672089057314.7893267208905551.78E-1414passed
/fits/[curve=growth]/coefficients/1/p_valuetail_probability4.0050322357171839e-084.005032235717234e-085.02E-2213passed
/fits/[curve=growth]/coefficients/1/lowertolerance0.189421872263415420.189421872263415485.50E-1715passed
/fits/[curve=growth]/coefficients/1/uppertolerance0.24233446146962020.242334461469620321.16E-1615passed
/fits/[curve=growth]/coefficients/1/parametertolerance0.215878166866517810.21587816686651798.54E-1715passed
/fits/[curve=growth]/coefficients/1/parameter_lowertolerance0.189421872263415420.189421872263415485.50E-1715passed
/fits/[curve=growth]/coefficients/1/parameter_uppertolerance0.24233446146962020.242334461469620321.16E-1615passed
/fits/[curve=exponential]/r_squaredtolerance0.956279218980532250.95627921898053211.08E-1615passed
/fits/[curve=exponential]/adjusted_r_squaredtolerance0.951907140878585520.95190714087858532.19E-1615passed
/fits/[curve=exponential]/standard_errortolerance0.174553409823995030.174553409823995282.51E-1614passed
/fits/[curve=exponential]/f_valuetolerance218.72418485724816218.72418485724756.52E-1314passed
/fits/[curve=exponential]/df1exact110.00E+2passed
/fits/[curve=exponential]/df2exact10100.00E+2passed
/fits/[curve=exponential]/p_valuetail_probability4.00503223571717e-084.005032235717225e-085.49E-2213passed
/fits/[curve=exponential]/coefficients/0/estimatetolerance1.50704508555536741.50704508555536631.09E-1515passed
/fits/[curve=exponential]/coefficients/0/standard_errortolerance0.107430197353810260.107430197353810472.12E-1614passed
/fits/[curve=exponential]/coefficients/0/t_valuetolerance14.02813289630354114.0281328963035043.73E-1414passed
/fits/[curve=exponential]/coefficients/0/p_valuetail_probability6.6436794777500471e-086.643679477750222e-081.75E-2113passed
/fits/[curve=exponential]/coefficients/0/lowertolerance1.31233202943560051.31233202943559911.38E-1514passed
/fits/[curve=exponential]/coefficients/0/uppertolerance1.70175814167513441.70175814167513358.93E-1615passed
/fits/[curve=exponential]/coefficients/0/parametertolerance4.51337443527074374.5133744352707385.36E-1514passed
/fits/[curve=exponential]/coefficients/0/parameter_lowertolerance3.71482670399478783.7148267039947834.93E-1514passed
/fits/[curve=exponential]/coefficients/0/parameter_uppertolerance5.48357983188011575.4835798318801114.48E-1515passed
/fits/[curve=exponential]/coefficients/1/estimatetolerance0.215878166866517810.21587816686651798.54E-1715passed
/fits/[curve=exponential]/coefficients/1/standard_errortolerance0.0145968894284809070.0145968894284809312.43E-1714passed
/fits/[curve=exponential]/coefficients/1/t_valuetolerance14.78932672089057314.7893267208905551.78E-1414passed
/fits/[curve=exponential]/coefficients/1/p_valuetail_probability4.0050322357171839e-084.005032235717234e-085.02E-2213passed
/fits/[curve=exponential]/coefficients/1/lowertolerance0.189421872263415420.189421872263415485.50E-1715passed
/fits/[curve=exponential]/coefficients/1/uppertolerance0.24233446146962020.242334461469620321.16E-1615passed
/fits/[curve=exponential]/coefficients/1/parametertolerance0.215878166866517810.21587816686651798.54E-1715passed
/fits/[curve=exponential]/coefficients/1/parameter_lowertolerance0.189421872263415420.189421872263415485.50E-1715passed
/fits/[curve=exponential]/coefficients/1/parameter_uppertolerance0.24233446146962020.242334461469620321.16E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

123  curves <- c("power", "compound", "s", "growth", "exponential")
124  outputs <- do.call(c, lapply(curves, function(curve) curve_outputs(curve, d$t, d$y, level = 0.90)))
125  emit("STAT-REG-CURVE-ESTIMATION-002", "Curve estimation log-scale curves 90 percent", "axis_coverage",
126       list(response = "y", predictor = "t", curves = as.list(curves), confidence_level = 0.90),
127       outputs, "Curves fitted to ln Y: power (on ln t), compound, growth and exponential (on t), S (on 1 / t); confint(level = 0.90).")

Whole script generators/r/regression_curve_estimation.R


Curve estimation logistic with upper bound

Case STAT-REG-CURVE-ESTIMATION-003 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm
Citation
R Core Team, stats package help: lm, summary.lm, confint.lm, pf, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_curve_estimation.R · sha256 f830ec43141611cc
Dataset
reference_data/datasets/REG_CURVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[curve=logistic]/r_squaredtolerance0.992559996966713820.99255999696671384.35E-2019passed
/fits/[curve=logistic]/adjusted_r_squaredtolerance0.991815996663385160.99181599666338522.36E-1817passed
/fits/[curve=logistic]/standard_errortolerance0.13766453650179350.137664536501793445.59E-1715passed
/fits/[curve=logistic]/f_valuetolerance1334.0854735220561334.08547352205761.55E-1214passed
/fits/[curve=logistic]/df1exact110.00E+2passed
/fits/[curve=logistic]/df2exact10100.00E+2passed
/fits/[curve=logistic]/p_valuetail_probability5.6275172180001754e-125.627517218000141e-123.48E-2614passed
/fits/[curve=logistic]/coefficients/0/estimatetolerance-1.0278686058134641-1.02786860581346142.70E-1514passed
/fits/[curve=logistic]/coefficients/0/standard_errortolerance0.0847266652649226880.084726665264922694.30E-1917passed
/fits/[curve=logistic]/coefficients/0/t_valuetolerance-12.131583399388285-12.1315833993882533.19E-1414passed
/fits/[curve=logistic]/coefficients/0/p_valuetail_probability2.6364696590170015e-072.636469659017064e-076.24E-2113passed
/fits/[curve=logistic]/coefficients/0/lowertolerance-1.2166513804894743-1.21665138048947162.67E-1514passed
/fits/[curve=logistic]/coefficients/0/uppertolerance-0.83908583113745394-0.83908583113745122.77E-1514passed
/fits/[curve=logistic]/coefficients/0/parametertolerance0.357768694618850190.35776869461885121.00E-1514passed
/fits/[curve=logistic]/coefficients/0/parameter_lowertolerance0.296220437515437250.29622043751543818.34E-1614passed
/fits/[curve=logistic]/coefficients/0/parameter_uppertolerance0.432105360193472680.43210536019347391.22E-1514passed
/fits/[curve=logistic]/coefficients/1/estimatetolerance-0.42048052501100713-0.42048052501100742.81E-1615passed
/fits/[curve=logistic]/coefficients/1/standard_errortolerance0.0115120868710954330.0115120868710954285.07E-1815passed
/fits/[curve=logistic]/coefficients/1/t_valuetolerance-36.52513481867048-36.525134818670524.29E-1414passed
/fits/[curve=logistic]/coefficients/1/p_valuetail_probability5.6275172180001947e-125.627517218000137e-125.73E-2613passed
/fits/[curve=logistic]/coefficients/1/lowertolerance-0.44613105303593598-0.446131053035936252.74E-1615passed
/fits/[curve=logistic]/coefficients/1/uppertolerance-0.39482999698607829-0.394829996986078572.79E-1615passed
/fits/[curve=logistic]/coefficients/1/parametertolerance0.65673116823004740.65673116823004722.20E-1615passed
/fits/[curve=logistic]/coefficients/1/parameter_lowertolerance0.640099879532066930.64009987953206672.23E-1615passed
/fits/[curve=logistic]/coefficients/1/parameter_uppertolerance0.67379457662153230.67379457662153212.18E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

130  outputs <- curve_outputs("logistic", d$t, d$y, upper = 50)
131  emit("STAT-REG-CURVE-ESTIMATION-003", "Curve estimation logistic with upper bound", "axis_coverage",
132       list(response = "y", predictor = "t", curves = list("logistic"), upper_bound = 50, confidence_level = 0.95),
133       outputs, "Logistic: lm of ln(1 / Y - 1 / U) on t with U = 50; b0 and b1 are exp of the fitted coefficients.")

Whole script generators/r/regression_curve_estimation.R


Curve estimation sequence with input positions

Case STAT-REG-CURVE-ESTIMATION-004 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm
Citation
R Core Team, stats package help: lm, summary.lm, confint.lm, pf, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_curve_estimation.R · sha256 f830ec43141611cc
Dataset
reference_data/datasets/REG_CURVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[curve=linear]/r_squaredtolerance0.991772790577932330.99177279057793231.70E-1817passed
/fits/[curve=linear]/adjusted_r_squaredtolerance0.990858656197702640.99085865619770263.65E-1817passed
/fits/[curve=linear]/standard_errortolerance1.46615605857984991.46615605857985056.41E-1615passed
/fits/[curve=linear]/f_valuetolerance1084.93106924681431084.93106924681321.11E-1214passed
/fits/[curve=linear]/df1exact110.00E+2passed
/fits/[curve=linear]/df2exact990.00E+2passed
/fits/[curve=linear]/p_valuetail_probability1.0786558825635552e-101.0786558825635579e-102.73E-2514passed
/fits/[curve=linear]/coefficients/0/estimatetolerance-2.6495541401273908-2.64955414012739887.96E-1514passed
/fits/[curve=linear]/coefficients/0/standard_errortolerance0.916884347359569920.91688434735957055.53E-1615passed
/fits/[curve=linear]/coefficients/0/t_valuetolerance-2.8897364730432336-2.88973647304324077.06E-1514passed
/fits/[curve=linear]/coefficients/0/p_valuetail_probability0.0178919478734949310.0178919478734947471.84E-1613passed
/fits/[curve=linear]/coefficients/0/lowertolerance-4.7236906339643996-4.72369063396440858.87E-1514passed
/fits/[curve=linear]/coefficients/0/uppertolerance-0.57541764629038195-0.5754176462903897.10E-1513passed
/fits/[curve=linear]/coefficients/0/parametertolerance-2.6495541401273908-2.64955414012739887.96E-1514passed
/fits/[curve=linear]/coefficients/0/parameter_lowertolerance-4.7236906339643996-4.72369063396440858.87E-1514passed
/fits/[curve=linear]/coefficients/0/parameter_uppertolerance-0.57541764629038195-0.5754176462903897.10E-1513passed
/fits/[curve=linear]/coefficients/1/estimatetolerance4.04229299363057324.0422929936305751.73E-1515passed
/fits/[curve=linear]/coefficients/1/standard_errortolerance0.122723212649712270.122723212649712379.54E-1715passed
/fits/[curve=linear]/coefficients/1/t_valuetolerance32.93829183863082532.938291838630827.33E-1515passed
/fits/[curve=linear]/coefficients/1/p_valuetail_probability1.0786558825635552e-101.078655882563556e-107.94E-2615passed
/fits/[curve=linear]/coefficients/1/lowertolerance3.76467379909341913.7646737990934211.83E-1515passed
/fits/[curve=linear]/coefficients/1/uppertolerance4.31991218816772724.3199121881677291.73E-1515passed
/fits/[curve=linear]/coefficients/1/parametertolerance4.04229299363057324.0422929936305751.73E-1515passed
/fits/[curve=linear]/coefficients/1/parameter_lowertolerance3.76467379909341913.7646737990934211.83E-1515passed
/fits/[curve=linear]/coefficients/1/parameter_uppertolerance4.31991218816772724.3199121881677291.73E-1515passed
/fits/[curve=quadratic]/r_squaredtolerance0.9963395986012330.9963395986012337.85E-1918passed
/fits/[curve=quadratic]/adjusted_r_squaredtolerance0.995424498251541220.99542449825154123.23E-1817passed
/fits/[curve=quadratic]/standard_errortolerance1.03727650445800081.0372765044580032.21E-1514passed
/fits/[curve=quadratic]/f_valuetolerance1088.77632812276241088.77632812275784.57E-1214passed
/fits/[curve=quadratic]/df1exact220.00E+2passed
/fits/[curve=quadratic]/df2exact880.00E+2passed
/fits/[curve=quadratic]/p_valuetail_probability1.7952083125882702e-101.7952083125882969e-102.67E-2413passed
/fits/[curve=quadratic]/coefficients/0/estimatetolerance0.108289345063532020.108289345063523438.59E-1513passed
/fits/[curve=quadratic]/coefficients/0/standard_errortolerance1.08756635420516881.08756635420517374.87E-1514passed
/fits/[curve=quadratic]/coefficients/0/t_valuetolerance0.0995703339339452380.09957033393393698.34E-1513passed
/fits/[curve=quadratic]/coefficients/0/p_valuetail_probability0.923135357494086260.92313535749409276.44E-1514passed
/fits/[curve=quadratic]/coefficients/0/lowertolerance-2.3996431650425061-2.3996431650425251.91E-1414passed
/fits/[curve=quadratic]/coefficients/0/uppertolerance2.61622185516957022.61622185516957242.20E-1515passed
/fits/[curve=quadratic]/coefficients/0/parametertolerance0.108289345063532020.108289345063523438.59E-1513passed
/fits/[curve=quadratic]/coefficients/0/parameter_lowertolerance-2.3996431650425061-2.3996431650425251.91E-1414passed
/fits/[curve=quadratic]/coefficients/0/parameter_uppertolerance2.61622185516957022.61622185516957242.20E-1515passed
/fits/[curve=quadratic]/coefficients/1/estimatetolerance2.80944603783313562.8094460378331393.58E-1514passed
/fits/[curve=quadratic]/coefficients/1/standard_errortolerance0.399773629207999970.39977362920800141.44E-1514passed
/fits/[curve=quadratic]/coefficients/1/t_valuetolerance7.02759219861247167.0275921986124561.59E-1414passed
/fits/[curve=quadratic]/coefficients/1/p_valuetail_probability0.00010958321831702040.00010958321831702221.80E-1813passed
/fits/[curve=quadratic]/coefficients/1/lowertolerance1.88756639573391061.8875663957339114.51E-1615passed
/fits/[curve=quadratic]/coefficients/1/uppertolerance3.73132567993236063.73132567993236736.70E-1514passed
/fits/[curve=quadratic]/coefficients/1/parametertolerance2.80944603783313562.8094460378331393.58E-1514passed
/fits/[curve=quadratic]/coefficients/1/parameter_lowertolerance1.88756639573391061.8875663957339114.51E-1615passed
/fits/[curve=quadratic]/coefficients/1/parameter_uppertolerance3.73132567993236063.73132567993236736.70E-1514passed
/fits/[curve=quadratic]/coefficients/2/estimatetolerance0.0951607532252693040.095160753225269111.94E-1614passed
/fits/[curve=quadratic]/coefficients/2/standard_errortolerance0.030121101952452690.0301211019524527849.41E-1714passed
/fits/[curve=quadratic]/coefficients/2/t_valuetolerance3.1592719740295093.15927197402949261.64E-1414passed
/fits/[curve=quadratic]/coefficients/2/p_valuetail_probability0.0134096753098652680.0134096753098656033.35E-1613passed
/fits/[curve=quadratic]/coefficients/2/lowertolerance0.0257013675660071190.0257013675660067164.03E-1613passed
/fits/[curve=quadratic]/coefficients/2/uppertolerance0.164620138884531480.16462013888453152.28E-1715passed
/fits/[curve=quadratic]/coefficients/2/parametertolerance0.0951607532252693040.095160753225269111.94E-1614passed
/fits/[curve=quadratic]/coefficients/2/parameter_lowertolerance0.0257013675660071190.0257013675660067164.03E-1613passed
/fits/[curve=quadratic]/coefficients/2/parameter_uppertolerance0.164620138884531480.16462013888453152.28E-1715passed
/nexact11110.00E+2passed
/missing_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

137  keep <- is.finite(d$y_gap)
138  outputs <- curve_outputs("linear", d$t[keep], d$y_gap[keep])
139  outputs <- c(outputs, curve_outputs("quadratic", d$t[keep], d$y_gap[keep]))
140  outputs[["/n"]] <- exact(sum(keep), "complete rows")
141  outputs[["/missing_count"]] <- exact(sum(!keep), "rows with a missing response")
142  emit("STAT-REG-CURVE-ESTIMATION-004", "Curve estimation sequence with input positions", "missing_excluded",
143       list(response = "y_gap", predictor = "", sequence = "input_position", curves = list("linear", "quadratic"), confidence_level = 0.95),
144       outputs, "Blank predictor with input positions: t is the row's position 1..N in the table, so the missing row leaves the gap t = 6 (the t column of REG_CURVE equals the positions).")

Whole script generators/r/regression_curve_estimation.R


Curve estimation sequence with retained order

Case STAT-REG-CURVE-ESTIMATION-005 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm
Citation
R Core Team, stats package help: lm, summary.lm, confint.lm, pf, pt.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_curve_estimation.R · sha256 f830ec43141611cc
Dataset
reference_data/datasets/REG_CURVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[curve=linear]/r_squaredtolerance0.98318421050596860.98318421050596871.14E-1615passed
/fits/[curve=linear]/adjusted_r_squaredtolerance0.981315789451076230.98131578945107631.07E-1615passed
/fits/[curve=linear]/standard_errortolerance2.09610268466181492.09610268466181361.30E-1515passed
/fits/[curve=linear]/f_valuetolerance526.21126695802673526.21126695802735.65E-1314passed
/fits/[curve=linear]/df1exact110.00E+2passed
/fits/[curve=linear]/df2exact990.00E+2passed
/fits/[curve=linear]/p_valuetail_probability2.7009505595846655e-092.7009505595846506e-091.49E-2314passed
/fits/[curve=linear]/coefficients/0/estimatetolerance-3.6981818181818049-3.6981818181818221.69E-1414passed
/fits/[curve=linear]/coefficients/0/standard_errortolerance1.3554862843400431.35548628434004264.09E-1615passed
/fits/[curve=linear]/coefficients/0/t_valuetolerance-2.728306336188691-2.72830633618870431.33E-1414passed
/fits/[curve=linear]/coefficients/0/p_valuetail_probability0.0232890077927818680.023289007792781394.78E-1613passed
/fits/[curve=linear]/coefficients/0/lowertolerance-6.764504825376358-6.7645048253763741.60E-1414passed
/fits/[curve=linear]/coefficients/0/uppertolerance-0.63185881098725183-0.631858810987271.82E-1413passed
/fits/[curve=linear]/coefficients/0/parametertolerance-3.6981818181818049-3.6981818181818221.69E-1414passed
/fits/[curve=linear]/coefficients/0/parameter_lowertolerance-6.764504825376358-6.7645048253763741.60E-1414passed
/fits/[curve=linear]/coefficients/0/parameter_uppertolerance-0.63185881098725183-0.631858810987271.82E-1413passed
/fits/[curve=linear]/coefficients/1/estimatetolerance4.58454545454545314.5845454545454562.71E-1515passed
/fits/[curve=linear]/coefficients/1/standard_errortolerance0.199855549304229080.199855549304229025.57E-1715passed
/fits/[curve=linear]/coefficients/1/t_valuetolerance22.93929525853021222.9392952585302332.12E-1415passed
/fits/[curve=linear]/coefficients/1/p_valuetail_probability2.7009505595846655e-092.7009505595846386e-092.69E-2314passed
/fits/[curve=linear]/coefficients/1/lowertolerance4.13244079216192174.1324407921619242.69E-1515passed
/fits/[curve=linear]/coefficients/1/uppertolerance5.03665011692898465.0366501169289872.63E-1515passed
/fits/[curve=linear]/coefficients/1/parametertolerance4.58454545454545314.5845454545454562.71E-1515passed
/fits/[curve=linear]/coefficients/1/parameter_lowertolerance4.13244079216192174.1324407921619242.69E-1515passed
/fits/[curve=linear]/coefficients/1/parameter_uppertolerance5.03665011692898465.0366501169289872.63E-1515passed
/fits/[curve=quadratic]/r_squaredtolerance0.986024444462551660.98602444446255173.55E-1817passed
/fits/[curve=quadratic]/adjusted_r_squaredtolerance0.982530555578189580.98253055557818965.65E-1918passed
/fits/[curve=quadratic]/standard_errortolerance2.02681786946698232.02681786946698231.96E-1717passed
/fits/[curve=quadratic]/f_valuetolerance282.21402485802861282.214024858028675.52E-1415passed
/fits/[curve=quadratic]/df1exact220.00E+2passed
/fits/[curve=quadratic]/df2exact880.00E+2passed
/fits/[curve=quadratic]/p_valuetail_probability3.8148399458772192e-083.8148399458772205e-081.30E-2315passed
/fits/[curve=quadratic]/coefficients/0/estimatetolerance-1.404242424242413-1.40424242424242981.68E-1413passed
/fits/[curve=quadratic]/coefficients/0/standard_errortolerance2.22586740943314432.2258674094331494.90E-1514passed
/fits/[curve=quadratic]/coefficients/0/t_valuetolerance-0.63087424627868005-0.63087424627868626.11E-1514passed
/fits/[curve=quadratic]/coefficients/0/p_valuetail_probability0.545719713815548510.54571971381554473.77E-1514passed
/fits/[curve=quadratic]/coefficients/0/lowertolerance-6.5371018748114285-6.5371018748114562.75E-1414passed
/fits/[curve=quadratic]/coefficients/0/uppertolerance3.7286170263266033.72861702632659686.25E-1514passed
/fits/[curve=quadratic]/coefficients/0/parametertolerance-1.404242424242413-1.40424242424242981.68E-1413passed
/fits/[curve=quadratic]/coefficients/0/parameter_lowertolerance-6.5371018748114285-6.5371018748114562.75E-1414passed
/fits/[curve=quadratic]/coefficients/0/parameter_uppertolerance3.7286170263266033.72861702632659686.25E-1514passed
/fits/[curve=quadratic]/coefficients/1/estimatetolerance3.52580419580419413.5258041958041994.91E-1514passed
/fits/[curve=quadratic]/coefficients/1/standard_errortolerance0.852524691940110490.85252469194011148.89E-1614passed
/fits/[curve=quadratic]/coefficients/1/t_valuetolerance4.13572091123887384.13572091123887551.73E-1515passed
/fits/[curve=quadratic]/coefficients/1/p_valuetail_probability0.00327258873062784330.00327258873062783331.00E-1714passed
/fits/[curve=quadratic]/coefficients/1/lowertolerance1.55987873082664171.55987873082664462.93E-1514passed
/fits/[curve=quadratic]/coefficients/1/uppertolerance5.4917296607817475.4917296607817536.18E-1514passed
/fits/[curve=quadratic]/coefficients/1/parametertolerance3.52580419580419413.5258041958041994.91E-1514passed
/fits/[curve=quadratic]/coefficients/1/parameter_lowertolerance1.55987873082664171.55987873082664462.93E-1514passed
/fits/[curve=quadratic]/coefficients/1/parameter_uppertolerance5.4917296607817475.4917296607817536.18E-1514passed
/fits/[curve=quadratic]/coefficients/2/estimatetolerance0.0882284382284382880.088228438228438022.63E-1614passed
/fits/[curve=quadratic]/coefficients/2/standard_errortolerance0.0691944211677144460.06919442116771455.50E-1715passed
/fits/[curve=quadratic]/coefficients/2/t_valuetolerance1.27508022669326061.27508022669325574.89E-1514passed
/fits/[curve=quadratic]/coefficients/2/p_valuetail_probability0.238062438247615550.238062438247617221.67E-1514passed
/fits/[curve=quadratic]/coefficients/2/lowertolerance-0.071334183117369915-0.071334183117370293.74E-1614passed
/fits/[curve=quadratic]/coefficients/2/uppertolerance0.247791059574246490.247791059574246351.37E-1615passed
/fits/[curve=quadratic]/coefficients/2/parametertolerance0.0882284382284382880.088228438228438022.63E-1614passed
/fits/[curve=quadratic]/coefficients/2/parameter_lowertolerance-0.071334183117369915-0.071334183117370293.74E-1614passed
/fits/[curve=quadratic]/coefficients/2/parameter_uppertolerance0.247791059574246490.247791059574246351.37E-1615passed
/nexact11110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

146  outputs <- curve_outputs("linear", as.double(seq_len(sum(keep))), d$y_gap[keep])
147  outputs <- c(outputs, curve_outputs("quadratic", as.double(seq_len(sum(keep))), d$y_gap[keep]))
148  outputs[["/n"]] <- exact(sum(keep), "complete rows")
149  emit("STAT-REG-CURVE-ESTIMATION-005", "Curve estimation sequence with retained order", "axis_coverage",
150       list(response = "y_gap", predictor = "", sequence = "retained_order", curves = list("linear", "quadratic"), confidence_level = 0.95),
151       outputs, "Blank predictor with retained order: the 11 complete rows are renumbered t = 1..11 in table order.")

Whole script generators/r/regression_curve_estimation.R


Small-sample boundary STAT-REG-DA-MIN-001

Case STAT-REG-DA-MIN-001 · status: passed

Reference

Evidence type
R
Source
R nlme/MASS/pls/stats
Citation
R nlme::lme, MASS::lda, pls::plsr and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_b.R · sha256 e3e09e1965800ae1
Dataset
reference_data/datasets/REG_MIN_DA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact660.00E+2passed
/class_order/0exactAApassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

67  da <- read_dataset("REG_DISCRIMINANT")[c(1L,5L,9L,13L,17L,21L),]
68  export_dataset(da,"REG_MIN_DA","Two classes with three distinct predictor values each",
69                 "R MASS lda small stratified-split boundary",
70                 "Constructed validation data; no third-party license.",
71                 c(x="Float64",class="String"))
72  da <- read_dataset("REG_MIN_DA")
73  fit <- MASS::lda(class~x,data=da,prior=c(0.5,0.5))
74  emit("STAT-REG-DA-MIN-001",
75       "tabelz.analysis.discriminant_analysis:compute_discriminant_analysis",
76       "discriminantAnalysisMenuItem","REG_MIN_DA",
77       list(response="class",predictors=list("x"),model="lda",standardize=FALSE,
78            prior_mode="equal",test_fraction=0.25,seed=0L),
79       list("/source_count"=exact(nrow(da),"six usable rows"),
80            "/class_order/0"=list(comparison="exact",value="A",rationale="R factor class order")),
81       c("Three rows per class allow one stratified holdout and two training values per class.",
82         "R lda fits the full CSV; split-dependent fitted moments are deferred."),
83       deferred=list(list(path="/artifact/class_means",
84                          reason="The training split uses a NumPy stratified permutation not replayable in R.")))

Whole script generators/r/regression_advanced_minimum_b.R


Regression edge STAT-REG-DA-ONECLASS-001

Case STAT-REG-DA-ONECLASS-001 · status: passed

Reference

Evidence type
Analytical
Source
Analytic design condition
Citation
Tabelz manual regression model design conditions; R stats/plm references
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_b.R · sha256 4a5d01b584989ad6
Dataset
reference_data/datasets/REG_EDGE_DA_ONECLASS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

63  da <- read_dataset("REG_MIN_DA")
64  da$class <- rep("A",nrow(da))
65  export_dataset(da,"REG_EDGE_DA_ONECLASS","One-class discriminant response",
66                 "Analytic two-class stratified-discriminant requirement",
67                 "Constructed validation data; no third-party license.",
68                 c(x="Float64",class="String"))
69  emit("STAT-REG-DA-ONECLASS-001","STAT-REG-DISCRIMINANT-LDA-001",
70       "REG_EDGE_DA_ONECLASS",
71       list(response="class",predictors=list("x"),model="lda",standardize=FALSE,
72            prior_mode="equal",test_fraction=0.25,seed=0L),
73       error=err("ValueError","at least two classes"),
74       note="LDA requires at least two observed classes; no stratified split exists for one class.")

Whole script generators/r/regression_advanced_special_b.R


Retained scoring boundary STAT-REG-DA-SCORE-MIDPOINT-001

Case STAT-REG-DA-SCORE-MIDPOINT-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls and analytic retained tree
Citation
R MASS::lda, pls::plsr and rpart::rpart documentation; Tabelz retained-artifact contract
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_boundaries.R · sha256 9a546ee72b8c92a9
Dataset
reference_data/datasets/REG_SCORE_BOUND_DA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/probabilities/0tolerance0.50.50.00E+1passed

Convention

Reference code

Whole script generators/r/regression_advanced_score_boundaries.R


Corrected retained scoring boundary STAT-REG-DA-SCORE-MIDPOINT-CORRECTED-001

Case STAT-REG-DA-SCORE-MIDPOINT-CORRECTED-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls or strict scoring contract
Citation
R MASS::lda, pls::plsr and Tabelz manual retained-scoring semantics
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_corrections.R · sha256 8c11616dfc0c8448
Dataset
reference_data/datasets/REG_SCORE_MID_DA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/probabilities/0tolerance0.50.50.00E+1passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

38  empty <- structure(list(),names=character())
39  lda <- MASS::lda(class~x,data=read_dataset("REG_DISCRIMINANT"),prior=c(0.5,0.5))
40  posterior <- stats::predict(lda,newdata=da)$posterior[1L,1L]
41  emit("STAT-REG-DA-SCORE-MIDPOINT-CORRECTED-001",
42       "STAT-REG-DISCRIMINANT-SCORE-LDA-001","REG_SCORE_MID_DA",
43       list("/0/probabilities/0"=tol(posterior,"R MASS LDA equal-prior midpoint posterior")),
44       NULL,"A one-row, fully observed scoring table lets the midpoint posterior be evaluated.")

Whole script generators/r/regression_advanced_score_corrections.R


Retained scoring boundary STAT-REG-DA-SCORE-MISSING-001

Case STAT-REG-DA-SCORE-MISSING-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls and analytic retained tree
Citation
R MASS::lda, pls::plsr and rpart::rpart documentation; Tabelz retained-artifact contract
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_boundaries.R · sha256 9a546ee72b8c92a9
Dataset
reference_data/datasets/REG_SCORE_BOUND_DA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/1/predictedabsentNonepassed

Convention

Reference code

Whole script generators/r/regression_advanced_score_boundaries.R


Regression II STAT-REG-DISCRIMINANT-LDA-001

Case STAT-REG-DISCRIMINANT-LDA-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact24240.00E+2passed
/class_order/0exactAApassed
/class_order/1exactBBpassed
/artifact/priors/0tolerance0.50.50.00E+1passed

Convention

Reference code

Whole script generators/r/regression_advanced_scoring.R


Regression II STAT-REG-DISCRIMINANT-QDA-001

Case STAT-REG-DISCRIMINANT-QDA-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact24240.00E+2passed
/class_order/0exactAApassed
/class_order/1exactBBpassed
/artifact/priors/0tolerance0.50.50.00E+1passed

Convention

Reference code

Whole script generators/r/regression_advanced_scoring.R


Regression II STAT-REG-DISCRIMINANT-SCORE-LDA-001

Case STAT-REG-DISCRIMINANT-SCORE-LDA-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/predictedexactAApassed
/12/predictedexactBBpassed
/0/row_idexactSTAT-REG-DISCRIMINANT-SCORE-LDA-001:0STAT-REG-DISCRIMINANT-SCORE-LDA-001:0passed

Convention

Reference code

Whole script generators/r/regression_advanced_scoring.R


Regression II STAT-REG-DISCRIMINANT-SCORE-QDA-001

Case STAT-REG-DISCRIMINANT-SCORE-QDA-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/predictedexactAApassed
/12/predictedexactBBpassed
/0/row_idexactSTAT-REG-DISCRIMINANT-SCORE-QDA-001:0STAT-REG-DISCRIMINANT-SCORE-QDA-001:0passed

Convention

Reference code

Whole script generators/r/regression_advanced_scoring.R


STAT REG EFFECTS 001

Case STAT-REG-EFFECTS-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS and nnet
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_core.R · sha256 4b30ab4a6f513e54
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/residual_variancetolerance0.237964434289452490.23796443428945252.21E-1817passed
/residual_dfexact68680.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

143  effects_entry <- "tabelz.executor.linear_effects"
144  effects_fit <- stats::lm(y ~ x + group, data = d)
145  emit("STAT-REG-EFFECTS-001", effects_entry, "linearEffectsMenuItem", "nominal",
146       list(response_column = "y", factor_columns = list("group"), covariate_columns = list("x")),
147       list("/residual_variance" = tol(summary(effects_fit)$sigma^2, "linear-effects residual variance", "1e-9"),
148            "/residual_df" = exact(stats::df.residual(effects_fit), "full model residual df")),
149       list("Treatment contrasts use alphabetically first factor level A as reference; Type II tests."))

Whole script generators/r/regression_core.R


Regression forest study design on independent rows

Case STAT-REG-FOREST-001 · status: passed

Reference

Evidence type
Analytical
Source
documented study design counts (manual chapter 76) evaluated by hand
Citation
Breiman (2001) Random forests, Machine Learning 45, 5-32; Tabelz manual chapter 76 (study design: 20% test reservation rounded up, 5-fold CV, 12-candidate grid, 500 trees, 10 permutations).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_forest.R · sha256 94a0ffe8dda384b9
Dataset
reference_data/datasets/REG_FOREST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responseexactwearwearpassed
/numeric_predictorssequence['load', 'speed']['load', 'speed']passed
/categorical_predictorssequence['tool']['tool']passed
/seedexact770.00E+2passed
/encoded_feature_countexact550.00E+2passed
/treesexact5005000.00E+2passed
/source_countexact32320.00E+2passed
/excluded_countexact000.00E+2passed
/missing_response_countexact220.00E+2passed
/split/test_fractiontolerance0.200000000000000010.21.10E-1817passed
/split/seedexact770.00E+2passed
/test_metrics/row_countexact660.00E+2passed
/baseline_test_metrics/row_countexact660.00E+2passed
/importance/0/predictorexactloadloadpassed
/importance/1/predictorexactspeedspeedpassed
/importance/2/predictorexacttooltoolpassed
/importance/0/permutationsexact10100.00E+2passed
/artifact/fitexacttrainingtrainingpassed
/artifact/fit_row_countexact24240.00E+2passed
/deployment_artifact/fitexactdeploymentdeploymentpassed
/deployment_artifact/fit_row_countexact30300.00E+2passed
/candidates/0/settings/min_samples_leafexact110.00E+2passed
/candidates/0/settings/max_featurestolerance0.50.50.00E+1passed
/candidates/0/settings/max_depthexact10100.00E+2passed
/candidates/1/settings/min_samples_leafexact110.00E+2passed
/candidates/1/settings/max_featurestolerance0.50.50.00E+1passed
/candidates/1/settings/max_depthabsentNonepassed
/candidates/2/settings/min_samples_leafexact110.00E+2passed
/candidates/2/settings/max_featurestolerance11.00.00E+2passed
/candidates/2/settings/max_depthexact10100.00E+2passed
/candidates/3/settings/min_samples_leafexact110.00E+2passed
/candidates/3/settings/max_featurestolerance11.00.00E+2passed
/candidates/3/settings/max_depthabsentNonepassed
/candidates/4/settings/min_samples_leafexact330.00E+2passed
/candidates/4/settings/max_featurestolerance0.50.50.00E+1passed
/candidates/4/settings/max_depthexact10100.00E+2passed
/candidates/5/settings/min_samples_leafexact330.00E+2passed
/candidates/5/settings/max_featurestolerance0.50.50.00E+1passed
/candidates/5/settings/max_depthabsentNonepassed
/candidates/6/settings/min_samples_leafexact330.00E+2passed
/candidates/6/settings/max_featurestolerance11.00.00E+2passed
/candidates/6/settings/max_depthexact10100.00E+2passed
/candidates/7/settings/min_samples_leafexact330.00E+2passed
/candidates/7/settings/max_featurestolerance11.00.00E+2passed
/candidates/7/settings/max_depthabsentNonepassed
/candidates/8/settings/min_samples_leafexact550.00E+2passed
/candidates/8/settings/max_featurestolerance0.50.50.00E+1passed
/candidates/8/settings/max_depthexact10100.00E+2passed
/candidates/9/settings/min_samples_leafexact550.00E+2passed
/candidates/9/settings/max_featurestolerance0.50.50.00E+1passed
/candidates/9/settings/max_depthabsentNonepassed
/candidates/10/settings/min_samples_leafexact550.00E+2passed
/candidates/10/settings/max_featurestolerance11.00.00E+2passed
/candidates/10/settings/max_depthexact10100.00E+2passed
/candidates/11/settings/min_samples_leafexact550.00E+2passed
/candidates/11/settings/max_featurestolerance11.00.00E+2passed
/candidates/11/settings/max_depthabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

24  emit <- function(id, title, kind, configuration, outputs, deferred, notes, error = NULL) {
25    expected <- list(schema_version = "2.0.0", case_id = id, outputs = outputs,
26                     expected_error = error, deferred_outputs = deferred, generator_run = run)
27    write_expected("regression", id, expected)
28    case <- list(
29      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
30      entry_point = "tabelz.analysis.regression_forest:compute_regression_forest",
31      menu_items = list("regressionForestMenuItem"), title = title, case_kind = kind,
32      dataset = list(path = "reference_data/datasets/REG_FOREST.json", sha256 = ""),
33      excluded_rows = list(), configuration = configuration, "repeat" = 1L,
34      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"),
35                      sha256 = ""),
36      oracle = list(oracle_class = "analytic",
37                    source = "documented study design counts (manual chapter 76) evaluated by hand",
38                    citation = "Breiman (2001) Random forests, Machine Learning 45, 5-32; Tabelz manual chapter 76 (study design: 20% test reservation rounded up, 5-fold CV, 12-candidate grid, 500 trees, 10 permutations).",
39                    generator = list(path = "generators/r/regression_forest.R", sha256 = ""),
40                    independent = TRUE,
41                    convention_notes = c(notes, "Expected computed from the exported CSV.", manual)),
42      supported = TRUE, unsupported_reason = NULL,
43      requirement_ids = list("REQ-STAT-REG-FOREST-001"), risk_ids = list("RISK-REFERENCE-001"),
44      review_reference = NULL
45    )
46    write_case("regression", id, case)

Whole script generators/r/regression_forest.R


Regression forest refuses dependent rows

Case STAT-REG-FOREST-DEPENDENT-001 · status: passed

Reference

Evidence type
Analytical
Source
documented study design counts (manual chapter 76) evaluated by hand
Citation
Breiman (2001) Random forests, Machine Learning 45, 5-32; Tabelz manual chapter 76 (study design: 20% test reservation rounded up, 5-fold CV, 12-candidate grid, 500 trees, 10 permutations).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_forest.R · sha256 94a0ffe8dda384b9
Dataset
reference_data/datasets/REG_FOREST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

126  emit("STAT-REG-FOREST-DEPENDENT-001", "Regression forest refuses dependent rows", "invalid_input",
127       list(response = "wear", numeric_predictors = list("load", "speed"),
128            categorical_predictors = list("tool"), dependence = "clustered", seed = 7L),
129       setNames(list(), character(0)), list(),
130       list("The manual requires independent observations for the random split; declaring clustered rows is refused before any fitting."),
131       error = list(error_class = "ValueError", message_contains = "needs independent observations"))

Whole script generators/r/regression_forest.R


Binomial GEE with an exchangeable working correlation

Case STAT-REG-GEE-001 · status: passed

Reference

Evidence type
R
Source
Base-R GEE with explicit dense block matrices
Citation
Liang and Zeger (1986); Mancl and DeRouen (2001); formulas as stated in manual chapter 67
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1
Reference script
generators/r/regression_gee.R · sha256 1f1de72c0127d183
Dataset
reference_data/datasets/REG_GEE_CLUSTERED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/[term=Intercept]/estimatetolerance0.0690148355207556440.069014835520755781.38E-1614passed
/coefficients/[term=Intercept]/std_errortolerance0.333115483725523020.3331154837255235.53E-1917passed
/coefficients/[term=Intercept]/p_valuetolerance0.837865162700314640.83786516270031424.48E-1615passed
/coefficients/[term=Intercept]/lowertolerance-0.6237367363280002-0.62373673632799993.37E-1615passed
/coefficients/[term=Intercept]/uppertolerance0.76176640736951140.76176640736951151.11E-1615passed
/coefficients/[term=x]/estimatetolerance0.893642251008733960.8936422510087343.43E-1817passed
/coefficients/[term=x]/std_errortolerance0.328069042107264620.328069042107264565.65E-1715passed
/coefficients/[term=x]/p_valuetolerance0.0127155601639127260.0127155601639127042.24E-1714passed
/coefficients/[term=x]/lowertolerance0.211385329015918220.211385329015918553.35E-1614passed
/coefficients/[term=x]/uppertolerance1.57589917300154971.57589917300154924.53E-1615passed
/coefficients/[term=group = b (vs a)]/estimatetolerance0.897147776373452890.89714777637345281.13E-1615passed
/coefficients/[term=group = b (vs a)]/std_errortolerance0.545140449716977950.54514044971697781.13E-1615passed
/coefficients/[term=group = b (vs a)]/p_valuetolerance0.114707576129978340.114707576129978231.12E-1615passed
/coefficients/[term=group = b (vs a)]/lowertolerance-0.23653385017904804-0.23653385017904773.29E-1614passed
/coefficients/[term=group = b (vs a)]/uppertolerance2.03082940292595372.03082940292595334.36E-1615passed
/degrees_of_freedomexact21210.00E+2passed
/rhotolerance0.123957878946208260.123957878946208292.58E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

186  b <- gee(d$event, rep(1, nrow(d)), rep(0, nrow(d)), "binomial", "exchangeable")
187  emit("STAT-REG-GEE-001", "Binomial GEE with an exchangeable working correlation",
188       list(family = "binomial", working = "exchangeable"), b,
189       c(sprintf("Converged in %d iterations; rho = %.10g.", b$iterations, b$rho)))

Whole script generators/r/regression_gee.R


Poisson GEE with exposure and independence working correlation

Case STAT-REG-GEE-002 · status: passed

Reference

Evidence type
R
Source
Base-R GEE with explicit dense block matrices
Citation
Liang and Zeger (1986); Mancl and DeRouen (2001); formulas as stated in manual chapter 67
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1
Reference script
generators/r/regression_gee.R · sha256 1f1de72c0127d183
Dataset
reference_data/datasets/REG_GEE_CLUSTERED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/[term=Intercept]/estimatetolerance0.25200045397561610.25200045397561611.05E-1817passed
/coefficients/[term=Intercept]/std_errortolerance0.165380820544665990.16538082054466598.26E-1715passed
/coefficients/[term=Intercept]/p_valuetolerance0.142487562900935450.142487562900935261.95E-1614passed
/coefficients/[term=Intercept]/lowertolerance-0.091927790081495209-0.091927790081494992.22E-1614passed
/coefficients/[term=Intercept]/uppertolerance0.595928698032727410.59592869803272722.21E-1615passed
/coefficients/[term=x]/estimatetolerance0.550198077984701110.55019807798470113.95E-1817passed
/coefficients/[term=x]/std_errortolerance0.0902112331141043810.090211233114104384.96E-1917passed
/coefficients/[term=x]/p_valuetolerance4.7263012267512877e-064.726301226751279e-068.48E-2114passed
/coefficients/[term=x]/lowertolerance0.362593548650653520.36259354865065351.19E-1817passed
/coefficients/[term=x]/uppertolerance0.737802607318748690.73780260731874873.29E-1817passed
/coefficients/[term=group = b (vs a)]/estimatetolerance-0.07736028762671876-0.077360287626718825.55E-1715passed
/coefficients/[term=group = b (vs a)]/std_errortolerance0.171492304381805990.171492304381806078.11E-1715passed
/coefficients/[term=group = b (vs a)]/p_valuetolerance0.656540660048903320.6565406600489033.29E-1615passed
/coefficients/[term=group = b (vs a)]/lowertolerance-0.43399805808337588-0.4339980580833761.13E-1615passed
/coefficients/[term=group = b (vs a)]/uppertolerance0.279277482829938360.279277482829938361.60E-1817passed
/degrees_of_freedomexact21210.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

190  pz <- gee(d$count, rep(1, nrow(d)), log(d$exposure), "poisson", "independence")
191  emit("STAT-REG-GEE-002", "Poisson GEE with exposure and independence working correlation",
192       list(family = "poisson", working = "independence"), pz,
193       c("Independence coefficients equal the Poisson GLM; the covariance is the bias-reduced sandwich."),
194       kind = "axis_coverage")

Whole script generators/r/regression_gee.R


STAT REG GENERAL 001

Case STAT-REG-GENERAL-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS and nnet
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_core.R · sha256 4b30ab4a6f513e54
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/residual_variancetolerance0.237964434289452550.237964434289452552.28E-1817passed
/residual_dfexact68680.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

133  glm_entry <- "tabelz.executor.general_linear"
134  glm_fit <- stats::lm(y ~ x + group, data = d,
135                       contrasts = list(group = stats::contr.sum(3)))
136  emit("STAT-REG-GENERAL-001", glm_entry, "generalLinearMenuItem", "nominal",
137       list(response_column = "y", factor_columns = list("group"), covariate_columns = list("x"),
138            test_type = "III"),
139       list("/residual_variance" = tol(summary(glm_fit)$sigma^2, "GLM residual variance", "1e-9"),
140            "/residual_df" = exact(stats::df.residual(glm_fit), "full model residual df")),
141       list("Sum-to-zero factor contrasts and Type III partial tests per manual; residual variance is SSE/df."))

Whole script generators/r/regression_core.R


STAT REG GENERALIZED LINEAR 001

Case STAT-REG-GENERALIZED-LINEAR-001 · status: passed

Reference

Evidence type
R
Source
R stats::glm with contr.sum, MASS::glm.nb/theta.ml, MASS::gamma.dispersion, sandwich::vcovHC(type = 'HC0'); Wald/F/LR tests and marginal means from the fitted coefficients
Citation
R Core Team, stats package help: glm, summary.glm, logLik.glm; Venables & Ripley (2002) MASS: glm.nb, theta.ml, gamma.dispersion; Zeileis (2006) sandwich.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_glm_general.R · sha256 c450a6f63ea68bc5
Dataset
reference_data/datasets/REG_GLM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact40400.00E+2passed
/statistic_kindexactttpassed
/type3_kindexactFFpassed
/fit/df_residualexact36360.00E+2passed
/fit/df_nullexact39390.00E+2passed
/fit/deviancetolerance21.177148559770621.17714855977064.00E-1616passed
/fit/null_deviancetolerance36.38741697500000136.3874169752.82E-1617passed
/fit/pearson_chi_squaretolerance21.177148559770621.17714855977064.00E-1616passed
/coefficients/0/termexactInterceptInterceptpassed
/coefficients/0/estimatetolerance1.63511418276565571.6351141827656561.99E-1615passed
/coefficients/0/standard_errortolerance0.123908077054111780.123908077054111784.03E-1816passed
/coefficients/0/statistictolerance13.19618721910748813.196187219107491.60E-1515passed
/coefficients/0/p_valuetail_probability2.3310797888188366e-152.3310797888188244e-151.22E-2914passed
/coefficients/0/lowertolerance1.38381695501918811.38381695501918864.55E-1615passed
/coefficients/0/uppertolerance1.88641141051212321.88641141051212324.34E-1716passed
/coefficients/1/termexactg[sum:A]g[sum:A]passed
/coefficients/1/estimatetolerance-0.44071147143947842-0.44071147143947855.49E-1715passed
/coefficients/1/standard_errortolerance0.172444615908658890.172444615908658872.41E-1715passed
/coefficients/1/statistictolerance-2.5556696514834427-2.55566965148344369.05E-1615passed
/coefficients/1/p_valuetail_probability0.0149659275914000460.0149659275914000133.33E-1714passed
/coefficients/1/lowertolerance-0.79044536246520747-0.79044536246520751.64E-1817passed
/coefficients/1/uppertolerance-0.090977580413749315-0.090977580413749542.22E-1614passed
/coefficients/2/termexactg[sum:B]g[sum:B]passed
/coefficients/2/estimatetolerance-0.27039053616443137-0.27039053616443094.98E-1614passed
/coefficients/2/standard_errortolerance0.172674104614992390.172674104614992363.25E-1715passed
/coefficients/2/statistictolerance-1.565900901975517-1.56590090197551432.65E-1514passed
/coefficients/2/p_valuetail_probability0.126120902087290990.126120902087291545.52E-1614passed
/coefficients/2/lowertolerance-0.62058985185876825-0.62058985185876766.67E-1614passed
/coefficients/2/uppertolerance0.0798087795299054490.079808779529905783.33E-1614passed
/coefficients/3/termexactxxpassed
/coefficients/3/estimatetolerance0.413512732103154810.413512732103154875.75E-1715passed
/coefficients/3/standard_errortolerance0.112458873845996470.112458873845996451.57E-1715passed
/coefficients/3/statistictolerance3.67701292002468223.6770129200246838.51E-1615passed
/coefficients/3/p_valuetail_probability0.000764662667301685630.00076466266730168272.92E-1814passed
/coefficients/3/lowertolerance0.185435564699072080.185435564699072211.34E-1615passed
/coefficients/3/uppertolerance0.641589899507237550.64158989950723751.20E-1817passed
/type3/0/termexactggpassed
/type3/0/dfexact220.00E+2passed
/type3/0/statistictolerance8.42003166946829938.4200316694682963.57E-1515passed
/type3/0/denominator_dfexact36360.00E+2passed
/type3/0/p_valuetail_probability0.00100024199023223690.00100024199023223982.86E-1814passed
/type3/1/termexactxxpassed
/type3/1/dfexact110.00E+2passed
/type3/1/statistictolerance13.52042401402843713.5204240140284469.18E-1515passed
/type3/1/denominator_dfexact36360.00E+2passed
/type3/1/p_valuetail_probability0.000764662667301685630.00076466266730168272.92E-1814passed
/fit/scaletolerance0.588254126660294420.58825412666029441.45E-1817passed
/fit/ml_scaletolerance0.529428713994265050.5294287139942654.41E-1817passed
/fit/log_likelihoodtolerance-44.0384062899874-44.03840628998744.11E-1617passed
/fit/aictolerance98.07681257997479998.07681257997481.78E-1617passed
/fit/bictolerance106.52120985054448106.521209850544488.81E-1819passed
/fit/parameters_countedexact550.00E+2passed
/means/0/levelexactAApassed
/means/0/link_meantolerance1.2895313153465081.28953131534650822.22E-1615passed
/means/0/link_setolerance0.207304672307621760.207304672307621763.21E-1816passed
/means/0/link_lowertolerance0.869097953064202230.86909795306420252.25E-1615passed
/means/0/link_uppertolerance1.70996467762881381.7099646776288141.89E-1615passed
/means/0/meantolerance1.2895313153465081.28953131534650822.22E-1615passed
/means/0/mean_setolerance0.207304672307621760.207304672307621763.21E-1816passed
/means/1/levelexactBBpassed
/means/1/link_meantolerance1.4598522506215551.45985225062155598.80E-1615passed
/means/1/link_setolerance0.212753839361905870.212753839361905873.47E-1816passed
/means/1/link_lowertolerance1.02836746532611521.02836746532611619.15E-1615passed
/means/1/link_uppertolerance1.89133703591699481.89133703591699568.46E-1615passed
/means/1/meantolerance1.4598522506215551.45985225062155598.80E-1615passed
/means/1/mean_setolerance0.212753839361905870.212753839361905873.47E-1816passed
/means/2/levelexactCCpassed
/means/2/link_meantolerance2.4413447943898962.44134479438989654.89E-1615passed
/means/2/link_setolerance0.214769509470232660.214769509470232633.21E-1715passed
/means/2/link_lowertolerance2.0057720406398032.0057720406398048.55E-1615passed
/means/2/link_uppertolerance2.87691754813998912.8769175481399892.25E-1717passed
/means/2/meantolerance2.4413447943898962.44134479438989654.89E-1615passed
/means/2/mean_setolerance0.214769509470232660.214769509470232633.21E-1715passed
/comparisons/0/firstexactAApassed
/comparisons/0/secondexactBBpassed
/comparisons/0/differencetolerance-0.17032093527504705-0.17032093527504765.52E-1614passed
/comparisons/0/standard_errortolerance0.297438946688182710.297438946688182665.06E-1715passed
/comparisons/0/statistictolerance-0.57262486023258208-0.5726248602325841.89E-1514passed
/comparisons/0/p_valuetail_probability0.570458032096016980.57045803209601551.45E-1514passed
/comparisons/0/holm_p_valuetail_probability0.570458032096016980.57045803209601551.45E-1514passed
/comparisons/0/lowertolerance-0.77355507871129459-0.77355507871129482.18E-1615passed
/comparisons/0/uppertolerance0.432913208161200440.432913208161199667.81E-1614passed
/comparisons/1/firstexactAApassed
/comparisons/1/secondexactCCpassed
/comparisons/1/differencetolerance-1.1518134790433883-1.15181347904338784.78E-1615passed
/comparisons/1/standard_errortolerance0.301549429671366110.301549429671366055.62E-1715passed
/comparisons/1/statistictolerance-3.8196506632383782-3.81965066323837738.81E-1615passed
/comparisons/1/p_valuetail_probability0.000508518663947689920.00050851866394769099.73E-1914passed
/comparisons/1/holm_p_valuetail_probability0.00152555599184306980.00152555599184307262.77E-1814passed
/comparisons/1/lowertolerance-1.7633840683589623-1.76338406835896148.57E-1615passed
/comparisons/1/uppertolerance-0.54024288972781431-0.54024288972781421.08E-1615passed
/comparisons/2/firstexactBBpassed
/comparisons/2/secondexactCCpassed
/comparisons/2/differencetolerance-0.98149254376834116-0.98149254376834029.96E-1614passed
/comparisons/2/standard_errortolerance0.301943142133149770.30194314213314975.88E-1715passed
/comparisons/2/statistictolerance-3.2505873020806222-3.25058730208061952.68E-1515passed
/comparisons/2/p_valuetail_probability0.00250096094694980940.00250096094694982931.99E-1714passed
/comparisons/2/holm_p_valuetail_probability0.00500192189389961870.0050019218938996593.99E-1714passed
/comparisons/2/lowertolerance-1.5938616189657697-1.59386161896576841.31E-1515passed
/comparisons/2/uppertolerance-0.36912346857091261-0.369123468570911837.82E-1614passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

48  emit <- function(id, kind, config, outputs, notes) {
49    write_expected("regression", id, list(
50      schema_version = "2.0.0", case_id = id, outputs = outputs,
51      expected_error = NULL, deferred_outputs = list(), generator_run = run))
52    write_case("regression", id, list(
53      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
54      entry_point = "tabelz.analysis.generalized_linear_models:compute_generalized_linear",
55      menu_items = list("generalizedLinearMenuItem"), title = gsub("-", " ", id),
56      case_kind = kind,
57      dataset = list(path = "reference_data/datasets/REG_GLM_01.json", sha256 = ""),
58      excluded_rows = list(), configuration = config, "repeat" = 1L,
59      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"), sha256 = ""),
60      oracle = list(
61        oracle_class = "independent_implementation",
62        source = "R stats::glm with contr.sum, MASS::glm.nb/theta.ml, MASS::gamma.dispersion, sandwich::vcovHC(type = 'HC0'); Wald/F/LR tests and marginal means from the fitted coefficients",
63        citation = "R Core Team, stats package help: glm, summary.glm, logLik.glm; Venables & Ripley (2002) MASS: glm.nb, theta.ml, gamma.dispersion; Zeileis (2006) sandwich.",
64        generator = list(path = "generators/r/regression_glm_general.R", sha256 = ""),
65        independent = TRUE,
66        convention_notes = c(
67          notes,
68          "Factors are sum-to-zero coded on sorted levels (last level = -1), covariates enter as recorded; coefficient tests are t with Type III F when the dispersion is a moment estimate (normal: deviance/df; gamma: Pearson/df) and z with chi-square otherwise (docs/manual/chapters/86-regression-and-count-models.tex:107-170).",
69          "Log-likelihood, AIC and BIC of true-likelihood families are evaluated at the maximum-likelihood dispersion and count the dispersion (and theta) as parameters.",
70          "Marginal means average the linear predictor over the equal-weight factor grid at covariate means and the mean offset, then invert the link; pairwise differences are on the link scale with Holm-adjusted p-values and unadjusted intervals (manual lines 160-169).",
71          "Expected computed from the exported CSV."
72        )),
73      supported = TRUE, unsupported_reason = NULL,
74      requirement_ids = list("REQ-STAT-REG-GENERALIZED-LINEAR-001"),
75      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/regression_glm_general.R


STAT REG GENERALIZED LINEAR 002

Case STAT-REG-GENERALIZED-LINEAR-002 · status: passed

Reference

Evidence type
R
Source
R stats::glm with contr.sum, MASS::glm.nb/theta.ml, MASS::gamma.dispersion, sandwich::vcovHC(type = 'HC0'); Wald/F/LR tests and marginal means from the fitted coefficients
Citation
R Core Team, stats package help: glm, summary.glm, logLik.glm; Venables & Ripley (2002) MASS: glm.nb, theta.ml, gamma.dispersion; Zeileis (2006) sandwich.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_glm_general.R · sha256 c450a6f63ea68bc5
Dataset
reference_data/datasets/REG_GLM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact40400.00E+2passed
/statistic_kindexactzzpassed
/type3_kindexactchi-squarechi-squarepassed
/fit/df_residualexact36360.00E+2passed
/fit/df_nullexact39390.00E+2passed
/fit/deviancetolerance33.871679186080233.871679186080196.65E-1515passed
/fit/null_deviancetolerance68.55859159025747868.558591590257461.39E-1415passed
/fit/pearson_chi_squaretolerance28.79415998725660928.7941599872566271.82E-1415passed
/coefficients/0/termexactInterceptInterceptpassed
/coefficients/0/estimatetolerance0.383793053583731390.38379305358373058.87E-1614passed
/coefficients/0/standard_errortolerance0.113312520734946060.113312520734946948.77E-1614passed
/coefficients/0/statistictolerance3.38703129269780683.38703129269777263.42E-1413passed
/coefficients/0/p_valuetail_probability0.000706533285079508580.00070653328507959558.69E-1712passed
/coefficients/0/lowertolerance0.197410542873841550.197410542873839132.42E-1513passed
/coefficients/0/uppertolerance0.570175564293621260.57017556429362196.68E-1614passed
/coefficients/0/exp_estimatetolerance1.46784164568010331.4678416456801021.30E-1515passed
/coefficients/0/exp_lowertolerance1.21824407948063771.21824407948063482.86E-1514passed
/coefficients/0/exp_uppertolerance1.76857752324263841.76857752324263951.10E-1515passed
/coefficients/1/termexactg[sum:A]g[sum:A]passed
/coefficients/1/estimatetolerance-0.26775542055150375-0.26775542055150391.70E-1615passed
/coefficients/1/standard_errortolerance0.128289303703702030.12828930370370254.76E-1614passed
/coefficients/1/statistictolerance-2.0871219409679993-2.0871219409679936.22E-1514passed
/coefficients/1/p_valuetail_probability0.0368771132002783590.036877113200278915.49E-1613passed
/coefficients/1/lowertolerance-0.4787725470476169-0.478772547047617961.06E-1514passed
/coefficients/1/uppertolerance-0.056738294055390603-0.056738294055389916.93E-1613passed
/coefficients/1/exp_estimatetolerance0.765094884706031930.76509488470603181.11E-1615passed
/coefficients/1/exp_lowertolerance0.619543385640216830.61954338564021626.70E-1614passed
/coefficients/1/exp_uppertolerance0.944841307600165940.94484130760016666.62E-1615passed
/coefficients/2/termexactg[sum:B]g[sum:B]passed
/coefficients/2/estimatetolerance0.0133313688359572060.0133313688359573831.77E-1613passed
/coefficients/2/standard_errortolerance0.130683116059269470.130683116059269722.46E-1614passed
/coefficients/2/statistictolerance0.102012939681595540.10201293968159671.17E-1513passed
/coefficients/2/p_valuetail_probability0.918746404287393290.91874640428739248.86E-1615passed
/coefficients/2/lowertolerance-0.20162322859545229-0.201623228595452562.74E-1614passed
/coefficients/2/uppertolerance0.228285966267366690.228285966267367366.67E-1614passed
/coefficients/2/exp_estimatetolerance1.01342062774022291.01342062774022312.23E-1615passed
/coefficients/2/exp_lowertolerance0.817402843949216920.81740284394921672.25E-1615passed
/coefficients/2/exp_uppertolerance1.25644457482850851.25644457482850916.31E-1615passed
/coefficients/3/termexactxxpassed
/coefficients/3/estimatetolerance0.510545277110658670.51054527711065925.54E-1614passed
/coefficients/3/standard_errortolerance0.0931370698598690120.093137069859869585.69E-1614passed
/coefficients/3/statistictolerance5.48165491870002345.4816549186999962.75E-1414passed
/coefficients/3/p_valuetail_probability4.213655165680836e-084.21365516568148e-086.44E-2112passed
/coefficients/3/lowertolerance0.357348429948020520.35734842994802014.46E-1614passed
/coefficients/3/uppertolerance0.663742124273296820.66374212427329841.55E-1514passed
/coefficients/3/exp_estimatetolerance1.66619948773019931.66619948773020029.13E-1615passed
/coefficients/3/exp_lowertolerance1.42953387550273251.42953387550273186.90E-1615passed
/coefficients/3/exp_uppertolerance1.9420461316007981.9420461316008013.08E-1514passed
/type3/0/termexactggpassed
/type3/0/dfexact220.00E+2passed
/type3/0/statistictolerance5.78370012764639975.7837001276463633.64E-1414passed
/type3/0/p_valuetail_probability0.0554734881238291840.055473488123830189.93E-1613passed
/type3/1/termexactxxpassed
/type3/1/dfexact110.00E+2passed
/type3/1/statistictolerance30.04854064770815830.048540647707862.98E-1314passed
/type3/1/p_valuetail_probability4.21365516568084e-084.2136551656814885e-086.49E-2112passed
/fit/scaletolerance11.00.00E+2passed
/fit/log_likelihoodtolerance-69.225963817565287-69.225963817565292.82E-1617passed
/fit/aictolerance146.45192763513057146.451927635130573.44E-1516passed
/fit/bictolerance153.20744545158632153.207445451586323.77E-1516passed
/fit/parameters_countedexact440.00E+2passed
/lr_tests/0/termexactggpassed
/lr_tests/0/dfexact220.00E+2passed
/lr_tests/0/statistictolerance5.80109328594426195.8010932859442622.08E-1717passed
/lr_tests/0/p_valuetail_probability0.0549931502193407170.054993150219340722.11E-1917passed
/lr_tests/1/termexactxxpassed
/lr_tests/1/dfexact110.00E+2passed
/lr_tests/1/statistictolerance34.05010135597825634.0501013559782563.79E-1616passed
/lr_tests/1/p_valuetail_probability5.3711110602089626e-095.371111060208967e-094.15E-2415passed
/means/0/levelexactAApassed
/means/0/link_meantolerance0.68821622403153460.68821622403153378.92E-1614passed
/means/0/link_setolerance0.171591085618462670.17159108561846391.23E-1514passed
/means/0/link_lowertolerance0.40597400449946580.405974004499462753.05E-1514passed
/means/0/link_uppertolerance0.970458443563603450.97045844356360471.22E-1514passed
/means/0/meantolerance1.99016236136037541.99016236136037381.57E-1515passed
/means/0/mean_setolerance0.341494120142830040.34149412014283222.16E-1514passed
/means/1/levelexactBBpassed
/means/1/link_meantolerance0.96930301341899550.9693030134189954.48E-1615passed
/means/1/link_setolerance0.170996198830704470.170996198830704853.84E-1614passed
/means/1/link_lowertolerance0.688039295577396290.68803929557739511.22E-1514passed
/means/1/link_uppertolerance1.25056673126059481.2505667312605952.33E-1615passed
/means/1/meantolerance2.63610648805961742.6361064880596161.28E-1515passed
/means/1/mean_setolerance0.450764189171152410.45076418917115327.75E-1614passed
/means/2/levelexactCCpassed
/means/2/link_meantolerance1.21039569629858491.21039569629858426.66E-1615passed
/means/2/link_setolerance0.145475402261985080.14547540226198543.29E-1614passed
/means/2/link_lowertolerance0.971109953255734460.97110995325573311.33E-1514passed
/means/2/link_uppertolerance1.44968143934143551.44968143934143522.70E-1615passed
/means/2/meantolerance3.35481187658490043.35481187658489782.63E-1515passed
/means/2/mean_setolerance0.48804260725947340.48804260725947417.26E-1614passed
/comparisons/0/firstexactAApassed
/comparisons/0/secondexactBBpassed
/comparisons/0/differencetolerance-0.28108678938746096-0.28108678938746133.29E-1614passed
/comparisons/0/standard_errortolerance0.226747379051019830.226747379051020466.35E-1614passed
/comparisons/0/statistictolerance-1.2396473580592715-1.23964735805926951.99E-1514passed
/comparisons/0/p_valuetail_probability0.215105855951643250.21510585595164418.59E-1614passed
/comparisons/0/holm_p_valuetail_probability0.43021171190328650.43021171190328821.72E-1514passed
/comparisons/0/lowertolerance-0.65405303822127103-0.65405303822127261.56E-1514passed
/comparisons/0/uppertolerance0.0918794594463491210.091879459446350018.88E-1614passed
/comparisons/0/ratiotolerance0.754962809876960630.75496280987696042.24E-1615passed
/comparisons/1/firstexactAApassed
/comparisons/1/secondexactCCpassed
/comparisons/1/differencetolerance-0.5221794722670503-0.52217947226705041.15E-1615passed
/comparisons/1/standard_errortolerance0.217150536090869320.217150536090870077.47E-1614passed
/comparisons/1/statistictolerance-2.4046888470426704-2.40468884704266248.01E-1514passed
/comparisons/1/p_valuetail_probability0.0161862407783691440.016186240778369493.47E-1613passed
/comparisons/1/holm_p_valuetail_probability0.0485587223351074330.0485587223351084741.04E-1513passed
/comparisons/1/lowertolerance-0.87936031915057311-0.87936031915057461.44E-1514passed
/comparisons/1/uppertolerance-0.1649986253835275-0.164998625383526221.28E-1514passed
/comparisons/1/ratiotolerance0.593226217914282090.5932262179142821.06E-1615passed
/comparisons/2/firstexactBBpassed
/comparisons/2/secondexactCCpassed
/comparisons/2/differencetolerance-0.24109268287958935-0.241092682879589151.96E-1615passed
/comparisons/2/standard_errortolerance0.221391394275141870.22139139427514223.36E-1614passed
/comparisons/2/statistictolerance-1.0889885023261701-1.08898850232616742.70E-1514passed
/comparisons/2/p_valuetail_probability0.276158955980285170.27615895598028641.22E-1514passed
/comparisons/2/holm_p_valuetail_probability0.43021171190328650.43021171190328821.72E-1514passed
/comparisons/2/lowertolerance-0.60524912072889969-0.60524912072890025.57E-1615passed
/comparisons/2/uppertolerance0.1230637549697210.123063754969721919.11E-1614passed
/comparisons/2/ratiotolerance0.785768795698641780.7857687956986422.17E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

219  fit <- stats::glm(count ~ g + x + offset(log_exposure), family = stats::poisson(), data = d,
220                    contrasts = sum_contrasts, control = control)
221  p <- length(stats::coef(fit))
222  V <- summary(fit)$cov.unscaled
223  ll <- as.numeric(stats::logLik(fit))
224  drop_g <- stats::glm(count ~ x + offset(log_exposure), family = stats::poisson(), data = d, control = control)
225  drop_x <- stats::glm(count ~ g + offset(log_exposure), family = stats::poisson(), data = d,
226                       contrasts = sum_contrasts, control = control)
227  lr_g <- 2 * (ll - as.numeric(stats::logLik(drop_g)))
228  lr_x <- 2 * (ll - as.numeric(stats::logLik(drop_x)))
229  emit("STAT-REG-GENERALIZED-LINEAR-002", "axis_coverage",
230       list(response = "count", family = "poisson", factors = list("g"), covariates = list("x"),
231            offset = "log_exposure", link = "canonical", scale = "default", covariance = "model",
232            confidence_level = 0.90, comparisons = "holm", lr_tests = TRUE),
233       c(glm_outputs(fit, V, FALSE, 0.90, coef_terms, type3, exp_link = TRUE),
234         list("/fit/scale" = tol(1, "fixed dispersion", "1e-12"),
235              "/fit/log_likelihood" = tol(ll, "Poisson log-likelihood"),
236              "/fit/aic" = tol(-2 * ll + 2 * p, "AIC"),
237              "/fit/bic" = tol(-2 * ll + log(n) * p, "BIC"),
238              "/fit/parameters_counted" = exact(as.integer(p), "coefficients only"),
239              "/lr_tests/0/term" = exact("g", "LR test term"),
240              "/lr_tests/0/df" = exact(2L, "LR df"),
241              "/lr_tests/0/statistic" = tol(lr_g, "LR chi-square dropping g"),
242              "/lr_tests/0/p_value" = prob(stats::pchisq(lr_g, 2, lower.tail = FALSE), "LR tail"),
243              "/lr_tests/1/term" = exact("x", "LR test term"),
244              "/lr_tests/1/df" = exact(1L, "LR df"),
245              "/lr_tests/1/statistic" = tol(lr_x, "LR chi-square dropping x"),
246              "/lr_tests/1/p_value" = prob(stats::pchisq(lr_x, 1, lower.tail = FALSE), "LR tail")),
247         means_outputs(fit, V, FALSE, 0.90, exp, exp, base = mean(d$log_exposure), ratio = TRUE)),
248       list("stats::glm(count ~ g + x + offset(log_exposure), poisson); LR tests refit without each term; marginal means use the mean offset."))

Whole script generators/r/regression_glm_general.R


STAT REG GENERALIZED LINEAR 003

Case STAT-REG-GENERALIZED-LINEAR-003 · status: passed

Reference

Evidence type
R
Source
R stats::glm with contr.sum, MASS::glm.nb/theta.ml, MASS::gamma.dispersion, sandwich::vcovHC(type = 'HC0'); Wald/F/LR tests and marginal means from the fitted coefficients
Citation
R Core Team, stats package help: glm, summary.glm, logLik.glm; Venables & Ripley (2002) MASS: glm.nb, theta.ml, gamma.dispersion; Zeileis (2006) sandwich.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_glm_general.R · sha256 c450a6f63ea68bc5
Dataset
reference_data/datasets/REG_GLM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact40400.00E+2passed
/statistic_kindexactzzpassed
/type3_kindexactchi-squarechi-squarepassed
/fit/df_residualexact36360.00E+2passed
/fit/df_nullexact39390.00E+2passed
/fit/deviancetolerance42.18937966619692842.189379666196927.52E-1515passed
/fit/null_deviancetolerance96.99410646850020696.99410646850021.20E-1617passed
/fit/pearson_chi_squaretolerance41.50190689738364141.501906895599821.78E-910passed
/coefficients/0/termexactInterceptInterceptpassed
/coefficients/0/estimatetolerance-0.24149993886540047-0.241499938823861784.15E-119passed
/coefficients/0/standard_errortolerance0.083926941247404740.083926941263274461.59E-119passed
/coefficients/0/statistictolerance-2.8775019710713972-2.87750197003235231.04E-99passed
/coefficients/0/p_valuetail_probability0.00400837406124218260.0040083740744427131.32E-118passed
/coefficients/0/lowertolerance-0.40599372104292286-0.405993721032488261.04E-1110passed
/coefficients/0/uppertolerance-0.077006156687878108-0.07700615661523537.26E-119passed
/coefficients/1/termexactg[sum:A]g[sum:A]passed
/coefficients/1/estimatetolerance-0.46791316999676102-0.467913169950731844.60E-1110passed
/coefficients/1/standard_errortolerance0.0778595426154549550.077859542562848265.26E-119passed
/coefficients/1/statistictolerance-6.0097086918139855-6.0097086952833323.47E-99passed
/coefficients/1/p_valuetail_probability1.8585697828574615e-091.8585697430863236e-093.98E-177passed
/coefficients/1/lowertolerance-0.62051506937581424-0.62051506922667791.49E-109passed
/coefficients/1/uppertolerance-0.31531127061770781-0.31531127067478585.71E-119passed
/coefficients/2/termexactg[sum:B]g[sum:B]passed
/coefficients/2/estimatetolerance0.143722913946731980.14372291393034571.64E-119passed
/coefficients/2/standard_errortolerance0.101336807170298510.101336807146917012.34E-119passed
/coefficients/2/statistictolerance1.41826960963159991.41826960979713681.66E-109passed
/coefficients/2/p_valuetail_probability0.156112065760499820.156112065712188894.83E-119passed
/coefficients/2/lowertolerance-0.054893578415333377-0.054893578385892792.94E-119passed
/coefficients/2/uppertolerance0.342339406308797330.34233940624658426.22E-119passed
/coefficients/3/termexactxxpassed
/coefficients/3/estimatetolerance0.419756241537607710.41975624150171783.59E-1110passed
/coefficients/3/standard_errortolerance0.0686135407294495570.068613540746801661.74E-119passed
/coefficients/3/statistictolerance6.1176880988076536.117688096737442.07E-99passed
/coefficients/3/p_valuetail_probability9.4942662237600883e-109.49426634706466e-101.23E-177passed
/coefficients/3/lowertolerance0.285276172856114510.28527617278621516.99E-119passed
/coefficients/3/uppertolerance0.554236310219100910.55423631021722051.88E-1211passed
/type3/0/termexactggpassed
/type3/0/dfexact220.00E+2passed
/type3/0/statistictolerance36.20380905630952436.203809097705314.14E-88passed
/type3/0/p_valuetail_probability1.3754434879389093e-081.375443459470126e-082.85E-167passed
/type3/1/termexactxxpassed
/type3/1/dfexact110.00E+2passed
/type3/1/statistictolerance37.42610767429279437.426107648962952.53E-89passed
/type3/1/p_valuetail_probability9.4942662237601028e-109.494266347064684e-101.23E-177passed
/fit/log_likelihoodtolerance-69.10516944847366-69.105169448473665.43E-1718passed
/fit/aictolerance146.21033889694732146.210338896947321.09E-1618passed
/fit/parameters_countedexact440.00E+2passed
/means/0/levelexactAApassed
/means/0/link_meantolerance-0.61284818549643483-0.61284818541712357.93E-119passed
/means/0/link_setolerance0.0862265526565200310.086226552638339931.82E-119passed
/means/0/link_lowertolerance-0.78184912321426059-0.7818491230993171.15E-109passed
/means/0/link_uppertolerance-0.44384724777860907-0.443847247734934.37E-1110passed
/means/0/meantolerance0.269988363228416410.26998836325463982.62E-1110passed
/means/0/mean_setolerance0.028509808422924540.0285098084182992474.63E-129passed
/means/1/levelexactBBpassed
/means/1/link_meantolerance-0.0012121015529418522-0.00121210153604589531.69E-117passed
/means/1/link_setolerance0.129641546829851850.12964154680608062.38E-119passed
/means/1/link_lowertolerance-0.25530486423951426-0.255304864176027556.35E-119passed
/means/1/link_uppertolerance0.25288066113363050.252880661103935762.97E-119passed
/means/1/meantolerance0.499516441560797830.49951644156753836.74E-1210passed
/means/1/mean_setolerance0.0517194563341977290.051719456324715449.48E-129passed
/means/2/levelexactCCpassed
/means/2/link_meantolerance0.179255240550355220.179255240553994553.64E-1210passed
/means/2/link_setolerance0.142658782720079980.142658782707412751.27E-1110passed
/means/2/link_lowertolerance-0.10035083565932648-0.100350835630859892.85E-119passed
/means/2/link_uppertolerance0.458861316760036940.4588613167388492.12E-1110passed
/means/2/meantolerance0.571131354800545440.57113135480197431.43E-1211passed
/means/2/mean_setolerance0.0560055554479177780.056005555442908295.01E-1210passed
/comparisons/0/firstexactAApassed
/comparisons/0/secondexactBBpassed
/comparisons/0/differencetolerance-0.611636083943493-0.61163608388107756.24E-119passed
/comparisons/0/standard_errortolerance0.144218566685782840.14421856661515877.06E-119passed
/comparisons/0/statistictolerance-4.2410356585785465-4.24103566022260651.64E-99passed
/comparisons/0/p_valuetail_probability2.2249078239952836e-052.2249078076961617e-051.63E-138passed
/comparisons/0/holm_p_valuetail_probability4.4498156479905673e-054.4498156153923234e-053.26E-138passed
/comparisons/0/lowertolerance-0.89429928054961538-0.89429928034877912.01E-109passed
/comparisons/0/uppertolerance-0.32897288733737062-0.32897288741337587.60E-119passed
/comparisons/1/firstexactAApassed
/comparisons/1/secondexactCCpassed
/comparisons/1/differencetolerance-0.79210342604679007-0.79210342597111797.57E-1110passed
/comparisons/1/standard_errortolerance0.159944070744540870.159944070673808647.07E-119passed
/comparisons/1/statistictolerance-4.9523775552262901-4.9523775569432691.72E-99passed
/comparisons/1/p_valuetail_probability7.3312186683004034e-077.331218603594731e-076.47E-158passed
/comparisons/1/holm_p_valuetail_probability2.199365600490121e-062.199365581078419e-061.94E-148passed
/comparisons/1/lowertolerance-1.1055880442468167-1.10558804403251192.14E-109passed
/comparisons/1/uppertolerance-0.47861880784676353-0.478618807909723946.30E-119passed
/comparisons/2/firstexactBBpassed
/comparisons/2/secondexactCCpassed
/comparisons/2/differencetolerance-0.18046734210329707-0.180467342090040451.33E-1110passed
/comparisons/2/standard_errortolerance0.195456455764146970.195456455732765963.14E-119passed
/comparisons/2/statistictolerance-0.92331226102382136-0.92331226110423768.04E-1110passed
/comparisons/2/p_valuetail_probability0.355844500110136960.355844500068241754.19E-119passed
/comparisons/2/holm_p_valuetail_probability0.355844500110136960.355844500068241754.19E-119passed
/comparisons/2/lowertolerance-0.56355495594687133-0.56355495587210917.48E-119passed
/comparisons/2/uppertolerance0.202620271740277140.202620271692028234.82E-119passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

251  fit <- stats::glm(cbind(events, trials - events) ~ g + x, family = stats::binomial(link = "probit"),
252                    data = d, contrasts = sum_contrasts, control = control)
253  X <- stats::model.matrix(fit)
254  u <- stats::residuals(fit, type = "working") * stats::weights(fit, type = "working")
255  bread <- summary(fit)$cov.unscaled
256  V <- bread %*% crossprod(X * u) %*% bread
257  stopifnot(max(abs(V - sandwich::vcovHC(fit, type = "HC0"))) < 1e-10)
258  ll <- as.numeric(stats::logLik(fit))
259  p <- length(stats::coef(fit))
260  emit("STAT-REG-GENERALIZED-LINEAR-003", "axis_coverage",
261       list(response = "events", trials = "trials", family = "binomial", factors = list("g"),
262            covariates = list("x"), link = "probit", scale = "default", covariance = "hc0",
263            confidence_level = 0.95, comparisons = "holm", lr_tests = FALSE),
264       c(glm_outputs(fit, V, FALSE, 0.95, coef_terms, type3),
265         list("/fit/log_likelihood" = tol(ll, "binomial log-likelihood with the combinatorial term"),
266              "/fit/aic" = tol(-2 * ll + 2 * p, "AIC"),
267              "/fit/parameters_counted" = exact(as.integer(p), "coefficients only")),
268         means_outputs(fit, V, FALSE, 0.95, stats::pnorm, stats::dnorm)),
269       list("stats::glm(cbind(events, trials - events) ~ g + x, binomial(probit)); HC0 = (X'WX)^-1 [sum u_i^2 x_i x_i'] (X'WX)^-1 with u_i = m_i (y_i - mu_i) mu_eta_i / V(mu_i), equal to sandwich::vcovHC(type = 'HC0')."))

Whole script generators/r/regression_glm_general.R


STAT REG GENERALIZED LINEAR 004

Case STAT-REG-GENERALIZED-LINEAR-004 · status: passed

Reference

Evidence type
R
Source
R stats::glm with contr.sum, MASS::glm.nb/theta.ml, MASS::gamma.dispersion, sandwich::vcovHC(type = 'HC0'); Wald/F/LR tests and marginal means from the fitted coefficients
Citation
R Core Team, stats package help: glm, summary.glm, logLik.glm; Venables & Ripley (2002) MASS: glm.nb, theta.ml, gamma.dispersion; Zeileis (2006) sandwich.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_glm_general.R · sha256 c450a6f63ea68bc5
Dataset
reference_data/datasets/REG_GLM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact40400.00E+2passed
/statistic_kindexactttpassed
/type3_kindexactFFpassed
/fit/df_residualexact36360.00E+2passed
/fit/df_nullexact39390.00E+2passed
/fit/deviancetolerance7.40156346530492387.4015634653049238.71E-1615passed
/fit/null_deviancetolerance12.43516946209092412.4351694620909241.14E-1617passed
/fit/pearson_chi_squaretolerance6.39772420738428336.3977242073842829.20E-1615passed
/coefficients/0/termexactInterceptInterceptpassed
/coefficients/0/estimatetolerance0.68443768450972220.68443768450972231.11E-1615passed
/coefficients/0/standard_errortolerance0.049568729204178250.049568729204181142.89E-1513passed
/coefficients/0/statistictolerance13.80785215796957613.8078521579687738.03E-1313passed
/coefficients/0/p_valuetail_probability5.929338814244517e-165.929338814255024e-161.05E-2711passed
/coefficients/0/lowertolerance0.583907642174503860.58390764217449815.77E-1514passed
/coefficients/0/uppertolerance0.784967726844940540.78496772684494655.99E-1514passed
/coefficients/1/termexactg[sum:A]g[sum:A]passed
/coefficients/1/estimatetolerance0.151446563626162210.151446563626162182.61E-1715passed
/coefficients/1/standard_errortolerance0.0621112674184573920.062111267418460042.65E-1513passed
/coefficients/1/statistictolerance2.43831063059514012.43831063059503531.05E-1313passed
/coefficients/1/p_valuetail_probability0.0198203924371372210.0198203924371421134.89E-1512passed
/coefficients/1/lowertolerance0.0254790747814962470.025479074781490895.36E-1512passed
/coefficients/1/uppertolerance0.277414052470828180.27741405247083355.33E-1513passed
/coefficients/2/termexactg[sum:B]g[sum:B]passed
/coefficients/2/estimatetolerance0.0622085096320058530.062208509632005937.59E-1714passed
/coefficients/2/standard_errortolerance0.0614732516299828920.061473251629984421.53E-1513passed
/coefficients/2/statistictolerance1.01196061673211291.0119606167320892.40E-1413passed
/coefficients/2/p_valuetail_probability0.31831038945439460.318310389454405651.11E-1413passed
/coefficients/2/lowertolerance-0.062465023219524164-0.0624650232195271552.99E-1513passed
/coefficients/2/uppertolerance0.186882042483535880.1868820424835393.13E-1513passed
/coefficients/3/termexactxxpassed
/coefficients/3/estimatetolerance-0.14860982010117738-0.148609820101177743.56E-1614passed
/coefficients/3/standard_errortolerance0.0387846346128335250.03878463461283875.18E-1512passed
/coefficients/3/statistictolerance-3.8316674008836369-3.83166740088313465.02E-1312passed
/coefficients/3/p_valuetail_probability0.00049122523681026050.00049122523681097047.10E-1611passed
/coefficients/3/lowertolerance-0.22726870488968379-0.227268704889694641.08E-1413passed
/coefficients/3/uppertolerance-0.069950935312670964-0.069950935312660831.01E-1412passed
/type3/0/termexactggpassed
/type3/0/dfexact220.00E+2passed
/type3/0/statistictolerance9.65496013122593779.654960131222833.11E-1212passed
/type3/0/denominator_dfexact36360.00E+2passed
/type3/0/p_valuetail_probability0.000439531931103986180.000439531931104875178.89E-1611passed
/type3/1/termexactxxpassed
/type3/1/dfexact110.00E+2passed
/type3/1/statistictolerance14.68167507099436414.6816750709905193.85E-1212passed
/type3/1/denominator_dfexact36360.00E+2passed
/type3/1/p_valuetail_probability0.000491225236810260830.00049122523681097047.10E-1611passed
/fit/scaletolerance0.177714561316230090.177714561316230062.70E-1715passed
/fit/ml_scaletolerance0.179675642319136270.17967564231913674.44E-1614passed
/fit/log_likelihoodtolerance-39.078154278025956-39.0781542780259563.52E-1617passed
/fit/aictolerance88.15630855605191188.156308556051912.96E-1617passed
/fit/bictolerance96.60070582662159296.600705826621591.09E-1617passed
/fit/parameters_countedexact550.00E+2passed
/means/0/levelexactAApassed
/means/0/link_meantolerance0.801696559021608520.80169655902160853.57E-1817passed
/means/0/link_setolerance0.0851762368120732440.085176236812075452.21E-1513passed
/means/0/link_lowertolerance0.628951144116952520.62895114411694814.44E-1514passed
/means/0/link_uppertolerance0.974441973926264530.9744419739262694.44E-1514passed
/means/0/meantolerance1.24735473633615351.24735473633615354.21E-1716passed
/means/0/mean_setolerance0.132525182022099370.132525182022102783.41E-1513passed
/means/1/levelexactBBpassed
/means/1/link_meantolerance0.712458505027452230.71245850502745231.08E-1615passed
/means/1/link_setolerance0.0802553101651462130.08025531016514674.85E-1614passed
/means/1/link_lowertolerance0.549693191934693790.5496931919346937.75E-1614passed
/means/1/link_uppertolerance0.875223818120210660.87522381812021171.00E-1514passed
/means/1/meantolerance1.4035905150173881.40359051501738772.59E-1615passed
/means/1/mean_setolerance0.158108284668785540.158108284668786438.91E-1614passed
/means/2/levelexactCCpassed
/means/2/link_meantolerance0.43659492213727830.43659492213727833.76E-1817passed
/means/2/link_setolerance0.0493488913734245260.049348891373439491.50E-1412passed
/means/2/link_lowertolerance0.336510731587800090.33651073158776983.03E-1413passed
/means/2/link_uppertolerance0.53667911268675650.53667911268678683.03E-1413passed
/means/2/meantolerance2.29045265827798742.29045265827798741.44E-1717passed
/means/2/mean_setolerance0.258892840246528310.25889284024660687.85E-1412passed
/comparisons/0/firstexactAApassed
/comparisons/0/secondexactBBpassed
/comparisons/0/differencetolerance0.0892380539941563520.089238053994156269.69E-1714passed
/comparisons/0/standard_errortolerance0.113327609460562630.113327609460563721.09E-1514passed
/comparisons/0/statistictolerance0.787434363249413630.78743436324940518.55E-1513passed
/comparisons/0/p_valuetail_probability0.436181560219830460.43618156021983544.94E-1513passed
/comparisons/0/holm_p_valuetail_probability0.436181560219830460.43618156021983544.94E-1513passed
/comparisons/0/lowertolerance-0.14060099089826608-0.140600990898268332.25E-1513passed
/comparisons/0/uppertolerance0.319077098886578760.319077098886580872.11E-1514passed
/comparisons/1/firstexactAApassed
/comparisons/1/secondexactCCpassed
/comparisons/1/differencetolerance0.365101636884330280.36510163688433032.84E-1817passed
/comparisons/1/standard_errortolerance0.0937946965865380650.093794696586548961.09E-1412passed
/comparisons/1/statistictolerance3.89256162844426433.89256162844381234.52E-1312passed
/comparisons/1/p_valuetail_probability0.000412009360660478660.000412009360661019245.41E-1611passed
/comparisons/1/holm_p_valuetail_probability0.00123602808198143610.00123602808198305761.62E-1511passed
/comparisons/1/lowertolerance0.174877175413390870.174877175413368832.20E-1412passed
/comparisons/1/uppertolerance0.555326098355269690.55532609835529172.20E-1413passed
/comparisons/2/firstexactBBpassed
/comparisons/2/secondexactCCpassed
/comparisons/2/differencetolerance0.275863582890173930.275863582890174041.12E-1615passed
/comparisons/2/standard_errortolerance0.092525123532308420.09252512353231728.78E-1513passed
/comparisons/2/statistictolerance2.98149921187456092.9814992118742792.82E-1313passed
/comparisons/2/p_valuetail_probability0.00511877190958125430.0051187719095850253.77E-1512passed
/comparisons/2/holm_p_valuetail_probability0.0102375438191625090.010237543819170057.54E-1512passed
/comparisons/2/lowertolerance0.088213934914324110.088213934914306431.77E-1412passed
/comparisons/2/uppertolerance0.463513230866023750.46351323086604171.79E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

273  fit <- stats::glm(ypos ~ g + x, family = stats::Gamma(link = "inverse"), data = d,
274                    contrasts = sum_contrasts, control = control)
275  p <- length(stats::coef(fit)); df <- n - p
276  pearson <- sum(stats::residuals(fit, type = "pearson")^2)
277  scale <- pearson / df
278  stopifnot(abs(scale - summary(fit)$dispersion) < 1e-12 * scale)
279  V <- scale * summary(fit)$cov.unscaled
280  mu <- stats::fitted(fit)
281  target <- stats::deviance(fit) / (2 * n)
282  shape <- stats::uniroot(function(s) log(s) - digamma(s) - target, c(1e-3, 1e6), tol = 1e-14)$root
283  ml_scale <- 1 / shape
284  stopifnot(abs(ml_scale - MASS::gamma.dispersion(fit)) < 1e-6)
285  ll <- sum(stats::dgamma(d$ypos, shape = shape, scale = mu * ml_scale, log = TRUE))
286  emit("STAT-REG-GENERALIZED-LINEAR-004", "axis_coverage",
287       list(response = "ypos", family = "gamma", factors = list("g"), covariates = list("x"),
288            link = "canonical", scale = "default", covariance = "model", confidence_level = 0.95,
289            comparisons = "holm", lr_tests = FALSE),
290       c(glm_outputs(fit, V, TRUE, 0.95, coef_terms, type3),
291         list("/fit/scale" = tol(scale, "Pearson chi-square / residual df"),
292              "/fit/ml_scale" = tol(ml_scale, "ML dispersion solving log(nu) - digamma(nu) = D/(2n)"),
293              "/fit/log_likelihood" = tol(ll, "gamma log-likelihood at the ML dispersion"),
294              "/fit/aic" = tol(-2 * ll + 2 * (p + 1), "AIC counting the dispersion"),
295              "/fit/bic" = tol(-2 * ll + log(n) * (p + 1), "BIC counting the dispersion"),
296              "/fit/parameters_counted" = exact(as.integer(p + 1L), "coefficients plus dispersion")),
297         means_outputs(fit, V, TRUE, 0.95, function(eta) 1 / eta, function(eta) -1 / eta^2)),
298       list("stats::glm(ypos ~ g + x, Gamma(inverse)); the moment dispersion is Pearson/df (summary.glm) and the ML dispersion solves log(nu) - digamma(nu) = D/(2n) (MASS::gamma.dispersion)."))

Whole script generators/r/regression_glm_general.R


STAT REG GENERALIZED LINEAR 005

Case STAT-REG-GENERALIZED-LINEAR-005 · status: passed

Reference

Evidence type
R
Source
R stats::glm with contr.sum, MASS::glm.nb/theta.ml, MASS::gamma.dispersion, sandwich::vcovHC(type = 'HC0'); Wald/F/LR tests and marginal means from the fitted coefficients
Citation
R Core Team, stats package help: glm, summary.glm, logLik.glm; Venables & Ripley (2002) MASS: glm.nb, theta.ml, gamma.dispersion; Zeileis (2006) sandwich.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_glm_general.R · sha256 c450a6f63ea68bc5
Dataset
reference_data/datasets/REG_GLM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact40400.00E+2passed
/statistic_kindexactzzpassed
/type3_kindexactchi-squarechi-squarepassed
/fit/df_residualexact36360.00E+2passed
/fit/df_nullexact39390.00E+2passed
/fit/deviancetolerance42.24820622358665642.248206221919381.67E-910passed
/fit/null_deviancetolerance49.57402854315162749.574028541144822.01E-910passed
/fit/pearson_chi_squaretolerance37.29857820461871437.298578201173583.45E-910passed
/coefficients/0/termexactInterceptInterceptpassed
/coefficients/0/estimatetolerance0.848147760567031340.84814776034447812.23E-109passed
/coefficients/0/standard_errortolerance0.154100110529095030.154100110867575443.38E-108passed
/coefficients/0/statistictolerance5.50387509557883095.5038750820453751.35E-88passed
/coefficients/0/p_valuetail_probability3.715328188713317e-083.7153284740682296e-082.85E-157passed
/coefficients/0/lowertolerance0.546117093916363540.54611709303040098.86E-108passed
/coefficients/0/uppertolerance1.15017842721769911.15017842765855544.41E-109passed
/coefficients/0/exp_estimatetolerance2.3353172767104262.33531727619069375.20E-109passed
/coefficients/0/exp_lowertolerance1.72653600837768041.7265360068480341.53E-99passed
/coefficients/0/exp_uppertolerance3.15875646753913533.15875646893169251.39E-99passed
/coefficients/1/termexactg[sum:A]g[sum:A]passed
/coefficients/1/estimatetolerance-0.4393930221659208-0.439393021663697045.02E-108passed
/coefficients/1/standard_errortolerance0.218159438564279190.218159438350310382.14E-109passed
/coefficients/1/statistictolerance-2.0140912768092618-2.01409127648256943.27E-109passed
/coefficients/1/p_valuetail_probability0.0439999512104873310.043999951244780473.43E-119passed
/coefficients/1/lowertolerance-0.86697766463938652-0.86697766371779169.22E-108passed
/coefficients/1/uppertolerance-0.011808379692455129-0.011808379609602468.29E-118passed
/coefficients/1/exp_estimatetolerance0.644427455577708310.64442745590135513.24E-109passed
/coefficients/1/exp_lowertolerance0.42021967671725030.42021967710452263.87E-109passed
/coefficients/1/exp_uppertolerance0.988261065608761920.9882610656906428.19E-1110passed
/coefficients/2/termexactg[sum:B]g[sum:B]passed
/coefficients/2/estimatetolerance0.343226983603190150.343226983914784343.12E-109passed
/coefficients/2/standard_errortolerance0.198777885875755840.198777885830344454.54E-119passed
/coefficients/2/statistictolerance1.72668595448147991.7266859564434961.96E-98passed
/coefficients/2/p_valuetail_probability0.0842240795049515460.084224079152393833.53E-108passed
/coefficients/2/lowertolerance-0.046370513636304345-0.046370513235705514.01E-108passed
/coefficients/2/uppertolerance0.73282448084268470.73282448106527422.23E-109passed
/coefficients/2/exp_estimatetolerance1.40948865643356651.4094886568727554.39E-109passed
/coefficients/2/exp_lowertolerance0.954688171666634220.95468817204908133.82E-109passed
/coefficients/2/exp_uppertolerance2.08094991807295322.0809499185361514.63E-109passed
/coefficients/3/termexactxxpassed
/coefficients/3/estimatetolerance0.263040818305194160.263040819295589159.90E-108passed
/coefficients/3/standard_errortolerance0.138979259619402660.138979259752335161.33E-109passed
/coefficients/3/statistictolerance1.89266239455827031.89266239987416165.32E-98passed
/coefficients/3/p_valuetail_probability0.0584027804135924670.058402779706204597.07E-107passed
/coefficients/3/lowertolerance-0.0093535251468768732-0.0093535244170248617.30E-107passed
/coefficients/3/uppertolerance0.535435161757265240.53543516300820321.25E-98passed
/coefficients/3/exp_estimatetolerance1.30087981757743371.30087981886581861.29E-99passed
/coefficients/3/exp_lowertolerance0.990690083000247630.99069008372330487.23E-109passed
/coefficients/3/exp_uppertolerance1.7081914201213161.70819142225815732.14E-98passed
/type3/0/termexactggpassed
/type3/0/dfexact220.00E+2passed
/type3/0/statistictolerance4.73974714809897614.7397471573118269.21E-98passed
/type3/0/p_valuetail_probability0.0934925454148397840.093492544984173384.31E-108passed
/type3/1/termexactxxpassed
/type3/1/dfexact110.00E+2passed
/type3/1/statistictolerance3.58217093977504543.5821709598974212.01E-88passed
/type3/1/p_valuetail_probability0.0584027804135925780.058402779706204577.07E-107passed
/fit/thetatolerance2.35718711056722492.3571871103525372.15E-1010passed
/fit/theta_setolerance1.05014089593052521.05014089574795481.83E-109passed
/fit/log_likelihoodtolerance-81.251286229650262-81.251286229650232.86E-1415passed
/fit/aictolerance172.50257245930052172.502572459300475.32E-1415passed
/fit/bictolerance180.9469697298702180.946969729870155.24E-1415passed
/fit/parameters_countedexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

312  fit <- stats::glm(nbcount ~ g + x, family = MASS::negative.binomial(theta), data = d,
313                    contrasts = sum_contrasts, control = control)
314  mu <- stats::fitted(fit)
315  y <- d$nbcount
316  p <- length(stats::coef(fit))
317  V <- summary(fit, dispersion = 1)$cov.unscaled
318  ll <- sum(stats::dnbinom(y, size = theta, mu = mu, log = TRUE))
319  info <- sum(-trigamma(y + theta) + trigamma(theta) - 1 / theta + 2 / (mu + theta) - (y + theta) / (mu + theta)^2)
320  emit("STAT-REG-GENERALIZED-LINEAR-005", "axis_coverage",
321       list(response = "nbcount", family = "negative_binomial", factors = list("g"),
322            covariates = list("x"), link = "canonical", theta = "estimate", scale = "default",
323            covariance = "model", confidence_level = 0.95, comparisons = "holm", lr_tests = FALSE),
324       c(glm_outputs(fit, V, FALSE, 0.95, coef_terms, type3, exp_link = TRUE),
325         list("/fit/theta" = tol(theta, "ML theta", "1e-7"),
326              "/fit/theta_se" = tol(1 / sqrt(info), "theta SE from the observed information", "1e-6"),
327              "/fit/log_likelihood" = tol(ll, "negative binomial log-likelihood"),
328              "/fit/aic" = tol(-2 * ll + 2 * (p + 1), "AIC counting theta"),
329              "/fit/bic" = tol(-2 * ll + log(n) * (p + 1), "BIC counting theta"),
330              "/fit/parameters_counted" = exact(as.integer(p + 1L), "coefficients plus theta"))),
331       list("MASS::glm.nb polished by alternating MASS::theta.ml (eps 1e-14) and glm(negative.binomial(theta)) steps; theta SE = 1/sqrt(observed information) as in theta.ml; coefficient covariance is the expected information at fixed theta with dispersion 1.",
332            "Theta and its SE carry 1e-7 and 1e-6 relative tolerances because both sides reach theta by an alternating search rather than in closed form."))

Whole script generators/r/regression_glm_general.R


Binomial random-intercept GLMM at the zero-variance boundary (closed form)

Case STAT-REG-GLMM-001 · status: passed

Reference

Evidence type
Analytical
Source
Closed-form maximum likelihood of a saturated design at the tau = 0 boundary
Citation
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/REG_GLMM_BINOMIAL_BOUNDARY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/variance_statusexactboundaryboundarypassed
/random_intercept_sdtolerance0.00.00.00E+1passed
/coefficients/[term=Intercept]/estimatetolerance-1.0986122886681098-1.09861228866810981.79E-1716passed
/coefficients/[term=x]/estimatetolerance2.19722457733621962.19722457733621963.58E-1716passed
/log_likelihoodtolerance-26.992086941702798-26.9920869417028057.09E-1515passed
/inference_availableexactFalseFalsepassed

Convention


Poisson random-intercept GLMM with exposure at the zero-variance boundary (closed form)

Case STAT-REG-GLMM-002 · status: passed

Reference

Evidence type
Analytical
Source
Closed-form maximum likelihood of a saturated design at the tau = 0 boundary
Citation
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/REG_GLMM_POISSON_BOUNDARY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/variance_statusexactboundaryboundarypassed
/random_intercept_sdtolerance0.00.00.00E+1passed
/coefficients/[term=Intercept]/estimatetolerance0.69314718055994530.69314718055994531.38E-1716passed
/coefficients/[term=level = B (vs A)]/estimatetolerance0.40546510810816440.40546510810816462.07E-1615passed
/log_likelihoodtolerance-33.87021657780064-33.870216577800657.43E-1515passed

Convention


Regression II large-offset numeric STAT-REG-HARD-ADV-DA-SCORE-001

Case STAT-REG-HARD-ADV-DA-SCORE-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/MASS/pls
Citation
R nlme::lme, plm::plm, stats::lm/nls, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_hard.R · sha256 57a7311aff0ed329
Dataset
reference_data/datasets/REG_ADV_HARD_DA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/probabilities/0tolerance0.999999931743970950.9999999317439711.03E-1817passed
/12/probabilities/0tolerance0.00407013771589611640.0040701377158961271.04E-1714passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

101  base_da$artifact$class_means <- list(list(as.double(lda$means[1L, 1L])),
102                                       list(as.double(lda$means[2L, 1L])))
103  posterior <- stats::predict(lda, newdata = da)$posterior
104  emit("STAT-REG-HARD-ADV-DA-SCORE-001", "STAT-REG-DISCRIMINANT-SCORE-LDA-001",
105       "REG_ADV_HARD_DA", base_da,
106       list("/0/probabilities/0" = tol(posterior[1L, 1L],
107                                        "LDA posterior under million-unit offset", "1e-9"),
108            "/12/probabilities/0" = tol(posterior[13L, 1L],
109                                         "opposite-class LDA posterior", "1e-9")),
110       "LDA class means carry the same million-unit translation as the CSV; pooled covariance is unchanged.")

Whole script generators/r/regression_advanced_hard.R


Regression II large-offset numeric STAT-REG-HARD-ADV-MIXED-001

Case STAT-REG-HARD-ADV-MIXED-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/MASS/pls
Citation
R nlme::lme, plm::plm, stats::lm/nls, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_hard.R · sha256 57a7311aff0ed329
Dataset
reference_data/datasets/REG_ADV_HARD_ORTH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.660185185185213410.6601851851851852.84E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

60  mixed <- nlme::lme(distance ~ age, random = ~1 | subject, data = orth,
61                     method = "REML", control = nlme::lmeControl(opt = "optim"))
62  emit("STAT-REG-HARD-ADV-MIXED-001", "STAT-REG-MIXED-REML-001", "REG_ADV_HARD_ORTH",
63       list(response_column = "distance", group_column = "subject",
64            fixed_numeric_columns = list("age"), estimation = "reml"),
65       list("/coefficients/1/estimate" = tol(nlme::fixef(mixed)[["age"]],
66                                                 "REML fixed slope after large offsets")),
67       "Both response and age are translated; fixed slope should remain stable.")

Whole script generators/r/regression_advanced_hard.R


Regression II large-offset numeric STAT-REG-HARD-ADV-NONLINEAR-001

Case STAT-REG-HARD-ADV-NONLINEAR-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/MASS/pls
Citation
R nlme::lme, plm::plm, stats::lm/nls, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_hard.R · sha256 57a7311aff0ed329
Dataset
reference_data/datasets/REG_ADV_HARD_DECAY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/estimates/1tolerance0.812588320858578240.81258823271788468.81E-86passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

84  nls_fit <- stats::nls(y ~ c + a * exp(-k * x), data = decay,
85                        start = list(a = 4, k = 0.8, c = 10001), algorithm = "port",
86                        lower = c(a = 0, k = 0.01, c = 9990),
87                        upper = c(a = 10, k = 3, c = 10010),
88                        control = stats::nls.control(maxiter = 1000, tol = 1e-12))
89  emit("STAT-REG-HARD-ADV-NONLINEAR-001", "STAT-REG-NONLINEAR-DECAY-001",
90       "REG_ADV_HARD_DECAY",
91       list(x_column = "x", response = "y", curve = "exponential_decay",
92            parameters = list(list(name = "a", start = 4, lower = 0, upper = 10),
93                              list(name = "k", start = 0.8, lower = 0.01, upper = 3),
94                              list(name = "c", start = 10001, lower = 9990, upper = 10010))),
95       list("/artifact/estimates/1" = tol(coef(nls_fit)[["k"]],
96                                            "decay rate with ten-thousand-unit baseline")),
97       "R nls port estimates rate on the original response scale with a large fitted c.")

Whole script generators/r/regression_advanced_hard.R


Regression II large-offset numeric STAT-REG-HARD-ADV-PANEL-001

Case STAT-REG-HARD-ADV-PANEL-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/MASS/pls
Citation
R nlme::lme, plm::plm, stats::lm/nls, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_hard.R · sha256 57a7311aff0ed329
Dataset
reference_data/datasets/REG_ADV_HARD_ORTH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/estimatetolerance0.66018518518518510.66018518518518512.81E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

68  panel <- plm::plm(distance ~ age, data = orth, index = c("subject", "time"),
69                    model = "within", effect = "individual")
70  emit("STAT-REG-HARD-ADV-PANEL-001", "STAT-REG-PANEL-001", "REG_ADV_HARD_ORTH",
71       list(response_column = "distance", predictor_columns = list("age"),
72            entity_column = "subject", time_column = "time"),
73       list("/coefficients/0/estimate" = tol(coef(panel)[["age"]],
74                                                 "within slope after large offsets", "1e-9")),
75       "Entity demeaning should cancel large age/response offsets before OLS.")

Whole script generators/r/regression_advanced_hard.R


Regression II large-offset numeric STAT-REG-HARD-ADV-PLS-SCORE-001

Case STAT-REG-HARD-ADV-PLS-SCORE-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/MASS/pls
Citation
R nlme::lme, plm::plm, stats::lm/nls, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_hard.R · sha256 57a7311aff0ed329
Dataset
reference_data/datasets/REG_ADV_HARD_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/predictedtolerance5.00836069881916055.0083606987632125.59E-1110passed
/29/predictedtolerance62.9925728030502862.992572802424396.26E-1011passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

114  pls_cfg$artifact$x_means <- list(mean(pls_data$x))
115  pls_cfg$artifact$coefficients <- list(as.double(stats::coef(pls_fit, ncomp = 1L))[1L])
116  pls_cfg$artifact$intercept <- mean(pls_data$y)
117  pls_pred <- as.double(stats::predict(pls_fit, newdata = pls_data, ncomp = 1L))
118  emit("STAT-REG-HARD-ADV-PLS-SCORE-001", "STAT-REG-PLS-SCORE-001",
119       "REG_ADV_HARD_PLS", pls_cfg,
120       list("/0/predicted" = tol(pls_pred[[1L]],
121                                  "PLS prediction after ten-million-unit x offset", "1e-9"),
122            "/29/predicted" = tol(pls_pred[[30L]],
123                                   "last PLS prediction after large offset", "1e-9")),
124       "R pls::plsr centers the large predictor before extracting the original-unit coefficient.")

Whole script generators/r/regression_advanced_hard.R


Regression II large-offset numeric STAT-REG-HARD-ADV-SELECTION-001

Case STAT-REG-HARD-ADV-SELECTION-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/MASS/pls
Citation
R nlme::lme, plm::plm, stats::lm/nls, MASS::lda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_hard.R · sha256 57a7311aff0ed329
Dataset
reference_data/datasets/REG_ADV_HARD_ORTH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance0.660185185188273180.66018518518507113.20E-1211passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

76  ols <- stats::lm(distance ~ age, data = orth)
77  emit("STAT-REG-HARD-ADV-SELECTION-001", "STAT-REG-SELECTION-001",
78       "REG_ADV_HARD_ORTH", list(response = "distance",
79                                 candidate_predictors = list("age")),
80       list("/coefficients/1/coefficient" = tol(coef(ols)[["age"]],
81                                                  "large-offset OLS slope", "1e-9")),
82       "One strong candidate is retained; R lm is the independent refit.")

Whole script generators/r/regression_advanced_hard.R


Extreme scale STAT-REG-HARD-BINARY-001

Case STAT-REG-HARD-BINARY-001 · status: passed

Reference

Evidence type
R
Source
R scaled-data models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_hard.R · sha256 6c8d7cf7a33ed329
Dataset
reference_data/datasets/REG_SCALE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance722.22061481396645722.22061481396661.13E-1315passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

51  b <- stats::glm(I(binary == "event") ~ x, family = stats::binomial(), data = d)
52  emit("STAT-REG-HARD-BINARY-001",
53       "tabelz.analysis.logistic_regression:compute_binary_logistic_regression",
54       "logisticRegressionMenuItem",
55       list(outcome_column = "binary", event_value = "event", predictors = list("x")),
56       list("/coefficients/1/coefficient" = tol(coef(b)[["x"]], "rescaled logit slope")),
57       list("The binomial logit slope changes inversely with predictor units."))

Whole script generators/r/regression_hard.R


Extreme scale STAT-REG-HARD-EFFECTS-001

Case STAT-REG-HARD-EFFECTS-001 · status: passed

Reference

Evidence type
R
Source
R scaled-data models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_hard.R · sha256 6c8d7cf7a33ed329
Dataset
reference_data/datasets/REG_SCALE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/xtolerance13976708874.54239113976708874.54249.36E-615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

109  e <- stats::lm(y ~ x + group, data = d)
110  emit("STAT-REG-HARD-EFFECTS-001", "tabelz.executor.linear_effects",
111       "linearEffectsMenuItem",
112       list(response_column = "y", factor_columns = list("group"), covariate_columns = list("x")),
113       list("/coefficients/x" = tol(coef(e)[["x"]], "rescaled effects slope", "1e-9")),
114       list("Treatment factor contrasts; large response and small predictor units."))

Whole script generators/r/regression_hard.R


Extreme scale STAT-REG-HARD-GENERAL-001

Case STAT-REG-HARD-GENERAL-001 · status: passed

Reference

Evidence type
R
Source
R scaled-data models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_hard.R · sha256 6c8d7cf7a33ed329
Dataset
reference_data/datasets/REG_SCALE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/xtolerance13976708874.54239113976708874.54249.36E-615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

100  g <- stats::lm(y ~ x + group, data = d,
101                 contrasts = list(group = stats::contr.sum(3)))
102  emit("STAT-REG-HARD-GENERAL-001", "tabelz.executor.general_linear",
103       "generalLinearMenuItem",
104       list(response_column = "y", factor_columns = list("group"), covariate_columns = list("x"),
105            test_type = "III"),
106       list("/coefficients/x" = tol(coef(g)[["x"]], "rescaled general-linear slope", "1e-9")),
107       list("Sum-to-zero factor contrasts; large response and small predictor units."))

Whole script generators/r/regression_hard.R


Extreme scale STAT-REG-HARD-NB-001

Case STAT-REG-HARD-NB-001 · status: passed

Reference

Evidence type
R
Source
R scaled-data models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_hard.R · sha256 6c8d7cf7a33ed329
Dataset
reference_data/datasets/REG_SCALE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance1888.53010916972631888.53010916983771.11E-1013passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

92  nb <- MASS::glm.nb(nb_count ~ x + offset(log(exposure)), data = d,
93                     control = stats::glm.control(maxit = 1000, epsilon = 1e-12))
94  emit("STAT-REG-HARD-NB-001", "tabelz.analysis.negative_binomial:compute_negative_binomial",
95       "nbRegressionMenuItem",
96       list(count = "nb_count", predictors = list("x"), exposure = "exposure"),
97       list("/coefficients/1/coefficient" = tol(coef(nb)[["x"]], "rescaled NB2 slope")),
98       list("MASS glm.nb theta is inverse alpha; compare only coefficient here."))

Whole script generators/r/regression_hard.R


Extreme scale STAT-REG-HARD-NOMINAL-001

Case STAT-REG-HARD-NOMINAL-001 · status: passed

Reference

Evidence type
R
Source
R scaled-data models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_hard.R · sha256 6c8d7cf7a33ed329
Dataset
reference_data/datasets/REG_SCALE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance-72.643547352671533-72.64354791586195.63E-78passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

76  n <- nnet::multinom(nom ~ x, data = core, trace = FALSE, maxit = 1000, reltol = 1e-15)
77  emit("STAT-REG-HARD-NOMINAL-001",
78       "tabelz.analysis.nominal_logistic:compute_nominal_logistic_regression",
79       "nominalRegressionMenuItem",
80       list(outcome_column = "nominal", reference_value = "A", predictors = list("x")),
81       list("/coefficients/1/coefficient" = tol(1e4 * coef(n)["B", "x"], "REG_CORE nominal slope times 1e4 (exact rescaling)")),
82       list("R nnet B-versus-A log relative-risk slope; reference category A."))

Whole script generators/r/regression_hard.R


Extreme scale STAT-REG-HARD-OLS-001

Case STAT-REG-HARD-OLS-001 · status: passed

Reference

Evidence type
R
Source
R scaled-data models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_hard.R · sha256 6c8d7cf7a33ed329
Dataset
reference_data/datasets/REG_SCALE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/estimatetolerance13987626254.07081213987626254.0708229.76E-615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

37  fit <- stats::lm(y ~ x + z, data = d)
38  emit("STAT-REG-HARD-OLS-001", "tabelz.analysis.fit_y_by_x:compute_fits",
39       "fitYByXMenuItem", list(y_columns = list("y"), x_columns = list("x", "z")),
40       list("/responses/0/parameters/1/estimate" = tol(coef(fit)[["x"]], "x slope", "1e-9")),
41       list("OLS slope changes by 1e10 under the opposing x/y unit scales."))

Whole script generators/r/regression_hard.R


Extreme scale STAT-REG-HARD-ORDINAL-001

Case STAT-REG-HARD-ORDINAL-001 · status: passed

Reference

Evidence type
R
Source
R scaled-data models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_hard.R · sha256 6c8d7cf7a33ed329
Dataset
reference_data/datasets/REG_SCALE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/coefficienttolerance379.42551798623015379.425643078711971.25E-46passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

63  core <- read_dataset("REG_CORE")
64  core$ord <- ordered(core$ordinal, levels = c("low", "middle", "high"))
65  core$nom <- factor(core$nominal, levels = c("A", "B", "C"))
66  o <- MASS::polr(ord ~ x, data = core, method = "logistic", Hess = TRUE,
67                  control = list(maxit = 1000, reltol = 1e-15))
68  emit("STAT-REG-HARD-ORDINAL-001",
69       "tabelz.analysis.ordinal_logistic:compute_ordinal_logistic_regression",
70       "ordinalRegressionMenuItem",
71       list(outcome_column = "ordinal", predictors = list("x"),
72            level_order = list("low", "middle", "high")),
73       list("/coefficients/0/coefficient" = tol(1e4 * coef(o)[["x"]], "REG_CORE ordered-logit slope times 1e4 (exact rescaling)")),
74       list("MASS polr uses the same low-to-high slope sign."))

Whole script generators/r/regression_hard.R


Extreme scale STAT-REG-HARD-POISSON-001

Case STAT-REG-HARD-POISSON-001 · status: passed

Reference

Evidence type
R
Source
R scaled-data models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_hard.R · sha256 6c8d7cf7a33ed329
Dataset
reference_data/datasets/REG_SCALE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance1851.93666890311011851.93666876707311.36E-710passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

84  p <- stats::glm(count ~ x + offset(log(exposure)), data = d, family = stats::poisson())
85  emit("STAT-REG-HARD-POISSON-001",
86       "tabelz.analysis.poisson_regression:compute_poisson_regression",
87       "poissonRegressionMenuItem",
88       list(count_column = "count", predictors = list("x"), exposure_column = "exposure"),
89       list("/coefficients/1/coefficient" = tol(coef(p)[["x"]], "rescaled Poisson slope")),
90       list("Log exposure is an offset; x slope changes inversely with x units."))

Whole script generators/r/regression_hard.R


Extreme scale STAT-REG-HARD-VIF-001

Case STAT-REG-HARD-VIF-001 · status: passed

Reference

Evidence type
R
Source
R scaled-data models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_hard.R · sha256 6c8d7cf7a33ed329
Dataset
reference_data/datasets/REG_SCALE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance1.00010902110332341.00010902110332344.78E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

43  vif <- 1 / (1 - summary(stats::lm(x ~ z, data = d))$r.squared)
44  emit("STAT-REG-HARD-VIF-001", "tabelz.analysis.fit_y_by_x:compute_fit_vif",
45       "fitYByXMenuItem",
46       list(y_columns = list("y"), x_columns = list("x", "z"), missing = "drop",
47            model_type = "linear", degree = 2L, interactions = FALSE, split_percent = 100L),
48       list("/0" = tol(vif, "scale-invariant VIF", "1e-9")),
49       list("Auxiliary R lm R-squared is unchanged by linear rescaling of x."))

Whole script generators/r/regression_hard.R


Regression completion STAT-REG-INTERACT-GENERAL-TYPEII-001

Case STAT-REG-INTERACT-GENERAL-TYPEII-001 · status: passed

Reference

Evidence type
R
Source
R regression fit
Citation
R stats::glm/lm and car::Anova help pages; Tabelz manual chapters 07 and 08
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_lastmile.R · sha256 512832af1db86a5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/term_tests/0/f_statistictolerance27.26621705324004827.266217053240022.83E-1414passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

81  interacting <- stats::lm(y ~ group * x, data = d,
82                           contrasts = list(group = stats::contr.sum(3)))
83  type2 <- car::Anova(interacting, type = 2)
84  emit("STAT-REG-INTERACT-GENERAL-TYPEII-001", general_entry,
85       "generalLinearMenuItem", "axis_coverage", "REG_CORE",
86       list(response_column = "y", factor_columns = list("group"),
87            covariate_columns = list("x"), interactions = list(list("group", "x")),
88            test_type = "II"),
89       list("/term_tests/0/f_statistic" = tol(type2["group", "F value"],
90                                               "Type II group effect in interaction model")),
91       list("Manual Type II preserves marginality: group main effect is not adjusted for group:x interaction.",
92            "R car::Anova(type=2) supplies the comparison on the same sum-coded design."))

Whole script generators/r/regression_lastmile.R


Regression II invalid design STAT-REG-INVALID-ADV-DISCRIMINANT-CONFIG-001

Case STAT-REG-INVALID-ADV-DISCRIMINANT-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-DISCRIMINANT-MIN-001

Case STAT-REG-INVALID-ADV-DISCRIMINANT-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-DISCRIMINANT-SCORE-CONFIG-001

Case STAT-REG-INVALID-ADV-DISCRIMINANT-SCORE-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-DISCRIMINANT-SCORE-ROLE-001

Case STAT-REG-INVALID-ADV-DISCRIMINANT-SCORE-ROLE-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

95  lda$artifact$predictors <- list("absent_predictor")
96  emit("STAT-REG-INVALID-ADV-DISCRIMINANT-SCORE-ROLE-001",
97       "STAT-REG-DISCRIMINANT-SCORE-LDA-001", lda, "REG_DISCRIMINANT",
98       "Unknown scoring predictor", "Stored artifact requires a predictor absent from the table.")

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-MIXED-CONFIG-001

Case STAT-REG-INVALID-ADV-MIXED-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-MIXED-MIN-001

Case STAT-REG-INVALID-ADV-MIXED-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-MIXED-STRUCT-CONFIG-001

Case STAT-REG-INVALID-ADV-MIXED-STRUCT-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-MIXED-STRUCT-MIN-001

Case STAT-REG-INVALID-ADV-MIXED-STRUCT-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-NONLINEAR-CONFIG-001

Case STAT-REG-INVALID-ADV-NONLINEAR-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_DECAY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-NONLINEAR-MIN-001

Case STAT-REG-INVALID-ADV-NONLINEAR-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-PANEL-CONFIG-001

Case STAT-REG-INVALID-ADV-PANEL-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-PANEL-MIN-001

Case STAT-REG-INVALID-ADV-PANEL-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-PLS-CONFIG-001

Case STAT-REG-INVALID-ADV-PLS-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-PLS-MIN-001

Case STAT-REG-INVALID-ADV-PLS-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-PLS-SCORE-CONFIG-001

Case STAT-REG-INVALID-ADV-PLS-SCORE-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-PLS-SCORE-ROLE-001

Case STAT-REG-INVALID-ADV-PLS-SCORE-ROLE-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

100  pls$artifact$predictors <- list("absent_predictor")
101  emit("STAT-REG-INVALID-ADV-PLS-SCORE-ROLE-001", "STAT-REG-PLS-SCORE-001",
102       pls, "REG_PLS", "Unknown scoring predictor",
103       "Stored artifact requires a predictor absent from the table.")

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-SELECTION-CONFIG-001

Case STAT-REG-INVALID-ADV-SELECTION-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-SELECTION-MIN-001

Case STAT-REG-INVALID-ADV-SELECTION-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-TREE-CLASS-CONFIG-001

Case STAT-REG-INVALID-ADV-TREE-CLASS-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-TREE-CLASS-MIN-001

Case STAT-REG-INVALID-ADV-TREE-CLASS-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-TREE-REG-CONFIG-001

Case STAT-REG-INVALID-ADV-TREE-REG-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-TREE-REG-MIN-001

Case STAT-REG-INVALID-ADV-TREE-REG-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-TREE-SCORE-CLASS-CONFIG-001

Case STAT-REG-INVALID-ADV-TREE-SCORE-CLASS-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-TREE-SCORE-CLASS-ROLE-001

Case STAT-REG-INVALID-ADV-TREE-SCORE-CLASS-ROLE-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

105  class_op$score_name <- "__tabelz_reserved"
106  emit("STAT-REG-INVALID-ADV-TREE-SCORE-CLASS-ROLE-001",
107       "STAT-REG-TREE-SCORE-CLASS-001", class_op, "REG_TREE", "score-column",
108       "Reserved internal score-column names are not user columns.")

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-TREE-SCORE-REG-CONFIG-001

Case STAT-REG-INVALID-ADV-TREE-SCORE-REG-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-TREE-SCORE-REG-ROLE-001

Case STAT-REG-INVALID-ADV-TREE-SCORE-REG-ROLE-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

110  reg_op$score_name <- "__tabelz_reserved"
111  emit("STAT-REG-INVALID-ADV-TREE-SCORE-REG-ROLE-001",
112       "STAT-REG-TREE-SCORE-REG-001", reg_op, "REG_TREE", "score-column",
113       "Reserved internal score-column names are not user columns.")

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-VALIDATED-CONFIG-001

Case STAT-REG-INVALID-ADV-VALIDATED-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Regression II invalid design STAT-REG-INVALID-ADV-VALIDATED-MIN-001

Case STAT-REG-INVALID-ADV-VALIDATED-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual input/design condition
Citation
Tabelz manual chapters for mixed, panel, nonlinear, validated and tree regression
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_invalid.R · sha256 9a17aa6da915a401
Dataset
reference_data/datasets/REG_ADV_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

Whole script generators/r/regression_advanced_invalid.R


Invalid design: STAT-REG-INVALID-BINARY-CONFIG-001

Case STAT-REG-INVALID-BINARY-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

84  emit("STAT-REG-INVALID-BINARY-CONFIG-001", binary, "logisticRegressionMenuItem", "REG_CORE",
85       list(outcome_column = "binary", event_value = "event", predictors = list("x"), threshold = 1),
86       "ValueError", "threshold",
87       list("Manual: classification threshold is strictly between zero and one."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-BINARY-MIN-001

Case STAT-REG-INVALID-BINARY-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

80  emit("STAT-REG-INVALID-BINARY-MIN-001", binary, "logisticRegressionMenuItem", "REG_ONE",
81       list(outcome_column = "binary", event_value = "event", predictors = list("x")),
82       "ValueError", "exactly two",
83       list("A one-row population cannot contain two observed outcome categories."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-EFFECTS-CONFIG-001

Case STAT-REG-INVALID-EFFECTS-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

140  emit("STAT-REG-INVALID-EFFECTS-CONFIG-001", effects, "linearEffectsMenuItem", "REG_CORE",
141       list(response_column = "y", covariate_columns = list("x"), confidence_level = 1),
142       "ValidationError", "confidence_level",
143       list("Confidence level one violates the strict linear-effects configuration bound."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-EFFECTS-MIN-001

Case STAT-REG-INVALID-EFFECTS-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

136  emit("STAT-REG-INVALID-EFFECTS-MIN-001", effects, "linearEffectsMenuItem", "REG_ONE",
137       list(response_column = "y", covariate_columns = list("x")),
138       "LinearEffectsExecutionError", "at least two",
139       list("The manual requires positive residual df; one row is below even the two-observation gate."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-GENERAL-CONFIG-001

Case STAT-REG-INVALID-GENERAL-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

131  emit("STAT-REG-INVALID-GENERAL-CONFIG-001", general, "generalLinearMenuItem", "REG_CORE",
132       list(response_column = "y", covariate_columns = list("x"), confidence_level = 1),
133       "ValidationError", "confidence_level",
134       list("Confidence level one violates the strict general-linear configuration bound."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-GENERAL-MIN-001

Case STAT-REG-INVALID-GENERAL-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

127  emit("STAT-REG-INVALID-GENERAL-MIN-001", general, "generalLinearMenuItem", "REG_ONE",
128       list(response_column = "y", covariate_columns = list("x")),
129       "GeneralLinearExecutionError", "at least two",
130       list("The manual requires positive residual df; one row is below even the two-observation gate."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-NB-CONFIG-001

Case STAT-REG-INVALID-NB-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

122  emit("STAT-REG-INVALID-NB-CONFIG-001", nb, "nbRegressionMenuItem", "REG_CORE",
123       list(count = "nb_count", predictors = list("x"), confidence_level = 1),
124       "ValueError", "confidence_level",
125       list("Confidence level one violates the strict NB2 configuration bound."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-NB-MIN-001

Case STAT-REG-INVALID-NB-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

117  emit("STAT-REG-INVALID-NB-MIN-001", nb, "nbRegressionMenuItem", "REG_TWO",
118       list(count = "nb_count", predictors = list("x")),
119       "ValueError", "More complete rows",
120       list("Two rows cannot identify NB2 intercept, slope and dispersion.",
121            "Until 2026-09-25 this case used REG_ONE, where a constant response is refused first."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-NOMINAL-CONFIG-001

Case STAT-REG-INVALID-NOMINAL-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

102  emit("STAT-REG-INVALID-NOMINAL-CONFIG-001", nominal, "nominalRegressionMenuItem", "REG_CORE",
103       list(outcome_column = "nominal", predictors = list("x"), confidence_level = 1),
104       "ValueError", "Confidence level",
105       list("Confidence level one lies outside the supported half-open interval."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-NOMINAL-MIN-001

Case STAT-REG-INVALID-NOMINAL-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 98  emit("STAT-REG-INVALID-NOMINAL-MIN-001", nominal, "nominalRegressionMenuItem", "REG_ONE",
 99       list(outcome_column = "nominal", predictors = list("x")),
100       "ValueError", "three",
101       list("A one-row population cannot supply three nominal categories."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-OLS-CONFIG-001

Case STAT-REG-INVALID-OLS-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

63  emit("STAT-REG-INVALID-OLS-CONFIG-001", fit, fit_menu, "REG_CORE",
64       list(y_columns = list("y"), x_columns = list("x"), alpha_level = 2),
65       "ValueError", "alpha level",
66       list("A significance level of 2 lies outside the open unit interval."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-OLS-MIN-001

Case STAT-REG-INVALID-OLS-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

60  emit("STAT-REG-INVALID-OLS-MIN-001", fit, fit_menu, "REG_ONE",
61       list(y_columns = list("y"), x_columns = list("x")), "ValueError", "row",
62       list("One source row cannot supply positive residual degrees of freedom for intercept and slope."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-ORDINAL-CONFIG-001

Case STAT-REG-INVALID-ORDINAL-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

93  emit("STAT-REG-INVALID-ORDINAL-CONFIG-001", ordinal, "ordinalRegressionMenuItem", "REG_CORE",
94       list(outcome_column = "ordinal", predictors = list("x"), confidence_level = 1),
95       "ValueError", "Confidence level",
96       list("Confidence level one lies outside the supported half-open interval."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-ORDINAL-MIN-001

Case STAT-REG-INVALID-ORDINAL-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

89  emit("STAT-REG-INVALID-ORDINAL-MIN-001", ordinal, "ordinalRegressionMenuItem", "REG_ONE",
90       list(outcome_column = "ordinal", predictors = list("x")),
91       "ValueError", "three",
92       list("A one-row population cannot supply three ordered outcome levels."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-POISSON-CONFIG-001

Case STAT-REG-INVALID-POISSON-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

112  emit("STAT-REG-INVALID-POISSON-CONFIG-001", poisson, "poissonRegressionMenuItem", "REG_CORE",
113       list(count_column = "count", predictors = list("x"), confidence_level = 1),
114       "ValueError", "Confidence level",
115       list("Confidence level one lies outside the supported half-open interval."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-POISSON-MIN-001

Case STAT-REG-INVALID-POISSON-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

107  emit("STAT-REG-INVALID-POISSON-MIN-001", poisson, "poissonRegressionMenuItem", "REG_TWO",
108       list(count_column = "count", predictors = list("x")),
109       "ValueError", "More complete rows",
110       list("Two rows cannot fit an intercept and a predictor with positive residual degrees of freedom.",
111            "Until 2026-09-25 this case used REG_ONE, where a constant predictor is refused first."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-VIF-CONFIG-001

Case STAT-REG-INVALID-VIF-CONFIG-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

74  vif_invalid <- vif_config
75  vif_invalid$missing <- "bad"
76  emit("STAT-REG-INVALID-VIF-CONFIG-001", vif, fit_menu, "REG_CORE",
77       vif_invalid, "ValueError", "missing",
78       list("A nonexistent missing-value strategy is not an estimator."))

Whole script generators/r/regression_invalid.R


Invalid design: STAT-REG-INVALID-VIF-MIN-001

Case STAT-REG-INVALID-VIF-MIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Manual design condition
Citation
Tabelz user manual, chapters/07-models-and-tests.tex, 07-linear-effects.tex and 08-general-linear.tex
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_invalid.R · sha256 16490144367b4b31
Dataset
reference_data/datasets/REG_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

68  vif_config <- list(y_columns = list("y"), x_columns = list("x", "z"),
69                     missing = "drop", model_type = "linear", degree = 2L,
70                     interactions = FALSE, split_percent = 100L)
71  emit("STAT-REG-INVALID-VIF-MIN-001", vif, fit_menu, "REG_ONE",
72       vif_config, "ValueError", "row",
73       list("One row cannot estimate the auxiliary regression for a VIF."))

Whole script generators/r/regression_invalid.R


Nearest-neighbour classification with city-block distance

Case STAT-REG-KNN-CITYBLOCK-001 · status: passed

Reference

Evidence type
Analytical
Source
separable-cluster argument checked with FNN::knn / FNN::knn.reg
Citation
Cover and Hart (1967) Nearest neighbor pattern classification, IEEE Trans. Inf. Theory 13; R FNN package documentation (knn, knn.reg).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · FNN 1.1.4.1, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_nearest_neighbour.R · sha256 112a4856125b2f29
Dataset
reference_data/datasets/REG_KNN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact18180.00E+2passed
/validation_rowsexact660.00E+2passed
/test_rowsexact660.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact000.00E+2passed
/chosen_kexact110.00E+2passed
/validation_errortolerance00.00.00E+2passed
/test_errortolerance00.00.00E+2passed
/error_nameexactmisclassification ratemisclassification ratepassed
/grid/0/kexact110.00E+2passed
/grid/0/validation_errortolerance00.00.00E+2passed
/grid/1/kexact330.00E+2passed
/grid/1/validation_errortolerance00.00.00E+2passed
/importance/0/featureexactf1f1passed
/importance/0/reduced_validation_errortolerance00.00.00E+2passed
/importance/0/changetolerance00.00.00E+2passed
/importance/1/featureexactf2f2passed
/importance/1/changetolerance00.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

117  outputs <- common("misclassification rate")
118  emit("STAT-REG-KNN-CITYBLOCK-001", "Nearest-neighbour classification with city-block distance", "axis_coverage",
119       list(target = "outcome", task = "classification", features = list("f1", "f2"),
120            distance = "cityblock", k_values = list(1L, 3L)),
121       outputs, list(separation_note, "City-block neighbourhoods coincide with the Euclidean ones here because the between-cluster gap dominates on each feature.", split_note))

Whole script generators/r/regression_nearest_neighbour.R


Nearest-neighbour classification of separable clusters

Case STAT-REG-KNN-CLASS-001 · status: passed

Reference

Evidence type
Analytical
Source
separable-cluster argument checked with FNN::knn / FNN::knn.reg
Citation
Cover and Hart (1967) Nearest neighbor pattern classification, IEEE Trans. Inf. Theory 13; R FNN package documentation (knn, knn.reg).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · FNN 1.1.4.1, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_nearest_neighbour.R · sha256 112a4856125b2f29
Dataset
reference_data/datasets/REG_KNN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact18180.00E+2passed
/validation_rowsexact660.00E+2passed
/test_rowsexact660.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact000.00E+2passed
/chosen_kexact110.00E+2passed
/validation_errortolerance00.00.00E+2passed
/test_errortolerance00.00.00E+2passed
/error_nameexactmisclassification ratemisclassification ratepassed
/grid/0/kexact110.00E+2passed
/grid/0/validation_errortolerance00.00.00E+2passed
/grid/1/kexact330.00E+2passed
/grid/1/validation_errortolerance00.00.00E+2passed
/importance/0/featureexactf1f1passed
/importance/0/reduced_validation_errortolerance00.00.00E+2passed
/importance/0/changetolerance00.00.00E+2passed
/importance/1/featureexactf2f2passed
/importance/1/changetolerance00.00.00E+2passed
/classessequence['fail', 'pass']['fail', 'pass']passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

109  outputs <- common("misclassification rate")
110  outputs[["/classes"]] <- list(comparison = "sequence", values = list("fail", "pass"),
111                                rationale = "classes in code-point order")
112  emit("STAT-REG-KNN-CLASS-001", "Nearest-neighbour classification of separable clusters", "nominal",
113       list(target = "outcome", task = "classification", features = list("f1", "f2"),
114            distance = "euclidean", k_values = list(1L, 3L)),
115       outputs, list(separation_note, split_note))

Whole script generators/r/regression_nearest_neighbour.R


Nearest-neighbour regression with constant within-cluster targets

Case STAT-REG-KNN-REG-001 · status: passed

Reference

Evidence type
Analytical
Source
separable-cluster argument checked with FNN::knn / FNN::knn.reg
Citation
Cover and Hart (1967) Nearest neighbor pattern classification, IEEE Trans. Inf. Theory 13; R FNN package documentation (knn, knn.reg).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · FNN 1.1.4.1, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_nearest_neighbour.R · sha256 112a4856125b2f29
Dataset
reference_data/datasets/REG_KNN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact18180.00E+2passed
/validation_rowsexact660.00E+2passed
/test_rowsexact660.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact000.00E+2passed
/chosen_kexact110.00E+2passed
/validation_errortolerance00.00.00E+2passed
/test_errortolerance00.00.00E+2passed
/error_nameexactmean squared errormean squared errorpassed
/grid/0/kexact110.00E+2passed
/grid/0/validation_errortolerance00.00.00E+2passed
/grid/1/kexact330.00E+2passed
/grid/1/validation_errortolerance00.00.00E+2passed
/importance/0/featureexactf1f1passed
/importance/0/reduced_validation_errortolerance00.00.00E+2passed
/importance/0/changetolerance00.00.00E+2passed
/importance/1/featureexactf2f2passed
/importance/1/changetolerance00.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

123  outputs <- common("mean squared error")
124  emit("STAT-REG-KNN-REG-001", "Nearest-neighbour regression with constant within-cluster targets", "axis_coverage",
125       list(target = "hours", task = "regression", features = list("f1", "f2"),
126            distance = "euclidean", k_values = list(1L, 3L)),
127       outputs, list(separation_note, "Repair hours are 0 in one cluster and 6 in the other, so the mean of within-cluster neighbours reproduces the target exactly and the mean squared error is 0.", split_note))

Whole script generators/r/regression_nearest_neighbour.R


STAT REG LOGLINEAR GENERAL 001

Case STAT-REG-LOGLINEAR-GENERAL-001 · status: passed

Reference

Evidence type
R
Source
R stats::glm(family = poisson) on the cell table with contr.treatment at the declared reference; rstandard/residuals for adjusted and deviance residuals; profile limits by uniroot on the constrained deviance; logit contrasts from the Poisson coefficients
Citation
R Core Team, stats package help: glm, rstandard.glm, residuals.glm, uniroot; Agresti (2013) Categorical Data Analysis, ch. 9 (loglinear models, sampling schemes) and 10 (logit-loglinear equivalence).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear.R · sha256 29efcb27edfe009b
Dataset
reference_data/datasets/REG_LOGLINEAR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/dfexact550.00E+2passed
/fit/rankexact770.00E+2passed
/fit/allowable_cellsexact12120.00E+2passed
/fit/structural_zero_cellsexact000.00E+2passed
/fit/deviancetolerance15.8773868450360315.8773868450360218.84E-1515passed
/fit/pearsontolerance13.44394755301195813.4439475530119678.53E-1515passed
/fit/deviance_ptail_probability0.00720299987756788110.0072029998775679143.29E-1714passed
/fit/pearson_ptail_probability0.0195553488951060520.0195553488951065.21E-1714passed
/fit/log_likelihoodtolerance-30.41969177532701-30.4196917753269961.42E-1415passed
/fit/boundaryexactFalseFalsepassed
/totaltolerance109109.00.00E+2passed
/fit/parameters/0/termexact(intercept)(intercept)passed
/fit/parameters/0/labelexactpassed
/fit/parameters/0/estimatetolerance1.85973245281426051.85973245281425986.93E-1615passed
/fit/parameters/0/standard_errortolerance0.286800033998153680.28680003399924141.09E-1211passed
/fit/parameters/0/z_valuetolerance6.4844220096090816.4844220095844862.46E-1111passed
/fit/parameters/0/p_valuetail_probability8.9072584526885216e-118.907258454141441e-111.45E-209passed
/fit/parameters/0/lowertolerance1.29761471541301621.29761471541088372.13E-1211passed
/fit/parameters/0/uppertolerance2.42185019021550472.4218501902176362.13E-1212passed
/fit/parameters/1/termexactAApassed
/fit/parameters/1/labelexactA=a1A=a1passed
/fit/parameters/1/estimatetolerance-0.84729786038720323-0.84729786038720284.46E-1615passed
/fit/parameters/1/standard_errortolerance0.487950036295847820.48795003647426661.78E-109passed
/fit/parameters/1/z_valuetolerance-1.7364438925330463-1.73644389189811536.35E-109passed
/fit/parameters/1/p_valuetail_probability0.082485377003684870.0824853771158651.12E-108passed
/fit/parameters/1/lowertolerance-1.803662357782077-1.8036623581317713.50E-109passed
/fit/parameters/1/uppertolerance0.10906663700767050.109066637357365443.50E-108passed
/fit/parameters/2/termexactBBpassed
/fit/parameters/2/labelexactB=b1B=b1passed
/fit/parameters/2/estimatetolerance0.405465108108165330.40546510810816494.46E-1614passed
/fit/parameters/2/standard_errortolerance0.345032779671177180.345032779671177245.66E-1715passed
/fit/parameters/2/z_valuetolerance1.17514952780597051.17514952780596921.35E-1514passed
/fit/parameters/2/p_valuetail_probability0.239934898752295680.239934898752296235.53E-1614passed
/fit/parameters/2/lowertolerance-0.27078671353308553-0.27078671353308626.69E-1614passed
/fit/parameters/2/uppertolerance1.08171692974941621.0817169297494161.77E-1615passed
/fit/parameters/3/termexactBBpassed
/fit/parameters/3/labelexactB=b2B=b2passed
/fit/parameters/3/estimatetolerance0.451985123743057990.451985123743056661.33E-1514passed
/fit/parameters/3/standard_errortolerance0.341881729378913920.34188172937891411.70E-1615passed
/fit/parameters/3/z_valuetolerance1.32205112149211821.32205112149211364.63E-1514passed
/fit/parameters/3/p_valuetail_probability0.186151128385110080.18615112838511151.42E-1514passed
/fit/parameters/3/lowertolerance-0.21809075281188239-0.218090752811884281.89E-1514passed
/fit/parameters/3/uppertolerance1.12206100029799851.12206100029799768.98E-1615passed
/fit/parameters/4/termexactCCpassed
/fit/parameters/4/labelexactC=c1C=c1passed
/fit/parameters/4/estimatetolerance0.165514438477573160.165514438477573473.07E-1614passed
/fit/parameters/4/standard_errortolerance0.19222162349810620.192221623503634255.53E-1210passed
/fit/parameters/4/z_valuetolerance0.861060454414504740.86106045438974332.48E-1110passed
/fit/parameters/4/p_valuetail_probability0.389204747915471720.389204747929108641.36E-1110passed
/fit/parameters/4/lowertolerance-0.21123302062853308-0.21123302063936761.08E-1110passed
/fit/parameters/4/uppertolerance0.542261897583679380.54226189759451461.08E-1110passed
/fit/parameters/5/termexactA*BA*Bpassed
/fit/parameters/5/labelexactA=a1, B=b1A=a1, B=b1passed
/fit/parameters/5/estimatetolerance1.02165124753198141.0216512475319814.01E-1615passed
/fit/parameters/5/standard_errortolerance0.570713838574260610.5707138387268051.53E-109passed
/fit/parameters/5/z_valuetolerance1.79012874487893711.7901287444004584.78E-109passed
/fit/parameters/5/p_valuetail_probability0.0734332167273604170.073433216804263087.69E-118passed
/fit/parameters/5/lowertolerance-0.096927321552175316-0.096927321851157712.99E-108passed
/fit/parameters/5/uppertolerance2.1402298166161382.14022981691511972.99E-109passed
/fit/parameters/6/termexactA*BA*Bpassed
/fit/parameters/6/labelexactA=a1, B=b2A=a1, B=b2passed
/fit/parameters/6/estimatetolerance0.800777844752310510.80077784475231148.86E-1614passed
/fit/parameters/6/standard_errortolerance0.575472702215068190.57547270236635131.51E-109passed
/fit/parameters/6/z_valuetolerance1.39151317112004441.3915131707542383.66E-109passed
/fit/parameters/6/p_valuetail_probability0.164069868453632330.164069868564480141.11E-109passed
/fit/parameters/6/lowertolerance-0.32712792567516635-0.327127925971674952.97E-109passed
/fit/parameters/6/uppertolerance1.92868361517978751.92868361547629792.97E-109passed
/fit/cells/0/observedtolerance1818.00.00E+2passed
/fit/cells/0/structuralexactFalseFalsepassed
/fit/cells/0/expectedtolerance13.5321100917431313.5321100917431171.26E-1415passed
/fit/cells/0/residualtolerance1.21456203596223871.21456203596224273.96E-1514passed
/fit/cells/0/adjusted_residualtolerance2.0427798006371522.0427798006547031.76E-1111passed
/fit/cells/0/deviance_residualtolerance1.15552998755385321.15552998755385542.19E-1514passed
/fit/cells/1/observedtolerance77.00.00E+2passed
/fit/cells/1/structuralexactFalseFalsepassed
/fit/cells/1/expectedtolerance11.4678899082568911.467889908256889.18E-1515passed
/fit/cells/1/residualtolerance-1.3193520789603079-1.31935207896030572.19E-1514passed
/fit/cells/1/adjusted_residualtolerance-2.0427798006372817-2.0427798006547071.74E-1111passed
/fit/cells/1/deviance_residualtolerance-1.4229577050451325-1.422957705045132.42E-1514passed
/fit/cells/2/observedtolerance1212.00.00E+2passed
/fit/cells/2/structuralexactFalseFalsepassed
/fit/cells/2/expectedtolerance11.36697247706422211.3669724770642224.64E-1616passed
/fit/cells/2/residualtolerance0.187758751420332490.18775875142033252.54E-1816passed
/fit/cells/2/adjusted_residualtolerance0.308532095101449690.308532095103576432.13E-1211passed
/fit/cells/2/deviance_residualtolerance0.186055269854091210.18605526985409122.31E-1917passed
/fit/cells/3/observedtolerance99.00.00E+2passed
/fit/cells/3/structuralexactFalseFalsepassed
/fit/cells/3/expectedtolerance9.6330275229357829.6330275229357775.31E-1515passed
/fit/cells/3/residualtolerance-0.20395821019809099-0.203958210198089321.67E-1514passed
/fit/cells/3/adjusted_residualtolerance-0.30853209510146967-0.3085320951035762.11E-1211passed
/fit/cells/3/deviance_residualtolerance-0.20625552433777888-0.206255524337778883.84E-1816passed
/fit/cells/4/observedtolerance66.00.00E+2passed
/fit/cells/4/structuralexactFalseFalsepassed
/fit/cells/4/expectedtolerance3.24770642201834743.2477064220183484.05E-1615passed
/fit/cells/4/residualtolerance1.52723678083335071.52723678083335024.62E-1615passed
/fit/cells/4/adjusted_residualtolerance2.31968630321318822.3196863031465626.66E-1110passed
/fit/cells/4/deviance_residualtolerance1.36423535503145141.36423535503145121.87E-1615passed
/fit/cells/5/observedtolerance00.00.00E+2passed
/fit/cells/5/structuralexactFalseFalsepassed
/fit/cells/5/expectedtolerance2.75229357798165092.75229357798165044.81E-1615passed
/fit/cells/5/residualtolerance-1.6590037908279929-1.65900379082799272.47E-1615passed
/fit/cells/5/adjusted_residualtolerance-2.3196863032133228-2.319686303146566.68E-1110passed
/fit/cells/5/deviance_residualtolerance-2.3461856610173251-2.34618566101732464.80E-1615passed
/fit/cells/6/observedtolerance1010.00.00E+2passed
/fit/cells/6/structuralexactFalseFalsepassed
/fit/cells/6/expectedtolerance11.36697247706422711.366972477064211.61E-1414passed
/fit/cells/6/residualtolerance-0.40545005741492363-0.40545005741491924.45E-1513passed
/fit/cells/6/adjusted_residualtolerance-0.66625046623356976-0.66625046623815534.59E-1211passed
/fit/cells/6/deviance_residualtolerance-0.4140130746840992-0.41401307468409385.38E-1513passed
/fit/cells/7/observedtolerance1111.00.00E+2passed
/fit/cells/7/structuralexactFalseFalsepassed
/fit/cells/7/expectedtolerance9.63302752293578569.6330275229357731.25E-1414passed
/fit/cells/7/residualtolerance0.440431497384280.44043149738428434.28E-1514passed
/fit/cells/7/adjusted_residualtolerance0.666250466233603180.6662504662381634.56E-1211passed
/fit/cells/7/deviance_residualtolerance0.430586468833327520.4305864688333284.98E-1614passed
/fit/cells/8/observedtolerance88.00.00E+2passed
/fit/cells/8/structuralexactFalseFalsepassed
/fit/cells/8/expectedtolerance11.90825688073394811.9082568807339212.69E-1414passed
/fit/cells/8/residualtolerance-1.1325542233003489-1.13255422330034256.40E-1514passed
/fit/cells/8/adjusted_residualtolerance-1.8717199672159517-1.8717199672296111.37E-1111passed
/fit/cells/8/deviance_residualtolerance-1.2049341010142423-1.20493410101423566.71E-1514passed
/fit/cells/9/observedtolerance1414.00.00E+2passed
/fit/cells/9/structuralexactFalseFalsepassed
/fit/cells/9/expectedtolerance10.09174311926605910.0917431192660362.27E-1414passed
/fit/cells/9/residualtolerance1.23026879220934181.23026879220935058.68E-1514passed
/fit/cells/9/adjusted_residualtolerance1.87171996721605941.8717199672296311.36E-1111passed
/fit/cells/9/deviance_residualtolerance1.16146430171333351.16146430171333955.98E-1514passed
/fit/cells/10/observedtolerance55.00.00E+2passed
/fit/cells/10/structuralexactFalseFalsepassed
/fit/cells/10/expectedtolerance7.57798165137614277.5779816513761393.59E-1515passed
/fit/cells/10/residualtolerance-0.93648979085397577-0.93648979085397451.22E-1514passed
/fit/cells/10/adjusted_residualtolerance-1.4810944017760885-1.48109440178241016.32E-1211passed
/fit/cells/10/deviance_residualtolerance-0.9989361912429392-0.99893619124293838.91E-1615passed
/fit/cells/11/observedtolerance99.00.00E+2passed
/fit/cells/11/structuralexactFalseFalsepassed
/fit/cells/11/expectedtolerance6.42201834862385026.4220183486238464.40E-1515passed
/fit/cells/11/residualtolerance1.01728830302967641.01728830302967872.25E-1514passed
/fit/cells/11/adjusted_residualtolerance1.48109440177618021.48109440178241956.24E-1211passed
/fit/cells/11/deviance_residualtolerance0.958590078523120060.9585900785231218.88E-1615passed
/fit/generators/0exactA*BA*Bpassed
/fit/generators/1exactCCpassed
/reference_levels/0/1exacta2a2passed
/reference_levels/1/1exactb3b3passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

35  emit <- function(id, menu, kind, config, outputs, notes) {
36    write_expected("regression", id, list(
37      schema_version = "2.0.0", case_id = id, outputs = outputs,
38      expected_error = NULL, deferred_outputs = list(), generator_run = run))
39    write_case("regression", id, list(
40      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
41      entry_point = "tabelz.analysis.loglinear:compute_loglinear",
42      menu_items = list(menu), title = gsub("-", " ", id),
43      case_kind = kind,
44      dataset = list(path = "reference_data/datasets/REG_LOGLINEAR_01.json", sha256 = ""),
45      excluded_rows = list(), configuration = config, "repeat" = 1L,
46      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"), sha256 = ""),
47      oracle = list(
48        oracle_class = "independent_implementation",
49        source = "R stats::glm(family = poisson) on the cell table with contr.treatment at the declared reference; rstandard/residuals for adjusted and deviance residuals; profile limits by uniroot on the constrained deviance; logit contrasts from the Poisson coefficients",
50        citation = "R Core Team, stats package help: glm, rstandard.glm, residuals.glm, uniroot; Agresti (2013) Categorical Data Analysis, ch. 9 (loglinear models, sampling schemes) and 10 (logit-loglinear equivalence).",
51        generator = list(path = "generators/r/regression_loglinear.R", sha256 = ""),
52        independent = TRUE,
53        convention_notes = c(
54          notes,
55          "log mu_c = o_c + x_c' lambda over the allowable cells in reference coding (last level by default); G^2 and X^2 on df = allowable cells - rank; the Poisson, multinomial and product-multinomial fitted counts coincide and each scheme's own log-likelihood is reported (docs/manual/chapters/86-regression-and-count-models.tex:171-197).",
56          "Adjusted residual (y - mu)/sqrt(mu (1 - h)) with the Poisson leverage (R rstandard type pearson); deviance residual sign(y - mu) sqrt(d_c).",
57          "The logit model fixes the explanatory factors' interaction; each logit parameter equals the loglinear parameter whose term contains the dependent factor (manual lines 193-197; chapters/106-loglinear-extensions.tex:74-118).",
58          "Expected computed from the exported CSV."
59        )),
60      supported = TRUE, unsupported_reason = NULL,
61      requirement_ids = list(if (menu == "loglinearLogitMenuItem") "REQ-STAT-REG-LOGLINEAR-LOGIT-001" else "REQ-STAT-REG-LOGLINEAR-GENERAL-001"),
62      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/regression_loglinear.R


STAT REG LOGLINEAR GENERAL 002

Case STAT-REG-LOGLINEAR-GENERAL-002 · status: passed

Reference

Evidence type
R
Source
R stats::glm(family = poisson) on the cell table with contr.treatment at the declared reference; rstandard/residuals for adjusted and deviance residuals; profile limits by uniroot on the constrained deviance; logit contrasts from the Poisson coefficients
Citation
R Core Team, stats package help: glm, rstandard.glm, residuals.glm, uniroot; Agresti (2013) Categorical Data Analysis, ch. 9 (loglinear models, sampling schemes) and 10 (logit-loglinear equivalence).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear.R · sha256 29efcb27edfe009b
Dataset
reference_data/datasets/REG_LOGLINEAR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/dfexact550.00E+2passed
/fit/rankexact770.00E+2passed
/fit/allowable_cellsexact12120.00E+2passed
/fit/structural_zero_cellsexact000.00E+2passed
/fit/deviancetolerance15.87738684503601815.8773868450360256.71E-1515passed
/fit/pearsontolerance13.44394755301195913.4439475530119612.20E-1515passed
/fit/deviance_ptail_probability0.00720299987756792010.0072029998775678912.95E-1714passed
/fit/pearson_ptail_probability0.0195553488951060380.0195553488951060353.39E-1815passed
/fit/log_likelihoodtolerance-27.154314777158675-27.154314777158791.13E-1314passed
/fit/boundaryexactFalseFalsepassed
/totaltolerance109109.00.00E+2passed
/fit/parameters/0/termexact(intercept)(intercept)passed
/fit/parameters/0/labelexactpassed
/fit/parameters/0/estimatetolerance2.60506538654477642.6050653865447774.18E-1615passed
/fit/parameters/0/standard_errortolerance0.218574565344006130.218574565345026561.02E-1211passed
/fit/parameters/0/z_valuetolerance11.91842876340512811.9184287633494895.56E-1111passed
/fit/parameters/0/p_valuetail_probability9.4881035478281526e-339.488103554163704e-336.34E-429passed
/fit/parameters/0/lowertolerance2.24554221997934672.24554221997766851.68E-1212passed
/fit/parameters/0/uppertolerance2.96458855311020612.9645885531118851.68E-1212passed
/fit/parameters/1/termexactAApassed
/fit/parameters/1/labelexactA=a2A=a2passed
/fit/parameters/1/estimatetolerance-0.17435338714477838-0.174353387144778773.91E-1614passed
/fit/parameters/1/standard_errortolerance0.296005147960382020.29600514796038191.11E-1615passed
/fit/parameters/1/z_valuetolerance-0.58902146920807685-0.58902146920807841.56E-1514passed
/fit/parameters/1/p_valuetail_probability0.555846871688522650.55584687168852161.00E-1514passed
/fit/parameters/1/lowertolerance-0.66123852836371966-0.66123852836372014.42E-1615passed
/fit/parameters/1/uppertolerance0.312531754074162950.312531754074162563.90E-1614passed
/fit/parameters/2/termexactBBpassed
/fit/parameters/2/labelexactB=b2B=b2passed
/fit/parameters/2/estimatetolerance-0.17435338714477783-0.174353387144777775.83E-1715passed
/fit/parameters/2/standard_errortolerance0.296005147960382020.296005147960381965.54E-1715passed
/fit/parameters/2/z_valuetolerance-0.58902146920807497-0.58902146920807491.15E-1615passed
/fit/parameters/2/p_valuetail_probability0.555846871688523980.5558468716885242.04E-1918passed
/fit/parameters/2/lowertolerance-0.66123852836371921-0.66123852836371923.49E-1817passed
/fit/parameters/2/uppertolerance0.31253175407416350.312531754074163671.70E-1615passed
/fit/parameters/3/termexactBBpassed
/fit/parameters/3/labelexactB=b3B=b3passed
/fit/parameters/3/estimatetolerance-1.4271163556401456-1.42711635564014542.16E-1615passed
/fit/parameters/3/standard_errortolerance0.454606056374689480.454606056566194851.92E-109passed
/fit/parameters/3/z_valuetolerance-3.1392374466386483-3.1392374453162261.32E-99passed
/fit/parameters/3/p_valuetail_probability0.00169388131327584160.00169388132092062087.64E-128passed
/fit/parameters/3/lowertolerance-2.1748767763021588-2.1748767766171573.15E-109passed
/fit/parameters/3/uppertolerance-0.67935593497813251-0.67935593466313373.15E-109passed
/fit/parameters/4/termexactCCpassed
/fit/parameters/4/labelexactC=c2C=c2passed
/fit/parameters/4/estimatetolerance-0.16551443847757336-0.165514438477574056.90E-1614passed
/fit/parameters/4/standard_errortolerance0.192221623498106180.192221623503634235.53E-1210passed
/fit/parameters/4/z_valuetolerance-0.86106045441450585-0.86106045438974642.48E-1110passed
/fit/parameters/4/p_valuetail_probability0.389204747915471160.3892047479291071.36E-1110passed
/fit/parameters/4/lowertolerance-0.48169087306693348-0.48169087307602729.09E-1210passed
/fit/parameters/4/uppertolerance0.150661996111786770.150661996120879089.09E-1210passed
/fit/parameters/5/termexactA*BA*Bpassed
/fit/parameters/5/labelexactA=a2, B=b2A=a2, B=b2passed
/fit/parameters/5/estimatetolerance0.220873402779671320.220873402779670666.62E-1614passed
/fit/parameters/5/standard_errortolerance0.425079569836802170.42507956983680215.20E-1715passed
/fit/parameters/5/z_valuetolerance0.519604842134543720.51960484213454221.55E-1514passed
/fit/parameters/5/p_valuetail_probability0.603339022500488210.60333902250048929.98E-1614passed
/fit/parameters/5/lowertolerance-0.47832026940936406-0.478320269409364958.91E-1614passed
/fit/parameters/5/uppertolerance0.920067074968706770.92006707496870634.48E-1615passed
/fit/parameters/6/termexactA*BA*Bpassed
/fit/parameters/6/labelexactA=a2, B=b3A=a2, B=b3passed
/fit/parameters/6/estimatetolerance1.02165124753198211.02165124753198174.34E-1615passed
/fit/parameters/6/standard_errortolerance0.570713838574259950.57071383872680471.53E-109passed
/fit/parameters/6/z_valuetolerance1.79012874487894021.79012874440046034.78E-109passed
/fit/parameters/6/p_valuetail_probability0.0734332167273599180.073433216804262727.69E-118passed
/fit/parameters/6/lowertolerance0.0829105202017139660.082910519950799342.51E-108passed
/fit/parameters/6/uppertolerance1.96039197486225031.9603919751131642.51E-109passed
/fit/cells/0/observedtolerance1818.00.00E+2passed
/fit/cells/0/structuralexactFalseFalsepassed
/fit/cells/0/expectedtolerance13.53211009174311713.5321100917431235.74E-1515passed
/fit/cells/0/residualtolerance1.21456203596224271.21456203596224091.82E-1514passed
/fit/cells/0/adjusted_residualtolerance2.04277980063715782.04277980065469931.75E-1111passed
/fit/cells/0/deviance_residualtolerance1.15552998755385541.15552998755385541.33E-1716passed
/fit/cells/1/observedtolerance77.00.00E+2passed
/fit/cells/1/structuralexactFalseFalsepassed
/fit/cells/1/expectedtolerance11.46788990825688111.4678899082568755.51E-1515passed
/fit/cells/1/residualtolerance-1.3193520789603057-1.31935207896030441.32E-1514passed
/fit/cells/1/adjusted_residualtolerance-2.0427798006372786-2.0427798006547011.74E-1111passed
/fit/cells/1/deviance_residualtolerance-1.4229577050451301-1.42295770504512881.35E-1515passed
/fit/cells/2/observedtolerance1212.00.00E+2passed
/fit/cells/2/structuralexactFalseFalsepassed
/fit/cells/2/expectedtolerance11.36697247706421611.3669724770642225.54E-1515passed
/fit/cells/2/residualtolerance0.18775875142033410.18775875142033251.61E-1514passed
/fit/cells/2/adjusted_residualtolerance0.308532095101452190.308532095103576652.12E-1211passed
/fit/cells/2/deviance_residualtolerance0.186055269854089410.18605526985409121.80E-1514passed
/fit/cells/3/observedtolerance99.00.00E+2passed
/fit/cells/3/structuralexactFalseFalsepassed
/fit/cells/3/expectedtolerance9.63302752293577679.6330275229357733.57E-1515passed
/fit/cells/3/residualtolerance-0.20395821019808932-0.20395821019808821.11E-1514passed
/fit/cells/3/adjusted_residualtolerance-0.30853209510146717-0.308532095103574042.11E-1211passed
/fit/cells/3/deviance_residualtolerance-0.20625552433777888-0.206255524337777271.61E-1514passed
/fit/cells/4/observedtolerance66.00.00E+2passed
/fit/cells/4/structuralexactFalseFalsepassed
/fit/cells/4/expectedtolerance3.24770642201834873.2477064220183512.21E-1515passed
/fit/cells/4/residualtolerance1.52723678083334961.52723678083334781.80E-1514passed
/fit/cells/4/adjusted_residualtolerance2.3196863032131862.3196863031465586.66E-1110passed
/fit/cells/4/deviance_residualtolerance1.36423535503145051.36423535503144921.29E-1515passed
/fit/cells/5/observedtolerance00.00.00E+2passed
/fit/cells/5/structuralexactFalseFalsepassed
/fit/cells/5/expectedtolerance2.75229357798165182.75229357798165184.90E-1716passed
/fit/cells/5/residualtolerance-1.6590037908279933-1.65900379082799331.88E-1716passed
/fit/cells/5/adjusted_residualtolerance-2.3196863032133237-2.31968630314656076.68E-1110passed
/fit/cells/5/deviance_residualtolerance-2.3461856610173251-2.3461856610173253.63E-1716passed
/fit/cells/6/observedtolerance1010.00.00E+2passed
/fit/cells/6/structuralexactFalseFalsepassed
/fit/cells/6/expectedtolerance11.36697247706421111.366972477064211.22E-1616passed
/fit/cells/6/residualtolerance-0.40545005741491918-0.40545005741491924.94E-1816passed
/fit/cells/6/adjusted_residualtolerance-0.6662504662335621-0.66625046623815554.59E-1211passed
/fit/cells/6/deviance_residualtolerance-0.41401307468409382-0.41401307468409381.93E-1817passed
/fit/cells/7/observedtolerance1111.00.00E+2passed
/fit/cells/7/structuralexactFalseFalsepassed
/fit/cells/7/expectedtolerance9.63302752293577319.6330275229357648.85E-1515passed
/fit/cells/7/residualtolerance0.440431497384284280.440431497384287333.05E-1514passed
/fit/cells/7/adjusted_residualtolerance0.666250466233609840.66625046623816744.56E-1211passed
/fit/cells/7/deviance_residualtolerance0.430586468833328020.43058646883333224.16E-1514passed
/fit/cells/8/observedtolerance88.00.00E+2passed
/fit/cells/8/structuralexactFalseFalsepassed
/fit/cells/8/expectedtolerance11.90825688073394211.9082568807339321.03E-1415passed
/fit/cells/8/residualtolerance-1.1325542233003476-1.1325542233003452.66E-1514passed
/fit/cells/8/adjusted_residualtolerance-1.8717199672159488-1.87171996722961651.37E-1111passed
/fit/cells/8/deviance_residualtolerance-1.2049341010142411-1.20493410101423783.29E-1514passed
/fit/cells/9/observedtolerance1414.00.00E+2passed
/fit/cells/9/structuralexactFalseFalsepassed
/fit/cells/9/expectedtolerance10.09174311926605410.0917431192660361.77E-1414passed
/fit/cells/9/residualtolerance1.23026879220934381.23026879220935056.68E-1514passed
/fit/cells/9/adjusted_residualtolerance1.87171996721606341.87171996722962991.36E-1111passed
/fit/cells/9/deviance_residualtolerance1.16146430171333351.16146430171333955.98E-1514passed
/fit/cells/10/observedtolerance55.00.00E+2passed
/fit/cells/10/structuralexactFalseFalsepassed
/fit/cells/10/expectedtolerance7.57798165137614537.5779816513761452.23E-1717passed
/fit/cells/10/residualtolerance-0.93648979085397654-0.93648979085397654.76E-1817passed
/fit/cells/10/adjusted_residualtolerance-1.4810944017760894-1.4810944017824146.32E-1211passed
/fit/cells/10/deviance_residualtolerance-0.99893619124294053-0.99893619124294058.92E-2019passed
/fit/cells/11/observedtolerance99.00.00E+2passed
/fit/cells/11/structuralexactFalseFalsepassed
/fit/cells/11/expectedtolerance6.42201834862385476.422018348623854.46E-1515passed
/fit/cells/11/residualtolerance1.01728830302967441.01728830302967642.03E-1514passed
/fit/cells/11/adjusted_residualtolerance1.48109440177617761.48109440178241576.24E-1211passed
/fit/cells/11/deviance_residualtolerance0.958590078523117280.95859007852312012.78E-1514passed
/reference_levels/0/1exacta1a1passed
/reference_levels/1/1exactb1b1passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

199  m <- loglinear(d, n ~ A * B + C, first)
200  rows <- parameter_table(m$fit, list("A", "B", "C", c("A", "B")), first)
201  emit("STAT-REG-LOGLINEAR-GENERAL-002", "loglinearGeneralMenuItem", "axis_coverage",
202       list(factors = list("A", "B", "C"), count = "n", terms = list("A*B", "C"),
203            sampling = "multinomial", coding = "first", confidence_level = 0.90,
204            conditional_test = "off"),
205       c(fit_outputs(m$fit, rows, 0.90, multinomial_loglik(m$fit)),
206         list("/reference_levels/0/1" = exact("a1", "reference of A"),
207              "/reference_levels/1/1" = exact("b1", "reference of B"))),
208       list("Multinomial log-likelihood = Poisson kernel - N log N + lgamma(N + 1) with N the table total; first-level reference coding."))

Whole script generators/r/regression_loglinear.R


STAT REG LOGLINEAR GENERAL 003

Case STAT-REG-LOGLINEAR-GENERAL-003 · status: passed

Reference

Evidence type
R
Source
R stats::glm(family = poisson) on the cell table with contr.treatment at the declared reference; rstandard/residuals for adjusted and deviance residuals; profile limits by uniroot on the constrained deviance; logit contrasts from the Poisson coefficients
Citation
R Core Team, stats package help: glm, rstandard.glm, residuals.glm, uniroot; Agresti (2013) Categorical Data Analysis, ch. 9 (loglinear models, sampling schemes) and 10 (logit-loglinear equivalence).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear.R · sha256 29efcb27edfe009b
Dataset
reference_data/datasets/REG_LOGLINEAR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/dfexact550.00E+2passed
/fit/rankexact770.00E+2passed
/fit/allowable_cellsexact12120.00E+2passed
/fit/structural_zero_cellsexact000.00E+2passed
/fit/deviancetolerance17.45357820043575417.4535782004357573.14E-1515passed
/fit/pearsontolerance13.95246977575352113.9524697757534893.22E-1414passed
/fit/deviance_ptail_probability0.00371525983300387230.00371525983300386675.62E-1814passed
/fit/pearson_ptail_probability0.0159141698833358930.0159141698833360942.01E-1613passed
/fit/log_likelihoodtolerance-25.369698375754297-25.3696983757540122.85E-1313passed
/fit/boundaryexactFalseFalsepassed
/totaltolerance109109.00.00E+2passed
/fit/parameters/0/termexact(intercept)(intercept)passed
/fit/parameters/0/labelexactpassed
/fit/parameters/0/estimatetolerance1.54892796294162591.54892796294162771.81E-1514passed
/fit/parameters/0/standard_errortolerance0.345659743230707310.34565974430591141.08E-98passed
/fit/parameters/0/z_valuetolerance4.48107710913795554.4810770951991891.39E-88passed
/fit/parameters/0/p_valuetail_probability7.4267270779119837e-067.426727563017821e-064.85E-137passed
/fit/parameters/0/lowertolerance0.871447315304077040.87144731319671732.11E-98passed
/fit/parameters/0/uppertolerance2.22640861057917492.2264086126865382.11E-99passed
/fit/parameters/1/termexactAApassed
/fit/parameters/1/labelexactA=a1A=a1passed
/fit/parameters/1/estimatetolerance-0.091807549253122359-0.091807549253121518.47E-1614passed
/fit/parameters/1/standard_errortolerance0.191767121425175850.19176712147782775.27E-119passed
/fit/parameters/1/z_valuetolerance-0.47874499325445652-0.47874499312300721.31E-109passed
/fit/parameters/1/p_valuetail_probability0.63212005308487940.63212005317840479.35E-119passed
/fit/parameters/1/lowertolerance-0.46766420066538628-0.46766420076858131.03E-109passed
/fit/parameters/1/uppertolerance0.284049102159141590.284049102262338261.03E-109passed
/fit/parameters/2/termexactBBpassed
/fit/parameters/2/labelexactB=b1B=b1passed
/fit/parameters/2/estimatetolerance0.693147180559944620.69314718055994558.88E-1614passed
/fit/parameters/2/standard_errortolerance0.408248289559092760.408248290463862359.05E-108passed
/fit/parameters/2/z_valuetolerance1.69785691278349751.69785690902066883.76E-98passed
/fit/parameters/2/p_valuetail_probability0.0895347711674239210.089534771877786737.10E-108passed
/fit/parameters/2/lowertolerance-0.10700476372595635-0.107004765499271511.77E-97passed
/fit/parameters/2/uppertolerance1.49329912484584561.49329912661916261.77E-98passed
/fit/parameters/3/termexactBBpassed
/fit/parameters/3/labelexactB=b2B=b2passed
/fit/parameters/3/estimatetolerance0.938269638592930550.93826963859293023.28E-1615passed
/fit/parameters/3/standard_errortolerance0.393178548806944310.393178549746391779.39E-108passed
/fit/parameters/3/z_valuetolerance2.38637036898376922.3863703632818565.70E-98passed
/fit/parameters/3/p_valuetail_probability0.0170156049985747820.0170156052624256052.64E-107passed
/fit/parameters/3/lowertolerance0.167653843437596040.167653841596312361.84E-97passed
/fit/parameters/3/uppertolerance1.70888543374826511.7088854355895481.84E-98passed
/fit/parameters/4/termexactCCpassed
/fit/parameters/4/labelexactC=c1C=c1passed
/fit/parameters/4/estimatetolerance0.200670695462151130.200670695462151163.07E-1715passed
/fit/parameters/4/standard_errortolerance0.449466574153696230.44946657497549418.22E-108passed
/fit/parameters/4/z_valuetolerance0.446464113243560780.446464112427252658.16E-108passed
/fit/parameters/4/p_valuetail_probability0.655262029725877880.65526203031541375.90E-109passed
/fit/parameters/4/lowertolerance-0.68026760213369486-0.68026760374438911.61E-98passed
/fit/parameters/4/uppertolerance1.08160899305799711.08160899466869151.61E-98passed
/fit/parameters/5/termexactB*CB*Cpassed
/fit/parameters/5/labelexactB=b1, C=c1B=b1, C=c1passed
/fit/parameters/5/estimatetolerance0.241162056816889350.24116205681688945.30E-1715passed
/fit/parameters/5/standard_errortolerance0.541562593382611590.54156259406465886.82E-108passed
/fit/parameters/5/z_valuetolerance0.445307818087261110.445307817526437835.61E-108passed
/fit/parameters/5/p_valuetail_probability0.65609731874671040.65609731915194464.05E-109passed
/fit/parameters/5/lowertolerance-0.82028112158713906-0.8202811229239271.34E-98passed
/fit/parameters/5/uppertolerance1.30260523522091761.30260523655770571.34E-98passed
/fit/parameters/6/termexactB*CB*Cpassed
/fit/parameters/6/labelexactB=b2, C=c1B=b2, C=c1passed
/fit/parameters/6/estimatetolerance-0.34043263783731026-0.340432637837309378.89E-1614passed
/fit/parameters/6/standard_errortolerance0.54359770248872830.5435977031682226.79E-108passed
/fit/parameters/6/z_valuetolerance-0.62625841919258163-0.62625841840976097.83E-108passed
/fit/parameters/6/p_valuetail_probability0.531145454909156210.53114545542253415.13E-109passed
/fit/parameters/6/lowertolerance-1.405864556793937-1.40586455812571921.33E-99passed
/fit/parameters/6/uppertolerance0.724999281119316330.72499928245110061.33E-98passed
/fit/cells/0/observedtolerance1818.00.00E+2passed
/fit/cells/0/structuralexactFalseFalsepassed
/fit/cells/0/expectedtolerance13.35779816513763113.3577981651376784.75E-1414passed
/fit/cells/0/residualtolerance1.27015458532893041.27015458532891511.53E-1413passed
/fit/cells/0/adjusted_residualtolerance2.03752591419655092.0375259143883571.92E-1010passed
/fit/cells/0/deviance_residualtolerance1.20555677574620731.20555677574619561.17E-1414passed
/fit/cells/1/observedtolerance77.00.00E+2passed
/fit/cells/1/structuralexactFalseFalsepassed
/fit/cells/1/expectedtolerance8.58715596330275058.587155963302783.02E-1414passed
/fit/cells/1/residualtolerance-0.54162029550168522-0.54162029550169479.44E-1513passed
/fit/cells/1/adjusted_residualtolerance-0.81971579533464922-0.81971579537856664.39E-1110passed
/fit/cells/1/deviance_residualtolerance-0.55973895156725872-0.55973895156726869.89E-1513passed
/fit/cells/2/observedtolerance1212.00.00E+2passed
/fit/cells/2/structuralexactFalseFalsepassed
/fit/cells/2/expectedtolerance9.5412844036697289.5412844036697562.84E-1414passed
/fit/cells/2/residualtolerance0.795984984768817850.79598498476880751.03E-1413passed
/fit/cells/2/adjusted_residualtolerance1.218144874399851.21814487447407757.42E-1110passed
/fit/cells/2/deviance_residualtolerance0.765018800096352150.76501880009634239.88E-1513passed
/fit/cells/3/observedtolerance99.00.00E+2passed
/fit/cells/3/structuralexactFalseFalsepassed
/fit/cells/3/expectedtolerance10.97247706422019310.9724770642202182.51E-1414passed
/fit/cells/3/residualtolerance-0.59546963619613547-0.59546963619614226.77E-1513passed
/fit/cells/3/adjusted_residualtolerance-0.92704211761337618-0.92704211768075976.74E-1110passed
/fit/cells/3/deviance_residualtolerance-0.61479718560393504-0.61479718560394196.88E-1513passed
/fit/cells/4/observedtolerance66.00.00E+2passed
/fit/cells/4/structuralexactFalseFalsepassed
/fit/cells/4/expectedtolerance5.24770642201835185.2477064220183651.33E-1414passed
/fit/cells/4/residualtolerance0.328399576976292410.328399576976286136.28E-1513passed
/fit/cells/4/adjusted_residualtolerance0.478937255578044620.478937255592439941.44E-1110passed
/fit/cells/4/deviance_residualtolerance0.320987795664424090.320987795664418545.55E-1513passed
/fit/cells/5/observedtolerance00.00.00E+2passed
/fit/cells/5/structuralexactFalseFalsepassed
/fit/cells/5/expectedtolerance4.29357798165137884.293577981651391.16E-1414passed
/fit/cells/5/residualtolerance-2.0720950706112347-2.07209507061123783.07E-1514passed
/fit/cells/5/adjusted_residualtolerance-2.9915670958774569-2.99156709505412138.23E-109passed
/fit/cells/5/deviance_residualtolerance-2.9303849513848443-2.93038495138484834.04E-1514passed
/fit/cells/6/observedtolerance1010.00.00E+2passed
/fit/cells/6/structuralexactFalseFalsepassed
/fit/cells/6/expectedtolerance14.64220183486239614.6422018348624353.92E-1414passed
/fit/cells/6/residualtolerance-1.2131676593527723-1.21316765935278078.40E-1514passed
/fit/cells/6/adjusted_residualtolerance-2.0375259141934263-2.03752591438840641.95E-1010passed
/fit/cells/6/deviance_residualtolerance-1.2876131460937652-1.28761314609377521.00E-1414passed
/fit/cells/7/observedtolerance1111.00.00E+2passed
/fit/cells/7/structuralexactFalseFalsepassed
/fit/cells/7/expectedtolerance9.41284403669724069.4128440366972652.48E-1414passed
/fit/cells/7/residualtolerance0.517319886682591550.51731988668258278.89E-1513passed
/fit/cells/7/adjusted_residualtolerance0.819715795333391670.81971579537854324.52E-1110passed
/fit/cells/7/deviance_residualtolerance0.503717101083617110.50371710108361284.33E-1514passed
/fit/cells/8/observedtolerance88.00.00E+2passed
/fit/cells/8/structuralexactFalseFalsepassed
/fit/cells/8/expectedtolerance10.45871559633027410.4587155963302972.29E-1414passed
/fit/cells/8/residualtolerance-0.76027221568613812-0.76027221568614446.33E-1514passed
/fit/cells/8/adjusted_residualtolerance-1.2181448743979761-1.21814487447410417.61E-1110passed
/fit/cells/8/deviance_residualtolerance-0.79342631226067029-0.79342631226067585.55E-1514passed
/fit/cells/9/observedtolerance1414.00.00E+2passed
/fit/cells/9/structuralexactFalseFalsepassed
/fit/cells/9/expectedtolerance12.02752293577982112.0275229357798371.60E-1414passed
/fit/cells/9/residualtolerance0.568753215635262820.56875321563525785.00E-1514passed
/fit/cells/9/adjusted_residualtolerance0.927042117611942220.92704211768073366.88E-1110passed
/fit/cells/9/deviance_residualtolerance0.554182513094107820.5541825130941034.77E-1514passed
/fit/cells/10/observedtolerance55.00.00E+2passed
/fit/cells/10/structuralexactFalseFalsepassed
/fit/cells/10/expectedtolerance5.75229357798165185.75229357798166159.72E-1514passed
/fit/cells/10/residualtolerance-0.31366555751132813-0.31366555751133193.77E-1513passed
/fit/cells/10/adjusted_residualtolerance-0.47893725557730971-0.478937255592456761.51E-1110passed
/fit/cells/10/deviance_residualtolerance-0.32090449711173696-0.320904497111741464.50E-1513passed
/fit/cells/11/observedtolerance99.00.00E+2passed
/fit/cells/11/structuralexactFalseFalsepassed
/fit/cells/11/expectedtolerance4.70642201834862474.7064220183486338.02E-1514passed
/fit/cells/11/residualtolerance1.97912817526760891.97912817526760355.36E-1514passed
/fit/cells/11/adjusted_residualtolerance2.99156709587284952.99156709505410538.19E-109passed
/fit/cells/11/deviance_residualtolerance1.75561473444486961.75561473444486524.45E-1514passed
/fit/generators/0exactB*CB*Cpassed
/fit/generators/1exactAApassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

211  m <- loglinear(d, n ~ B * C + A, last)
212  rows <- parameter_table(m$fit, list("A", "B", "C", c("B", "C")), last)
213  emit("STAT-REG-LOGLINEAR-GENERAL-003", "loglinearGeneralMenuItem", "axis_coverage",
214       list(factors = list("A", "B", "C"), count = "n", terms = list("B*C"),
215            sampling = "product_multinomial", fixed_margins = list("A"), coding = "last",
216            confidence_level = 0.95, conditional_test = "off"),
217       c(fit_outputs(m$fit, rows, 0.95, product_multinomial_loglik(m$fit, d$A)),
218         list("/fit/generators/0" = exact("B*C", "generating term"),
219              "/fit/generators/1" = exact("A", "the fixed stratum factor"))),
220       list("Product-multinomial: the stratum factor A enters the generating class; log-likelihood adds lgamma(n_a + 1) - n_a log n_a per stratum."))

Whole script generators/r/regression_loglinear.R


STAT REG LOGLINEAR GENERAL 004

Case STAT-REG-LOGLINEAR-GENERAL-004 · status: passed

Reference

Evidence type
R
Source
R stats::glm(family = poisson) on the cell table with contr.treatment at the declared reference; rstandard/residuals for adjusted and deviance residuals; profile limits by uniroot on the constrained deviance; logit contrasts from the Poisson coefficients
Citation
R Core Team, stats package help: glm, rstandard.glm, residuals.glm, uniroot; Agresti (2013) Categorical Data Analysis, ch. 9 (loglinear models, sampling schemes) and 10 (logit-loglinear equivalence).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear.R · sha256 29efcb27edfe009b
Dataset
reference_data/datasets/REG_LOGLINEAR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/dfexact550.00E+2passed
/fit/rankexact770.00E+2passed
/fit/allowable_cellsexact12120.00E+2passed
/fit/structural_zero_cellsexact000.00E+2passed
/fit/deviancetolerance13.31921982672039913.3192198267203963.38E-1515passed
/fit/pearsontolerance12.25383683824797712.2538368382479562.09E-1414passed
/fit/deviance_ptail_probability0.0205640885895092270.0205640885895092764.89E-1714passed
/fit/pearson_ptail_probability0.0314703986421750040.031470398642175292.85E-1614passed
/fit/log_likelihoodtolerance-30.164942828811924-30.1649428288120169.23E-1414passed
/fit/boundaryexactFalseFalsepassed
/totaltolerance109109.00.00E+2passed
/fit/parameters/0/termexact(intercept)(intercept)passed
/fit/parameters/0/labelexactpassed
/fit/parameters/0/estimatetolerance2.3608540011180212.36085400111802231.35E-1515passed
/fit/parameters/0/standard_errortolerance0.231625847486745830.231625847486747581.75E-1514passed
/fit/parameters/0/z_valuetolerance10.19253259830215810.1925325983020877.11E-1414passed
/fit/parameters/0/p_valuetail_probability2.1411235199461188e-242.1411235199476715e-241.55E-3612passed
/fit/parameters/0/lowertolerance1.9068756821554321.90687568215542982.20E-1514passed
/fit/parameters/0/uppertolerance2.814832320080612.8148323200806154.90E-1514passed
/fit/parameters/1/termexactAApassed
/fit/parameters/1/labelexactA=a1A=a1passed
/fit/parameters/1/estimatetolerance-0.044451762570833747-0.044451762570834113.61E-1614passed
/fit/parameters/1/standard_errortolerance0.298216039682829060.298216039682829066.13E-1917passed
/fit/parameters/1/z_valuetolerance-0.14905892593205552-0.149058925932056721.20E-1514passed
/fit/parameters/1/p_valuetail_probability0.881507136093534930.88150713609353391.00E-1514passed
/fit/parameters/1/lowertolerance-0.62894445996134607-0.62894445996134676.66E-1614passed
/fit/parameters/1/uppertolerance0.540040934819678630.54004093481967842.21E-1615passed
/fit/parameters/2/termexactBBpassed
/fit/parameters/2/labelexactB=b1B=b1passed
/fit/parameters/2/estimatetolerance-0.73759894313077934-0.73759894313077894.45E-1615passed
/fit/parameters/2/standard_errortolerance0.2982160396828290.298216039682828955.16E-1715passed
/fit/parameters/2/z_valuetolerance-2.4733711302559747-2.47337113025597331.37E-1515passed
/fit/parameters/2/p_valuetail_probability0.0133845066581579580.0133845066581580064.84E-1714passed
/fit/parameters/2/lowertolerance-1.3220916405212915-1.32209164052129122.55E-1615passed
/fit/parameters/2/uppertolerance-0.1531062457402671-0.153106245740266545.56E-1614passed
/fit/parameters/3/termexactBBpassed
/fit/parameters/3/labelexactB=b3B=b3passed
/fit/parameters/3/estimatetolerance-0.83290912293510433-0.83290912293510421.12E-1615passed
/fit/parameters/3/standard_errortolerance0.331880893599242780.33188089359924261.70E-1615passed
/fit/parameters/3/z_valuetolerance-2.5096627705868171-2.5096627705868188.70E-1615passed
/fit/parameters/3/p_valuetail_probability0.0120846506929702390.0120846506929702013.81E-1714passed
/fit/parameters/3/lowertolerance-1.4833837215465899-1.48338372154658954.16E-1615passed
/fit/parameters/3/uppertolerance-0.18243452432361884-0.182434524323618951.12E-1615passed
/fit/parameters/4/termexactCCpassed
/fit/parameters/4/labelexactC=c1C=c1passed
/fit/parameters/4/estimatetolerance0.15684247149296990.15684247149297023.04E-1614passed
/fit/parameters/4/standard_errortolerance0.187074735831334410.18707473583134197.49E-1513passed
/fit/parameters/4/z_valuetolerance0.838394723883925440.83839472388389353.20E-1413passed
/fit/parameters/4/p_valuetail_probability0.401809051688235050.40180905168825311.80E-1413passed
/fit/parameters/4/lowertolerance-0.20981727315379026-0.209817273153804721.45E-1413passed
/fit/parameters/4/uppertolerance0.523502216139730030.52350221613974511.51E-1413passed
/fit/parameters/5/termexactA*BA*Bpassed
/fit/parameters/5/labelexactA=a1, B=b1A=a1, B=b1passed
/fit/parameters/5/estimatetolerance0.904653027793945190.90465302779394711.88E-1514passed
/fit/parameters/5/standard_errortolerance0.415750995477093820.41575099547709389.73E-1917passed
/fit/parameters/5/z_valuetolerance2.17594915619098692.1759491561909924.93E-1514passed
/fit/parameters/5/p_valuetail_probability0.0295590625851626270.029559062585162273.57E-1613passed
/fit/parameters/5/lowertolerance0.0897960501221666260.089796050122168291.67E-1513passed
/fit/parameters/5/uppertolerance1.71951000546572371.71951000546572572.04E-1514passed
/fit/parameters/6/termexactA*BA*Bpassed
/fit/parameters/6/labelexactA=a1, B=b3A=a1, B=b3passed
/fit/parameters/6/estimatetolerance-0.31222318136789851-0.312223181367897955.59E-1614passed
/fit/parameters/6/standard_errortolerance0.546311830228606520.54631183022878051.74E-1312passed
/fit/parameters/6/z_valuetolerance-0.5715109285428569-0.57151092854267381.83E-1312passed
/fit/parameters/6/p_valuetail_probability0.567653354434369480.56765335443449351.24E-1312passed
/fit/parameters/6/lowertolerance-1.3829746929441273-1.38297469294446823.41E-1312passed
/fit/parameters/6/uppertolerance0.758528330208330440.75852833020867233.42E-1312passed
/fit/cells/0/observedtolerance1818.00.00E+2passed
/fit/cells/0/structuralexactFalseFalsepassed
/fit/cells/0/expectedtolerance14.01739130434782214.0173913043478714.93E-1414passed
/fit/cells/0/residualtolerance1.19728414306553281.19728414306551721.56E-1413passed
/fit/cells/0/adjusted_residualtolerance2.00475882043467732.00475882043467658.42E-1615passed
/fit/cells/0/deviance_residualtolerance1.14072095263960651.14072095263959251.40E-1413passed
/fit/cells/1/observedtolerance77.00.00E+2passed
/fit/cells/1/structuralexactFalseFalsepassed
/fit/cells/1/expectedtolerance11.98260869565216411.9826086956522074.29E-1414passed
/fit/cells/1/residualtolerance-1.2949563829810757-1.29495638298108571.00E-1414passed
/fit/cells/1/adjusted_residualtolerance-2.0047588204346702-2.00475882043470933.91E-1413passed
/fit/cells/1/deviance_residualtolerance-1.3917319610050651-1.39173196100507671.16E-1414passed
/fit/cells/2/observedtolerance1212.00.00E+2passed
/fit/cells/2/structuralexactFalseFalsepassed
/fit/cells/2/expectedtolerance11.86086956521738911.8608695652174051.55E-1414passed
/fit/cells/2/residualtolerance0.185580027074745220.185580027074740444.78E-1513passed
/fit/cells/2/adjusted_residualtolerance0.303983257789307240.30398325778930234.94E-1513passed
/fit/cells/2/deviance_residualtolerance0.183949686428752090.18394968642873941.27E-1413passed
/fit/cells/3/observedtolerance99.00.00E+2passed
/fit/cells/3/structuralexactFalseFalsepassed
/fit/cells/3/expectedtolerance10.13913043478260610.1391304347826159.00E-1515passed
/fit/cells/3/residualtolerance-0.20071930460795151-0.200719304607954232.72E-1513passed
/fit/cells/3/adjusted_residualtolerance-0.30398325778930452-0.303983257789311577.05E-1513passed
/fit/cells/3/deviance_residualtolerance-0.20288546200272842-0.20288546200272795.30E-1614passed
/fit/cells/4/observedtolerance66.00.00E+2passed
/fit/cells/4/structuralexactFalseFalsepassed
/fit/cells/4/expectedtolerance3.773913043478263.77391304347826666.62E-1514passed
/fit/cells/4/residualtolerance1.403278141059041.40327814105903544.63E-1514passed
/fit/cells/4/adjusted_residualtolerance2.1330045803601762.13300458036008239.37E-1413passed
/fit/cells/4/deviance_residualtolerance1.27113851019241751.27113851019241443.08E-1514passed
/fit/cells/5/observedtolerance00.00.00E+2passed
/fit/cells/5/structuralexactFalseFalsepassed
/fit/cells/5/expectedtolerance3.22608695652173833.22608695652174274.41E-1514passed
/fit/cells/5/residualtolerance-1.5177549927367275-1.51775499273672891.37E-1515passed
/fit/cells/5/adjusted_residualtolerance-2.1330045803601738-2.13300458036008948.44E-1413passed
/fit/cells/5/deviance_residualtolerance-1.8941374813458887-1.89413748134589071.96E-1514passed
/fit/cells/6/observedtolerance1010.00.00E+2passed
/fit/cells/6/structuralexactFalseFalsepassed
/fit/cells/6/expectedtolerance11.86086956521738311.860869565217412.69E-1414passed
/fit/cells/6/residualtolerance-0.39514658826118826-0.395146588261195537.27E-1513passed
/fit/cells/6/adjusted_residualtolerance-0.64725686862620546-0.64725686862622361.81E-1413passed
/fit/cells/6/deviance_residualtolerance-0.40308884879702322-0.40308884879703047.17E-1513passed
/fit/cells/7/observedtolerance1111.00.00E+2passed
/fit/cells/7/structuralexactFalseFalsepassed
/fit/cells/7/expectedtolerance10.13913043478260310.139130434782621.73E-1414passed
/fit/cells/7/residualtolerance0.427381920695846130.427381920695840136.00E-1513passed
/fit/cells/7/adjusted_residualtolerance0.647256868626212010.64725686862620922.77E-1514passed
/fit/cells/7/deviance_residualtolerance0.418319490905700750.41831949090569495.83E-1513passed
/fit/cells/8/observedtolerance88.00.00E+2passed
/fit/cells/8/structuralexactFalseFalsepassed
/fit/cells/8/expectedtolerance12.39999999999999512.4000000000000162.13E-1414passed
/fit/cells/8/residualtolerance-1.1075255767817753-1.10752557678178024.91E-1514passed
/fit/cells/8/adjusted_residualtolerance-1.8239786743555648-1.8239786743555932.82E-1413passed
/fit/cells/8/deviance_residualtolerance-1.1748552017375373-1.17485520173754275.35E-1514passed
/fit/cells/9/observedtolerance1414.00.00E+2passed
/fit/cells/9/structuralexactFalseFalsepassed
/fit/cells/9/expectedtolerance10.59999999999999610.600000000000011.43E-1414passed
/fit/cells/9/residualtolerance1.19787547782620621.1978754778262015.13E-1514passed
/fit/cells/9/adjusted_residualtolerance1.82397867435556821.823978674355581.17E-1414passed
/fit/cells/9/deviance_residualtolerance1.13381868080690621.13381868080690154.69E-1514passed
/fit/cells/10/observedtolerance55.00.00E+2passed
/fit/cells/10/structuralexactFalseFalsepassed
/fit/cells/10/expectedtolerance8.08695652173912738.0869565217391411.42E-1414passed
/fit/cells/10/residualtolerance-0.90969660853967738-0.90969660853968164.21E-1514passed
/fit/cells/10/adjusted_residualtolerance-1.43699905046902-1.43699905046903531.53E-1413passed
/fit/cells/10/deviance_residualtolerance-0.96610818362974726-0.96610818362975184.56E-1514passed
/fit/cells/11/observedtolerance99.00.00E+2passed
/fit/cells/11/structuralexactFalseFalsepassed
/fit/cells/11/expectedtolerance6.91304347826086476.91304347826087459.81E-1514passed
/fit/cells/11/residualtolerance0.983907985942666370.9839079859426624.33E-1514passed
/fit/cells/11/adjusted_residualtolerance1.43699905046902471.43699905046902641.73E-1514passed
/fit/cells/11/deviance_residualtolerance0.930505986286275570.93050598628627134.22E-1514passed
/fit/parameters/1/profile_lowertolerance-0.63438998382583967-0.63438998382583992.17E-1615passed
/fit/parameters/1/profile_uppertolerance0.542934330985664020.54293433098565251.16E-1413passed
/fit/parameters/4/profile_lowertolerance-0.20908774176712683-0.209087741767038628.82E-1412passed
/fit/parameters/4/profile_uppertolerance0.526291558993360440.52629155899335832.11E-1514passed
/reference_levels/1/1exactb2b2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

250  out[["/reference_levels/1/1"]] <- exact("b2", "recorded reference of B")
251  emit("STAT-REG-LOGLINEAR-GENERAL-004", "loglinearGeneralMenuItem", "axis_coverage",
252       list(factors = list("A", "B", "C"), count = "n", terms = list("A*B", "C"),
253            sampling = "poisson", coding = "last", confidence_level = 0.95,
254            cell_structure = "weight", pseudocount = TRUE, reference_levels = list("B=b2"),
255            intervals = "wald_and_profile", conditional_test = "off"),
256       out,
257       list("Counts + 0.5 in every cell (named pseudocount analysis), offset log(weight) from the cell structure, B's reference recorded as b2 (A and C keep the last level).",
258            "Profile limits for A [A=a1] and C [C=c1]: the constrained deviance rises by chi-square(1, 0.95) = 3.841; found by uniroot on refits with the parameter fixed through the offset (manual chapters/86 lines 202-212)."))

Whole script generators/r/regression_loglinear.R


STAT REG LOGLINEAR GENERAL 005

Case STAT-REG-LOGLINEAR-GENERAL-005 · status: passed

Reference

Evidence type
R
Source
R stats::glm(family = poisson) on the cell table with contr.treatment at the declared reference; rstandard/residuals for adjusted and deviance residuals; profile limits by uniroot on the constrained deviance; logit contrasts from the Poisson coefficients
Citation
R Core Team, stats package help: glm, rstandard.glm, residuals.glm, uniroot; Agresti (2013) Categorical Data Analysis, ch. 9 (loglinear models, sampling schemes) and 10 (logit-loglinear equivalence).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear.R · sha256 29efcb27edfe009b
Dataset
reference_data/datasets/REG_LOGLINEAR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/dfexact440.00E+2passed
/fit/rankexact770.00E+2passed
/fit/allowable_cellsexact11110.00E+2passed
/fit/structural_zero_cellsexact110.00E+2passed
/fit/deviancetolerance8.21619563873764858.2161956387376453.59E-1515passed
/fit/pearsontolerance8.01483188760924048.0148318876092231.78E-1414passed
/fit/deviance_ptail_probability0.0839720955507005930.083972095550700699.70E-1714passed
/fit/pearson_ptail_probability0.0910363918974241990.091036391897424846.38E-1614passed
/fit/log_likelihoodtolerance-26.589096172177822-26.5890961721778321.05E-1415passed
/fit/boundaryexactFalseFalsepassed
/totaltolerance109109.00.00E+2passed
/fit/parameters/0/termexact(intercept)(intercept)passed
/fit/parameters/0/labelexactpassed
/fit/parameters/0/estimatetolerance1.91635134681376851.91635134681377031.78E-1515passed
/fit/parameters/0/standard_errortolerance0.28586681088293020.285866810883134152.04E-1312passed
/fit/parameters/0/z_valuetolerance6.70365104957414466.7036510495693684.78E-1212passed
/fit/parameters/0/p_valuetail_probability2.0327535981994157e-112.0327535982658884e-116.65E-2210passed
/fit/parameters/0/lowertolerance1.35606269310790271.35606269310750463.98E-1312passed
/fit/parameters/0/uppertolerance2.47664000051963432.4766400005200364.02E-1312passed
/fit/parameters/1/termexactAApassed
/fit/parameters/1/labelexactA=a1A=a1passed
/fit/parameters/1/estimatetolerance-0.1828607857096895-0.1828607857096895.00E-1614passed
/fit/parameters/1/standard_errortolerance0.497246844896364390.497246844896437177.28E-1412passed
/fit/parameters/1/z_valuetolerance-0.36774649771342666-0.367746497713371875.48E-1412passed
/fit/parameters/1/p_valuetail_probability0.713062266154201870.71306226615424274.09E-1413passed
/fit/parameters/1/lowertolerance-1.1574466931327378-1.15744669313288021.42E-1312passed
/fit/parameters/1/uppertolerance0.791725121713358960.79172512171350221.43E-1312passed
/fit/parameters/2/termexactBBpassed
/fit/parameters/2/labelexactB=b1B=b1passed
/fit/parameters/2/estimatetolerance0.405465108108164610.40546510810816571.11E-1514passed
/fit/parameters/2/standard_errortolerance0.345032779671139820.34503277967117673.69E-1412passed
/fit/parameters/2/z_valuetolerance1.17514952780609571.17514952780597341.22E-1312passed
/fit/parameters/2/p_valuetail_probability0.239934898752245610.239934898752294574.90E-1412passed
/fit/parameters/2/lowertolerance-0.27078671353301298-0.270786713533084267.13E-1412passed
/fit/parameters/2/uppertolerance1.08171692974934211.08171692974941587.37E-1413passed
/fit/parameters/3/termexactBBpassed
/fit/parameters/3/labelexactB=b2B=b2passed
/fit/parameters/3/estimatetolerance0.451985123743057770.45198512374305785.61E-1715passed
/fit/parameters/3/standard_errortolerance0.341881729378868520.34188172937891354.50E-1412passed
/fit/parameters/3/z_valuetolerance1.32205112149229321.32205112149211931.74E-1312passed
/fit/parameters/3/p_valuetail_probability0.186151128385051770.18615112838510965.78E-1412passed
/fit/parameters/3/lowertolerance-0.21809075281179369-0.21809075281188198.82E-1412passed
/fit/parameters/3/uppertolerance1.12206100029790921.12206100029799768.84E-1413passed
/fit/parameters/4/termexactCCpassed
/fit/parameters/4/labelexactC=c1C=c1passed
/fit/parameters/4/estimatetolerance0.0582689081239762960.058268908123976855.55E-1614passed
/fit/parameters/4/standard_errortolerance0.197149497914677620.197149497915418947.41E-1311passed
/fit/parameters/4/z_valuetolerance0.295556969408027230.295556969406918671.11E-1211passed
/fit/parameters/4/p_valuetail_probability0.767568439565838090.76756843956668498.47E-1311passed
/fit/parameters/4/lowertolerance-0.32813700735894624-0.328137007360398751.45E-1211passed
/fit/parameters/4/uppertolerance0.444674823606898810.44467482360835241.45E-1211passed
/fit/parameters/5/termexactA*BA*Bpassed
/fit/parameters/5/labelexactA=a1, B=b1A=a1, B=b1passed
/fit/parameters/5/estimatetolerance0.357214172854468440.3572141728544671.45E-1514passed
/fit/parameters/5/standard_errortolerance0.578682531598476050.57868253159958861.11E-1211passed
/fit/parameters/5/z_valuetolerance0.617288674444254060.61728867444306481.19E-1211passed
/fit/parameters/5/p_valuetail_probability0.537044335713791180.53704433571457557.84E-1311passed
/fit/parameters/5/lowertolerance-0.77698274756100605-0.77698274756318842.18E-1211passed
/fit/parameters/5/uppertolerance1.49141109326994291.49141109327212232.18E-1211passed
/fit/parameters/6/termexactA*BA*Bpassed
/fit/parameters/6/labelexactA=a1, B=b2A=a1, B=b2passed
/fit/parameters/6/estimatetolerance0.136340770074796360.13634077007479567.74E-1614passed
/fit/parameters/6/standard_errortolerance0.583376394648401960.58337639464847597.39E-1412passed
/fit/parameters/6/z_valuetolerance0.23370978209868820.233709782098657253.09E-1412passed
/fit/parameters/6/p_valuetail_probability0.815210293030152130.81521029303017612.40E-1413passed
/fit/parameters/6/lowertolerance-1.0070559528668965-1.00705595286704221.46E-1312passed
/fit/parameters/6/uppertolerance1.27973749301648911.27973749301663341.44E-1312passed
/fit/cells/0/observedtolerance1818.00.00E+2passed
/fit/cells/0/structuralexactFalseFalsepassed
/fit/cells/0/expectedtolerance12.86407766990293312.8640776699029622.85E-1414passed
/fit/cells/0/residualtolerance1.43195417672581821.43195417672580869.60E-1514passed
/fit/cells/0/adjusted_residualtolerance2.36175064012920372.361750640121657.55E-1211passed
/fit/cells/0/deviance_residualtolerance1.34971982915085591.34971982915084728.69E-1514passed
/fit/cells/1/observedtolerance77.00.00E+2passed
/fit/cells/1/structuralexactFalseFalsepassed
/fit/cells/1/expectedtolerance12.13592233009709912.135922330097122.12E-1414passed
/fit/cells/1/residualtolerance-1.4742870582410657-1.47428705824107034.61E-1514passed
/fit/cells/1/adjusted_residualtolerance-2.3617506401313415-2.3617506401216869.66E-1211passed
/fit/cells/1/deviance_residualtolerance-1.6025632461886226-1.60256324618862835.73E-1514passed
/fit/cells/2/observedtolerance1212.00.00E+2passed
/fit/cells/2/structuralexactFalseFalsepassed
/fit/cells/2/expectedtolerance10.80582524271845210.8058252427184772.49E-1414passed
/fit/cells/2/residualtolerance0.363277848594009690.363277848594001757.94E-1513passed
/fit/cells/2/adjusted_residualtolerance0.584365022724154910.5843650227249648.09E-1311passed
/fit/cells/2/deviance_residualtolerance0.356877243946823190.356877243946819473.72E-1513passed
/fit/cells/3/observedtolerance99.00.00E+2passed
/fit/cells/3/structuralexactFalseFalsepassed
/fit/cells/3/expectedtolerance10.19417475728155210.194174757281571.73E-1414passed
/fit/cells/3/residualtolerance-0.37401743675374111-0.374017436753746275.16E-1513passed
/fit/cells/3/adjusted_residualtolerance-0.58436502272468205-0.58436502272498763.06E-1312passed
/fit/cells/3/deviance_residualtolerance-0.38170108464642338-0.3817010846464284.61E-1513passed
/fit/cells/4/observedtolerance66.00.00E+2passed
/fit/cells/4/structuralexactFalseFalsepassed
/fit/cells/4/expectedtolerance6.00000000000000096.0000000000000191.78E-1414passed
/fit/cells/4/residualtolerance0-7.61454375085889e-157.61E-15passed
/fit/cells/4/deviance_residualtolerance0-0.00.00E+2passed
/fit/cells/5/observedtolerance00.00.00E+2passed
/fit/cells/5/structuralexactTrueTruepassed
/fit/cells/6/observedtolerance1010.00.00E+2passed
/fit/cells/6/structuralexactFalseFalsepassed
/fit/cells/6/expectedtolerance10.80582524271845210.8058252427184863.38E-1414passed
/fit/cells/6/residualtolerance-0.24513871083986297-0.245138710839872889.91E-1513passed
/fit/cells/6/adjusted_residualtolerance-0.39432761696020646-0.39432761696077765.71E-1311passed
/fit/cells/6/deviance_residualtolerance-0.2482842814187553-0.248284281418760665.36E-1513passed
/fit/cells/7/observedtolerance1111.00.00E+2passed
/fit/cells/7/structuralexactFalseFalsepassed
/fit/cells/7/expectedtolerance10.19417475728155210.1941747572815782.62E-1414passed
/fit/cells/7/residualtolerance0.252385750004557980.252385750004549278.71E-1513passed
/fit/cells/7/adjusted_residualtolerance0.394327616960559460.39432761696074631.87E-1312passed
/fit/cells/7/deviance_residualtolerance0.249165623320567740.249165623320556581.12E-1413passed
/fit/cells/8/observedtolerance88.00.00E+2passed
/fit/cells/8/structuralexactFalseFalsepassed
/fit/cells/8/expectedtolerance11.32038834951457111.3203883495145962.53E-1414passed
/fit/cells/8/residualtolerance-0.98686607224095824-0.98686607224096466.32E-1514passed
/fit/cells/8/adjusted_residualtolerance-1.5972317390169091-1.5972317390192332.32E-1211passed
/fit/cells/8/deviance_residualtolerance-1.0421877503663506-1.04218775036635756.87E-1514passed
/fit/cells/9/observedtolerance1414.00.00E+2passed
/fit/cells/9/structuralexactFalseFalsepassed
/fit/cells/9/expectedtolerance10.67961165048543710.6796116504854572.00E-1414passed
/fit/cells/9/residualtolerance1.0160408078481441.0160408078481376.89E-1514passed
/fit/cells/9/adjusted_residualtolerance1.59723173901834131.59723173901920758.66E-1312passed
/fit/cells/9/deviance_residualtolerance0.969230289769661520.96923028976965466.88E-1514passed
/fit/cells/10/observedtolerance55.00.00E+2passed
/fit/cells/10/structuralexactFalseFalsepassed
/fit/cells/10/expectedtolerance7.20388349514563147.2038834951456491.78E-1414passed
/fit/cells/10/residualtolerance-0.82111746995073709-0.82111746995074285.66E-1514passed
/fit/cells/10/adjusted_residualtolerance-1.2678343420780387-1.26783434207934941.31E-1211passed
/fit/cells/10/deviance_residualtolerance-0.8694501511040289-0.86945015110403586.88E-1514passed
/fit/cells/11/observedtolerance99.00.00E+2passed
/fit/cells/11/structuralexactFalseFalsepassed
/fit/cells/11/expectedtolerance6.79611650485436596.7961165048543781.24E-1414passed
/fit/cells/11/residualtolerance0.845392177291580890.84539217729157535.55E-1514passed
/fit/cells/11/adjusted_residualtolerance1.26783434207917981.2678343420793331.53E-1312passed
/fit/cells/11/deviance_residualtolerance0.80495413478834310.80495413478833755.55E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

261  support <- !(d$A == "a1" & d$B == "b3" & d$C == "c2")
262  m <- loglinear(d[support, ], n ~ A * B + C, last)
263  rows <- parameter_table(m$fit, list("A", "B", "C", c("A", "B")), last)
264  emit("STAT-REG-LOGLINEAR-GENERAL-005", "loglinearGeneralMenuItem", "axis_coverage",
265       list(factors = list("A", "B", "C"), count = "n", terms = list("A*B", "C"),
266            sampling = "poisson", coding = "last", confidence_level = 0.95,
267            structural_zeros = list("A=a1; B=b3; C=c2"), conditional_test = "off"),
268       fit_outputs(m$fit, rows, 0.95, poisson_loglik(m$fit), support),
269       list("The structural zero is dropped from the fit: 11 allowable cells, rank 7, df 4; every parameter stays estimable on the support (manual lines 178-183)."))

Whole script generators/r/regression_loglinear.R


STAT REG LOGLINEAR LOGIT 001

Case STAT-REG-LOGLINEAR-LOGIT-001 · status: passed

Reference

Evidence type
R
Source
R stats::glm(family = poisson) on the cell table with contr.treatment at the declared reference; rstandard/residuals for adjusted and deviance residuals; profile limits by uniroot on the constrained deviance; logit contrasts from the Poisson coefficients
Citation
R Core Team, stats package help: glm, rstandard.glm, residuals.glm, uniroot; Agresti (2013) Categorical Data Analysis, ch. 9 (loglinear models, sampling schemes) and 10 (logit-loglinear equivalence).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear.R · sha256 29efcb27edfe009b
Dataset
reference_data/datasets/REG_LOGLINEAR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/dfexact220.00E+2passed
/fit/rankexact10100.00E+2passed
/fit/allowable_cellsexact12120.00E+2passed
/fit/structural_zero_cellsexact000.00E+2passed
/fit/deviancetolerance4.60613660374246874.6061366037424767.10E-1514passed
/fit/pearsontolerance3.15069498063674923.15069498063674974.87E-1615passed
/fit/deviance_ptail_probability0.0999516907849418990.099951690784941553.47E-1614passed
/fit/pearson_ptail_probability0.206935632057349440.206935632057349385.65E-1715passed
/fit/log_likelihoodtolerance-24.784066654680235-24.7840666546802451.03E-1415passed
/fit/boundaryexactFalseFalsepassed
/totaltolerance109109.00.00E+2passed
/fit/parameters/0/termexact(intercept)(intercept)passed
/fit/parameters/0/labelexactpassed
/fit/parameters/0/estimatetolerance2.01495626510421212.01495626510421264.52E-1615passed
/fit/parameters/0/standard_errortolerance0.347039697342839240.34703969734544642.61E-1211passed
/fit/parameters/0/z_valuetolerance5.80612615943369815.8061261593900814.36E-1111passed
/fit/parameters/0/p_valuetail_probability6.3934794883663958e-096.393479490031075e-091.66E-189passed
/fit/parameters/0/lowertolerance1.33477095710656671.3347709571014575.11E-1211passed
/fit/parameters/0/uppertolerance2.69514157310185752.6951415731069685.11E-1211passed
/fit/parameters/1/termexactAApassed
/fit/parameters/1/labelexactA=a1A=a1passed
/fit/parameters/1/estimatetolerance-1.6097574223388402-1.60975742233884119.15E-1615passed
/fit/parameters/1/standard_errortolerance0.579564217152829690.57956421716313381.03E-1110passed
/fit/parameters/1/z_valuetolerance-2.7775307285997455-2.77753072855036544.94E-1110passed
/fit/parameters/1/p_valuetail_probability0.00547736606687787820.0054773660677101568.32E-139passed
/fit/parameters/1/lowertolerance-2.7456824146865371-2.74568241470673422.02E-1111passed
/fit/parameters/1/uppertolerance-0.47383242999114317-0.47383242997094822.02E-1110passed
/fit/parameters/2/termexactBBpassed
/fit/parameters/2/labelexactB=b1B=b1passed
/fit/parameters/2/estimatetolerance0.428510895842057480.42851089584205761.13E-1615passed
/fit/parameters/2/standard_errortolerance0.43369057091940960.433690570921742962.33E-1211passed
/fit/parameters/2/z_valuetolerance0.988056749616733950.98805674961141825.32E-1211passed
/fit/parameters/2/p_valuetail_probability0.3231248559388280.323124855941431252.60E-1211passed
/fit/parameters/2/lowertolerance-0.42150700359459925-0.42150700359917264.57E-1210passed
/fit/parameters/2/uppertolerance1.27852879527871411.27852879528328784.57E-1211passed
/fit/parameters/3/termexactBBpassed
/fit/parameters/3/labelexactB=b2B=b2passed
/fit/parameters/3/estimatetolerance0.692207676312930560.69220767631292951.11E-1514passed
/fit/parameters/3/standard_errortolerance0.417760661242049140.41776066124440692.36E-1211passed
/fit/parameters/3/z_valuetolerance1.65694796215354461.65694796214419039.35E-1211passed
/fit/parameters/3/p_valuetail_probability0.0975299963592630990.09752999636115451.89E-1210passed
/fit/parameters/3/lowertolerance-0.12658817387912358-0.12658817388374614.62E-1210passed
/fit/parameters/3/uppertolerance1.51100352650498461.5110035265096054.62E-1211passed
/fit/parameters/4/termexactCCpassed
/fit/parameters/4/labelexactC=c1C=c1passed
/fit/parameters/4/estimatetolerance-0.14321552775860485-0.14321552775860545.60E-1614passed
/fit/parameters/4/standard_errortolerance0.476842288347926520.47684228835087872.95E-1211passed
/fit/parameters/4/z_valuetolerance-0.30034149918789937-0.30034149918604111.86E-1211passed
/fit/parameters/4/p_valuetail_probability0.763916681711620970.76391668171303821.42E-1211passed
/fit/parameters/4/lowertolerance-1.0778092392262042-1.07780923923199115.79E-1211passed
/fit/parameters/4/uppertolerance0.791378183708994380.79137818371478025.79E-1211passed
/fit/parameters/5/termexactA*BA*Bpassed
/fit/parameters/5/labelexactA=a1, B=b1A=a1, B=b1passed
/fit/parameters/5/estimatetolerance1.03610759884166571.0361075988416662.22E-1615passed
/fit/parameters/5/standard_errortolerance0.594005727804850280.59400572781284427.99E-1210passed
/fit/parameters/5/z_valuetolerance1.74427206732602391.74427206730255052.35E-1110passed
/fit/parameters/5/p_valuetail_probability0.0811116599317016630.081111659935792894.09E-1210passed
/fit/parameters/5/lowertolerance-0.12812223426634306-0.128122234282010971.57E-119passed
/fit/parameters/5/uppertolerance2.20033743194967452.2003374319653431.57E-1111passed
/fit/parameters/6/termexactA*BA*Bpassed
/fit/parameters/6/labelexactA=a1, B=b2A=a1, B=b2passed
/fit/parameters/6/estimatetolerance0.983694645427167340.9836946454271686.62E-1615passed
/fit/parameters/6/standard_errortolerance0.605044539011738490.60504453902002348.28E-1210passed
/fit/parameters/6/z_valuetolerance1.62582187260776621.62582187258550472.23E-1110passed
/fit/parameters/6/p_valuetail_probability0.103987553970905910.103987553975643024.74E-1210passed
/fit/parameters/6/lowertolerance-0.20217086007847951-0.20217086009471741.62E-1110passed
/fit/parameters/6/uppertolerance2.16956015093281442.16956015094905341.62E-1111passed
/fit/parameters/7/termexactA*CA*Cpassed
/fit/parameters/7/labelexactA=a1, C=c1A=a1, C=c1passed
/fit/parameters/7/estimatetolerance1.24218282113116121.24218282113116164.30E-1615passed
/fit/parameters/7/standard_errortolerance0.415947820256192650.415947820258132041.94E-1211passed
/fit/parameters/7/z_valuetolerance2.98639098617242382.98639098615850031.39E-1111passed
/fit/parameters/7/p_valuetail_probability0.00282291465730631110.0028229146574348551.29E-1310passed
/fit/parameters/7/lowertolerance0.42694007398108380.426940073977282953.80E-1211passed
/fit/parameters/7/uppertolerance2.05742556828123832.057425568285043.80E-1211passed
/fit/parameters/8/termexactB*CB*Cpassed
/fit/parameters/8/labelexactB=b1, C=c1B=b1, C=c1passed
/fit/parameters/8/estimatetolerance-0.050317456917490816-0.0503174569174918159.99E-1613passed
/fit/parameters/8/standard_errortolerance0.571533599427957010.57153359943229764.34E-1211passed
/fit/parameters/8/z_valuetolerance-0.088039368057893919-0.088039368057227046.67E-1311passed
/fit/parameters/8/p_valuetail_probability0.929845386421102260.92984538642163235.30E-1312passed
/fit/parameters/8/lowertolerance-1.1705027277508284-1.1705027277593378.51E-1211passed
/fit/parameters/8/uppertolerance1.06986781391584661.06986781392435348.51E-1211passed
/fit/parameters/9/termexactB*CB*Cpassed
/fit/parameters/9/labelexactB=b2, C=c1B=b2, C=c1passed
/fit/parameters/9/estimatetolerance-0.61614177748740528-0.61614177748740521.07E-1615passed
/fit/parameters/9/standard_errortolerance0.576156308440477050.57615630844498494.51E-1211passed
/fit/parameters/9/z_valuetolerance-1.0694003839950303-1.06940038398666338.37E-1211passed
/fit/parameters/9/p_valuetail_probability0.284889294374922890.28488929437869153.77E-1210passed
/fit/parameters/9/lowertolerance-1.745387391496291-1.74538739150512618.84E-1211passed
/fit/parameters/9/uppertolerance0.513103836521480280.51310383653031588.84E-1210passed
/fit/cells/0/observedtolerance1818.00.00E+2passed
/fit/cells/0/structuralexactFalseFalsepassed
/fit/cells/0/expectedtolerance18.51288866124820518.512888661248182.50E-1414passed
/fit/cells/0/residualtolerance-0.11920267552505996-0.119202675525054275.69E-1513passed
/fit/cells/0/adjusted_residualtolerance-0.43066374589324136-0.43066374589520731.97E-1211passed
/fit/cells/0/deviance_residualtolerance-0.11975953493796322-0.119759534937962299.28E-1614passed
/fit/cells/1/observedtolerance77.00.00E+2passed
/fit/cells/1/structuralexactFalseFalsepassed
/fit/cells/1/expectedtolerance6.48711133875177566.4871113387517758.97E-1615passed
/fit/cells/1/residualtolerance0.201371229648845550.201371229648845933.84E-1614passed
/fit/cells/1/adjusted_residualtolerance0.430663745894748480.43066374589523434.86E-1311passed
/fit/cells/1/deviance_residualtolerance0.198801509024969130.198801509024971932.80E-1513passed
/fit/cells/2/observedtolerance1212.00.00E+2passed
/fit/cells/2/structuralexactFalseFalsepassed
/fit/cells/2/expectedtolerance12.98671199390579112.986711993905781.05E-1415passed
/fit/cells/2/residualtolerance-0.27380463939262156-0.273804639392618732.83E-1513passed
/fit/cells/2/adjusted_residualtolerance-0.83132462103370519-0.83132462103764523.94E-1211passed
/fit/cells/2/deviance_residualtolerance-0.27738646929634997-0.277386469296348361.61E-1514passed
/fit/cells/3/observedtolerance99.00.00E+2passed
/fit/cells/3/structuralexactFalseFalsepassed
/fit/cells/3/expectedtolerance8.01328800609428.0132880060941937.11E-1515passed
/fit/cells/3/residualtolerance0.348566006891597560.34856600689160022.66E-1514passed
/fit/cells/3/adjusted_residualtolerance0.831324621036285790.83132462103766561.38E-1211passed
/fit/cells/3/deviance_residualtolerance0.341756409072151380.341756409072152056.65E-1614passed
/fit/cells/4/observedtolerance66.00.00E+2passed
/fit/cells/4/structuralexactFalseFalsepassed
/fit/cells/4/expectedtolerance4.50039934484598454.5003993448459831.80E-1515passed
/fit/cells/4/residualtolerance0.706887163167161940.70688716316716277.73E-1614passed
/fit/cells/4/adjusted_residualtolerance1.76104365004221311.76104365004313579.23E-1312passed
/fit/cells/4/deviance_residualtolerance0.672248976773852310.6722489767738541.67E-1514passed
/fit/cells/5/observedtolerance00.00.00E+2passed
/fit/cells/5/structuralexactFalseFalsepassed
/fit/cells/5/expectedtolerance1.49960065515401691.49960065515401627.07E-1615passed
/fit/cells/5/residualtolerance-1.224581828688478-1.22458182868847782.42E-1615passed
/fit/cells/5/adjusted_residualtolerance-1.7610436500704356-1.76104365004313572.73E-1110passed
/fit/cells/5/deviance_residualtolerance-1.7318202303668917-1.73182023036689134.44E-1615passed
/fit/cells/6/observedtolerance1010.00.00E+2passed
/fit/cells/6/structuralexactFalseFalsepassed
/fit/cells/6/expectedtolerance9.48711133875178099.4871113387517671.42E-1414passed
/fit/cells/6/residualtolerance0.166516049617902650.16651604961790744.74E-1513passed
/fit/cells/6/adjusted_residualtolerance0.430663745894328210.43066374589524079.12E-1311passed
/fit/cells/6/deviance_residualtolerance0.165048514022362010.165048514022378131.61E-1413passed
/fit/cells/7/observedtolerance1111.00.00E+2passed
/fit/cells/7/structuralexactFalseFalsepassed
/fit/cells/7/expectedtolerance11.51288866124821411.5128886612482195.08E-1515passed
/fit/cells/7/residualtolerance-0.15115793336323219-0.151157933363233741.55E-1513passed
/fit/cells/7/adjusted_residualtolerance-0.43066374589305817-0.43066374589522792.17E-1211passed
/fit/cells/7/deviance_residualtolerance-0.15230161930408551-0.152301619304086981.47E-1514passed
/fit/cells/8/observedtolerance88.00.00E+2passed
/fit/cells/8/structuralexactFalseFalsepassed
/fit/cells/8/expectedtolerance7.01328800609420447.01328800609419647.95E-1514passed
/fit/cells/8/residualtolerance0.372588606480208690.37258860648021193.22E-1514passed
/fit/cells/8/adjusted_residualtolerance0.831324621036569680.83132462103766151.09E-1211passed
/fit/cells/8/deviance_residualtolerance0.364327463043432160.364327463043435833.67E-1513passed
/fit/cells/9/observedtolerance1414.00.00E+2passed
/fit/cells/9/structuralexactFalseFalsepassed
/fit/cells/9/expectedtolerance14.98671199390580414.986711993905791.47E-1415passed
/fit/cells/9/residualtolerance-0.25488086162173729-0.254880861621733733.56E-1513passed
/fit/cells/9/adjusted_residualtolerance-0.83132462103382854-0.83132462103764963.82E-1211passed
/fit/cells/9/deviance_residualtolerance-0.25775736140573008-0.25775736140572623.89E-1513passed
/fit/cells/10/observedtolerance55.00.00E+2passed
/fit/cells/10/structuralexactFalseFalsepassed
/fit/cells/10/expectedtolerance6.49960065515401296.4996006551540133.74E-1717passed
/fit/cells/10/residualtolerance-0.58820983887710521-0.58820983887710523.34E-1817passed
/fit/cells/10/adjusted_residualtolerance-1.7610436499981927-1.76104365004313464.49E-1110passed
/fit/cells/10/deviance_residualtolerance-0.61332948827930844-0.61332948827930843.49E-1817passed
/fit/cells/11/observedtolerance99.00.00E+2passed
/fit/cells/11/structuralexactFalseFalsepassed
/fit/cells/11/expectedtolerance7.50039934484598187.50039934484598452.67E-1515passed
/fit/cells/11/residualtolerance0.547562159837208350.54756215983720721.11E-1514passed
/fit/cells/11/adjusted_residualtolerance1.76104365001514451.76104365004314942.80E-1110passed
/fit/cells/11/deviance_residualtolerance0.530686639993973250.53068663999397161.66E-1514passed
/logit/0/logitexactlog(C=c1 / C=c2)log(C=c1 / C=c2)passed
/logit/0/termexact(intercept)(intercept)passed
/logit/0/estimatetolerance-0.14321552775860485-0.14321552775860545.60E-1614passed
/logit/0/standard_errortolerance0.476842288347926520.47684228835087872.95E-1211passed
/logit/1/logitexactlog(C=c1 / C=c2)log(C=c1 / C=c2)passed
/logit/1/termexactAApassed
/logit/1/estimatetolerance1.24218282113116121.24218282113116164.30E-1615passed
/logit/1/standard_errortolerance0.415947820256192650.415947820258132041.94E-1211passed
/logit/2/logitexactlog(C=c1 / C=c2)log(C=c1 / C=c2)passed
/logit/2/termexactBBpassed
/logit/2/estimatetolerance-0.050317456917490816-0.0503174569174918159.99E-1613passed
/logit/2/standard_errortolerance0.571533599427957010.57153359943229764.34E-1211passed
/logit/3/logitexactlog(C=c1 / C=c2)log(C=c1 / C=c2)passed
/logit/3/termexactBBpassed
/logit/3/estimatetolerance-0.61614177748740528-0.61614177748740521.07E-1615passed
/logit/3/standard_errortolerance0.576156308440477050.57615630844498494.51E-1211passed
/joint/total_associationtolerance11.27125024129356111.2712502412935451.56E-1414passed
/joint/total_dfexact330.00E+2passed
/joint/partition/0/addedexactAApassed
/joint/partition/0/lr_chi_squaretolerance9.28638042380944039.286380423809441.24E-1717passed
/joint/partition/0/dfexact110.00E+2passed
/joint/partition/0/p_valuetail_probability0.00230863879974669290.0023086387997466948.92E-1915passed
/joint/partition/1/addedexactBBpassed
/joint/partition/1/lr_chi_squaretolerance1.98486981748412111.9848698174841051.60E-1414passed
/joint/partition/1/dfexact220.00E+2passed
/joint/partition/1/p_valuetail_probability0.370673036293593440.370673036293596443.00E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

35  emit <- function(id, menu, kind, config, outputs, notes) {
36    write_expected("regression", id, list(
37      schema_version = "2.0.0", case_id = id, outputs = outputs,
38      expected_error = NULL, deferred_outputs = list(), generator_run = run))
39    write_case("regression", id, list(
40      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
41      entry_point = "tabelz.analysis.loglinear:compute_loglinear",
42      menu_items = list(menu), title = gsub("-", " ", id),
43      case_kind = kind,
44      dataset = list(path = "reference_data/datasets/REG_LOGLINEAR_01.json", sha256 = ""),
45      excluded_rows = list(), configuration = config, "repeat" = 1L,
46      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"), sha256 = ""),
47      oracle = list(
48        oracle_class = "independent_implementation",
49        source = "R stats::glm(family = poisson) on the cell table with contr.treatment at the declared reference; rstandard/residuals for adjusted and deviance residuals; profile limits by uniroot on the constrained deviance; logit contrasts from the Poisson coefficients",
50        citation = "R Core Team, stats package help: glm, rstandard.glm, residuals.glm, uniroot; Agresti (2013) Categorical Data Analysis, ch. 9 (loglinear models, sampling schemes) and 10 (logit-loglinear equivalence).",
51        generator = list(path = "generators/r/regression_loglinear.R", sha256 = ""),
52        independent = TRUE,
53        convention_notes = c(
54          notes,
55          "log mu_c = o_c + x_c' lambda over the allowable cells in reference coding (last level by default); G^2 and X^2 on df = allowable cells - rank; the Poisson, multinomial and product-multinomial fitted counts coincide and each scheme's own log-likelihood is reported (docs/manual/chapters/86-regression-and-count-models.tex:171-197).",
56          "Adjusted residual (y - mu)/sqrt(mu (1 - h)) with the Poisson leverage (R rstandard type pearson); deviance residual sign(y - mu) sqrt(d_c).",
57          "The logit model fixes the explanatory factors' interaction; each logit parameter equals the loglinear parameter whose term contains the dependent factor (manual lines 193-197; chapters/106-loglinear-extensions.tex:74-118).",
58          "Expected computed from the exported CSV."
59        )),
60      supported = TRUE, unsupported_reason = NULL,
61      requirement_ids = list(if (menu == "loglinearLogitMenuItem") "REQ-STAT-REG-LOGLINEAR-LOGIT-001" else "REQ-STAT-REG-LOGLINEAR-GENERAL-001"),
62      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/regression_loglinear.R


STAT REG LOGLINEAR LOGIT 002

Case STAT-REG-LOGLINEAR-LOGIT-002 · status: passed

Reference

Evidence type
R
Source
R stats::glm(family = poisson) on the cell table with contr.treatment at the declared reference; rstandard/residuals for adjusted and deviance residuals; profile limits by uniroot on the constrained deviance; logit contrasts from the Poisson coefficients
Citation
R Core Team, stats package help: glm, rstandard.glm, residuals.glm, uniroot; Agresti (2013) Categorical Data Analysis, ch. 9 (loglinear models, sampling schemes) and 10 (logit-loglinear equivalence).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear.R · sha256 29efcb27edfe009b
Dataset
reference_data/datasets/REG_LOGLINEAR_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/dfexact440.00E+2passed
/fit/rankexact880.00E+2passed
/fit/allowable_cellsexact12120.00E+2passed
/fit/structural_zero_cellsexact000.00E+2passed
/fit/deviancetolerance6.5910064212265886.59100642122658354.47E-1515passed
/fit/pearsontolerance4.92840967216926364.9284096721692576.19E-1514passed
/fit/deviance_ptail_probability0.159145802501620110.15914580250162042.77E-1614passed
/fit/pearson_ptail_probability0.294722302244146360.29472230224414695.56E-1614passed
/fit/log_likelihoodtolerance-22.511124565254022-22.511124565253851.71E-1314passed
/fit/boundaryexactFalseFalsepassed
/totaltolerance109109.00.00E+2passed
/fit/parameters/0/termexact(intercept)(intercept)passed
/fit/parameters/0/labelexactpassed
/fit/parameters/0/estimatetolerance2.85115104474288162.85115104474288564.01E-1514passed
/fit/parameters/0/standard_errortolerance0.220333857014775920.220333857014777561.64E-1514passed
/fit/parameters/0/z_valuetolerance12.9401404004454912.9401404004454117.85E-1414passed
/fit/parameters/0/p_valuetail_probability2.671408502104514e-382.671408502107246e-382.73E-5011passed
/fit/parameters/0/lowertolerance2.41930462041912312.41930462041912354.30E-1615passed
/fit/parameters/0/uppertolerance3.28299746906664013.28299746906664777.59E-1514passed
/fit/parameters/1/termexactAApassed
/fit/parameters/1/labelexactA=a2A=a2passed
/fit/parameters/1/estimatetolerance-0.71418565892485986-0.71418565892486262.77E-1514passed
/fit/parameters/1/standard_errortolerance0.349428758671202710.3494287586712041.28E-1514passed
/fit/parameters/1/z_valuetolerance-2.0438662851928497-2.043866285192854.85E-1615passed
/fit/parameters/1/p_valuetail_probability0.0409667586338079150.0409667586338078456.96E-1714passed
/fit/parameters/1/lowertolerance-1.3990534410829552-1.39905344108296075.50E-1514passed
/fit/parameters/1/uppertolerance-0.029317876766764561-0.0293178767667646721.11E-1614passed
/fit/parameters/2/termexactBBpassed
/fit/parameters/2/labelexactB=b2B=b2passed
/fit/parameters/2/estimatetolerance-0.17435338714477694-0.174353387144779772.83E-1513passed
/fit/parameters/2/standard_errortolerance0.296005147960381740.296005147960381851.14E-1615passed
/fit/parameters/2/z_valuetolerance-0.58902146920807252-0.58902146920808189.33E-1513passed
/fit/parameters/2/p_valuetail_probability0.555846871688525650.55584687168851926.44E-1513passed
/fit/parameters/2/lowertolerance-0.75451281638557477-0.7545128163855783.22E-1514passed
/fit/parameters/2/uppertolerance0.405806042096020950.405806042096018452.50E-1514passed
/fit/parameters/3/termexactBBpassed
/fit/parameters/3/labelexactB=b3B=b3passed
/fit/parameters/3/estimatetolerance-1.4271163556401445-1.42711635564014581.33E-1515passed
/fit/parameters/3/standard_errortolerance0.454606056566060570.45460605656619461.34E-1312passed
/fit/parameters/3/z_valuetolerance-3.1392374453171517-3.1392374453162299.23E-1312passed
/fit/parameters/3/p_valuetail_probability0.00169388132091527170.00169388132092060345.33E-1511passed
/fit/parameters/3/lowertolerance-2.3181278536634018-2.31812785366366562.64E-1312passed
/fit/parameters/3/uppertolerance-0.53610485761688742-0.53610485761662592.62E-1312passed
/fit/parameters/4/termexactCCpassed
/fit/parameters/4/labelexactC=c2C=c2passed
/fit/parameters/4/estimatetolerance-0.81093021621632722-0.81093021621632972.44E-1514passed
/fit/parameters/4/standard_errortolerance0.300462606288652570.30046260628866571.32E-1413passed
/fit/parameters/4/z_valuetolerance-2.698938900361104-2.6989389003609941.10E-1313passed
/fit/parameters/4/p_valuetail_probability0.00695609461236025230.006956094612362552.30E-1512passed
/fit/parameters/4/lowertolerance-1.3998261032431241-1.39982610324315232.82E-1413passed
/fit/parameters/4/uppertolerance-0.22203432918953037-0.222034329189506942.34E-1412passed
/fit/parameters/5/termexactA*BA*Bpassed
/fit/parameters/5/labelexactA=a2, B=b2A=a2, B=b2passed
/fit/parameters/5/estimatetolerance0.220873402779669940.220873402779672182.24E-1513passed
/fit/parameters/5/standard_errortolerance0.425079569836801730.425079569836801952.22E-1615passed
/fit/parameters/5/z_valuetolerance0.519604842134540950.51960484213454594.99E-1514passed
/fit/parameters/5/p_valuetail_probability0.60333902250049010.60333902250048653.56E-1514passed
/fit/parameters/5/lowertolerance-0.61226724466423998-0.61226724466423831.66E-1514passed
/fit/parameters/5/uppertolerance1.054014050223581.05401405022358262.63E-1514passed
/fit/parameters/6/termexactA*BA*Bpassed
/fit/parameters/6/labelexactA=a2, B=b3A=a2, B=b3passed
/fit/parameters/6/estimatetolerance1.02165124753198061.02165124753198342.84E-1514passed
/fit/parameters/6/standard_errortolerance0.570713838726697790.57071383872680451.07E-1312passed
/fit/parameters/6/z_valuetolerance1.79012874440079361.7901287444004643.29E-1312passed
/fit/parameters/6/p_valuetail_probability0.0734332168042091810.073433216804262145.30E-1412passed
/fit/parameters/6/lowertolerance-0.096927321850947656-0.096927321851154162.07E-1311passed
/fit/parameters/6/uppertolerance2.14022981691490882.1402298169151212.12E-1313passed
/fit/parameters/7/termexactA*CA*Cpassed
/fit/parameters/7/labelexactA=a2, C=c2A=a2, C=c2passed
/fit/parameters/7/estimatetolerance1.20179652490333891.20179652490334182.87E-1514passed
/fit/parameters/7/standard_errortolerance0.403940346280508730.40394034628051869.88E-1513passed
/fit/parameters/7/z_valuetolerance2.97518318229289712.9751831822928326.52E-1413passed
/fit/parameters/7/p_valuetail_probability0.00292813637129891960.00292813637129954156.22E-1612passed
/fit/parameters/7/lowertolerance0.410087994290903920.410087994290887271.67E-1413passed
/fit/parameters/7/uppertolerance1.99350505551577381.99350505551579632.25E-1413passed
/fit/cells/0/observedtolerance1818.00.00E+2passed
/fit/cells/0/structuralexactFalseFalsepassed
/fit/cells/0/expectedtolerance17.30769230769227817.3076923076923466.77E-1414passed
/fit/cells/0/residualtolerance0.166410058867575980.16641005886755941.66E-1413passed
/fit/cells/0/adjusted_residualtolerance0.416333199893227590.41633319989320352.41E-1413passed
/fit/cells/0/deviance_residualtolerance0.165318749398686130.165318749398676051.01E-1413passed
/fit/cells/1/observedtolerance77.00.00E+2passed
/fit/cells/1/structuralexactFalseFalsepassed
/fit/cells/1/expectedtolerance7.69230769230769257.6923076923077029.78E-1514passed
/fit/cells/1/residualtolerance-0.24961508830135318-0.249615088301356563.38E-1513passed
/fit/cells/1/adjusted_residualtolerance-0.41633319989321094-0.41633319989323262.16E-1413passed
/fit/cells/1/deviance_residualtolerance-0.25350714391907564-0.253507143919076971.33E-1514passed
/fit/cells/2/observedtolerance1212.00.00E+2passed
/fit/cells/2/structuralexactFalseFalsepassed
/fit/cells/2/expectedtolerance14.53846153846152214.5384615384615421.99E-1414passed
/fit/cells/2/residualtolerance-0.66575028593567875-0.66575028593568354.78E-1514passed
/fit/cells/2/adjusted_residualtolerance-1.5544396553385154-1.55443965533857225.68E-1413passed
/fit/cells/2/deviance_residualtolerance-0.68668689333679878-0.6866868933368045.21E-1514passed
/fit/cells/3/observedtolerance99.00.00E+2passed
/fit/cells/3/structuralexactFalseFalsepassed
/fit/cells/3/expectedtolerance6.46153846153846436.4615384615384577.07E-1514passed
/fit/cells/3/residualtolerance0.998625428903522680.9986254289035263.33E-1514passed
/fit/cells/3/adjusted_residualtolerance1.55443965533852451.55443965533857294.84E-1413passed
/fit/cells/3/deviance_residualtolerance0.942075034302940930.94207503430294241.44E-1514passed
/fit/cells/4/observedtolerance66.00.00E+2passed
/fit/cells/4/structuralexactFalseFalsepassed
/fit/cells/4/expectedtolerance4.15384615384615154.1538461538461631.16E-1414passed
/fit/cells/4/residualtolerance0.905821627315678060.90582162731567117.00E-1514passed
/fit/cells/4/adjusted_residualtolerance1.73622946456493541.73622946456485158.39E-1413passed
/fit/cells/4/deviance_residualtolerance0.848757721140796510.84875772114079075.78E-1514passed
/fit/cells/5/observedtolerance00.00.00E+2passed
/fit/cells/5/structuralexactFalseFalsepassed
/fit/cells/5/expectedtolerance1.8461538461538481.84615384615384854.77E-1615passed
/fit/cells/5/residualtolerance-1.3587324409735158-1.35873244097351584.62E-1716passed
/fit/cells/5/adjusted_residualtolerance-1.7362294645649374-1.73622946456486357.39E-1413passed
/fit/cells/5/deviance_residualtolerance-1.9215378456610466-1.92153784566104681.95E-1615passed
/fit/cells/6/observedtolerance1010.00.00E+2passed
/fit/cells/6/structuralexactFalseFalsepassed
/fit/cells/6/expectedtolerance8.47368421052631158.4736842105263197.09E-1515passed
/fit/cells/6/residualtolerance0.524334288933274050.52433428893327152.55E-1514passed
/fit/cells/6/adjusted_residualtolerance0.85426459259071930.85426459259071553.78E-1514passed
/fit/cells/6/deviance_residualtolerance0.509669073589027350.5096690735890234.32E-1514passed
/fit/cells/7/observedtolerance1111.00.00E+2passed
/fit/cells/7/structuralexactFalseFalsepassed
/fit/cells/7/expectedtolerance12.5263157894736812.526315789473691.03E-1415passed
/fit/cells/7/residualtolerance-0.43125358551541532-0.43125358551541822.88E-1514passed
/fit/cells/7/adjusted_residualtolerance-0.85426459259071441-0.85426459259072046.00E-1514passed
/fit/cells/7/deviance_residualtolerance-0.44048870746111735-0.44048870746111992.55E-1514passed
/fit/cells/8/observedtolerance88.00.00E+2passed
/fit/cells/8/structuralexactFalseFalsepassed
/fit/cells/8/expectedtolerance8.87719298245613698.877192982456143.51E-1515passed
/fit/cells/8/residualtolerance-0.29441322739314013-0.294413227393141241.11E-1514passed
/fit/cells/8/adjusted_residualtolerance-0.48647294600317692-0.486472946003178361.44E-1514passed
/fit/cells/8/deviance_residualtolerance-0.29947321840693619-0.29947321840693811.89E-1514passed
/fit/cells/9/observedtolerance1414.00.00E+2passed
/fit/cells/9/structuralexactFalseFalsepassed
/fit/cells/9/expectedtolerance13.12280701754385113.1228070175438631.21E-1415passed
/fit/cells/9/residualtolerance0.242148496896446770.242148496896443223.55E-1513passed
/fit/cells/9/adjusted_residualtolerance0.486472946003183970.48647294600317776.27E-1513passed
/fit/cells/9/deviance_residualtolerance0.239523235609831490.239523235609834743.25E-1513passed
/fit/cells/10/observedtolerance55.00.00E+2passed
/fit/cells/10/structuralexactFalseFalsepassed
/fit/cells/10/expectedtolerance5.64912280701754285.6491228070175581.51E-1414passed
/fit/cells/10/residualtolerance-0.27310921904406743-0.273109219044073375.94E-1513passed
/fit/cells/10/adjusted_residualtolerance-0.40713414780260815-0.407134147802617038.88E-1513passed
/fit/cells/10/deviance_residualtolerance-0.27860718095108344-0.27860718095109026.78E-1513passed
/fit/cells/11/observedtolerance99.00.00E+2passed
/fit/cells/11/structuralexactFalseFalsepassed
/fit/cells/11/expectedtolerance8.35087719298245558.350877192982482.48E-1414passed
/fit/cells/11/residualtolerance0.224626411882550810.224626411882541848.97E-1513passed
/fit/cells/11/adjusted_residualtolerance0.407134147802608870.407134147802593161.57E-1413passed
/fit/cells/11/deviance_residualtolerance0.221806724138843640.221806724138831121.25E-1413passed
/logit/0/logitexactlog(C=c2 / C=c1)log(C=c2 / C=c1)passed
/logit/0/estimatetolerance-0.81093021621632722-0.81093021621632972.44E-1514passed
/logit/0/standard_errortolerance0.300462606288652570.30046260628866571.32E-1413passed
/logit/1/logitexactlog(C=c2 / C=c1)log(C=c2 / C=c1)passed
/logit/1/estimatetolerance1.20179652490333891.20179652490334182.87E-1514passed
/logit/1/standard_errortolerance0.403940346280508730.40394034628051869.88E-1513passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

320  emit("STAT-REG-LOGLINEAR-LOGIT-002", "loglinearLogitMenuItem", "axis_coverage",
321       list(factors = list("A", "B", "C"), count = "n", dependent = list("C"),
322            terms = list("A*C"), sampling = "multinomial", coding = "first",
323            confidence_level = 0.95, conditional_test = "off"),
324       out,
325       list("glm(n ~ A*B + C + A:C, poisson) with first-level references; the fitted counts are those of the multinomial model and the logit is log(C=c2 / C=c1)."))

Whole script generators/r/regression_loglinear.R


STAT REG LOGLINEAR SELECTION 001

Case STAT-REG-LOGLINEAR-SELECTION-001 · status: passed

Reference

Evidence type
R
Source
Base R implementation of hierarchical backward elimination over stats::glm(family = poisson) fits of each generating class; stats::pchisq
Citation
R Core Team, stats package help: glm, pchisq; Agresti (2013) Categorical Data Analysis, section 9.4 (hierarchical loglinear model selection by likelihood-ratio comparisons).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear_selection.R · sha256 0288dc1141c43c92
Dataset
reference_data/datasets/REG_LOGLINEAR_02.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/final/deviancetolerance0.280520069522240840.280520069522244343.50E-1513passed
/final/dfexact440.00E+2passed
/final/deviance_ptail_probability0.991036714211564250.9910367142115642.21E-1615passed
/final/boundaryexactFalseFalsepassed
/final/generators/0exactA*BA*Bpassed
/final/generators/1exactA*CA*Cpassed
/steps/0/stepexact110.00E+2passed
/steps/0/model/0exactA*B*CA*B*Cpassed
/steps/0/removedexactA*B*CA*B*Cpassed
/steps/0/candidates/0/termexactA*B*CA*B*Cpassed
/steps/0/candidates/0/orderexact330.00E+2passed
/steps/0/candidates/0/dfexact220.00E+2passed
/steps/0/candidates/0/lr_chi_squaretolerance0.259266397067031140.25926639706703382.66E-1513passed
/steps/0/candidates/0/p_valuetail_probability0.878417576690494920.87841757669049361.34E-1514passed
/steps/1/stepexact220.00E+2passed
/steps/1/model/0exactA*BA*Bpassed
/steps/1/model/1exactA*CA*Cpassed
/steps/1/model/2exactB*CB*Cpassed
/steps/1/removedexactB*CB*Cpassed
/steps/1/candidates/0/termexactA*BA*Bpassed
/steps/1/candidates/0/orderexact220.00E+2passed
/steps/1/candidates/0/dfexact220.00E+2passed
/steps/1/candidates/0/lr_chi_squaretolerance35.60609961441016935.606099614410166.75E-1515passed
/steps/1/candidates/0/p_valuetail_probability1.8545293361962089e-081.8545293361962152e-086.27E-2314passed
/steps/1/candidates/1/termexactA*CA*Cpassed
/steps/1/candidates/1/orderexact220.00E+2passed
/steps/1/candidates/1/dfexact110.00E+2passed
/steps/1/candidates/1/lr_chi_squaretolerance10.98024771724198910.9802477172419923.33E-1515passed
/steps/1/candidates/1/p_valuetail_probability0.000920881199029118840.00092088119902911652.39E-1814passed
/steps/1/candidates/2/termexactB*CB*Cpassed
/steps/1/candidates/2/orderexact220.00E+2passed
/steps/1/candidates/2/dfexact220.00E+2passed
/steps/1/candidates/2/lr_chi_squaretolerance0.0212536724552050420.0212536724552138688.83E-1512passed
/steps/1/candidates/2/p_valuetail_probability0.989429429112607560.98942942911260314.44E-1514passed
/steps/2/stepexact330.00E+2passed
/steps/2/model/0exactA*BA*Bpassed
/steps/2/model/1exactA*CA*Cpassed
/steps/2/candidates/0/termexactA*BA*Bpassed
/steps/2/candidates/0/orderexact220.00E+2passed
/steps/2/candidates/0/dfexact220.00E+2passed
/steps/2/candidates/0/lr_chi_squaretolerance38.27231274687171238.2723127468717056.74E-1515passed
/steps/2/candidates/0/p_valuetail_probability4.8895949292945122e-094.889594929294536e-092.40E-2314passed
/steps/2/candidates/1/termexactA*CA*Cpassed
/steps/2/candidates/1/orderexact220.00E+2passed
/steps/2/candidates/1/dfexact110.00E+2passed
/steps/2/candidates/1/lr_chi_squaretolerance13.64646084970353513.6464608497035286.76E-1515passed
/steps/2/candidates/1/p_valuetail_probability0.000220656843754802550.000220656843754803581.03E-1814passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

34  emit <- function(id, kind, config, outputs, notes) {
35    write_expected("regression", id, list(
36      schema_version = "2.0.0", case_id = id, outputs = outputs,
37      expected_error = NULL, deferred_outputs = list(), generator_run = run))
38    write_case("regression", id, list(
39      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
40      entry_point = "tabelz.analysis.loglinear_selection:compute_loglinear_selection",
41      menu_items = list("loglinearSelectionMenuItem"), title = gsub("-", " ", id),
42      case_kind = kind,
43      dataset = list(path = "reference_data/datasets/REG_LOGLINEAR_02.json", sha256 = ""),
44      excluded_rows = list(), configuration = config, "repeat" = 1L,
45      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"), sha256 = ""),
46      oracle = list(
47        oracle_class = "independent_implementation",
48        source = "Base R implementation of hierarchical backward elimination over stats::glm(family = poisson) fits of each generating class; stats::pchisq",
49        citation = "R Core Team, stats package help: glm, pchisq; Agresti (2013) Categorical Data Analysis, section 9.4 (hierarchical loglinear model selection by likelihood-ratio comparisons).",
50        generator = list(path = "generators/r/regression_loglinear_selection.R", sha256 = ""),
51        independent = TRUE,
52        convention_notes = c(
53          notes,
54          "Backward elimination from the saturated model: each maximal term whose removal keeps the model hierarchical (and that is not inside a term fixed by the sampling design) is tested by G^2(reduced) - G^2(current) on the df difference; the candidate with the largest p above alpha is removed, ties going to the higher-order term and then the term earlier in table order; the search stops when no candidate exceeds alpha (docs/manual/chapters/86-regression-and-count-models.tex:193-197).",
55          "Candidates are listed by order (highest first) and then table order; df and deviances come from Poisson fits on the full positive table, so every fit is interior and the chi-square reference applies.",
56          "Expected computed from the exported CSV."
57        )),
58      supported = TRUE, unsupported_reason = NULL,
59      requirement_ids = list("REQ-STAT-REG-LOGLINEAR-SELECTION-001"),
60      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/regression_loglinear_selection.R


STAT REG LOGLINEAR SELECTION 002

Case STAT-REG-LOGLINEAR-SELECTION-002 · status: passed

Reference

Evidence type
R
Source
Base R implementation of hierarchical backward elimination over stats::glm(family = poisson) fits of each generating class; stats::pchisq
Citation
R Core Team, stats package help: glm, pchisq; Agresti (2013) Categorical Data Analysis, section 9.4 (hierarchical loglinear model selection by likelihood-ratio comparisons).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear_selection.R · sha256 0288dc1141c43c92
Dataset
reference_data/datasets/REG_LOGLINEAR_02.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/final/deviancetolerance0.280520069522240840.280520069522244343.50E-1513passed
/final/dfexact440.00E+2passed
/final/deviance_ptail_probability0.991036714211564250.9910367142115642.21E-1615passed
/final/boundaryexactFalseFalsepassed
/final/generators/0exactA*BA*Bpassed
/final/generators/1exactA*CA*Cpassed
/steps/0/stepexact110.00E+2passed
/steps/0/model/0exactA*B*CA*B*Cpassed
/steps/0/removedexactA*B*CA*B*Cpassed
/steps/0/candidates/0/termexactA*B*CA*B*Cpassed
/steps/0/candidates/0/orderexact330.00E+2passed
/steps/0/candidates/0/dfexact220.00E+2passed
/steps/0/candidates/0/lr_chi_squaretolerance0.259266397067031140.25926639706703382.66E-1513passed
/steps/0/candidates/0/p_valuetail_probability0.878417576690494920.87841757669049361.34E-1514passed
/steps/1/stepexact220.00E+2passed
/steps/1/model/0exactA*BA*Bpassed
/steps/1/model/1exactA*CA*Cpassed
/steps/1/model/2exactB*CB*Cpassed
/steps/1/removedexactB*CB*Cpassed
/steps/1/candidates/0/termexactA*BA*Bpassed
/steps/1/candidates/0/orderexact220.00E+2passed
/steps/1/candidates/0/dfexact220.00E+2passed
/steps/1/candidates/0/lr_chi_squaretolerance35.60609961441016935.606099614410166.75E-1515passed
/steps/1/candidates/0/p_valuetail_probability1.8545293361962089e-081.8545293361962152e-086.27E-2314passed
/steps/1/candidates/1/termexactA*CA*Cpassed
/steps/1/candidates/1/orderexact220.00E+2passed
/steps/1/candidates/1/dfexact110.00E+2passed
/steps/1/candidates/1/lr_chi_squaretolerance10.98024771724198910.9802477172419923.33E-1515passed
/steps/1/candidates/1/p_valuetail_probability0.000920881199029118840.00092088119902911652.39E-1814passed
/steps/1/candidates/2/termexactB*CB*Cpassed
/steps/1/candidates/2/orderexact220.00E+2passed
/steps/1/candidates/2/dfexact220.00E+2passed
/steps/1/candidates/2/lr_chi_squaretolerance0.0212536724552050420.0212536724552138688.83E-1512passed
/steps/1/candidates/2/p_valuetail_probability0.989429429112607560.98942942911260314.44E-1514passed
/steps/2/stepexact330.00E+2passed
/steps/2/model/0exactA*BA*Bpassed
/steps/2/model/1exactA*CA*Cpassed
/steps/2/candidates/0/termexactA*BA*Bpassed
/steps/2/candidates/0/orderexact220.00E+2passed
/steps/2/candidates/0/dfexact220.00E+2passed
/steps/2/candidates/0/lr_chi_squaretolerance38.27231274687171238.2723127468717056.74E-1515passed
/steps/2/candidates/0/p_valuetail_probability4.8895949292945122e-094.889594929294536e-092.40E-2314passed
/steps/2/candidates/1/termexactA*CA*Cpassed
/steps/2/candidates/1/orderexact220.00E+2passed
/steps/2/candidates/1/dfexact110.00E+2passed
/steps/2/candidates/1/lr_chi_squaretolerance13.64646084970353513.6464608497035286.76E-1515passed
/steps/2/candidates/1/p_valuetail_probability0.000220656843754802550.000220656843754803581.03E-1814passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

158  r <- backward(list(), 0.30)
159  emit("STAT-REG-LOGLINEAR-SELECTION-002", "axis_coverage", c(common, list(alpha = 0.30, highest_order_only = FALSE)),
160       selection_outputs(r),
161       list("Alpha 0.30: a candidate is removed only when its likelihood-ratio p exceeds 0.30."))

Whole script generators/r/regression_loglinear_selection.R


STAT REG LOGLINEAR SELECTION 003

Case STAT-REG-LOGLINEAR-SELECTION-003 · status: passed

Reference

Evidence type
R
Source
Base R implementation of hierarchical backward elimination over stats::glm(family = poisson) fits of each generating class; stats::pchisq
Citation
R Core Team, stats package help: glm, pchisq; Agresti (2013) Categorical Data Analysis, section 9.4 (hierarchical loglinear model selection by likelihood-ratio comparisons).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear_selection.R · sha256 0288dc1141c43c92
Dataset
reference_data/datasets/REG_LOGLINEAR_02.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/final/deviancetolerance0.280520069522240840.280520069522244343.50E-1513passed
/final/dfexact440.00E+2passed
/final/deviance_ptail_probability0.991036714211564250.9910367142115642.21E-1615passed
/final/boundaryexactFalseFalsepassed
/final/generators/0exactA*BA*Bpassed
/final/generators/1exactA*CA*Cpassed
/steps/0/stepexact110.00E+2passed
/steps/0/model/0exactA*B*CA*B*Cpassed
/steps/0/removedexactA*B*CA*B*Cpassed
/steps/0/candidates/0/termexactA*B*CA*B*Cpassed
/steps/0/candidates/0/orderexact330.00E+2passed
/steps/0/candidates/0/dfexact220.00E+2passed
/steps/0/candidates/0/lr_chi_squaretolerance0.259266397067031140.25926639706703382.66E-1513passed
/steps/0/candidates/0/p_valuetail_probability0.878417576690494920.87841757669049361.34E-1514passed
/steps/1/stepexact220.00E+2passed
/steps/1/model/0exactA*BA*Bpassed
/steps/1/model/1exactA*CA*Cpassed
/steps/1/model/2exactB*CB*Cpassed
/steps/1/removedexactB*CB*Cpassed
/steps/1/candidates/0/termexactA*BA*Bpassed
/steps/1/candidates/0/orderexact220.00E+2passed
/steps/1/candidates/0/dfexact220.00E+2passed
/steps/1/candidates/0/lr_chi_squaretolerance35.60609961441016935.606099614410166.75E-1515passed
/steps/1/candidates/0/p_valuetail_probability1.8545293361962089e-081.8545293361962152e-086.27E-2314passed
/steps/1/candidates/1/termexactA*CA*Cpassed
/steps/1/candidates/1/orderexact220.00E+2passed
/steps/1/candidates/1/dfexact110.00E+2passed
/steps/1/candidates/1/lr_chi_squaretolerance10.98024771724198910.9802477172419923.33E-1515passed
/steps/1/candidates/1/p_valuetail_probability0.000920881199029118840.00092088119902911652.39E-1814passed
/steps/1/candidates/2/termexactB*CB*Cpassed
/steps/1/candidates/2/orderexact220.00E+2passed
/steps/1/candidates/2/dfexact220.00E+2passed
/steps/1/candidates/2/lr_chi_squaretolerance0.0212536724552050420.0212536724552138688.83E-1512passed
/steps/1/candidates/2/p_valuetail_probability0.989429429112607560.98942942911260314.44E-1514passed
/steps/2/stepexact330.00E+2passed
/steps/2/model/0exactA*BA*Bpassed
/steps/2/model/1exactA*CA*Cpassed
/steps/2/candidates/0/termexactA*BA*Bpassed
/steps/2/candidates/0/orderexact220.00E+2passed
/steps/2/candidates/0/dfexact220.00E+2passed
/steps/2/candidates/0/lr_chi_squaretolerance38.27231274687171238.2723127468717056.74E-1515passed
/steps/2/candidates/0/p_valuetail_probability4.8895949292945122e-094.889594929294536e-092.40E-2314passed
/steps/2/candidates/1/termexactA*CA*Cpassed
/steps/2/candidates/1/orderexact220.00E+2passed
/steps/2/candidates/1/dfexact110.00E+2passed
/steps/2/candidates/1/lr_chi_squaretolerance13.64646084970353513.6464608497035286.76E-1515passed
/steps/2/candidates/1/p_valuetail_probability0.000220656843754802550.000220656843754803581.03E-1814passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

164  r <- backward(list(), 0.05, highest_only = TRUE)
165  emit("STAT-REG-LOGLINEAR-SELECTION-003", "axis_coverage", c(common, list(alpha = 0.05, highest_order_only = TRUE)),
166       selection_outputs(r),
167       list("highest_order_only: at each step only terms of the highest order present are candidates."))

Whole script generators/r/regression_loglinear_selection.R


STAT REG LOGLINEAR SELECTION 004

Case STAT-REG-LOGLINEAR-SELECTION-004 · status: passed

Reference

Evidence type
R
Source
Base R implementation of hierarchical backward elimination over stats::glm(family = poisson) fits of each generating class; stats::pchisq
Citation
R Core Team, stats package help: glm, pchisq; Agresti (2013) Categorical Data Analysis, section 9.4 (hierarchical loglinear model selection by likelihood-ratio comparisons).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_loglinear_selection.R · sha256 0288dc1141c43c92
Dataset
reference_data/datasets/REG_LOGLINEAR_02.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/final/deviancetolerance0.280520069522240840.280520069522244343.50E-1513passed
/final/dfexact440.00E+2passed
/final/deviance_ptail_probability0.991036714211564250.9910367142115642.21E-1615passed
/final/boundaryexactFalseFalsepassed
/final/generators/0exactA*BA*Bpassed
/final/generators/1exactA*CA*Cpassed
/steps/0/stepexact110.00E+2passed
/steps/0/model/0exactA*B*CA*B*Cpassed
/steps/0/removedexactA*B*CA*B*Cpassed
/steps/0/candidates/0/termexactA*B*CA*B*Cpassed
/steps/0/candidates/0/orderexact330.00E+2passed
/steps/0/candidates/0/dfexact220.00E+2passed
/steps/0/candidates/0/lr_chi_squaretolerance0.259266397067031140.25926639706703382.66E-1513passed
/steps/0/candidates/0/p_valuetail_probability0.878417576690494920.87841757669049361.34E-1514passed
/steps/1/stepexact220.00E+2passed
/steps/1/model/0exactA*BA*Bpassed
/steps/1/model/1exactA*CA*Cpassed
/steps/1/model/2exactB*CB*Cpassed
/steps/1/removedexactB*CB*Cpassed
/steps/1/candidates/0/termexactA*CA*Cpassed
/steps/1/candidates/0/orderexact220.00E+2passed
/steps/1/candidates/0/dfexact110.00E+2passed
/steps/1/candidates/0/lr_chi_squaretolerance10.98024771724198910.9802477172419923.33E-1515passed
/steps/1/candidates/0/p_valuetail_probability0.000920881199029118840.00092088119902911652.39E-1814passed
/steps/1/candidates/1/termexactB*CB*Cpassed
/steps/1/candidates/1/orderexact220.00E+2passed
/steps/1/candidates/1/dfexact220.00E+2passed
/steps/1/candidates/1/lr_chi_squaretolerance0.0212536724552050420.0212536724552138688.83E-1512passed
/steps/1/candidates/1/p_valuetail_probability0.989429429112607560.98942942911260314.44E-1514passed
/steps/2/stepexact330.00E+2passed
/steps/2/model/0exactA*BA*Bpassed
/steps/2/model/1exactA*CA*Cpassed
/steps/2/candidates/0/termexactA*CA*Cpassed
/steps/2/candidates/0/orderexact220.00E+2passed
/steps/2/candidates/0/dfexact110.00E+2passed
/steps/2/candidates/0/lr_chi_squaretolerance13.64646084970353513.6464608497035286.76E-1515passed
/steps/2/candidates/0/p_valuetail_probability0.000220656843754802550.000220656843754803581.03E-1814passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

170  r <- backward(list(c("A", "B")), 0.05)
171  emit("STAT-REG-LOGLINEAR-SELECTION-004", "axis_coverage",
172       list(factors = list("A", "B", "C"), count = "n", sampling = "product_multinomial",
173            fixed_margins = list("A", "B"), coding = "last", confidence_level = 0.95,
174            conditional_test = "off", alpha = 0.05, highest_order_only = FALSE),
175       selection_outputs(r),
176       list("Product-multinomial with A, B totals fixed: A*B and its subterms are never candidates; the Poisson fitted counts and G^2 are unchanged."))

Whole script generators/r/regression_loglinear_selection.R


No missing values bypass imputation and give the ordinary OLS fit

Case STAT-REG-MI-NOMISS-001 · status: passed

Reference

Evidence type
Analytical
Source
Hand calculation
Citation
Least-squares normal equations; predictive mean matching donor rule
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/REG_MI_COMPLETE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/modeexactno_missingno_missingpassed
/terms/[term=Intercept]/estimatetolerance2.22.19999999999999841.60E-1515passed
/terms/[term=x]/estimatetolerance0.60.60000000000000044.22E-1615passed
/terms/[term=x]/std_errortolerance0.282842712474619010.282842712474618955.86E-1715passed
/terms/[term=Intercept]/std_errortolerance0.938083151964685950.93808315196468563.37E-1615passed
/terms/[term=x]/dftolerance33.00.00E+215passed
/terms/[term=x]/between_variancetolerance00.00.00E+215passed

Convention


Predictive mean matching copies an observed donor value

Case STAT-REG-MI-PMM-001 · status: passed

Reference

Evidence type
Analytical
Source
Hand calculation
Citation
Least-squares normal equations; predictive mean matching donor rule
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/REG_MI_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/1/xtolerance220.00E+215passed
/4/xtolerance220.00E+215passed
/0/xtolerance220.00E+215passed
/5/ytolerance660.00E+215passed

Convention


Replaying the completed-dataset operation recreates the recorded table

Case STAT-REG-MI-TABLE-001 · status: passed

Reference

Evidence type
Analytical
Source
Hand calculation
Citation
Least-squares normal equations; predictive mean matching donor rule
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/REG_MI_AUXILIARY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/frame/1/xtolerance220.00E+215passed
/frame/4/xtolerance220.00E+215passed
/frame/3/xtolerance220.00E+215passed
/frame/4/ztolerance550.00E+215passed
/row_ids_uniqueexactTrueTruepassed

Convention


Regression setting STAT-REG-MISSING-EFFECTS-001

Case STAT-REG-MISSING-EFFECTS-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observationsexact68680.00E+2passed
/residual_variancetolerance0.239793702533092010.239793702533091965.31E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

168  effects_pop <- stats::lm(y ~ x + group, data = linear_pop)
169  emit("STAT-REG-MISSING-EFFECTS-001", "tabelz.executor.linear_effects",
170       "linearEffectsMenuItem", "missing_excluded",
171       list(response_column = "y", factor_columns = list("group"), covariate_columns = list("x")),
172       list("/observations" = exact(nrow(linear_pop), "retained effects rows"),
173            "/residual_variance" = tol(summary(effects_pop)$sigma^2, "retained effects residual variance")),
174       list("Missing selected values are omitted listwise after active exclusions."),
175       dataset = "REG_MISSING", excluded = list(2L, 7L))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-MISSING-GENERAL-001

Case STAT-REG-MISSING-GENERAL-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observationsexact68680.00E+2passed
/residual_variancetolerance0.239793702533091960.239793702533091982.47E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

157  linear_pop <- remaining[stats::complete.cases(remaining[c("y", "x", "group")]), , drop = FALSE]
158  general_pop <- stats::lm(y ~ x + group, data = linear_pop,
159                           contrasts = list(group = stats::contr.sum(3)))
160  emit("STAT-REG-MISSING-GENERAL-001", "tabelz.executor.general_linear",
161       "generalLinearMenuItem", "missing_excluded",
162       list(response_column = "y", factor_columns = list("group"),
163            covariate_columns = list("x"), test_type = "III"),
164       list("/observations" = exact(nrow(linear_pop), "retained general-linear rows"),
165            "/residual_variance" = tol(summary(general_pop)$sigma^2, "retained GLM residual variance")),
166       list("Missing selected values are omitted listwise after active exclusions."),
167       dataset = "REG_MISSING", excluded = list(2L, 7L))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-MISSING-NB-001

Case STAT-REG-MISSING-NB-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact68680.00E+2passed
/coefficients/1/coefficienttolerance0.160411116989466560.160411116989492952.64E-1412passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

146  nb_pop <- remaining[stats::complete.cases(remaining[c("nb_count", "x", "exposure")]), , drop = FALSE]
147  nb_pop_fit <- MASS::glm.nb(nb_count ~ x + offset(log(exposure)), data = nb_pop,
148                             control = stats::glm.control(maxit = 1000, epsilon = 1e-12))
149  emit("STAT-REG-MISSING-NB-001", "tabelz.analysis.negative_binomial:compute_negative_binomial",
150       "nbRegressionMenuItem", "missing_excluded",
151       list(count = "nb_count", predictors = list("x"), exposure = "exposure"),
152       list("/n" = exact(nrow(nb_pop), "retained NB2 rows"),
153            "/coefficients/1/coefficient" = tol(coef(nb_pop_fit)[["x"]], "retained NB2 slope")),
154       list("Two active exclusions are applied before complete-case NB2 fitting."),
155       dataset = "REG_MISSING", excluded = list(2L, 7L))

Whole script generators/r/regression_options.R


Small-sample boundary STAT-REG-MIXED-MIN-001

Case STAT-REG-MIXED-MIN-001 · status: passed

Reference

Evidence type
R
Source
R nlme/MASS/pls/stats
Citation
R nlme::lme, MASS::lda, pls::plsr and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_b.R · sha256 e3e09e1965800ae1
Dataset
reference_data/datasets/REG_MIN_MIXED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance00.00.00E+2passed
/analyzed_countexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

22  types <- c(distance="Float64",age="Float64",subject="String",sex="String",
23             time="Int64",age_sq="Float64")
24  four <- orth[c(1L,2L,5L,6L),]
25  export_dataset(four,"REG_MIN_MIXED","Two repeated subjects with two observations each",
26                 "R nlme random-intercept small-sample boundary",
27                 "Derived from R nlme::Orthodont; retain attribution.",types)
28  m <- read_dataset("REG_MIN_MIXED")
29  fit <- nlme::lme(distance~age,random=~1|subject,data=m,method="REML",
30                   control=nlme::lmeControl(opt="optim"))
31  emit("STAT-REG-MIXED-MIN-001","tabelz.executor.mixed_models","mixedModelsMenuItem",
32       "REG_MIN_MIXED",
33       list(response_column="distance",group_column="subject",
34            fixed_numeric_columns=list("age"),estimation="reml"),
35       list("/coefficients/1/estimate"=list(comparison = "tolerance", value = "0", absolute_tolerance = "1e-12", relative_tolerance = "0",
36                  rationale = paste("Exactly zero by the design's symmetry; any implementation returns rounding",
37                                    "noise of order 1e-15, so an absolute 1e-12 replaces a relative tier that",
38                                    "cannot compare with zero. R gives", format(nlme::fixef(fit)[["age"]], digits = 17))),
39            "/analyzed_count"=exact(nrow(m),"four analyzed rows")),
40       c("Two groups have two complete observations each, meeting the random-intercept minimum."))

Whole script generators/r/regression_advanced_minimum_b.R


Regression II STAT-REG-MIXED-ML-001

Case STAT-REG-MIXED-ML-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats
Citation
R nlme::Orthodont, nlme::lme, plm::plm, stats::nls/lm/step documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_core.R · sha256 b7543c65b2778e63
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.660185185185185540.66018518518518514.43E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

60  mixed_ml <- nlme::lme(distance ~ age, random = ~1 | subject, data = d,
61                        method = "ML", control = nlme::lmeControl(opt = "optim"))
62  emit("STAT-REG-MIXED-ML-001", "tabelz.executor.mixed_models",
63       "mixedModelsMenuItem", "REG_ORTHODONT",
64       list(response_column = "distance", group_column = "subject",
65            fixed_numeric_columns = list("age"), estimation = "ml"),
66       list("/coefficients/1/estimate" = tol(nlme::fixef(mixed_ml)[["age"]], "ML age slope")),
67       list("R nlme maximum likelihood selected explicitly; compare only fixed slope."),
68       kind = "nominal")

Whole script generators/r/regression_advanced_core.R


Regression II STAT-REG-MIXED-REML-001

Case STAT-REG-MIXED-REML-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats
Citation
R nlme::Orthodont, nlme::lme, plm::plm, stats::nls/lm/step documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_core.R · sha256 b7543c65b2778e63
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.660185185185185430.6601851851851854.44E-1615passed
/analyzed_countexact1081080.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

50  mixed <- nlme::lme(distance ~ age, random = ~1 | subject, data = d,
51                     method = "REML", control = nlme::lmeControl(opt = "optim"))
52  emit("STAT-REG-MIXED-REML-001", "tabelz.executor.mixed_models",
53       "mixedModelsMenuItem", "REG_ORTHODONT",
54       list(response_column = "distance", group_column = "subject",
55            fixed_numeric_columns = list("age"), estimation = "reml"),
56       list("/coefficients/1/estimate" = tol(nlme::fixef(mixed)[["age"]], "REML age slope"),
57            "/analyzed_count" = exact(nrow(d), "complete repeated-measures rows")),
58       list("R nlme random intercept; REML is the manual default. No fallback estimator."))

Whole script generators/r/regression_advanced_core.R


Regression II STAT-REG-MIXED-STRUCT-SLOPE-001

Case STAT-REG-MIXED-STRUCT-SLOPE-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats
Citation
R nlme::Orthodont, nlme::lme, plm::plm, stats::nls/lm/step documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_core.R · sha256 b7543c65b2778e63
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.660185185185185210.66018518518518521.79E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

70  random_slope <- nlme::lme(distance ~ age, random = ~age | subject, data = d,
71                            method = "REML", control = nlme::lmeControl(opt = "optim"))
72  emit("STAT-REG-MIXED-STRUCT-SLOPE-001", "tabelz.executor.mixed_model_structures",
73       "mixedModelStructuresMenuItem", "REG_ORTHODONT",
74       list(response_column = "distance", fixed_numeric_columns = list("age"),
75            model_mode = "random_slope", group_column = "subject",
76            random_slope_column = "age", estimation = "reml"),
77       list("/coefficients/1/estimate" = tol(nlme::fixef(random_slope)[["age"]],
78                                                "random-slope fixed age effect")),
79       list("R nlme correlated random intercept and age slope per subject, REML."))

Whole script generators/r/regression_advanced_core.R


Regression edge STAT-REG-MIXED-UNBALANCED-001

Case STAT-REG-MIXED-UNBALANCED-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/rpart
Citation
R nlme::lme, plm::plm, stats::nls/lm/step, rpart::rpart documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_a.R · sha256 b8ea61b61080b6f8
Dataset
reference_data/datasets/REG_EDGE_MIXED_UNBALANCED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.181188118906150690.181188118865240474.09E-119passed
/analyzed_countexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

31  fit <- lme4::lmer(distance~age+(1|subject),data=u,REML=TRUE,
32                    control=lme4::lmerControl(optCtrl=list(ftol_abs=1e-15,xtol_abs=1e-15),
33                                              check.conv.singular="ignore"))
34  emit("STAT-REG-MIXED-UNBALANCED-001","tabelz.executor.mixed_models","mixedModelsMenuItem",
35       "REG_EDGE_MIXED_UNBALANCED",
36       list(response_column="distance",group_column="subject",
37            fixed_numeric_columns=list("age"),estimation="reml"),
38       list("/coefficients/1/estimate"=tol(lme4::fixef(fit)[["age"]],"REML fixed age slope (lme4, converged)"),
39            "/analyzed_count"=exact(as.integer(nrow(u)),"five complete rows")),
40       c("One group is exactly at the two-observation minimum while the other has three rows."))

Whole script generators/r/regression_advanced_special_a.R


STAT REG NB 001

Case STAT-REG-NB-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS and nnet
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_core.R · sha256 4b30ab4a6f513e54
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance0.188853010916972570.18885301091698381.12E-1413passed
/alphatolerance0.21725322805809980.2172532280558472.25E-1210passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

124  nb_entry <- "tabelz.analysis.negative_binomial:compute_negative_binomial"
125  nb_fit <- MASS::glm.nb(nb_count ~ x + offset(log(exposure)), data = d,
126                         control = stats::glm.control(maxit = 1000, epsilon = 1e-12))
127  emit("STAT-REG-NB-001", nb_entry, "nbRegressionMenuItem", "nominal",
128       list(count = "nb_count", predictors = list("x"), exposure = "exposure"),
129       list("/coefficients/1/coefficient" = tol(coef(nb_fit)[["x"]], "NB2 log-rate slope"),
130            "/alpha" = tol(1 / nb_fit$theta, "NB2 dispersion alpha")),
131       list("MASS glm.nb theta is inverse of Tabelz NB2 alpha; offset is log exposure."))

Whole script generators/r/regression_core.R


STAT REG NOMINAL 001

Case STAT-REG-NOMINAL-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS and nnet
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_core.R · sha256 4b30ab4a6f513e54
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/coefficienttolerance8.9213382441815576e-058.921330315026448e-057.93E-116passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

105  nominal_entry <- "tabelz.analysis.nominal_logistic:compute_nominal_logistic_regression"
106  d$nominal_factor <- factor(d$nominal, levels = c("A", "B", "C"))
107  nominal_fit <- nnet::multinom(nominal_factor ~ x, data = d, trace = FALSE,
108                                maxit = 1000, reltol = 1e-15)
109  emit("STAT-REG-NOMINAL-001", nominal_entry, "nominalRegressionMenuItem", "nominal",
110       list(outcome_column = "nominal", reference_value = "A", predictors = list("x")),
111       list("/coefficients/0/coefficient" = tol(coef(nominal_fit)["B", "(Intercept)"],
112                                                "B vs A intercept")),
113       list("R nnet multinom uses A as first factor level and reference; tight convergence."))

Whole script generators/r/regression_core.R


Regression edge STAT-REG-NONLINEAR-BOUND-001

Case STAT-REG-NONLINEAR-BOUND-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/rpart
Citation
R nlme::lme, plm::plm, stats::nls/lm/step, rpart::rpart documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_a.R · sha256 b8ea61b61080b6f8
Dataset
reference_data/datasets/REG_EDGE_DECAY_BOUND.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/estimates/0tolerance109.9999999998489121.51E-1010passed
/eligible_countexact12120.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

56  curve <- data.frame(x=seq(0,5,length.out=12L))
57  curve$y <- 1+10*exp(-0.8*curve$x)
58  export_dataset(curve,"REG_EDGE_DECAY_BOUND","Exact exponential decay with amplitude at upper bound",
59                 "R nls port bound-active fit",
60                 "Constructed validation data; no third-party license.",
61                 c(x="Float64",y="Float64"))
62  c <- read_dataset("REG_EDGE_DECAY_BOUND")
63  fit <- stats::nls(y~c0+a*exp(-k*x),data=c,
64                    start=list(a=9.5,k=0.8,c0=1),algorithm="port",
65                    lower=c(a=0,k=0.01,c0=0),upper=c(a=10,k=3,c0=5),
66                    control=stats::nls.control(maxiter=1000,tol=1e-12))
67  emit("STAT-REG-NONLINEAR-BOUND-001","tabelz.executor.nonlinear_regression",
68       "nonlinearRegressionMenuItem","REG_EDGE_DECAY_BOUND",
69       list(x_column="x",response="y",curve="exponential_decay",
70            parameters=list(list(name="a",start=9.5,lower=0,upper=10),
71                            list(name="k",start=0.8,lower=0.01,upper=3),
72                            list(name="c",start=1,lower=0,upper=5))),
73       list("/artifact/estimates/0"=tol(coef(fit)[["a"]],"bound-active amplitude"),
74            "/eligible_count"=exact(as.integer(nrow(c)),"all twelve rows eligible")),
75       c("True amplitude equals upper bound 10; R port and Tabelz use original response scale."))

Whole script generators/r/regression_advanced_special_a.R


Regression II STAT-REG-NONLINEAR-DECAY-001

Case STAT-REG-NONLINEAR-DECAY-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats
Citation
R nlme::Orthodont, nlme::lme, plm::plm, stats::nls/lm/step documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_core.R · sha256 b7543c65b2778e63
Dataset
reference_data/datasets/REG_DECAY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/estimates/0tolerance4.01261153829097594.0126115383263853.54E-1111passed
/artifact/estimates/1tolerance0.812588245136491970.81258824555635924.20E-109passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 91  decay <- stats::nls(y ~ c + a * exp(-k * x), data = curve_data,
 92                      start = list(a = 4, k = 0.8, c = 1),
 93                      algorithm = "port", lower = c(a = 0, k = 0.01, c = 0),
 94                      upper = c(a = 10, k = 3, c = 5),
 95                      control = stats::nls.control(maxiter = 1000, tol = 1e-12))
 96  emit("STAT-REG-NONLINEAR-DECAY-001", "tabelz.executor.nonlinear_regression",
 97       "nonlinearRegressionMenuItem", "REG_DECAY",
 98       list(x_column = "x", response = "y", curve = "exponential_decay",
 99            parameters = list(list(name = "a", start = 4, lower = 0, upper = 10),
100                              list(name = "k", start = 0.8, lower = 0.01, upper = 3),
101                              list(name = "c", start = 1, lower = 0, upper = 5))),
102       list("/artifact/estimates/0" = tol(coef(decay)[["a"]], "decay amplitude"),
103            "/artifact/estimates/1" = tol(coef(decay)[["k"]], "decay rate")),
104       list("R nls port bounds match the manual's strictly positive rate and original-response-scale loss."))

Whole script generators/r/regression_advanced_core.R


Small-sample boundary STAT-REG-NONLINEAR-MIN-001

Case STAT-REG-NONLINEAR-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats, plm, rpart
Citation
R stats::nls/lm, plm::plm, rpart::rpart documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_a.R · sha256 e69dea823e2a8f59
Dataset
reference_data/datasets/REG_MIN_DECAY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/estimates/0tolerance44.00.00E+2passed
/eligible_countexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

35  curve <- data.frame(x=c(0,1,2,3))
36  curve$y <- 1+4*exp(-0.8*curve$x)
37  export_dataset(curve,"REG_MIN_DECAY","Four exact exponential-decay observations",
38                 "R nls port three-parameter small-sample boundary",
39                 "Constructed validation data; no third-party license.",
40                 c(x="Float64",y="Float64"))
41  c <- read_dataset("REG_MIN_DECAY")
42  fit <- stats::nls(y~c0+a*exp(-k*x),data=c,start=list(a=4,k=0.8,c0=1),
43                    algorithm="port",lower=c(a=0,k=0.01,c0=0),
44                    upper=c(a=10,k=3,c0=5),
45                    control=stats::nls.control(maxiter=1000,tol=1e-12))
46  emit("STAT-REG-NONLINEAR-MIN-001","tabelz.executor.nonlinear_regression",
47       "nonlinearRegressionMenuItem","REG_MIN_DECAY",
48       list(x_column="x",response="y",curve="exponential_decay",
49            parameters=list(list(name="a",start=4,lower=0,upper=10),
50                            list(name="k",start=0.8,lower=0.01,upper=3),
51                            list(name="c",start=1,lower=0,upper=5))),
52       list("/artifact/estimates/0"=tol(coef(fit)[["a"]],"exact decay amplitude","1e-6"),
53            "/eligible_count"=exact(nrow(c),"all four fit rows")),
54       c("Four observations identify a three-parameter exponential-decay curve; R port uses matching bounds."))

Whole script generators/r/regression_advanced_minimum_a.R


Regression setting STAT-REG-NUMERIC-BINARY-001

Case STAT-REG-NUMERIC-BINARY-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/accuracytolerance0.930555555555555580.93055555555555562.27E-1918passed
/coefficients/1/odds_ratio_ci_lowertolerance0.66810746196598170.66810746196598152.25E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

50  binary <- stats::glm(I(binary == "event") ~ x, family = stats::binomial(), data = d)
51  binary_p <- as.vector(stats::predict(binary, type = "response"))
52  binary_actual <- d$binary == "event"
53  binary_estimate <- coef(binary)[["x"]]
54  binary_se <- summary(binary)$coefficients["x", "Std. Error"]
55  emit("STAT-REG-NUMERIC-BINARY-001",
56       "tabelz.analysis.logistic_regression:compute_binary_logistic_regression",
57       "logisticRegressionMenuItem", "axis_coverage",
58       list(outcome_column = "binary", event_value = "event", predictors = list("x"),
59            threshold = 0.3, confidence_level = 0.9),
60       list("/accuracy" = tol(mean((binary_p >= 0.3) == binary_actual), "0.3-threshold accuracy"),
61            "/coefficients/1/odds_ratio_ci_lower" = tol(exp(binary_estimate -
62               stats::qnorm(0.95) * binary_se), "90 percent Wald OR lower")),
63       list("Threshold 0.3 changes only classification; Wald interval uses z[0.95]."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-NUMERIC-EFFECTS-CONFIDENCE-001

Case STAT-REG-NUMERIC-EFFECTS-CONFIDENCE-001 · status: passed

Reference

Evidence type
Analytical
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

324  emit("STAT-REG-NUMERIC-EFFECTS-CONFIDENCE-001", effects_entry, "linearEffectsMenuItem",
325       "invalid_input",
326       list(response_column = "y", factor_columns = list("group"),
327            covariate_columns = list("x"), confidence_level = 0.9, max_rows = 100L),
328       structure(list(), names = character()),
329       list("Manual chapter 07: individual 95% t intervals; a 90% request must be refused.",
330            "Until 2026-09-25 this case wrongly expected a 90% interval; -095-001 checks the 95% interval."),
331       error = list(error_class = "ValidationError", message_contains = "fixed at 95%"))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-NUMERIC-EFFECTS-CONFIDENCE-095-001

Case STAT-REG-NUMERIC-EFFECTS-CONFIDENCE-095-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/adjusted_means/0/lowertolerance2.67476551854089632.6747655185408951.35E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

332  emit("STAT-REG-NUMERIC-EFFECTS-CONFIDENCE-095-001", effects_entry, "linearEffectsMenuItem",
333       "axis_coverage",
334       list(response_column = "y", factor_columns = list("group"),
335            covariate_columns = list("x"), confidence_level = 0.95, max_rows = 100L),
336       list("/adjusted_means/0/lower" = tol(effects_ci, "95 percent effects mean A lower")),
337       list("Individual t interval from R predict.lm fit and se.fit at complete-case mean x."))

Whole script generators/r/regression_options.R


Regression completion STAT-REG-NUMERIC-EFFECTS-MAXROWS-001

Case STAT-REG-NUMERIC-EFFECTS-MAXROWS-001 · status: passed

Reference

Evidence type
R
Source
R regression fit
Citation
R stats::glm/lm and car::Anova help pages; Tabelz manual chapters 07 and 08
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_lastmile.R · sha256 512832af1db86a5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observationsexact72720.00E+2passed
/residual_variancetolerance0.237964434289452710.23796443428945252.18E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

71  effects_fit <- stats::lm(y ~ group + x, data = d)
72  emit("STAT-REG-NUMERIC-EFFECTS-MAXROWS-001", "tabelz.executor.linear_effects",
73       "linearEffectsMenuItem", "axis_coverage", "REG_CORE",
74       list(response_column = "y", factor_columns = list("group"),
75            covariate_columns = list("x"), max_rows = 100L),
76       list("/observations" = exact(nrow(d), "72 below explicit row cap 100"),
77            "/residual_variance" = tol(summary(effects_fit)$sigma^2,
78                                        "linear-effects residual variance")),
79       list("Manual fixes individual intervals at 95%; max_rows=100 is the valid non-default numeric setting."))

Whole script generators/r/regression_lastmile.R


Fit axis STAT-REG-NUMERIC-FIT-ALPHA-001

Case STAT-REG-NUMERIC-FIT-ALPHA-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/lower_citolerance1.3404681433560991.34046814335609922.07E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

132  alpha_fit <- stats::lm(y ~ x + z, data = d)
133  ci <- stats::confint(alpha_fit, level = 0.9)
134  emit("STAT-REG-NUMERIC-FIT-ALPHA-001", fit_entry, "REG_CORE",
135       list(y_columns = list("y"), x_columns = list("x", "z"), alpha_level = 0.1),
136       list("/responses/0/parameters/1/lower_ci" = tol(ci["x", 1], "90 percent x lower CI")),
137       list("The manual passes alpha=1-confidence to the OLS t interval; 0.1 gives 90 percent."),
138       kind = "axis_coverage")

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-NUMERIC-FIT-DEGREE3-001

Case STAT-REG-NUMERIC-FIT-DEGREE3-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/3/estimatetolerance0.00549718289065148380.0054971828906519985.14E-1613passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

72  poly3 <- stats::lm(y ~ x + I(x^2) + I(x^3), data = p)
73  emit("STAT-REG-NUMERIC-FIT-DEGREE3-001", fit_entry, "REG_POSITIVE",
74       list(y_columns = list("y"), x_columns = list("x"), model_type = "polynomial", degree = 3L),
75       list("/responses/0/parameters/3/estimate" = tol(coef(poly3)[[4]], "cubic coefficient")),
76       list("Non-default polynomial degree three, with powers x, x-squared and x-cubed."),
77       kind = "axis_coverage")

Whole script generators/r/regression_fit_axes.R


Fit axis STAT-REG-NUMERIC-FIT-HOLDOUT-001

Case STAT-REG-NUMERIC-FIT-HOLDOUT-001 · status: passed

Reference

Evidence type
R
Source
R lm and matrix algebra
Citation
R stats::lm, sandwich::vcovHC documentation; Tabelz manual chapter 07
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, sandwich 3.1.3, jsonlite 2.0.0
Reference script
generators/r/regression_fit_axes.R · sha256 2e4f649c93f83e5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/training_observationsexact57570.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

140  emit("STAT-REG-NUMERIC-FIT-HOLDOUT-001", fit_entry, "REG_CORE",
141       list(y_columns = list("y"), x_columns = list("x", "z"), split_percent = 80L),
142       list("/responses/0/training_observations" = exact(floor(0.8 * nrow(d)),
143                                                           "80 percent training share")),
144       list("The holdout split is Python-seeded; only its deterministic training count is referenceable in R."),
145       kind = "axis_coverage",
146       deferred = list(list(path = "/responses/0/parameters/1/estimate",
147                            reason = "Python's seeded row split is not the same as an R RNG split.")))

Whole script generators/r/regression_fit_axes.R


Regression setting STAT-REG-NUMERIC-GENERAL-CONFIDENCE-001

Case STAT-REG-NUMERIC-GENERAL-CONFIDENCE-001 · status: passed

Reference

Evidence type
Analytical
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

306  emit("STAT-REG-NUMERIC-GENERAL-CONFIDENCE-001", general_entry, "generalLinearMenuItem",
307       "invalid_input", config_general_conf, structure(list(), names = character()),
308       list("Manual chapter 08: individual 95% t intervals; a 90% request must be refused.",
309            "Until 2026-09-25 this case wrongly expected a 90% interval; -095-001 checks the 95% interval."),
310       error = list(error_class = "ValidationError", message_contains = "fixed at 95%"))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-NUMERIC-GENERAL-CONFIDENCE-095-001

Case STAT-REG-NUMERIC-GENERAL-CONFIDENCE-095-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/adjusted_means/0/lowertolerance2.67476551854089582.6747655185408963.67E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

311  config_general_conf$confidence_level <- 0.95
312  emit("STAT-REG-NUMERIC-GENERAL-CONFIDENCE-095-001", general_entry, "generalLinearMenuItem",
313       "axis_coverage", config_general_conf,
314       list("/adjusted_means/0/lower" = tol(general_ci, "95 percent adjusted mean A lower")),
315       list("Individual t interval from R predict.lm fit and se.fit at complete-case mean x; no simultaneous adjustment."))

Whole script generators/r/regression_options.R


Regression completion STAT-REG-NUMERIC-GENERAL-MAXROWS-001

Case STAT-REG-NUMERIC-GENERAL-MAXROWS-001 · status: passed

Reference

Evidence type
R
Source
R regression fit
Citation
R stats::glm/lm and car::Anova help pages; Tabelz manual chapters 07 and 08
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_lastmile.R · sha256 512832af1db86a5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observationsexact72720.00E+2passed
/residual_variancetolerance0.237964434289452550.237964434289452552.28E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

59  general_entry <- "tabelz.executor.general_linear"
60  general_fit <- stats::lm(y ~ group + x, data = d,
61                           contrasts = list(group = stats::contr.sum(3)))
62  emit("STAT-REG-NUMERIC-GENERAL-MAXROWS-001", general_entry,
63       "generalLinearMenuItem", "axis_coverage", "REG_CORE",
64       list(response_column = "y", factor_columns = list("group"),
65            covariate_columns = list("x"), max_rows = 100L),
66       list("/observations" = exact(nrow(d), "72 below explicit row cap 100"),
67            "/residual_variance" = tol(summary(general_fit)$sigma^2,
68                                        "general-linear residual variance")),
69       list("Manual fixes individual intervals at 95%; max_rows=100 is the valid non-default numeric setting."))

Whole script generators/r/regression_lastmile.R


Regression setting STAT-REG-NUMERIC-NB-001

Case STAT-REG-NUMERIC-NB-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/rate_ratio_ci_lowertolerance1.10188057683256081.1018805768327742.13E-1312passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

133  nb_se <- sqrt(diag(nb_joint_covariance(nb)))[["x"]]
134  stopifnot(abs(nb_se - 0.05583151840027438) < 1e-12)
135  emit("STAT-REG-NUMERIC-NB-001", "tabelz.analysis.negative_binomial:compute_negative_binomial",
136       "nbRegressionMenuItem", "axis_coverage",
137       list(count = "nb_count", predictors = list("x"), exposure = "exposure",
138            confidence_level = 0.9, max_rows = 100L),
139       list("/coefficients/1/rate_ratio_ci_lower" = tol(exp(coef(nb)[["x"]] -
140           stats::qnorm(0.95) * nb_se), "90 percent NB2 Wald rate-ratio lower, joint observed information")),
141       list("MASS theta=1/alpha; confidence 0.9 and row cap 100 are explicit.",
142            "Issue #11 ruling: joint observed-information SE from the analytic NB2 Hessian; until 2026-09-25 this case used glm.nb's conditional expected-information SE."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-NUMERIC-NOMINAL-001

Case STAT-REG-NUMERIC-NOMINAL-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/rrr_ci_lowertolerance0.721948315282813110.72194831702213621.74E-98passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

79  nom <- d
80  nom$response <- factor(nom$nominal, levels = c("C", "A", "B"))
81  nom_fit <- nnet::multinom(response ~ x, data = nom, trace = FALSE,
82                            maxit = 1000, reltol = 1e-15)
83  nom_se <- summary(nom_fit)$standard.errors["A", "x"]
84  emit("STAT-REG-NUMERIC-NOMINAL-001",
85       "tabelz.analysis.nominal_logistic:compute_nominal_logistic_regression",
86       "nominalRegressionMenuItem", "axis_coverage",
87       list(outcome_column = "nominal", reference_value = "C", predictors = list("x"),
88            confidence_level = 0.9),
89       list("/coefficients/1/rrr_ci_lower" = tol(exp(coef(nom_fit)["A", "x"] -
90           stats::qnorm(0.95) * nom_se), "A-vs-C 90 percent Wald RRR lower")),
91       list("R first level C is explicit reference; numeric confidence setting is non-default."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-NUMERIC-ORDINAL-001

Case STAT-REG-NUMERIC-ORDINAL-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/odds_ratio_ci_lowertolerance0.827512144147965260.8275122505595461.06E-76passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

65  ord <- d
66  ord$response <- ordered(ord$ordinal, levels = c("low", "middle", "high"))
67  ord_fit <- MASS::polr(response ~ x, data = ord, method = "logistic", Hess = TRUE,
68                        control = list(maxit = 1000, reltol = 1e-15))
69  ord_se <- summary(ord_fit)$coefficients["x", "Std. Error"]
70  emit("STAT-REG-NUMERIC-ORDINAL-001",
71       "tabelz.analysis.ordinal_logistic:compute_ordinal_logistic_regression",
72       "ordinalRegressionMenuItem", "axis_coverage",
73       list(outcome_column = "ordinal", predictors = list("x"),
74            level_order = list("low", "middle", "high"), confidence_level = 0.9),
75       list("/coefficients/0/odds_ratio_ci_lower" = tol(exp(coef(ord_fit)[["x"]] -
76           stats::qnorm(0.95) * ord_se), "90 percent ordered-logit Wald OR lower")),
77       list("MASS polr uses the manual's positive-at-higher-level slope sign."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-NUMERIC-POISSON-001

Case STAT-REG-NUMERIC-POISSON-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/rate_ratio_ci_lowertolerance1.10500495023020771.10500518440182852.34E-76passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 93  pois <- stats::glm(count ~ x + offset(log(exposure)), data = d,
 94                     family = stats::poisson())
 95  pois_se <- summary(pois)$coefficients["x", "Std. Error"]
 96  emit("STAT-REG-NUMERIC-POISSON-001",
 97       "tabelz.analysis.poisson_regression:compute_poisson_regression",
 98       "poissonRegressionMenuItem", "axis_coverage",
 99       list(count_column = "count", predictors = list("x"), exposure_column = "exposure",
100            confidence_level = 0.9),
101       list("/coefficients/1/rate_ratio_ci_lower" = tol(exp(coef(pois)[["x"]] -
102           stats::qnorm(0.95) * pois_se), "90 percent Poisson Wald rate-ratio lower")),
103       list("Poisson uses a log-exposure offset and a Wald rather than profile-likelihood interval."))

Whole script generators/r/regression_options.R


Regression setting STAT-REG-NUMERIC-VIF-DEGREE3-001

Case STAT-REG-NUMERIC-VIF-DEGREE3-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_POSITIVE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance478.08473800010495478.0847380000871.80E-1113passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

211  positive <- read_dataset("REG_POSITIVE")
212  poly3_design <- data.frame(x = positive$x, z = positive$z,
213                             x2 = positive$x^2, z2 = positive$z^2,
214                             x3 = positive$x^3, z3 = positive$z^3)
215  poly3_vif <- 1 / (1 - summary(stats::lm(x ~ z + x2 + z2 + x3 + z3,
216                                        data = poly3_design))$r.squared)
217  emit("STAT-REG-NUMERIC-VIF-DEGREE3-001", vif_entry, "fitYByXMenuItem", "axis_coverage",
218       list(y_columns = list("y"), x_columns = list("x", "z"), missing = "drop",
219            model_type = "polynomial", degree = 3L, interactions = FALSE,
220            split_percent = 100L),
221       list("/0" = tol(poly3_vif, "cubic-design x VIF")),
222       list("R auxiliary OLS conditions x on z, x^2, z^2, x^3 and z^3."),
223       dataset = "REG_POSITIVE")

Whole script generators/r/regression_options.R


Regression setting STAT-REG-NUMERIC-VIF-HOLDOUT-001

Case STAT-REG-NUMERIC-VIF-HOLDOUT-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

205  vif_entry <- "tabelz.analysis.fit_y_by_x:compute_fit_vif"
206  emit("STAT-REG-NUMERIC-VIF-HOLDOUT-001", vif_entry, "fitYByXMenuItem", "axis_coverage",
207       list(y_columns = list("y"), x_columns = list("x"), missing = "drop",
208            model_type = "linear", degree = 2L, interactions = FALSE, split_percent = 80L),
209       list("/0" = tol(1, "single-predictor VIF invariant", "1e-9")),
210       list("For one selected X, the auxiliary regression has no other predictors, so VIF=1 under any training-row split."))

Whole script generators/r/regression_options.R


STAT REG OLS 001

Case STAT-REG-OLS-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS and nnet
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_core.R · sha256 4b30ab4a6f513e54
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/parameters/1/estimatetolerance1.3987626254070821.39876262540708221.83E-1615passed
/responses/0/metrics/ssetolerance15.71564104374702415.7156410437470244.69E-1616passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

67  fit <- stats::lm(y ~ x + z, data = d)
68  fit_entry <- "tabelz.analysis.fit_y_by_x:compute_fits"
69  fit_menu <- "fitYByXMenuItem"
70  emit("STAT-REG-OLS-001", fit_entry, fit_menu, "nominal",
71       list(y_columns = list("y"), x_columns = list("x", "z")),
72       list("/responses/0/parameters/1/estimate" = tol(coef(fit)[["x"]], "OLS x slope", "1e-9"),
73            "/responses/0/metrics/sse" = tol(sum(resid(fit)^2), "OLS SSE", "1e-9")),
74       list("Classical OLS with intercept; R lm coefficients and residual sum of squares."))

Whole script generators/r/regression_core.R


STAT REG ORDINAL 001

Case STAT-REG-ORDINAL-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS and nnet
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_core.R · sha256 4b30ab4a6f513e54
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/coefficienttolerance0.0379425517986230130.037942564420134011.26E-86passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 95  ordinal_entry <- "tabelz.analysis.ordinal_logistic:compute_ordinal_logistic_regression"
 96  d$ordinal_ordered <- ordered(d$ordinal, levels = c("low", "middle", "high"))
 97  ordinal_fit <- MASS::polr(ordinal_ordered ~ x, data = d, method = "logistic",
 98                            Hess = TRUE, control = list(maxit = 1000, reltol = 1e-15))
 99  emit("STAT-REG-ORDINAL-001", ordinal_entry, "ordinalRegressionMenuItem", "nominal",
100       list(outcome_column = "ordinal", predictors = list("x"),
101            level_order = list("low", "middle", "high")),
102       list("/coefficients/0/coefficient" = tol(coef(ordinal_fit)[["x"]], "proportional-odds slope")),
103       list("MASS polr convention logit P(Y <= k) = zeta[k] - eta; positive x slope raises higher-level odds."))

Whole script generators/r/regression_core.R


Regression II STAT-REG-PANEL-001

Case STAT-REG-PANEL-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats
Citation
R nlme::Orthodont, nlme::lme, plm::plm, stats::nls/lm/step documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_core.R · sha256 b7543c65b2778e63
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/estimatetolerance0.66018518518518510.66018518518518512.81E-1817passed
/observationsexact1081080.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

81  panel <- plm::plm(distance ~ age, data = d, index = c("subject", "time"),
82                    model = "within", effect = "individual")
83  emit("STAT-REG-PANEL-001", "tabelz.executor.panel_fixed_effects",
84       "panelFixedEffectsMenuItem", "REG_ORTHODONT",
85       list(response_column = "distance", predictor_columns = list("age"),
86            entity_column = "subject", time_column = "time"),
87       list("/coefficients/0/estimate" = tol(coef(panel)[["age"]], "within-entity age slope", "1e-9"),
88            "/observations" = exact(nrow(d), "balanced complete panel rows")),
89       list("R plm within estimator uses entity demeaning; slope does not depend on covariance correction."))

Whole script generators/r/regression_advanced_core.R


Small-sample boundary STAT-REG-PANEL-MIN-001

Case STAT-REG-PANEL-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats, plm, rpart
Citation
R stats::nls/lm, plm::plm, rpart::rpart documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_a.R · sha256 e69dea823e2a8f59
Dataset
reference_data/datasets/REG_MIN_PANEL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/estimatetolerance2.52.49999999999999964.44E-1615passed
/observationsexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

21  panel <- data.frame(y=c(1,3,2,5),x=c(1,2,1,2),entity=c("A","A","B","B"),
22                      time=c(1L,2L,1L,2L))
23  export_dataset(panel,"REG_MIN_PANEL","Two entities each observed twice",
24                 "R plm within-estimator small-panel boundary",
25                 "Constructed validation data; no third-party license.",
26                 c(y="Float64",x="Float64",entity="String",time="Int64"))
27  p <- read_dataset("REG_MIN_PANEL")
28  fit <- plm::plm(y~x,data=p,index=c("entity","time"),model="within",effect="individual")
29  emit("STAT-REG-PANEL-MIN-001","tabelz.executor.panel_fixed_effects","panelFixedEffectsMenuItem",
30       "REG_MIN_PANEL",list(response_column="y",predictor_columns=list("x"),
31                            entity_column="entity",time_column="time"),
32       list("/coefficients/0/estimate"=tol(coef(fit)[["x"]],"within slope for two two-period entities"),
33            "/observations"=exact(nrow(p),"all four panel rows")),
34       c("Two retained entities have two time points each; plm removes entity means."))

Whole script generators/r/regression_advanced_minimum_a.R


Regression edge STAT-REG-PANEL-SINGLETON-CORRECTED-001

Case STAT-REG-PANEL-SINGLETON-CORRECTED-001 · status: passed

Reference

Evidence type
R
Source
R plm::plm and analytic singleton partition
Citation
Tabelz manual regression model design conditions; R stats/plm references
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_b.R · sha256 4a5d01b584989ad6
Dataset
reference_data/datasets/REG_EDGE_PANEL_SINGLETON.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/estimatetolerance2.52.49999999999999964.44E-1615passed
/observationsexact440.00E+2passed
/singleton_observation_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

87  p <- read_dataset("REG_EDGE_PANEL_SINGLETON")
88  fit <- plm::plm(y~x,data=p,index=c("entity","time"),model="within",effect="individual")
89  emit("STAT-REG-PANEL-SINGLETON-CORRECTED-001","STAT-REG-PANEL-001",
90       "REG_EDGE_PANEL_SINGLETON",
91       list(response_column="y",predictor_columns=list("x"),entity_column="entity",
92            time_column="time"),
93       outputs=list("/coefficients/0/estimate"=tol(coef(fit)[["x"]],"R within slope after singleton removal"),
94                    "/observations"=exact(4L,"two retained two-period entities"),
95                    "/singleton_observation_count"=exact(1L,"one singleton source row")),
96       note="The manual excludes singleton C from the within fit and reports it separately.",
97       source="R plm::plm and analytic singleton partition")

Whole script generators/r/regression_advanced_special_b.R


STAT REG PENALIZED LOGISTIC 001

Case STAT-REG-PENALIZED-LOGISTIC-001 · status: passed

Reference

Evidence type
Analytical
Source
Base R Newton minimisation of the contracted elastic-net objective on a table whose split and folds are determined by class counts alone
Citation
Zou & Hastie (2005) Regularization and variable selection via the elastic net, JRSS-B 67, 301-320; Hastie, Tibshirani & Friedman (2009) ESL section 18.4.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_penalized_logistic.R · sha256 aab20acf7da335c9
Dataset
reference_data/datasets/REG_ENET_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/analyzed_countexact24240.00E+2passed
/train_eventsexact10100.00E+2passed
/train_non_eventsexact10100.00E+2passed
/test_eventsexact220.00E+2passed
/test_non_eventsexact220.00E+2passed
/train_prevalencetolerance0.50.50.00E+1passed
/baseline_cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/selected_lambdatolerance0.00010.00014.79E-2116passed
/selected_rhotolerance11.00.00E+2passed
/final_objectivetolerance0.00102102903703094330.00102102903703094331.95E-2016passed
/final_convergedexactTrueTruepassed
/coefficients/0/termexactInterceptInterceptpassed
/coefficients/0/fitted_scaletolerance02.710089127613145e-132.71E-13passed
/coefficients/0/raw_scaletolerance-70.611842813478702-70.611842778570113.49E-89passed
/coefficients/1/termexactloadloadpassed
/coefficients/1/fitted_scaletolerance9.21024036697542859.2102403624222234.55E-99passed
/coefficients/1/raw_scaletolerance6.14016024465028566.1401602416148153.04E-99passed
/coefficients/1/statusexactestimatedestimatedpassed
/artifact/numeric/0/meantolerance11.511.50.00E+1passed
/artifact/numeric/0/sdtolerance1.51.50.00E+1passed
/evaluation/test_countexact440.00E+2passed
/evaluation/model/log_losstolerance0.000100005000333400420.0001000050007887214.55E-138passed
/evaluation/model/briertolerance1.0000000000010001e-081.0000000091063576e-089.11E-178passed
/evaluation/model/auctolerance11.00.00E+2passed
/evaluation/baseline/log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/evaluation/baseline/briertolerance0.250.250.00E+0passed
/evaluation/confusion/true_positiveexact220.00E+2passed
/evaluation/confusion/false_positiveexact000.00E+2passed
/evaluation/confusion/true_negativeexact220.00E+2passed
/evaluation/confusion/false_negativeexact000.00E+2passed
/candidates/0/lamtolerance0.00010.00014.79E-2116passed
/candidates/0/rhotolerance00.00.00E+2passed
/candidates/0/eligibleexactTrueTruepassed
/candidates/0/cv_log_losstolerance0.000723382631635116020.00072338263163604499.29E-1611passed
/candidates/1/lamtolerance0.00010.00014.79E-2116passed
/candidates/1/rhotolerance0.250.250.00E+0passed
/candidates/1/eligibleexactTrueTruepassed
/candidates/1/cv_log_losstolerance0.000583618157385986670.00058361815738745621.47E-1511passed
/candidates/2/lamtolerance0.00010.00014.79E-2116passed
/candidates/2/rhotolerance0.50.50.00E+1passed
/candidates/2/eligibleexactTrueTruepassed
/candidates/2/cv_log_losstolerance0.000436877393065301080.00043687739306809762.80E-1511passed
/candidates/3/lamtolerance0.00010.00014.79E-2116passed
/candidates/3/rhotolerance0.750.750.00E+0passed
/candidates/3/eligibleexactTrueTruepassed
/candidates/3/cv_log_losstolerance0.000279590224510831180.00027959022451964658.82E-1510passed
/candidates/4/lamtolerance0.00010.00014.79E-2116passed
/candidates/4/rhotolerance11.00.00E+2passed
/candidates/4/eligibleexactTrueTruepassed
/candidates/4/cv_log_losstolerance0.000100005000333400250.000100005000788720924.55E-138passed
/candidates/5/lamtolerance0.0010.0012.08E-2016passed
/candidates/5/rhotolerance00.00.00E+2passed
/candidates/5/eligibleexactTrueTruepassed
/candidates/5/cv_log_losstolerance0.00525898992994778580.00525898992994791.15E-1613passed
/candidates/6/lamtolerance0.0010.0012.08E-2016passed
/candidates/6/rhotolerance0.250.250.00E+0passed
/candidates/6/eligibleexactTrueTruepassed
/candidates/6/cv_log_losstolerance0.00433802103878637260.0043380210387866312.58E-1613passed
/candidates/7/lamtolerance0.0010.0012.08E-2016passed
/candidates/7/rhotolerance0.50.50.00E+1passed
/candidates/7/eligibleexactTrueTruepassed
/candidates/7/cv_log_losstolerance0.00335363389453366630.00335363389453525751.59E-1512passed
/candidates/8/lamtolerance0.0010.0012.08E-2016passed
/candidates/8/rhotolerance0.750.750.00E+0passed
/candidates/8/eligibleexactTrueTruepassed
/candidates/8/cv_log_losstolerance0.00227386730069098180.00227386730070713861.62E-1411passed
/candidates/9/lamtolerance0.0010.0012.08E-2016passed
/candidates/9/rhotolerance11.00.00E+2passed
/candidates/9/eligibleexactTrueTruepassed
/candidates/9/cv_log_losstolerance0.00100050033358352680.00100050033780191784.22E-128passed
/candidates/10/lamtolerance0.010.012.08E-1916passed
/candidates/10/rhotolerance00.00.00E+2passed
/candidates/10/eligibleexactTrueTruepassed
/candidates/10/cv_log_losstolerance0.034169950229657090.034169950241245461.16E-119passed
/candidates/11/lamtolerance0.010.012.08E-1916passed
/candidates/11/rhotolerance0.250.250.00E+0passed
/candidates/11/eligibleexactTrueTruepassed
/candidates/11/cv_log_losstolerance0.0292928492878032420.0292928493622665427.45E-118passed
/candidates/12/lamtolerance0.010.012.08E-1916passed
/candidates/12/rhotolerance0.50.50.00E+1passed
/candidates/12/eligibleexactTrueTruepassed
/candidates/12/cv_log_losstolerance0.0238939891211923520.0238939891211924156.26E-1714passed
/candidates/13/lamtolerance0.010.012.08E-1916passed
/candidates/13/rhotolerance0.750.750.00E+0passed
/candidates/13/eligibleexactTrueTruepassed
/candidates/13/cv_log_losstolerance0.0177169297103924730.0177169297103946282.15E-1512passed
/candidates/14/lamtolerance0.010.012.08E-1916passed
/candidates/14/rhotolerance11.00.00E+2passed
/candidates/14/eligibleexactTrueTruepassed
/candidates/14/cv_log_losstolerance0.0100503358535014160.0100503358716219891.81E-118passed
/candidates/15/lamtolerance0.100000000000000010.14.45E-1816passed
/candidates/15/rhotolerance00.00.00E+2passed
/candidates/15/eligibleexactTrueTruepassed
/candidates/15/cv_log_losstolerance0.178350193525343560.178350193526430331.09E-1211passed
/candidates/16/lamtolerance0.100000000000000010.14.45E-1816passed
/candidates/16/rhotolerance0.250.250.00E+0passed
/candidates/16/eligibleexactTrueTruepassed
/candidates/16/cv_log_losstolerance0.166312456974446160.16631245699969622.53E-119passed
/candidates/17/lamtolerance0.100000000000000010.14.45E-1816passed
/candidates/17/rhotolerance0.50.50.00E+1passed
/candidates/17/eligibleexactTrueTruepassed
/candidates/17/cv_log_losstolerance0.151451493392274340.151451493392274322.33E-1715passed
/candidates/18/lamtolerance0.100000000000000010.14.45E-1816passed
/candidates/18/rhotolerance0.750.750.00E+0passed
/candidates/18/eligibleexactTrueTruepassed
/candidates/18/cv_log_losstolerance0.132275165739942910.132275165739943364.47E-1614passed
/candidates/19/lamtolerance0.100000000000000010.14.45E-1816passed
/candidates/19/rhotolerance11.00.00E+2passed
/candidates/19/eligibleexactTrueTruepassed
/candidates/19/cv_log_losstolerance0.105360515657826310.105360515660912173.09E-1210passed
/candidates/20/lamtolerance11.00.00E+2passed
/candidates/20/rhotolerance00.00.00E+2passed
/candidates/20/eligibleexactTrueTruepassed
/candidates/20/cv_log_losstolerance0.512590743242441690.51259074324244173.18E-1817passed
/candidates/21/lamtolerance11.00.00E+2passed
/candidates/21/rhotolerance0.250.250.00E+0passed
/candidates/21/eligibleexactTrueTruepassed
/candidates/21/cv_log_losstolerance0.57579724695810930.57579724695810931.09E-1817passed
/candidates/22/lamtolerance11.00.00E+2passed
/candidates/22/rhotolerance0.50.50.00E+1passed
/candidates/22/eligibleexactTrueTruepassed
/candidates/22/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/23/lamtolerance11.00.00E+2passed
/candidates/23/rhotolerance0.750.750.00E+0passed
/candidates/23/eligibleexactTrueTruepassed
/candidates/23/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/24/lamtolerance11.00.00E+2passed
/candidates/24/rhotolerance11.00.00E+2passed
/candidates/24/eligibleexactTrueTruepassed
/candidates/24/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/25/lamtolerance1010.00.00E+2passed
/candidates/25/rhotolerance00.00.00E+2passed
/candidates/25/eligibleexactTrueTruepassed
/candidates/25/cv_log_losstolerance0.669054234110365580.66905423411036564.54E-1817passed
/candidates/26/lamtolerance1010.00.00E+2passed
/candidates/26/rhotolerance0.250.250.00E+0passed
/candidates/26/eligibleexactTrueTruepassed
/candidates/26/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/27/lamtolerance1010.00.00E+2passed
/candidates/27/rhotolerance0.50.50.00E+1passed
/candidates/27/eligibleexactTrueTruepassed
/candidates/27/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/28/lamtolerance1010.00.00E+2passed
/candidates/28/rhotolerance0.750.750.00E+0passed
/candidates/28/eligibleexactTrueTruepassed
/candidates/28/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/29/lamtolerance1010.00.00E+2passed
/candidates/29/rhotolerance11.00.00E+2passed
/candidates/29/eligibleexactTrueTruepassed
/candidates/29/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/30/lamtolerance100100.00.00E+2passed
/candidates/30/rhotolerance00.00.00E+2passed
/candidates/30/eligibleexactTrueTruepassed
/candidates/30/cv_log_losstolerance0.690656524391227270.69065652439122741.08E-1615passed
/candidates/31/lamtolerance100100.00.00E+2passed
/candidates/31/rhotolerance0.250.250.00E+0passed
/candidates/31/eligibleexactTrueTruepassed
/candidates/31/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/32/lamtolerance100100.00.00E+2passed
/candidates/32/rhotolerance0.50.50.00E+1passed
/candidates/32/eligibleexactTrueTruepassed
/candidates/32/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/33/lamtolerance100100.00.00E+2passed
/candidates/33/rhotolerance0.750.750.00E+0passed
/candidates/33/eligibleexactTrueTruepassed
/candidates/33/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed
/candidates/34/lamtolerance100100.00.00E+2passed
/candidates/34/rhotolerance11.00.00E+2passed
/candidates/34/eligibleexactTrueTruepassed
/candidates/34/cv_log_losstolerance0.693147180559945290.69314718055994533.77E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

180  id <- "STAT-REG-PENALIZED-LOGISTIC-001"

Whole script generators/r/regression_penalized_logistic.R


Regression edge STAT-REG-PLS-CONSTANT-001

Case STAT-REG-PLS-CONSTANT-001 · status: passed

Reference

Evidence type
Analytical
Source
Analytic design condition
Citation
Tabelz manual regression model design conditions; R stats/plm references
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_b.R · sha256 4a5d01b584989ad6
Dataset
reference_data/datasets/REG_EDGE_PLS_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

75  pls_data <- read_dataset("REG_MIN_PLS")
76  pls_data$x <- rep(7,nrow(pls_data))
77  export_dataset(pls_data,"REG_EDGE_PLS_CONSTANT","Zero-variance PLS predictor",
78                 "Analytic positive training-variance requirement",
79                 "Constructed validation data; no third-party license.",
80                 c(x="Float64",y="Float64"))
81  emit("STAT-REG-PLS-CONSTANT-001","STAT-REG-PLS-FIT-001",
82       "REG_EDGE_PLS_CONSTANT",
83       list(response="y",predictors=list("x"),scale=FALSE,cv_folds=2L,
84            test_fraction=0.25,seed=0L),
85       error=err("ValueError","positive training variance"),
86       note="A constant predictor has zero variance in every training split, so no PLS component exists.")

Whole script generators/r/regression_advanced_special_b.R


Regression II STAT-REG-PLS-FIT-001

Case STAT-REG-PLS-FIT-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact30300.00E+2passed
/training_metrics/row_countexact23230.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

164  pls_fit <- pls::plsr(y ~ x, data = pls_data, ncomp = 1L, validation = "none", scale = FALSE)
165  pls_slope <- as.double(stats::coef(pls_fit, ncomp = 1L))[1L]
166  pls_mean_x <- mean(pls_data$x)
167  pls_mean_y <- mean(pls_data$y)
168  pls_art <- list(response = "y", predictors = list("x"), imputation = "drop", scale = FALSE,
169                  component_count = 1L, imputation_values = list(NULL),
170                  x_means = list(pls_mean_x), coefficients = list(pls_slope),
171                  intercept = pls_mean_y)
172  emit("STAT-REG-PLS-FIT-001", "tabelz.analysis.partial_least_squares:compute_partial_least_squares",
173       "partialLeastSquaresMenuItem", "REG_PLS",
174       list(response = "y", predictors = list("x"), scale = FALSE, cv_folds = 3L),
175       list("/source_count" = exact(as.integer(nrow(pls_data)), "complete PLS population"),
176            "/training_metrics/row_count" = exact(23L, "three-quarter training share")),
177       list("R pls::plsr reference is full-population; NumPy random split prevents coefficient comparison."),
178       kind = "determinism", deferred = list(list(path = "/artifact/coefficients",
179         reason = "PLS training split and cross-validation folds are NumPy-specific.")),
180       repeat_count = 2L)

Whole script generators/r/regression_advanced_scoring.R


Small-sample boundary STAT-REG-PLS-MIN-001

Case STAT-REG-PLS-MIN-001 · status: passed

Reference

Evidence type
R
Source
R nlme/MASS/pls/stats
Citation
R nlme::lme, MASS::lda, pls::plsr and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_b.R · sha256 e3e09e1965800ae1
Dataset
reference_data/datasets/REG_MIN_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 85  pls_data <- read_dataset("REG_PLS")[seq_len(8L),]
 86  export_dataset(pls_data,"REG_MIN_PLS","Eight-row one-predictor PLS design",
 87                 "R pls plsr small training/CV boundary",
 88                 "Constructed validation data; no third-party license.",
 89                 c(x="Float64",y="Float64"))
 90  pls_data <- read_dataset("REG_MIN_PLS")
 91  fit <- pls::plsr(y~x,data=pls_data,ncomp=1L,validation="none",scale=FALSE)
 92  emit("STAT-REG-PLS-MIN-001",
 93       "tabelz.analysis.partial_least_squares:compute_partial_least_squares",
 94       "partialLeastSquaresMenuItem","REG_MIN_PLS",
 95       list(response="y",predictors=list("x"),scale=FALSE,cv_folds=2L,
 96            test_fraction=0.25,seed=0L),
 97       list("/source_count"=exact(nrow(pls_data),"eight usable PLS rows")),
 98       c("Two-fold CV has six training rows; R full-sample plsr confirms one-component estimability.",
 99         "NumPy holdout and fold assignments are not reproduced by R."),
100       deferred=list(list(path="/artifact/coefficients",
101                          reason="NumPy split and CV folds cannot be independently replayed in R.")))

Whole script generators/r/regression_advanced_minimum_b.R


Regression II STAT-REG-PLS-SCORE-001

Case STAT-REG-PLS-SCORE-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/predictedtolerance5.00836069876324875.0083606987632123.64E-1414passed
/29/predictedtolerance62.99257280242433462.992572802424395.69E-1415passed
/0/row_idexactSTAT-REG-PLS-SCORE-001:0STAT-REG-PLS-SCORE-001:0passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

181  pls_pred <- as.double(stats::predict(pls_fit, newdata = pls_data, ncomp = 1L))
182  pls_score_id <- "STAT-REG-PLS-SCORE-001"

Whole script generators/r/regression_advanced_scoring.R


Retained scoring boundary STAT-REG-PLS-SCORE-CENTER-001

Case STAT-REG-PLS-SCORE-CENTER-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls and analytic retained tree
Citation
R MASS::lda, pls::plsr and rpart::rpart documentation; Tabelz retained-artifact contract
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_boundaries.R · sha256 9a546ee72b8c92a9
Dataset
reference_data/datasets/REG_SCORE_BOUND_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/predictedtolerance34.00046675059380234.00046675059383.65E-1616passed

Convention

Reference code

Whole script generators/r/regression_advanced_score_boundaries.R


Corrected retained scoring boundary STAT-REG-PLS-SCORE-CENTER-CORRECTED-001

Case STAT-REG-PLS-SCORE-CENTER-CORRECTED-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls or strict scoring contract
Citation
R MASS::lda, pls::plsr and Tabelz manual retained-scoring semantics
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_corrections.R · sha256 8c11616dfc0c8448
Dataset
reference_data/datasets/REG_SCORE_CENTER_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/predictedtolerance34.00046675059380234.00046675059383.65E-1616passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

50  emit("STAT-REG-PLS-SCORE-CENTER-CORRECTED-001",
51       "STAT-REG-PLS-SCORE-001","REG_SCORE_CENTER_PLS",
52       list("/0/predicted"=tol(pls_cfg$artifact$intercept,
53                               "At retained x_mean, centered PLS prediction equals intercept")),
54       NULL,"A one-row, fully observed scoring table tests the retained centered intercept.")

Whole script generators/r/regression_advanced_score_corrections.R


Retained scoring boundary STAT-REG-PLS-SCORE-MISSING-001

Case STAT-REG-PLS-SCORE-MISSING-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls and analytic retained tree
Citation
R MASS::lda, pls::plsr and rpart::rpart documentation; Tabelz retained-artifact contract
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_boundaries.R · sha256 9a546ee72b8c92a9
Dataset
reference_data/datasets/REG_SCORE_BOUND_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/1/predictedabsentNonepassed

Convention

Reference code

Whole script generators/r/regression_advanced_score_boundaries.R


True minimum sample STAT-REG-PLS-TRUE-MIN-001

Case STAT-REG-PLS-TRUE-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats/rpart/pls and analytic sample conditions
Citation
R stats::lm, rpart::rpart, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_true_min.R · sha256 f10363ad1ec76730
Dataset
reference_data/datasets/REG_TRUE_MIN_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

37  pls_data <- read_dataset("REG_PLS")[1:5,]
38  export_dataset(pls_data,"REG_TRUE_MIN_PLS","Five-row one-predictor PLS CV design",
39                 "R pls plsr estimability with one component",
40                 "Constructed validation data; no third-party license.",
41                 c(x="Float64",y="Float64"))
42  pls_data <- read_dataset("REG_TRUE_MIN_PLS")
43  fit <- pls::plsr(y~x,data=pls_data,ncomp=1L,validation="none",scale=FALSE)
44  emit("STAT-REG-PLS-TRUE-MIN-001","STAT-REG-PLS-FIT-001",
45       "REG_TRUE_MIN_PLS",
46       list(response="y",predictors=list("x"),scale=FALSE,cv_folds=2L,
47            test_fraction=0.2,seed=0L),
48       list("/source_count"=exact(as.integer(nrow(pls_data)),"five analyzed rows")),
49       "Four training rows leave two rows per two-fold training partition, enough for one component.",
50       deferred=list(list(path="/artifact/coefficients",
51         reason="NumPy holdout and CV fold identities are not replayable in R.")))

Whole script generators/r/regression_advanced_true_min.R


STAT REG POISSON 001

Case STAT-REG-POISSON-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS and nnet
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_core.R · sha256 4b30ab4a6f513e54
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance0.1851936668903110.18519366687670731.36E-1110passed
/deviancetolerance87.6867161581957587.686716158195761.46E-1415passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

115  poisson_entry <- "tabelz.analysis.poisson_regression:compute_poisson_regression"
116  poisson_fit <- stats::glm(count ~ x + offset(log(exposure)),
117                            family = stats::poisson(), data = d)
118  emit("STAT-REG-POISSON-001", poisson_entry, "poissonRegressionMenuItem", "nominal",
119       list(count_column = "count", predictors = list("x"), exposure_column = "exposure"),
120       list("/coefficients/1/coefficient" = tol(coef(poisson_fit)[["x"]], "Poisson log-rate slope"),
121            "/deviance" = tol(stats::deviance(poisson_fit), "Poisson deviance")),
122       list("Exposure enters only as log offset; glm Poisson uses log link."))

Whole script generators/r/regression_core.R


Regression II population STAT-REG-POP-ADV-DA-001

Case STAT-REG-POP-ADV-DA-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_DISCRIMINANT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact24240.00E+2passed
/excluded_countexact110.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_predictor_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

163  da_config <- list(response = "class", predictors = list("x"), model = "lda",
164                    imputation = "drop", standardize = FALSE, prior_mode = "equal")
165  emit("STAT-REG-POP-ADV-DA-001", "STAT-REG-DISCRIMINANT-LDA-001",
166       "REG_DISCRIMINANT_MISSING", da_config,
167       list("/source_count" = exact(as.integer(nrow(da)), "source class rows"),
168            "/excluded_count" = exact(1L, "excluded row"),
169            "/missing_response_count" = exact(1L, "missing class"),
170            "/incomplete_predictor_count" = exact(1L, "missing x")),
171       list("Class and predictor missing in different rows; original row 3 excluded."),
172       excluded = list(2L),
173       deferred = list(list(path = "/artifact/class_means",
174         reason = "NumPy stratified train split cannot be reproduced with R RNG.")))

Whole script generators/r/regression_advanced_population.R


Regression II population STAT-REG-POP-ADV-MIXED-001

Case STAT-REG-POP-ADV-MIXED-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_ORTHODONT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.671090493814609720.67109049389557048.10E-119passed
/analyzed_countexact1051050.00E+2passed
/excluded_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

60  orth_keep <- orth[-9L, , drop = FALSE]
61  orth_fit <- orth_keep[stats::complete.cases(orth_keep[c("distance", "age", "subject")]),,
62                        drop = FALSE]
63  random_intercept <- nlme::lme(distance ~ age, random = ~1 | subject, data = orth_fit,
64                                method = "REML", control = nlme::lmeControl(opt = "optim"))
65  emit("STAT-REG-POP-ADV-MIXED-001", "STAT-REG-MIXED-REML-001",
66       "REG_ORTHODONT_MISSING", list(response_column = "distance", group_column = "subject",
67                                     fixed_numeric_columns = list("age"), estimation = "reml"),
68       list("/coefficients/1/estimate" = tol(nlme::fixef(random_intercept)[["age"]],
69                                                 "R lme REML after omissions"),
70            "/analyzed_count" = exact(as.integer(nrow(orth_fit)), "complete nonexcluded rows"),
71            "/excluded_count" = exact(1L, "explicitly excluded source row")),
72       list("Drop missing distance/age in different subjects, then source row 9; REML."),
73       excluded = list(8L))

Whole script generators/r/regression_advanced_population.R


Regression II population STAT-REG-POP-ADV-NONLINEAR-001

Case STAT-REG-POP-ADV-NONLINEAR-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_DECAY_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/estimates/0tolerance3.96052998635809013.9605299864175575.95E-1110passed
/artifact/estimates/1tolerance0.796330809890557490.79633080994918275.86E-1110passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

104  decay_keep <- decay[-3L, , drop = FALSE]
105  decay_fit <- decay_keep[stats::complete.cases(decay_keep[c("x", "y")]), , drop = FALSE]
106  nls_fit <- stats::nls(y ~ c + a * exp(-k * x), data = decay_fit,
107                        start = list(a = 4, k = 0.8, c = 1), algorithm = "port",
108                        lower = c(a = 0, k = 0.01, c = 0),
109                        upper = c(a = 10, k = 3, c = 5),
110                        control = stats::nls.control(maxiter = 1000, tol = 1e-12))
111  emit("STAT-REG-POP-ADV-NONLINEAR-001", "STAT-REG-NONLINEAR-DECAY-001",
112       "REG_DECAY_MISSING",
113       list(x_column = "x", response = "y", curve = "exponential_decay",
114            parameters = list(list(name = "a", start = 4, lower = 0, upper = 10),
115                              list(name = "k", start = 0.8, lower = 0.01, upper = 3),
116                              list(name = "c", start = 1, lower = 0, upper = 5))),
117       list("/artifact/estimates/0" = tol(coef(nls_fit)[["a"]], "R nls amplitude"),
118            "/artifact/estimates/1" = tol(coef(nls_fit)[["k"]], "R nls rate")),
119       list("R nls port on complete nonexcluded rows; missing rows 1/2, excluded row 3."),
120       excluded = list(2L))

Whole script generators/r/regression_advanced_population.R


Regression II population STAT-REG-POP-ADV-PANEL-001

Case STAT-REG-POP-ADV-PANEL-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_ORTHODONT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/estimatetolerance0.668650793650793610.66865079365079382.19E-1615passed
/excluded_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 92  panel <- plm::plm(distance ~ age, data = orth_fit,
 93                    index = c("subject", "time"), model = "within", effect = "individual")
 94  emit("STAT-REG-POP-ADV-PANEL-001", "STAT-REG-PANEL-001",
 95       "REG_ORTHODONT_MISSING",
 96       list(response_column = "distance", predictor_columns = list("age"),
 97            entity_column = "subject", time_column = "time"),
 98       list("/coefficients/0/estimate" = tol(coef(panel)[["age"]],
 99                                                 "R plm within slope", "1e-9"),
100            "/excluded_count" = exact(1L, "explicitly excluded source row")),
101       list("R plm within estimator on complete nonexcluded rows."),
102       excluded = list(8L))

Whole script generators/r/regression_advanced_population.R


Regression II population STAT-REG-POP-ADV-PLS-001

Case STAT-REG-POP-ADV-PLS-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_PLS_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact30300.00E+2passed
/excluded_countexact110.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_predictor_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

176  emit("STAT-REG-POP-ADV-PLS-001", "STAT-REG-PLS-FIT-001",
177       "REG_PLS_MISSING", list(response = "y", predictors = list("x"),
178                               imputation = "drop", scale = FALSE, cv_folds = 3L),
179       list("/source_count" = exact(as.integer(nrow(pls_data)), "source PLS rows"),
180            "/excluded_count" = exact(1L, "excluded row"),
181            "/missing_response_count" = exact(1L, "missing response"),
182            "/incomplete_predictor_count" = exact(1L, "missing predictor")),
183       list("R complete cases after original row 3 exclusion; split moments deferred."),
184       excluded = list(2L),
185       deferred = list(list(path = "/artifact/coefficients",
186         reason = "NumPy split and selected component count cannot be reproduced with R RNG.")))

Whole script generators/r/regression_advanced_population.R


Regression II population STAT-REG-POP-ADV-SELECTION-001

Case STAT-REG-POP-ADV-SELECTION-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_ORTHODONT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance0.686523873270862170.68652387327086183.30E-1615passed
/nexact1051050.00E+2passed
/droppedexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

200  selection_clean <- orth_fit
201  selection_ref <- stats::lm(distance ~ age, data = selection_clean)
202  emit("STAT-REG-POP-ADV-SELECTION-001", "STAT-REG-SELECTION-001",
203       "REG_ORTHODONT_MISSING", list(response = "distance",
204                                     candidate_predictors = list("age")),
205       list("/coefficients/1/coefficient" = tol(coef(selection_ref)[["age"]],
206                                                  "complete-case OLS slope", "1e-9"),
207            "/n" = exact(as.integer(nrow(selection_clean)), "complete nonexcluded rows"),
208            "/dropped" = exact(2L, "two missing rows after exclusion")),
209       list("R lm on complete rows after original row 9 is excluded; one predictor wins."),
210       excluded = list(8L))

Whole script generators/r/regression_advanced_population.R


Regression II population STAT-REG-POP-ADV-STRUCT-001

Case STAT-REG-POP-ADV-STRUCT-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_ORTHODONT_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.68113653790483530.681136494066684.38E-87passed
/analyzed_countexact1051050.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

77  random_slope <- lme4::lmer(distance ~ age + (age | subject), data = orth_fit, REML = TRUE,
78                             control = lme4::lmerControl(
79                               optCtrl = list(ftol_abs = 1e-15, xtol_abs = 1e-15),
80                               check.conv.singular = "ignore"))
81  emit("STAT-REG-POP-ADV-STRUCT-001", "STAT-REG-MIXED-STRUCT-SLOPE-001",
82       "REG_ORTHODONT_MISSING",
83       list(response_column = "distance", fixed_numeric_columns = list("age"),
84            model_mode = "random_slope", group_column = "subject",
85            random_slope_column = "age", estimation = "reml"),
86       list("/coefficients/1/estimate" = tol(lme4::fixef(random_slope)[["age"]],
87                                                 "R random-slope age effect (lme4, converged)"),
88            "/analyzed_count" = exact(as.integer(nrow(orth_fit)), "complete nonexcluded rows")),
89       list("Subject random slope plus intercept; omit original rows 1, 5, 9."),
90       excluded = list(8L))

Whole script generators/r/regression_advanced_population.R


Regression II population STAT-REG-POP-ADV-TREE-CLASS-001

Case STAT-REG-POP-ADV-TREE-CLASS-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_TREE_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/threshold/0tolerance00.00.00E+2passed
/source_countexact40400.00E+2passed
/excluded_countexact110.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_predictor_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

122  tree_class <- tree[-4L, , drop = FALSE]
123  tree_class_fit <- tree_class[stats::complete.cases(tree_class[c("x", "class")]),,
124                               drop = FALSE]
125  rpart_class <- rpart::rpart(class ~ x, data = tree_class_fit, method = "class",
126                             control = rpart::rpart.control(cp = 0, xval = 0,
127                                                           minsplit = 2, minbucket = 1,
128                                                           maxdepth = 1))
129  stopifnot(nrow(rpart_class$frame) == 3L)
130  emit("STAT-REG-POP-ADV-TREE-CLASS-001", "STAT-REG-TREE-CLASS-001",
131       "REG_TREE_MISSING", list(response = "class", numeric_predictors = list("x"),
132                                 positive_class = "B", cv_folds = 2L,
133                                 max_depth = 1L, min_samples_leaf = 1L),
134       list("/artifact/threshold/0" = tol(rpart_class$splits[1L, "index"],
135                                              "unique split despite omissions", "1e-12"),
136            "/source_count" = exact(as.integer(nrow(tree)), "source row count"),
137            "/excluded_count" = exact(1L, "excluded row"),
138            "/missing_response_count" = exact(1L, "missing class label"),
139            "/incomplete_predictor_count" = exact(1L, "missing numeric predictor")),
140       list("Omit source class row 1, x row 2 and excluded row 4; leaves remain pure."),
141       excluded = list(3L))

Whole script generators/r/regression_advanced_population.R


Regression II population STAT-REG-POP-ADV-TREE-REG-001

Case STAT-REG-POP-ADV-TREE-REG-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_TREE_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/threshold/0tolerance00.00.00E+2passed
/source_countexact40400.00E+2passed
/excluded_countexact110.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_numeric_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

143  tree_reg <- tree[-4L, , drop = FALSE]
144  tree_reg_fit <- tree_reg[stats::complete.cases(tree_reg[c("x", "y")]),, drop = FALSE]
145  rpart_reg <- rpart::rpart(y ~ x, data = tree_reg_fit, method = "anova",
146                           control = rpart::rpart.control(cp = 0, xval = 0,
147                                                         minsplit = 2, minbucket = 1,
148                                                         maxdepth = 1))
149  stopifnot(nrow(rpart_reg$frame) == 3L)
150  emit("STAT-REG-POP-ADV-TREE-REG-001", "STAT-REG-TREE-REG-001",
151       "REG_TREE_MISSING", list(response = "y", numeric_predictors = list("x"),
152                                 cv_folds = 2L, max_depth = 1L,
153                                 depth_grid = list(1L), min_leaf_grid = list(1L)),
154       list("/artifact/threshold/0" = tol(rpart_reg$splits[1L, "index"],
155                                              "unique regression split", "1e-12"),
156            "/source_count" = exact(as.integer(nrow(tree)), "source row count"),
157            "/excluded_count" = exact(1L, "excluded row"),
158            "/missing_response_count" = exact(1L, "missing numeric response"),
159            "/incomplete_numeric_count" = exact(1L, "missing numeric predictor")),
160       list("Omit source x row 2, y row 3 and excluded row 4; leaves remain pure."),
161       excluded = list(3L))

Whole script generators/r/regression_advanced_population.R


Regression II population STAT-REG-POP-ADV-VALIDATED-001

Case STAT-REG-POP-ADV-VALIDATED-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/rpart/MASS/pls
Citation
R nlme::lme, plm::plm, rpart::rpart, MASS::lda, pls::plsr, stats::nls/lm
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_population.R · sha256 c1c329b2471524f0
Dataset
reference_data/datasets/REG_PLS_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact30300.00E+2passed
/excluded_countexact110.00E+2passed
/missing_response_countexact110.00E+2passed
/incomplete_numeric_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

188  emit("STAT-REG-POP-ADV-VALIDATED-001", "STAT-REG-VALIDATED-COUNT-001",
189       "REG_PLS_MISSING", list(response = "y", numeric_predictors = list("x"),
190                               algorithm = "ols", imputation = "drop", cv_folds = 3L),
191       list("/source_count" = exact(as.integer(nrow(pls_data)), "source regression rows"),
192            "/excluded_count" = exact(1L, "excluded row"),
193            "/missing_response_count" = exact(1L, "missing response"),
194            "/incomplete_numeric_count" = exact(1L, "missing predictor")),
195       list("Complete-case OLS after excluding original row 3; split coefficients deferred."),
196       excluded = list(2L),
197       deferred = list(list(path = "/recipe/coefficients",
198         reason = "No explicit split assignment; NumPy chooses training identities.")))

Whole script generators/r/regression_advanced_population.R


Missing and excluded STAT-REG-POP-BINARY-001

Case STAT-REG-POP-BINARY-001 · status: passed

Reference

Evidence type
R
Source
R complete-case models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_population.R · sha256 d1c98b1023443b9f
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact68680.00E+2passed
/coefficients/1/coefficienttolerance0.0204397047502816420.020439704750281738.66E-1714passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

68  binary_data <- kept[stats::complete.cases(kept[c("binary", "x")]), , drop = FALSE]
69  binary_fit <- stats::glm(I(binary == "event") ~ x, data = binary_data,
70                           family = stats::binomial())
71  emit("STAT-REG-POP-BINARY-001",
72       "tabelz.analysis.logistic_regression:compute_binary_logistic_regression",
73       "logisticRegressionMenuItem", "missing_excluded",
74       list(outcome_column = "binary", event_value = "event", predictors = list("x")),
75       list("/n" = exact(nrow(binary_data), "complete non-excluded binary rows"),
76            "/coefficients/1/coefficient" = tol(coef(binary_fit)[["x"]], "binary slope")),
77       list("Event is explicit; no imputation of missing outcome or predictor."))

Whole script generators/r/regression_population.R


Missing and excluded STAT-REG-POP-EFFECTS-001

Case STAT-REG-POP-EFFECTS-001 · status: passed

Reference

Evidence type
R
Source
R complete-case models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_population.R · sha256 d1c98b1023443b9f
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observationsexact68680.00E+2passed
/retained_row_idssequence['STAT-REG-POP-EFFECTS-001:0', 'STAT-REG-POP-EFFECTS-001:1',['STAT-REG-POP-EFFECTS-001:0', 'STAT-REG-POP-EFFECTS-001:1',passed
/residual_variancetolerance0.239793702533092010.239793702533091965.31E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

140  effects_fit <- stats::lm(y ~ x + group, data = linear_data)
141  emit("STAT-REG-POP-EFFECTS-001", "tabelz.executor.linear_effects",
142       "linearEffectsMenuItem", "row_alignment",
143       list(response_column = "y", factor_columns = list("group"),
144            covariate_columns = list("x")),
145       list("/observations" = exact(nrow(linear_data), "complete non-excluded effects rows"),
146            "/retained_row_ids" = seq_ids("STAT-REG-POP-EFFECTS-001", linear_indexes),
147            "/residual_variance" = tol(summary(effects_fit)$sigma^2, "effects residual variance")),
148       list("Linear effects treatment contrasts; first alphabetical level reference."))

Whole script generators/r/regression_population.R


Missing and excluded STAT-REG-POP-GENERAL-001

Case STAT-REG-POP-GENERAL-001 · status: passed

Reference

Evidence type
R
Source
R complete-case models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_population.R · sha256 d1c98b1023443b9f
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/observationsexact68680.00E+2passed
/retained_row_idssequence['STAT-REG-POP-GENERAL-001:0', 'STAT-REG-POP-GENERAL-001:1',['STAT-REG-POP-GENERAL-001:0', 'STAT-REG-POP-GENERAL-001:1',passed
/residual_variancetolerance0.239793702533091960.239793702533091982.47E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

127  linear_data <- kept[stats::complete.cases(kept[c("y", "x", "group")]), , drop = FALSE]
128  linear_fit <- stats::lm(y ~ x + group, data = linear_data,
129                          contrasts = list(group = stats::contr.sum(3)))
130  linear_indexes <- as.integer(rownames(linear_data))
131  emit("STAT-REG-POP-GENERAL-001", "tabelz.executor.general_linear",
132       "generalLinearMenuItem", "row_alignment",
133       list(response_column = "y", factor_columns = list("group"),
134            covariate_columns = list("x"), test_type = "III"),
135       list("/observations" = exact(nrow(linear_data), "complete non-excluded GLM rows"),
136            "/retained_row_ids" = seq_ids("STAT-REG-POP-GENERAL-001", linear_indexes),
137            "/residual_variance" = tol(summary(linear_fit)$sigma^2, "GLM residual variance")),
138       list("General linear precision weights absent; sum contrasts per manual."))

Whole script generators/r/regression_population.R


Missing and excluded STAT-REG-POP-NB-001

Case STAT-REG-POP-NB-001 · status: passed

Reference

Evidence type
R
Source
R complete-case models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_population.R · sha256 d1c98b1023443b9f
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact68680.00E+2passed
/analyzed_row_idssequence['STAT-REG-POP-NB-001:0', 'STAT-REG-POP-NB-001:1', 'STAT-REG['STAT-REG-POP-NB-001:0', 'STAT-REG-POP-NB-001:1', 'STAT-REGpassed
/coefficients/1/coefficienttolerance0.160411116989466560.160411116989492952.64E-1412passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

115  nb_data <- kept[stats::complete.cases(kept[c("nb_count", "x", "exposure")]), , drop = FALSE]
116  nb_fit <- MASS::glm.nb(nb_count ~ x + offset(log(exposure)), data = nb_data,
117                         control = stats::glm.control(maxit = 1000, epsilon = 1e-12))
118  nb_indexes <- as.integer(rownames(nb_data))
119  emit("STAT-REG-POP-NB-001", "tabelz.analysis.negative_binomial:compute_negative_binomial",
120       "nbRegressionMenuItem", "row_alignment",
121       list(count = "nb_count", predictors = list("x"), exposure = "exposure"),
122       list("/n" = exact(nrow(nb_data), "complete non-excluded NB2 rows"),
123            "/analyzed_row_ids" = seq_ids("STAT-REG-POP-NB-001", nb_indexes),
124            "/coefficients/1/coefficient" = tol(coef(nb_fit)[["x"]], "NB2 slope")),
125       list("An explicit row-ID sequence checks population and prediction alignment."))

Whole script generators/r/regression_population.R


Missing and excluded STAT-REG-POP-NOMINAL-001

Case STAT-REG-POP-NOMINAL-001 · status: passed

Reference

Evidence type
R
Source
R complete-case models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_population.R · sha256 d1c98b1023443b9f
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact68680.00E+2passed
/coefficients/1/coefficienttolerance-0.012377674727291463-0.0123776746267614731.01E-108passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 92  nom_data <- kept[stats::complete.cases(kept[c("nominal", "x")]), , drop = FALSE]
 93  nom_data$response <- factor(nom_data$nominal, levels = c("A", "B", "C"))
 94  nom_fit <- nnet::multinom(response ~ x, data = nom_data, trace = FALSE,
 95                            maxit = 1000, reltol = 1e-15)
 96  emit("STAT-REG-POP-NOMINAL-001",
 97       "tabelz.analysis.nominal_logistic:compute_nominal_logistic_regression",
 98       "nominalRegressionMenuItem", "missing_excluded",
 99       list(outcome_column = "nominal", reference_value = "A", predictors = list("x")),
100       list("/n" = exact(nrow(nom_data), "complete non-excluded nominal rows"),
101            "/coefficients/1/coefficient" = tol(coef(nom_fit)["B", "x"], "B versus A x slope")),
102       list("R multinom reference A and tight convergence."))

Whole script generators/r/regression_population.R


Missing and excluded STAT-REG-POP-OLS-001

Case STAT-REG-POP-OLS-001 · status: passed

Reference

Evidence type
R
Source
R complete-case models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_population.R · sha256 d1c98b1023443b9f
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/observationsexact68680.00E+2passed
/responses/0/parameters/1/estimatetolerance1.41225914801089721.4122591480108987.13E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

50  fit_rows <- which(stats::complete.cases(kept[c("y", "x", "z")]))
51  fit_data <- kept[fit_rows, , drop = FALSE]
52  fit <- stats::lm(y ~ x + z, data = fit_data)
53  emit("STAT-REG-POP-OLS-001", "tabelz.analysis.fit_y_by_x:compute_fits",
54       "fitYByXMenuItem", "missing_excluded",
55       list(y_columns = list("y"), x_columns = list("x", "z")),
56       list("/responses/0/observations" = exact(nrow(fit_data), "complete, non-excluded rows"),
57            "/responses/0/parameters/1/estimate" = tol(coef(fit)[["x"]], "x slope")),
58       list("Listwise OLS after population exclusion."))

Whole script generators/r/regression_population.R


Missing and excluded STAT-REG-POP-ORDINAL-001

Case STAT-REG-POP-ORDINAL-001 · status: passed

Reference

Evidence type
R
Source
R complete-case models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_population.R · sha256 d1c98b1023443b9f
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact68680.00E+2passed
/coefficients/0/coefficienttolerance0.0724614361563067630.072461452112100031.60E-86passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

79  ord_data <- kept[stats::complete.cases(kept[c("ordinal", "x")]), , drop = FALSE]
80  ord_data$response <- ordered(ord_data$ordinal, levels = c("low", "middle", "high"))
81  ord_fit <- MASS::polr(response ~ x, data = ord_data, method = "logistic", Hess = TRUE,
82                        control = list(maxit = 1000, reltol = 1e-15))
83  emit("STAT-REG-POP-ORDINAL-001",
84       "tabelz.analysis.ordinal_logistic:compute_ordinal_logistic_regression",
85       "ordinalRegressionMenuItem", "missing_excluded",
86       list(outcome_column = "ordinal", predictors = list("x"),
87            level_order = list("low", "middle", "high")),
88       list("/n" = exact(nrow(ord_data), "complete non-excluded ordinal rows"),
89            "/coefficients/0/coefficient" = tol(coef(ord_fit)[["x"]], "ordered slope")),
90       list("Low-to-high outcome order is explicit in R and Tabelz."))

Whole script generators/r/regression_population.R


Missing and excluded STAT-REG-POP-POISSON-001

Case STAT-REG-POP-POISSON-001 · status: passed

Reference

Evidence type
R
Source
R complete-case models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_population.R · sha256 d1c98b1023443b9f
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact68680.00E+2passed
/coefficients/1/coefficienttolerance0.2006482275606610.200648227540143942.05E-119passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

104  pois_data <- kept[stats::complete.cases(kept[c("count", "x", "exposure")]), , drop = FALSE]
105  pois_fit <- stats::glm(count ~ x + offset(log(exposure)), data = pois_data,
106                         family = stats::poisson())
107  emit("STAT-REG-POP-POISSON-001",
108       "tabelz.analysis.poisson_regression:compute_poisson_regression",
109       "poissonRegressionMenuItem", "missing_excluded",
110       list(count_column = "count", predictors = list("x"), exposure_column = "exposure"),
111       list("/n" = exact(nrow(pois_data), "complete non-excluded Poisson rows"),
112            "/coefficients/1/coefficient" = tol(coef(pois_fit)[["x"]], "Poisson rate slope")),
113       list("Positive exposure enters as log offset."))

Whole script generators/r/regression_population.R


Missing and excluded STAT-REG-POP-VIF-001

Case STAT-REG-POP-VIF-001 · status: passed

Reference

Evidence type
R
Source
R complete-case models
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_population.R · sha256 d1c98b1023443b9f
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance1.00098587156680831.0009858715668082.28E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

60  vif <- 1 / (1 - summary(stats::lm(x ~ z, data = fit_data))$r.squared)
61  emit("STAT-REG-POP-VIF-001", "tabelz.analysis.fit_y_by_x:compute_fit_vif",
62       "fitYByXMenuItem", "missing_excluded",
63       list(y_columns = list("y"), x_columns = list("x", "z"), missing = "drop",
64            model_type = "linear", degree = 2L, interactions = FALSE, split_percent = 100L),
65       list("/0" = tol(vif, "x variance inflation", "1e-9")),
66       list("VIF uses the same complete design rows as the OLS fit."))

Whole script generators/r/regression_population.R


STAT REG QUANTILE REGRESSION 001

Case STAT-REG-QUANTILE-REGRESSION-001 · status: passed

Reference

Evidence type
R
Source
R quantreg::rq(method = 'br') and the check loss sum rho_tau(y - X beta); base R sort() for the intercept-only order statistic
Citation
Koenker, R. (2005) Quantile Regression, CUP; quantreg package help: rq, rq.fit.br.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, quantreg 6.1, jsonlite 2.0.0
Reference script
generators/r/regression_quantile.R · sha256 9b1ef07b4d0ba8ce
Dataset
reference_data/datasets/REG_QUANT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit_rowsexact30300.00E+2passed
/source_countexact30300.00E+2passed
/excluded_countexact000.00E+2passed
/incomplete_countexact000.00E+2passed
/crossing_rowsexact000.00E+2passed
/fits/0/quantiletolerance0.50.50.00E+1passed
/fits/0/check_losstolerance16.12622424501280316.1262242450128074.05E-1515passed
/fits/0/tail_informationtolerance1515.00.00E+2passed
/fits/0/interval_statusexactavailableavailablepassed
/fits/0/resamplesexact9999990.00E+2passed
/fits/0/coefficients/0/termexact(Intercept)(Intercept)passed
/fits/0/coefficients/0/estimatetolerance1.4535887941080751.4535887941080783.06E-1514passed
/fits/0/coefficients/1/termexactx1x1passed
/fits/0/coefficients/1/estimatetolerance0.575792131247382490.57579213124738195.57E-1615passed
/fits/0/coefficients/2/termexactx2x2passed
/fits/0/coefficients/2/estimatetolerance-1.0945152954491282-1.09451529544912822.17E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

32  emit <- function(id, kind, config, outputs, notes) {
33    write_expected("regression", id, list(
34      schema_version = "2.0.0", case_id = id, outputs = outputs,
35      expected_error = NULL, deferred_outputs = list(), generator_run = run))
36    write_case("regression", id, list(
37      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
38      entry_point = "tabelz.analysis.quantile_regression:compute_quantile_regression",
39      menu_items = list("quantileRegressionMenuItem"), title = gsub("-", " ", id),
40      case_kind = kind,
41      dataset = list(path = "reference_data/datasets/REG_QUANT_01.json", sha256 = ""),
42      excluded_rows = list(), configuration = config, "repeat" = 1L,
43      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"), sha256 = ""),
44      oracle = list(
45        oracle_class = "independent_implementation",
46        source = "R quantreg::rq(method = 'br') and the check loss sum rho_tau(y - X beta); base R sort() for the intercept-only order statistic",
47        citation = "Koenker, R. (2005) Quantile Regression, CUP; quantreg package help: rq, rq.fit.br.",
48        generator = list(path = "generators/r/regression_quantile.R", sha256 = ""),
49        independent = TRUE,
50        convention_notes = c(
51          notes,
52          "Check loss rho_tau(u) = u (tau - 1(u < 0)) with an intercept, no weights, complete cases (docs/manual/chapters/76-forest-penalty-quantiles.tex:121-146); categorical predictors are reference-cell indicators with the first sorted level as reference, as in rq's default treatment contrasts.",
53          "Intercept-only fits report the declared lower empirical quantile y_(ceil(n tau)) and, when n tau is an integer, the minimizing interval [y_(n tau), y_(n tau + 1)] as the sensitivity range (manual lines 133-146).",
54          "Uniqueness of each compared vertex was checked in R by agreement of the simplex (br) and interior-point (fn) solutions to 1e-5.",
55          "Bootstrap percentile intervals (999 NumPy-seeded resamples), the held-out split and the near-optimal sensitivity probes are not reproduced in R; only the interval status rule min(n tau, n(1 - tau)) >= 10 is checked.",
56          "Expected computed from the exported CSV."
57        )),
58      supported = TRUE, unsupported_reason = NULL,
59      requirement_ids = list("REQ-STAT-REG-QUANTILE-REGRESSION-001"),
60      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/regression_quantile.R


STAT REG QUANTILE REGRESSION 002

Case STAT-REG-QUANTILE-REGRESSION-002 · status: passed

Reference

Evidence type
R
Source
R quantreg::rq(method = 'br') and the check loss sum rho_tau(y - X beta); base R sort() for the intercept-only order statistic
Citation
Koenker, R. (2005) Quantile Regression, CUP; quantreg package help: rq, rq.fit.br.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, quantreg 6.1, jsonlite 2.0.0
Reference script
generators/r/regression_quantile.R · sha256 9b1ef07b4d0ba8ce
Dataset
reference_data/datasets/REG_QUANT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit_rowsexact30300.00E+2passed
/source_countexact30300.00E+2passed
/excluded_countexact000.00E+2passed
/incomplete_countexact000.00E+2passed
/fits/0/quantiletolerance0.250.250.00E+0passed
/fits/0/check_losstolerance11.99945875904058911.9994587590405989.11E-1515passed
/fits/0/tail_informationtolerance7.57.50.00E+1passed
/fits/0/interval_statusexacttail_withheldtail_withheldpassed
/fits/0/resamplesexact000.00E+2passed
/fits/0/coefficients/0/termexact(Intercept)(Intercept)passed
/fits/0/coefficients/0/estimatetolerance1.51773048420228921.5177304842022966.86E-1514passed
/fits/0/coefficients/1/termexactx1x1passed
/fits/0/coefficients/1/estimatetolerance0.360619289738465240.36061928973846552.80E-1615passed
/fits/0/coefficients/2/termexactx2x2passed
/fits/0/coefficients/2/estimatetolerance-1.0963790036147474-1.09637900361475032.87E-1514passed
/fits/1/quantiletolerance0.750.750.00E+0passed
/fits/1/check_losstolerance12.52446027906566312.5244602790656684.83E-1515passed
/fits/1/tail_informationtolerance7.57.50.00E+1passed
/fits/1/interval_statusexacttail_withheldtail_withheldpassed
/fits/1/resamplesexact000.00E+2passed
/fits/1/coefficients/0/termexact(Intercept)(Intercept)passed
/fits/1/coefficients/0/estimatetolerance3.36561728262240713.36561728262240764.58E-1615passed
/fits/1/coefficients/1/termexactx1x1passed
/fits/1/coefficients/1/estimatetolerance0.361324174465001420.361324174465000766.62E-1614passed
/fits/1/coefficients/2/termexactx2x2passed
/fits/1/coefficients/2/estimatetolerance-1.0671678966335363-1.06716789663353674.28E-1615passed
/quantiles/0tolerance0.250.250.00E+0passed
/quantiles/1tolerance0.750.750.00E+0passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

111  out <- c(fit_outputs(0L, 0.25, y ~ x1 + x2, c("(Intercept)", "x1", "x2"), FALSE),
112           fit_outputs(1L, 0.75, y ~ x1 + x2, c("(Intercept)", "x1", "x2"), FALSE))
113  out[["/quantiles/0"]] <- tol(0.25, "sorted quantiles", "1e-12")
114  out[["/quantiles/1"]] <- tol(0.75, "sorted quantiles", "1e-12")
115  emit("STAT-REG-QUANTILE-REGRESSION-002", "axis_coverage",
116       list(response = "y", numeric_predictors = list("x1", "x2"), quantiles = list(0.75, 0.25),
117            confidence_level = 0.90, dependence = "independent", held_out_evaluation = FALSE,
118            seed = 3L),
119       c(common[c("/fit_rows", "/source_count", "/excluded_count", "/incomplete_count")], out),
120       list("Quantiles 0.75 and 0.25 are fitted separately (rq per tau) and reported in increasing order; both tails have min(n tau, n(1 - tau)) = 7.5 < 10 so intervals are withheld."))

Whole script generators/r/regression_quantile.R


STAT REG QUANTILE REGRESSION 003

Case STAT-REG-QUANTILE-REGRESSION-003 · status: passed

Reference

Evidence type
R
Source
R quantreg::rq(method = 'br') and the check loss sum rho_tau(y - X beta); base R sort() for the intercept-only order statistic
Citation
Koenker, R. (2005) Quantile Regression, CUP; quantreg package help: rq, rq.fit.br.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, quantreg 6.1, jsonlite 2.0.0
Reference script
generators/r/regression_quantile.R · sha256 9b1ef07b4d0ba8ce
Dataset
reference_data/datasets/REG_QUANT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit_rowsexact30300.00E+2passed
/source_countexact30300.00E+2passed
/excluded_countexact000.00E+2passed
/incomplete_countexact000.00E+2passed
/crossing_rowsexact000.00E+2passed
/fits/0/quantiletolerance0.50.50.00E+1passed
/fits/0/check_losstolerance24.14033521961735624.140335219617357.37E-1515passed
/fits/0/tail_informationtolerance1515.00.00E+2passed
/fits/0/interval_statusexactavailableavailablepassed
/fits/0/resamplesexact9999990.00E+2passed
/fits/0/coefficients/0/termexact(Intercept)(Intercept)passed
/fits/0/coefficients/0/estimatetolerance-1.3289517650229055-1.32895176502290553.76E-1716passed
/fits/0/coefficients/1/termexactx1x1passed
/fits/0/coefficients/1/estimatetolerance0.597143627054702360.59714362705470231.07E-1615passed
/fits/0/coefficients/2/termexactg=B (vs A)g=B (vs A)passed
/fits/0/coefficients/2/estimatetolerance1.1237240635947191.12372406359471922.14E-1615passed
/fits/0/coefficients/3/termexactg=C (vs A)g=C (vs A)passed
/fits/0/coefficients/3/estimatetolerance-1.2135289679331718-1.2135289679331716.90E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

123  emit("STAT-REG-QUANTILE-REGRESSION-003", "axis_coverage",
124       list(response = "y", numeric_predictors = list("x1"), categorical_predictors = list("g"),
125            quantiles = list(0.5), confidence_level = 0.95, dependence = "independent",
126            held_out_evaluation = FALSE, seed = 0L),
127       c(common, fit_outputs(0L, 0.5, y ~ x1 + g, c("(Intercept)", "x1", "g=B (vs A)", "g=C (vs A)"), TRUE)),
128       list("rq(y ~ x1 + g) with R's default treatment contrasts (first sorted level A as reference)."))

Whole script generators/r/regression_quantile.R


STAT REG QUANTILE REGRESSION 004

Case STAT-REG-QUANTILE-REGRESSION-004 · status: passed

Reference

Evidence type
R
Source
R quantreg::rq(method = 'br') and the check loss sum rho_tau(y - X beta); base R sort() for the intercept-only order statistic
Citation
Koenker, R. (2005) Quantile Regression, CUP; quantreg package help: rq, rq.fit.br.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, quantreg 6.1, jsonlite 2.0.0
Reference script
generators/r/regression_quantile.R · sha256 9b1ef07b4d0ba8ce
Dataset
reference_data/datasets/REG_QUANT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit_rowsexact30300.00E+2passed
/source_countexact30300.00E+2passed
/excluded_countexact000.00E+2passed
/incomplete_countexact000.00E+2passed
/crossing_rowsexact000.00E+2passed
/fits/0/coefficients/0/termexact(Intercept)(Intercept)passed
/fits/0/coefficients/0/estimatetolerance1.33600000000000011.3362.36E-1716passed
/fits/0/coefficients/0/range_lowtolerance1.33600000000000011.3362.36E-1716passed
/fits/0/coefficients/0/range_hightolerance1.70700000000000011.7072.72E-1716passed
/fits/0/coefficients/0/ambiguousexactTrueTruepassed
/fits/0/check_losstolerance30.54899999999999930.5494.88E-1616passed
/fits/0/interval_statusexactavailableavailablepassed
/fits/0/tail_informationtolerance1515.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

131  sorted <- sort(d$y)
132  estimate <- sorted[[15]]
133  stopifnot(sorted[[16]] > estimate)
134  emit("STAT-REG-QUANTILE-REGRESSION-004", "axis_coverage",
135       list(response = "y", quantiles = list(0.5), confidence_level = 0.95,
136            dependence = "independent", held_out_evaluation = FALSE, seed = 0L),
137       c(common,
138         list("/fits/0/coefficients/0/term" = exact("(Intercept)", "term label"),
139              "/fits/0/coefficients/0/estimate" = tol(estimate, "order statistic y_(15) = y_(ceil(30 * 0.5))", "1e-12"),
140              "/fits/0/coefficients/0/range_low" = tol(estimate, "lower end of the minimizing interval", "1e-12"),
141              "/fits/0/coefficients/0/range_high" = tol(sorted[[16]], "upper end of the minimizing interval y_(16)", "1e-12"),
142              "/fits/0/coefficients/0/ambiguous" = exact(TRUE, "an interval of minimizers is flagged ambiguous"),
143              "/fits/0/check_loss" = tol(check_loss(d$y, rep(estimate, n), 0.5), "check loss of the median", "1e-12"),
144              "/fits/0/interval_status" = exact("available", "15 >= 10 tail rows"),
145              "/fits/0/tail_information" = tol(15, "min(n tau, n (1 - tau))", "1e-12"))),
146       list("Intercept-only median: n tau = 15 is an integer and y_(15) < y_(16), so the declared estimate is y_(15) and every value up to y_(16) minimizes the loss (analytic, base R sort)."))

Whole script generators/r/regression_quantile.R


QUEST root split between two point-mass classes, no pruning

Case STAT-REG-QUEST-001 · status: passed

Reference

Evidence type
Analytical
Source
Loh and Shih (1997) split rule evaluated by hand for two point-mass classes
Citation
Loh and Shih (1997) Split selection methods for classification trees, Statistica Sinica 7, 815-840, Algorithm 1 (equal-variance case).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_quest.R · sha256 8069251c8d3970f1
Dataset
reference_data/datasets/REG_QUEST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact21210.00E+2passed
/test_rowsexact990.00E+2passed
/missing_targetexact000.00E+2passed
/excluded_countexact000.00E+2passed
/classessequence['high', 'low']['high', 'low']passed
/grown_leavesexact220.00E+2passed
/training_errortolerance00.00.00E+2passed
/test_errortolerance00.00.00E+2passed
/training_rows_routedexact000.00E+2passed
/test_rows_routedexact000.00E+2passed
/nodes/0/nexact21210.00E+2passed
/nodes/0/split/predictorexactvibrationvibrationpassed
/nodes/0/split/kindexactnumericnumericpassed
/nodes/0/split/stageexactbonferronibonferronipassed
/nodes/0/split/thresholdtolerance22.00.00E+2passed
/nodes/0/split/methodexactlinearlinearpassed
/nodes/0/split/superclasses/0/0exactlowlowpassed
/nodes/0/split/superclasses/1/0exacthighhighpassed
/nodes/0/split/missing_rowsexact000.00E+2passed
/nodes/0/tests/0/predictorexactvibrationvibrationpassed
/nodes/0/tests/0/testexactFFpassed
/nodes/0/tests/0/p_valuetolerance00.00.00E+2passed
/nodes/0/tests/0/rowsexact21210.00E+2passed
/nodes/1/stop_reasonexactpure nodepure nodepassed
/nodes/2/stop_reasonexactpure nodepure nodepassed
/nodes/1/predictionexactlowlowpassed
/nodes/2/predictionexacthighhighpassed
/selected_alphaabsentNonepassed
/cv_folds_usedabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

27  emit <- function(id, title, kind, configuration, outputs, notes) {
28    expected <- list(schema_version = "2.0.0", case_id = id, outputs = outputs,
29                     expected_error = NULL, deferred_outputs = list(), generator_run = run)
30    write_expected("regression", id, expected)
31    case <- list(
32      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
33      entry_point = "tabelz.analysis.quest:compute_quest",
34      menu_items = list("questMenuItem"), title = title, case_kind = kind,
35      dataset = list(path = "reference_data/datasets/REG_QUEST.json", sha256 = ""),
36      excluded_rows = list(), configuration = configuration, "repeat" = 1L,
37      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"),
38                      sha256 = ""),
39      oracle = list(oracle_class = "analytic",
40                    source = "Loh and Shih (1997) split rule evaluated by hand for two point-mass classes",
41                    citation = "Loh and Shih (1997) Split selection methods for classification trees, Statistica Sinica 7, 815-840, Algorithm 1 (equal-variance case).",
42                    generator = list(path = "generators/r/regression_quest.R", sha256 = ""),
43                    independent = TRUE,
44                    convention_notes = c(notes, "Expected computed from the exported CSV.", manual)),
45      supported = TRUE, unsupported_reason = NULL,
46      requirement_ids = list("REQ-STAT-REG-QUEST-001"), risk_ids = list("RISK-REFERENCE-001"),
47      review_reference = NULL
48    )
49    write_case("regression", id, case)

Whole script generators/r/regression_quest.R


QUEST root split with equal priors

Case STAT-REG-QUEST-EQUAL-001 · status: passed

Reference

Evidence type
Analytical
Source
Loh and Shih (1997) split rule evaluated by hand for two point-mass classes
Citation
Loh and Shih (1997) Split selection methods for classification trees, Statistica Sinica 7, 815-840, Algorithm 1 (equal-variance case).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_quest.R · sha256 8069251c8d3970f1
Dataset
reference_data/datasets/REG_QUEST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact21210.00E+2passed
/test_rowsexact990.00E+2passed
/missing_targetexact000.00E+2passed
/excluded_countexact000.00E+2passed
/classessequence['high', 'low']['high', 'low']passed
/grown_leavesexact220.00E+2passed
/training_errortolerance00.00.00E+2passed
/test_errortolerance00.00.00E+2passed
/training_rows_routedexact000.00E+2passed
/test_rows_routedexact000.00E+2passed
/nodes/0/nexact21210.00E+2passed
/nodes/0/split/predictorexactvibrationvibrationpassed
/nodes/0/split/kindexactnumericnumericpassed
/nodes/0/split/stageexactbonferronibonferronipassed
/nodes/0/split/thresholdtolerance22.00.00E+2passed
/nodes/0/split/methodexactlinearlinearpassed
/nodes/0/split/superclasses/0/0exactlowlowpassed
/nodes/0/split/superclasses/1/0exacthighhighpassed
/nodes/0/split/missing_rowsexact000.00E+2passed
/nodes/0/tests/0/predictorexactvibrationvibrationpassed
/nodes/0/tests/0/testexactFFpassed
/nodes/0/tests/0/p_valuetolerance00.00.00E+2passed
/nodes/0/tests/0/rowsexact21210.00E+2passed
/nodes/1/stop_reasonexactpure nodepure nodepassed
/nodes/2/stop_reasonexactpure nodepure nodepassed
/nodes/1/predictionexactlowlowpassed
/nodes/2/predictionexacthighhighpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

114  emit("STAT-REG-QUEST-EQUAL-001", "QUEST root split with equal priors", "axis_coverage",
115       list(target = "class", predictors = list("vibration"), min_parent = 2L, min_child = 1L,
116            max_depth = 5L, prune = FALSE, priors = "equal"),
117       root_outputs(),
118       list(invariance, "Equal priors change the prior ratio in the linear rule, whose variance factor is 0 here, so the threshold stays 2."))

Whole script generators/r/regression_quest.R


QUEST root split kept by cost-complexity pruning

Case STAT-REG-QUEST-PRUNED-001 · status: passed

Reference

Evidence type
Analytical
Source
Loh and Shih (1997) split rule evaluated by hand for two point-mass classes
Citation
Loh and Shih (1997) Split selection methods for classification trees, Statistica Sinica 7, 815-840, Algorithm 1 (equal-variance case).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_quest.R · sha256 8069251c8d3970f1
Dataset
reference_data/datasets/REG_QUEST.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/training_rowsexact21210.00E+2passed
/test_rowsexact990.00E+2passed
/missing_targetexact000.00E+2passed
/excluded_countexact000.00E+2passed
/classessequence['high', 'low']['high', 'low']passed
/grown_leavesexact220.00E+2passed
/training_errortolerance00.00.00E+2passed
/test_errortolerance00.00.00E+2passed
/training_rows_routedexact000.00E+2passed
/test_rows_routedexact000.00E+2passed
/nodes/0/nexact21210.00E+2passed
/nodes/0/split/predictorexactvibrationvibrationpassed
/nodes/0/split/kindexactnumericnumericpassed
/nodes/0/split/stageexactbonferronibonferronipassed
/nodes/0/split/thresholdtolerance22.00.00E+2passed
/nodes/0/split/methodexactlinearlinearpassed
/nodes/0/split/superclasses/0/0exactlowlowpassed
/nodes/0/split/superclasses/1/0exacthighhighpassed
/nodes/0/split/missing_rowsexact000.00E+2passed
/nodes/0/tests/0/predictorexactvibrationvibrationpassed
/nodes/0/tests/0/testexactFFpassed
/nodes/0/tests/0/p_valuetolerance00.00.00E+2passed
/nodes/0/tests/0/rowsexact21210.00E+2passed
/nodes/1/stop_reasonexactpure nodepure nodepassed
/nodes/2/stop_reasonexactpure nodepure nodepassed
/nodes/1/predictionexactlowlowpassed
/nodes/2/predictionexacthighhighpassed
/cv_folds_usedexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

120  emit("STAT-REG-QUEST-PRUNED-001", "QUEST root split kept by cost-complexity pruning", "axis_coverage",
121       list(target = "class", predictors = list("vibration"), min_parent = 2L, min_child = 1L,
122            max_depth = 5L, prune = TRUE, cv_folds = 5L),
123       c(root_outputs(), list("/cv_folds_used" = exact(5L, "five folds requested on 21 training rows"))),
124       list(invariance, "Cross-validated pruning with the 1-SE rule keeps the split: the two-leaf tree has cross-validated error 0 (every fold refits the same midpoint rule) and the root alone misclassifies the minority class, so the smallest tree within one SE of the minimum is the two-leaf tree."))

Whole script generators/r/regression_quest.R


Blockwise regression enter all with casewise table

Case STAT-REG-REGRESSION-BLOCKS-001 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm and step
Citation
R Core Team, stats package help: lm, anova.lm, confint.lm, step, AIC, BIC, rstandard, rstudent, hatvalues, cooks.distance, dfbetas; Belsley, Kuh and Welsch (1980), Regression Diagnostics, ch. 3.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_blocks.R · sha256 86184660dac4e79f
Dataset
reference_data/datasets/REG_BLOCKS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/models/0/r_squaredtolerance0.898605402867828310.89860540286782833.23E-1817passed
/models/0/adjusted_r_squaredtolerance0.888465943154611140.88846594315461112.56E-1817passed
/models/0/standard_errortolerance0.792783919632095490.79278391963209532.21E-1615passed
/models/0/df2exact20200.00E+2passed
/models/0/residual_sstolerance12.57012686454457512.5701268645445687.38E-1515passed
/models/0/regression_sstolerance111.40222689074416111.402226890744171.42E-1415passed
/models/0/f_valuetolerance88.62458437469426588.624584374694314.31E-1415passed
/models/0/p_valuetail_probability1.148545328910703e-101.1485453289106956e-107.37E-2514passed
/models/0/durbin_watsontolerance2.71686806562813032.71686806562813032.43E-1717passed
/models/0/r_squared_changetolerance0.898605402867828310.89860540286782833.23E-1817passed
/models/0/df1exact220.00E+2passed
/models/0/f_changetolerance88.62458437469426588.624584374694274.21E-1617passed
/models/0/p_changetail_probability1.148545328910703e-101.1485453289107003e-102.72E-2514passed
/models/0/predictorsset['x1', 'x2']['x1', 'x2']passed
/models/1/r_squaredtolerance0.972408034755074310.97240803475507433.08E-1817passed
/models/1/adjusted_r_squaredtolerance0.96627648692286860.96627648692286864.88E-1817passed
/models/1/standard_errortolerance0.435930733294201070.435930733294201231.64E-1615passed
/models/1/df2exact18180.00E+2passed
/models/1/residual_sstolerance3.42064087614755773.42064087614756042.71E-1515passed
/models/1/regression_sstolerance120.55171287914118120.551712879141182.16E-1516passed
/models/1/f_valuetolerance158.59095637280046158.590956372800351.11E-1315passed
/models/1/p_valuetail_probability9.0390504021975662e-149.039050402197648e-148.21E-2814passed
/models/1/durbin_watsontolerance2.85395344278626872.85395344278626834.20E-1615passed
/models/1/r_squared_changetolerance0.0738026318872460060.0738026318872463.15E-1917passed
/models/1/df1exact220.00E+2passed
/models/1/f_changetolerance24.07308363464139224.073083634641372.10E-1415passed
/models/1/p_changetail_probability8.1829591465124059e-068.182959146512443e-063.73E-2014passed
/models/1/predictorsset['x1', 'x2', 'x3', 'x4']['x1', 'x2', 'x3', 'x4']passed
/models/1/coefficients/0/estimatetolerance1.98277534060737051.98277534060737051.55E-1717passed
/models/1/coefficients/0/standard_errortolerance0.136794457021216060.13679445702121613.26E-1715passed
/models/1/coefficients/0/t_valuetolerance14.49455909094220714.4945590909422061.49E-1515passed
/models/1/coefficients/0/p_valuetail_probability2.2842318698562376e-112.284231869856236e-111.63E-2615passed
/models/1/coefficients/0/lowertolerance1.69538085086869231.6953808508686922.42E-1615passed
/models/1/coefficients/0/uppertolerance2.27016983034604892.2701698303460491.15E-1717passed
/models/1/coefficients/1/estimatetolerance1.53997514876274181.53997514876274071.14E-1515passed
/models/1/coefficients/1/standard_errortolerance0.0664075805126408960.06640758051264092.13E-1917passed
/models/1/coefficients/1/t_valuetolerance23.18974937612133823.1897493761213231.46E-1415passed
/models/1/coefficients/1/p_valuetail_probability7.3813486664787184e-157.381348666478785e-156.62E-2914passed
/models/1/coefficients/1/lowertolerance1.40045799922465331.40045799922465221.11E-1515passed
/models/1/coefficients/1/uppertolerance1.67949229830083021.67949229830082911.07E-1515passed
/models/1/coefficients/1/standardizedtolerance0.917923972011284060.91792397201128355.53E-1615passed
/models/1/coefficients/1/zero_ordertolerance0.91986923051148550.91986923051148552.79E-1817passed
/models/1/coefficients/1/partialtolerance0.983672801122149280.98367280112214952.18E-1615passed
/models/1/coefficients/1/parttolerance0.907927393550236660.90792739355023642.21E-1615passed
/models/1/coefficients/1/tolerancetolerance0.978337759356080030.978337759356082.17E-1817passed
/models/1/coefficients/1/viftolerance1.02214188345155721.02214188345155729.38E-1817passed
/models/1/coefficients/2/estimatetolerance-0.72143265472075979-0.72143265472075934.42E-1615passed
/models/1/coefficients/2/standard_errortolerance0.116676159166633070.116676159166633081.38E-1715passed
/models/1/coefficients/2/t_valuetolerance-6.183205376947944-6.183205376947944.41E-1515passed
/models/1/coefficients/2/p_valuetail_probability7.7610299092293112e-067.761029909229387e-067.62E-2014passed
/models/1/coefficients/2/lowertolerance-0.96656016908461062-0.96656016908461024.43E-1615passed
/models/1/coefficients/2/uppertolerance-0.47630514035690896-0.476305140356908464.97E-1614passed
/models/1/coefficients/2/standardizedtolerance-0.24376812595595629-0.243768125955956151.40E-1615passed
/models/1/coefficients/2/zero_ordertolerance-0.31369485297341543-0.31369485297341545.77E-1715passed
/models/1/coefficients/2/partialtolerance-0.82455911297364481-0.82455911297364471.07E-1615passed
/models/1/coefficients/2/parttolerance-0.24208547710562292-0.242085477105622781.38E-1615passed
/models/1/coefficients/2/tolerancetolerance0.986242323797951740.98624232379795152.26E-1615passed
/models/1/coefficients/2/viftolerance1.01394959014643421.01394959014643441.76E-1615passed
/models/1/coefficients/3/estimatetolerance1.12206056889958331.12206056889958332.28E-1716passed
/models/1/coefficients/3/standard_errortolerance0.16171229678455190.161712296784551988.43E-1715passed
/models/1/coefficients/3/t_valuetolerance6.93862242519810576.9386224251981023.57E-1515passed
/models/1/coefficients/3/p_valuetail_probability1.7476517496610364e-061.747651749661046e-069.50E-2114passed
/models/1/coefficients/3/lowertolerance0.782315640406918260.7823156404069182.26E-1615passed
/models/1/coefficients/3/uppertolerance1.46180549739224831.46180549739224852.21E-1615passed
/models/1/coefficients/3/standardizedtolerance0.276282131413588480.27628213141358853.15E-1816passed
/models/1/coefficients/3/zero_ordertolerance0.203740894675598780.203740894675598782.91E-1816passed
/models/1/coefficients/3/partialtolerance0.853152068947611640.85315206894761133.32E-1615passed
/models/1/coefficients/3/parttolerance0.271661641148491890.271661641148491845.08E-1715passed
/models/1/coefficients/3/tolerancetolerance0.966832063023997710.96683206302399792.24E-1615passed
/models/1/coefficients/3/viftolerance1.03430578923113251.03430578923113232.45E-1615passed
/models/1/coefficients/4/estimatetolerance-0.27315446256399445-0.27315446256399461.66E-1615passed
/models/1/coefficients/4/standard_errortolerance0.226412591685981630.22641259168598181.66E-1615passed
/models/1/coefficients/4/t_valuetolerance-1.2064455449670419-1.20644554496704171.95E-1615passed
/models/1/coefficients/4/p_valuetail_probability0.243263579034017130.243263579034017241.11E-1615passed
/models/1/coefficients/4/lowertolerance-0.74882966662516814-0.74882966662516864.48E-1615passed
/models/1/coefficients/4/uppertolerance0.202520741497179190.20252074149717942.23E-1614passed
/models/1/coefficients/4/standardizedtolerance-0.048199735544381943-0.0481997355443819642.06E-1715passed
/models/1/coefficients/4/zero_ordertolerance0.0979415771086462190.097941577108646329.71E-1715passed
/models/1/coefficients/4/partialtolerance-0.27351831016121358-0.27351831016121332.73E-1615passed
/models/1/coefficients/4/parttolerance-0.047234876985351462-0.047234876985351433.45E-1715passed
/models/1/coefficients/4/tolerancetolerance0.960364872693618810.96036487269361884.15E-1918passed
/models/1/coefficients/4/viftolerance1.04127090487515761.04127090487515761.02E-1717passed
/models/1/collinearity/0/dimensionexact110.00E+2passed
/models/1/collinearity/0/eigenvaluetolerance2.36952618282843242.36952618282843241.46E-1717passed
/models/1/collinearity/0/condition_indextolerance11.00.00E+2passed
/models/1/collinearity/1/dimensionexact220.00E+2passed
/models/1/collinearity/1/eigenvaluetolerance1.04122092951935131.041220929519351.33E-1514passed
/models/1/collinearity/1/condition_indextolerance1.50854861807036421.5085486180703659.10E-1615passed
/models/1/collinearity/2/dimensionexact330.00E+2passed
/models/1/collinearity/2/eigenvaluetolerance0.752728124823528580.75272812482352842.25E-1615passed
/models/1/collinearity/2/condition_indextolerance1.77423721466447961.77423721466447982.11E-1615passed
/models/1/collinearity/3/dimensionexact440.00E+2passed
/models/1/collinearity/3/eigenvaluetolerance0.524362156229643110.52436215622964332.18E-1615passed
/models/1/collinearity/3/condition_indextolerance2.12576418441085262.1257641844108524.69E-1615passed
/models/1/collinearity/4/dimensionexact550.00E+2passed
/models/1/collinearity/4/eigenvaluetolerance0.31216260659904860.31216260659904862.36E-1817passed
/models/1/collinearity/4/condition_indextolerance2.75511871417570082.75511871417570084.41E-1716passed
/casewise/0/observedtolerance-0.21069097886642679-0.21069097886642682.40E-1816passed
/casewise/0/predictedtolerance-0.12356514775308289-0.12356514775308012.79E-1513passed
/casewise/0/residualtolerance-0.087125831113343905-0.08712583111334672.79E-1513passed
/casewise/0/standardized_residualtolerance-0.19986163961177841-0.199861639611784746.33E-1513passed
/casewise/0/studentized_residualtolerance-0.23290413679985744-0.232904136799855921.52E-1514passed
/casewise/0/deleted_studentized_residualtolerance-0.22668395849222167-0.226683958492220231.44E-1514passed
/casewise/0/leveragetolerance0.263615757095421510.26361575709542141.11E-1615passed
/casewise/0/cooks_distancetolerance0.0038837501176763440.0038837501176762925.21E-1713passed
/casewise/0/dfbetas/0tolerance0.000396815803297777870.00039681580329771496.30E-1712passed
/casewise/0/dfbetas/1tolerance0.0684775745344376880.068477574534437264.30E-1614passed
/casewise/0/dfbetas/2tolerance-0.028332668196641663-0.0283326681966414342.29E-1614passed
/casewise/0/dfbetas/3tolerance-0.085376857530706077-0.085376857530705476.11E-1614passed
/casewise/0/dfbetas/4tolerance0.0355291772404992340.0355291772404990052.29E-1614passed
/casewise/1/observedtolerance-0.81604839374920402-0.8160483937492044.82E-1817passed
/casewise/1/predictedtolerance-1.1730252071763674-1.17302520717636676.87E-1615passed
/casewise/1/residualtolerance0.356976813427163420.35697681342716277.22E-1614passed
/casewise/1/standardized_residualtolerance0.818884254224504950.8188842542245032.00E-1514passed
/casewise/1/studentized_residualtolerance0.893384146293340620.8933841462933487.33E-1514passed
/casewise/1/deleted_studentized_residualtolerance0.88812672230360690.88812672230361467.66E-1514passed
/casewise/1/leveragetolerance0.159827314697315450.159827314697315432.43E-1715passed
/casewise/1/cooks_distancetolerance0.0303660933669845750.0303660933669850644.89E-1613passed
/casewise/1/dfbetas/0tolerance0.242836769517125130.242836769517127132.00E-1514passed
/casewise/1/dfbetas/1tolerance-0.2807445538058298-0.28074455380583212.28E-1514passed
/casewise/1/dfbetas/2tolerance0.036296902589411720.0362969025894120253.05E-1614passed
/casewise/1/dfbetas/3tolerance-0.12280507298380892-0.122805072983809859.34E-1614passed
/casewise/1/dfbetas/4tolerance-0.078792508864626576-0.078792508864627236.52E-1614passed
/casewise/4/observedtolerance0.785198072135295470.78519807213529553.81E-1817passed
/casewise/4/predictedtolerance0.829375808534418830.82937580853442021.34E-1514passed
/casewise/4/residualtolerance-0.044177736399123327-0.04417773639912471.37E-1513passed
/casewise/4/standardized_residualtolerance-0.10134118341527584-0.101341183415278953.11E-1513passed
/casewise/4/studentized_residualtolerance-0.12379589786241581-0.123795897862416496.77E-1614passed
/casewise/4/deleted_studentized_residualtolerance-0.12035923612338065-0.120359236123381316.63E-1614passed
/casewise/4/leveragetolerance0.329869422438199440.329869422438199444.43E-1816passed
/casewise/4/cooks_distancetolerance0.00150877725650024940.00150877725650026631.69E-1713passed
/casewise/4/dfbetas/0tolerance-0.067837757478916252-0.067837757478916613.61E-1614passed
/casewise/4/dfbetas/1tolerance0.0344227421586826840.0344227421586828441.60E-1614passed
/casewise/4/dfbetas/2tolerance0.0581075949233358990.05810759492333622.99E-1614passed
/casewise/4/dfbetas/3tolerance0.0112401891681597860.0112401891681598455.88E-1714passed
/casewise/4/dfbetas/4tolerance0.0398815471133840280.0398815471133842642.36E-1614passed
/casewise/10/observedtolerance2.58580070911278392.5858007091127842.39E-1717passed
/casewise/10/predictedtolerance2.2216983257180522.2216983257180523.61E-1716passed
/casewise/10/residualtolerance0.3641023833947320.36410238339473191.12E-1615passed
/casewise/10/standardized_residualtolerance0.835229901418779950.83522990141877945.53E-1615passed
/casewise/10/studentized_residualtolerance0.915532214589072750.91553221458907052.22E-1514passed
/casewise/10/deleted_studentized_residualtolerance0.911206193694754970.91120619369475262.33E-1514passed
/casewise/10/leveragetolerance0.167728914196566160.167728914196566161.03E-1817passed
/casewise/10/cooks_distancetolerance0.0337847247428110090.0337847247428108361.73E-1614passed
/casewise/10/dfbetas/0tolerance0.368502604520445590.368502604520444649.47E-1614passed
/casewise/10/dfbetas/1tolerance-0.004592300234572735-0.0045923002345726666.93E-1713passed
/casewise/10/dfbetas/2tolerance-0.13253234281647136-0.132532342816470944.16E-1614passed
/casewise/10/dfbetas/3tolerance-0.089617569520224066-0.089617569520223772.91E-1614passed
/casewise/10/dfbetas/4tolerance-0.29826067979237059-0.29826067979236987.81E-1614passed
/casewise/22/observedtolerance5.59122296556430955.59122296556430953.27E-1717passed
/casewise/22/predictedtolerance5.33653647983824445.3365364798382441.78E-1717passed
/casewise/22/residualtolerance0.254686485726064950.25468648572606511.65E-1615passed
/casewise/22/standardized_residualtolerance0.584236132656841270.58423613265684152.19E-1615passed
/casewise/22/studentized_residualtolerance0.650309944709990930.65030994470998882.11E-1514passed
/casewise/22/deleted_studentized_residualtolerance0.639545255854397920.63954525585439582.11E-1514passed
/casewise/22/leveragetolerance0.192883854740395130.192883854740395075.79E-1715passed
/casewise/22/cooks_distancetolerance0.0202129931276872830.0202129931276871481.35E-1614passed
/casewise/22/dfbetas/0tolerance0.148378933129668270.148378933129667854.20E-1614passed
/casewise/22/dfbetas/1tolerance0.243561185839166750.243561185839165868.89E-1614passed
/casewise/22/dfbetas/2tolerance0.0390227783270928680.0390227783270926941.74E-1614passed
/casewise/22/dfbetas/3tolerance-0.054013342288140136-0.054013342288139922.15E-1614passed
/casewise/22/dfbetas/4tolerance-0.1153840245752258-0.115384024575225423.75E-1614passed
/nexact23230.00E+2passed
/missing_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

50  emit <- function(id, title, kind, config, outputs, notes) {
51    write_expected("regression", id, list(
52      schema_version = "2.0.0", case_id = id, outputs = outputs, expected_error = NULL,
53      deferred_outputs = list(), generator_run = run))
54    write_case("regression", id, list(
55      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
56      entry_point = entry, menu_items = list(menu), title = title, case_kind = kind,
57      dataset = list(path = "reference_data/datasets/REG_BLOCKS.json", sha256 = ""),
58      excluded_rows = list(), configuration = config, "repeat" = 1L,
59      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"), sha256 = ""),
60      oracle = list(oracle_class = "independent_implementation", source = "R 4.6.1 stats::lm and step",
61                    citation = "R Core Team, stats package help: lm, anova.lm, confint.lm, step, AIC, BIC, rstandard, rstudent, hatvalues, cooks.distance, dfbetas; Belsley, Kuh and Welsch (1980), Regression Diagnostics, ch. 3.",
62                    generator = list(path = "generators/r/regression_blocks.R", sha256 = ""),
63                    independent = TRUE,
64                    convention_notes = c(notes, notes_common, "Expected computed from the exported CSV.")),
65      supported = TRUE, unsupported_reason = NULL,
66      requirement_ids = list("REQ-STAT-REG-REGRESSION-BLOCKS-001"),
67      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/regression_blocks.R


Blockwise regression forward AIC

Case STAT-REG-REGRESSION-BLOCKS-002 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm and step
Citation
R Core Team, stats package help: lm, anova.lm, confint.lm, step, AIC, BIC, rstandard, rstudent, hatvalues, cooks.distance, dfbetas; Belsley, Kuh and Welsch (1980), Regression Diagnostics, ch. 3.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_blocks.R · sha256 86184660dac4e79f
Dataset
reference_data/datasets/REG_BLOCKS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/models/0/r_squaredtolerance0.898605402867828310.89860540286782833.23E-1817passed
/models/0/adjusted_r_squaredtolerance0.888465943154611140.88846594315461112.56E-1817passed
/models/0/standard_errortolerance0.792783919632095490.79278391963209532.21E-1615passed
/models/0/df2exact20200.00E+2passed
/models/0/residual_sstolerance12.57012686454457512.5701268645445687.38E-1515passed
/models/0/regression_sstolerance111.40222689074416111.402226890744171.42E-1415passed
/models/0/f_valuetolerance88.62458437469426588.624584374694314.31E-1415passed
/models/0/p_valuetail_probability1.148545328910703e-101.1485453289106956e-107.37E-2514passed
/models/0/durbin_watsontolerance2.71686806562813032.71686806562813032.43E-1717passed
/models/0/r_squared_changetolerance0.898605402867828310.89860540286782833.23E-1817passed
/models/0/df1exact220.00E+2passed
/models/0/f_changetolerance88.62458437469426588.624584374694274.21E-1617passed
/models/0/p_changetail_probability1.148545328910703e-101.1485453289107003e-102.72E-2514passed
/models/0/predictorsset['x1', 'x2']['x1', 'x2']passed
/models/1/r_squaredtolerance0.970176901151252990.9701769011512535.52E-1918passed
/models/1/adjusted_r_squaredtolerance0.965467990806713970.9654679908067141.53E-1817passed
/models/1/standard_errortolerance0.441125340855682670.44112534085568281.09E-1615passed
/models/1/df2exact19190.00E+2passed
/models/1/residual_sstolerance3.69723976055580253.69723976055580348.90E-1615passed
/models/1/regression_sstolerance120.27511399473293120.275113994732933.52E-1516passed
/models/1/f_valuetolerance206.03002184494426206.030021844944233.02E-1415passed
/models/1/p_valuetail_probability1.1486630672467477e-141.1486630672467534e-145.72E-2914passed
/models/1/durbin_watsontolerance2.79630531838030552.7963053183803054.51E-1615passed
/models/1/r_squared_changetolerance0.0715714982834246840.071571498283424682.16E-1917passed
/models/1/df1exact110.00E+2passed
/models/1/f_changetolerance45.59749053181324345.5974905318131361.07E-1314passed
/models/1/p_changetail_probability1.8851110917256922e-061.8851110917257178e-062.56E-2013passed
/models/1/predictorsset['x1', 'x2', 'x3']['x1', 'x2', 'x3']passed
/models/1/enteredset['x3']['x3']passed
/models/1/selection/0/criteriontolerance59.37523721055206559.3752372105520442.14E-1415passed
/models/1/selection/0/actionexactstartstartpassed
/models/1/selection/1/actionexactaddaddpassed
/models/1/selection/1/predictorexactx3x3passed
/models/1/selection/1/criteriontolerance33.22929576450321333.229295764503227.31E-1515passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

187  forced <- stats::lm(y ~ x1 + x2, data = cc)
188  fwd <- stats::step(forced, scope = list(lower = ~ x1 + x2, upper = ~ x1 + x2 + x3 + x4 + x5),
189                     direction = "forward", k = 2, trace = 0)
190  chosen_fwd <- setdiff(attr(stats::terms(fwd), "term.labels"), c("x1", "x2"))
191  outputs <- c(model_outputs(0, c("x1", "x2"), NULL, with_coefficients = FALSE),
192               model_outputs(1, c("x1", "x2", chosen_fwd), c("x1", "x2"), with_coefficients = FALSE))
193  outputs[["/models/1/entered"]] <- setof(chosen_fwd, "predictors step() adds by AIC")
194  outputs[["/models/1/selection/0/criterion"]] <- tol(stats::AIC(forced), "AIC of the forced model (R's AIC with the variance counted)")
195  outputs[["/models/1/selection/0/action"]] <- list(comparison = "exact", value = "start", rationale = "the first move records the forced model")
196  first_added <- sub("^[+] ", "", as.character(fwd$anova$Step[2]))
197  stopifnot(length(chosen_fwd) >= 1, first_added %in% chosen_fwd)
198  outputs[["/models/1/selection/1/action"]] <- list(comparison = "exact", value = "add", rationale = "forward entry adds")
199  outputs[["/models/1/selection/1/predictor"]] <- list(comparison = "exact", value = first_added, rationale = "the first predictor step() adds")
200  outputs[["/models/1/selection/1/criterion"]] <- tol(stats::AIC(fit_of(c("x1", "x2", first_added))), "AIC after the first addition")
201  emit("STAT-REG-REGRESSION-BLOCKS-002", "Blockwise regression forward AIC", "axis_coverage",
202       list(response = "y", blocks = list(list("x1", "x2"), list("x3", "x4", "x5")), method = "forward",
203            criterion = "aic", confidence_level = 0.95, casewise = "none", casewise_threshold = 3.0),
204       outputs, "step(direction = 'forward', k = 2) from y ~ x1 + x2 with scope x3 + x4 + x5; the reported criterion is AIC() of the fitted lm, which differs from extractAIC by a constant and selects the same moves.")

Whole script generators/r/regression_blocks.R


Blockwise regression backward BIC

Case STAT-REG-REGRESSION-BLOCKS-003 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm and step
Citation
R Core Team, stats package help: lm, anova.lm, confint.lm, step, AIC, BIC, rstandard, rstudent, hatvalues, cooks.distance, dfbetas; Belsley, Kuh and Welsch (1980), Regression Diagnostics, ch. 3.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_blocks.R · sha256 86184660dac4e79f
Dataset
reference_data/datasets/REG_BLOCKS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/models/0/r_squaredtolerance0.898605402867828310.89860540286782833.23E-1817passed
/models/0/adjusted_r_squaredtolerance0.888465943154611140.88846594315461112.56E-1817passed
/models/0/standard_errortolerance0.792783919632095490.79278391963209532.21E-1615passed
/models/0/df2exact20200.00E+2passed
/models/0/residual_sstolerance12.57012686454457512.5701268645445687.38E-1515passed
/models/0/regression_sstolerance111.40222689074416111.402226890744171.42E-1415passed
/models/0/f_valuetolerance88.62458437469426588.624584374694314.31E-1415passed
/models/0/p_valuetail_probability1.148545328910703e-101.1485453289106956e-107.37E-2514passed
/models/0/durbin_watsontolerance2.71686806562813032.71686806562813032.43E-1717passed
/models/0/r_squared_changetolerance0.898605402867828310.89860540286782833.23E-1817passed
/models/0/df1exact220.00E+2passed
/models/0/f_changetolerance88.62458437469426588.624584374694274.21E-1617passed
/models/0/p_changetail_probability1.148545328910703e-101.1485453289107003e-102.72E-2514passed
/models/0/predictorsset['x1', 'x2']['x1', 'x2']passed
/models/1/r_squaredtolerance0.970176901151252990.9701769011512535.52E-1918passed
/models/1/adjusted_r_squaredtolerance0.965467990806713970.9654679908067141.53E-1817passed
/models/1/standard_errortolerance0.441125340855682670.44112534085568281.09E-1615passed
/models/1/df2exact19190.00E+2passed
/models/1/residual_sstolerance3.69723976055580253.69723976055580348.90E-1615passed
/models/1/regression_sstolerance120.27511399473293120.275113994732933.52E-1516passed
/models/1/f_valuetolerance206.03002184494426206.030021844944233.02E-1415passed
/models/1/p_valuetail_probability1.1486630672467477e-141.1486630672467534e-145.72E-2914passed
/models/1/durbin_watsontolerance2.79630531838030552.7963053183803054.51E-1615passed
/models/1/r_squared_changetolerance0.0715714982834246840.071571498283424682.16E-1917passed
/models/1/df1exact110.00E+2passed
/models/1/f_changetolerance45.59749053181324345.5974905318131361.07E-1314passed
/models/1/p_changetail_probability1.8851110917256922e-061.8851110917257178e-062.56E-2013passed
/models/1/predictorsset['x1', 'x2', 'x3']['x1', 'x2', 'x3']passed
/models/1/enteredset['x3']['x3']passed
/models/1/selection/0/criteriontolerance43.22193530936868943.221935309368722.85E-1415passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

207  full <- stats::lm(y ~ x1 + x2 + x3 + x4 + x5, data = cc)
208  bwd <- stats::step(full, scope = list(lower = ~ x1 + x2, upper = ~ x1 + x2 + x3 + x4 + x5),
209                     direction = "backward", k = log(n_complete), trace = 0)
210  chosen_bwd <- setdiff(attr(stats::terms(bwd), "term.labels"), c("x1", "x2"))
211  outputs <- c(model_outputs(0, c("x1", "x2"), NULL, with_coefficients = FALSE),
212               model_outputs(1, c("x1", "x2", chosen_bwd), c("x1", "x2"), with_coefficients = FALSE))
213  outputs[["/models/1/entered"]] <- setof(chosen_bwd, "predictors step() keeps by BIC")
214  outputs[["/models/1/selection/0/criterion"]] <- tol(stats::BIC(full), "BIC of the full model (R's BIC with the variance counted)")
215  emit("STAT-REG-REGRESSION-BLOCKS-003", "Blockwise regression backward BIC", "axis_coverage",
216       list(response = "y", blocks = list(list("x1", "x2"), list("x3", "x4", "x5")), method = "backward",
217            criterion = "bic", confidence_level = 0.95, casewise = "none", casewise_threshold = 3.0),
218       outputs, "step(direction = 'backward', k = log(n)) from the full model with x1 + x2 forced; the start criterion is BIC() of the full block-2 model.")

Whole script generators/r/regression_blocks.R


Blockwise regression stepwise AIC 90 percent outliers

Case STAT-REG-REGRESSION-BLOCKS-004 · status: passed

Reference

Evidence type
R
Source
R 4.6.1 stats::lm and step
Citation
R Core Team, stats package help: lm, anova.lm, confint.lm, step, AIC, BIC, rstandard, rstudent, hatvalues, cooks.distance, dfbetas; Belsley, Kuh and Welsch (1980), Regression Diagnostics, ch. 3.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_blocks.R · sha256 86184660dac4e79f
Dataset
reference_data/datasets/REG_BLOCKS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/models/0/r_squaredtolerance0.898605402867828310.89860540286782833.23E-1817passed
/models/0/adjusted_r_squaredtolerance0.888465943154611140.88846594315461112.56E-1817passed
/models/0/standard_errortolerance0.792783919632095490.79278391963209532.21E-1615passed
/models/0/df2exact20200.00E+2passed
/models/0/residual_sstolerance12.57012686454457512.5701268645445687.38E-1515passed
/models/0/regression_sstolerance111.40222689074416111.402226890744171.42E-1415passed
/models/0/f_valuetolerance88.62458437469426588.624584374694314.31E-1415passed
/models/0/p_valuetail_probability1.148545328910703e-101.1485453289106956e-107.37E-2514passed
/models/0/durbin_watsontolerance2.71686806562813032.71686806562813032.43E-1717passed
/models/0/r_squared_changetolerance0.898605402867828310.89860540286782833.23E-1817passed
/models/0/df1exact220.00E+2passed
/models/0/f_changetolerance88.62458437469426588.624584374694274.21E-1617passed
/models/0/p_changetail_probability1.148545328910703e-101.1485453289107003e-102.72E-2514passed
/models/0/predictorsset['x1', 'x2']['x1', 'x2']passed
/models/0/coefficients/0/estimatetolerance2.42012673835808962.4201267383580894.87E-1615passed
/models/0/coefficients/0/standard_errortolerance0.190127127054539570.19012712705453957.91E-1715passed
/models/0/coefficients/0/t_valuetolerance12.72899231083345512.7289923108334583.39E-1515passed
/models/0/coefficients/0/p_valuetail_probability4.7621401128199224e-114.76214011281992e-112.58E-2615passed
/models/0/coefficients/0/lowertolerance2.09221101385300662.09221101385300663.74E-1716passed
/models/0/coefficients/0/uppertolerance2.74804246286317212.74804246286317174.36E-1615passed
/models/0/coefficients/1/estimatetolerance1.50729632942692751.50729632942692641.13E-1515passed
/models/0/coefficients/1/standard_errortolerance0.119975210764557370.11997521076455736.48E-1715passed
/models/0/coefficients/1/t_valuetolerance12.56339805382702812.5633980538270243.99E-1515passed
/models/0/coefficients/1/p_valuetail_probability6.0181899344203938e-116.018189934420422e-112.84E-2514passed
/models/0/coefficients/1/lowertolerance1.30037289472302891.3003728947230289.28E-1615passed
/models/0/coefficients/1/uppertolerance1.71421976413082591.71421976413082481.13E-1515passed
/models/0/coefficients/1/standardizedtolerance0.898445299469412320.89844529946941157.80E-1615passed
/models/0/coefficients/1/zero_ordertolerance0.91986923051148550.91986923051148552.79E-1817passed
/models/0/coefficients/1/partialtolerance0.942092729138495620.94209272913849561.48E-1817passed
/models/0/coefficients/1/parttolerance0.894539513987959830.89453951398795962.19E-1615passed
/models/0/coefficients/1/tolerancetolerance0.991324356131298660.99132435613129873.74E-1817passed
/models/0/coefficients/1/viftolerance1.00875156936782881.00875156936782883.75E-1716passed
/models/0/coefficients/2/estimatetolerance-0.68071791648298607-0.68071791648298592.18E-1615passed
/models/0/coefficients/2/standard_errortolerance0.211642751245254420.211642751245254338.53E-1715passed
/models/0/coefficients/2/t_valuetolerance-3.2163535603171267-3.2163535603171274.58E-1615passed
/models/0/coefficients/2/p_valuetail_probability0.00433141435750920930.0043314143575092054.32E-1815passed
/models/0/coefficients/2/lowertolerance-1.0457420305376226-1.0457420305376226.64E-1615passed
/models/0/coefficients/2/uppertolerance-0.31569380242834966-0.315693802428349663.11E-1817passed
/models/0/coefficients/2/standardizedtolerance-0.23001083984745457-0.230010839847454552.43E-1715passed
/models/0/coefficients/2/zero_ordertolerance-0.31369485297341543-0.31369485297341545.77E-1715passed
/models/0/coefficients/2/partialtolerance-0.58387611718549526-0.58387611718549541.09E-1615passed
/models/0/coefficients/2/parttolerance-0.22901092032048578-0.229010920320485762.42E-1715passed
/models/0/coefficients/2/tolerancetolerance0.991324356131298660.99132435613129842.26E-1615passed
/models/0/coefficients/2/viftolerance1.00875156936782881.0087515693678292.60E-1615passed
/models/0/collinearity/0/dimensionexact110.00E+2passed
/models/0/collinearity/0/eigenvaluetolerance1.48491475042452281.48491475042452216.55E-1615passed
/models/0/collinearity/0/condition_indextolerance11.00.00E+2passed
/models/0/collinearity/1/dimensionexact220.00E+2passed
/models/0/collinearity/1/eigenvaluetolerance1.01311417351271051.01311417351271058.08E-1817passed
/models/0/collinearity/1/condition_indextolerance1.21065824779452311.2106582477945231.87E-1615passed
/models/0/collinearity/2/dimensionexact330.00E+2passed
/models/0/collinearity/2/eigenvaluetolerance0.501971076062768360.50197107606276884.46E-1615passed
/models/0/collinearity/2/condition_indextolerance1.71993254308602261.71993254308602151.10E-1515passed
/models/1/r_squaredtolerance0.970176901151252990.9701769011512535.52E-1918passed
/models/1/adjusted_r_squaredtolerance0.965467990806713970.9654679908067141.53E-1817passed
/models/1/standard_errortolerance0.441125340855682670.44112534085568281.09E-1615passed
/models/1/df2exact19190.00E+2passed
/models/1/residual_sstolerance3.69723976055580253.69723976055580348.90E-1615passed
/models/1/regression_sstolerance120.27511399473293120.275113994732933.52E-1516passed
/models/1/f_valuetolerance206.03002184494426206.030021844944233.02E-1415passed
/models/1/p_valuetail_probability1.1486630672467477e-141.1486630672467534e-145.72E-2914passed
/models/1/durbin_watsontolerance2.79630531838030552.7963053183803054.51E-1615passed
/models/1/r_squared_changetolerance0.0715714982834246840.071571498283424682.16E-1917passed
/models/1/df1exact110.00E+2passed
/models/1/f_changetolerance45.59749053181324345.5974905318131361.07E-1314passed
/models/1/p_changetail_probability1.8851110917256922e-061.8851110917257178e-062.56E-2013passed
/models/1/predictorsset['x1', 'x2', 'x3']['x1', 'x2', 'x3']passed
/models/1/coefficients/0/estimatetolerance1.92266247098375991.92266247098375993.63E-1716passed
/models/1/coefficients/0/standard_errortolerance0.128915298276295830.12891529827629582.67E-1715passed
/models/1/coefficients/0/t_valuetolerance14.91415291040976114.9141529104097632.21E-1515passed
/models/1/coefficients/0/p_valuetail_probability6.0884761849485466e-126.088476184948554e-127.27E-2714passed
/models/1/coefficients/0/lowertolerance1.69975079882715251.69975079882715272.42E-1615passed
/models/1/coefficients/0/uppertolerance2.1455741431403672.1455741431403671.50E-1717passed
/models/1/coefficients/1/estimatetolerance1.53189940625773071.53189940625772988.99E-1615passed
/models/1/coefficients/1/standard_errortolerance0.0668566438199420660.066856643819942051.39E-1715passed
/models/1/coefficients/1/t_valuetolerance22.91319633667872322.9131963366787131.04E-1415passed
/models/1/coefficients/1/p_valuetail_probability2.6526790729496149e-152.652679072949644e-152.92E-2913passed
/models/1/coefficients/1/lowertolerance1.41629538976047151.41629538976047068.53E-1615passed
/models/1/coefficients/1/uppertolerance1.64750342275498981.6475034227549898.45E-1615passed
/models/1/coefficients/1/standardizedtolerance0.913110311451179710.91311031145117925.53E-1615passed
/models/1/coefficients/1/zero_ordertolerance0.91986923051148550.91986923051148552.79E-1817passed
/models/1/coefficients/1/partialtolerance0.982382035967663940.98238203596766422.18E-1615passed
/models/1/coefficients/1/parttolerance0.907789719149625140.90778971914962553.34E-1615passed
/models/1/coefficients/1/tolerancetolerance0.988380174914204050.9883801749142043.97E-1817passed
/models/1/coefficients/1/viftolerance1.01175643277831281.01175643277831284.19E-1716passed
/models/1/coefficients/2/estimatetolerance-0.71313718804413928-0.71313718804413884.43E-1615passed
/models/1/coefficients/2/standard_errortolerance0.117861288084085820.117861288084085831.01E-1716passed
/models/1/coefficients/2/t_valuetolerance-6.0506481783515342-6.050648178351534.41E-1515passed
/models/1/coefficients/2/p_valuetail_probability8.0652344809386006e-068.065234480938704e-061.03E-1913passed
/models/1/coefficients/2/lowertolerance-0.91693500847850462-0.91693500847850415.54E-1615passed
/models/1/coefficients/2/uppertolerance-0.50933936760977394-0.50933936760977363.32E-1615passed
/models/1/coefficients/2/standardizedtolerance-0.24096513339323042-0.240965133393230281.40E-1615passed
/models/1/coefficients/2/zero_ordertolerance-0.31369485297341543-0.31369485297341545.77E-1715passed
/models/1/coefficients/2/partialtolerance-0.81137965315309635-0.81137965315309632.60E-1817passed
/models/1/coefficients/2/parttolerance-0.23971846309833097-0.239718463098330974.99E-1816passed
/models/1/coefficients/2/tolerancetolerance0.989679458128792990.98967945812879282.24E-1615passed
/models/1/coefficients/2/viftolerance1.0104281661970841.01042816619708421.99E-1615passed
/models/1/coefficients/3/estimatetolerance1.08927642564819331.08927642564819266.97E-1615passed
/models/1/coefficients/3/standard_errortolerance0.161312355823283490.16131235582328352.46E-1816passed
/models/1/coefficients/3/t_valuetolerance6.75259139381417046.7525913938141664.45E-1515passed
/models/1/coefficients/3/p_valuetail_probability1.8851110917256994e-061.8851110917257148e-061.54E-2014passed
/models/1/coefficients/3/lowertolerance0.810345938290343510.81034593829034314.39E-1615passed
/models/1/coefficients/3/uppertolerance1.3682069130060431.36820691300604218.64E-1615passed
/models/1/coefficients/3/standardizedtolerance0.2682097748714230.26820977487142291.10E-1615passed
/models/1/coefficients/3/zero_ordertolerance0.203740894675598780.203740894675598782.91E-1816passed
/models/1/coefficients/3/partialtolerance0.840161250670185880.84016125067018544.47E-1615passed
/models/1/coefficients/3/parttolerance0.267528499946126710.267528499946126655.78E-1715passed
/models/1/coefficients/3/tolerancetolerance0.994926287243608650.99492628724360873.55E-1817passed
/models/1/coefficients/3/viftolerance1.00509958659394541.00509958659394541.78E-1716passed
/models/1/collinearity/0/dimensionexact110.00E+2passed
/models/1/collinearity/0/eigenvaluetolerance1.97815829468260171.97815829468260173.66E-1716passed
/models/1/collinearity/0/condition_indextolerance11.00.00E+2passed
/models/1/collinearity/1/dimensionexact220.00E+2passed
/models/1/collinearity/1/eigenvaluetolerance1.01326791324768941.01326791324768941.09E-1716passed
/models/1/collinearity/1/condition_indextolerance1.39723152424181031.39723152424181033.48E-1817passed
/models/1/collinearity/2/dimensionexact330.00E+2passed
/models/1/collinearity/2/eigenvaluetolerance0.675195668122346770.67519566812234713.34E-1615passed
/models/1/collinearity/2/condition_indextolerance1.7116528874935991.71165288749359864.37E-1615passed
/models/1/collinearity/3/dimensionexact440.00E+2passed
/models/1/collinearity/3/eigenvaluetolerance0.333378123947359330.333378123947359331.69E-1817passed
/models/1/collinearity/3/condition_indextolerance2.4359141127845862.4359141127845862.75E-1716passed
/models/1/enteredset['x3']['x3']passed
/casewise_truncatedexactFalseFalsepassed
/casewise/0/observedtolerance-2.0459930473501053-2.04599304735010534.27E-1716passed
/casewise/0/standardized_residualtolerance-1.594846084387211-1.5948460843872165.09E-1514passed
/casewise/1/observedtolerance2.97012312794563422.97012312794563374.88E-1615passed
/casewise/1/standardized_residualtolerance1.55742406057240191.55742406057242.00E-1514passed
/casewise/2/observedtolerance4.72671578027658254.72671578027658254.30E-1717passed
/casewise/2/standardized_residualtolerance1.56782734162600731.56782734162600871.37E-1515passed
/casewise/3absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

235  outputs[[paste0("/casewise/", length(flagged))]] <- list(comparison = "absent", rationale = paste("exactly", length(flagged), "rows exceed |e / s| > 1.5"))
236  emit("STAT-REG-REGRESSION-BLOCKS-004", "Blockwise regression stepwise AIC 90 percent outliers", "axis_coverage",
237       list(response = "y", blocks = list(list("x1", "x2"), list("x3", "x4", "x5")), method = "stepwise",
238            criterion = "aic", confidence_level = 0.90, casewise = "outliers", casewise_threshold = 1.5),
239       outputs, "step(direction = 'both', k = 2) from y ~ x1 + x2; confint(level = 0.90); the casewise table lists rows with |e / s| > 1.5 in retained order.")

Whole script generators/r/regression_blocks.R


Independent equal-variance REML repeated-measures model reproduces OLS

Case STAT-REG-REPCOV-001 · status: passed

Reference

Evidence type
Analytical
Source
Closed-form OLS and balanced-design variance-component algebra
Citation
Searle, Casella and McCulloch (1992), Variance Components; Satterthwaite (1946)
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/REG_REPCOV_OLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/[term=Intercept]/estimatetolerance3.9797404916250363.97974049162503362.37E-1515passed
/coefficients/[term=Intercept]/std_errortolerance0.20277570671987980.20277570269844974.02E-97passed
/coefficients/[term=Intercept]/dftolerance26.026.0000000000000181.78E-1415passed
/coefficients/[term=x]/estimatetolerance0.80969406215787290.80969406215787431.42E-1514passed
/coefficients/[term=x]/std_errortolerance0.17447338530697140.174473381846830453.46E-97passed
/coefficients/[term=x]/dftolerance26.026.000000000000032.84E-1414passed
/variance_parameters/[name=Residual variance]/estimatetolerance1.07452926426966561.0745292216497254.26E-87passed

Convention


Balanced random-intercept Satterthwaite df equal the classical stratum df

Case STAT-REG-REPCOV-002 · status: passed

Reference

Evidence type
Analytical
Source
Closed-form OLS and balanced-design variance-component algebra
Citation
Searle, Casella and McCulloch (1992), Variance Components; Satterthwaite (1946)
Reference software
not recorded
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/REG_REPCOV_BALANCED.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/[term=arm]/estimatetolerance0.67772912500000020.67772912500000036.54E-1716passed
/coefficients/[term=w]/estimatetolerance0.37753965833333360.37753965833333341.83E-1615passed
/coefficients/[term=arm]/dftolerance10.09.9999999999999415.86E-1414passed
/coefficients/[term=w]/dftolerance35.035.0000000000000646.39E-1414passed

Convention


Regression II complete row alignment STAT-REG-ROWS-ADV-DA-LDA-001

Case STAT-REG-ROWS-ADV-DA-LDA-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_rows.R · sha256 504917bf3c718b60
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:0STAT-REG-ROWS-ADV-DA-LDA-001:0passed
/0/predictedexactAApassed
/0/probabilities/0tolerance0.999999931743970950.9999999317439711.03E-1817passed
/1/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:1STAT-REG-ROWS-ADV-DA-LDA-001:1passed
/1/predictedexactAApassed
/1/probabilities/0tolerance0.999999931743970950.9999999317439711.03E-1817passed
/2/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:2STAT-REG-ROWS-ADV-DA-LDA-001:2passed
/2/predictedexactAApassed
/2/probabilities/0tolerance0.999999931743970950.9999999317439711.03E-1817passed
/3/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:3STAT-REG-ROWS-ADV-DA-LDA-001:3passed
/3/predictedexactAApassed
/3/probabilities/0tolerance0.999999931743970950.9999999317439711.03E-1817passed
/4/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:4STAT-REG-ROWS-ADV-DA-LDA-001:4passed
/4/predictedexactAApassed
/4/probabilities/0tolerance0.999983298578152050.9999832985781523.31E-1817passed
/5/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:5STAT-REG-ROWS-ADV-DA-LDA-001:5passed
/5/predictedexactAApassed
/5/probabilities/0tolerance0.999983298578152050.9999832985781523.31E-1817passed
/6/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:6STAT-REG-ROWS-ADV-DA-LDA-001:6passed
/6/predictedexactAApassed
/6/probabilities/0tolerance0.999983298578152050.9999832985781523.31E-1817passed
/7/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:7STAT-REG-ROWS-ADV-DA-LDA-001:7passed
/7/predictedexactAApassed
/7/probabilities/0tolerance0.999983298578152050.9999832985781523.31E-1817passed
/8/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:8STAT-REG-ROWS-ADV-DA-LDA-001:8passed
/8/predictedexactAApassed
/8/probabilities/0tolerance0.995929862284103960.9959298622841041.66E-1817passed
/9/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:9STAT-REG-ROWS-ADV-DA-LDA-001:9passed
/9/predictedexactAApassed
/9/probabilities/0tolerance0.995929862284103960.9959298622841041.66E-1817passed
/10/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:10STAT-REG-ROWS-ADV-DA-LDA-001:10passed
/10/predictedexactAApassed
/10/probabilities/0tolerance0.995929862284103960.9959298622841041.66E-1817passed
/11/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:11STAT-REG-ROWS-ADV-DA-LDA-001:11passed
/11/predictedexactAApassed
/11/probabilities/0tolerance0.995929862284103960.9959298622841041.66E-1817passed
/12/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:12STAT-REG-ROWS-ADV-DA-LDA-001:12passed
/12/predictedexactBBpassed
/12/probabilities/0tolerance0.00407013771589611640.0040701377158961166.56E-2117passed
/13/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:13STAT-REG-ROWS-ADV-DA-LDA-001:13passed
/13/predictedexactBBpassed
/13/probabilities/0tolerance0.00407013771589611640.0040701377158961166.56E-2117passed
/14/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:14STAT-REG-ROWS-ADV-DA-LDA-001:14passed
/14/predictedexactBBpassed
/14/probabilities/0tolerance0.00407013771589611640.0040701377158961166.56E-2117passed
/15/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:15STAT-REG-ROWS-ADV-DA-LDA-001:15passed
/15/predictedexactBBpassed
/15/probabilities/0tolerance0.00407013771589611640.0040701377158961166.56E-2117passed
/16/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:16STAT-REG-ROWS-ADV-DA-LDA-001:16passed
/16/predictedexactBBpassed
/16/probabilities/0tolerance1.6701421848095093e-051.6701421848095093e-054.88E-2216passed
/17/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:17STAT-REG-ROWS-ADV-DA-LDA-001:17passed
/17/predictedexactBBpassed
/17/probabilities/0tolerance1.6701421848095093e-051.6701421848095093e-054.88E-2216passed
/18/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:18STAT-REG-ROWS-ADV-DA-LDA-001:18passed
/18/predictedexactBBpassed
/18/probabilities/0tolerance1.6701421848095093e-051.6701421848095093e-054.88E-2216passed
/19/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:19STAT-REG-ROWS-ADV-DA-LDA-001:19passed
/19/predictedexactBBpassed
/19/probabilities/0tolerance1.6701421848095093e-051.6701421848095093e-054.88E-2216passed
/20/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:20STAT-REG-ROWS-ADV-DA-LDA-001:20passed
/20/predictedexactBBpassed
/20/probabilities/0tolerance6.8256029104462388e-086.825602910446239e-081.61E-2517passed
/21/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:21STAT-REG-ROWS-ADV-DA-LDA-001:21passed
/21/predictedexactBBpassed
/21/probabilities/0tolerance6.8256029104462388e-086.825602910446239e-081.61E-2517passed
/22/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:22STAT-REG-ROWS-ADV-DA-LDA-001:22passed
/22/predictedexactBBpassed
/22/probabilities/0tolerance6.8256029104462388e-086.825602910446239e-081.61E-2517passed
/23/row_idexactSTAT-REG-ROWS-ADV-DA-LDA-001:23STAT-REG-ROWS-ADV-DA-LDA-001:23passed
/23/predictedexactBBpassed
/23/probabilities/0tolerance6.8256029104462388e-086.825602910446239e-081.61E-2517passed

Convention

Reference code

Whole script generators/r/regression_advanced_rows.R


Regression II complete row alignment STAT-REG-ROWS-ADV-DA-QDA-001

Case STAT-REG-ROWS-ADV-DA-QDA-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_rows.R · sha256 504917bf3c718b60
Dataset
reference_data/datasets/REG_DISCRIMINANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:0STAT-REG-ROWS-ADV-DA-QDA-001:0passed
/0/predictedexactAApassed
/0/probabilities/0tolerance0.999999931743970950.9999999317439711.03E-1817passed
/1/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:1STAT-REG-ROWS-ADV-DA-QDA-001:1passed
/1/predictedexactAApassed
/1/probabilities/0tolerance0.999999931743970950.9999999317439711.03E-1817passed
/2/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:2STAT-REG-ROWS-ADV-DA-QDA-001:2passed
/2/predictedexactAApassed
/2/probabilities/0tolerance0.999999931743970950.9999999317439711.03E-1817passed
/3/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:3STAT-REG-ROWS-ADV-DA-QDA-001:3passed
/3/predictedexactAApassed
/3/probabilities/0tolerance0.999999931743970950.9999999317439711.03E-1817passed
/4/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:4STAT-REG-ROWS-ADV-DA-QDA-001:4passed
/4/predictedexactAApassed
/4/probabilities/0tolerance0.999983298578152050.9999832985781523.31E-1817passed
/5/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:5STAT-REG-ROWS-ADV-DA-QDA-001:5passed
/5/predictedexactAApassed
/5/probabilities/0tolerance0.999983298578152050.9999832985781523.31E-1817passed
/6/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:6STAT-REG-ROWS-ADV-DA-QDA-001:6passed
/6/predictedexactAApassed
/6/probabilities/0tolerance0.999983298578152050.9999832985781523.31E-1817passed
/7/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:7STAT-REG-ROWS-ADV-DA-QDA-001:7passed
/7/predictedexactAApassed
/7/probabilities/0tolerance0.999983298578152050.9999832985781523.31E-1817passed
/8/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:8STAT-REG-ROWS-ADV-DA-QDA-001:8passed
/8/predictedexactAApassed
/8/probabilities/0tolerance0.995929862284103960.9959298622841041.66E-1817passed
/9/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:9STAT-REG-ROWS-ADV-DA-QDA-001:9passed
/9/predictedexactAApassed
/9/probabilities/0tolerance0.995929862284103960.9959298622841041.66E-1817passed
/10/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:10STAT-REG-ROWS-ADV-DA-QDA-001:10passed
/10/predictedexactAApassed
/10/probabilities/0tolerance0.995929862284103960.9959298622841041.66E-1817passed
/11/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:11STAT-REG-ROWS-ADV-DA-QDA-001:11passed
/11/predictedexactAApassed
/11/probabilities/0tolerance0.995929862284103960.9959298622841041.66E-1817passed
/12/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:12STAT-REG-ROWS-ADV-DA-QDA-001:12passed
/12/predictedexactBBpassed
/12/probabilities/0tolerance0.00407013771589612680.0040701377158961161.04E-1714passed
/13/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:13STAT-REG-ROWS-ADV-DA-QDA-001:13passed
/13/predictedexactBBpassed
/13/probabilities/0tolerance0.00407013771589612680.0040701377158961161.04E-1714passed
/14/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:14STAT-REG-ROWS-ADV-DA-QDA-001:14passed
/14/predictedexactBBpassed
/14/probabilities/0tolerance0.00407013771589612680.0040701377158961161.04E-1714passed
/15/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:15STAT-REG-ROWS-ADV-DA-QDA-001:15passed
/15/predictedexactBBpassed
/15/probabilities/0tolerance0.00407013771589612680.0040701377158961161.04E-1714passed
/16/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:16STAT-REG-ROWS-ADV-DA-QDA-001:16passed
/16/predictedexactBBpassed
/16/probabilities/0tolerance1.6701421848095181e-051.6701421848095093e-058.85E-2014passed
/17/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:17STAT-REG-ROWS-ADV-DA-QDA-001:17passed
/17/predictedexactBBpassed
/17/probabilities/0tolerance1.6701421848095181e-051.6701421848095093e-058.85E-2014passed
/18/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:18STAT-REG-ROWS-ADV-DA-QDA-001:18passed
/18/predictedexactBBpassed
/18/probabilities/0tolerance1.6701421848095181e-051.6701421848095093e-058.85E-2014passed
/19/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:19STAT-REG-ROWS-ADV-DA-QDA-001:19passed
/19/predictedexactBBpassed
/19/probabilities/0tolerance1.6701421848095181e-051.6701421848095093e-058.85E-2014passed
/20/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:20STAT-REG-ROWS-ADV-DA-QDA-001:20passed
/20/predictedexactBBpassed
/20/probabilities/0tolerance6.8256029104462878e-086.825602910446264e-082.38E-2214passed
/21/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:21STAT-REG-ROWS-ADV-DA-QDA-001:21passed
/21/predictedexactBBpassed
/21/probabilities/0tolerance6.8256029104462878e-086.825602910446264e-082.38E-2214passed
/22/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:22STAT-REG-ROWS-ADV-DA-QDA-001:22passed
/22/predictedexactBBpassed
/22/probabilities/0tolerance6.8256029104462878e-086.825602910446264e-082.38E-2214passed
/23/row_idexactSTAT-REG-ROWS-ADV-DA-QDA-001:23STAT-REG-ROWS-ADV-DA-QDA-001:23passed
/23/predictedexactBBpassed
/23/probabilities/0tolerance6.8256029104462878e-086.825602910446264e-082.38E-2214passed

Convention

Reference code

Whole script generators/r/regression_advanced_rows.R


Regression II complete row alignment STAT-REG-ROWS-ADV-PLS-001

Case STAT-REG-ROWS-ADV-PLS-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_rows.R · sha256 504917bf3c718b60
Dataset
reference_data/datasets/REG_PLS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/row_idexactSTAT-REG-ROWS-ADV-PLS-001:0STAT-REG-ROWS-ADV-PLS-001:0passed
/0/predictedtolerance5.00836069876324875.0083606987632123.64E-1414passed
/1/row_idexactSTAT-REG-ROWS-ADV-PLS-001:1STAT-REG-ROWS-ADV-PLS-001:1passed
/1/predictedtolerance7.00781628854466557.0078162885446323.37E-1414passed
/2/row_idexactSTAT-REG-ROWS-ADV-PLS-001:2STAT-REG-ROWS-ADV-PLS-001:2passed
/2/predictedtolerance9.00727187832608149.0072718783260513.01E-1414passed
/3/row_idexactSTAT-REG-ROWS-ADV-PLS-001:3STAT-REG-ROWS-ADV-PLS-001:3passed
/3/predictedtolerance11.00672746810749911.0067274681074742.47E-1414passed
/4/row_idexactSTAT-REG-ROWS-ADV-PLS-001:4STAT-REG-ROWS-ADV-PLS-001:4passed
/4/predictedtolerance13.00618305788891513.0061830578888942.13E-1414passed
/5/row_idexactSTAT-REG-ROWS-ADV-PLS-001:5STAT-REG-ROWS-ADV-PLS-001:5passed
/5/predictedtolerance15.00563864767033115.0056386476703131.78E-1414passed
/6/row_idexactSTAT-REG-ROWS-ADV-PLS-001:6STAT-REG-ROWS-ADV-PLS-001:6passed
/6/predictedtolerance17.0050942374517517.0050942374517331.73E-1414passed
/7/row_idexactSTAT-REG-ROWS-ADV-PLS-001:7STAT-REG-ROWS-ADV-PLS-001:7passed
/7/predictedtolerance19.00454982723316319.0045498272331521.09E-1415passed
/8/row_idexactSTAT-REG-ROWS-ADV-PLS-001:8STAT-REG-ROWS-ADV-PLS-001:8passed
/8/predictedtolerance21.00400541701458221.004005417014571.04E-1415passed
/9/row_idexactSTAT-REG-ROWS-ADV-PLS-001:9STAT-REG-ROWS-ADV-PLS-001:9passed
/9/predictedtolerance23.00346100679599823.0034610067959953.41E-1515passed
/10/row_idexactSTAT-REG-ROWS-ADV-PLS-001:10STAT-REG-ROWS-ADV-PLS-001:10passed
/10/predictedtolerance25.00291659657741425.0029165965774145.45E-1717passed
/11/row_idexactSTAT-REG-ROWS-ADV-PLS-001:11STAT-REG-ROWS-ADV-PLS-001:11passed
/11/predictedtolerance27.0023721863588327.0023721863588343.52E-1515passed
/12/row_idexactSTAT-REG-ROWS-ADV-PLS-001:12STAT-REG-ROWS-ADV-PLS-001:12passed
/12/predictedtolerance29.00182777614024629.0018277761402536.98E-1515passed
/13/row_idexactSTAT-REG-ROWS-ADV-PLS-001:13STAT-REG-ROWS-ADV-PLS-001:13passed
/13/predictedtolerance31.00128336592166531.0012833659216727.44E-1515passed
/14/row_idexactSTAT-REG-ROWS-ADV-PLS-001:14STAT-REG-ROWS-ADV-PLS-001:14passed
/14/predictedtolerance33.00073895570308533.000738955703096.90E-1515passed
/15/row_idexactSTAT-REG-ROWS-ADV-PLS-001:15STAT-REG-ROWS-ADV-PLS-001:15passed
/15/predictedtolerance35.00019454548449735.000194545484511.44E-1415passed
/16/row_idexactSTAT-REG-ROWS-ADV-PLS-001:16STAT-REG-ROWS-ADV-PLS-001:16passed
/16/predictedtolerance36.9996501352659136.999650135265932.08E-1415passed
/17/row_idexactSTAT-REG-ROWS-ADV-PLS-001:17STAT-REG-ROWS-ADV-PLS-001:17passed
/17/predictedtolerance38.99910572504732938.999105725047352.13E-1415passed
/18/row_idexactSTAT-REG-ROWS-ADV-PLS-001:18STAT-REG-ROWS-ADV-PLS-001:18passed
/18/predictedtolerance40.99856131482874840.998561314828772.18E-1415passed
/19/row_idexactSTAT-REG-ROWS-ADV-PLS-001:19STAT-REG-ROWS-ADV-PLS-001:19passed
/19/predictedtolerance42.99801690461016842.998016904610192.12E-1415passed
/20/row_idexactSTAT-REG-ROWS-ADV-PLS-001:20STAT-REG-ROWS-ADV-PLS-001:20passed
/20/predictedtolerance44.99747249439157344.997472494391613.57E-1415passed
/21/row_idexactSTAT-REG-ROWS-ADV-PLS-001:21STAT-REG-ROWS-ADV-PLS-001:21passed
/21/predictedtolerance46.99692808417299346.9969280841730354.22E-1415passed
/22/row_idexactSTAT-REG-ROWS-ADV-PLS-001:22STAT-REG-ROWS-ADV-PLS-001:22passed
/22/predictedtolerance48.99638367395441248.9963836739544554.27E-1415passed
/23/row_idexactSTAT-REG-ROWS-ADV-PLS-001:23STAT-REG-ROWS-ADV-PLS-001:23passed
/23/predictedtolerance50.99583926373583250.9958392637358744.22E-1415passed
/24/row_idexactSTAT-REG-ROWS-ADV-PLS-001:24STAT-REG-ROWS-ADV-PLS-001:24passed
/24/predictedtolerance52.99529485351725152.9952948535172944.26E-1415passed
/25/row_idexactSTAT-REG-ROWS-ADV-PLS-001:25STAT-REG-ROWS-ADV-PLS-001:25passed
/25/predictedtolerance54.99475044329865654.994750443298715.71E-1414passed
/26/row_idexactSTAT-REG-ROWS-ADV-PLS-001:26STAT-REG-ROWS-ADV-PLS-001:26passed
/26/predictedtolerance56.99420603308007656.994206033080135.66E-1415passed
/27/row_idexactSTAT-REG-ROWS-ADV-PLS-001:27STAT-REG-ROWS-ADV-PLS-001:27passed
/27/predictedtolerance58.99366162286149558.993661622861555.70E-1415passed
/28/row_idexactSTAT-REG-ROWS-ADV-PLS-001:28STAT-REG-ROWS-ADV-PLS-001:28passed
/28/predictedtolerance60.99311721264291560.993117212642975.65E-1415passed
/29/row_idexactSTAT-REG-ROWS-ADV-PLS-001:29STAT-REG-ROWS-ADV-PLS-001:29passed
/29/predictedtolerance62.99257280242433462.992572802424395.69E-1415passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

90  pls_id <- "STAT-REG-ROWS-ADV-PLS-001"

Whole script generators/r/regression_advanced_rows.R


Regression II complete row alignment STAT-REG-ROWS-ADV-TREE-CLASS-001

Case STAT-REG-ROWS-ADV-TREE-CLASS-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_rows.R · sha256 504917bf3c718b60
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:0STAT-REG-ROWS-ADV-TREE-CLASS-001:0passed
/0/tree_scoreexactAApassed
/1/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:1STAT-REG-ROWS-ADV-TREE-CLASS-001:1passed
/1/tree_scoreabsentNonepassed
/2/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:2STAT-REG-ROWS-ADV-TREE-CLASS-001:2passed
/2/tree_scoreexactAApassed
/3/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:3STAT-REG-ROWS-ADV-TREE-CLASS-001:3passed
/3/tree_scoreexactBBpassed
/4/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:4STAT-REG-ROWS-ADV-TREE-CLASS-001:4passed
/4/tree_scoreexactAApassed
/5/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:5STAT-REG-ROWS-ADV-TREE-CLASS-001:5passed
/5/tree_scoreexactBBpassed
/6/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:6STAT-REG-ROWS-ADV-TREE-CLASS-001:6passed
/6/tree_scoreexactAApassed
/7/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:7STAT-REG-ROWS-ADV-TREE-CLASS-001:7passed
/7/tree_scoreexactBBpassed
/8/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:8STAT-REG-ROWS-ADV-TREE-CLASS-001:8passed
/8/tree_scoreexactAApassed
/9/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:9STAT-REG-ROWS-ADV-TREE-CLASS-001:9passed
/9/tree_scoreexactBBpassed
/10/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:10STAT-REG-ROWS-ADV-TREE-CLASS-001:10passed
/10/tree_scoreexactAApassed
/11/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:11STAT-REG-ROWS-ADV-TREE-CLASS-001:11passed
/11/tree_scoreexactBBpassed
/12/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:12STAT-REG-ROWS-ADV-TREE-CLASS-001:12passed
/12/tree_scoreexactAApassed
/13/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:13STAT-REG-ROWS-ADV-TREE-CLASS-001:13passed
/13/tree_scoreexactBBpassed
/14/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:14STAT-REG-ROWS-ADV-TREE-CLASS-001:14passed
/14/tree_scoreexactAApassed
/15/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:15STAT-REG-ROWS-ADV-TREE-CLASS-001:15passed
/15/tree_scoreexactBBpassed
/16/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:16STAT-REG-ROWS-ADV-TREE-CLASS-001:16passed
/16/tree_scoreexactAApassed
/17/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:17STAT-REG-ROWS-ADV-TREE-CLASS-001:17passed
/17/tree_scoreexactBBpassed
/18/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:18STAT-REG-ROWS-ADV-TREE-CLASS-001:18passed
/18/tree_scoreexactAApassed
/19/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:19STAT-REG-ROWS-ADV-TREE-CLASS-001:19passed
/19/tree_scoreexactBBpassed
/20/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:20STAT-REG-ROWS-ADV-TREE-CLASS-001:20passed
/20/tree_scoreexactAApassed
/21/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:21STAT-REG-ROWS-ADV-TREE-CLASS-001:21passed
/21/tree_scoreexactBBpassed
/22/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:22STAT-REG-ROWS-ADV-TREE-CLASS-001:22passed
/22/tree_scoreexactAApassed
/23/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:23STAT-REG-ROWS-ADV-TREE-CLASS-001:23passed
/23/tree_scoreexactBBpassed
/24/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:24STAT-REG-ROWS-ADV-TREE-CLASS-001:24passed
/24/tree_scoreexactAApassed
/25/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:25STAT-REG-ROWS-ADV-TREE-CLASS-001:25passed
/25/tree_scoreexactBBpassed
/26/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:26STAT-REG-ROWS-ADV-TREE-CLASS-001:26passed
/26/tree_scoreexactAApassed
/27/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:27STAT-REG-ROWS-ADV-TREE-CLASS-001:27passed
/27/tree_scoreexactBBpassed
/28/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:28STAT-REG-ROWS-ADV-TREE-CLASS-001:28passed
/28/tree_scoreexactAApassed
/29/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:29STAT-REG-ROWS-ADV-TREE-CLASS-001:29passed
/29/tree_scoreexactBBpassed
/30/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:30STAT-REG-ROWS-ADV-TREE-CLASS-001:30passed
/30/tree_scoreexactAApassed
/31/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:31STAT-REG-ROWS-ADV-TREE-CLASS-001:31passed
/31/tree_scoreexactBBpassed
/32/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:32STAT-REG-ROWS-ADV-TREE-CLASS-001:32passed
/32/tree_scoreexactAApassed
/33/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:33STAT-REG-ROWS-ADV-TREE-CLASS-001:33passed
/33/tree_scoreexactBBpassed
/34/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:34STAT-REG-ROWS-ADV-TREE-CLASS-001:34passed
/34/tree_scoreexactAApassed
/35/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:35STAT-REG-ROWS-ADV-TREE-CLASS-001:35passed
/35/tree_scoreexactBBpassed
/36/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:36STAT-REG-ROWS-ADV-TREE-CLASS-001:36passed
/36/tree_scoreexactAApassed
/37/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:37STAT-REG-ROWS-ADV-TREE-CLASS-001:37passed
/37/tree_scoreexactBBpassed
/38/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:38STAT-REG-ROWS-ADV-TREE-CLASS-001:38passed
/38/tree_scoreexactAApassed
/39/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-CLASS-001:39STAT-REG-ROWS-ADV-TREE-CLASS-001:39passed
/39/tree_scoreexactBBpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

36  class_id <- "STAT-REG-ROWS-ADV-TREE-CLASS-001"

Whole script generators/r/regression_advanced_rows.R


Regression II complete row alignment STAT-REG-ROWS-ADV-TREE-REG-001

Case STAT-REG-ROWS-ADV-TREE-REG-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_rows.R · sha256 504917bf3c718b60
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:0STAT-REG-ROWS-ADV-TREE-REG-001:0passed
/0/tree_scoretolerance-2-2.00.00E+2passed
/1/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:1STAT-REG-ROWS-ADV-TREE-REG-001:1passed
/1/tree_scoretolerance22.00.00E+2passed
/2/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:2STAT-REG-ROWS-ADV-TREE-REG-001:2passed
/2/tree_scoretolerance-2-2.00.00E+2passed
/3/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:3STAT-REG-ROWS-ADV-TREE-REG-001:3passed
/3/tree_scoretolerance22.00.00E+2passed
/4/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:4STAT-REG-ROWS-ADV-TREE-REG-001:4passed
/4/tree_scoretolerance-2-2.00.00E+2passed
/5/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:5STAT-REG-ROWS-ADV-TREE-REG-001:5passed
/5/tree_scoretolerance22.00.00E+2passed
/6/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:6STAT-REG-ROWS-ADV-TREE-REG-001:6passed
/6/tree_scoretolerance-2-2.00.00E+2passed
/7/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:7STAT-REG-ROWS-ADV-TREE-REG-001:7passed
/7/tree_scoretolerance22.00.00E+2passed
/8/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:8STAT-REG-ROWS-ADV-TREE-REG-001:8passed
/8/tree_scoretolerance-2-2.00.00E+2passed
/9/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:9STAT-REG-ROWS-ADV-TREE-REG-001:9passed
/9/tree_scoretolerance22.00.00E+2passed
/10/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:10STAT-REG-ROWS-ADV-TREE-REG-001:10passed
/10/tree_scoretolerance-2-2.00.00E+2passed
/11/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:11STAT-REG-ROWS-ADV-TREE-REG-001:11passed
/11/tree_scoretolerance22.00.00E+2passed
/12/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:12STAT-REG-ROWS-ADV-TREE-REG-001:12passed
/12/tree_scoretolerance-2-2.00.00E+2passed
/13/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:13STAT-REG-ROWS-ADV-TREE-REG-001:13passed
/13/tree_scoretolerance22.00.00E+2passed
/14/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:14STAT-REG-ROWS-ADV-TREE-REG-001:14passed
/14/tree_scoretolerance-2-2.00.00E+2passed
/15/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:15STAT-REG-ROWS-ADV-TREE-REG-001:15passed
/15/tree_scoretolerance22.00.00E+2passed
/16/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:16STAT-REG-ROWS-ADV-TREE-REG-001:16passed
/16/tree_scoretolerance-2-2.00.00E+2passed
/17/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:17STAT-REG-ROWS-ADV-TREE-REG-001:17passed
/17/tree_scoretolerance22.00.00E+2passed
/18/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:18STAT-REG-ROWS-ADV-TREE-REG-001:18passed
/18/tree_scoretolerance-2-2.00.00E+2passed
/19/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:19STAT-REG-ROWS-ADV-TREE-REG-001:19passed
/19/tree_scoretolerance22.00.00E+2passed
/20/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:20STAT-REG-ROWS-ADV-TREE-REG-001:20passed
/20/tree_scoretolerance-2-2.00.00E+2passed
/21/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:21STAT-REG-ROWS-ADV-TREE-REG-001:21passed
/21/tree_scoretolerance22.00.00E+2passed
/22/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:22STAT-REG-ROWS-ADV-TREE-REG-001:22passed
/22/tree_scoretolerance-2-2.00.00E+2passed
/23/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:23STAT-REG-ROWS-ADV-TREE-REG-001:23passed
/23/tree_scoretolerance22.00.00E+2passed
/24/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:24STAT-REG-ROWS-ADV-TREE-REG-001:24passed
/24/tree_scoretolerance-2-2.00.00E+2passed
/25/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:25STAT-REG-ROWS-ADV-TREE-REG-001:25passed
/25/tree_scoretolerance22.00.00E+2passed
/26/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:26STAT-REG-ROWS-ADV-TREE-REG-001:26passed
/26/tree_scoretolerance-2-2.00.00E+2passed
/27/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:27STAT-REG-ROWS-ADV-TREE-REG-001:27passed
/27/tree_scoretolerance22.00.00E+2passed
/28/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:28STAT-REG-ROWS-ADV-TREE-REG-001:28passed
/28/tree_scoretolerance-2-2.00.00E+2passed
/29/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:29STAT-REG-ROWS-ADV-TREE-REG-001:29passed
/29/tree_scoretolerance22.00.00E+2passed
/30/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:30STAT-REG-ROWS-ADV-TREE-REG-001:30passed
/30/tree_scoretolerance-2-2.00.00E+2passed
/31/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:31STAT-REG-ROWS-ADV-TREE-REG-001:31passed
/31/tree_scoretolerance22.00.00E+2passed
/32/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:32STAT-REG-ROWS-ADV-TREE-REG-001:32passed
/32/tree_scoretolerance-2-2.00.00E+2passed
/33/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:33STAT-REG-ROWS-ADV-TREE-REG-001:33passed
/33/tree_scoretolerance22.00.00E+2passed
/34/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:34STAT-REG-ROWS-ADV-TREE-REG-001:34passed
/34/tree_scoretolerance-2-2.00.00E+2passed
/35/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:35STAT-REG-ROWS-ADV-TREE-REG-001:35passed
/35/tree_scoretolerance22.00.00E+2passed
/36/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:36STAT-REG-ROWS-ADV-TREE-REG-001:36passed
/36/tree_scoretolerance-2-2.00.00E+2passed
/37/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:37STAT-REG-ROWS-ADV-TREE-REG-001:37passed
/37/tree_scoretolerance22.00.00E+2passed
/38/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:38STAT-REG-ROWS-ADV-TREE-REG-001:38passed
/38/tree_scoretolerance-2-2.00.00E+2passed
/39/__tabelz_row_idexactSTAT-REG-ROWS-ADV-TREE-REG-001:39STAT-REG-ROWS-ADV-TREE-REG-001:39passed
/39/tree_scoreabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

53  reg_id <- "STAT-REG-ROWS-ADV-TREE-REG-001"

Whole script generators/r/regression_advanced_rows.R


Regression setting STAT-REG-ROWS-OLS-001

Case STAT-REG-ROWS-OLS-001 · status: passed

Reference

Evidence type
R
Source
R regression fits
Citation
R stats::lm/glm/predict.lm, MASS::polr/glm.nb, nnet::multinom and car::Anova help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_options.R · sha256 b22011f9ca86fc99
Dataset
reference_data/datasets/REG_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/responses/0/actual_predicted/xsequence['-0.79564612573140181', '-0.67321627912056847', '0.06657318[-0.7956461257314031, -0.6732162791205702, 0.066573186344934passed
/responses/0/actual_predicted/ysequence['-0.66396923222151238', '-0.27142121217491777', '0.06951530[-0.6639692322215123, -0.27142121217491777, 0.06951530968505passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

36  remaining <- m[-c(3L, 8L), , drop = FALSE]
37  usable <- remaining[stats::complete.cases(remaining[c("y", "x", "z")]), , drop = FALSE]
38  fit <- stats::lm(y ~ x + z, data = usable)
39  emit("STAT-REG-ROWS-OLS-001", "tabelz.analysis.fit_y_by_x:compute_fits",
40       "fitYByXMenuItem", "row_alignment",
41       list(y_columns = list("y"), x_columns = list("x", "z")),
42       list("/responses/0/actual_predicted/x" = sequence(stats::predict(fit),
43                                                          "all fitted means in retained row order"),
44            "/responses/0/actual_predicted/y" = sequence(usable$y,
45                                                          "all observed responses in retained row order")),
46       list("The Actual by Predicted payload has x=fitted and y=observed; all retained rows are below the plot cap."),
47       dataset = "REG_MISSING", excluded = list(2L, 7L))

Whole script generators/r/regression_options.R


Regression II STAT-REG-SELECTION-001

Case STAT-REG-SELECTION-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats
Citation
R nlme::Orthodont, nlme::lme, plm::plm, stats::nls/lm/step documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_core.R · sha256 b7543c65b2778e63
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance0.660185185185184760.66018518518518484.12E-1817passed
/nexact1081080.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

106  selection_data <- d
107  selection_fit <- stats::lm(distance ~ age, data = selection_data)
108  emit("STAT-REG-SELECTION-001", "tabelz.analysis.model_selection:compute_model_selection",
109       "modelSelectionMenuItem", "REG_ORTHODONT",
110       list(response = "distance", candidate_predictors = list("age")),
111       list("/coefficients/1/coefficient" = tol(coef(selection_fit)[["age"]],
112                                                 "single candidate selected slope", "1e-9"),
113            "/n" = exact(nrow(selection_data), "all complete rows")),
114       list("With one strong numeric candidate, R step and best-subset search both select age; refit is R lm."))

Whole script generators/r/regression_advanced_core.R


Regression edge STAT-REG-SELECTION-ALIAS-001

Case STAT-REG-SELECTION-ALIAS-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/rpart
Citation
R nlme::lme, plm::plm, stats::nls/lm/step, rpart::rpart documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_a.R · sha256 b8ea61b61080b6f8
Dataset
reference_data/datasets/REG_EDGE_SELECTION_ALIAS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 92  sel <- read_dataset("REG_MIN_SELECTION")
 93  sel$z <- sel$x
 94  export_dataset(sel,"REG_EDGE_SELECTION_ALIAS","Two exactly aliased candidate predictors",
 95                 "R stats lm/step AIC alias boundary",
 96                 "Constructed validation data; no third-party license.",
 97                 c(x="Float64",y="Float64",z="Float64"))
 98  s <- read_dataset("REG_EDGE_SELECTION_ALIAS")
 99  fit <- stats::step(stats::lm(y~1,data=s),scope=list(lower=~1,upper=~x+z),
100                     direction="forward",trace=0)
101  emit("STAT-REG-SELECTION-ALIAS-001",
102       "tabelz.analysis.model_selection:compute_model_selection",
103       "modelSelectionMenuItem","REG_EDGE_SELECTION_ALIAS",
104       list(response="y",candidate_predictors=list("x","z"),direction="forward"),
105       list("/n"=exact(as.integer(nrow(s)),"five source rows")),
106       c("x and z are exactly aliased; R step breaks the tie by candidate order.",
107         "Only population size is asserted until Tabelz and R alias-selection conventions are aligned."),
108       deferred=list(list(path="/selected",
109         reason="Exact-alias tie breaking is not specified in the manual.")))

Whole script generators/r/regression_advanced_special_a.R


Small-sample boundary STAT-REG-SELECTION-MIN-001

Case STAT-REG-SELECTION-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats, plm, rpart
Citation
R stats::nls/lm, plm::plm, rpart::rpart documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_a.R · sha256 e69dea823e2a8f59
Dataset
reference_data/datasets/REG_MIN_SELECTION.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/coefficienttolerance1.991.99000000000000022.13E-1615passed
/nexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

55  sel <- data.frame(x=c(1,2,3,4,5),y=c(2.1,4.0,6.2,8.1,10.0))
56  export_dataset(sel,"REG_MIN_SELECTION","Five-row one-predictor selection design",
57                 "R lm and step AIC small-sample boundary",
58                 "Constructed validation data; no third-party license.",
59                 c(x="Float64",y="Float64"))
60  s <- read_dataset("REG_MIN_SELECTION")
61  fit <- stats::lm(y~x,data=s)
62  emit("STAT-REG-SELECTION-MIN-001","tabelz.analysis.model_selection:compute_model_selection",
63       "modelSelectionMenuItem","REG_MIN_SELECTION",
64       list(response="y",candidate_predictors=list("x")),
65       list("/coefficients/1/coefficient"=tol(coef(fit)[["x"]],"single selected OLS slope"),
66            "/n"=exact(nrow(s),"five selected rows")),
67       c("One strong candidate gives the unique one-term AIC choice."))

Whole script generators/r/regression_advanced_minimum_a.R


True minimum sample STAT-REG-SELECTION-TRUE-MIN-001

Case STAT-REG-SELECTION-TRUE-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats/rpart/pls and analytic sample conditions
Citation
R stats::lm, rpart::rpart, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_true_min.R · sha256 f10363ad1ec76730
Dataset
reference_data/datasets/REG_TRUE_MIN_SELECTION.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact330.00E+2passed
/coefficients/1/coefficienttolerance2.05000000000000072.058.78E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

25  sel <- read_dataset("REG_MIN_SELECTION")[1:3,]
26  export_dataset(sel,"REG_TRUE_MIN_SELECTION","Three rows for one candidate predictor",
27                 "R lm/step AIC minimal one-predictor model",
28                 "Constructed validation data; no third-party license.",
29                 c(x="Float64",y="Float64"))
30  s <- read_dataset("REG_TRUE_MIN_SELECTION")
31  fit <- stats::lm(y~x,data=s)
32  emit("STAT-REG-SELECTION-TRUE-MIN-001","STAT-REG-SELECTION-001",
33       "REG_TRUE_MIN_SELECTION",list(response="y",candidate_predictors=list("x")),
34       list("/n"=exact(as.integer(nrow(s)),"three rows exceed one predictor plus intercept"),
35            "/coefficients/1/coefficient"=tol(coef(fit)[["x"]],"three-point OLS slope")),
36       "Three complete rows are the smallest accepted for one candidate predictor.")

Whole script generators/r/regression_advanced_true_min.R


Large-offset random-slope mixed model

Case STAT-REG-STRUCT-HARD-001 · status: passed

Reference

Evidence type
R
Source
R nlme::lme
Citation
R nlme::lme REML random-slope documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · nlme 3.1.169, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_structure_hard.R · sha256 b4579cd3289ab12d
Dataset
reference_data/datasets/REG_ADV_HARD_ORTH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.660185185185220290.66018518518518523.51E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

1  source("common.R",local=TRUE)
2  d <- read_dataset("REG_ADV_HARD_ORTH")
3  fit <- nlme::lme(distance~age,random=~age|subject,data=d,method="REML",
4                   control=nlme::lmeControl(opt="optim"))
5  id <- "STAT-REG-STRUCT-HARD-001"

Whole script generators/r/regression_advanced_structure_hard.R


Small-sample boundary STAT-REG-STRUCT-MIN-001

Case STAT-REG-STRUCT-MIN-001 · status: passed

Reference

Evidence type
R
Source
R nlme/MASS/pls/stats
Citation
R nlme::lme, MASS::lda, pls::plsr and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_b.R · sha256 e3e09e1965800ae1
Dataset
reference_data/datasets/REG_MIN_SLOPE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance0.562500000000000220.56252.20E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

41  six <- orth[c(1:3,5:7),]
42  export_dataset(six,"REG_MIN_SLOPE","Two repeated subjects with three observations each",
43                 "R nlme random-slope minimum-group-size boundary",
44                 "Derived from R nlme::Orthodont; retain attribution.",types)
45  m <- read_dataset("REG_MIN_SLOPE")
46  fit <- nlme::lme(distance~age,random=~age|subject,data=m,method="REML",
47                   control=nlme::lmeControl(opt="optim"))
48  emit("STAT-REG-STRUCT-MIN-001","tabelz.executor.mixed_model_structures",
49       "mixedModelStructuresMenuItem","REG_MIN_SLOPE",
50       list(response_column="distance",fixed_numeric_columns=list("age"),
51            model_mode="random_slope",group_column="subject",random_slope_column="age",
52            estimation="reml"),
53       list("/coefficients/1/estimate"=tol(nlme::fixef(fit)[["age"]],"random-slope fixed age effect")),
54       c("Two groups each have three complete rows, the declared random-slope minimum."))

Whole script generators/r/regression_advanced_minimum_b.R


Regression edge STAT-REG-STRUCT-ONEGROUP-001

Case STAT-REG-STRUCT-ONEGROUP-001 · status: passed

Reference

Evidence type
Analytical
Source
Analytic design condition
Citation
Tabelz manual regression model design conditions; R stats/plm references
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_b.R · sha256 4a5d01b584989ad6
Dataset
reference_data/datasets/REG_EDGE_STRUCT_ONE_GROUP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

24  err <- function(cls,message) list(error_class=cls,message_contains=message)
25  exact <- function(x,note) list(comparison="exact",value=as.integer(x),rationale=note)
26  tol <- function(x,note) list(comparison="tolerance",value=dec(unname(x)),
27                               absolute_tolerance="0",relative_tolerance="1e-9",rationale=note)
28  orth <- read_dataset("REG_ORTHODONT")[1:4,]
29  export_dataset(orth,"REG_EDGE_STRUCT_ONE_GROUP","One repeated group with four observations",
30                 "Analytic random-slope group-count design condition",
31                 "Derived from R nlme::Orthodont; retain attribution.",
32                 c(distance="Float64",age="Float64",subject="String",sex="String",
33                   time="Int64",age_sq="Float64"))
34  emit("STAT-REG-STRUCT-ONEGROUP-001","STAT-REG-MIXED-STRUCT-SLOPE-001",
35       "REG_EDGE_STRUCT_ONE_GROUP",
36       list(response_column="distance",fixed_numeric_columns=list("age"),
37            model_mode="random_slope",group_column="subject",random_slope_column="age",
38            estimation="reml"),error=err("RuntimeError","at least two groups"),
39       note="A random-slope variance cannot be identified from one repeated group.")

Whole script generators/r/regression_advanced_special_b.R


Survey Cox regression (Taylor)

Case STAT-REG-SURVEY-COX-001 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance0.429667407055990440.42966740705599127.73E-1614passed
/fit/coefficients/0/standard_errortolerance0.57531359971158080.57531359971158062.19E-1615passed
/fit/coefficients/0/statistictolerance0.746840344590139260.74684034459014091.67E-1514passed
/fit/coefficients/0/p_valuetail_probability0.455159939183731540.455159939183730549.96E-1614passed
/fit/coefficients/0/lowertolerance-0.69792652819480083-0.69792652819479998.86E-1614passed
/fit/coefficients/0/uppertolerance1.55726134230678181.55726134230678234.59E-1615passed
/fit/coefficients/0/ratiotolerance1.53674632755762561.5367463275576271.37E-1515passed
/fit/coefficients/0/ratio_lowertolerance0.497616027636832160.49761602763683264.43E-1615passed
/fit/coefficients/0/ratio_uppertolerance4.74580629260875324.7458062926087562.66E-1515passed
/fit/coefficients/1/estimatetolerance0.891120989758563690.89112098975856479.96E-1614passed
/fit/coefficients/1/standard_errortolerance0.532717927205486160.53271792720548583.29E-1615passed
/fit/coefficients/1/statistictolerance1.67278205641243631.67278205641243922.86E-1514passed
/fit/coefficients/1/p_valuetail_probability0.094370208519162530.09437020851916196.25E-1614passed
/fit/coefficients/1/lowertolerance-0.15298696148301927-0.152986961483017821.45E-1514passed
/fit/coefficients/1/uppertolerance1.93522894100014661.9352289410001474.85E-1615passed
/fit/coefficients/1/ratiotolerance2.43786093727553872.43786093727554132.62E-1514passed
/fit/coefficients/1/ratio_lowertolerance0.85814091062328590.85814091062328711.22E-1514passed
/fit/coefficients/1/ratio_uppertolerance6.92562943442158126.9256294344215853.54E-1515passed
/fit/coefficients/2/estimatetolerance-0.056981380091591177-0.0569813800915912052.80E-1715passed
/fit/coefficients/2/standard_errortolerance0.285647836623884820.28564783662388489.95E-1917passed
/fit/coefficients/2/statistictolerance-0.19948122403118038-0.199481224031180467.96E-1715passed
/fit/coefficients/2/p_valuetail_probability0.841886329257302780.84188632925730289.94E-1917passed
/fit/coefficients/2/lowertolerance-0.61684085213618667-0.61684085213618681.09E-1615passed
/fit/coefficients/2/uppertolerance0.502878091953004370.50287809195300451.10E-1615passed
/fit/coefficients/2/ratiotolerance0.94461165778532530.94461165778532531.76E-1817passed
/fit/coefficients/2/ratio_lowertolerance0.539646570845225380.53964657084522541.77E-1817passed
/fit/coefficients/2/ratio_uppertolerance1.65347327719767231.65347327719767234.20E-1716passed
/fit/term_tests/[term=grp]/dfexact220.00E+2passed
/fit/term_tests/[term=grp]/statistictolerance2.86167791976413532.86167791976414551.02E-1414passed
/fit/term_tests/[term=grp]/p_valuetail_probability0.239108235854257170.239108235854255951.22E-1514passed
/fit/term_tests/[term=x]/dfexact110.00E+2passed
/fit/term_tests/[term=x]/statistictolerance0.0397927587409779710.039792758740978014.16E-1714passed
/fit/term_tests/[term=x]/p_valuetail_probability0.841886329257302890.84188632925730272.22E-1615passed
/fit/term_tests/[term=All predictors]/dfexact330.00E+2passed
/fit/term_tests/[term=All predictors]/statistictolerance3.06445802224064153.0644580222406551.37E-1414passed
/fit/term_tests/[term=All predictors]/p_valuetail_probability0.381793394478414040.381793394478410653.39E-1514passed
/eventsexact23230.00E+2passed
/fit/statistics/[label=Weighted events]/valuetolerance51.30149999999999751.301500000000011.44E-1415passed
/fit/rows_usedexact30300.00E+2passed
/fit/rankexact330.00E+2passed
/fit/design_dfexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

422  COX <- "tabelz.analysis.survey_models:compute_survey_cox"
423  COX_REQ <- "REQ-STAT-REG-SURVEY-COX-001"

Whole script generators/r/regression_survey_models.R


Survey Cox regression from supplied replicate weights with 90% intervals

Case STAT-REG-SURVEY-COX-002 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance0.429667407055990440.42966740705599127.73E-1614passed
/fit/coefficients/0/standard_errortolerance0.154601162141556840.154601162168106462.65E-119passed
/fit/coefficients/0/statistictolerance2.77919907654106662.77919907606380034.77E-109passed
/fit/coefficients/0/p_valuetail_probability0.00544931174157618860.00544931174958313358.01E-128passed
/fit/coefficients/0/lowertolerance0.175371124776538160.17537112473286864.37E-119passed
/fit/coefficients/0/uppertolerance0.683963689335442710.68396368937911394.37E-1110passed
/fit/coefficients/0/ratiotolerance1.53674632755762561.5367463275576271.37E-1515passed
/fit/coefficients/0/ratio_lowertolerance1.19168839964576481.19168839959372425.20E-1110passed
/fit/coefficients/0/ratio_uppertolerance1.98171709648582911.98171709657237328.65E-1110passed
/fit/coefficients/1/estimatetolerance0.891120989758564130.89112098975856475.56E-1615passed
/fit/coefficients/1/standard_errortolerance0.144944386103955990.144944386103165657.90E-1311passed
/fit/coefficients/1/statistictolerance6.14802003521157846.1480200352451053.35E-1111passed
/fit/coefficients/1/p_valuetail_probability7.8456125557156509e-107.845612554057671e-101.66E-199passed
/fit/coefficients/1/lowertolerance0.652708690569217650.65270869057051821.30E-1211passed
/fit/coefficients/1/uppertolerance1.12953328894791061.12953328894661121.30E-1211passed
/fit/coefficients/1/ratiotolerance2.437860937275542.43786093727554131.32E-1515passed
/fit/coefficients/1/ratio_lowertolerance1.9207364699074391.9207364699099372.50E-1211passed
/fit/coefficients/1/ratio_uppertolerance3.09421206011690763.0942120601128874.02E-1211passed
/fit/coefficients/2/estimatetolerance-0.056981380091591198-0.0569813800915912056.98E-1815passed
/fit/coefficients/2/standard_errortolerance0.117906410942929950.117906410924807361.81E-119passed
/fit/coefficients/2/statistictolerance-0.48327635143751263-0.483276351511793777.43E-119passed
/fit/coefficients/2/p_valuetail_probability0.628899529112960210.62889952906022475.27E-1110passed
/fit/coefficients/2/lowertolerance-0.25092016777190018-0.250920167742091252.98E-119passed
/fit/coefficients/2/uppertolerance0.13695740758871780.136957407558908872.98E-119passed
/fit/coefficients/2/ratiotolerance0.94461165778532530.94461165778532531.76E-1817passed
/fit/coefficients/2/ratio_lowertolerance0.778084485297696470.77808448532089042.32E-1110passed
/fit/coefficients/2/ratio_uppertolerance1.14677930338444981.14677930335026563.42E-1110passed
/fit/term_tests/[term=grp]/dfexact220.00E+2passed
/fit/term_tests/[term=grp]/statistictolerance39.14355499211859539.1435549915937245.25E-1010passed
/fit/term_tests/[term=grp]/p_valuetail_probability3.1628968770117634e-093.1628968778418213e-098.30E-199passed
/fit/term_tests/[term=x]/dfexact110.00E+2passed
/fit/term_tests/[term=x]/statistictolerance0.23355603185875420.233556031930550887.18E-119passed
/fit/term_tests/[term=x]/p_valuetail_probability0.628899529112959990.62889952906022475.27E-1110passed
/fit/term_tests/[term=All predictors]/dfexact330.00E+2passed
/fit/term_tests/[term=All predictors]/statistictolerance41.73027125432819441.730271249338364.99E-99passed
/fit/term_tests/[term=All predictors]/p_valuetail_probability4.5772952643513363e-094.577295275509918e-091.12E-178passed
/eventsexact23230.00E+2passed
/fit/statistics/[label=Weighted events]/valuetolerance51.30149999999999751.301500000000011.44E-1415passed
/fit/rows_usedexact30300.00E+2passed
/fit/rankexact330.00E+2passed
/fit/design_dfexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

448  cox_case("STAT-REG-SURVEY-COX-002", "Survey Cox regression from supplied replicate weights with 90% intervals",
449           replicate, design_rep, D_R, 0.90,
450           c("svycoxph on svrepdesign(type = 'other', scale = 0.75, mse = TRUE): the model is refitted under each replicate weight column.",
451             "degf = qr(replicate weights)$rank - 1 = 3 (reference only; normal inference)."), "axis_coverage")

Whole script generators/r/regression_survey_models.R


Survey linear regression with design-df t and F inference

Case STAT-REG-SURVEY-LINEAR-001 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance4.75210749103756584.7521074910375678.89E-1615passed
/fit/coefficients/0/standard_errortolerance0.371825429288772770.37182542928877243.90E-1614passed
/fit/coefficients/0/statistictolerance12.78048007670532712.7804800767053431.63E-1414passed
/fit/coefficients/0/p_valuetail_probability5.2178844394329544e-055.217884439432926e-052.85E-1914passed
/fit/coefficients/0/lowertolerance3.79629979648017193.79629979648017371.78E-1515passed
/fit/coefficients/0/uppertolerance5.70791518559495975.707915185594965.02E-1818passed
/fit/coefficients/0/dfexact550.00E+2passed
/fit/coefficients/1/estimatetolerance0.977182234311349760.97718223431135078.90E-1615passed
/fit/coefficients/1/standard_errortolerance0.483648848153509780.483648848153509344.44E-1615passed
/fit/coefficients/1/statistictolerance2.02043742695153262.0204374269515363.56E-1514passed
/fit/coefficients/1/p_valuetail_probability0.0993148579334519190.099314857933451533.88E-1614passed
/fit/coefficients/1/lowertolerance-0.26607670957848806-0.26607670957848651.56E-1514passed
/fit/coefficients/1/uppertolerance2.22044117820118772.22044117820118776.98E-1817passed
/fit/coefficients/1/dfexact550.00E+2passed
/fit/coefficients/2/estimatetolerance-0.25080171609245849-0.250801716092458555.87E-1715passed
/fit/coefficients/2/standard_errortolerance0.439873787254217160.439873787254216664.96E-1614passed
/fit/coefficients/2/statistictolerance-0.57016745111822753-0.57016745111822837.78E-1614passed
/fit/coefficients/2/p_valuetail_probability0.593228281639430420.59322828163942967.76E-1614passed
/fit/coefficients/2/lowertolerance-1.3815332835807015-1.38153328358070041.12E-1515passed
/fit/coefficients/2/uppertolerance0.87992985139578450.87992985139578341.11E-1514passed
/fit/coefficients/2/dfexact550.00E+2passed
/fit/coefficients/3/estimatetolerance1.17073874095277391.17073874095277142.48E-1514passed
/fit/coefficients/3/standard_errortolerance0.154886113808293980.154886113808294343.60E-1614passed
/fit/coefficients/3/statistictolerance7.55870692450727697.5587069245072433.38E-1414passed
/fit/coefficients/3/p_valuetail_probability0.000642521331040617510.0006425213310406311.34E-1713passed
/fit/coefficients/3/lowertolerance0.772591310204874480.7725913102048713.44E-1514passed
/fit/coefficients/3/uppertolerance1.56888617170067321.5688861717006721.29E-1515passed
/fit/coefficients/3/dfexact550.00E+2passed
/fit/term_tests/[term=grp]/statistictolerance2.94290971404652212.9429097140465252.69E-1515passed
/fit/term_tests/[term=grp]/p_valuetail_probability0.142978980396142990.142978980396142762.27E-1614passed
/fit/term_tests/[term=grp]/denominator_dfexact550.00E+2passed
/fit/term_tests/[term=x]/statistictolerance57.1340503705942657.134050370593764.98E-1314passed
/fit/term_tests/[term=x]/p_valuetail_probability0.000642521331040617510.00064252133104063051.30E-1713passed
/fit/term_tests/[term=All predictors]/statistictolerance50.19836280061457550.198362800614035.47E-1313passed
/fit/term_tests/[term=All predictors]/p_valuetail_probability0.000372752355326178110.00037275235532618769.48E-1813passed
/fit/statistics/[label=Weighted R-squared]/valuetolerance0.567362755195379710.56736275519537971.21E-1817passed
/fit/statistics/[label=Sum of weights (analysed rows)]/valuetolerance65.84120000000000165.84123.86E-1617passed
/fit/rows_usedexact30300.00E+2passed
/fit/rankexact440.00E+2passed
/fit/design_dfexact880.00E+2passed
/fit/test_dfexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

167  LINEAR <- "tabelz.analysis.survey_models:compute_survey_linear"
168  LIN_REQ <- "REQ-STAT-REG-SURVEY-LINEAR-001"

Whole script generators/r/regression_survey_models.R


Survey linear regression with normal inference

Case STAT-REG-SURVEY-LINEAR-002 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance4.75210749103756584.7521074910375678.89E-1615passed
/fit/coefficients/0/standard_errortolerance0.371825429288772770.37182542928877243.90E-1614passed
/fit/coefficients/0/statistictolerance12.78048007670532712.7804800767053431.63E-1414passed
/fit/coefficients/0/p_valuetail_probability2.1076453902737243e-372.107645390273297e-374.27E-5012passed
/fit/coefficients/0/lowertolerance4.02334304109542634.0233430410954281.76E-1515passed
/fit/coefficients/0/uppertolerance5.48087194097970535.4808719409797051.46E-1717passed
/fit/coefficients/1/estimatetolerance0.977182234311349760.97718223431135078.90E-1615passed
/fit/coefficients/1/standard_errortolerance0.483648848153509780.483648848153509344.44E-1615passed
/fit/coefficients/1/statistictolerance2.02043742695153262.0204374269515363.56E-1514passed
/fit/coefficients/1/p_valuetail_probability0.0433380345242187980.043338034524218443.61E-1614passed
/fit/coefficients/1/lowertolerance0.0292479107661893780.0292479107661909321.55E-1513passed
/fit/coefficients/1/uppertolerance1.92511655785651011.92511655785651042.69E-1615passed
/fit/coefficients/2/estimatetolerance-0.25080171609245849-0.250801716092458555.87E-1715passed
/fit/coefficients/2/standard_errortolerance0.439873787254217160.439873787254216664.96E-1614passed
/fit/coefficients/2/statistictolerance-0.57016745111822753-0.57016745111822837.78E-1614passed
/fit/coefficients/2/p_valuetail_probability0.568564130022210510.568564130022215.53E-1615passed
/fit/coefficients/2/lowertolerance-1.1129384968539577-1.1129384968539576.29E-1615passed
/fit/coefficients/2/uppertolerance0.611335064669040860.61133506466904017.75E-1614passed
/fit/coefficients/3/estimatetolerance1.17073874095277391.17073874095277142.48E-1514passed
/fit/coefficients/3/standard_errortolerance0.154886113808293980.154886113808294343.60E-1614passed
/fit/coefficients/3/statistictolerance7.55870692450727697.5587069245072433.38E-1414passed
/fit/coefficients/3/p_valuetail_probability4.0709599079711517e-144.07095990797219e-141.04E-2612passed
/fit/coefficients/3/lowertolerance0.867167536183145790.86716753618314263.22E-1514passed
/fit/coefficients/3/uppertolerance1.4743099457224021.47430994572240031.74E-1514passed
/fit/term_tests/[term=grp]/dfexact220.00E+2passed
/fit/term_tests/[term=grp]/statistictolerance5.88581942809304435.885819428093055.29E-1515passed
/fit/term_tests/[term=grp]/p_valuetail_probability0.0527121281601115530.052712128160111431.25E-1614passed
/fit/term_tests/[term=x]/dfexact110.00E+2passed
/fit/term_tests/[term=x]/statistictolerance57.1340503705942657.134050370593764.98E-1314passed
/fit/term_tests/[term=x]/p_valuetail_probability4.0709599079711429e-144.070959907972173e-141.03E-2612passed
/fit/term_tests/[term=All predictors]/statistictolerance150.59508840184373150.595088401842081.65E-1213passed
/fit/term_tests/[term=All predictors]/p_valuetail_probability1.9606188963164449e-321.960618896318026e-321.58E-4412passed
/fit/statistics/[label=Weighted R-squared]/valuetolerance0.567362755195379710.56736275519537971.21E-1817passed
/fit/statistics/[label=Sum of weights (analysed rows)]/valuetolerance65.84120000000000165.84123.86E-1617passed
/fit/rows_usedexact30300.00E+2passed
/fit/rankexact440.00E+2passed
/fit/design_dfexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

229  linear_case("STAT-REG-SURVEY-LINEAR-002", "Survey linear regression with normal inference",
230              taylor, design_taylor, D_T, "normal", 0.95, list(), y ~ grp + x,
231              c("Normal inference: z = estimate / SE, 2 * pnorm(-|z|), qnorm intervals; term tests are Wald chi-square(q) (regTermTest df = Inf)."),
232              "axis_coverage")

Whole script generators/r/regression_survey_models.R


Survey linear regression with an interaction and 90% intervals

Case STAT-REG-SURVEY-LINEAR-003 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance4.70421931678230724.7042193167823113.56E-1515passed
/fit/coefficients/0/standard_errortolerance0.315950236269715520.315950236269715752.25E-1615passed
/fit/coefficients/0/statistictolerance14.88911472997438114.889114729974383.20E-1616passed
/fit/coefficients/0/p_valuetail_probability0.000657439998051100320.00065743999805110034.27E-2218passed
/fit/coefficients/0/lowertolerance3.96067358352816173.9606735835281653.09E-1515passed
/fit/coefficients/0/uppertolerance5.44776505003645275.4477650500364563.58E-1515passed
/fit/coefficients/0/dfexact330.00E+2passed
/fit/coefficients/1/estimatetolerance1.33382631351532851.3338263135153271.51E-1514passed
/fit/coefficients/1/standard_errortolerance0.477113979253593280.47711397925359331.13E-1817passed
/fit/coefficients/1/statistictolerance2.79561356722767452.7956135672276713.54E-1514passed
/fit/coefficients/1/p_valuetail_probability0.0681004315147124840.068100431514712661.80E-1614passed
/fit/coefficients/1/lowertolerance0.211003720507125790.21100372050712498.87E-1614passed
/fit/coefficients/1/uppertolerance2.45664890652353132.4566489065235292.23E-1515passed
/fit/coefficients/1/dfexact330.00E+2passed
/fit/coefficients/2/estimatetolerance-0.15506232265492842-0.155062322654929036.10E-1614passed
/fit/coefficients/2/standard_errortolerance0.410609147608492430.41060914760849231.10E-1615passed
/fit/coefficients/2/statistictolerance-0.3776397178632202-0.37763971786322181.61E-1514passed
/fit/coefficients/2/p_valuetail_probability0.73083268968121440.73083268968121331.11E-1514passed
/fit/coefficients/2/lowertolerance-1.1213748766318996-1.12137487663189942.13E-1615passed
/fit/coefficients/2/uppertolerance0.811250231322042660.81125023132204131.33E-1514passed
/fit/coefficients/2/dfexact330.00E+2passed
/fit/coefficients/3/estimatetolerance0.82724615878415930.827246158784166.68E-1615passed
/fit/coefficients/3/standard_errortolerance0.179456687419075860.179456687419075721.38E-1615passed
/fit/coefficients/3/statistictolerance4.60972600509634054.6097260050963487.06E-1514passed
/fit/coefficients/3/p_valuetail_probability0.0192026138903406210.0192026138903405388.33E-1714passed
/fit/coefficients/3/lowertolerance0.404919352481445560.40491935248144681.22E-1514passed
/fit/coefficients/3/uppertolerance1.2495729650868731.24957296508687322.11E-1615passed
/fit/coefficients/3/dfexact330.00E+2passed
/fit/coefficients/4/estimatetolerance1.15767973710291111.15767973710290822.93E-1514passed
/fit/coefficients/4/standard_errortolerance0.643458899153745540.64345889915374554.91E-1817passed
/fit/coefficients/4/statistictolerance1.79915102367136531.79915102367136084.48E-1514passed
/fit/coefficients/4/p_valuetail_probability0.169824250435189340.169824250435190127.76E-1614passed
/fit/coefficients/4/lowertolerance-0.35661290796334844-0.356612907963350431.99E-1514passed
/fit/coefficients/4/uppertolerance2.67197238216917032.67197238216916683.53E-1514passed
/fit/coefficients/4/dfexact330.00E+2passed
/fit/coefficients/5/estimatetolerance0.594107281838006980.59410728183800482.22E-1514passed
/fit/coefficients/5/standard_errortolerance0.316282962290248490.316282962290247668.30E-1614passed
/fit/coefficients/5/statistictolerance1.87840431724805651.87840431724805422.27E-1514passed
/fit/coefficients/5/p_valuetail_probability0.156942555778674360.15694255577867473.32E-1614passed
/fit/coefficients/5/lowertolerance-0.15022147666666819-0.150221476666668071.15E-1615passed
/fit/coefficients/5/uppertolerance1.33843604034268231.33843604034267764.70E-1514passed
/fit/coefficients/5/dfexact330.00E+2passed
/fit/term_tests/[term=grp]/statistictolerance5.14149425718221755.1414942571822116.19E-1514passed
/fit/term_tests/[term=grp]/p_valuetail_probability0.107334215500319920.107334215500320071.54E-1614passed
/fit/term_tests/[term=grp]/denominator_dfexact330.00E+2passed
/fit/term_tests/[term=x]/statistictolerance21.24957384206146921.2495738420615376.78E-1414passed
/fit/term_tests/[term=x]/p_valuetail_probability0.0192026138903406140.019202613890340538.32E-1714passed
/fit/term_tests/[term=grp:x]/statistictolerance2.50658548520388272.50658548520387567.11E-1514passed
/fit/term_tests/[term=grp:x]/p_valuetail_probability0.229073720009705380.2290737200097066.14E-1614passed
/fit/statistics/[label=Weighted R-squared]/valuetolerance0.628942227021475290.62894222702147533.64E-1817passed
/fit/statistics/[label=Sum of weights (analysed rows)]/valuetolerance65.84120000000000165.84123.86E-1617passed
/fit/rows_usedexact30300.00E+2passed
/fit/rankexact660.00E+2passed
/fit/design_dfexact880.00E+2passed
/fit/test_dfexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

233  domain_a <- subset(taylor, dom == "A")
234  linear_case("STAT-REG-SURVEY-LINEAR-003", "Survey linear regression with an interaction and 90% intervals",
235              taylor, design_taylor, D_T, "design_df", 0.90,
236              list(interactions = list("grp:x")), y ~ grp + x + grp:x,
237              c("svyglm(y ~ grp + x + grp:x): interaction columns grp = b (vs a):x and grp = c (vs a):x follow the main effects (indices 4 and 5).",
238                "Test df = D - rank(X) + 1 = 8 - 6 + 1 = 3; 90% intervals use qt(0.95, 3)."),
239              "axis_coverage")

Whole script generators/r/regression_survey_models.R


Survey logistic regression with design-df inference

Case STAT-REG-SURVEY-LOGIT-001 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance-1.0714957947790638-1.07149579477906261.15E-1514passed
/fit/coefficients/0/standard_errortolerance0.770739098787651230.77073909878771766.64E-1413passed
/fit/coefficients/0/statistictolerance-1.3902185531582523-1.3902185531581311.21E-1313passed
/fit/coefficients/0/p_valuetail_probability0.223173451593168150.22317345159320283.46E-1412passed
/fit/coefficients/0/lowertolerance-3.0527437221373024-3.0527437221374731.71E-1313passed
/fit/coefficients/0/uppertolerance0.909752132579174870.90975213257934741.73E-1312passed
/fit/coefficients/0/dfexact550.00E+2passed
/fit/coefficients/1/estimatetolerance3.08138293989121343.08138293989121259.20E-1615passed
/fit/coefficients/1/standard_errortolerance1.28423184770729491.28423184770740621.11E-1313passed
/fit/coefficients/1/statistictolerance2.3993976986260892.39939769862588032.09E-1313passed
/fit/coefficients/1/p_valuetail_probability0.0616666925980688120.061666692598084731.59E-1412passed
/fit/coefficients/1/lowertolerance-0.21984012057081959-0.21984012057110782.88E-1311passed
/fit/coefficients/1/uppertolerance6.38260600035324636.3826060003535332.87E-1313passed
/fit/coefficients/1/dfexact550.00E+2passed
/fit/coefficients/1/ratiotolerance21.78851377515436421.7885137751543432.15E-1415passed
/fit/coefficients/1/ratio_lowertolerance0.802647114467114680.80264711446688342.31E-1312passed
/fit/coefficients/1/ratio_uppertolerance591.46706438392357591.46706438409331.70E-1012passed
/fit/coefficients/2/estimatetolerance-0.14917538856652648-0.149175388566526873.93E-1614passed
/fit/coefficients/2/standard_errortolerance1.14942597837132411.14942597837139547.13E-1413passed
/fit/coefficients/2/statistictolerance-0.12978251002983257-0.129782510029824857.72E-1513passed
/fit/coefficients/2/p_valuetail_probability0.901797972335839360.90179797233584525.88E-1514passed
/fit/coefficients/2/lowertolerance-3.1038689299763509-3.1038689299765351.84E-1313passed
/fit/coefficients/2/uppertolerance2.80551815284329782.8055181528434811.83E-1313passed
/fit/coefficients/2/dfexact550.00E+2passed
/fit/coefficients/2/ratiotolerance0.861418018777586060.86141801877758573.34E-1615passed
/fit/coefficients/2/ratio_lowertolerance0.0448752469117850580.044875246911776798.27E-1512passed
/fit/coefficients/2/ratio_uppertolerance16.5356416764322716.5356416764353043.03E-1212passed
/fit/coefficients/3/estimatetolerance2.41179128105566992.41179128105567088.55E-1615passed
/fit/coefficients/3/standard_errortolerance0.48887545959322520.48887545959326133.61E-1413passed
/fit/coefficients/3/statistictolerance4.93334495264382964.9333449526434673.62E-1313passed
/fit/coefficients/3/p_valuetail_probability0.00434700917349081350.0043470091734921751.36E-1512passed
/fit/coefficients/3/lowertolerance1.15509690473697061.15509690473687829.24E-1413passed
/fit/coefficients/3/uppertolerance3.66848565737436923.66848565737446339.41E-1413passed
/fit/coefficients/3/dfexact550.00E+2passed
/fit/coefficients/3/ratiotolerance11.15392307211226911.153923072112281.05E-1415passed
/fit/coefficients/3/ratio_lowertolerance3.17433101033537353.17433101033508042.93E-1313passed
/fit/coefficients/3/ratio_uppertolerance39.19251001030742339.192510010311113.69E-1213passed
/fit/term_tests/[term=grp]/statistictolerance3.8627106957874563.8627106957871023.54E-1313passed
/fit/term_tests/[term=grp]/p_valuetail_probability0.0967708552885043340.096770855288517751.34E-1412passed
/fit/term_tests/[term=grp]/denominator_dfexact550.00E+2passed
/fit/statistics/[label=Event rows]/valuetolerance1414.00.00E+2passed
/fit/rows_usedexact30300.00E+2passed
/fit/rankexact440.00E+2passed
/fit/design_dfexact880.00E+2passed
/fit/test_dfexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

275  logistic_case("STAT-REG-SURVEY-LOGIT-001", "Survey logistic regression with design-df inference",
276                taylor, design_taylor, D_T, "design_df", 0.95,
277                c("Coefficient t with D - rank(X) + 1 = 8 - 4 + 1 = 5 df; term F(2, 5)."), "nominal")

Whole script generators/r/regression_survey_models.R


Survey logistic regression with normal inference

Case STAT-REG-SURVEY-LOGIT-002 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance-1.0714957947790638-1.07149579477906261.15E-1514passed
/fit/coefficients/0/standard_errortolerance0.770739098787651230.77073909878771766.64E-1413passed
/fit/coefficients/0/statistictolerance-1.3902185531582523-1.3902185531581311.21E-1313passed
/fit/coefficients/0/p_valuetail_probability0.164462521233593890.16446252123363073.68E-1412passed
/fit/coefficients/0/lowertolerance-2.5821166698797189-2.5821166698798481.29E-1313passed
/fit/coefficients/0/uppertolerance0.439125080321591140.43912508032172281.32E-1312passed
/fit/coefficients/1/estimatetolerance3.08138293989121343.08138293989121259.20E-1615passed
/fit/coefficients/1/standard_errortolerance1.28423184770729491.28423184770740621.11E-1313passed
/fit/coefficients/1/statistictolerance2.3993976986260892.39939769862588032.09E-1313passed
/fit/coefficients/1/p_valuetail_probability0.0164220678672757570.0164220678672851149.36E-1512passed
/fit/coefficients/1/lowertolerance0.564334770585588340.5643347705853692.19E-1312passed
/fit/coefficients/1/uppertolerance5.5984311091968385.5984311091970562.18E-1313passed
/fit/coefficients/1/ratiotolerance21.78851377515436421.7885137751543432.15E-1415passed
/fit/coefficients/1/ratio_lowertolerance1.75827773552168251.75827773552129683.86E-1312passed
/fit/coefficients/1/ratio_uppertolerance270.00247056488814270.00247056494715.89E-1112passed
/fit/coefficients/2/estimatetolerance-0.14917538856652648-0.149175388566526873.93E-1614passed
/fit/coefficients/2/standard_errortolerance1.14942597837132411.14942597837139547.13E-1413passed
/fit/coefficients/2/statistictolerance-0.12978251002983257-0.129782510029824857.72E-1513passed
/fit/coefficients/2/p_valuetail_probability0.896738500844849180.89673850084485536.11E-1514passed
/fit/coefficients/2/lowertolerance-2.4020089090690364-2.4020089090691771.41E-1313passed
/fit/coefficients/2/uppertolerance2.10365813193598332.1036581319361231.40E-1313passed
/fit/coefficients/2/ratiotolerance0.861418018777586060.86141801877758573.34E-1615passed
/fit/coefficients/2/ratio_lowertolerance0.0905358921037374140.090535892103724671.27E-1412passed
/fit/coefficients/2/ratio_uppertolerance8.19609754576073978.1960975457618871.15E-1212passed
/fit/coefficients/3/estimatetolerance2.41179128105566992.41179128105567088.55E-1615passed
/fit/coefficients/3/standard_errortolerance0.48887545959322520.48887545959326133.61E-1413passed
/fit/coefficients/3/statistictolerance4.93334495264382964.9333449526434673.62E-1313passed
/fit/coefficients/3/p_valuetail_probability8.0833212135920942e-078.083321213607072e-071.50E-1811passed
/fit/coefficients/3/lowertolerance1.45361298732748211.45361298732741226.99E-1413passed
/fit/coefficients/3/uppertolerance3.36996957478385763.3699695747839297.15E-1413passed
/fit/coefficients/3/ratiotolerance11.15392307211226911.153923072112281.05E-1415passed
/fit/coefficients/3/ratio_lowertolerance4.27854495208892694.2785449520886282.99E-1313passed
/fit/coefficients/3/ratio_uppertolerance29.07764235078500929.0776423507870872.08E-1213passed
/fit/term_tests/[term=grp]/statistictolerance7.72542139157491197.7254213915742047.08E-1313passed
/fit/term_tests/[term=grp]/p_valuetail_probability0.0210109679180716070.0210109679180790457.44E-1512passed
/fit/statistics/[label=Event rows]/valuetolerance1414.00.00E+2passed
/fit/rows_usedexact30300.00E+2passed
/fit/rankexact440.00E+2passed
/fit/design_dfexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

278  logistic_case("STAT-REG-SURVEY-LOGIT-002", "Survey logistic regression with normal inference",
279                taylor, design_taylor, D_T, "normal", 0.95,
280                c("Normal inference: z tests, qnorm intervals, chi-square term tests."), "axis_coverage")

Whole script generators/r/regression_survey_models.R


Survey logistic regression with a first-stage FPC and 90% intervals

Case STAT-REG-SURVEY-LOGIT-003 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance-1.0714957947790638-1.07149579477906261.15E-1514passed
/fit/coefficients/0/standard_errortolerance0.719134642425143130.71913464242520446.13E-1413passed
/fit/coefficients/0/statistictolerance-1.4899793885129082-1.489979388512781.28E-1313passed
/fit/coefficients/0/p_valuetail_probability0.196414732483400580.196414732483433163.26E-1412passed
/fit/coefficients/0/lowertolerance-2.5205868862052734-2.5205868862053961.23E-1313passed
/fit/coefficients/0/uppertolerance0.377595296647145640.37759529664727091.25E-1312passed
/fit/coefficients/0/dfexact550.00E+2passed
/fit/coefficients/1/estimatetolerance3.08138293989121343.08138293989121259.20E-1615passed
/fit/coefficients/1/standard_errortolerance1.16747659740003961.16747659740013981.00E-1313passed
/fit/coefficients/1/statistictolerance2.63935306862118412.6393530686209572.27E-1313passed
/fit/coefficients/1/p_valuetail_probability0.0460120064385511620.046012006438563721.26E-1412passed
/fit/coefficients/1/lowertolerance0.728861121395891760.72886112139568792.04E-1312passed
/fit/coefficients/1/uppertolerance5.4339047583865355.4339047583867372.02E-1313passed
/fit/coefficients/1/dfexact550.00E+2passed
/fit/coefficients/1/ratiotolerance21.78851377515436421.7885137751543432.15E-1415passed
/fit/coefficients/1/ratio_lowertolerance2.07271868799342982.07271868799300754.22E-1312passed
/fit/coefficients/1/ratio_uppertolerance229.04185468105175229.041854681097934.62E-1112passed
/fit/coefficients/2/estimatetolerance-0.14917538856652648-0.149175388566526873.93E-1614passed
/fit/coefficients/2/standard_errortolerance1.0612161967490191.0612161967490856.60E-1413passed
/fit/coefficients/2/statistictolerance-0.14057021464949138-0.140570214649483028.36E-1513passed
/fit/coefficients/2/p_valuetail_probability0.893696989576269420.89369698957627566.22E-1514passed
/fit/coefficients/2/lowertolerance-2.2875773595802942-2.28757735958042831.34E-1313passed
/fit/coefficients/2/uppertolerance1.9892265824472411.98922658244737451.33E-1313passed
/fit/coefficients/2/dfexact550.00E+2passed
/fit/coefficients/2/ratiotolerance0.861418018777586060.86141801877758573.34E-1615passed
/fit/coefficients/2/ratio_lowertolerance0.101512091493540710.101512091493527091.36E-1412passed
/fit/coefficients/2/ratio_uppertolerance7.3098779875096767.3098779875106529.76E-1312passed
/fit/coefficients/3/estimatetolerance2.41179128105566992.41179128105567088.55E-1615passed
/fit/coefficients/3/standard_errortolerance0.446948933071176830.44694893307120813.13E-1413passed
/fit/coefficients/3/statistictolerance5.39612269456204535.396122694561673.76E-1313passed
/fit/coefficients/3/p_valuetail_probability0.00295151523457439250.00295151523457529379.01E-1612passed
/fit/coefficients/3/lowertolerance1.51116756050766511.51116756050760256.26E-1413passed
/fit/coefficients/3/uppertolerance3.31241500160367463.3124150016037396.44E-1413passed
/fit/coefficients/3/dfexact550.00E+2passed
/fit/coefficients/3/ratiotolerance11.15392307211226911.153923072112281.05E-1415passed
/fit/coefficients/3/ratio_lowertolerance4.53201911303381794.5320191130335342.84E-1313passed
/fit/coefficients/3/ratio_uppertolerance27.45134051637219127.451340516373961.77E-1213passed
/fit/term_tests/[term=grp]/statistictolerance4.66302831849009624.66302831848967354.23E-1313passed
/fit/term_tests/[term=grp]/p_valuetail_probability0.071962849451077060.071962849451087681.06E-1412passed
/fit/term_tests/[term=grp]/denominator_dfexact550.00E+2passed
/fit/statistics/[label=Event rows]/valuetolerance1414.00.00E+2passed
/fit/rows_usedexact30300.00E+2passed
/fit/rankexact440.00E+2passed
/fit/design_dfexact880.00E+2passed
/fit/test_dfexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

281  logistic_case("STAT-REG-SURVEY-LOGIT-003", "Survey logistic regression with a first-stage FPC and 90% intervals",
282                with_fpc, design_fpc, D_F, "design_df", 0.90,
283                c("svydesign(fpc = ~N) with stratum PSU population counts 25 and 40; 90% intervals use qt(0.95, 5)."),
284                "axis_coverage")

Whole script generators/r/regression_survey_models.R


Survey proportional-odds regression (Taylor)

Case STAT-REG-SURVEY-ORD-001 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance1.35246378640243271.3524637864024326.67E-1615passed
/fit/coefficients/0/standard_errortolerance1.34614269389997051.34614269389997171.15E-1515passed
/fit/coefficients/0/statistictolerance1.00469570761785221.00469570761785091.33E-1514passed
/fit/coefficients/0/p_valuetail_probability0.315043395669114990.315043395669115666.66E-1614passed
/fit/coefficients/0/lowertolerance-1.2859274116932351-1.28592741169323893.78E-1514passed
/fit/coefficients/0/uppertolerance3.99085498449810053.9908549844981032.66E-1515passed
/fit/coefficients/0/ratiotolerance3.86694112064496983.8669411206449672.71E-1515passed
/fit/coefficients/0/ratio_lowertolerance0.276394133581541410.276394133581540361.05E-1514passed
/fit/coefficients/0/ratio_uppertolerance54.10112521843192554.1011252184320741.49E-1314passed
/fit/coefficients/1/estimatetolerance0.0886686952764496330.088668695276449831.94E-1614passed
/fit/coefficients/1/standard_errortolerance0.779205749443789180.7792057494437892.27E-1615passed
/fit/coefficients/1/statistictolerance0.113793687148410940.113793687148411232.92E-1614passed
/fit/coefficients/1/p_valuetail_probability0.909401343012788410.90940134301278813.33E-1615passed
/fit/coefficients/1/lowertolerance-1.438546510179918-1.4385465101799183.01E-1716passed
/fit/coefficients/1/uppertolerance1.61588390073281741.61588390073281742.00E-1817passed
/fit/coefficients/1/ratiotolerance1.09271857352485081.0927185735248512.22E-1615passed
/fit/coefficients/1/ratio_lowertolerance0.237272381159420620.237272381159420622.26E-2019passed
/fit/coefficients/1/ratio_uppertolerance5.03233404196305315.0323340419630531.73E-1717passed
/fit/coefficients/2/estimatetolerance0.979782344861366280.97978234486136584.41E-1615passed
/fit/coefficients/2/standard_errortolerance0.627895651090221720.6278956510902223.31E-1615passed
/fit/coefficients/2/statistictolerance1.56042225035347841.56042225035347681.59E-1514passed
/fit/coefficients/2/p_valuetail_probability0.118660130130485220.11866013013048563.77E-1614passed
/fit/coefficients/2/lowertolerance-0.25087051732479604-0.250870517324797371.33E-1514passed
/fit/coefficients/2/uppertolerance2.21043520704752882.2104352070475292.72E-1716passed
/fit/coefficients/2/ratiotolerance2.66387637244477252.6638763724447711.34E-1515passed
/fit/coefficients/2/ratio_lowertolerance0.778123118499346230.77812311849934529.96E-1614passed
/fit/coefficients/2/ratio_uppertolerance9.11968448046500599.1196844804650064.65E-1818passed
/fit/thresholds/0/estimatetolerance0.0304781311245355680.0304781311245355234.52E-1714passed
/fit/thresholds/0/standard_errortolerance0.571521367020598190.57152136702059831.13E-1615passed
/fit/thresholds/0/statistictolerance0.0533280694008367790.053328069400836699.00E-1714passed
/fit/thresholds/0/p_valuetail_probability0.957470515868448110.95747051586844821.14E-1615passed
/fit/thresholds/0/lowertolerance-1.0896831646309346-1.0896831646309354.64E-1615passed
/fit/thresholds/0/uppertolerance1.15063942688000551.1506394268800064.78E-1615passed
/fit/thresholds/1/estimatetolerance1.46418141849070091.46418141849070071.85E-1615passed
/fit/thresholds/1/standard_errortolerance0.722672436891823260.72267243689182334.50E-1817passed
/fit/thresholds/1/statistictolerance2.02606512127135962.02606512127135962.59E-1716passed
/fit/thresholds/1/p_valuetail_probability0.0427581117347082260.042758111734708193.45E-1715passed
/fit/thresholds/1/lowertolerance0.0477694695629324870.047769469562932044.44E-1614passed
/fit/thresholds/1/uppertolerance2.88059336741846922.8805933674184693.60E-1716passed
/fit/term_tests/[term=grp]/dfexact220.00E+2passed
/fit/term_tests/[term=grp]/statistictolerance1.19989096490054231.19989096490053962.71E-1514passed
/fit/term_tests/[term=grp]/p_valuetail_probability0.548841556775282680.54884155677528357.77E-1614passed
/fit/term_tests/[term=x]/dfexact110.00E+2passed
/fit/term_tests/[term=x]/statistictolerance2.43491759939821372.43491759939820934.41E-1514passed
/fit/term_tests/[term=x]/p_valuetail_probability0.11866013013048510.118660130130485796.91E-1614passed
/fit/term_tests/[term=All predictors]/dfexact330.00E+2passed
/fit/term_tests/[term=All predictors]/statistictolerance7.33198007371413947.33198007371413858.87E-1615passed
/fit/term_tests/[term=All predictors]/p_valuetail_probability0.0620364571486919580.062036457148692015.53E-1715passed
/fit/rows_usedexact30300.00E+2passed
/fit/rankexact330.00E+2passed
/fit/design_dfexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

413  ordinal_case("STAT-REG-SURVEY-ORD-001", "Survey proportional-odds regression (Taylor)",
414               taylor, design_taylor, D_T, 0.95, list(),
415               c("Normal reference for every coefficient; D = 8 shown for reference only."), "nominal")

Whole script generators/r/regression_survey_models.R


Survey proportional-odds regression in a domain with 90% intervals

Case STAT-REG-SURVEY-ORD-002 · status: passed

Reference

Evidence type
R
Source
R survey package: svydesign, svrepdesign, svyglm, svyolr, svycoxph, regTermTest
Citation
Lumley, T. (2024). survey: analysis of complex survey samples. R package version 4.5; Binder (1983, 1992); Lumley (2010) Complex Surveys, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survey 4.5, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/regression_survey_models.R · sha256 6c4d50bfb5529493
Dataset
reference_data/datasets/REG_SURVEY_B7.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fit/coefficients/0/estimatetolerance1.25750854934943131.25750854934943229.36E-1615passed
/fit/coefficients/0/standard_errortolerance1.93399738608375091.93399738608374278.18E-1514passed
/fit/coefficients/0/statistictolerance0.650212124586075020.65021212458607823.22E-1514passed
/fit/coefficients/0/p_valuetail_probability0.515555210302384380.51555521030238242.00E-1514passed
/fit/coefficients/0/lowertolerance-1.9236340656650919-1.92363406566507861.33E-1414passed
/fit/coefficients/0/uppertolerance4.43865116436395464.4386511643639431.16E-1414passed
/fit/coefficients/0/ratiotolerance3.51664900634673133.51664900634673443.09E-1515passed
/fit/coefficients/0/ratio_lowertolerance0.146075149708917940.146075149708919881.94E-1513passed
/fit/coefficients/0/ratio_uppertolerance84.66067129475926884.66067129475839.66E-1313passed
/fit/coefficients/1/estimatetolerance-0.00019148310635745192-0.000191483106356581748.70E-1611passed
/fit/coefficients/1/standard_errortolerance1.01050614770034141.01050614770033684.65E-1514passed
/fit/coefficients/1/statistictolerance-0.000189492272553927-0.000189492272553066778.60E-1611passed
/fit/coefficients/1/p_valuetail_probability0.999848807042242550.99984880704224337.76E-1615passed
/fit/coefficients/1/lowertolerance-1.6623261852080233-1.66232618520801567.74E-1514passed
/fit/coefficients/1/uppertolerance1.66194321899530851.66194321899530255.99E-1514passed
/fit/coefficients/1/ratiotolerance0.99980853522536250.99980853522536348.93E-1615passed
/fit/coefficients/1/ratio_lowertolerance0.189697195653976250.189697195653977721.47E-1514passed
/fit/coefficients/1/ratio_uppertolerance5.26954077345914575.2695407734591153.11E-1414passed
/fit/coefficients/2/estimatetolerance0.379379958796138450.379379958796138672.22E-1615passed
/fit/coefficients/2/standard_errortolerance1.37573066497594421.37573066497594132.88E-1514passed
/fit/coefficients/2/statistictolerance0.275766157180753260.275766157180754037.75E-1614passed
/fit/coefficients/2/p_valuetail_probability0.782727678792985550.7827276787929855.57E-1615passed
/fit/coefficients/2/lowertolerance-1.883495615197903-1.88349561519789923.79E-1514passed
/fit/coefficients/2/uppertolerance2.642255532790182.64225553279017643.56E-1514passed
/fit/coefficients/2/ratiotolerance1.46137819376737241.46137819376737294.79E-1615passed
/fit/coefficients/2/ratio_lowertolerance0.152057640652902340.15205764065290295.52E-1614passed
/fit/coefficients/2/ratio_uppertolerance14.04484652036474714.0448465203646975.02E-1414passed
/fit/thresholds/0/estimatetolerance-0.52360386930507496-0.5236038693050749.94E-1614passed
/fit/thresholds/0/standard_errortolerance1.35500158714482181.35500158714481566.24E-1514passed
/fit/thresholds/0/statistictolerance-0.38642306715550179-0.386423067155502841.05E-1514passed
/fit/thresholds/0/p_valuetail_probability0.699183369213885910.69918336921388517.79E-1614passed
/fit/thresholds/0/lowertolerance-2.7523831444452358-2.75238314444522561.02E-1414passed
/fit/thresholds/0/uppertolerance1.70517540583508561.70517540583507748.17E-1514passed
/fit/thresholds/1/estimatetolerance1.24385102520358261.24385102520358358.82E-1615passed
/fit/thresholds/1/standard_errortolerance1.31502449931864061.3150244993186346.63E-1514passed
/fit/thresholds/1/statistictolerance0.945876693436559220.94587669343656475.44E-1514passed
/fit/thresholds/1/p_valuetail_probability0.344211482229700770.3442114822296982.77E-1514passed
/fit/thresholds/1/lowertolerance-0.9191717920307263-0.91917179203071521.11E-1413passed
/fit/thresholds/1/uppertolerance3.40687384243789153.4068738424378829.34E-1514passed
/fit/term_tests/[term=grp]/dfexact220.00E+2passed
/fit/term_tests/[term=grp]/statistictolerance0.507052248806121320.50705224880612239.99E-1614passed
/fit/term_tests/[term=grp]/p_valuetail_probability0.776059470569901990.77605947056990154.46E-1615passed
/fit/term_tests/[term=x]/dfexact110.00E+2passed
/fit/term_tests/[term=x]/statistictolerance0.0760469734462398960.076046973446240344.44E-1614passed
/fit/term_tests/[term=x]/p_valuetail_probability0.782727678792985660.7827276787929856.67E-1615passed
/fit/term_tests/[term=All predictors]/dfexact330.00E+2passed
/fit/term_tests/[term=All predictors]/statistictolerance1.11291084725028911.11291084725030351.44E-1413passed
/fit/term_tests/[term=All predictors]/p_valuetail_probability0.773958303491963150.77395830349195973.44E-1514passed
/fit/rows_usedexact17170.00E+2passed
/fit/rankexact330.00E+2passed
/fit/design_dfexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

416  ordinal_case("STAT-REG-SURVEY-ORD-002", "Survey proportional-odds regression in a domain with 90% intervals",
417               domain_a, design_taylor, D_T, 0.90,
418               list(domain_column = "dom", domain_values = list("A")),
419               c("svyolr on subset(design, dom == 'A'); qnorm(0.95) intervals."), "axis_coverage")

Whole script generators/r/regression_survey_models.R


Regression completion STAT-REG-THRESHOLD-BINARY-ABOVE-001

Case STAT-REG-THRESHOLD-BINARY-ABOVE-001 · status: passed

Reference

Evidence type
R
Source
R regression fit
Citation
R stats::glm/lm and car::Anova help pages; Tabelz manual chapters 07 and 08
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_lastmile.R · sha256 512832af1db86a5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/accuracytolerance0.930555555555555580.93055555555555562.27E-1918passed

Convention

Reference code

Whole script generators/r/regression_lastmile.R


Regression completion STAT-REG-THRESHOLD-BINARY-BELOW-001

Case STAT-REG-THRESHOLD-BINARY-BELOW-001 · status: passed

Reference

Evidence type
R
Source
R regression fit
Citation
R stats::glm/lm and car::Anova help pages; Tabelz manual chapters 07 and 08
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, car 3.1.5, jsonlite 2.0.0
Reference script
generators/r/regression_lastmile.R · sha256 512832af1db86a5f
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/accuracytolerance0.916666666666666630.91666666666666663.41E-1918passed

Convention

Reference code

Whole script generators/r/regression_lastmile.R


Regression II STAT-REG-TREE-CLASS-001

Case STAT-REG-TREE-CLASS-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/threshold/0tolerance00.00.00E+2passed
/training_metrics/accuracytolerance11.00.00E+2passed
/test_metrics/accuracytolerance11.00.00E+2passed
/source_countexact40400.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

51  tree_class <- rpart::rpart(class ~ x, data = tree, method = "class",
52                            control = rpart::rpart.control(cp = 0, xval = 0,
53                                                          minsplit = 2, minbucket = 1,
54                                                          maxdepth = 1))
55  tree_reg <- rpart::rpart(y ~ x, data = tree, method = "anova",
56                          control = rpart::rpart.control(cp = 0, xval = 0,
57                                                        minsplit = 2, minbucket = 1,
58                                                        maxdepth = 1))
59  stopifnot(nrow(tree_class$frame) == 3L, nrow(tree_reg$frame) == 3L,
60            abs(tree_class$splits[1L, "index"]) < 1e-12,
61            abs(tree_reg$splits[1L, "index"]) < 1e-12)
62  tree_id <- function(id, i) paste0(id, ":", i - 1L)
63  test_positions <- which((seq_len(nrow(tree)) - 1L) %% 8L %in% c(2L, 3L))
64  train <- sort(vapply(setdiff(seq_len(nrow(tree)), test_positions),
65                       function(i) tree_id("STAT-REG-TREE-CLASS-001", i), ""))

Whole script generators/r/regression_advanced_scoring.R


Small-sample boundary STAT-REG-TREE-CLASS-MIN-001

Case STAT-REG-TREE-CLASS-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats, plm, rpart
Citation
R stats::nls/lm, plm::plm, rpart::rpart documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_a.R · sha256 e69dea823e2a8f59
Dataset
reference_data/datasets/REG_MIN_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/threshold/0tolerance00.00.00E+2passed
/source_countexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

68  tree <- data.frame(x=rep(c(-1,1),each=4L),class=rep(c("A","B"),each=4L),
69                     y=rep(c(-2,2),each=4L))
70  export_dataset(tree,"REG_MIN_TREE","Eight rows in two pure CART classes",
71                 "R rpart unique-split classification and regression tree",
72                 "Constructed validation data; no third-party license.",
73                 c(x="Float64",class="String",y="Float64"))
74  t <- read_dataset("REG_MIN_TREE")
75  train <- c(1L,2L,3L,5L,6L,7L)
76  fit <- rpart::rpart(class~x,data=t[train,],method="class",
77                      control=rpart::rpart.control(cp=0,xval=0,minsplit=2,
78                                                    minbucket=1,maxdepth=1))
79  id <- "STAT-REG-TREE-CLASS-MIN-001"

Whole script generators/r/regression_advanced_minimum_a.R


Regression edge STAT-REG-TREE-CLASS-ONECLASS-001

Case STAT-REG-TREE-CLASS-ONECLASS-001 · status: passed

Reference

Evidence type
Analytical
Source
Analytic design condition
Citation
Tabelz manual regression model design conditions; R stats/plm references
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_b.R · sha256 4a5d01b584989ad6
Dataset
reference_data/datasets/REG_EDGE_TREE_ONECLASS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

51  tree <- read_dataset("REG_MIN_TREE")
52  tree$class <- rep("A",nrow(tree))
53  export_dataset(tree,"REG_EDGE_TREE_ONECLASS","Constant class response for a classification tree",
54                 "Analytic two-class requirement",
55                 "Constructed validation data; no third-party license.",
56                 c(x="Float64",class="String",y="Float64"))
57  emit("STAT-REG-TREE-CLASS-ONECLASS-001","STAT-REG-TREE-CLASS-001",
58       "REG_EDGE_TREE_ONECLASS",
59       list(response="class",numeric_predictors=list("x"),positive_class="B",
60            cv_folds=2L,max_depth=1L,min_samples_leaf=1L),
61       error=err("RuntimeError","at least two classes"),
62       note="A classification tree requires two observed response classes before splitting.")

Whole script generators/r/regression_advanced_special_b.R


True minimum sample STAT-REG-TREE-CLASS-TRUE-MIN-001

Case STAT-REG-TREE-CLASS-TRUE-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats/rpart/pls and analytic sample conditions
Citation
R stats::lm, rpart::rpart, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_true_min.R · sha256 f10363ad1ec76730
Dataset
reference_data/datasets/REG_TRUE_MIN_CLASS_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/threshold/0tolerance00.00.00E+2passed
/source_countexact660.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

52  tree <- data.frame(x=c(-1,-1,-1,1,1,1),class=c("A","A","A","B","B","B"),
53                     y=c(-2,-2,-2,2,2,2))
54  export_dataset(tree,"REG_TRUE_MIN_CLASS_TREE","Three rows in each pure class",
55                 "R rpart unique two-leaf split at the smallest CV-valid stratified sample",
56                 "Constructed validation data; no third-party license.",
57                 c(x="Float64",class="String",y="Float64"))
58  t <- read_dataset("REG_TRUE_MIN_CLASS_TREE")
59  train <- c(1L,2L,4L,5L); test <- c(3L,6L)
60  fit <- rpart::rpart(class~x,data=t[train,],method="class",
61                      control=rpart::rpart.control(cp=0,xval=0,minsplit=2,
62                                                    minbucket=1,maxdepth=1))
63  id <- "STAT-REG-TREE-CLASS-TRUE-MIN-001"

Whole script generators/r/regression_advanced_true_min.R


Regression II STAT-REG-TREE-REG-001

Case STAT-REG-TREE-REG-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/artifact/threshold/0tolerance00.00.00E+2passed
/artifact/value/1tolerance-2-2.00.00E+2passed
/artifact/value/2tolerance22.00.00E+2passed
/source_countexact40400.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

80  emit("STAT-REG-TREE-REG-001", "tabelz.executor.regression_tree",
81       "regressionTreeMenuItem", "REG_TREE",
82       list(response = "y", numeric_predictors = list("x"), cv_folds = 2L,
83            max_depth = 1L, depth_grid = list(1L), min_leaf_grid = list(1L)),
84       list("/artifact/threshold/0" = tol(tree_reg$splits[1L, "index"], "unique CART split", "1e-12"),
85            "/artifact/value/1" = tol(-2, "left-leaf mean", "1e-12"),
86            "/artifact/value/2" = tol(2, "right-leaf mean", "1e-12"),
87            "/source_count" = exact(as.integer(nrow(tree)), "complete source rows")),
88       list("Response is exactly constant on each side; random holdout cannot change leaf means."),
89       kind = "hard_numeric", deferred = list(list(path = "/split",
90         reason = "No explicit split assignment exists for regression trees; NumPy chooses held-out identities.")))

Whole script generators/r/regression_advanced_scoring.R


Regression edge STAT-REG-TREE-REG-CONSTANT-001

Case STAT-REG-TREE-REG-CONSTANT-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats/rpart
Citation
R nlme::lme, plm::plm, stats::nls/lm/step, rpart::rpart documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_a.R · sha256 b8ea61b61080b6f8
Dataset
reference_data/datasets/REG_EDGE_TREE_CONSTANT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact880.00E+2passed
/artifact/value/0tolerance77.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

76  tree <- data.frame(x=c(-2,-1,0,1,2,3,4,5),y=rep(7,8L))
77  export_dataset(tree,"REG_EDGE_TREE_CONSTANT","Constant-response regression tree",
78                 "R rpart no-split leaf mean boundary",
79                 "Constructed validation data; no third-party license.",
80                 c(x="Float64",y="Float64"))
81  t <- read_dataset("REG_EDGE_TREE_CONSTANT")
82  fit <- rpart::rpart(y~x,data=t,method="anova",
83                      control=rpart::rpart.control(cp=0,xval=0,minsplit=2,minbucket=1))
84  emit("STAT-REG-TREE-REG-CONSTANT-001","tabelz.executor.regression_tree",
85       "regressionTreeMenuItem","REG_EDGE_TREE_CONSTANT",
86       list(response="y",numeric_predictors=list("x"),cv_folds=2L,max_depth=1L,
87            depth_grid=list(1L),min_leaf_grid=list(1L)),
88       list("/source_count"=exact(as.integer(nrow(t)),"all eight constant-response rows"),
89            "/artifact/value/0"=tol(fit$frame$yval[1L],"constant-response root leaf mean","1e-9")),
90       c("With zero response variance, R rpart produces one leaf of value seven.",
91         "The random holdout cannot change the constant training response."))

Whole script generators/r/regression_advanced_special_a.R


Small-sample boundary STAT-REG-TREE-REG-MIN-001

Case STAT-REG-TREE-REG-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats, plm, rpart
Citation
R stats::nls/lm, plm::plm, rpart::rpart documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_a.R · sha256 e69dea823e2a8f59
Dataset
reference_data/datasets/REG_MIN_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact880.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 80  make_ids <- function(idx) as.list(sort(paste0(id,":",idx-1L)))
 81  emit(id,"tabelz.executor.classification_tree","classificationTreeMenuItem",
 82       "REG_MIN_TREE",
 83       list(response="class",numeric_predictors=list("x"),positive_class="B",cv_folds=2L,
 84            max_depth=1L,min_samples_leaf=1L,
 85            split_assignment=list(train_row_ids=make_ids(train),
 86                                  test_row_ids=make_ids(c(4L,8L)),test_fraction=0.25,seed=0L)),
 87       list("/artifact/threshold/0"=tol(fit$splits[1L,"index"],"unique split at zero"),
 88            "/source_count"=exact(nrow(t),"eight complete rows")),
 89       c("Six explicit training rows contain three per class; two held-out rows contain one per class.",
 90         "rpart uses cp=0, xval=0, minsplit=2, minbucket=1, maxdepth=1."))
 91  fit <- rpart::rpart(y~x,data=t,method="anova",
 92                      control=rpart::rpart.control(cp=0,xval=0,minsplit=2,
 93                                                    minbucket=1,maxdepth=1))
 94  emit("STAT-REG-TREE-REG-MIN-001","tabelz.executor.regression_tree",
 95       "regressionTreeMenuItem","REG_MIN_TREE",
 96       list(response="y",numeric_predictors=list("x"),cv_folds=2L,max_depth=1L,
 97            depth_grid=list(1L),min_leaf_grid=list(1L)),
 98       list("/source_count"=exact(nrow(t),"eight complete rows")),
 99       c("R full-sample split is uniquely at zero, but the engine's split and CV folds use NumPy RNG."),
100       deferred=list(list(path="/artifact/threshold/0",
101                          reason="The regression-tree API does not expose an explicit split assignment.")))

Whole script generators/r/regression_advanced_minimum_a.R


True minimum sample STAT-REG-TREE-REG-TRUE-MIN-001

Case STAT-REG-TREE-REG-TRUE-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats/rpart/pls and analytic sample conditions
Citation
R stats::lm, rpart::rpart, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_true_min.R · sha256 f10363ad1ec76730
Dataset
reference_data/datasets/REG_TRUE_MIN_REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

73  reg <- data.frame(x=c(-1,1,2),y=c(-2,2,3))
74  export_dataset(reg,"REG_TRUE_MIN_REG_TREE","Three finite regression-tree rows",
75                 "R rpart small-sample regression-tree estimability",
76                 "Constructed validation data; no third-party license.",
77                 c(x="Float64",y="Float64"))
78  r <- read_dataset("REG_TRUE_MIN_REG_TREE")
79  emit("STAT-REG-TREE-REG-TRUE-MIN-001","STAT-REG-TREE-REG-001",
80       "REG_TRUE_MIN_REG_TREE",
81       list(response="y",numeric_predictors=list("x"),cv_folds=2L,max_depth=1L,
82            depth_grid=list(1L),min_leaf_grid=list(1L)),
83       list("/source_count"=exact(as.integer(nrow(r)),"three finite tree rows")),
84       "One held-out row leaves two training rows for two-fold CV; fitted split depends on NumPy holdout.",
85       deferred=list(list(path="/artifact/threshold/0",
86         reason="The API does not expose the regression-tree training split.")))

Whole script generators/r/regression_advanced_true_min.R


Regression II STAT-REG-TREE-SCORE-CLASS-001

Case STAT-REG-TREE-SCORE-CLASS-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/tree_scoreexactAApassed
/1/tree_scoreabsentNonepassed
/2/tree_scoreexactAApassed
/2/__tabelz_row_idexactSTAT-REG-TREE-SCORE-CLASS-001:2STAT-REG-TREE-SCORE-CLASS-001:2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 93  reg_rows <- as.list(vapply(seq_len(nrow(tree) - 1L), function(i) tree_id(reg_op_id, i), ""))
 94  reg_art <- list(feature_order = list("x"), children_left = list(1L, -1L, -1L),
 95                  children_right = list(2L, -1L, -1L), feature = list(0L, -2L, -2L),
 96                  threshold = list(0, 0, 0), value = list(0, -2, 2),
 97                  n_node_samples = list(40L, 20L, 20L))
 98  emit(reg_op_id, "operation:tabelz.regression_tree_scores", "regressionTreeMenuItem",
 99       "REG_TREE", list(artifact = reg_art, row_ids = reg_rows),
100       list("/0/tree_score" = tol(-2, "first row follows left leaf", "1e-12"),
101            "/1/tree_score" = tol(2, "second row follows right leaf", "1e-12"),
102            "/0/__tabelz_row_id" = exact(tree_id(reg_op_id, 1L), "retained row identity"),
103            "/39/tree_score" = absent("unfitted final row must receive null")),
104       list("R rpart two-leaf model stored as the documented immutable numeric artifact."),
105       kind = "row_alignment")
106  class_op_id <- "STAT-REG-TREE-SCORE-CLASS-001"

Whole script generators/r/regression_advanced_scoring.R


Retained scoring boundary STAT-REG-TREE-SCORE-CLASS-MISSING-001

Case STAT-REG-TREE-SCORE-CLASS-MISSING-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls and analytic retained tree
Citation
R MASS::lda, pls::plsr and rpart::rpart documentation; Tabelz retained-artifact contract
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_boundaries.R · sha256 9a546ee72b8c92a9
Dataset
reference_data/datasets/REG_SCORE_BOUND_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/1/tree_scoreabsentNonepassed

Convention

Reference code

Whole script generators/r/regression_advanced_score_boundaries.R


Retained scoring boundary STAT-REG-TREE-SCORE-CLASS-THRESHOLD-001

Case STAT-REG-TREE-SCORE-CLASS-THRESHOLD-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls and analytic retained tree
Citation
R MASS::lda, pls::plsr and rpart::rpart documentation; Tabelz retained-artifact contract
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_boundaries.R · sha256 9a546ee72b8c92a9
Dataset
reference_data/datasets/REG_SCORE_BOUND_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/tree_scoreexactAApassed

Convention

Reference code

Whole script generators/r/regression_advanced_score_boundaries.R


True minimum sample STAT-REG-TREE-SCORE-CLASS-TRUE-MIN-001

Case STAT-REG-TREE-SCORE-CLASS-TRUE-MIN-001 · status: passed

Reference

Evidence type
R
Source
R stats/rpart/pls and analytic sample conditions
Citation
R stats::lm, rpart::rpart, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_true_min.R · sha256 f10363ad1ec76730
Dataset
reference_data/datasets/REG_SCORE_MID_DA.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/tree_scoreexactAApassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

87  id <- "STAT-REG-TREE-SCORE-CLASS-TRUE-MIN-001"

Whole script generators/r/regression_advanced_true_min.R


Regression II STAT-REG-TREE-SCORE-REG-001

Case STAT-REG-TREE-SCORE-REG-001 · status: passed

Reference

Evidence type
R
Source
R rpart/MASS/pls
Citation
R rpart::rpart, MASS::lda/qda, pls::plsr documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_scoring.R · sha256 b719391a082cd335
Dataset
reference_data/datasets/REG_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/tree_scoretolerance-2-2.00.00E+2passed
/1/tree_scoretolerance22.00.00E+2passed
/0/__tabelz_row_idexactSTAT-REG-TREE-SCORE-REG-001:0STAT-REG-TREE-SCORE-REG-001:0passed
/39/tree_scoreabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

92  reg_op_id <- "STAT-REG-TREE-SCORE-REG-001"

Whole script generators/r/regression_advanced_scoring.R


Large-offset retained tree scoring

Case STAT-REG-TREE-SCORE-REG-HARD-001 · status: passed

Reference

Evidence type
R
Source
R rpart and analytic tree traversal
Citation
R rpart::rpart documentation and Tabelz retained tree scoring contract
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, rpart 4.1.27, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_hard_closure.R · sha256 305566a494a873a8
Dataset
reference_data/datasets/REG_TREE_SCORE_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/tree_scoretolerance999999.875999999.8750.00E-1passed
/1/tree_scoretolerance999999.875999999.8750.00E-1passed
/2/tree_scoretolerance999999.875999999.8750.00E-1passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 1  source("common.R",local=TRUE)
 2  run <- generator_run(c("stats","rpart","jsonlite"))
 3  tol <- function(x,note) list(comparison="tolerance",value=dec(unname(x)),
 4    absolute_tolerance="0",relative_tolerance="1e-9",rationale=note)
 5  score <- data.frame(x=c(1e8-1,1e8,1e8+1),y=c(1e6-0.125,1e6-0.125,1e6+0.125))
 6  export_dataset(score,"REG_TREE_SCORE_HARD","Large-offset predictor and leaf values",
 7                 "R rpart unique threshold and analytic retained tree scoring",
 8                 "Constructed validation data; no third-party license.",
 9                 c(x="Float64",y="Float64"))
10  score <- read_dataset("REG_TREE_SCORE_HARD")
11  training <- data.frame(x=rep(c(1e8-1,1e8+1),each=4L),
12                         y=rep(c(1e6-0.125,1e6+0.125),each=4L))
13  tree <- rpart::rpart(y~x,data=training,method="anova",
14                       control=rpart::rpart.control(cp=0,xval=0,minsplit=2,
15                                                     minbucket=1,maxdepth=1))
16  stopifnot(nrow(tree$frame)==3L,abs(tree$splits[1L,"index"]-1e8)<1e-12)
17  id <- "STAT-REG-TREE-SCORE-REG-HARD-001"

Whole script generators/r/regression_advanced_hard_closure.R


Retained scoring boundary STAT-REG-TREE-SCORE-REG-MISSING-001

Case STAT-REG-TREE-SCORE-REG-MISSING-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls and analytic retained tree
Citation
R MASS::lda, pls::plsr and rpart::rpart documentation; Tabelz retained-artifact contract
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_boundaries.R · sha256 9a546ee72b8c92a9
Dataset
reference_data/datasets/REG_SCORE_BOUND_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/1/tree_scoreabsentNonepassed

Convention

Reference code

Whole script generators/r/regression_advanced_score_boundaries.R


Retained scoring boundary STAT-REG-TREE-SCORE-REG-THRESHOLD-001

Case STAT-REG-TREE-SCORE-REG-THRESHOLD-001 · status: passed

Reference

Evidence type
R
Source
R MASS/pls and analytic retained tree
Citation
R MASS::lda, pls::plsr and rpart::rpart documentation; Tabelz retained-artifact contract
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_score_boundaries.R · sha256 9a546ee72b8c92a9
Dataset
reference_data/datasets/REG_SCORE_BOUND_TREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/tree_scoretolerance-2-2.00.00E+2passed

Convention

Reference code

Whole script generators/r/regression_advanced_score_boundaries.R


STAT REG TWO STAGE 001

Case STAT-REG-TWO-STAGE-001 · status: passed

Reference

Evidence type
R
Source
Base R: solve/crossprod for the IV estimator, stats::lm for first stages and the Sargan auxiliary regression, stats::pt/pf/pchisq
Citation
R Core Team, stats package help: lm, summary.lm, pt, pf, pchisq. Wooldridge, Introductory Econometrics, ch. 15 (2SLS, first-stage F, Sargan n R^2).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_two_stage.R · sha256 7494053441fd504b
Dataset
reference_data/datasets/REG_IV_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact29290.00E+2passed
/residual_standard_errortolerance0.640129403632124120.64012940363212475.52E-1615passed
/r_squaredtolerance0.921927945960833270.9219279459608332.19E-1615passed
/wald_ftolerance102.0859343727042102.085934372703972.26E-1314passed
/wald_df1exact220.00E+2passed
/wald_df2exact26260.00E+2passed
/wald_ptail_probability4.8749970266896481e-134.874997026689796e-131.47E-2613passed
/coefficients/0/termexact(Intercept)(Intercept)passed
/coefficients/0/roleexactconstantconstantpassed
/coefficients/0/estimatetolerance1.4488385471805071.44883854718050922.20E-1514passed
/coefficients/0/standard_errortolerance0.236233571344321320.236233571344321663.38E-1614passed
/coefficients/0/t_valuetolerance6.13307642489457196.1330764248945739.31E-1615passed
/coefficients/0/p_valuetail_probability1.7483277866324688e-061.748327786632463e-065.76E-2114passed
/coefficients/0/lowertolerance0.963253486886513130.96325348688651451.34E-1514passed
/coefficients/0/uppertolerance1.93442360747450071.93442360747450383.11E-1514passed
/coefficients/1/termexactwwpassed
/coefficients/1/roleexactexogenousexogenouspassed
/coefficients/1/estimatetolerance-0.90159386918586648-0.90159386918586661.12E-1615passed
/coefficients/1/standard_errortolerance0.167151215060091870.167151215060091981.09E-1615passed
/coefficients/1/t_valuetolerance-5.3938816350317174-5.3938816350317143.59E-1515passed
/coefficients/1/p_valuetail_probability1.1919832986486634e-051.1919832986486771e-051.37E-1913passed
/coefficients/1/lowertolerance-1.2451781124468111-1.24517811244681186.58E-1615passed
/coefficients/1/uppertolerance-0.55800962592492176-0.55800962592492143.34E-1615passed
/coefficients/2/termexactxxpassed
/coefficients/2/roleexactendogenousendogenouspassed
/coefficients/2/estimatetolerance1.21491638918434491.21491638918434371.15E-1515passed
/coefficients/2/standard_errortolerance0.104072631609591490.10407263160959151.11E-1715passed
/coefficients/2/t_valuetolerance11.67373564398631411.6737356439863011.26E-1414passed
/coefficients/2/p_valuetail_probability7.7023964838325291e-127.702396483832708e-121.79E-2513passed
/coefficients/2/lowertolerance1.00099203115379481.00099203115379361.15E-1514passed
/coefficients/2/uppertolerance1.4288407472148951.42884074721489381.15E-1515passed
/first_stages/0/r_squaredtolerance0.699859158367293510.69985915836729341.11E-1615passed
/first_stages/0/partial_r_squaredtolerance0.688773485921089670.68877348592108942.24E-1615passed
/first_stages/0/f_valuetolerance27.66367319151565227.6636731915156351.74E-1415passed
/first_stages/0/df1exact220.00E+2passed
/first_stages/0/df2exact25250.00E+2passed
/first_stages/0/p_valuetail_probability4.6074152462148602e-074.607415246214892e-073.18E-2114passed
/sargan/statistictolerance0.679183734877409570.67918373487741596.33E-1514passed
/sargan/dfexact110.00E+2passed
/sargan/p_valuetail_probability0.40986794382816250.40986794382816042.11E-1514passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

34  emit <- function(id, kind, config, outputs, notes, excluded = list()) {
35    write_expected("regression", id, list(
36      schema_version = "2.0.0", case_id = id, outputs = outputs,
37      expected_error = NULL, deferred_outputs = list(), generator_run = run))
38    write_case("regression", id, list(
39      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
40      entry_point = "tabelz.analysis.two_stage_least_squares:compute_two_stage",
41      menu_items = list("twoStageLeastSquaresMenuItem"), title = gsub("-", " ", id),
42      case_kind = kind,
43      dataset = list(path = "reference_data/datasets/REG_IV_01.json", sha256 = ""),
44      excluded_rows = excluded, configuration = config, "repeat" = 1L,
45      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"), sha256 = ""),
46      oracle = list(
47        oracle_class = "independent_implementation",
48        source = "Base R: solve/crossprod for the IV estimator, stats::lm for first stages and the Sargan auxiliary regression, stats::pt/pf/pchisq",
49        citation = "R Core Team, stats package help: lm, summary.lm, pt, pf, pchisq. Wooldridge, Introductory Econometrics, ch. 15 (2SLS, first-stage F, Sargan n R^2).",
50        generator = list(path = "generators/r/regression_two_stage.R", sha256 = ""),
51        independent = TRUE,
52        convention_notes = c(
53          notes,
54          "Instruments Z = [1, exogenous, excluded instruments]; X = [1, exogenous, endogenous]; beta = (X'P_Z X)^-1 X'P_Z y; residual variance from u = y - X beta on n - k df; t intervals (docs/manual/chapters/86-regression-and-count-models.tex:90-105).",
55          "HC1 covariance: (X_hat'X_hat)^-1 [sum u_i^2 x_hat_i x_hat_i'] (X_hat'X_hat)^-1 n/(n-k); the Sargan statistic n R^2 of u on Z is unchanged by it; robust first-stage F is the HC1 joint Wald statistic of the excluded instruments divided by their number.",
56          "First-stage F compares the first-stage regression on Z with the regression on the included exogenous regressors; partial R^2 = (RSS_restricted - RSS_full)/RSS_restricted.",
57          "Rows with a missing value in any role are dropped; excluded rows are removed first.",
58          "Expected computed from the exported CSV."
59        )),
60      supported = TRUE, unsupported_reason = NULL,
61      requirement_ids = list("REQ-STAT-REG-TWO-STAGE-001"),
62      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/regression_two_stage.R


STAT REG TWO STAGE 002

Case STAT-REG-TWO-STAGE-002 · status: passed

Reference

Evidence type
R
Source
Base R: solve/crossprod for the IV estimator, stats::lm for first stages and the Sargan auxiliary regression, stats::pt/pf/pchisq
Citation
R Core Team, stats package help: lm, summary.lm, pt, pf, pchisq. Wooldridge, Introductory Econometrics, ch. 15 (2SLS, first-stage F, Sargan n R^2).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_two_stage.R · sha256 7494053441fd504b
Dataset
reference_data/datasets/REG_IV_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact29290.00E+2passed
/residual_standard_errortolerance0.640129403632124120.64012940363212475.52E-1615passed
/r_squaredtolerance0.921927945960833270.9219279459608332.19E-1615passed
/wald_ftolerance128.71116087984973128.711160879849611.15E-1315passed
/wald_df1exact220.00E+2passed
/wald_df2exact26260.00E+2passed
/wald_ptail_probability3.2584484556997336e-143.258448455699767e-143.34E-2813passed
/coefficients/0/termexact(Intercept)(Intercept)passed
/coefficients/0/roleexactconstantconstantpassed
/coefficients/0/estimatetolerance1.4488385471805071.44883854718050922.20E-1514passed
/coefficients/0/standard_errortolerance0.299863334461044850.29986333446104534.44E-1614passed
/coefficients/0/t_valuetolerance4.83166289664775664.8316628966477571.09E-1717passed
/coefficients/0/p_valuetail_probability5.2406693160626065e-055.240669316062602e-054.77E-2015passed
/coefficients/0/lowertolerance0.832460635626215530.83246063562621659.97E-1614passed
/coefficients/0/uppertolerance2.06521645873479852.06521645873480163.14E-1514passed
/coefficients/1/termexactwwpassed
/coefficients/1/roleexactexogenousexogenouspassed
/coefficients/1/estimatetolerance-0.90159386918586648-0.90159386918586661.12E-1615passed
/coefficients/1/standard_errortolerance0.192105937971434580.19210593797143462.38E-1715passed
/coefficients/1/t_valuetolerance-4.6932118741687727-4.6932118741687732.37E-1717passed
/coefficients/1/p_valuetail_probability7.55420101092384e-057.554201010923822e-051.76E-1914passed
/coefficients/1/lowertolerance-1.296473280024252-1.29647328002425222.17E-1615passed
/coefficients/1/uppertolerance-0.50671445834748108-0.5067144583474811.13E-1615passed
/coefficients/2/termexactxxpassed
/coefficients/2/roleexactendogenousendogenouspassed
/coefficients/2/estimatetolerance1.21491638918434491.21491638918434371.15E-1515passed
/coefficients/2/standard_errortolerance0.107839850914837750.107839850914837782.60E-1715passed
/coefficients/2/t_valuetolerance11.26593164658375711.2659316465837431.39E-1414passed
/coefficients/2/p_valuetail_probability1.6756860531821621e-111.675686053182208e-114.59E-2513passed
/coefficients/2/lowertolerance0.99324840097003730.99324840097003611.22E-1514passed
/coefficients/2/uppertolerance1.43658437739865241.43658437739865131.10E-1515passed
/first_stages/0/r_squaredtolerance0.699859158367293510.69985915836729341.11E-1615passed
/first_stages/0/partial_r_squaredtolerance0.688773485921089670.68877348592108942.24E-1615passed
/first_stages/0/f_valuetolerance27.66367319151565227.6636731915156351.74E-1415passed
/first_stages/0/df1exact220.00E+2passed
/first_stages/0/df2exact25250.00E+2passed
/first_stages/0/p_valuetail_probability4.6074152462148602e-074.607415246214892e-073.18E-2114passed
/first_stages/0/robust_ftolerance32.64107955102336732.641079551023367.22E-1515passed
/first_stages/0/robust_ptail_probability1.0696110960674005e-071.069611096067404e-073.56E-2214passed
/sargan/statistictolerance0.679183734877409570.67918373487741596.33E-1514passed
/sargan/dfexact110.00E+2passed
/sargan/p_valuetail_probability0.40986794382816250.40986794382816042.11E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

195  r <- two_stage(d, "x", "w", c("z1", "z2"), "hc1", 0.95)
196  emit("STAT-REG-TWO-STAGE-002", "axis_coverage",
197       list(response = "y", endogenous = list("x"), exogenous = list("w"),
198            instruments = list("z1", "z2"), covariance = "hc1", confidence_level = 0.95),
199       outputs_for(r, c("(Intercept)", "w", "x"), c("constant", "exogenous", "endogenous")),
200       list("HC1 coefficient covariance and the robust first-stage F; the Sargan statistic is the homoskedastic n R^2."))

Whole script generators/r/regression_two_stage.R


STAT REG TWO STAGE 003

Case STAT-REG-TWO-STAGE-003 · status: passed

Reference

Evidence type
R
Source
Base R: solve/crossprod for the IV estimator, stats::lm for first stages and the Sargan auxiliary regression, stats::pt/pf/pchisq
Citation
R Core Team, stats package help: lm, summary.lm, pt, pf, pchisq. Wooldridge, Introductory Econometrics, ch. 15 (2SLS, first-stage F, Sargan n R^2).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_two_stage.R · sha256 7494053441fd504b
Dataset
reference_data/datasets/REG_IV_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact30300.00E+2passed
/residual_standard_errortolerance0.86353898233374360.86353898233374342.23E-1615passed
/r_squaredtolerance0.847533720309247230.84753372030924726.48E-1918passed
/wald_ftolerance90.59396120742118890.593961207421361.70E-1314passed
/wald_df1exact110.00E+2passed
/wald_df2exact28280.00E+2passed
/wald_ptail_probability2.8309367585977058e-102.830936758597643e-106.31E-2413passed
/coefficients/0/termexact(Intercept)(Intercept)passed
/coefficients/0/roleexactconstantconstantpassed
/coefficients/0/estimatetolerance0.470285381211777760.470285381211776651.11E-1514passed
/coefficients/0/standard_errortolerance0.217776313448102890.21777631344810271.97E-1615passed
/coefficients/0/t_valuetolerance2.15948820955612762.15948820955612453.07E-1514passed
/coefficients/0/p_valuetail_probability0.0395270298778121960.039527029877812442.43E-1614passed
/coefficients/0/lowertolerance0.099819357654816310.099819357654815428.88E-1614passed
/coefficients/0/uppertolerance0.840751404768739220.84075140476873791.33E-1514passed
/coefficients/1/termexactxxpassed
/coefficients/1/roleexactendogenousendogenouspassed
/coefficients/1/estimatetolerance1.36843067747579371.36843067747579424.66E-1615passed
/coefficients/1/standard_errortolerance0.143771623485030310.143771623485030238.45E-1715passed
/coefficients/1/t_valuetolerance9.51808600546460869.5180860054646167.10E-1515passed
/coefficients/1/p_valuetail_probability2.8309367585976857e-102.830936758597643e-104.30E-2413passed
/coefficients/1/lowertolerance1.12385632129577441.1238563212957756.63E-1615passed
/coefficients/1/uppertolerance1.6130050336558131.61300503365581332.70E-1615passed
/first_stages/0/r_squaredtolerance0.629693944924425870.62969394492442612.27E-1615passed
/first_stages/0/partial_r_squaredtolerance0.629693944924425870.62969394492442612.27E-1615passed
/first_stages/0/f_valuetolerance47.6131303180705647.6131303180705963.56E-1415passed
/first_stages/0/df1exact110.00E+2passed
/first_stages/0/df2exact28280.00E+2passed
/first_stages/0/p_valuetail_probability1.6850561116548277e-071.6850561116548115e-071.62E-2114passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

203  r <- two_stage(d, "x", character(0), "z1", "classical", 0.90)
204  stopifnot(is.null(r$sargan))
205  emit("STAT-REG-TWO-STAGE-003", "axis_coverage",
206       list(response = "y", endogenous = list("x"), exogenous = list(),
207            instruments = list("z1"), covariance = "classical", confidence_level = 0.90),
208       outputs_for(r, c("(Intercept)", "x"), c("constant", "endogenous")),
209       list("Exactly identified (one instrument, one endogenous regressor): the result carries no Sargan test, which the expected outputs therefore omit; 90% t intervals."))

Whole script generators/r/regression_two_stage.R


STAT REG TWO STAGE 004

Case STAT-REG-TWO-STAGE-004 · status: passed

Reference

Evidence type
R
Source
Base R: solve/crossprod for the IV estimator, stats::lm for first stages and the Sargan auxiliary regression, stats::pt/pf/pchisq
Citation
R Core Team, stats package help: lm, summary.lm, pt, pf, pchisq. Wooldridge, Introductory Econometrics, ch. 15 (2SLS, first-stage F, Sargan n R^2).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_two_stage.R · sha256 7494053441fd504b
Dataset
reference_data/datasets/REG_IV_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact27270.00E+2passed
/residual_standard_errortolerance0.66347963970591650.6634796397059174.46E-1615passed
/r_squaredtolerance0.916301512201451170.91630151220145111.07E-1615passed
/wald_ftolerance98.15211040844113298.152110408441058.53E-1415passed
/wald_df1exact220.00E+2passed
/wald_df2exact24240.00E+2passed
/wald_ptail_probability2.7942231458926052e-122.7942231458926165e-121.13E-2614passed
/coefficients/0/termexact(Intercept)(Intercept)passed
/coefficients/0/roleexactconstantconstantpassed
/coefficients/0/estimatetolerance1.44824918494338651.4482491849433874.47E-1615passed
/coefficients/0/standard_errortolerance0.31431524517925330.314315245179253744.44E-1614passed
/coefficients/0/t_valuetolerance4.60763264638167064.6076326463816655.35E-1514passed
/coefficients/0/p_valuetail_probability0.00011262869300620020.0001126286930062021.80E-1813passed
/coefficients/0/lowertolerance0.569128458748664130.56912845874866328.84E-1614passed
/coefficients/0/uppertolerance2.3273699111381092.32736991113811081.76E-1515passed
/coefficients/1/termexactwwpassed
/coefficients/1/roleexactexogenousexogenouspassed
/coefficients/1/estimatetolerance-0.89752491265364487-0.89752491265364491.50E-1817passed
/coefficients/1/standard_errortolerance0.19272822858720540.19272822858720573.02E-1614passed
/coefficients/1/t_valuetolerance-4.6569457895864694-4.6569457895864627.07E-1514passed
/coefficients/1/p_valuetail_probability9.9349751095310378e-059.934975109531268e-052.30E-1813passed
/coefficients/1/lowertolerance-1.4365741088744013-1.43657410887440228.78E-1615passed
/coefficients/1/uppertolerance-0.35847571643288856-0.358475716432887561.00E-1514passed
/coefficients/2/termexactxxpassed
/coefficients/2/roleexactendogenousendogenouspassed
/coefficients/2/estimatetolerance1.21785544296772891.2178554429677289.36E-1615passed
/coefficients/2/standard_errortolerance0.118920856488143160.118920856488143215.17E-1715passed
/coefficients/2/t_valuetolerance10.24089027721687710.2408902772168651.22E-1414passed
/coefficients/2/p_valuetail_probability3.0816876640204913e-103.081687664020575e-108.37E-2413passed
/coefficients/2/lowertolerance0.885241001514427640.88524100151442651.11E-1514passed
/coefficients/2/uppertolerance1.55046988442103011.55046988442102947.02E-1615passed
/first_stages/0/r_squaredtolerance0.686273398191322710.68627339819132274.37E-1817passed
/first_stages/0/partial_r_squaredtolerance0.67883747421447660.67883747421447651.12E-1615passed
/first_stages/0/f_valuetolerance24.30741548806928224.307415488069271.10E-1415passed
/first_stages/0/df1exact220.00E+2passed
/first_stages/0/df2exact23230.00E+2passed
/first_stages/0/p_valuetail_probability2.1248875153376344e-062.124887515337641e-066.81E-2114passed
/first_stages/0/robust_ftolerance26.62348426236773226.62348426236773.16E-1414passed
/first_stages/0/robust_ptail_probability1.0334998740138908e-061.0334998740139014e-061.06E-2013passed
/sargan/statistictolerance0.677728716885390050.67772871688538782.22E-1514passed
/sargan/dfexact110.00E+2passed
/sargan/p_valuetail_probability0.410369931112179290.410369931112180341.05E-1514passed
/missing_countexact110.00E+2passed
/excluded_countexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

212  r <- two_stage(d[-c(1, 6), ], "x", "w", c("z1", "z2"), "hc1", 0.99)
213  emit("STAT-REG-TWO-STAGE-004", "axis_coverage",
214       list(response = "y", endogenous = list("x"), exogenous = list("w"),
215            instruments = list("z1", "z2"), covariance = "hc1", confidence_level = 0.99),
216       c(outputs_for(r, c("(Intercept)", "w", "x"), c("constant", "exogenous", "endogenous")),
217         list("/missing_count" = exact(1L, "one row has a missing w"),
218              "/excluded_count" = exact(2L, "two rows excluded by identity"))),
219       list("Rows 0 and 5 (0-based) are excluded before the fit; HC1 covariance with 99% intervals."),
220       excluded = list(0L, 5L))

Whole script generators/r/regression_two_stage.R


Regression II STAT-REG-VALIDATED-COUNT-001

Case STAT-REG-VALIDATED-COUNT-001 · status: passed

Reference

Evidence type
R
Source
R nlme/plm/stats
Citation
R nlme::Orthodont, nlme::lme, plm::plm, stats::nls/lm/step documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_core.R · sha256 b7543c65b2778e63
Dataset
reference_data/datasets/REG_ORTHODONT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact1081080.00E+2passed
/training_metrics/row_countexact81810.00E+2passed
/test_metrics/row_countexact27270.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

118  emit("STAT-REG-VALIDATED-COUNT-001",
119       "tabelz.analysis.predictive_validation:compute_validated_regression",
120       "validatedRegressionMenuItem", "REG_ORTHODONT",
121       list(response = "distance", numeric_predictors = list("age"),
122            algorithm = "ols", imputation = "drop", test_fraction = 0.25, seed = 0),
123       list("/source_count" = exact(nrow(d), "complete source population"),
124            "/training_metrics/row_count" = exact(nrow(d) - floor(nrow(d) * 0.25),
125                                                    "training share after 25 percent holdout"),
126            "/test_metrics/row_count" = exact(floor(nrow(d) * 0.25),
127                                                "test share after 25 percent holdout")),
128       list("R checks population and deterministic split sizes only; Python NumPy RNG assigns row identities."),
129       kind = "determinism",
130       deferred = list(list(path = "/recipe/coefficients",
131                            reason = "No explicit split assignment exists; R cannot reproduce NumPy's row permutation.")),
132       repeat_count = 2L)

Whole script generators/r/regression_advanced_core.R


Small-sample boundary STAT-REG-VALIDATED-MIN-001

Case STAT-REG-VALIDATED-MIN-001 · status: passed

Reference

Evidence type
R
Source
R nlme/MASS/pls/stats
Citation
R nlme::lme, MASS::lda, pls::plsr and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, nlme 3.1.169, MASS 7.3.65, pls 2.9.0, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_minimum_b.R · sha256 e3e09e1965800ae1
Dataset
reference_data/datasets/REG_MIN_SELECTION.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact550.00E+2passed
/training_metrics/row_countexact440.00E+2passed
/test_metrics/row_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

55  v <- read_dataset("REG_MIN_SELECTION")
56  emit("STAT-REG-VALIDATED-MIN-001",
57       "tabelz.analysis.predictive_validation:compute_validated_regression",
58       "validatedRegressionMenuItem","REG_MIN_SELECTION",
59       list(response="y",numeric_predictors=list("x"),algorithm="ols",
60            imputation="drop",test_fraction=0.2,seed=0L),
61       list("/source_count"=exact(nrow(v),"five complete rows"),
62            "/training_metrics/row_count"=exact(4L,"four training rows"),
63            "/test_metrics/row_count"=exact(1L,"one held-out row")),
64       c("R checks population and split sizes; NumPy determines the held-out identity."),
65       deferred=list(list(path="/recipe/coefficients",
66                          reason="No explicit row split exists to replay the NumPy permutation in R.")))

Whole script generators/r/regression_advanced_minimum_b.R


Regression edge STAT-REG-VALIDATED-SMALLTRAIN-001

Case STAT-REG-VALIDATED-SMALLTRAIN-001 · status: passed

Reference

Evidence type
Analytical
Source
Analytic design condition
Citation
Tabelz manual regression model design conditions; R stats/plm references
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, plm 2.6.7, jsonlite 2.0.0
Reference script
generators/r/regression_advanced_special_b.R · sha256 4a5d01b584989ad6
Dataset
reference_data/datasets/REG_EDGE_VALIDATED_SMALLTRAIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

40  v <- read_dataset("REG_MIN_SELECTION")[1:4,]
41  export_dataset(v,"REG_EDGE_VALIDATED_SMALLTRAIN","Four-row OLS population with two-row holdout",
42                 "Analytic train-rank boundary for one-predictor validated regression",
43                 "Constructed validation data; no third-party license.",
44                 c(x="Float64",y="Float64"))
45  emit("STAT-REG-VALIDATED-SMALLTRAIN-001","STAT-REG-VALIDATED-COUNT-001",
46       "REG_EDGE_VALIDATED_SMALLTRAIN",
47       list(response="y",numeric_predictors=list("x"),algorithm="ols",
48            imputation="drop",test_fraction=0.5,seed=0L),
49       error=err("ValueError","cannot estimate"),
50       note="Two training rows cannot estimate the one-predictor model plus the required residual degrees of freedom.")

Whole script generators/r/regression_advanced_special_b.R


STAT REG VIF 001

Case STAT-REG-VIF-001 · status: passed

Reference

Evidence type
R
Source
R stats, MASS and nnet
Citation
R stats::lm/glm, MASS::polr/glm.nb, nnet::multinom help pages
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, MASS 7.3.65, nnet 7.3.20, jsonlite 2.0.0
Reference script
generators/r/regression_core.R · sha256 4b30ab4a6f513e54
Dataset
reference_data/datasets/REG_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0tolerance1.00010902110332341.00010902110332372.70E-1615passed
/1tolerance1.00010902110332341.00010902110332321.74E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

76  vif_entry <- "tabelz.analysis.fit_y_by_x:compute_fit_vif"
77  r2_x <- summary(stats::lm(x ~ z, data = d))$r.squared
78  r2_z <- summary(stats::lm(z ~ x, data = d))$r.squared
79  emit("STAT-REG-VIF-001", vif_entry, fit_menu, "nominal",
80       list(y_columns = list("y"), x_columns = list("x", "z"),
81            missing = "drop", model_type = "linear", degree = 2L,
82            interactions = FALSE, split_percent = 100L),
83       list("/0" = tol(1 / (1 - r2_x), "x VIF", "1e-9"),
84            "/1" = tol(1 / (1 - r2_z), "z VIF", "1e-9")),
85       list("Each VIF = 1/(1-R-squared) from the auxiliary R lm regression."))

Whole script generators/r/regression_core.R


STAT REG WEIGHT ESTIMATION 001

Case STAT-REG-WEIGHT-ESTIMATION-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm with weights = z^-p, stats::qt/pt/pf
Citation
R Core Team, stats package help: lm, summary.lm, confint.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_weight_estimation.R · sha256 8d75fc39f6126fa0
Dataset
reference_data/datasets/REG_WEIGHT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact23230.00E+2passed
/selected_powertolerance11.00.00E+2passed
/at_boundaryexactFalseFalsepassed
/r_squaredtolerance0.893387641483249780.89338764148324983.65E-1817passed
/adjusted_r_squaredtolerance0.882726405631574780.88272640563157484.22E-1817passed
/standard_errortolerance0.476152141849362050.4761521418493625.16E-1715passed
/f_valuetolerance83.79775608686931383.797756086869342.86E-1415passed
/df1exact220.00E+2passed
/df2exact20200.00E+2passed
/p_valuetail_probability1.8970357299823228e-101.8970357299823135e-109.32E-2514passed
/grid/0/log_likelihoodtolerance-40.058970301800926-40.0589703018009265.06E-1818passed
/grid/0/weighted_ssetolerance595.09247849633539595.09247849633541.37E-1517passed
/grid/1/log_likelihoodtolerance-36.920447624680385-36.920447624680387.12E-1515passed
/grid/1/weighted_ssetolerance236.00997405712687236.009974057126871.73E-1517passed
/grid/2/log_likelihoodtolerance-34.24487367233089-34.2448736723309041.40E-1415passed
/grid/2/weighted_ssetolerance97.44496192202137897.44496192202142.81E-1415passed
/grid/3/log_likelihoodtolerance-32.090104159755647-32.090104159755651.17E-1617passed
/grid/3/weighted_ssetolerance42.0975155769397742.0975155769397561.40E-1415passed
/grid/4/log_likelihoodtolerance-30.487057642566857-30.487057642566863.24E-1515passed
/grid/4/weighted_ssetolerance19.08047651287821419.0804765128782241.03E-1415passed
/grid/5/log_likelihoodtolerance-29.444361939337856-29.4443619393378637.28E-1515passed
/grid/5/weighted_ssetolerance9.0799502322497459.0799502322497493.51E-1515passed
/grid/6/log_likelihoodtolerance-28.956251325968481-28.9562513259684773.94E-1515passed
/grid/6/weighted_ssetolerance4.53441724375470034.53441724375469958.44E-1615passed
/grid/7/log_likelihoodtolerance-29.003524596610543-29.003524596610561.76E-1415passed
/grid/7/weighted_ssetolerance2.37234651380394322.37234651380394633.14E-1514passed
/grid/8/log_likelihoodtolerance-29.548849229465603-29.548849229465616.77E-1515passed
/grid/8/weighted_ssetolerance1.29611503013446111.29611503013446176.21E-1615passed
/coefficients/0/termexact(Intercept)(Intercept)passed
/coefficients/0/estimatetolerance1.96327385068364761.9632738506836476.96E-1615passed
/coefficients/0/standard_errortolerance0.417366346597343510.41736634659734352.84E-1817passed
/coefficients/0/t_valuetolerance4.70395820527841214.703958205278411.75E-1515passed
/coefficients/0/p_valuetail_probability0.000136165485585732430.000136165485585733167.28E-1914passed
/coefficients/0/lowertolerance1.09266290756269321.09266290756269284.02E-1615passed
/coefficients/0/uppertolerance2.83388479380460192.8338847938046018.90E-1615passed
/coefficients/1/termexactx1x1passed
/coefficients/1/estimatetolerance0.773602981321068310.77360298132106852.18E-1615passed
/coefficients/1/standard_errortolerance0.0917796692625374720.091779669262537461.37E-1715passed
/coefficients/1/t_valuetolerance8.42891445934679068.4289144593467943.57E-1515passed
/coefficients/1/p_valuetail_probability5.1536697811237143e-085.153669781123687e-082.72E-2214passed
/coefficients/1/lowertolerance0.582153946037264670.5821539460372653.32E-1615passed
/coefficients/1/uppertolerance0.965052016604871940.9650520166048721.14E-1615passed
/coefficients/2/termexactx2x2passed
/coefficients/2/estimatetolerance-1.1218811778017941-1.12188117780179392.39E-1615passed
/coefficients/2/standard_errortolerance0.123836518420608890.123836518420608892.41E-1816passed
/coefficients/2/t_valuetolerance-9.059372728740172-9.059372728740171.81E-1515passed
/coefficients/2/p_valuetail_probability1.6213705158206247e-081.6213705158206293e-084.64E-2314passed
/coefficients/2/lowertolerance-1.38019962866385-1.38019962866384981.80E-1615passed
/coefficients/2/uppertolerance-0.86356272693973801-0.86356272693973791.09E-1615passed
/missing_countexact110.00E+2passed
/excluded_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

33  emit <- function(id, kind, config, outputs, notes, excluded = list()) {
34    write_expected("regression", id, list(
35      schema_version = "2.0.0", case_id = id, outputs = outputs,
36      expected_error = NULL, deferred_outputs = list(), generator_run = run))
37    write_case("regression", id, list(
38      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = "regression",
39      entry_point = "tabelz.analysis.weight_estimation:compute_weight_estimation",
40      menu_items = list("weightEstimationMenuItem"), title = gsub("-", " ", id),
41      case_kind = kind,
42      dataset = list(path = "reference_data/datasets/REG_WEIGHT_01.json", sha256 = ""),
43      excluded_rows = excluded, configuration = config, "repeat" = 1L,
44      expected = list(path = paste0("reference_data/expected/regression/", id, ".json"), sha256 = ""),
45      oracle = list(
46        oracle_class = "independent_implementation",
47        source = "R stats::lm with weights = z^-p, stats::qt/pt/pf",
48        citation = "R Core Team, stats package help: lm, summary.lm, confint.",
49        generator = list(path = "generators/r/regression_weight_estimation.R", sha256 = ""),
50        independent = TRUE,
51        convention_notes = c(
52          notes,
53          "Weights are z^(-p); for each grid power p the profile log-likelihood is -n/2 [ln(2 pi) + ln(WSSE/n) + 1] + 1/2 sum ln w_i and the largest value selects p (docs/manual/chapters/86-regression-and-count-models.tex:71-84).",
54          "The final fit is stats::lm(y ~ predictors, weights = z^-p_selected); s^2 = WSSE/(n-k); t intervals on n-k df; R-squared and F from weighted sums of squares about the weighted mean, as summary.lm reports for a weighted fit.",
55          "Rows with a missing value in any role are dropped (R na.omit); excluded rows are removed before anything is computed.",
56          "Expected computed from the exported CSV."
57        )),
58      supported = TRUE, unsupported_reason = NULL,
59      requirement_ids = list("REQ-STAT-REG-WEIGHT-ESTIMATION-001"),
60      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/regression_weight_estimation.R


STAT REG WEIGHT ESTIMATION 002

Case STAT-REG-WEIGHT-ESTIMATION-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm with weights = z^-p, stats::qt/pt/pf
Citation
R Core Team, stats package help: lm, summary.lm, confint.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_weight_estimation.R · sha256 8d75fc39f6126fa0
Dataset
reference_data/datasets/REG_WEIGHT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact24240.00E+2passed
/selected_powertolerance00.00.00E+2passed
/at_boundaryexactFalseFalsepassed
/r_squaredtolerance0.475506442857456020.47550644285745591.07E-1615passed
/adjusted_r_squaredtolerance0.451665826623704030.45166582662370391.16E-1615passed
/standard_errortolerance1.83759482546382351.83759482546382353.31E-1716passed
/f_valuetolerance19.94522449399880819.94522449399887.40E-1515passed
/df1exact110.00E+2passed
/df2exact22220.00E+2passed
/p_valuetail_probability0.00019326894068704220.000193268940687042145.64E-2015passed
/grid/0/log_likelihoodtolerance-49.050588581031512-49.050588581031519.21E-1717passed
/grid/0/weighted_ssetolerance299.61203346720413299.612033467204075.87E-1415passed
/grid/1/log_likelihoodtolerance-48.437045697714709-48.437045697714714.19E-1617passed
/grid/1/weighted_ssetolerance206.98973267683368206.98973267683368.28E-1415passed
/grid/2/log_likelihoodtolerance-47.991264554924911-47.991264554924913.88E-1617passed
/grid/2/weighted_ssetolerance145.01397425902388145.01397425902393.09E-1415passed
/grid/3/log_likelihoodtolerance-47.716161729161954-47.716161729161954.69E-1617passed
/grid/3/weighted_ssetolerance103.04999157700033103.049991577000341.12E-1415passed
/grid/4/log_likelihoodtolerance-47.613369628266774-47.6133696282667742.14E-1617passed
/grid/4/weighted_ssetolerance74.28860433657122574.288604336571241.46E-1415passed
/grid/5/log_likelihoodtolerance-47.6836073553646-47.68360735536462.67E-1617passed
/grid/5/weighted_ssetolerance54.33236333893382454.3323633389338248.18E-1717passed
/grid/6/log_likelihoodtolerance-47.92676653104126-47.926766531041276.72E-1515passed
/grid/6/weighted_ssetolerance40.31375287182282840.3137528718228357.29E-1515passed
/grid/7/log_likelihoodtolerance-48.341655991568501-48.34165599156853.49E-1617passed
/grid/7/weighted_ssetolerance30.34331239729527930.343312397295284.85E-1616passed
/grid/8/log_likelihoodtolerance-48.925519974388422-48.925519974388427.12E-1717passed
/grid/8/weighted_ssetolerance23.16264437856836623.1626443785683737.33E-1515passed
/grid/9/log_likelihoodtolerance-49.673564124456675-49.673564124456691.42E-1415passed
/grid/9/weighted_ssetolerance17.92483509083420417.924835090834217.12E-1515passed
/grid/10/log_likelihoodtolerance-50.578724901158331-50.578724901158347.41E-1515passed
/grid/10/weighted_ssetolerance14.05427472392022314.054274723920236.92E-1515passed
/grid/11/log_likelihoodtolerance-51.6318064802919-51.63180648029193.25E-1617passed
/grid/11/weighted_ssetolerance11.15617044173542511.1561704417354141.11E-1415passed
/grid/12/log_likelihoodtolerance-52.821956944820592-52.8219569448205857.43E-1515passed
/grid/12/weighted_ssetolerance8.9574116075084768.9574116075084741.79E-1515passed
/coefficients/0/termexact(Intercept)(Intercept)passed
/coefficients/0/estimatetolerance1.29271290723443141.292712907234431.29E-1515passed
/coefficients/0/standard_errortolerance0.895605592232900820.89560559223290113.29E-1615passed
/coefficients/0/t_valuetolerance1.44339530530562321.44339530530562121.99E-1514passed
/coefficients/0/p_valuetail_probability0.162998572863859490.162998572863860055.57E-1614passed
/coefficients/0/lowertolerance-0.24517119713177982-0.245171197131781371.55E-1514passed
/coefficients/0/uppertolerance2.83059701160064232.83059701160064149.34E-1615passed
/coefficients/1/termexactx1x1passed
/coefficients/1/estimatetolerance0.806945556113415670.80694555611341581.15E-1615passed
/coefficients/1/standard_errortolerance0.18068611073539370.180686110735393732.90E-1715passed
/coefficients/1/t_valuetolerance4.46600766837662194.4660076683766221.36E-1717passed
/coefficients/1/p_valuetail_probability0.000193268940687042930.000193268940687043071.35E-1915passed
/coefficients/1/lowertolerance0.496681417535488830.496681417535488941.15E-1615passed
/coefficients/1/uppertolerance1.11720969469134261.11720969469134263.03E-1716passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

148  r <- profile(d, "x1", grid_of(-1, 2, 0.25), 0.90)
149  emit("STAT-REG-WEIGHT-ESTIMATION-002", "axis_coverage",
150       list(response = "y", predictors = list("x1"), weight_variable = "z",
151            power_from = -1, power_to = 2, power_step = 0.25, confidence_level = 0.90),
152       outputs_for(r, c("(Intercept)", "x1")),
153       list("Grid -1 to 2 by 0.25 with 90% t intervals; one predictor."))

Whole script generators/r/regression_weight_estimation.R


STAT REG WEIGHT ESTIMATION 003

Case STAT-REG-WEIGHT-ESTIMATION-003 · status: passed

Reference

Evidence type
R
Source
R stats::lm with weights = z^-p, stats::qt/pt/pf
Citation
R Core Team, stats package help: lm, summary.lm, confint.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_weight_estimation.R · sha256 8d75fc39f6126fa0
Dataset
reference_data/datasets/REG_WEIGHT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact23230.00E+2passed
/selected_powertolerance00.00.00E+2passed
/at_boundaryexactTrueTruepassed
/r_squaredtolerance0.860280282520136890.86028028252013681.16E-1615passed
/adjusted_r_squaredtolerance0.846308310772150520.84630831077215051.72E-1817passed
/standard_errortolerance0.976741432337090210.97674143233709053.33E-1615passed
/f_valuetolerance61.57185957981366261.5718595798136263.57E-1415passed
/df1exact220.00E+2passed
/df2exact20200.00E+2passed
/p_valuetail_probability2.8351561812214512e-092.8351561812214553e-094.12E-2414passed
/grid/0/log_likelihoodtolerance-40.058970301800926-40.0589703018009265.06E-1818passed
/grid/0/weighted_ssetolerance595.09247849633539595.09247849633541.37E-1517passed
/grid/1/log_likelihoodtolerance-34.24487367233089-34.2448736723309041.40E-1415passed
/grid/1/weighted_ssetolerance97.44496192202137897.44496192202142.81E-1415passed
/grid/2/log_likelihoodtolerance-30.487057642566857-30.487057642566863.24E-1515passed
/grid/2/weighted_ssetolerance19.08047651287821419.0804765128782241.03E-1415passed
/coefficients/0/termexact(Intercept)(Intercept)passed
/coefficients/0/estimatetolerance2.03675762870101852.0367576287010153.55E-1514passed
/coefficients/0/standard_errortolerance0.49122940674120280.491229406741203133.34E-1615passed
/coefficients/0/t_valuetolerance4.14624531990621484.1462453199062059.76E-1514passed
/coefficients/0/p_valuetail_probability0.000499745753754653210.00049974575375466391.07E-1713passed
/coefficients/0/lowertolerance1.01207104201677361.01207104201676964.04E-1514passed
/coefficients/0/uppertolerance3.06144421538526323.06144421538526062.63E-1515passed
/coefficients/1/termexactx1x1passed
/coefficients/1/estimatetolerance0.718553837719815740.7185538377198162.18E-1615passed
/coefficients/1/standard_errortolerance0.100080556350239260.100080556350239315.08E-1715passed
/coefficients/1/t_valuetolerance7.17975462891297147.179754628912971.73E-1515passed
/coefficients/1/p_valuetail_probability5.9441133146721278e-075.94411331467215e-072.23E-2114passed
/coefficients/1/lowertolerance0.509789455391185030.50978945539118532.23E-1615passed
/coefficients/1/uppertolerance0.927318220048446440.92731822004844672.22E-1615passed
/coefficients/2/termexactx2x2passed
/coefficients/2/estimatetolerance-1.0607241722640557-1.06072417226405552.35E-1615passed
/coefficients/2/standard_errortolerance0.139457810398159850.139457810398159832.40E-1715passed
/coefficients/2/t_valuetolerance-7.6060578409744766-7.6060578409744773.07E-1717passed
/coefficients/2/p_valuetail_probability2.5204172062267379e-072.5204172062267374e-075.32E-2315passed
/coefficients/2/lowertolerance-1.3516280671903504-1.35162806719035021.76E-1615passed
/coefficients/2/uppertolerance-0.76982027733776093-0.76982027733776072.24E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

156  r <- profile(d, c("x1", "x2"), grid_of(-2, 0, 1), 0.95)
157  stopifnot(isTRUE(r$at_boundary))
158  emit("STAT-REG-WEIGHT-ESTIMATION-003", "axis_coverage",
159       list(response = "y", predictors = list("x1", "x2"), weight_variable = "z",
160            power_from = -2, power_to = 0, power_step = 1, confidence_level = 0.95),
161       outputs_for(r, c("(Intercept)", "x1", "x2")),
162       list("Grid -2 to 0 by 1: the likelihood keeps rising to the upper edge, so at_boundary is true and the edge power is used for the final fit."))

Whole script generators/r/regression_weight_estimation.R


STAT REG WEIGHT ESTIMATION 004

Case STAT-REG-WEIGHT-ESTIMATION-004 · status: passed

Reference

Evidence type
R
Source
R stats::lm with weights = z^-p, stats::qt/pt/pf
Citation
R Core Team, stats package help: lm, summary.lm, confint.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/regression_weight_estimation.R · sha256 8d75fc39f6126fa0
Dataset
reference_data/datasets/REG_WEIGHT_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact21210.00E+2passed
/selected_powertolerance1.51.50.00E+1passed
/at_boundaryexactFalseFalsepassed
/r_squaredtolerance0.908461686870389420.90846168687038962.20E-1615passed
/adjusted_r_squaredtolerance0.898290763189321550.89829076318932182.23E-1615passed
/standard_errortolerance0.358148576024816420.358148576024816152.75E-1615passed
/f_valuetolerance89.31948713383823789.31948713383841.56E-1314passed
/df1exact220.00E+2passed
/df2exact18180.00E+2passed
/p_valuetail_probability4.5125927859333113e-104.5125927859332255e-108.58E-2413passed
/grid/0/log_likelihoodtolerance-31.902550451695301-31.90255045169533.48E-1616passed
/grid/0/weighted_ssetolerance92.73783206622275192.737832066222678.54E-1415passed
/grid/1/log_likelihoodtolerance-29.878893590474561-29.878893590474563.93E-1616passed
/grid/1/weighted_ssetolerance40.23152052383027740.2315205238302846.76E-1515passed
/grid/2/log_likelihoodtolerance-28.362207649610589-28.36220764961061.07E-1415passed
/grid/2/weighted_ssetolerance18.31660461775934618.3166046177593641.75E-1415passed
/grid/3/log_likelihoodtolerance-27.358015154513701-27.35801515451373.61E-1616passed
/grid/3/weighted_ssetolerance8.75630601125549658.7563060112554964.24E-1717passed
/grid/4/log_likelihoodtolerance-26.860648097399881-26.860648097399883.51E-1616passed
/grid/4/weighted_ssetolerance4.39298671263474284.3929867126347432.97E-1717passed
/grid/5/log_likelihoodtolerance-26.851662094538995-26.8516620945389771.82E-1415passed
/grid/5/weighted_ssetolerance2.30886724515486692.30886724515486333.60E-1514passed
/grid/6/log_likelihoodtolerance-27.294686326020635-27.294686326020633.62E-1515passed
/grid/6/weighted_ssetolerance1.2668751059129511.26687510591295028.39E-1615passed
/coefficients/0/termexact(Intercept)(Intercept)passed
/coefficients/0/estimatetolerance1.90777291129299691.90777291129299693.53E-1716passed
/coefficients/0/standard_errortolerance0.413520234849840520.41352023484984032.24E-1615passed
/coefficients/0/t_valuetolerance4.61349348958886294.61349348958886552.61E-1515passed
/coefficients/0/p_valuetail_probability0.000215753638141196060.000215753638141194681.38E-1814passed
/coefficients/0/lowertolerance0.717479531004842070.71747953100484276.71E-1615passed
/coefficients/0/uppertolerance3.09806629158115143.0980662915811514.12E-1615passed
/coefficients/1/termexactx1x1passed
/coefficients/1/estimatetolerance0.802764712097635890.80276471209763572.21E-1615passed
/coefficients/1/standard_errortolerance0.0929069125544526720.09290691255445266.89E-1715passed
/coefficients/1/t_valuetolerance8.64052727645142768.6405272764514313.55E-1515passed
/coefficients/1/p_valuetail_probability8.0454982564817056e-088.045498256481657e-084.89E-2214passed
/coefficients/1/lowertolerance0.535337694803712740.53533769480371271.92E-1817passed
/coefficients/1/uppertolerance1.0701917293915591.07019172939155864.01E-1615passed
/coefficients/2/termexactx2x2passed
/coefficients/2/estimatetolerance-1.1339094985876883-1.1339094985876871.33E-1514passed
/coefficients/2/standard_errortolerance0.121904820223668110.1219048202236681.12E-1615passed
/coefficients/2/t_valuetolerance-9.3015969057434944-9.3015969057434931.77E-1515passed
/coefficients/2/p_valuetail_probability2.688861583363372e-082.6888615833633733e-081.29E-2315passed
/coefficients/2/lowertolerance-1.4848052669414185-1.48480526694141671.82E-1514passed
/coefficients/2/uppertolerance-0.78301373023395815-0.78301373023395721.00E-1514passed
/missing_countexact110.00E+2passed
/excluded_countexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

165  r <- profile(d[-c(3, 8), ], c("x1", "x2"), grid_of(-1, 2, 0.5), 0.99)
166  emit("STAT-REG-WEIGHT-ESTIMATION-004", "axis_coverage",
167       list(response = "y", predictors = list("x1", "x2"), weight_variable = "z",
168            power_from = -1, power_to = 2, power_step = 0.5, confidence_level = 0.99),
169       c(outputs_for(r, c("(Intercept)", "x1", "x2")),
170         list("/missing_count" = exact(1L, "one row has a missing x2"),
171              "/excluded_count" = exact(2L, "two rows excluded by identity"))),
172       list("Rows 2 and 7 (0-based) are excluded; the grid -1 to 2 by 0.5 and 99% intervals."),
173       excluded = list(2L, 7L))

Whole script generators/r/regression_weight_estimation.R


Parametric survival STAT-SURV-ALT-001

Case STAT-SURV-ALT-001 · status: passed

Reference

Evidence type
R
Source
R survival::survreg
Citation
R survival::survreg and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_parametric.R · sha256 7c99eb709767fd04
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_nexact32320.00E+2passed
/fits/[family=weibull]/median_usetolerance14.90189190382716714.9018919038299632.80E-1212passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

62  fit_aft <- survival::survreg(survival::Surv(time,status)~stress,data=d,dist="weibull",
63                               control=survival::survreg.control(maxiter=200L))
64  median50 <- exp(coef(fit_aft)[[1]]+coef(fit_aft)[[2]]*50)*(log(2))^fit_aft$scale
65  emit("STAT-SURV-ALT-001","tabelz.executor.reliability","acceleratedLifeMenuItem",
66       "SURV_CORE",list(mode="accelerated",time_column="time",event_column="status",
67                        stress_column="stress",stress_type="linear",stress_unit="native",
68                        use_stress=50,high_stress=80),
69       list("/total_n"=exact(nrow(d),"All finite stress observations"),
70            "/fits/[family=weibull]/median_use"=tol(median50,"Weibull median at stress 50")),
71       c("R Weibull survreg scale=1/shape; intercept+slope*stress is log scale.",
72         "Median at stress s is exp(beta0+beta1*s)*(log 2)^(survreg scale)."))

Whole script generators/r/survival_parametric.R


Survival option STAT-SURV-ALT-ARRHENIUS-001

Case STAT-SURV-ALT-ARRHENIUS-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_options.R · sha256 95af0dbeed660def
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[family=weibull]/median_usetolerance12.45271981199398512.4527198117597422.34E-1010passed

Convention

Reference code

Whole script generators/r/survival_options.R


Survival boundary STAT-SURV-ALT-CONSTANT-001

Case STAT-SURV-ALT-CONSTANT-001 · status: passed

Reference

Evidence type
R
Source
R survival and analytic boundary
Citation
R survival::survfit/coxph/survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_boundaries.R · sha256 528007972452fb15
Dataset
reference_data/datasets/SURV_CONST_STRESS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

78  emit("STAT-SURV-ALT-CONSTANT-001","tabelz.executor.reliability",
79       "acceleratedLifeMenuItem","SURV_CONST_STRESS",
80       list(mode="accelerated",time_column="time",event_column="status",stress_column="stress",
81            stress_type="linear",stress_unit="native",use_stress=40,high_stress=60),list(),
82       c("Constant stress cannot identify an accelerated-life stress slope."),
83       error=list(error_class="ReliabilityExecutionError",message_contains="Varying stress levels"))

Whole script generators/r/survival_boundaries.R


Hard-numeric survival STAT-SURV-ALT-HARD-001

Case STAT-SURV-ALT-HARD-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_hard.R · sha256 bc20ccd17540f4e7
Dataset
reference_data/datasets/SURV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[family=weibull]/median_usetolerance1002928.99463064831002928.99463056468.37E-813passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

75  fit_alt <- survival::survreg(survival::Surv(time,status)~stress,data=d,dist="weibull",
76                               control=survival::survreg.control(rel.tolerance=1e-14,maxiter=1000))
77  use_stress <- 102500
78  median_use <- exp(coef(fit_alt)[[1]]+coef(fit_alt)[[2]]*use_stress)*(log(2))^fit_alt$scale
79  emit("STAT-SURV-ALT-HARD-001","tabelz.executor.reliability","acceleratedLifeMenuItem",
80       "SURV_HARD",list(mode="accelerated",time_column="time",event_column="status",
81                        stress_column="stress",stress_type="linear",stress_unit="native",
82                        use_stress=use_stress,high_stress=104000),
83       list("/fits/[family=weibull]/median_use"=tol(median_use,"Large-offset linear-stress Weibull median")),
84       c("Stress is near 1e5 with a 3000-unit span; R survreg Weibull scale=1/shape."))

Whole script generators/r/survival_hard.R


Survival corrected option STAT-SURV-ALT-INVERSE-001

Case STAT-SURV-ALT-INVERSE-001 · status: passed

Reference

Evidence type
R
Source
R survival
Citation
R survival::survreg and coxph documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_corrections.R · sha256 69d4b35a58e7b5d7
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[family=weibull]/median_usetolerance15.44920097510062915.4492009750901361.05E-1112passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

23  x <- log(d$stress)
24  fit <- survival::survreg(survival::Surv(d$time,d$status)~x,dist="weibull")
25  median50 <- exp(coef(fit)[[1]]+coef(fit)[[2]]*log(50))*(log(2))^fit$scale
26  emit("STAT-SURV-ALT-INVERSE-001","tabelz.executor.reliability",
27       "acceleratedLifeMenuItem",
28       list(mode="accelerated",time_column="time",event_column="status",stress_column="stress",
29            stress_type="inverse_power",stress_unit="native",use_stress=50,high_stress=80,
30            confidence_level=0.9),
31       list("/fits/[family=weibull]/median_use"=tol(median50,"R inverse-power use-stress Weibull median")),
32       c("R stress covariate is log(S); Weibull survreg scale=1/shape.",
33         "Median=exp(beta0+beta1*log(50))*(log 2)^(survreg scale)."))

Whole script generators/r/survival_corrections.R


Survival option STAT-SURV-ALT-INVERSE-POWER-001

Case STAT-SURV-ALT-INVERSE-POWER-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_options.R · sha256 95af0dbeed660def
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[family=weibull]/median_usetolerance15.4492009750973515.4492009750901367.21E-1212passed

Convention

Reference code

Whole script generators/r/survival_options.R


Small-sample survival boundary STAT-SURV-ALT-MIN-001

Case STAT-SURV-ALT-MIN-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::survfit/coxph and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_minimum.R · sha256 2997d37929a6dae8
Dataset
reference_data/datasets/SURV_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_nexact550.00E+2passed
/failuresexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

58  emit("STAT-SURV-ALT-MIN-001","tabelz.executor.reliability","acceleratedLifeMenuItem",
59       "SURV_MIN",integer(),
60       list(mode="accelerated",time_column="time",event_column="status",stress_column="stress",
61            stress_type="linear",stress_unit="native",use_stress=40,high_stress=60),
62       list("/total_n"=exact(5L,"Minimum five stress/lifetime rows"),
63            "/failures"=exact(3L,"Minimum three failures")),
64       c("Five rows, three events and five stress levels meet physical ALT minimum."))

Whole script generators/r/survival_minimum.R


Survival population STAT-SURV-ALT-POP-001

Case STAT-SURV-ALT-POP-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::Surv/survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_population.R · sha256 3b3b4629e272fcd9
Dataset
reference_data/datasets/SURV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/input_row_countexact31310.00E+2passed
/omitted_row_countexact330.00E+2passed
/total_nexact28280.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

86  altdata <- eligible[stats::complete.cases(eligible[c("time","status","stress")]),,drop=FALSE]
87  emit("STAT-SURV-ALT-POP-001","tabelz.executor.reliability",
88       "acceleratedLifeMenuItem","SURV_POP",7L,
89       list(mode="accelerated",time_column="time",event_column="status",stress_column="stress",
90            stress_type="linear",stress_unit="native",use_stress=50,high_stress=80),
91       list("/input_row_count"=exact(nrow(eligible),"Rows sent to worker after exclusion"),
92            "/omitted_row_count"=exact(3L,"Missing time, status or stress"),
93            "/total_n"=exact(nrow(altdata),"Complete AFT rows")),
94       c("Physical ALT omits incomplete time/event/stress rows after exclusion."))

Whole script generators/r/survival_population.R


Survival STAT-SURV-COX-001

Case STAT-SURV-COX-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/survdiff/coxph and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, MASS 7.3.65, jsonlite 2.0.0
Reference script
generators/r/survival_core.R · sha256 1ed8945e6864eb84
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/log_hazardtolerance-0.021273688367235425-0.021273688367235423.44E-1815passed
/coefficients/0/hazard_ratiotolerance0.978951000400891110.97895100040089113.71E-1817passed
/event_countexact26260.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

55  i <- seq_len(32L)
56  raw <- data.frame(time=exp(2.2+0.025*i+0.18*sin(i*1.17)-0.31*cos(i*0.41)),
57                    status=as.integer(i%%5L!=0L),
58                    stress=as.double(rep(c(20,40,60,80),each=8L)),
59                    x=as.double(sin(i*0.39)+0.1*cos(i*1.21)))
60  export_dataset(raw,"SURV_CORE","Constructed censored duration and stress population",
61                 "R survival::coxph/survreg documentation",
62                 "Constructed validation data; no third-party license.",
63                 c(time="Float64",status="Int64",stress="Float64",x="Float64"))
64  d <- read_dataset("SURV_CORE")
65  cox <- survival::coxph(survival::Surv(time,status)~x,data=d,ties="efron",x=TRUE,
66                         control=survival::coxph.control(eps=1e-12,iter.max=200L))
67  emit("STAT-SURV-COX-001","tabelz.executor.cox_regression","coxRegressionMenuItem",
68       "SURV_CORE",list(duration_column="time",event_column="status",event_value="1",
69                        censored_value="0",predictors=list("x")),
70       list("/coefficients/0/log_hazard"=tol(coef(cox)[[1]],"Efron-tie log hazard", "1e-6"),
71            "/coefficients/0/hazard_ratio"=tol(exp(coef(cox)[[1]]),"Exponentiated Efron hazard ratio","1e-6"),
72            "/event_count"=exact(sum(d$status),"Observed failures")),
73       c("R coxph uses Efron ties, eps=1e-12 and iter.max=200; independent subjects."))

Whole script generators/r/survival_core.R


Survival option STAT-SURV-COX-CONFIDENCE-001

Case STAT-SURV-COX-CONFIDENCE-001 · status: passed

Reference

Evidence type
Analytical
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_options.R · sha256 95af0dbeed660def
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

40  emit("STAT-SURV-COX-CONFIDENCE-001","tabelz.executor.cox_regression",
41       "coxRegressionMenuItem","SURV_CORE",cox_config,structure(list(),names=character()),
42       c("Manual chapter 14: fixed 95% asymptotic intervals; a 90% request must be refused.",
43         "Until 2026-09-25 this case wrongly expected a 90% interval; -095-001 checks the 95% interval."),
44       error=list(error_class="ValidationError",message_contains="fixed at 95%"),kind="invalid_input")

Whole script generators/r/survival_options.R


Survival option STAT-SURV-COX-CONFIDENCE-095-001

Case STAT-SURV-COX-CONFIDENCE-095-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_options.R · sha256 95af0dbeed660def
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/hazard_ratio_ci_lowertolerance0.5475111266929340.54751112669298124.72E-1413passed
/analyzed_countexact32320.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

45  cox_config$confidence_level <- 0.95
46  emit("STAT-SURV-COX-CONFIDENCE-095-001","tabelz.executor.cox_regression",
47       "coxRegressionMenuItem","SURV_CORE",cox_config,
48       list("/coefficients/0/hazard_ratio_ci_lower"=tol(lower95,"95% Wald hazard-ratio lower"),
49            "/analyzed_count"=exact(nrow(d),"All finite rows")),
50       c("R coxph Efron ties; 95% Wald interval exp(beta-z_0.975*SE)."))

Whole script generators/r/survival_options.R


Hard-numeric survival STAT-SURV-COX-HARD-001

Case STAT-SURV-COX-HARD-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_hard.R · sha256 bc20ccd17540f4e7
Dataset
reference_data/datasets/SURV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/log_hazardtolerance-0.35940520742719329-0.35940520655245888.75E-108passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

63  cox <- survival::coxph(survival::Surv(time,status)~x,data=d,ties="efron")
64  emit("STAT-SURV-COX-HARD-001","tabelz.executor.cox_regression","coxRegressionMenuItem",
65       "SURV_HARD",list(duration_column="time",event_column="status",event_value="1",
66                        censored_value="0",predictors=list("x")),
67       list("/coefficients/0/log_hazard"=tol(coef(cox)[[1]],"Large-offset covariate Efron log hazard")),
68       c("Covariate is near 1e7 with O(1) spread; Cox partial likelihood is shift-invariant."))

Whole script generators/r/survival_hard.R


Invalid survival request STAT-SURV-COX-INVALID-001

Case STAT-SURV-COX-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid survival request STAT-SURV-COX-INVALID-002

Case STAT-SURV-COX-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Small-sample survival boundary STAT-SURV-COX-MIN-001

Case STAT-SURV-COX-MIN-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::survfit/coxph and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_minimum.R · sha256 2997d37929a6dae8
Dataset
reference_data/datasets/SURV_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/analyzed_countexact550.00E+2passed
/coefficients/0/log_hazardtolerance-0.59193918723529437-0.59193918723928713.99E-1211passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

42  cox <- survival::coxph(survival::Surv(time,status)~x,data=d,ties="efron")
43  emit("STAT-SURV-COX-MIN-001","tabelz.executor.cox_regression","coxRegressionMenuItem",
44       "SURV_MIN",integer(),
45       list(duration_column="time",event_column="status",event_value="1",censored_value="0",
46            predictors=list("x")),
47       list("/analyzed_count"=exact(nrow(d),"Five identifiable rows"),
48            "/coefficients/0/log_hazard"=tol(coef(cox)[[1]],"R Efron small-sample log hazard")),
49       c("Five rows with three failures avoid complete separation; Efron ties pinned."))

Whole script generators/r/survival_minimum.R


Survival corrected option STAT-SURV-COX-OPTIONS-001

Case STAT-SURV-COX-OPTIONS-001 · status: passed

Reference

Evidence type
R
Source
R survival
Citation
R survival::survreg and coxph documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_corrections.R · sha256 69d4b35a58e7b5d7
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/log_hazardtolerance-0.02127368836732162-0.021273688367235428.62E-1411passed
/analyzed_countexact32320.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

35  cox <- survival::coxph(survival::Surv(time,status)~x,data=d,ties="efron",
36                         control=survival::coxph.control(iter.max=300L))
37  emit("STAT-SURV-COX-OPTIONS-001","tabelz.executor.cox_regression",
38       "coxRegressionMenuItem",
39       list(duration_column="time",event_column="status",event_value="1",censored_value="0",
40            predictors=list("x"),confidence_level=0.95,max_iterations=300L,max_diagnostic_rows=100L),
41       list("/coefficients/0/log_hazard"=tol(coef(cox)[[1]],"R Efron Cox log hazard"),
42            "/analyzed_count"=exact(nrow(d),"All finite rows")),
43       c("Cox intervals are fixed at 95% by the manual; max_iterations and diagnostic bound are nondefaults."))

Whole script generators/r/survival_corrections.R


Survival population STAT-SURV-COX-POP-001

Case STAT-SURV-COX-POP-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::Surv/survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_population.R · sha256 3b3b4629e272fcd9
Dataset
reference_data/datasets/SURV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact32320.00E+2passed
/excluded_countexact110.00E+2passed
/missing_countexact330.00E+2passed
/coefficients/0/log_hazardtolerance-0.25036455063421509-0.250364550634234241.91E-1413passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

64  coxdata <- eligible[stats::complete.cases(eligible[c("time","status","x")]),,drop=FALSE]
65  cox <- survival::coxph(survival::Surv(time,status)~x,data=coxdata,ties="efron")
66  emit("STAT-SURV-COX-POP-001","tabelz.executor.cox_regression","coxRegressionMenuItem",
67       "SURV_POP",7L,list(duration_column="time",event_column="status",event_value="1",
68                          censored_value="0",predictors=list("x")),
69       list("/source_count"=exact(nrow(d),"CSV population"),
70            "/excluded_count"=exact(1L,"One source row excluded"),
71            "/missing_count"=exact(3L,"Missing time, status, and predictor rows"),
72            "/coefficients/0/log_hazard"=tol(coef(cox)[[1]],"Efron-tie complete-case log hazard","1e-6")),
73       c("R coxph Efron fit uses the same complete predictor/time/event rows."))

Whole script generators/r/survival_population.R


Survival boundary STAT-SURV-COX-TIES-001

Case STAT-SURV-COX-TIES-001 · status: passed

Reference

Evidence type
R
Source
R survival and analytic boundary
Citation
R survival::survfit/coxph/survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_boundaries.R · sha256 528007972452fb15
Dataset
reference_data/datasets/SURV_TIES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/log_hazardtolerance-0.02018576583744534-0.0201857658373860855.93E-1411passed
/event_countexact26260.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

66  cox <- survival::coxph(survival::Surv(time,status)~x,data=ties,ties="efron")
67  emit("STAT-SURV-COX-TIES-001","tabelz.executor.cox_regression","coxRegressionMenuItem",
68       "SURV_TIES",list(duration_column="time",event_column="status",event_value="1",
69                         censored_value="0",predictors=list("x")),
70       list("/coefficients/0/log_hazard"=tol(coef(cox)[[1]],"R Efron tied-event log hazard"),
71            "/event_count"=exact(as.integer(sum(ties$status)),"All failures retained")),
72       c("Integer-rounded durations create exact ties; R coxph ties=efron matches manual."))

Whole script generators/r/survival_boundaries.R


Counting-process Cox model with Efron ties

Case STAT-SURV-COXTD-001 · status: passed

Reference

Evidence type
R
Source
R survival::coxph (counting process, cluster, strata, tt) and cox.zph
Citation
R survival package documentation: coxph, coxph.control, cox.zph; Therneau and Grambsch (2000).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_cox_time_dependent.R · sha256 4b0d7390bdbbf36e
Dataset
reference_data/datasets/SURV_COXTD_EPISODES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/coefficienttolerance0.986763836851640370.98676383685164161.23E-1514passed
/coefficients/0/standard_errortolerance0.524018792009535160.52401879200953531.12E-1615passed
/coefficients/0/ztolerance1.88306956143222641.88306956143222811.75E-1515passed
/coefficients/0/p_valuetail_probability0.0596909416113026930.059690941611302492.01E-1614passed
/coefficients/0/hazard_ratiotolerance2.68253927565236162.68253927565236473.12E-1514passed
/coefficients/0/lowertolerance0.960506890713469510.96050689071347059.97E-1614passed
/coefficients/0/uppertolerance7.49189520136837137.4918952013683819.75E-1514passed
/coefficients/1/coefficienttolerance1.15516981685471421.15516981685471551.34E-1514passed
/coefficients/1/standard_errortolerance0.615308029274176490.61530802927417661.14E-1615passed
/coefficients/1/ztolerance1.87738459746310161.87738459746310341.78E-1515passed
/coefficients/1/p_valuetail_probability0.0604654009779486480.060465400977948412.36E-1614passed
/coefficients/1/hazard_ratiotolerance3.17456246596961393.17456246596961834.43E-1514passed
/coefficients/1/lowertolerance0.950457567901749840.9504575679017511.11E-1514passed
/coefficients/1/uppertolerance10.60315282942208810.6031528294221021.42E-1414passed
/tests/0/chisqtolerance6.05164642172472616.0516464217247197.10E-1514passed
/tests/0/dfexact220.00E+2passed
/tests/0/p_valuetail_probability0.0485178643928119640.048517864392812121.53E-1614passed
/tests/1/chisqtolerance5.90422826669447255.9042282666944841.16E-1414passed
/log_likelihoodtolerance-23.600482497165643-23.6004824971656432.81E-1616passed
/null_log_likelihoodtolerance-26.626305708028006-26.6263057080280033.22E-1515passed
/ph_test/statusexactcomputedcomputedpassed
/ph_test/rows/0/chisqtolerance0.465806029728458360.46580602972845758.33E-1614passed
/ph_test/rows/0/dfexact110.00E+2passed
/ph_test/rows/0/p_valuetail_probability0.494922944777157040.49492294477715754.41E-1615passed
/ph_test/rows/1/chisqtolerance0.0146000371023622870.0146000371023623728.48E-1714passed
/ph_test/rows/1/dfexact110.00E+2passed
/ph_test/rows/1/p_valuetail_probability0.903825201867918590.90382520186791833.30E-1615passed
/ph_test/rows/2/chisqtolerance0.491833516502980630.49183351650297988.35E-1614passed
/ph_test/rows/2/dfexact220.00E+2passed
/ph_test/rows/2/p_valuetail_probability0.781987316207797160.78198731620779753.29E-1615passed
/subjectsexact16160.00E+2passed
/episodesexact28280.00E+2passed
/eventsexact13130.00E+2passed
/gapsexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

140  fit <- survival::coxph(survival::Surv(start, stop, status) ~ x + trt, data = d, ties = "efron", control = control)
141  emit("STAT-SURV-COXTD-001", "Counting-process Cox model with Efron ties", "nominal",
142       "SURV_COXTD_EPISODES", episode_config,
143       c(coefficient_outputs(fit, 0.95), zph_outputs(fit), counts(d)),
144       list("coxph(Surv(start, stop, status) ~ x + trt, ties = 'efron'); terms in order x, trt = B (vs A).",
145            "cox.zph(transform = 'km', terms = TRUE): rows x, trt, GLOBAL."))

Whole script generators/r/survival_cox_time_dependent.R


Breslow ties with the robust sandwich variance

Case STAT-SURV-COXTD-002 · status: passed

Reference

Evidence type
R
Source
R survival::coxph (counting process, cluster, strata, tt) and cox.zph
Citation
R survival package documentation: coxph, coxph.control, cox.zph; Therneau and Grambsch (2000).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_cox_time_dependent.R · sha256 4b0d7390bdbbf36e
Dataset
reference_data/datasets/SURV_COXTD_EPISODES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/coefficienttolerance0.860131739818786210.86013173981878642.24E-1615passed
/coefficients/0/standard_errortolerance0.498232121712455730.49823212171245561.08E-1615passed
/coefficients/0/robust_standard_errortolerance0.371122637295155160.3711226372951551.71E-1615passed
/coefficients/0/ztolerance2.31764827413295292.31764827413295471.81E-1515passed
/coefficients/0/p_valuetail_probability0.0204684443373109560.020468444337310878.70E-1714passed
/coefficients/0/hazard_ratiotolerance2.36347203657502282.36347203657502324.50E-1615passed
/coefficients/0/lowertolerance1.14195846123514971.14195846123515036.24E-1615passed
/coefficients/0/uppertolerance4.89159654864348694.8915965486434871.92E-1717passed
/coefficients/1/coefficienttolerance1.07183999275031131.07183999275031174.21E-1615passed
/coefficients/1/standard_errortolerance0.619115363828169230.61911536382816924.79E-1918passed
/coefficients/1/robust_standard_errortolerance0.469431547046817550.46943154704681782.26E-1615passed
/coefficients/1/ztolerance2.28327218205344452.28327218205344451.01E-1717passed
/coefficients/1/p_valuetail_probability0.0224143405622359740.022414340562235987.34E-1815passed
/coefficients/1/hazard_ratiotolerance2.92074871526514062.9207487152651421.30E-1515passed
/coefficients/1/lowertolerance1.16389375262736271.16389375262736272.10E-1716passed
/coefficients/1/uppertolerance7.32951185489713367.329511854897146.26E-1515passed
/tests/0/chisqtolerance4.95432543685536334.9543254368553567.10E-1514passed
/tests/0/dfexact220.00E+2passed
/tests/0/p_valuetail_probability0.0839811660954560970.08398116609545643.06E-1614passed
/tests/1/chisqtolerance9.16633279355497529.166332793554985.34E-1515passed
/log_likelihoodtolerance-25.50863915756549-25.5086391575654863.50E-1515passed
/null_log_likelihoodtolerance-27.985801875993172-27.9858018759931657.40E-1515passed
/subjectsexact16160.00E+2passed
/episodesexact28280.00E+2passed
/eventsexact13130.00E+2passed
/gapsexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

148  fit2 <- survival::coxph(survival::Surv(start, stop, status) ~ x + trt, data = d, ties = "breslow",
149                          cluster = id, control = control)
150  emit("STAT-SURV-COXTD-002", "Breslow ties with the robust sandwich variance", "axis_coverage",
151       "SURV_COXTD_EPISODES", c(episode_config, list(ties = "breslow", robust = TRUE)),
152       c(coefficient_outputs(fit2, 0.95, robust = TRUE), counts(d)),
153       list("coxph(..., ties = 'breslow', cluster = id): fit$var is the robust sandwich with the subject as the cluster, fit$naive.var the model-based covariance; z, p, the interval and the Wald test use the robust SE."))

Whole script generators/r/survival_cox_time_dependent.R


Stratified baseline hazards by sex with 90% intervals

Case STAT-SURV-COXTD-003 · status: passed

Reference

Evidence type
R
Source
R survival::coxph (counting process, cluster, strata, tt) and cox.zph
Citation
R survival package documentation: coxph, coxph.control, cox.zph; Therneau and Grambsch (2000).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_cox_time_dependent.R · sha256 4b0d7390bdbbf36e
Dataset
reference_data/datasets/SURV_COXTD_EPISODES.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/coefficienttolerance0.936297318369149760.93629731836915383.99E-1514passed
/coefficients/0/standard_errortolerance0.54590315777142240.54590315777142316.64E-1614passed
/coefficients/0/ztolerance1.71513446119538851.71513446119539385.31E-1514passed
/coefficients/0/p_valuetail_probability0.0863205774319589780.086320577431958029.58E-1613passed
/coefficients/0/hazard_ratiotolerance2.55052014899976152.55052014899977181.03E-1414passed
/coefficients/0/lowertolerance1.03911202813851841.03911202813852153.15E-1514passed
/coefficients/0/uppertolerance6.26030000067191856.26030000067195053.20E-1414passed
/coefficients/1/coefficienttolerance1.34594607603390041.34594607603390666.19E-1514passed
/coefficients/1/standard_errortolerance0.75951248639401880.75951248639401991.11E-1514passed
/coefficients/1/ztolerance1.77211843142187961.77211843142188545.81E-1514passed
/coefficients/1/p_valuetail_probability0.0763748999821170640.076374899982116129.43E-1613passed
/coefficients/1/hazard_ratiotolerance3.84181947514482363.84181947514484762.40E-1414passed
/coefficients/1/lowertolerance1.10148493247913451.10148493247913944.87E-1514passed
/coefficients/1/uppertolerance13.3997083794713113.3997083794714161.06E-1314passed
/tests/0/chisqtolerance5.67025559803450825.6702555980345123.53E-1515passed
/tests/0/dfexact220.00E+2passed
/tests/0/p_valuetail_probability0.0587110221434684470.058711022143468331.18E-1614passed
/tests/1/chisqtolerance4.90876850534330344.908768505343332.67E-1414passed
/log_likelihoodtolerance-15.378345732919254-15.3783457329192522.11E-1515passed
/null_log_likelihoodtolerance-18.213473531936508-18.2134735319365082.46E-1616passed
/ph_test/statusexactcomputedcomputedpassed
/ph_test/rows/0/chisqtolerance0.391496132064136420.39149613206414579.27E-1513passed
/ph_test/rows/0/dfexact110.00E+2passed
/ph_test/rows/0/p_valuetail_probability0.53151401787937880.53151401787937375.11E-1514passed
/ph_test/rows/1/chisqtolerance0.772624925585925880.77262492558593529.33E-1513passed
/ph_test/rows/1/dfexact110.00E+2passed
/ph_test/rows/1/p_valuetail_probability0.37940623357390890.37940623357390612.77E-1514passed
/ph_test/rows/2/chisqtolerance1.37780068068008221.37780068068010552.33E-1413passed
/ph_test/rows/2/dfexact220.00E+2passed
/ph_test/rows/2/p_valuetail_probability0.502127935412575610.50212793541256966.00E-1513passed
/subjectsexact16160.00E+2passed
/episodesexact28280.00E+2passed
/eventsexact13130.00E+2passed
/gapsexact110.00E+2passed
/strata_levels/0exactFFpassed
/strata_levels/1exactMMpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

156  fit3 <- survival::coxph(survival::Surv(start, stop, status) ~ x + trt + survival::strata(sex), data = d,
157                          ties = "efron", control = control)
158  emit("STAT-SURV-COXTD-003", "Stratified baseline hazards by sex with 90% intervals", "axis_coverage",
159       "SURV_COXTD_EPISODES", c(episode_config, list(strata = "sex", confidence_level = 0.90)),
160       c(coefficient_outputs(fit3, 0.90), zph_outputs(fit3), counts(d),
161         list("/strata_levels/0" = exact("F", "strata levels sorted naturally"),
162              "/strata_levels/1" = exact("M", "strata levels sorted naturally"))),
163       list("coxph(... + strata(sex)): separate baseline hazards, common coefficients; 90% Wald limits on the log scale."))

Whole script generators/r/survival_cox_time_dependent.R


One row per subject with the covariate-by-time term x t

Case STAT-SURV-COXTD-004 · status: passed

Reference

Evidence type
R
Source
R survival::coxph (counting process, cluster, strata, tt) and cox.zph
Citation
R survival package documentation: coxph, coxph.control, cox.zph; Therneau and Grambsch (2000).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_cox_time_dependent.R · sha256 4b0d7390bdbbf36e
Dataset
reference_data/datasets/SURV_COXTD_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/coefficienttolerance0.375022258063264840.37502225806422959.65E-1311passed
/coefficients/0/standard_errortolerance2.32421777254627852.3242177725463719.24E-1413passed
/coefficients/0/ztolerance0.161354182251352530.161354182251761184.09E-1311passed
/coefficients/0/p_valuetail_probability0.871814451010288540.87181445100996673.22E-1312passed
/coefficients/0/hazard_ratiotolerance1.45502380026957721.4550238002709811.40E-1212passed
/coefficients/0/lowertolerance0.0152929878154723120.0152929878154843061.20E-1412passed
/coefficients/0/uppertolerance138.43562061881735138.435620618975861.59E-1011passed
/coefficients/1/coefficienttolerance1.28659321248013741.2865932124800696.84E-1413passed
/coefficients/1/standard_errortolerance0.635337455675808020.63533745567580295.11E-1514passed
/coefficients/1/ztolerance2.02505487593453642.0250548759344459.15E-1413passed
/coefficients/1/p_valuetail_probability0.0428617296884440110.0428617296884534149.40E-1512passed
/coefficients/1/hazard_ratiotolerance3.62043148135859343.62043148135834562.48E-1313passed
/coefficients/1/lowertolerance1.04222169657166461.04222169657160386.08E-1413passed
/coefficients/1/uppertolerance12.57652201477758112.5765220147765939.88E-1313passed
/coefficients/2/coefficienttolerance0.106088982285940280.106088982285829041.11E-1311passed
/coefficients/2/standard_errortolerance0.313431906992543970.313431906992553859.88E-1513passed
/coefficients/2/ztolerance0.338475375094673960.338475375094308363.66E-1311passed
/coefficients/2/p_valuetail_probability0.735004980990006110.73500498099028162.75E-1312passed
/coefficients/2/hazard_ratiotolerance1.11192081336040931.11192081336028561.24E-1312passed
/coefficients/2/lowertolerance0.601561642798579820.60156164279850127.86E-1412passed
/coefficients/2/uppertolerance2.05526384533470982.0552638453345211.89E-1313passed
/tests/0/chisqtolerance7.2961446068012357.2961446068012427.11E-1515passed
/tests/0/dfexact330.00E+2passed
/tests/0/p_valuetail_probability0.0630343366861025320.063034336686102322.08E-1614passed
/tests/1/chisqtolerance6.97000023290335956.9700002329040867.26E-1312passed
/log_likelihoodtolerance-23.111764797251904-23.1117647972519044.69E-1616passed
/null_log_likelihoodtolerance-26.759837100652522-26.7598371006525263.52E-1515passed
/ph_test/statusexactwithheldwithheldpassed
/split/cut_countexact660.00E+2passed
/split/functionexacttimetimepassed
/subjectsexact16160.00E+2passed
/eventsexact13130.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

166  fit4 <- survival::coxph(survival::Surv(time, status) ~ x + trt + tt(x), data = d1, ties = "efron",
167                          tt = function(x, t, ...) x * t, control = control)
168  one_config <- list(layout = "one_row", stop = "time", status = "status", event_value = "1",
169                     censored_value = "0", numeric_covariates = list("x"),
170                     categorical_covariates = list("trt"))
171  cuts <- length(unique(d1$time[d1$status == 1L]))
172  emit("STAT-SURV-COXTD-004", "One row per subject with the covariate-by-time term x t", "axis_coverage",
173       "SURV_COXTD_ONE", c(one_config, list(time_function = "time", time_covariates = list("x"))),
174       c(coefficient_outputs(fit4, 0.95, names = c("x", "trt = B (vs A)", "x * t")),
175         list("/ph_test/status" = exact("withheld", "the model already has a declared time interaction"),
176              "/split/cut_count" = exact(cuts, "distinct event times"),
177              "/split/function" = exact("time", "g(t) = t"),
178              "/subjects" = exact(nrow(d1), "16 subjects"),
179              "/events" = exact(as.integer(sum(d1$status)), "events"))),
180       list("coxph(Surv(time, status) ~ x + trt + tt(x), tt = function(x, t, ...) x * t): survival expands the data at each event time, the same expansion the product performs; terms x, trt = B (vs A), x x t."))

Whole script generators/r/survival_cox_time_dependent.R


One row per subject with the step function x 1{t > 8}

Case STAT-SURV-COXTD-005 · status: passed

Reference

Evidence type
R
Source
R survival::coxph (counting process, cluster, strata, tt) and cox.zph
Citation
R survival package documentation: coxph, coxph.control, cox.zph; Therneau and Grambsch (2000).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_cox_time_dependent.R · sha256 4b0d7390bdbbf36e
Dataset
reference_data/datasets/SURV_COXTD_ONE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/coefficienttolerance1.22469075363569191.22469075363769762.01E-1211passed
/coefficients/0/standard_errortolerance0.682502616599516390.68250261659972512.09E-1312passed
/coefficients/0/ztolerance1.79441180714816871.79441180715055882.39E-1211passed
/coefficients/0/p_valuetail_probability0.0727474637419140840.072747463741532863.81E-1311passed
/coefficients/0/hazard_ratiotolerance3.40311351956527863.40311351957210436.83E-1211passed
/coefficients/0/lowertolerance0.893159775431566660.89315977543299291.43E-1211passed
/coefficients/0/uppertolerance12.96652843714558912.9665284371769033.13E-1111passed
/coefficients/1/coefficienttolerance1.23474109972781341.23474109972793821.25E-1312passed
/coefficients/1/standard_errortolerance0.623652544784898950.62365254478489781.11E-1514passed
/coefficients/1/ztolerance1.97985418331561871.97985418331582232.04E-1312passed
/coefficients/1/p_valuetail_probability0.0477199159274245350.0477199159274016852.29E-1412passed
/coefficients/1/hazard_ratiotolerance3.43748843879642953.43748843879685854.29E-1312passed
/coefficients/1/lowertolerance1.01248182891071091.01248182891083931.28E-1312passed
/coefficients/1/uppertolerance11.67065563988623311.6706556398876681.44E-1212passed
/coefficients/2/coefficienttolerance-0.21624794810601-0.216247948107935271.93E-1211passed
/coefficients/2/standard_errortolerance1.09106987440494921.09106987440507751.28E-1312passed
/coefficients/2/ztolerance-0.19819807436617928-0.198198074367920531.74E-1211passed
/coefficients/2/p_valuetail_probability0.842890094035971480.84289009403460911.36E-1211passed
/coefficients/2/hazard_ratiotolerance0.80553554609699130.80553554609544041.55E-1211passed
/coefficients/2/lowertolerance0.0949214466732868770.094921446673080242.07E-1311passed
/coefficients/2/uppertolerance6.83604747680684186.8360474767954011.14E-1111passed
/tests/0/chisqtolerance7.22232728715559357.2223272871556017.07E-1515passed
/tests/0/dfexact330.00E+2passed
/tests/0/p_valuetail_probability0.0651391235642586330.06513912356425842.36E-1614passed
/tests/1/chisqtolerance6.99086668512470416.99086668513321558.51E-1211passed
/log_likelihoodtolerance-23.148673457074725-23.1486734570747252.35E-1616passed
/null_log_likelihoodtolerance-26.759837100652522-26.7598371006525263.52E-1515passed
/ph_test/statusexactwithheldwithheldpassed
/split/cut_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

183  fit5 <- survival::coxph(survival::Surv(time, status) ~ x + trt + tt(x), data = d1, ties = "efron",
184                          tt = function(x, t, ...) x * (t > 8), control = control)
185  emit("STAT-SURV-COXTD-005", "One row per subject with the step function x 1{t > 8}", "axis_coverage",
186       "SURV_COXTD_ONE", c(one_config, list(time_function = "step", time_covariates = list("x"), step_time = 8)),
187       c(coefficient_outputs(fit5, 0.95, names = c("x", "trt = B (vs A)", "x * 1{t > 8}")),
188         list("/ph_test/status" = exact("withheld", "the model already has a declared time interaction"),
189              "/split/cut_count" = exact(1L, "one cut at the step time"))),
190       list("coxph(..., tt = function(x, t, ...) x * (t > 8)): the step is in force strictly after 8, so an event at exactly 8 uses the pre-step value; the product splits each episode once at the step time."))

Whole script generators/r/survival_cox_time_dependent.R


Degradation-based lifetime: repeated measurements, lognormal rates, no temperature axis

Case STAT-SURV-DEG-001 · status: imprecise_reference

Reference

Evidence type
R
Source
R survival::survreg lognormal pseudo-failure fits and base R delta-method algebra
Citation
Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, chapter 21 (pseudo-failure limits); Therneau T (2024), A Package for Survival Analysis in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_degradation.R · sha256 b4e6ab8a59ddbd26
Dataset
reference_data/datasets/SURV_DEG_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/distancetolerance2525.00.00E+2passed
/unit_countexact10100.00E+2passed
/analyzed_countexact40400.00E+2passed
/source_countexact40400.00E+2passed
/excluded_countexact000.00E+2passed
/unobserved_countexact000.00E+2passed
/parameters/[name=beta0]/estimatetolerance0.170006365876530290.17000636590860546imprecise_reference
/parameters/[name=beta0]/standard_errortolerance0.191495456781113740.19149545773885918imprecise_reference
/parameters/[name=beta0]/lowertolerance-0.20531783261749903-0.20531783446257051imprecise_reference
/parameters/[name=beta0]/uppertolerance0.54533056437055960.5453305662797814imprecise_reference
/parameters/[name=tau]/estimatetolerance0.605561805002655330.6055618052716375imprecise_reference
/parameters/[name=tau]/standard_errortolerance0.223606797747812840.2236067996189082imprecise_reference
/parameters/[name=tau]/lowertolerance0.390682558803125580.3906825575439193imprecise_reference
/parameters/[name=tau]/uppertolerance0.938626748021442970.9386267518805747imprecise_reference
/parameters/[name=sigma_e]/estimatetolerance0.00109711630085728030.0010971162997829417imprecise_reference
/lifetime/distancetolerance2525.00.00E+2passed
/lifetime/eta_usetolerance0.170006365876530290.17000636590860546imprecise_reference
/lifetime/tautolerance0.605561805002655330.6055618052716375imprecise_reference
/lifetime/median/probabilitytolerance0.50.50.00E+1passed
/lifetime/median/estimatetolerance21.09148614868556521.09148614800905imprecise_reference
/lifetime/median/lowertolerance14.49125349956726114.491253471900238imprecise_reference
/lifetime/median/uppertolerance30.69788186187476630.697881918514557imprecise_reference
/lifetime/quantiles/0/probabilitytolerance0.100000000000000010.14.45E-1816passed
/lifetime/quantiles/0/estimatetolerance9.70664574975561589.706645746098252imprecise_reference
/lifetime/quantiles/0/lowertolerance5.84920429813359635.849204293392352imprecise_reference
/lifetime/quantiles/0/uppertolerance16.10799809835891616.107998099277065imprecise_reference
/lifetime/quantiles/1/probabilitytolerance0.50.50.00E+1passed
/lifetime/quantiles/1/estimatetolerance21.09148614868556521.09148614800905imprecise_reference
/lifetime/quantiles/1/lowertolerance14.49125349956726114.491253471900238imprecise_reference
/lifetime/quantiles/1/uppertolerance30.69788186187476630.697881918514557imprecise_reference
/lifetime/quantiles/2/probabilitytolerance0.900000000000000020.92.20E-1817passed
/lifetime/quantiles/2/estimatetolerance45.82950685837020945.82950687269831imprecise_reference
/lifetime/quantiles/2/lowertolerance27.61676437033578327.616764202961203imprecise_reference
/lifetime/quantiles/2/uppertolerance76.05321429824925976.05321480673297imprecise_reference
/lifetime/reliability/0/timetolerance1010.00.00E+2passed
/lifetime/reliability/0/reliabilitytolerance0.891097065932296320.8910970658202142imprecise_reference
/lifetime/reliability/0/lowertolerance0.627256814431365380.6272568141643394imprecise_reference
/lifetime/reliability/0/uppertolerance0.975482053476199980.9754820534482685imprecise_reference
/lifetime/reliability/1/timetolerance1515.00.00E+2passed
/lifetime/reliability/1/reliabilitytolerance0.713219562188659670.7132195620854989imprecise_reference
/lifetime/reliability/1/lowertolerance0.450302181334833520.45030218095934355imprecise_reference
/lifetime/reliability/1/uppertolerance0.883045800548309170.8830458006007963imprecise_reference
/lifetime/reliability/2/timetolerance2020.00.00E+2passed
/lifetime/reliability/2/reliabilitytolerance0.534961744370562760.5349617443340232imprecise_reference
/lifetime/reliability/2/lowertolerance0.299030528018476850.29903052718979084imprecise_reference
/lifetime/reliability/2/uppertolerance0.756220671147275540.7562206718219427imprecise_reference
/units/0/readingsexact440.00E+2passed
/units/0/distinct_positive_timesexact440.00E+2passed
/units/0/least_squares_ratetolerance0.756240222533499760.75624022253349981.86E-1817passed
/units/0/empirical_bayes_ratetolerance0.756240222533499760.7562402265972118imprecise_reference
/units/1/readingsexact440.00E+2passed
/units/1/distinct_positive_timesexact440.00E+2passed
/units/1/least_squares_ratetolerance0.763892431363699110.76389243136369917.52E-1918passed
/units/1/empirical_bayes_ratetolerance0.763892431363699110.7638924352965764imprecise_reference
/units/2/readingsexact440.00E+2passed
/units/2/distinct_positive_timesexact440.00E+2passed
/units/2/least_squares_ratetolerance0.476857291506800950.476857291506800951.01E-1817passed
/units/2/empirical_bayes_ratetolerance0.476857291506800950.47685730456429426imprecise_reference
/units/3/readingsexact440.00E+2passed
/units/3/distinct_positive_timesexact440.00E+2passed
/units/3/least_squares_ratetolerance0.572387021596267750.5723870215962682.25E-1615passed
/units/3/empirical_bayes_ratetolerance0.572387021596267750.5723870302930119imprecise_reference
/units/4/readingsexact440.00E+2passed
/units/4/distinct_positive_timesexact440.00E+2passed
/units/4/least_squares_ratetolerance1.20774374697379041.20774374697379022.33E-1615passed
/units/4/empirical_bayes_ratetolerance1.20774374697379041.2077437468676409imprecise_reference
/units/5/readingsexact440.00E+2passed
/units/5/distinct_positive_timesexact440.00E+2passed
/units/5/least_squares_ratetolerance1.96062902376928251.96062902376928252.63E-1817passed
/units/5/empirical_bayes_ratetolerance1.96062902376928251.960629022014015imprecise_reference
/units/6/readingsexact440.00E+2passed
/units/6/distinct_positive_timesexact440.00E+2passed
/units/6/least_squares_ratetolerance1.99046968044317051.99046968044317072.41E-1615passed
/units/6/empirical_bayes_ratetolerance1.99046968044317051.9904696786623235imprecise_reference
/units/7/readingsexact440.00E+2passed
/units/7/distinct_positive_timesexact440.00E+2passed
/units/7/least_squares_ratetolerance3.23461838643439363.23461838643439361.12E-1717passed
/units/7/empirical_bayes_ratetolerance3.23461838643439363.2346183843120464imprecise_reference
/units/8/readingsexact440.00E+2passed
/units/8/distinct_positive_timesexact440.00E+2passed
/units/8/least_squares_ratetolerance2.17360116421211342.17360116421211344.31E-1716passed
/units/8/empirical_bayes_ratetolerance2.17360116421211342.1736011623044074imprecise_reference
/units/9/readingsexact440.00E+2passed
/units/9/distinct_positive_timesexact440.00E+2passed
/units/9/least_squares_ratetolerance1.04768175284561351.04768175284561351.02E-1717passed
/units/9/empirical_bayes_ratetolerance1.04768175284561351.0476817536512257imprecise_reference
/readings/0/timetolerance44.00.00E+2passed
/readings/0/responsetolerance3.02633160205862423.0263316020586242.94E-1717passed
/readings/0/fittedtolerance3.0249608901339993.024960906388847imprecise_reference
/readings/0/residualtolerance0.00137071192462512310.001370695669776989imprecise_reference
/readings/1/timetolerance88.00.00E+2passed
/readings/1/responsetolerance6.05181797477948546.0518179747794854.60E-1717passed
/readings/1/fittedtolerance6.04992178026799816.049921812777694imprecise_reference
/readings/1/residualtolerance0.00189619451148725910.0018961620017909908imprecise_reference
/readings/2/timetolerance1212.00.00E+2passed
/readings/2/responsetolerance9.07480341647430329.0748034164743031.93E-1717passed
/readings/2/fittedtolerance9.07488267040199769.074882719166542imprecise_reference
/readings/2/residualtolerance-7.9253927694367121e-05-7.930269223876962e-05imprecise_reference
/readings/3/timetolerance1616.00.00E+2passed
/readings/3/responsetolerance12.09861222574486712.0986122257448673.88E-1616passed
/readings/3/fittedtolerance12.09984356053599612.099843625555389imprecise_reference
/readings/3/residualtolerance-0.001231334791128802-0.0012313998105213386imprecise_reference
/readings/4/timetolerance44.00.00E+2passed
/readings/4/responsetolerance3.05588588011175413.0558858801117544.86E-1817passed
/readings/4/fittedtolerance3.05556972545479643.0555697411863054imprecise_reference
/readings/4/residualtolerance0.000316154656957667870.0003161389254486835imprecise_reference
/readings/5/timetolerance88.00.00E+2passed
/readings/5/responsetolerance6.11108513702900966.111085137029012.70E-1717passed
/readings/5/fittedtolerance6.11113945090959296.111139482372611imprecise_reference
/readings/5/residualtolerance-5.4313880583301e-05-5.434534360126975e-05imprecise_reference
/readings/6/timetolerance1212.00.00E+2passed
/readings/6/responsetolerance9.16631147828433699.1663114782843375.35E-2020passed
/readings/6/fittedtolerance9.16670917636438899.166709223558916imprecise_reference
/readings/6/residualtolerance-0.00039769808005196694-0.00039774527457936415imprecise_reference
/readings/7/timetolerance1616.00.00E+2passed
/readings/7/responsetolerance12.22252529365527912.2225252936552791.56E-1616passed
/readings/7/fittedtolerance12.22227890181918612.222278964745222imprecise_reference
/readings/7/residualtolerance0.000246391836093096120.0002463289100571586imprecise_reference
/readings/8/timetolerance44.00.00E+2passed
/readings/8/responsetolerance1.90732404511663021.90732404511663052.51E-1615passed
/readings/8/fittedtolerance1.90742916602720381.907429218257177imprecise_reference
/readings/8/residualtolerance-0.00010512091057357509-0.00010517314054658833imprecise_reference
/readings/9/timetolerance88.00.00E+2passed
/readings/9/responsetolerance3.81557107578460693.81557107578460734.02E-1615passed
/readings/9/fittedtolerance3.81485833205440763.814858436514354imprecise_reference
/readings/9/residualtolerance0.000712743730199250080.0007126392702532236imprecise_reference
/readings/10/timetolerance1212.00.00E+2passed
/readings/10/responsetolerance5.72199985320000075.7219998532000017.43E-1817passed
/readings/10/fittedtolerance5.72228749808161125.722287654771531imprecise_reference
/readings/10/residualtolerance-0.00028764488161048263-0.0002878015715301885imprecise_reference
/readings/11/timetolerance1616.00.00E+2passed
/readings/11/responsetolerance7.62960230613256667.6296023061325672.48E-1717passed
/readings/11/fittedtolerance7.62971666410881527.629716873028708imprecise_reference
/readings/11/residualtolerance-0.00011435797624859134-0.00011456689614153248imprecise_reference
/readings/12/timetolerance44.00.00E+2passed
/readings/12/responsetolerance2.28971930146936442.28971930146936442.09E-1717passed
/readings/12/fittedtolerance2.2895480863850712.2895481211720474imprecise_reference
/readings/12/residualtolerance0.000171215084293407930.00017118029731699735imprecise_reference
/readings/13/timetolerance88.00.00E+2passed
/readings/13/responsetolerance4.58049359525540914.5804935952554091.16E-1717passed
/readings/13/fittedtolerance4.5790961727701424.579096242344095imprecise_reference
/readings/13/residualtolerance0.00139742248526708580.0013973529113142646imprecise_reference
/readings/14/timetolerance1212.00.00E+2passed
/readings/14/responsetolerance6.86857063927348136.8685706392734812.64E-1717passed
/readings/14/fittedtolerance6.86864425915521266.868644363516142imprecise_reference
/readings/14/residualtolerance-7.3619881731268322e-05-7.372424266094413e-05imprecise_reference
/readings/15/timetolerance1616.00.00E+2passed
/readings/15/responsetolerance9.15750604543787899.1575060454378793.87E-1717passed
/readings/15/fittedtolerance9.15819234554028419.15819248468819imprecise_reference
/readings/15/residualtolerance-0.00068630010240511297-0.0006864392503107553imprecise_reference
/readings/16/timetolerance44.00.00E+2passed
/readings/16/responsetolerance4.83172402901157844.8317240290115781.93E-1717passed
/readings/16/fittedtolerance4.83097498789516164.830974987470563imprecise_reference
/readings/16/residualtolerance0.000749041116416826470.0007490415410149609imprecise_reference
/readings/17/timetolerance88.00.00E+2passed
/readings/17/responsetolerance9.66300027617900889.6630002761790097.28E-1818passed
/readings/17/fittedtolerance9.66194997579032319.661949974941127imprecise_reference
/readings/17/residualtolerance0.00105030038868569870.0010503012378819676imprecise_reference
/readings/18/timetolerance1212.00.00E+2passed
/readings/18/responsetolerance14.49218291018782714.4921829101878275.19E-1818passed
/readings/18/fittedtolerance14.49292496368548614.49292496241169imprecise_reference
/readings/18/residualtolerance-0.00074205349765854578-0.0007420522238632543imprecise_reference
/readings/19/timetolerance1616.00.00E+2passed
/readings/19/responsetolerance19.32374408123044119.323744081230443.72E-1616passed
/readings/19/fittedtolerance19.32389995158064619.323899949882254imprecise_reference
/readings/19/residualtolerance-0.00015587035020558915-0.00015586865181305143imprecise_reference
/readings/20/timetolerance44.00.00E+2passed
/readings/20/responsetolerance7.84102177206024917.841021772060258.72E-1615passed
/readings/20/fittedtolerance7.842516095077137.84251608805606imprecise_reference
/readings/20/residualtolerance-0.0014943230168809052-0.0014943159958100694imprecise_reference
/readings/21/timetolerance88.00.00E+2passed
/readings/21/responsetolerance15.68400199722007915.684001997220082.61E-1616passed
/readings/21/fittedtolerance15.6850321901542615.68503217611212imprecise_reference
/readings/21/residualtolerance-0.0010301929341807181-0.0010301788920408228imprecise_reference
/readings/22/timetolerance1212.00.00E+2passed
/readings/22/responsetolerance23.52906693576921923.529066935769221.84E-1617passed
/readings/22/fittedtolerance23.5275482852313923.527548264168182imprecise_reference
/readings/22/residualtolerance0.00151865053782884730.001518671601036914imprecise_reference
/readings/23/timetolerance1616.00.00E+2passed
/readings/23/responsetolerance31.36981406962645731.3698140696264573.53E-1616passed
/readings/23/fittedtolerance31.3700643803085231.37006435222424imprecise_reference
/readings/23/residualtolerance-0.00025031068206260443-0.0002502825977828138imprecise_reference
/readings/24/timetolerance44.00.00E+2passed
/readings/24/responsetolerance7.95949714032209247.9594971403220923.44E-1717passed
/readings/24/fittedtolerance7.96187872177268217.961878714649294imprecise_reference
/readings/24/residualtolerance-0.0023815814505896427-0.002381574327201541imprecise_reference
/readings/25/timetolerance88.00.00E+2passed
/readings/25/responsetolerance15.92342777809903315.9234277780990341.31E-1516passed
/readings/25/fittedtolerance15.92375744354536415.923757429298588imprecise_reference
/readings/25/residualtolerance-0.00032966544633161732-0.00032965119955363775imprecise_reference
/readings/26/timetolerance1212.00.00E+2passed
/readings/26/responsetolerance23.88550328732781823.885503287327823.56E-1616passed
/readings/26/fittedtolerance23.88563616531804723.88563614394788imprecise_reference
/readings/26/residualtolerance-0.00013287799022876357-0.0001328566200626824imprecise_reference
/readings/27/timetolerance1616.00.00E+2passed
/readings/27/responsetolerance31.84837477366921431.8483747736692173.32E-1515passed
/readings/27/fittedtolerance31.84751488709072831.847514858597176imprecise_reference
/readings/27/residualtolerance0.000859886578485458130.0008599150720414173imprecise_reference
/readings/28/timetolerance44.00.00E+2passed
/readings/28/responsetolerance12.93787619159832412.9378761915983242.72E-1717passed
/readings/28/fittedtolerance12.93847354573757412.938473537248186imprecise_reference
/readings/28/residualtolerance-0.00059735413925032788-0.0005973456498615803imprecise_reference
/readings/29/timetolerance88.00.00E+2passed
/readings/29/responsetolerance25.87642016826196225.8764201682619621.18E-1617passed
/readings/29/fittedtolerance25.87694709147514925.87694707449637imprecise_reference
/readings/29/residualtolerance-0.00052692321318659197-0.0005269062344090969imprecise_reference
/readings/30/timetolerance1212.00.00E+2passed
/readings/30/responsetolerance38.81444853372082838.814448533720833.84E-1718passed
/readings/30/fittedtolerance38.81542063721272538.81542061174456imprecise_reference
/readings/30/residualtolerance-0.00097210349189680301-0.000972078023728784imprecise_reference
/readings/31/timetolerance1616.00.00E+2passed
/readings/31/responsetolerance51.75503606071062351.755036060710622.08E-1617passed
/readings/31/fittedtolerance51.75389418295029751.75389414899274imprecise_reference
/readings/31/residualtolerance0.00114187776032537160.0011419117178803617imprecise_reference
/readings/32/timetolerance44.00.00E+2passed
/readings/32/responsetolerance8.69208914108981658.6920891410898164.91E-1717passed
/readings/32/fittedtolerance8.69440465684845348.69440464921763imprecise_reference
/readings/32/residualtolerance-0.0023155157586369768-0.002315508127813004imprecise_reference
/readings/33/timetolerance88.00.00E+2passed
/readings/33/responsetolerance17.39022589687144617.3902258968714464.01E-1616passed
/readings/33/fittedtolerance17.38880931369690717.38880929843526imprecise_reference
/readings/33/residualtolerance0.00141658317453874360.001416598436186689imprecise_reference
/readings/34/timetolerance1212.00.00E+2passed
/readings/34/responsetolerance26.08356907545930226.08356907545933.00E-1616passed
/readings/34/fittedtolerance26.08321397054535926.083213947652887imprecise_reference
/readings/34/residualtolerance0.000355104913943193880.000355127806415112imprecise_reference
/readings/35/timetolerance1616.00.00E+2passed
/readings/35/responsetolerance34.77722288606074134.777222886060743.03E-1617passed
/readings/35/fittedtolerance34.77761862739381434.77761859687052imprecise_reference
/readings/35/residualtolerance-0.000395741333072408-0.0003957108097765172imprecise_reference
/readings/36/timetolerance44.00.00E+2passed
/readings/36/responsetolerance4.19057072309893374.1905707230989344.75E-1716passed
/readings/36/fittedtolerance4.1907270113824544.190727014604903imprecise_reference
/readings/36/residualtolerance-0.00015628828352021173-0.00015629150596918606imprecise_reference
/readings/37/timetolerance88.00.00E+2passed
/readings/37/responsetolerance8.38100389178307918.381003891783084.05E-1717passed
/readings/37/fittedtolerance8.38145402276490798.381454029209806imprecise_reference
/readings/37/residualtolerance-0.00045013098182877798-0.00045013742672672663imprecise_reference
/readings/38/timetolerance1212.00.00E+2passed
/readings/38/responsetolerance12.57175835389657712.5717583538965777.92E-1717passed
/readings/38/fittedtolerance12.57218103414736312.572181043814709imprecise_reference
/readings/38/residualtolerance-0.00042268025078584515-0.00042268991813187995imprecise_reference
/readings/39/timetolerance1616.00.00E+2passed
/readings/39/responsetolerance16.763489193279716.76348919327973.37E-1616passed
/readings/39/fittedtolerance16.76290804552981616.76290805841961imprecise_reference
/readings/39/residualtolerance0.000581147749883825780.0005811348600879285imprecise_reference

Convention

Reference code

The lines of the reference script that produce this case's expected values.

246  BASE_NOTES <- c(
247    paste0(MANUAL, ":97-104: y_ij = y0 + s exp(eta_i + u_i) t_ij + e_ij with u ~ N(0, tau^2) and ",
248           "log T ~ N(log D - eta, tau^2); failure is threshold crossing of the noise-free trajectory."),
249    paste0(MANUAL, ":110-122: inverse-kelvin temperature acceleration and delta-method lifetime intervals ",
250           "for population quantities."),
251    paste0("Oracle: per-unit least-squares rates r_ls,i = sum(t z) / sum(t^2) with z = s (y - y0), ",
252           "pseudo-failure times t~_i = D / r_ls,i, then R survival::survreg(Surv(t~) ~ 1, dist = \"lognormal\") ",
253           "for STAT-SURV-DEG-001. survreg's intercept a and Log(scale) map to the engine's ",
254           "beta0 = log D - a and log tau; the lifetime covariance of (eta_use, log tau) is survreg's ",
255           "(Intercept, Log(scale)) covariance with the intercept signs flipped. survreg fits the ",
256           "observed information of the uncensored lognormal, which at the ML point equals the ",
257           "engine's marginal-likelihood information in this limit."),
258    paste0("Pseudo-failure limit: the data carry measurement noise SD 0.001 while the log rates spread ",
259           "over about 0.3, so sigma_e^2 / (tau^2 S_tt r^2) is below 2e-8 and the engine's direct ",
260           "marginal ML, its covariance, the empirical-Bayes rates and the fitted trajectories agree with ",
261           "the two-stage reference to well under 1e-6 relative (1e-7 absolute); the per-unit LS rates are ",
262           "closed-form algebra and allow 1e-9."),
263    paste0("Engine contract note: repeated-measurement mode estimates sigma_e and refuses a supplied ",
264           "measurement_sd (degradation_reliability_contracts), so the small calibrated error is carried by ",
265           "the data and the ML sigma_e is compared with its pseudo-failure limit sqrt(RSS / (N - g))."),
266    paste0("Not compared: the quadrature log-likelihood and its diagnostics (integration constants and ",
267           "search internals without a two-stage counterpart), the raw covariance matrix (its entries are ",
268           "covered through the parameter and lifetime intervals), the reliability curve (plot data) and ",
269           "text/boolean labels.")

Whole script generators/r/survival_degradation.R


Degradation-based lifetime: inverse-kelvin temperature acceleration to a use temperature

Case STAT-SURV-DEG-002 · status: imprecise_reference

Reference

Evidence type
R
Source
R survival::survreg lognormal pseudo-failure fits and base R delta-method algebra
Citation
Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, chapter 21 (pseudo-failure limits); Therneau T (2024), A Package for Survival Analysis in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_degradation.R · sha256 b4e6ab8a59ddbd26
Dataset
reference_data/datasets/SURV_DEG_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/distancetolerance2525.00.00E+2passed
/unit_countexact10100.00E+2passed
/analyzed_countexact40400.00E+2passed
/source_countexact40400.00E+2passed
/excluded_countexact000.00E+2passed
/unobserved_countexact000.00E+2passed
/parameters/[name=beta0]/estimatetolerance0.184422919294014330.18442291959155713imprecise_reference
/parameters/[name=beta0]/standard_errortolerance0.107285194188771070.10728519510674801imprecise_reference
/parameters/[name=beta0]/lowertolerance-0.025852197390362819-0.02585219889202181imprecise_reference
/parameters/[name=beta0]/uppertolerance0.394698035978391480.39469803807513604imprecise_reference
/parameters/[name=beta1]/estimatetolerance6070.72299145763286070.722998408129imprecise_reference
/parameters/[name=beta1]/standard_errortolerance1297.65811244473551297.6581243496578imprecise_reference
/parameters/[name=beta1]/lowertolerance3527.35982681972433527.3598104370003imprecise_reference
/parameters/[name=beta1]/uppertolerance8614.08615609554188614.086186379256imprecise_reference
/parameters/[name=tau]/estimatetolerance0.339125587544234050.3391255857275501imprecise_reference
/parameters/[name=tau]/standard_errortolerance0.223606797736725090.22360680311450437imprecise_reference
/parameters/[name=tau]/lowertolerance0.218789314659125540.2187893111809839imprecise_reference
/parameters/[name=tau]/uppertolerance0.525647992939700550.5256479956642859imprecise_reference
/parameters/[name=sigma_e]/estimatetolerance0.00109711630085728030.0010971162989607591imprecise_reference
/lifetime/distancetolerance2525.00.00E+2passed
/lifetime/eta_usetolerance0.184422919294014330.18442291959155713imprecise_reference
/lifetime/tautolerance0.339125587544234050.3391255857275501imprecise_reference
/lifetime/median/probabilitytolerance0.50.50.00E+1passed
/lifetime/median/estimatetolerance20.78960091284217220.789600906656375imprecise_reference
/lifetime/median/lowertolerance16.8470874274931516.847087392169104imprecise_reference
/lifetime/median/uppertolerance25.65473159532125925.654731633845927imprecise_reference
/lifetime/quantiles/0/probabilitytolerance0.100000000000000010.14.45E-1816passed
/lifetime/quantiles/0/estimatetolerance13.46166448203341313.46166450936909imprecise_reference
/lifetime/quantiles/0/lowertolerance10.13636667648960110.136366690369554imprecise_reference
/lifetime/quantiles/0/uppertolerance17.87784680749121317.877846855617303imprecise_reference
/lifetime/quantiles/1/probabilitytolerance0.50.50.00E+1passed
/lifetime/quantiles/1/estimatetolerance20.78960091284217220.789600906656375imprecise_reference
/lifetime/quantiles/1/lowertolerance16.8470874274931516.847087392169104imprecise_reference
/lifetime/quantiles/1/uppertolerance25.65473159532125925.654731633845927imprecise_reference
/lifetime/quantiles/2/probabilitytolerance0.900000000000000020.92.20E-1817passed
/lifetime/quantiles/2/estimatetolerance32.106542745296732.106542660994imprecise_reference
/lifetime/quantiles/2/lowertolerance24.17559064966053824.175590422119463imprecise_reference
/lifetime/quantiles/2/uppertolerance42.63929274753989842.63929292494423imprecise_reference
/lifetime/reliability/0/timetolerance1010.00.00E+2passed
/lifetime/reliability/0/reliabilitytolerance0.984540072210966070.9845400726261883imprecise_reference
/lifetime/reliability/0/lowertolerance0.78008003700645390.780080032819208imprecise_reference
/lifetime/reliability/0/uppertolerance0.999126137790649180.9991261378595953imprecise_reference
/lifetime/reliability/1/timetolerance1515.00.00E+2passed
/lifetime/reliability/1/reliabilitytolerance0.832096548487204710.832096549561325imprecise_reference
/lifetime/reliability/1/lowertolerance0.562922573748204310.5629225725146134imprecise_reference
/lifetime/reliability/1/uppertolerance0.950172995675624410.9501729966409735imprecise_reference
/lifetime/reliability/2/timetolerance2020.00.00E+2passed
/lifetime/reliability/2/reliabilitytolerance0.545451596713355190.5454515966080308imprecise_reference
/lifetime/reliability/2/lowertolerance0.307436101556415490.307436098879358imprecise_reference
/lifetime/reliability/2/uppertolerance0.764363404396149470.7643634065076927imprecise_reference
/units/0/readingsexact440.00E+2passed
/units/0/distinct_positive_timesexact440.00E+2passed
/units/0/least_squares_ratetolerance0.756240222533499760.75624022253349981.86E-1817passed
/units/0/empirical_bayes_ratetolerance0.756240222533499760.7562402210257444imprecise_reference
/units/0/temperaturetolerance338338.00.00E+2passed
/units/1/readingsexact440.00E+2passed
/units/1/distinct_positive_timesexact440.00E+2passed
/units/1/least_squares_ratetolerance0.763892431363699110.76389243136369917.52E-1918passed
/units/1/empirical_bayes_ratetolerance0.763892431363699110.763892429583672imprecise_reference
/units/1/temperaturetolerance338338.00.00E+2passed
/units/2/readingsexact440.00E+2passed
/units/2/distinct_positive_timesexact440.00E+2passed
/units/2/least_squares_ratetolerance0.476857291506800950.476857291506800951.01E-1817passed
/units/2/empirical_bayes_ratetolerance0.476857291506800950.47685731020142397imprecise_reference
/units/2/temperaturetolerance338338.00.00E+2passed
/units/3/readingsexact440.00E+2passed
/units/3/distinct_positive_timesexact440.00E+2passed
/units/3/least_squares_ratetolerance0.572387021596267750.5723870215962682.25E-1615passed
/units/3/empirical_bayes_ratetolerance0.572387021596267750.5723870302149758imprecise_reference
/units/3/temperaturetolerance338338.00.00E+2passed
/units/4/readingsexact440.00E+2passed
/units/4/distinct_positive_timesexact440.00E+2passed
/units/4/least_squares_ratetolerance1.20774374697379041.20774374697379022.33E-1615passed
/units/4/empirical_bayes_ratetolerance1.20774374697379041.2077437375778144imprecise_reference
/units/4/temperaturetolerance338338.00.00E+2passed
/units/5/readingsexact440.00E+2passed
/units/5/distinct_positive_timesexact440.00E+2passed
/units/5/least_squares_ratetolerance1.96062902376928251.96062902376928252.63E-1817passed
/units/5/empirical_bayes_ratetolerance1.96062902376928251.960629023751901imprecise_reference
/units/5/temperaturetolerance358358.00.00E+2passed
/units/6/readingsexact440.00E+2passed
/units/6/distinct_positive_timesexact440.00E+2passed
/units/6/least_squares_ratetolerance1.99046968044317051.99046968044317072.41E-1615passed
/units/6/empirical_bayes_ratetolerance1.99046968044317051.990469680260581imprecise_reference
/units/6/temperaturetolerance358358.00.00E+2passed
/units/7/readingsexact440.00E+2passed
/units/7/distinct_positive_timesexact440.00E+2passed
/units/7/least_squares_ratetolerance3.23461838643439363.23461838643439361.12E-1717passed
/units/7/empirical_bayes_ratetolerance3.23461838643439363.2346183830490394imprecise_reference
/units/7/temperaturetolerance358358.00.00E+2passed
/units/8/readingsexact440.00E+2passed
/units/8/distinct_positive_timesexact440.00E+2passed
/units/8/least_squares_ratetolerance2.17360116421211342.17360116421211344.31E-1716passed
/units/8/empirical_bayes_ratetolerance2.17360116421211342.1736011631619934imprecise_reference
/units/8/temperaturetolerance358358.00.00E+2passed
/units/9/readingsexact440.00E+2passed
/units/9/distinct_positive_timesexact440.00E+2passed
/units/9/least_squares_ratetolerance1.04768175284561351.04768175284561351.02E-1717passed
/units/9/empirical_bayes_ratetolerance1.04768175284561351.0476817658556574imprecise_reference
/units/9/temperaturetolerance358358.00.00E+2passed
/readings/0/timetolerance44.00.00E+2passed
/readings/0/responsetolerance3.02633160205862423.0263316020586242.94E-1717passed
/readings/0/fittedtolerance3.0249608901339993.0249608841029776imprecise_reference
/readings/0/residualtolerance0.00137071192462512310.0013707179556465832imprecise_reference
/readings/1/timetolerance88.00.00E+2passed
/readings/1/responsetolerance6.05181797477948546.0518179747794854.60E-1717passed
/readings/1/fittedtolerance6.04992178026799816.049921768205955imprecise_reference
/readings/1/residualtolerance0.00189619451148725910.0018962065735301792imprecise_reference
/readings/2/timetolerance1212.00.00E+2passed
/readings/2/responsetolerance9.07480341647430329.0748034164743031.93E-1717passed
/readings/2/fittedtolerance9.07488267040199769.074882652308933imprecise_reference
/readings/2/residualtolerance-7.9253927694367121e-05-7.923583462954298e-05imprecise_reference
/readings/3/timetolerance1616.00.00E+2passed
/readings/3/responsetolerance12.09861222574486712.0986122257448673.88E-1616passed
/readings/3/fittedtolerance12.09984356053599612.09984353641191imprecise_reference
/readings/3/residualtolerance-0.001231334791128802-0.0012313106670429619imprecise_reference
/readings/4/timetolerance44.00.00E+2passed
/readings/4/responsetolerance3.05588588011175413.0558858801117544.86E-1817passed
/readings/4/fittedtolerance3.05556972545479643.055569718334688imprecise_reference
/readings/4/residualtolerance0.000316154656957667870.00031616177706617066imprecise_reference
/readings/5/timetolerance88.00.00E+2passed
/readings/5/responsetolerance6.11108513702900966.111085137029012.70E-1717passed
/readings/5/fittedtolerance6.11113945090959296.111139436669376imprecise_reference
/readings/5/residualtolerance-5.4313880583301e-05-5.429964036629542e-05imprecise_reference
/readings/6/timetolerance1212.00.00E+2passed
/readings/6/responsetolerance9.16631147828433699.1663114782843375.35E-2020passed
/readings/6/fittedtolerance9.16670917636438899.166709155004064imprecise_reference
/readings/6/residualtolerance-0.00039769808005196694-0.00039767671972690266imprecise_reference
/readings/7/timetolerance1616.00.00E+2passed
/readings/7/responsetolerance12.22252529365527912.2225252936552791.56E-1616passed
/readings/7/fittedtolerance12.22227890181918612.222278873338752imprecise_reference
/readings/7/residualtolerance0.000246391836093096120.0002464203165271073imprecise_reference
/readings/8/timetolerance44.00.00E+2passed
/readings/8/responsetolerance1.90732404511663021.90732404511663052.51E-1615passed
/readings/8/fittedtolerance1.90742916602720381.9074292408056959imprecise_reference
/readings/8/residualtolerance-0.00010512091057357509-0.00010519568906541821imprecise_reference
/readings/9/timetolerance88.00.00E+2passed
/readings/9/responsetolerance3.81557107578460693.81557107578460734.02E-1615passed
/readings/9/fittedtolerance3.81485833205440763.8148584816113917imprecise_reference
/readings/9/residualtolerance0.000712743730199250080.0007125941732155638imprecise_reference
/readings/10/timetolerance1212.00.00E+2passed
/readings/10/responsetolerance5.72199985320000075.7219998532000017.43E-1817passed
/readings/10/fittedtolerance5.72228749808161125.722287722417088imprecise_reference
/readings/10/residualtolerance-0.00028764488161048263-0.00028786921708690016imprecise_reference
/readings/11/timetolerance1616.00.00E+2passed
/readings/11/responsetolerance7.62960230613256667.6296023061325672.48E-1717passed
/readings/11/fittedtolerance7.62971666410881527.6297169632227835imprecise_reference
/readings/11/residualtolerance-0.00011435797624859134-0.000114657090216852imprecise_reference
/readings/12/timetolerance44.00.00E+2passed
/readings/12/responsetolerance2.28971930146936442.28971930146936442.09E-1717passed
/readings/12/fittedtolerance2.2895480863850712.2895481208599033imprecise_reference
/readings/12/residualtolerance0.000171215084293407930.0001711806094610857imprecise_reference
/readings/13/timetolerance88.00.00E+2passed
/readings/13/responsetolerance4.58049359525540914.5804935952554091.16E-1717passed
/readings/13/fittedtolerance4.5790961727701424.579096241719807imprecise_reference
/readings/13/residualtolerance0.00139742248526708580.0013973535356024414imprecise_reference
/readings/14/timetolerance1212.00.00E+2passed
/readings/14/responsetolerance6.86857063927348136.8685706392734812.64E-1717passed
/readings/14/fittedtolerance6.86864425915521266.86864436257971imprecise_reference
/readings/14/residualtolerance-7.3619881731268322e-05-7.372330622867906e-05imprecise_reference
/readings/15/timetolerance1616.00.00E+2passed
/readings/15/responsetolerance9.15750604543787899.1575060454378793.87E-1717passed
/readings/15/fittedtolerance9.15819234554028419.158192483439613imprecise_reference
/readings/15/residualtolerance-0.00068630010240511297-0.0006864380017344018imprecise_reference
/readings/16/timetolerance44.00.00E+2passed
/readings/16/responsetolerance4.83172402901157844.8317240290115781.93E-1717passed
/readings/16/fittedtolerance4.83097498789516164.830974950311258imprecise_reference
/readings/16/residualtolerance0.000749041116416826470.0007490787003208155imprecise_reference
/readings/17/timetolerance88.00.00E+2passed
/readings/17/responsetolerance9.66300027617900889.6630002761790097.28E-1818passed
/readings/17/fittedtolerance9.66194997579032319.661949900622515imprecise_reference
/readings/17/residualtolerance0.00105030038868569870.0010503755564936768imprecise_reference
/readings/18/timetolerance1212.00.00E+2passed
/readings/18/responsetolerance14.49218291018782714.4921829101878275.19E-1818passed
/readings/18/fittedtolerance14.49292496368548614.492924850933772imprecise_reference
/readings/18/residualtolerance-0.00074205349765854578-0.0007419407459448024imprecise_reference
/readings/19/timetolerance1616.00.00E+2passed
/readings/19/responsetolerance19.32374408123044119.323744081230443.72E-1616passed
/readings/19/fittedtolerance19.32389995158064619.32389980124503imprecise_reference
/readings/19/residualtolerance-0.00015587035020558915-0.00015572001458963314imprecise_reference
/readings/20/timetolerance44.00.00E+2passed
/readings/20/responsetolerance7.84102177206024917.841021772060258.72E-1615passed
/readings/20/fittedtolerance7.842516095077137.842516095007604imprecise_reference
/readings/20/residualtolerance-0.0014943230168809052-0.0014943229473542985imprecise_reference
/readings/21/timetolerance88.00.00E+2passed
/readings/21/responsetolerance15.68400199722007915.684001997220082.61E-1616passed
/readings/21/fittedtolerance15.6850321901542615.685032190015209imprecise_reference
/readings/21/residualtolerance-0.0010301929341807181-0.001030192795129281imprecise_reference
/readings/22/timetolerance1212.00.00E+2passed
/readings/22/responsetolerance23.52906693576921923.529066935769221.84E-1617passed
/readings/22/fittedtolerance23.5275482852313923.527548285022814imprecise_reference
/readings/22/residualtolerance0.00151865053782884730.0015186507464051147imprecise_reference
/readings/23/timetolerance1616.00.00E+2passed
/readings/23/responsetolerance31.36981406962645731.3698140696264573.53E-1616passed
/readings/23/fittedtolerance31.3700643803085231.370064380030417imprecise_reference
/readings/23/residualtolerance-0.00025031068206260443-0.00025031040395973037imprecise_reference
/readings/24/timetolerance44.00.00E+2passed
/readings/24/responsetolerance7.95949714032209247.9594971403220923.44E-1717passed
/readings/24/fittedtolerance7.96187872177268217.961878721042324imprecise_reference
/readings/24/residualtolerance-0.0023815814505896427-0.0023815807202316464imprecise_reference
/readings/25/timetolerance88.00.00E+2passed
/readings/25/responsetolerance15.92342777809903315.9234277780990341.31E-1516passed
/readings/25/fittedtolerance15.92375744354536415.923757442084648imprecise_reference
/readings/25/residualtolerance-0.00032966544633161732-0.00032966398561384835imprecise_reference
/readings/26/timetolerance1212.00.00E+2passed
/readings/26/responsetolerance23.88550328732781823.885503287327823.56E-1616passed
/readings/26/fittedtolerance23.88563616531804723.88563616312697imprecise_reference
/readings/26/residualtolerance-0.00013287799022876357-0.0001328757991529983imprecise_reference
/readings/27/timetolerance1616.00.00E+2passed
/readings/27/responsetolerance31.84837477366921431.8483747736692173.32E-1515passed
/readings/27/fittedtolerance31.84751488709072831.847514884169296imprecise_reference
/readings/27/residualtolerance0.000859886578485458130.0008598894999209961imprecise_reference
/readings/28/timetolerance44.00.00E+2passed
/readings/28/responsetolerance12.93787619159832412.9378761915983242.72E-1717passed
/readings/28/fittedtolerance12.93847354573757412.938473532196157imprecise_reference
/readings/28/residualtolerance-0.00059735413925032788-0.0005973405978334512imprecise_reference
/readings/29/timetolerance88.00.00E+2passed
/readings/29/responsetolerance25.87642016826196225.8764201682619621.18E-1617passed
/readings/29/fittedtolerance25.87694709147514925.876947064392315imprecise_reference
/readings/29/residualtolerance-0.00052692321318659197-0.0005268961303528386imprecise_reference
/readings/30/timetolerance1212.00.00E+2passed
/readings/30/responsetolerance38.81444853372082838.814448533720833.84E-1718passed
/readings/30/fittedtolerance38.81542063721272538.815420596588474imprecise_reference
/readings/30/residualtolerance-0.00097210349189680301-0.0009720628676461729imprecise_reference
/readings/31/timetolerance1616.00.00E+2passed
/readings/31/responsetolerance51.75503606071062351.755036060710622.08E-1617passed
/readings/31/fittedtolerance51.75389418295029751.75389412878463imprecise_reference
/readings/31/residualtolerance0.00114187776032537160.0011419319259928784imprecise_reference
/readings/32/timetolerance44.00.00E+2passed
/readings/32/responsetolerance8.69208914108981658.6920891410898164.91E-1717passed
/readings/32/fittedtolerance8.69440465684845348.694404652647973imprecise_reference
/readings/32/residualtolerance-0.0023155157586369768-0.002315511558157013imprecise_reference
/readings/33/timetolerance88.00.00E+2passed
/readings/33/responsetolerance17.39022589687144617.3902258968714464.01E-1616passed
/readings/33/fittedtolerance17.38880931369690717.388809305295947imprecise_reference
/readings/33/residualtolerance0.00141658317453874360.0014165915754986713imprecise_reference
/readings/34/timetolerance1212.00.00E+2passed
/readings/34/responsetolerance26.08356907545930226.08356907545933.00E-1616passed
/readings/34/fittedtolerance26.08321397054535926.08321395794392imprecise_reference
/readings/34/residualtolerance0.000355104913943193880.00035511751538308545imprecise_reference
/readings/35/timetolerance1616.00.00E+2passed
/readings/35/responsetolerance34.77722288606074134.777222886060743.03E-1617passed
/readings/35/fittedtolerance34.77761862739381434.777618610591894imprecise_reference
/readings/35/residualtolerance-0.000395741333072408-0.00039572453115255257imprecise_reference
/readings/36/timetolerance44.00.00E+2passed
/readings/36/responsetolerance4.19057072309893374.1905707230989344.75E-1716passed
/readings/36/fittedtolerance4.1907270113824544.19072706342263imprecise_reference
/readings/36/residualtolerance-0.00015628828352021173-0.00015634032369593598imprecise_reference
/readings/37/timetolerance88.00.00E+2passed
/readings/37/responsetolerance8.38100389178307918.381003891783084.05E-1717passed
/readings/37/fittedtolerance8.38145402276490798.38145412684526imprecise_reference
/readings/37/residualtolerance-0.00045013098182877798-0.00045023506218022646imprecise_reference
/readings/38/timetolerance1212.00.00E+2passed
/readings/38/responsetolerance12.57175835389657712.5717583538965777.92E-1717passed
/readings/38/fittedtolerance12.57218103414736312.572181190267889imprecise_reference
/readings/38/residualtolerance-0.00042268025078584515-0.0004228363713121297imprecise_reference
/readings/39/timetolerance1616.00.00E+2passed
/readings/39/responsetolerance16.763489193279716.76348919327973.37E-1616passed
/readings/39/fittedtolerance16.76290804552981616.76290825369052imprecise_reference
/readings/39/residualtolerance0.000581147749883825780.0005809395891809288imprecise_reference

Convention

Reference code

The lines of the reference script that produce this case's expected values.

292  emit("STAT-SURV-DEG-002",
293       "Degradation-based lifetime: inverse-kelvin temperature acceleration to a use temperature",
294       list(
295         mode = "repeated", response_column = "response", time_column = "time",
296         unit_column = "unit", direction = "increasing", baseline = 0, threshold = 25,
297         time_unit = "hours", temperature_column = "temp_k", use_temperature = 348,
298         dropout_ignorable_declared = TRUE,
299         lifetime_probabilities = as.list(probabilities),
300         reliability_times = as.list(r_times),
301         confidence_level = 0.95
302       ),
303       c(counts_outputs,
304         parameter_outputs(TRUE, sqrt(V2[1, 1]), sqrt(V2[2, 2]), eta2, tau2),
305         lifetime_outputs(eta2, tau2, cov2),
306         units_outputs(TRUE),
307         readings_outputs()),
308       c(BASE_NOTES,
309         paste0("Temperature oracle: x_i = 1/348 - 1/temp_k_i and R survival::survreg(Surv(t~) ~ x_i, ",
310                "dist = \"lognormal\"); the engine's beta0 = log D - a and beta1 = -b, with the ",
311                "covariance of (eta_use, log tau) taken from the (Intercept, Log(scale)) block with the ",
312                "intercept signs flipped; beta1's own covariance enters only through the parameter table.")),

Whole script generators/r/survival_degradation.R


Distribution identification of complete data across the seven families

Case STAT-SURV-DISTID-001 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_COMPLETE_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/plotted_pointsexact29290.00E+2passed
/families/0/parameters/0/estimatetolerance1.45953266119186381.45953266108494951.07E-1010passed
/families/0/parameters/0/standard_errortolerance0.187232842316485020.187232834093899688.22E-97passed
/families/0/parameters/0/lowertolerance1.13506140183567731.13506141426482741.24E-87passed
/families/0/parameters/0/uppertolerance1.87675801999846151.87675799917263092.08E-87passed
/families/0/parameters/1/estimatetolerance102.80061609192241102.80061608205559.87E-910passed
/families/0/parameters/1/standard_errortolerance13.92907291830159113.9290723093167876.09E-77passed
/families/0/parameters/1/lowertolerance78.82451510373955778.824516009373459.06E-77passed
/families/0/parameters/1/uppertolerance134.06954238754912134.06954082145551.57E-67passed
/families/0/parameters/2/estimatetolerance00.00.00E+2passed
/families/0/log_likelihoodtolerance-156.6748912336796-156.674891233679572.60E-1415passed
/families/0/anderson_darlingtolerance1.93864244316831421.9386424428925372.76E-109passed
/families/0/correlationtolerance0.915748920896303890.91574892089630393.75E-1817passed
/families/0/percentiles/0/estimatetolerance4.397320288747014.3973202873097091.44E-99passed
/families/0/percentiles/0/standard_errortolerance2.06027503859747262.06027493552585341.03E-77passed
/families/0/percentiles/0/lowertolerance1.75539524890301491.75539532844672237.95E-87passed
/families/0/percentiles/0/uppertolerance11.01542557661008311.0154250702578485.06E-77passed
/families/0/percentiles/1/estimatetolerance13.43360811015725713.4336081068653183.29E-99passed
/families/0/percentiles/1/standard_errortolerance4.47247706562276554.4724768348834032.31E-77passed
/families/0/percentiles/1/lowertolerance6.99520590833926016.9952061409992872.33E-77passed
/families/0/percentiles/1/uppertolerance25.79792921351280925.797928342831948.71E-77passed
/families/0/percentiles/2/estimatetolerance21.99797329013561921.9979732855396974.60E-99passed
/families/0/percentiles/2/standard_errortolerance6.06107710345452236.0610767861504733.17E-77passed
/families/0/percentiles/2/lowertolerance12.8189382741445512.8189386324241533.58E-77passed
/families/0/percentiles/2/uppertolerance37.74968086472146237.749679793872851.07E-67passed
/families/0/percentiles/3/estimatetolerance79.97185161498420379.97185160583749.15E-99passed
/families/0/percentiles/3/standard_errortolerance11.96894568128181811.9689451019142145.79E-77passed
/families/0/percentiles/3/lowertolerance59.64067006892553759.640670906955428.38E-77passed
/families/0/percentiles/3/uppertolerance107.23382288190092107.233821350594871.53E-67passed
/families/1/parameters/0/estimatetolerance4.31092652098065354.31092652098065354.24E-1717passed
/families/1/parameters/0/standard_errortolerance0.112241202939502850.112241191867327781.11E-87passed
/families/1/parameters/0/lowertolerance4.09093780563777724.0909378273388412.17E-88passed
/families/1/parameters/0/uppertolerance4.53091523632352994.5309152146224662.17E-88passed
/families/1/parameters/1/estimatetolerance0.604437375980252490.60443737596394411.63E-1110passed
/families/1/parameters/1/standard_errortolerance0.0793665157262890830.079366511809926753.92E-97passed
/families/1/parameters/1/lowertolerance0.467286326650499240.467286332568854055.92E-97passed
/families/1/parameters/1/uppertolerance0.781842995708170620.78184298576364919.94E-97passed
/families/1/parameters/2/estimatetolerance00.00.00E+2passed
/families/1/log_likelihoodtolerance-151.56582746173058-151.565827461730553.06E-1415passed
/families/1/anderson_darlingtolerance0.939030471938022910.93903047192964388.38E-1211passed
/families/1/correlationtolerance0.958752394125690580.95875239412569082.19E-1615passed
/families/1/percentiles/0/estimatetolerance18.26149813397270818.261498134665536.93E-1010passed
/families/1/percentiles/0/standard_errortolerance3.94582963703255323.94582941034567722.27E-77passed
/families/1/percentiles/0/lowertolerance11.95677697339018411.95677726494162.92E-77passed
/families/1/percentiles/0/uppertolerance27.89065271011191227.8906520321486836.78E-77passed
/families/1/percentiles/1/estimatetolerance27.56964947170123627.5696494724407837.40E-1010passed
/families/1/percentiles/1/standard_errortolerance4.74649965131632454.7464993450473043.06E-77passed
/families/1/percentiles/1/lowertolerance19.67365392849582119.673654357557734.29E-77passed
/families/1/percentiles/1/uppertolerance38.63469260743419738.6346917669245268.41E-77passed
/families/1/percentiles/2/estimatetolerance34.33993132179120534.339931322508917.18E-1010passed
/families/1/percentiles/2/standard_errortolerance5.20150512059882795.2015047533210653.67E-77passed
/families/1/percentiles/2/lowertolerance25.51919902615782825.5191995617960185.36E-77passed
/families/1/percentiles/2/uppertolerance46.20955704670020546.209556078710949.68E-77passed
/families/1/percentiles/3/estimatetolerance74.50949157641758874.509491576417592.13E-1718passed
/families/1/percentiles/3/standard_errortolerance8.36303496494786468.363034139965738.25E-77passed
/families/1/percentiles/3/lowertolerance59.79594238960201159.79594368723761.30E-67passed
/families/1/percentiles/3/uppertolerance92.84349594834094492.843493933538312.01E-67passed
/families/2/parameters/0/estimatetolerance91.91034482758615891.910345222115373.95E-78passed
/families/2/parameters/0/standard_errortolerance17.06732256472805917.0673215154616321.05E-67passed
/families/2/parameters/0/lowertolerance63.87048839414666263.870490197221671.80E-67passed
/families/2/parameters/0/uppertolerance132.26001082371491132.26000822545572.60E-67passed
/families/2/parameters/1/estimatetolerance00.00.00E+2passed
/families/2/log_likelihoodtolerance-160.10359408561541-160.103594085615382.57E-1415passed
/families/2/anderson_darlingtolerance3.06302781376965073.06302779873671851.50E-88passed
/families/2/correlationtolerance0.915748920896303890.91574892089630393.75E-1817passed
/families/2/percentiles/0/estimatetolerance0.923729833928370940.92372983789352133.97E-98passed
/families/2/percentiles/0/standard_errortolerance0.171532323895560750.171532313350080651.05E-87passed
/families/2/percentiles/0/lowertolerance0.641919859488340850.64191987760984931.81E-87passed
/families/2/percentiles/0/uppertolerance1.32925752876607151.32925750265269362.61E-87passed
/families/2/percentiles/1/estimatetolerance4.71438437450266354.7143843947393622.02E-88passed
/families/2/percentiles/1/standard_errortolerance0.875439200719881130.87543914689954865.38E-87passed
/families/2/percentiles/1/lowertolerance3.27612776387759873.27612785636325089.25E-87passed
/families/2/percentiles/1/uppertolerance6.78405167088143116.7840515376081561.33E-77passed
/families/2/percentiles/2/estimatetolerance9.68372132532310569.6837213668909074.16E-88passed
/families/2/percentiles/2/standard_errortolerance1.79822190631820261.7982217957669511.11E-77passed
/families/2/percentiles/2/lowertolerance6.72942759252450446.7294277824974131.90E-77passed
/families/2/percentiles/2/uppertolerance13.93498294129628313.9349826675423662.74E-77passed
/families/2/percentiles/3/estimatetolerance63.70739638153369863.707396655000512.73E-78passed
/families/2/percentiles/3/standard_errortolerance11.83016651544838811.8301657881523237.27E-77passed
/families/2/percentiles/3/lowertolerance44.2716489513894844.271650201185811.25E-67passed
/families/2/percentiles/3/uppertolerance91.67565360328582491.675651802309791.80E-67passed
/families/3/parameters/0/estimatetolerance4.26062129882267644.2606212988602363.76E-1111passed
/families/3/parameters/0/standard_errortolerance0.100823031105551340.10082302829265732.81E-97passed
/families/3/parameters/0/lowertolerance4.06301178904363444.0630117945943655.55E-98passed
/families/3/parameters/0/uppertolerance4.45823080860171844.4582308031261065.48E-98passed
/families/3/parameters/1/estimatetolerance0.319818332390454370.31981833238854351.91E-1211passed
/families/3/parameters/1/standard_errortolerance0.0514269704360228110.051426968215338712.22E-97passed
/families/3/parameters/1/lowertolerance0.23336174322143730.233361746395462343.17E-97passed
/families/3/parameters/1/uppertolerance0.438305629367680580.43830562340091315.97E-97passed
/families/3/parameters/2/estimatetolerance00.00.00E+2passed
/families/3/log_likelihoodtolerance-150.69088825062573-150.690888250625732.78E-1516passed
/families/3/anderson_darlingtolerance0.679915820484633840.6799158204843153.19E-1312passed
/families/3/correlationtolerance0.966016335166557270.96601633516655763.35E-1615passed
/families/3/percentiles/0/estimatetolerance16.29759793852010616.2975979392753357.55E-1010passed
/families/3/percentiles/0/standard_errortolerance4.08199204109599474.0819918778275831.63E-77passed
/families/3/percentiles/0/lowertolerance9.9753160583498959.975316254902491.97E-77passed
/families/3/percentiles/0/uppertolerance26.62689553012438626.6268950079385755.22E-77passed
/families/3/percentiles/1/estimatetolerance27.63092826602407827.630928267217331.19E-910passed
/families/3/percentiles/1/standard_errortolerance4.86455007290561974.8645498916117531.81E-77passed
/families/3/percentiles/1/lowertolerance19.56760118446710819.5676014372400432.53E-77passed
/families/3/percentiles/1/uppertolerance39.01695407857221239.01695357792385.01E-77passed
/families/3/percentiles/2/estimatetolerance35.08973363340215535.089733634867421.47E-910passed
/families/3/percentiles/2/standard_errortolerance5.12923136085789285.1292311816393261.79E-77passed
/families/3/percentiles/2/lowertolerance26.34850262592631226.348502891100592.65E-77passed
/families/3/percentiles/2/uppertolerance46.7309062660568246.7309057996545154.66E-78passed
/families/3/percentiles/3/estimatetolerance70.85399128322096470.85399128588222.66E-910passed
/families/3/percentiles/3/standard_errortolerance7.14371416710065087.1437139680641951.99E-77passed
/families/3/percentiles/3/lowertolerance58.14918067686369358.1491809996341263.23E-78passed
/families/3/percentiles/3/uppertolerance86.33463141399373386.33463094125884.73E-78passed
/families/4/parameters/0/estimatetolerance76.74578986643415176.745789788428927.80E-88passed
/families/4/parameters/0/standard_errortolerance9.70966224494761229.7096620755941961.69E-77passed
/families/4/parameters/0/lowertolerance57.71520156428850657.715201818209872.54E-78passed
/families/4/parameters/0/uppertolerance95.77637816857979795.776377758647984.10E-78passed
/families/4/parameters/1/estimatetolerance31.86086110560322431.860861149589544.40E-88passed
/families/4/parameters/1/standard_errortolerance5.33300585429231165.3330057750465617.92E-87passed
/families/4/parameters/1/lowertolerance22.9498467803042722.9498469342611561.54E-78passed
/families/4/parameters/1/uppertolerance44.23186263978543344.231862465190971.75E-78passed
/families/4/log_likelihoodtolerance-161.20621692962214-161.206216929622122.43E-1415passed
/families/4/anderson_darlingtolerance2.28508213744902112.2850821381961597.47E-109passed
/families/4/correlationtolerance0.8280839628298120.82808396282981182.18E-1615passed
/families/4/percentiles/0/estimatetolerance-69.658685442304304-69.658685722431942.80E-78passed
/families/4/percentiles/0/standard_errortolerance25.01923516801277425.0192348668601743.01E-77passed
/families/4/percentiles/0/lowertolerance-118.69548529234726-118.695484982226653.10E-78passed
/families/4/percentiles/0/uppertolerance-20.62188559226135-20.6218864626372288.70E-77passed
/families/4/percentiles/1/estimatetolerance-17.06657148271195-17.0665716902321952.08E-77passed
/families/4/percentiles/1/standard_errortolerance17.22601146210135817.226011266779091.95E-77passed
/families/4/percentiles/1/lowertolerance-50.828933545704764-50.82893337040041.75E-78passed
/families/4/percentiles/1/uppertolerance16.69579058028086416.695789989936015.90E-77passed
/families/4/percentiles/2/estimatetolerance6.7403227901071296.740322615454091.75E-77passed
/families/4/percentiles/2/standard_errortolerance14.09479143830946214.0947912797957551.59E-77passed
/families/4/percentiles/2/lowertolerance-20.884960798582917-20.8849606625548071.36E-78passed
/families/4/percentiles/2/uppertolerance34.36560637879717534.3656058934629844.85E-77passed
/families/4/percentiles/3/estimatetolerance76.74578986643415176.745789788428927.80E-88passed
/families/4/percentiles/3/standard_errortolerance9.70966224494761229.7096620755941961.69E-77passed
/families/4/percentiles/3/lowertolerance57.71520156428850657.715201818209872.54E-78passed
/families/4/percentiles/3/uppertolerance95.77637816857979795.776377758647984.10E-78passed
/families/5/parameters/0/estimatetolerance91.91034482758620191.910344821655855.93E-910passed
/families/5/parameters/0/standard_errortolerance13.60835613112760713.6083552362292018.95E-77passed
/families/5/parameters/0/lowertolerance65.23845692178126465.238458669819561.75E-67passed
/families/5/parameters/0/uppertolerance118.58223273339114118.582230973492141.76E-67passed
/families/5/parameters/1/estimatetolerance73.28324052030144773.283240523976543.68E-910passed
/families/5/parameters/1/standard_errortolerance9.62256090102864739.6225605851634343.16E-77passed
/families/5/parameters/1/lowertolerance56.65476297232730556.654763454509544.82E-78passed
/families/5/parameters/1/uppertolerance94.79226563492126494.79226483766267.97E-78passed
/families/5/log_likelihoodtolerance-165.68484373661835-165.684843736618323.09E-1415passed
/families/5/anderson_darlingtolerance3.43235662678574153.432356626420733.65E-109passed
/families/5/correlationtolerance0.817059835675540190.81705983567554021.42E-1817passed
/families/5/percentiles/0/estimatetolerance-78.571965959640679-78.571965974120561.45E-89passed
/families/5/percentiles/0/standard_errortolerance26.1972244427107726.1972233499544451.09E-67passed
/families/5/percentiles/0/lowertolerance-129.91758236226616-129.9175802349832.13E-67passed
/families/5/percentiles/0/uppertolerance-27.22634955701519-27.2263517132581042.16E-67passed
/families/5/percentiles/1/estimatetolerance-28.629859136988756-28.6298591489641151.20E-89passed
/families/5/percentiles/1/standard_errortolerance20.87351376243494320.87351278505229.77E-77passed
/families/5/percentiles/1/lowertolerance-69.541194342162385-69.541192438502781.90E-67passed
/families/5/percentiles/1/uppertolerance12.2814760681848812.2814741405745521.93E-66passed
/families/5/percentiles/2/estimatetolerance-2.0059067893876517-2.00590680002778751.06E-88passed
/families/5/percentiles/2/standard_errortolerance18.36466293799822618.364662003051479.35E-77passed
/families/5/percentiles/2/lowertolerance-37.9999847360817-37.999982914259881.82E-67passed
/families/5/percentiles/2/uppertolerance33.98817115730639733.9881693142043061.84E-67passed
/families/5/percentiles/3/estimatetolerance91.91034482758620191.910344821655855.93E-910passed
/families/5/percentiles/3/standard_errortolerance13.60835613112760713.6083552362292018.95E-77passed
/families/5/percentiles/3/lowertolerance65.23845692178126465.238458669819561.75E-67passed
/families/5/percentiles/3/uppertolerance118.58223273339114118.582230973492141.76E-67passed
/families/6/parameters/0/estimatetolerance135.08981907760682135.089819009336146.83E-89passed
/families/6/parameters/0/standard_errortolerance20.50323083496558120.503227755949013.08E-66passed
/families/6/parameters/0/lowertolerance94.90422507436318394.904231040854095.97E-67passed
/families/6/parameters/0/uppertolerance175.27541308085046175.27540697781826.10E-67passed
/families/6/parameters/1/estimatetolerance103.32865952037881103.328659573319815.29E-89passed
/families/6/parameters/1/standard_errortolerance12.41352499771568512.4135245856216054.12E-77passed
/families/6/parameters/1/lowertolerance81.65082432098948681.65082501091456.90E-78passed
/families/6/parameters/1/uppertolerance130.76183819412765130.761837223222369.71E-78passed
/families/6/log_likelihoodtolerance-175.6181871797304-175.618187179730373.00E-1415passed
/families/6/anderson_darlingtolerance4.70159428244912284.701594280826961.62E-99passed
/families/6/correlationtolerance0.715635752387925870.7156357523879261.13E-1615passed
/families/6/percentiles/0/estimatetolerance-340.23743411892411-340.237434430731473.12E-79passed
/families/6/percentiles/0/standard_errortolerance67.1307426075285667.130738841780783.77E-67passed
/families/6/percentiles/0/lowertolerance-471.81127188510857-471.811264816185947.07E-67passed
/families/6/percentiles/0/uppertolerance-208.66359635273969-208.663604045277037.69E-67passed
/families/6/percentiles/1/estimatetolerance-171.81647451971764-171.816474745233532.26E-78passed
/families/6/percentiles/1/standard_errortolerance48.09037762052977948.090374276255143.34E-67passed
/families/6/percentiles/1/lowertolerance-266.07188265888703-266.071876329745076.33E-67passed
/families/6/percentiles/1/uppertolerance-77.561066380548283-77.561073160721996.78E-67passed
/families/6/percentiles/2/estimatetolerance-97.437620282045714-97.43762046945311.87E-78passed
/families/6/percentiles/2/standard_errortolerance40.05581582249366340.055812605387213.22E-67passed
/families/6/percentiles/2/lowertolerance-175.94557666550293-175.945570547497546.12E-67passed
/families/6/percentiles/2/uppertolerance-18.929663898588515-18.929670391408656.49E-66passed
/families/6/percentiles/3/estimatetolerance97.21853029700437897.218530209330148.77E-89passed
/families/6/percentiles/3/standard_errortolerance22.51301497490210222.5130118520995353.12E-66passed
/families/6/percentiles/3/lowertolerance53.09383176278535653.093837795691686.03E-66passed
/families/6/percentiles/3/uppertolerance141.3432288312234141.34322262296866.21E-67passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

526  ALL7 <- c("weibull", "lognormal", "exponential", "loglogistic", "logistic", "normal", "sev")
527  DISTID_NOTES <- c(
528    paste0(MANUAL, ":53-62: probability-plot positions follow Benard median ranks (i - 0.3)/(n + 0.4) for complete data, with frequencies expanding ranks; the Anderson-Darling statistic is the complete-data form; no family is selected automatically."),
529    "R: survreg per family; A2 computed from its definition by the exact piecewise closed form c^2 log(u2/u1) - (1-c)^2 log((1-u2)/(1-u1)) - (u2-u1); the correlation is exact arithmetic on the plotted points (1e-9 tolerance).",
530    "Iterative maximum likelihood: 1e-7 absolute or 1e-6 relative elsewhere."
531  )
532  emit("STAT-SURV-DISTID-001",
533       "Distribution identification of complete data across the seven families",
534       DISTID, "lifeDistributionIdMenuItem", DISTID_REQ, "SURV_COMPLETE_B9",
535       list(layout = "arbitrary", start = "time", end = "time", frequency = "freq",
536            families = as.list(ALL7), threshold = 0.0, confidence_level = 0.95),
537       c(list("/plotted_points" = exact(sum(cd$freq), "frequencies expand the ranks")),
538         ident_complete(ALL7, 0.0, 0.95)),
539       DISTID_NOTES, "nominal")

Whole script generators/r/survival_life_families.R


Distribution identification of right-censored data with the restricted-domain AD

Case STAT-SURV-DISTID-002 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_KM_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/plotted_pointsexact13130.00E+2passed
/families/0/parameters/0/estimatetolerance1.7766991007375251.77669910038536033.52E-109passed
/families/0/parameters/0/standard_errortolerance0.378471264827343450.3784712347436123.01E-87passed
/families/0/parameters/0/lowertolerance1.17028461043808731.17028464894732353.85E-87passed
/families/0/parameters/0/uppertolerance2.697343591812132.69734350198439288.98E-87passed
/families/0/parameters/1/estimatetolerance156.74130136193452156.741301343717231.82E-89passed
/families/0/parameters/1/standard_errortolerance23.98465000044372923.9846481273575561.87E-67passed
/families/0/parameters/1/lowertolerance116.12669508527753116.126697787642412.70E-67passed
/families/0/parameters/1/uppertolerance211.56061949917211211.560614526803084.97E-67passed
/families/0/parameters/2/estimatetolerance00.00.00E+2passed
/families/0/log_likelihoodtolerance-84.270365465784479-84.270365465784471.38E-1415passed
/families/0/anderson_darlingtolerance0.905035657636814710.90503565701964346.17E-109passed
/families/0/ad_domain_probabilitytolerance0.916066281606997880.91606628161254025.54E-1211passed
/families/0/correlationtolerance0.97309520484033840.97309520484033841.38E-1817passed
/families/0/percentiles/0/estimatetolerance11.76867992819429711.7686799207867387.41E-99passed
/families/0/percentiles/0/standard_errortolerance6.40997116550469676.4099707410462274.24E-77passed
/families/0/percentiles/0/lowertolerance4.04677710662792484.04677738742668952.81E-77passed
/families/0/percentiles/0/uppertolerance34.22521764923527634.225215231322892.42E-67passed
/families/0/percentiles/1/estimatetolerance29.45452395641000429.454523943226521.32E-89passed
/families/0/percentiles/1/standard_errortolerance10.63007872682721610.6300780931528296.34E-77passed
/families/0/percentiles/1/lowertolerance14.51961414515272114.5196147462908326.01E-77passed
/families/0/percentiles/1/uppertolerance59.75151769362710859.751515166318882.53E-67passed
/families/0/percentiles/2/estimatetolerance44.16782335704149944.1678233408194461.62E-89passed
/families/0/percentiles/2/standard_errortolerance12.57586845465608912.5758677614179716.93E-77passed
/families/0/percentiles/2/lowertolerance25.2780793910309625.278080154187777.63E-77passed
/families/0/percentiles/2/uppertolerance77.17345095414931777.173448567558652.39E-67passed
/families/0/percentiles/3/estimatetolerance127.52443580132935127.524435781293432.00E-89passed
/families/0/percentiles/3/standard_errortolerance19.28958558916858119.289584349760751.24E-67passed
/families/0/percentiles/3/lowertolerance94.80670810920645494.806709895854311.79E-67passed
/families/0/percentiles/3/uppertolerance171.53302810297828171.53302481650993.29E-67passed
/families/1/parameters/0/estimatetolerance4.79434759182165944.794347592114412.93E-1010passed
/families/1/parameters/0/standard_errortolerance0.173131812437139380.17313175876986065.37E-86passed
/families/1/parameters/0/lowertolerance4.45501547486672234.4550155803454071.05E-77passed
/families/1/parameters/0/uppertolerance5.13367970877659645.1336796038834131.05E-77passed
/families/1/parameters/1/estimatetolerance0.736015741707313320.7360157418222571.15E-109passed
/families/1/parameters/1/standard_errortolerance0.145120241477383370.14512022433547281.71E-86passed
/families/1/parameters/1/lowertolerance0.500098640899526490.50009866383623082.29E-87passed
/families/1/parameters/1/uppertolerance1.083224643575471.08322459423240154.93E-87passed
/families/1/parameters/2/estimatetolerance00.00.00E+2passed
/families/1/log_likelihoodtolerance-83.589329184123116-83.589329184123092.83E-1415passed
/families/1/anderson_darlingtolerance0.827834800868237240.82783480171950458.51E-108passed
/families/1/ad_domain_probabilitytolerance0.852553682775485160.85255368264609951.29E-109passed
/families/1/correlationtolerance0.966431373732689840.966431373732691.07E-1615passed
/families/1/percentiles/0/estimatetolerance21.80455593578705421.804555936339845.53E-1010passed
/families/1/percentiles/0/standard_errortolerance7.06408548641886917.0640851354380893.51E-77passed
/families/1/percentiles/0/lowertolerance11.55529410362212811.5552944686578773.65E-77passed
/families/1/percentiles/0/uppertolerance41.14466107858174141.1446597808938461.30E-67passed
/families/1/percentiles/1/estimatetolerance36.00685833370140236.0068583374347643.73E-99passed
/families/1/percentiles/1/standard_errortolerance8.7745239037965498.774523369463985.34E-77passed
/families/1/percentiles/1/lowertolerance22.33344030582786422.3334409588257586.53E-77passed
/families/1/percentiles/1/uppertolerance58.05168524461166958.051683559301471.69E-67passed
/families/1/percentiles/2/estimatetolerance47.04504117909233747.04504118593486.84E-99passed
/families/1/percentiles/2/standard_errortolerance9.76496678583495389.7649659449366228.41E-77passed
/families/1/percentiles/2/lowertolerance31.3208159998928431.3208171035647321.10E-67passed
/families/1/percentiles/2/uppertolerance70.66341756709228270.663415097635012.47E-67passed
/families/1/percentiles/3/estimatetolerance120.82552850064192120.825528536013683.54E-89passed
/families/1/percentiles/3/standard_errortolerance20.91874273799137320.9187362597380266.48E-66passed
/families/1/percentiles/3/lowertolerance86.05748256290738186.057491640137919.08E-66passed
/families/1/percentiles/3/uppertolerance169.64019748994878169.64017969584941.78E-56passed
/families/2/parameters/0/estimatetolerance184.23571428571086184.235714299273841.36E-810passed
/families/2/parameters/0/standard_errortolerance49.23906580462121249.239065808246533.63E-910passed
/families/2/parameters/0/lowertolerance109.11404122323711109.114041231269248.03E-910passed
/families/2/parameters/0/uppertolerance311.07635679006972311.076356812972052.29E-810passed
/families/2/parameters/1/estimatetolerance00.00.00E+2passed
/families/2/log_likelihoodtolerance-87.027023911206484-87.027023911206482.86E-1617passed
/families/2/anderson_darlingtolerance1.77372729800315951.77372729806143455.83E-1110passed
/families/2/ad_domain_probabilitytolerance0.757741272412476350.75774127238719162.53E-1110passed
/families/2/correlationtolerance0.97309520484033840.97309520484033841.38E-1817passed
/families/2/percentiles/0/estimatetolerance1.85163080478112961.85163080491744061.36E-1010passed
/families/2/percentiles/0/standard_errortolerance0.49486914844910190.49486914848553723.64E-1110passed
/families/2/percentiles/0/lowertolerance1.09663276062633511.096632760707068.07E-1110passed
/families/2/percentiles/0/uppertolerance3.12642186182384843.1264218620540212.30E-1010passed
/families/2/percentiles/1/estimatetolerance9.4500567295576259.4500567302533076.96E-1010passed
/families/2/percentiles/1/standard_errortolerance2.52563389768441132.52563389787036341.86E-1010passed
/families/2/percentiles/1/lowertolerance5.59681863827882795.596818638690824.12E-1010passed
/families/2/percentiles/1/uppertolerance15.95613114583977115.9561311470144871.17E-910passed
/families/2/percentiles/2/estimatetolerance19.4111698597304519.4111698611594541.43E-910passed
/families/2/percentiles/2/standard_errortolerance5.18785336368453195.1878533640664983.82E-1010passed
/families/2/percentiles/2/lowertolerance11.49631164878957711.4963116496358468.46E-1010passed
/families/2/percentiles/2/uppertolerance32.77516536035970132.7751653627727042.41E-910passed
/families/2/percentiles/3/estimatetolerance127.70246591558811127.702465924989269.40E-910passed
/families/2/percentiles/3/standard_errortolerance34.12991963587880734.1299196383916842.51E-910passed
/families/2/percentiles/3/lowertolerance75.63209003338845175.63209003895595.57E-910passed
/families/2/percentiles/3/uppertolerance215.62169964789643215.621699663771121.59E-810passed
/families/3/parameters/0/estimatetolerance4.7966473157968834.7966473154212433.76E-1010passed
/families/3/parameters/0/standard_errortolerance0.166363312874681560.16636329567187211.72E-86passed
/families/3/parameters/0/lowertolerance4.47058121421373844.47058124755498553.33E-88passed
/families/3/parameters/0/uppertolerance5.12271341738002755.12271338328750053.41E-88passed
/families/3/parameters/1/estimatetolerance0.428498734481800920.42849873474641752.65E-109passed
/families/3/parameters/1/standard_errortolerance0.0939553100150310880.093955306562737573.45E-97passed
/families/3/parameters/1/lowertolerance0.278810916797955390.278810921446796374.65E-97passed
/families/3/parameters/1/uppertolerance0.658550847151948340.65855083698476261.02E-87passed
/families/3/parameters/2/estimatetolerance00.00.00E+2passed
/families/3/log_likelihoodtolerance-83.790055320148738-83.790055320148721.39E-1415passed
/families/3/anderson_darlingtolerance0.79050246086561660.79050246003642518.29E-108passed
/families/3/ad_domain_probabilitytolerance0.857391980067242150.85739198003898492.83E-1110passed
/families/3/correlationtolerance0.96798327895742320.96798327895742331.10E-1615passed
/families/3/percentiles/0/estimatetolerance16.90558896460522916.9055889376985462.69E-88passed
/families/3/percentiles/0/standard_errortolerance7.25045545553507867.2504552799545941.76E-77passed
/families/3/percentiles/0/lowertolerance7.29403208544075817.29403221255111.27E-77passed
/families/3/percentiles/0/uppertolerance39.18257212093303839.182571313388398.08E-77passed
/families/3/percentiles/1/estimatetolerance34.29360657149267434.2936065318908663.96E-88passed
/families/3/percentiles/1/standard_errortolerance9.99231436297768039.9923141021364542.61E-77passed
/families/3/percentiles/1/lowertolerance19.37286916427764519.372869417935322.54E-77passed
/families/3/percentiles/1/uppertolerance60.70610613779862860.706105202741149.35E-77passed
/families/3/percentiles/2/estimatetolerance47.23519941784800847.2351993726414.52E-89passed
/families/3/percentiles/2/standard_errortolerance11.16640247621570111.1664021109861783.65E-77passed
/families/3/percentiles/2/lowertolerance29.71951271176447729.7195131205334174.09E-77passed
/families/3/percentiles/2/uppertolerance75.07404598732405575.074044811037551.18E-67passed
/families/3/percentiles/3/estimatetolerance121.10371361724117121.103713571749784.55E-89passed
/families/3/percentiles/3/standard_errortolerance20.14721499879092520.147212907898722.09E-66passed
/families/3/percentiles/3/lowertolerance87.40751074036670387.407513654642162.91E-67passed
/families/3/percentiles/3/uppertolerance167.79003689340436167.79003117301815.72E-67passed
/families/4/parameters/0/estimatetolerance126.99970836938374126.999708756174133.87E-78passed
/families/4/parameters/0/standard_errortolerance15.90385051678009115.903849647712318.69E-77passed
/families/4/parameters/0/lowertolerance95.82873414098604395.828736231117972.09E-67passed
/families/4/parameters/0/uppertolerance158.17068259778142158.17068128123031.32E-68passed
/families/4/parameters/1/estimatetolerance40.35007999481484140.3500801681774061.73E-78passed
/families/4/parameters/1/standard_errortolerance8.81282711427046918.8128265011045926.13E-77passed
/families/4/parameters/1/lowertolerance26.29869420529735226.2986951499352139.45E-77passed
/families/4/parameters/1/uppertolerance61.90911772570071761.909116033931991.69E-67passed
/families/4/log_likelihoodtolerance-86.419610492035673-86.419610492035671.33E-1617passed
/families/4/anderson_darlingtolerance0.910034886091827030.91003489526391399.17E-97passed
/families/4/ad_domain_probabilitytolerance0.965306691058460340.96530669025887868.00E-109passed
/families/4/correlationtolerance0.946857779932885450.94685777993288531.15E-1615passed
/families/4/percentiles/0/estimatetolerance-58.413745169308555-58.413745579139914.10E-78passed
/families/4/percentiles/0/standard_errortolerance39.48185311947595739.481850643573942.48E-67passed
/families/4/percentiles/0/lowertolerance-135.7967553263818-135.79675088353444.44E-67passed
/families/4/percentiles/0/uppertolerance18.96926498776468918.9692597252545745.26E-66passed
/families/4/percentiles/1/estimatetolerance8.19136002016692548.1913598965018221.24E-77passed
/families/4/percentiles/1/standard_errortolerance26.6876153270788726.6876137780839071.55E-67passed
/families/4/percentiles/1/lowertolerance-44.115404854166783-44.1154019418575562.91E-67passed
/families/4/percentiles/1/uppertolerance60.49812489450063460.49812173486123.16E-67passed
/families/4/percentiles/2/estimatetolerance38.34152090729406838.341520913167975.87E-99passed
/families/4/percentiles/2/standard_errortolerance21.63702361136188321.6370224335652281.18E-67passed
/families/4/percentiles/2/lowertolerance-4.0662661036179912-4.0662637893050762.31E-66passed
/families/4/percentiles/2/uppertolerance80.74930791820612880.749305615641022.30E-67passed
/families/4/percentiles/3/estimatetolerance126.99970836938374126.999708756174133.87E-78passed
/families/4/percentiles/3/standard_errortolerance15.90385051678009115.903849647712318.69E-77passed
/families/4/percentiles/3/lowertolerance95.82873414098604395.828736231117972.09E-67passed
/families/4/percentiles/3/uppertolerance158.17068259778142158.17068128123031.32E-68passed
/families/5/parameters/0/estimatetolerance131.829561272468131.829561313363464.09E-89passed
/families/5/parameters/0/standard_errortolerance16.42097064063963816.420969885626077.55E-77passed
/families/5/parameters/0/lowertolerance99.64505022562470199.645051746319551.52E-67passed
/families/5/parameters/0/uppertolerance164.01407231931131164.014070880407361.44E-68passed
/families/5/parameters/1/estimatetolerance69.79489564438728269.794895670570262.62E-89passed
/families/5/parameters/1/standard_errortolerance13.71023828600524813.7102375812757157.05E-77passed
/families/5/parameters/1/lowertolerance47.49161476670870747.491615731245579.65E-77passed
/families/5/parameters/1/uppertolerance102.57236950017706102.572367493929152.01E-67passed
/families/5/log_likelihoodtolerance-86.276874204252863-86.276874204252869.01E-1717passed
/families/5/anderson_darlingtolerance1.13822964485616331.13822964645433181.60E-98passed
/families/5/ad_domain_probabilitytolerance0.968100380919538980.96810038082781299.17E-1110passed
/families/5/correlationtolerance0.960551368524957150.96055136852495711.70E-1918passed
/families/5/percentiles/0/estimatetolerance-30.537645828754677-30.5376458487699272.00E-89passed
/families/5/percentiles/0/standard_errortolerance30.71771306838724830.7177121296565589.39E-77passed
/families/5/percentiles/0/lowertolerance-90.743257130229026-90.743255310365951.82E-67passed
/families/5/percentiles/0/uppertolerance29.66796547271967229.6679636128260921.86E-67passed
/families/5/percentiles/1/estimatetolerance17.02717402909803517.0271740269263182.17E-99passed
/families/5/percentiles/1/standard_errortolerance23.1396993771563423.1396988071567125.70E-77passed
/families/5/percentiles/1/lowertolerance-28.325803363212302-28.3258022482052851.12E-67passed
/families/5/percentiles/1/uppertolerance62.38015142140837162.380150302057921.12E-67passed
/families/5/percentiles/2/estimatetolerance42.38380349238143242.383803499722097.34E-99passed
/families/5/percentiles/2/standard_errortolerance19.72868928494049419.7286888532723064.32E-77passed
/families/5/percentiles/2/lowertolerance3.71628303171679873.7162838851115518.53E-76passed
/families/5/percentiles/2/uppertolerance81.05132395304606581.051323114332638.39E-77passed
/families/5/percentiles/3/estimatetolerance131.829561272468131.829561313363464.09E-89passed
/families/5/percentiles/3/standard_errortolerance16.42097064063963816.420969885626077.55E-77passed
/families/5/percentiles/3/lowertolerance99.64505022562470199.645051746319551.52E-67passed
/families/5/percentiles/3/uppertolerance164.01407231931131164.014070880407361.44E-68passed
/families/6/parameters/0/estimatetolerance168.31672858951427168.316728597164177.65E-910passed
/families/6/parameters/0/standard_errortolerance18.59892216614508318.5989214477443747.18E-77passed
/families/6/parameters/0/lowertolerance131.86351099260622131.863512408295631.42E-67passed
/families/6/parameters/0/uppertolerance204.76994618642232204.76994478603271.40E-68passed
/families/6/parameters/1/estimatetolerance69.25219513550675369.252195137202421.70E-910passed
/families/6/parameters/1/standard_errortolerance13.0059143337831113.0059138189408155.15E-77passed
/families/6/parameters/1/lowertolerance47.92626299987876647.926263699816857.00E-77passed
/families/6/parameters/1/uppertolerance100.06760867415089100.067607217616031.46E-67passed
/families/6/log_likelihoodtolerance-89.301118539969181-89.301118539969152.83E-1415passed
/families/6/anderson_darlingtolerance2.0257488726693842.025748872999793.30E-109passed
/families/6/ad_domain_probabilitytolerance0.978154319947010740.97815431993504011.20E-1110passed
/families/6/correlationtolerance0.930593281657563050.93059328165756291.07E-1615passed
/families/6/percentiles/0/estimatetolerance-150.25370331566788-150.253703315818371.50E-1011passed
/families/6/percentiles/0/standard_errortolerance60.87835718805244860.8783556666339151.52E-67passed
/families/6/percentiles/0/lowertolerance-269.57309084221578-269.573087860440752.98E-67passed
/families/6/percentiles/0/uppertolerance-30.934315789119992-30.934318771195992.98E-67passed
/families/6/percentiles/1/estimatetolerance-37.37581238770872-37.3758123850953442.61E-910passed
/families/6/percentiles/1/standard_errortolerance41.19012674803354241.190125968226127.80E-77passed
/families/6/percentiles/1/lowertolerance-118.10697733249438-118.106975801486561.53E-67passed
/families/6/percentiles/1/uppertolerance43.35535255707694143.355351031295871.53E-67passed
/families/6/percentiles/2/estimatetolerance12.47385131190401412.4738513157380453.83E-99passed
/families/6/percentiles/2/standard_errortolerance33.09492233292996133.094921827491285.05E-77passed
/families/6/percentiles/2/lowertolerance-52.391004531788994-52.391003537313379.94E-77passed
/families/6/percentiles/2/uppertolerance77.33870715559702177.338706168789469.87E-77passed
/families/6/percentiles/3/estimatetolerance142.93490429370837142.934904300736787.03E-910passed
/families/6/percentiles/3/standard_errortolerance18.73959161144134518.7395911158920044.96E-77passed
/families/6/percentiles/3/lowertolerance106.20597965029444106.205980628581689.78E-78passed
/families/6/percentiles/3/uppertolerance179.6638289371223179.663827972891859.64E-78passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

589  emit("STAT-SURV-DISTID-002",
590       "Distribution identification of right-censored data with the restricted-domain AD",
591       DISTID, "lifeDistributionIdMenuItem", DISTID_REQ, "SURV_KM_B9",
592       list(layout = "right_censored", start = "time", status = "status", frequency = "freq",
593            families = as.list(ALL7), threshold = 0.0, confidence_level = 0.95),
594       c(list("/plotted_points" = exact(km_failure_times(), "one point per failure time")),
595         ident_censored(ALL7, 0.0, 0.95)),
596       c(DISTID_NOTES,
597         paste0(MANUAL, ":56-62: under right censoring the discrepancy integrates only to the largest observation; the covered fitted probability is reported and no p-value is given."),
598         "Plotting positions are the Kaplan-Meier midpoints (F(t-) + F(t))/2 at failure times."), "axis_coverage")

Whole script generators/r/survival_life_families.R


Parametric survival STAT-SURV-FAMILIES-001

Case STAT-SURV-FAMILIES-001 · status: passed

Reference

Evidence type
R
Source
R survival::survreg
Citation
R survival::survreg and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_parametric.R · sha256 7c99eb709767fd04
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_nexact32320.00E+2passed
/failuresexact26260.00E+2passed
/fits/[family=exponential]/param1_valtolerance17.38079623664851417.3807962366485142.52E-1616passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

24  exp_scale <- sum(d$time)/sum(d$status)
25  emit("STAT-SURV-FAMILIES-001","tabelz.executor.reliability","reliabilityFamiliesMenuItem",
26       "SURV_CORE",list(mode="families",time_column="time",event_column="status"),
27       list("/total_n"=exact(nrow(d),"All finite right-censored observations"),
28            "/failures"=exact(sum(d$status),"Observed failures"),
29            "/fits/[family=exponential]/param1_val"=tol(exp_scale,
30                  "Right-censored exponential scale=sum time on test/failures","1e-9")),
31       c("The exponential MLE is analytic; right-censored records contribute survival probability."))

Whole script generators/r/survival_parametric.R


Hard-numeric survival STAT-SURV-FAMILIES-HARD-001

Case STAT-SURV-FAMILIES-HARD-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_hard.R · sha256 bc20ccd17540f4e7
Dataset
reference_data/datasets/SURV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[family=exponential]/param1_valtolerance1253510.87505676411253510.8750567641.09E-1117passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

69  emit("STAT-SURV-FAMILIES-HARD-001","tabelz.executor.reliability",
70       "reliabilityFamiliesMenuItem","SURV_HARD",
71       list(mode="families",time_column="time",event_column="status"),
72       list("/fits/[family=exponential]/param1_val"=tol(sum(d$time)/sum(d$status),
73          "Large-offset exponential time-on-test scale","1e-9")),
74       c("Exponential scale=sum time on test/failures remains analytic at large lifetime offset."))

Whole script generators/r/survival_hard.R


Small-sample survival boundary STAT-SURV-FAMILIES-MIN-001

Case STAT-SURV-FAMILIES-MIN-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::survfit/coxph and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_minimum.R · sha256 2997d37929a6dae8
Dataset
reference_data/datasets/SURV_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_nexact330.00E+2passed
/failuresexact220.00E+2passed
/fits/[family=exponential]/param1_valtolerance33.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

50  emit("STAT-SURV-FAMILIES-MIN-001","tabelz.executor.reliability",
51       "reliabilityFamiliesMenuItem","SURV_MIN",3:4,
52       list(mode="families",time_column="time",event_column="status"),
53       list("/total_n"=exact(3L,"Minimum three retained lifetimes"),
54            "/failures"=exact(2L,"Minimum two observed failures"),
55            "/fits/[family=exponential]/param1_val"=tol(sum(d$time[1:3])/sum(d$status[1:3]),
56               "Analytic exponential total-time/failure scale","1e-9")),
57       c("Three rows and two failures meet censored-families minimum."))

Whole script generators/r/survival_minimum.R


Survival boundary STAT-SURV-FAMILIES-ONE-FAIL-001

Case STAT-SURV-FAMILIES-ONE-FAIL-001 · status: passed

Reference

Evidence type
R
Source
R survival and analytic boundary
Citation
R survival::survfit/coxph/survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_boundaries.R · sha256 528007972452fb15
Dataset
reference_data/datasets/SURV_ONE_FAIL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

73  emit("STAT-SURV-FAMILIES-ONE-FAIL-001","tabelz.executor.reliability",
74       "reliabilityFamiliesMenuItem","SURV_ONE_FAIL",
75       list(mode="families",time_column="time",event_column="status"),list(),
76       c("Censored life families require at least two failures for parameter estimation."),
77       error=list(error_class="ReliabilityExecutionError",message_contains="At least two observed failures"))

Whole script generators/r/survival_boundaries.R


Survival population STAT-SURV-FAMILIES-POP-001

Case STAT-SURV-FAMILIES-POP-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::Surv/survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_population.R · sha256 3b3b4629e272fcd9
Dataset
reference_data/datasets/SURV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/input_row_countexact31310.00E+2passed
/omitted_row_countexact220.00E+2passed
/total_nexact29290.00E+2passed
/fits/[family=exponential]/param1_valtolerance17.97677123903956117.976771239039561.32E-1617passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

75  exp_scale <- sum(kmdata$time)/sum(kmdata$status)
76  emit("STAT-SURV-FAMILIES-POP-001","tabelz.executor.reliability",
77       "reliabilityFamiliesMenuItem","SURV_POP",7L,
78       list(mode="families",time_column="time",event_column="status"),
79       list("/input_row_count"=exact(nrow(eligible),"Rows sent to worker after exclusion"),
80            "/omitted_row_count"=exact(2L,"Rows with missing lifetime or status"),
81            "/total_n"=exact(nrow(kmdata),"Finite complete lifetimes"),
82            "/fits/[family=exponential]/param1_val"=tol(exp_scale,
83                 "Exponential total time/failures","1e-9")),
84       c("The exponential right-censored MLE is total time on test divided by observed failures."))

Whole script generators/r/survival_population.R


Parametric survival STAT-SURV-INTERVAL-001

Case STAT-SURV-INTERVAL-001 · status: passed

Reference

Evidence type
R
Source
R survival::survreg
Citation
R survival::survreg and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_parametric.R · sha256 7c99eb709767fd04
Dataset
reference_data/datasets/SURV_INTERVAL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_nexact20200.00E+2passed
/exact_countexact550.00E+2passed
/right_countexact550.00E+2passed
/left_countexact550.00E+2passed
/interval_countexact550.00E+2passed
/fits/[family=exponential]/param1_valtolerance5.04308579226238775.0430857276036686.47E-87passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

42  interval <- data.frame(lower=lower,upper=upper)
43  export_dataset(interval,"SURV_INTERVAL","Constructed exact/right/left/interval lifetime matrix",
44                 "R survival::Surv(type=interval2) and survreg exponential model",
45                 "Constructed validation data; no third-party license.",
46                 c(lower="Float64",upper="Float64"))
47  interval <- read_dataset("SURV_INTERVAL")
48  fit_int <- survival::survreg(survival::Surv(lower,upper,type="interval2")~1,
49                              data=interval,dist="exponential",control=survival::survreg.control(maxiter=200L))
50  emit("STAT-SURV-INTERVAL-001","tabelz.executor.reliability","intervalLifeMenuItem",
51       "SURV_INTERVAL",list(mode="interval",lower_column="lower",upper_column="upper"),
52       list("/total_n"=exact(nrow(interval),"All four censoring types retained"),
53            "/exact_count"=exact(5L,"Five exact failures"),
54            "/right_count"=exact(5L,"Five right-censored rows"),
55            "/left_count"=exact(5L,"Five left-censored rows"),
56            "/interval_count"=exact(5L,"Five interval-censored rows"),
57            "/fits/[family=exponential]/param1_val"=tol(exp(coef(fit_int)[[1]]),
58                                                       "Interval2 exponential scale MLE")),
59       c("R Surv(type=interval2): NA lower is left censoring, NA upper is right censoring.",
60         "survreg exponential intercept is log scale, so exp(intercept) is Tabelz scale."))

Whole script generators/r/survival_parametric.R


Hard-numeric survival STAT-SURV-INTERVAL-HARD-001

Case STAT-SURV-INTERVAL-HARD-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_hard.R · sha256 bc20ccd17540f4e7
Dataset
reference_data/datasets/SURV_INTERVAL_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/fits/[family=exponential]/param1_valtolerance1142666.76794608641142666.76794915743.07E-611passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 98  fit_int <- survival::survreg(survival::Surv(lower,upper,type="interval2")~1,
 99                               data=interval,dist="exponential",control=survival::survreg.control(rel.tolerance=1e-14,maxiter=1000))
100  emit("STAT-SURV-INTERVAL-HARD-001","tabelz.executor.reliability","intervalLifeMenuItem",
101       "SURV_INTERVAL_HARD",list(mode="interval",lower_column="lower",upper_column="upper"),
102       list("/fits/[family=exponential]/param1_val"=tol(exp(coef(fit_int)[[1]]),
103            "Large-offset mixed-censoring exponential scale")),
104       c("All positive endpoints have a 1e6 offset but retain their censoring interval widths."))

Whole script generators/r/survival_hard.R


Small-sample survival boundary STAT-SURV-INTERVAL-MIN-001

Case STAT-SURV-INTERVAL-MIN-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::survfit/coxph and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_minimum.R · sha256 2997d37929a6dae8
Dataset
reference_data/datasets/SURV_INTERVAL.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/total_nexact330.00E+2passed
/exact_countexact110.00E+2passed
/left_countexact110.00E+2passed
/right_countexact110.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

78  valid_int <- inter[1:3,,drop=FALSE]
79  emit("STAT-SURV-INTERVAL-MIN-001","tabelz.executor.reliability","intervalLifeMenuItem",
80       "SURV_INTERVAL",3:19,
81       list(mode="interval",lower_column="lower",upper_column="upper"),
82       list("/total_n"=exact(nrow(valid_int),"Minimum three retained interval records"),
83            "/exact_count"=exact(1L,"One exact failure"),
84            "/left_count"=exact(1L,"One left-censored failure-indicative row"),
85            "/right_count"=exact(1L,"One right-censored row")),
86       c("Exact plus left censoring supplies two failure-indicative records."))

Whole script generators/r/survival_minimum.R


Survival boundary STAT-SURV-INTERVAL-NARROW-001

Case STAT-SURV-INTERVAL-NARROW-001 · status: passed

Reference

Evidence type
R
Source
R survival and analytic boundary
Citation
R survival::survfit/coxph/survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_boundaries.R · sha256 528007972452fb15
Dataset
reference_data/datasets/SURV_NARROW.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/interval_countexact550.00E+2passed
/fits/[family=exponential]/param1_valtolerance5.02826807770398345.0282680776130619.09E-1110passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

105  narrow_scale <- stats::uniroot(exp_score, c(1, 20), tol = 1e-15)$root
106  emit("STAT-SURV-INTERVAL-NARROW-001","tabelz.executor.reliability",
107       "intervalLifeMenuItem","SURV_NARROW",
108       list(mode="interval",lower_column="lower",upper_column="upper"),
109       list("/interval_count"=exact(5L,"Narrow positive interval remains interval-censored"),
110            "/fits/[family=exponential]/param1_val"=tol(narrow_scale,
111              "Exponential MLE from the width-first score equation (uniroot)")),
112       c("A 1e-8 interval width is not rounded to an exact failure."))

Whole script generators/r/survival_boundaries.R


Survival population STAT-SURV-INTERVAL-POP-001

Case STAT-SURV-INTERVAL-POP-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::Surv/survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_population.R · sha256 3b3b4629e272fcd9
Dataset
reference_data/datasets/SURV_INTERVAL_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/input_row_countexact19190.00E+2passed
/omitted_row_countexact110.00E+2passed
/exact_countexact550.00E+2passed
/right_countexact440.00E+2passed
/left_countexact440.00E+2passed
/interval_countexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

118  inter_eligible <- inter[-10L,,drop=FALSE]
119  valid <- !(is.na(inter_eligible$lower)&is.na(inter_eligible$upper))
120  inter_complete <- inter_eligible[valid,,drop=FALSE]
121  exact_n <- sum(!is.na(inter_complete$lower)&!is.na(inter_complete$upper)&inter_complete$lower==inter_complete$upper)
122  right_n <- sum(!is.na(inter_complete$lower)&is.na(inter_complete$upper))
123  left_n <- sum(is.na(inter_complete$lower)&!is.na(inter_complete$upper))
124  interval_n <- nrow(inter_complete)-exact_n-right_n-left_n
125  emit("STAT-SURV-INTERVAL-POP-001","tabelz.executor.reliability","intervalLifeMenuItem",
126       "SURV_INTERVAL_POP",9L,
127       list(mode="interval",lower_column="lower",upper_column="upper"),
128       list("/input_row_count"=exact(nrow(inter_eligible),"Rows after exclusion"),
129            "/omitted_row_count"=exact(1L,"Both endpoints missing"),
130            "/exact_count"=exact(exact_n,"Exact endpoint pairs"),
131            "/right_count"=exact(right_n,"Finite lower, absent upper"),
132            "/left_count"=exact(left_n,"Absent lower, finite upper"),
133            "/interval_count"=exact(interval_n,"Finite ordered interval pairs")),
134       c("R interval2 endpoint types match the worker's missing-bound decoding."))

Whole script generators/r/survival_population.R


Survival STAT-SURV-KM-001

Case STAT-SURV-KM-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/survdiff/coxph and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, MASS 7.3.65, jsonlite 2.0.0
Reference script
generators/r/survival_core.R · sha256 1ed8945e6864eb84
Dataset
reference_data/datasets/HARVEST_GEHAN_MP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/curves/0/nexact21210.00E+2passed
/curves/0/eventsexact990.00E+2passed
/curves/0/survival/0tolerance0.85714285714285710.85714285714285721.06E-1615passed
/curves/0/lower/0tolerance0.619717955265983940.61971795526598381.07E-1615passed
/curves/0/upper/0tolerance0.951551747646997010.9515517476469974.84E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

24  mp <- read_dataset("HARVEST_GEHAN_MP")
25  both <- read_dataset("HARVEST_GEHAN_BOTH")
26  stopifnot(nrow(mp)==21L,nrow(both)==42L)
27  km <- survival::survfit(survival::Surv(time,status)~1,data=mp,conf.type="log-log")
28  config <- list(time_column="time",event_column="status",event_value="1",censored_value="0")
29  emit("STAT-SURV-KM-001","tabelz.executor.survival","survivalMenuItem",
30       "HARVEST_GEHAN_MP",config,
31       list("/curves/0/n"=exact(nrow(mp),"Gehan 6-MP row count"),
32            "/curves/0/events"=exact(sum(mp$status),"Gehan 6-MP failures"),
33            "/curves/0/survival/0"=tol(km$surv[1],"Kaplan-Meier first failure survival"),
34            "/curves/0/lower/0"=tol(km$lower[1],"Greenwood log-log lower pointwise limit"),
35            "/curves/0/upper/0"=tol(km$upper[1],"Greenwood log-log upper pointwise limit")),
36       c("R survfit conf.type=log-log matches Tabelz; this independently checks V14 Gehan values."))

Whole script generators/r/survival_core.R


Survival boundary STAT-SURV-KM-ALL-CENS-001

Case STAT-SURV-KM-ALL-CENS-001 · status: passed

Reference

Evidence type
R
Source
R survival and analytic boundary
Citation
R survival::survfit/coxph/survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_boundaries.R · sha256 528007972452fb15
Dataset
reference_data/datasets/SURV_ALL_CENS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/curves/0/survival/0tolerance11.00.00E+2passed
/curves/0/inference_availableexactFalseFalsepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

54  cfg <- list(time_column="time",event_column="status",event_value="1",censored_value="0")
55  emit("STAT-SURV-KM-ALL-CENS-001","tabelz.executor.survival","survivalMenuItem",
56       "SURV_ALL_CENS",cfg,
57       list("/curves/0/survival/0"=tol(1,"R all-censored KM remains at one","1e-9"),
58            "/curves/0/inference_available"=exact(FALSE,"No event means no Greenwood inference")),
59       c("R survfit(Surv(time,status)~1) stays at survival 1 when all rows are censored."))

Whole script generators/r/survival_boundaries.R


Survival option STAT-SURV-KM-CONFIDENCE-001

Case STAT-SURV-KM-CONFIDENCE-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_options.R · sha256 95af0dbeed660def
Dataset
reference_data/datasets/HARVEST_GEHAN_MP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/curves/0/lower/0tolerance0.671106780592172880.67110678059217281.11E-1615passed
/curves/0/upper/0tolerance0.942159405720121360.94215940572012143.21E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

24  mp <- read_dataset("HARVEST_GEHAN_MP")
25  km90 <- survival::survfit(survival::Surv(time,status)~1,data=mp,
26                            conf.type="log-log",conf.int=0.9)
27  emit("STAT-SURV-KM-CONFIDENCE-001","tabelz.executor.survival","survivalMenuItem",
28       "HARVEST_GEHAN_MP",list(time_column="time",event_column="status",event_value="1",
29                              censored_value="0",confidence_level=0.9),
30       list("/curves/0/lower/0"=tol(km90$lower[1],"90% Greenwood log-log lower", "1e-9"),
31            "/curves/0/upper/0"=tol(km90$upper[1],"90% Greenwood log-log upper", "1e-9")),
32       c("R survfit conf.type=log-log and conf.int=0.9; pointwise, not simultaneous."))

Whole script generators/r/survival_options.R


Hard-numeric survival STAT-SURV-KM-HARD-001

Case STAT-SURV-KM-HARD-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_hard.R · sha256 bc20ccd17540f4e7
Dataset
reference_data/datasets/SURV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/curves/0/survival/0tolerance0.974999999999999980.9752.20E-1817passed
/curves/0/median/timetolerance1003154.66792142581003154.66792142582.53E-1116passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

47  km <- survival::survfit(survival::Surv(time,status)~1,data=d,conf.type="log-log")
48  cfg <- list(time_column="time",event_column="status",event_value="1",censored_value="0")
49  emit("STAT-SURV-KM-HARD-001","tabelz.executor.survival","survivalMenuItem",
50       "SURV_HARD",cfg,
51       list("/curves/0/survival/0"=tol(km$surv[1],"Large-offset first KM step","1e-9"),
52            "/curves/0/median/time"=tol(summary(km)$table[["median"]],
53                                           "First crossing of one-half survival","1e-9")),
54       c("Lifetimes are near 1e6 with small relative spread; events remain distinct."))

Whole script generators/r/survival_hard.R


Invalid survival request STAT-SURV-KM-INVALID-001

Case STAT-SURV-KM-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/HARVEST_GEHAN_MP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid survival request STAT-SURV-KM-INVALID-002

Case STAT-SURV-KM-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/HARVEST_GEHAN_MP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Small-sample survival boundary STAT-SURV-KM-MIN-001

Case STAT-SURV-KM-MIN-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::survfit/coxph and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_minimum.R · sha256 2997d37929a6dae8
Dataset
reference_data/datasets/SURV_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/analyzed_countexact110.00E+2passed
/curves/0/survival/0tolerance00.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

30  cfg <- list(time_column="time",event_column="status",event_value="1",censored_value="0")
31  emit("STAT-SURV-KM-MIN-001","tabelz.executor.survival","survivalMenuItem",
32       "SURV_MIN",1:4,cfg,
33       list("/analyzed_count"=exact(1L,"One observed failure is the smallest KM sample"),
34            "/curves/0/survival/0"=tol(0,"Single event reduces survival to zero","1e-9")),
35       c("R survfit on the first row has one event at time 1 and S(1)=0."))

Whole script generators/r/survival_minimum.R


Survival population STAT-SURV-KM-POP-001

Case STAT-SURV-KM-POP-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::Surv/survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_population.R · sha256 3b3b4629e272fcd9
Dataset
reference_data/datasets/SURV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact32320.00E+2passed
/excluded_countexact110.00E+2passed
/missing_countexact220.00E+2passed
/curves/0/survival/0tolerance11.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

44  eligible <- d[-8L,,drop=FALSE]
45  kmdata <- eligible[stats::complete.cases(eligible[c("time","status")]),,drop=FALSE]
46  km <- survival::survfit(survival::Surv(time,status)~1,data=kmdata,conf.type="log-log")
47  cfg <- list(time_column="time",event_column="status",event_value="1",censored_value="0")
48  emit("STAT-SURV-KM-POP-001","tabelz.executor.survival","survivalMenuItem",
49       "SURV_POP",7L,cfg,
50       list("/source_count"=exact(nrow(d),"CSV population"),
51            "/excluded_count"=exact(1L,"One source row excluded"),
52            "/missing_count"=exact(2L,"Missing time and status rows"),
53            "/curves/0/survival/0"=tol(km$surv[1],"First complete-case Kaplan-Meier step")),
54       c("Kaplan-Meier removes missing time/status after exclusion; log-log Greenwood bands."))

Whole script generators/r/survival_population.R


Parametric Weibull life distribution under mixed censoring

Case STAT-SURV-LIFE-001 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_LIFE_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact24240.00E+2passed
/counts/missing_countexact000.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact24240.00E+2passed
/counts/exacttolerance1616.00.00E+2passed
/counts/righttolerance55.00.00E+2passed
/counts/lefttolerance22.00.00E+2passed
/counts/intervaltolerance55.00.00E+2passed
/log_likelihoodtolerance-105.02050939613645-105.020509396136453.13E-1516passed
/aictolerance214.04101879227289214.04101879227293.73E-1516passed
/bictolerance216.7054278126233216.70542781262331.32E-1517passed
/covariance/0/0tolerance0.0202207705818092110.0202207678775478272.70E-96passed
/covariance/0/1tolerance-0.0035441834844420019-0.00354418301238248334.72E-106passed
/covariance/1/1tolerance0.027455971661040190.027455969997135881.66E-97passed
/parameters/0/estimatetolerance1.48535712391663391.48535712392576489.13E-1211passed
/parameters/0/standard_errortolerance0.246121352191886860.246121344735598167.46E-97passed
/parameters/0/lowertolerance1.0734655689252351.07346557949553631.06E-88passed
/parameters/0/uppertolerance2.05529254913955972.0552925289265822.02E-88passed
/parameters/1/estimatetolerance88.16339558639637688.163395589965523.57E-910passed
/parameters/1/standard_errortolerance12.5368133328754812.5368124950662638.38E-77passed
/parameters/1/lowertolerance66.71871248932858166.718713735443071.25E-67passed
/parameters/1/uppertolerance116.50081410918483116.500811942715912.17E-67passed
/parameters/2/estimatetolerance00.00.00E+2passed
/percentiles/0/estimatetolerance3.98362840383470523.9836284040718162.37E-1010passed
/percentiles/0/standard_errortolerance2.20188617394135782.20188609626869177.77E-87passed
/percentiles/0/lowertolerance1.34831365721978311.34831370891321475.17E-87passed
/percentiles/0/uppertolerance11.76973560629880411.7697351564562354.50E-77passed
/percentiles/1/estimatetolerance11.93578786917905411.9357878698089666.30E-1010passed
/percentiles/1/standard_errortolerance4.53215152984532034.53215135331776951.77E-77passed
/percentiles/1/lowertolerance5.67074853961860375.670748704520951.65E-77passed
/percentiles/1/uppertolerance25.12243860977540125.1224378818799687.28E-77passed
/percentiles/2/estimatetolerance19.37836686996022819.3783668709252059.65E-1010passed
/percentiles/2/standard_errortolerance5.94067521415109215.9406749630317462.51E-77passed
/percentiles/2/lowertolerance10.62602753328976110.6260278040238742.71E-77passed
/percentiles/2/uppertolerance35.33974492069783235.339744023817538.97E-77passed
/percentiles/3/estimatetolerance68.88518642189573968.885186424788932.89E-910passed
/percentiles/3/standard_errortolerance10.59164612801448310.59164544780526.80E-77passed
/percentiles/3/lowertolerance50.96204187997189250.962042869063249.89E-77passed
/percentiles/3/uppertolerance93.11182859500338893.111826795673861.80E-67passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

221  life_case("STAT-SURV-LIFE-001", "Parametric Weibull life distribution under mixed censoring",
222            "weibull", 0.0, 0.95, c(1, 5, 10, 50), LIFE_NOTES, "nominal")

Whole script generators/r/survival_life_families.R


Lognormal life distribution with a declared threshold at 90%

Case STAT-SURV-LIFE-002 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_LIFE_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact24240.00E+2passed
/counts/missing_countexact000.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact24240.00E+2passed
/counts/exacttolerance1616.00.00E+2passed
/counts/righttolerance55.00.00E+2passed
/counts/lefttolerance22.00.00E+2passed
/counts/intervaltolerance55.00.00E+2passed
/log_likelihoodtolerance-101.50970779291677-101.509707792916771.04E-1516passed
/aictolerance207.01941558583354207.019415585833542.08E-1516passed
/bictolerance209.68382460618395209.683824606183952.97E-1516passed
/covariance/0/0tolerance0.0210866847726037530.0210866848380305846.54E-118passed
/covariance/0/1tolerance0.00135733915897756890.00135733909577511356.32E-117passed
/covariance/1/1tolerance0.0263865668680594750.0263865664019972764.66E-107passed
/parameters/0/estimatetolerance4.07422022146808074.0742202229582521.49E-99passed
/parameters/0/standard_errortolerance0.145212550327455360.145212550552734882.25E-108passed
/parameters/0/lowertolerance3.83536683138309263.8353668325027121.12E-99passed
/parameters/0/uppertolerance4.31307361155306884.3130736134137921.86E-99passed
/parameters/1/estimatetolerance0.740976845322904750.74097684649382671.17E-98passed
/parameters/1/standard_errortolerance0.12036385286072580.12036385198794488.73E-108passed
/parameters/1/lowertolerance0.567238754678669690.56723875691353512.23E-98passed
/parameters/1/uppertolerance0.967928726265730450.9679287255113077.54E-109passed
/parameters/2/estimatetolerance55.00.00E+2passed
/percentiles/0/estimatetolerance22.38176676878277922.381766761207397.58E-99passed
/percentiles/0/standard_errortolerance4.14891098712631754.1489109724222981.47E-88passed
/percentiles/0/lowertolerance16.73770437442102516.737704383629499.21E-99passed
/percentiles/0/uppertolerance30.73977043268854830.739770390059164.26E-88passed
/percentiles/1/estimatetolerance63.80460813449324363.80460822212228.76E-88passed
/percentiles/1/standard_errortolerance8.53916711821639048.5391671441886882.60E-88passed
/percentiles/1/lowertolerance51.3104126651745151.3104127170245555.19E-88passed
/percentiles/1/uppertolerance79.66964207062045579.669642209559981.39E-78passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

223  life_case("STAT-SURV-LIFE-002", "Lognormal life distribution with a declared threshold at 90%",
224            "lognormal", 5.0, 0.90, c(5, 50), c(LIFE_NOTES,
225              "Threshold 5 shifts the data before fitting; survreg fits the shifted records."), "axis_coverage")

Whole script generators/r/survival_life_families.R


Exponential life distribution from time-and-status data

Case STAT-SURV-LIFE-003 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_LIFE_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact24240.00E+2passed
/counts/missing_countexact660.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact18180.00E+2passed
/counts/exacttolerance1616.00.00E+2passed
/counts/righttolerance55.00.00E+2passed
/counts/lefttolerance00.00.00E+2passed
/counts/intervaltolerance00.00.00E+2passed
/log_likelihoodtolerance-90.51945764131159-90.519457641311581.45E-1415passed
/aictolerance183.03891528262318183.038915282623152.91E-1415passed
/bictolerance184.08343772034661184.083437720346582.94E-1415passed
/covariance/0/0tolerance0.0625000000000000140.062499992549420297.45E-96passed
/parameters/0/estimatetolerance105.36875000000001105.36874999892621.07E-910passed
/parameters/0/standard_errortolerance26.34218750000000126.3421859296149631.57E-67passed
/parameters/0/lowertolerance64.55225858942351664.552260474061951.88E-67passed
/parameters/0/uppertolerance171.99357108755896171.993566062607335.02E-67passed
/parameters/1/estimatetolerance00.00.00E+2passed
/percentiles/0/estimatetolerance1.05899132596363121.0589913259528381.08E-1110passed
/percentiles/0/standard_errortolerance0.26474783149090780.26474781570801051.58E-87passed
/percentiles/0/lowertolerance0.648771878925780140.64877189786702881.89E-87passed
/percentiles/0/uppertolerance1.72859315407304481.72859310357059155.05E-87passed
/percentiles/1/estimatetolerance5.40471031299822085.4047103129431375.51E-1110passed
/percentiles/1/standard_errortolerance1.35117757824955521.35117749769933188.06E-87passed
/percentiles/1/lowertolerance3.31109800320859023.3110980998779029.67E-87passed
/percentiles/1/uppertolerance8.82211687456026898.8221166168139382.58E-77passed
/percentiles/2/estimatetolerance73.03605198162573273.036051980881427.44E-1010passed
/percentiles/2/standard_errortolerance18.25901299540643318.259011906898471.09E-67passed
/percentiles/2/lowertolerance44.74421604003541844.7442173463672361.31E-67passed
/percentiles/2/uppertolerance119.21685887377801119.216855390746963.48E-67passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

231  dk <- d[!is.na(d$status), ]
232  fit3 <- survreg(Surv(start, status) ~ 1, data = dk, weights = dk$freq, dist = "exponential")
233  outputs3 <- c(
234    list("/counts/source_count" = exact(nrow(d), "source rows"),
235         "/counts/missing_count" = exact(sum(is.na(d$status)), "rows without a status code"),
236         "/counts/zero_frequency_rows" = exact(0, "no zero-frequency rows"),
237         "/counts/rows_used" = exact(nrow(dk), "rows with a status code"),
238         "/counts/exact" = tol(sum(dk$freq[dk$status == 1]), "weighted failures"),
239         "/counts/right" = tol(sum(dk$freq[dk$status == 0]), "weighted right-censored rows"),
240         "/counts/left" = tol(0, "no left censoring in this layout"),
241         "/counts/interval" = tol(0, "no interval records in this layout")),
242    fit_info_outputs(fit3, dk$freq),
243    family_params_outputs(fit3, "exponential", 0.0, 0.95),
244    percentile_outputs(fit3, "exponential", c(1, 5, 50), 0.0, 0.95)
245  )
246  emit("STAT-SURV-LIFE-003", "Exponential life distribution from time-and-status data",
247       LIFE, "lifeDistributionMenuItem", LIFE_REQ, "SURV_LIFE_B9",
248       list(layout = "right_censored", start = "start", status = "status", frequency = "freq",
249            family = "exponential", threshold = 0.0, confidence_level = 0.95,
250            percents = list(1, 5, 50)),
251       outputs3, c(LIFE_NOTES,
252         "The time-and-status layout reads only rows with a status code; the interval and left-censored rows count as missing."), "axis_coverage")

Whole script generators/r/survival_life_families.R


Smallest-extreme-value life distribution with interval records

Case STAT-SURV-LIFE-004 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_LIFE_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact24240.00E+2passed
/counts/missing_countexact000.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact24240.00E+2passed
/counts/exacttolerance1616.00.00E+2passed
/counts/righttolerance55.00.00E+2passed
/counts/lefttolerance22.00.00E+2passed
/counts/intervaltolerance55.00.00E+2passed
/log_likelihoodtolerance-116.88243551898506-116.882435518985081.63E-1415passed
/aictolerance237.76487103797012237.764871037970153.27E-1415passed
/bictolerance240.42928005832053240.429280058320562.76E-1415passed
/covariance/0/0tolerance202.2294633367369202.22944802249891.53E-57passed
/covariance/0/1tolerance-0.38289316377223187-0.382893121216343674.26E-86passed
/covariance/1/1tolerance0.0234882911096165190.023488290462490596.47E-107passed
/parameters/0/estimatetolerance104.78839065699613104.788390682093252.51E-89passed
/parameters/0/standard_errortolerance14.22074060436856714.2207400659212845.38E-77passed
/parameters/0/lowertolerance76.91625123894738276.916252319381781.08E-67passed
/parameters/0/uppertolerance132.66053007504487132.660529044804721.03E-68passed
/parameters/1/estimatetolerance67.06869782118182967.068697839137291.80E-89passed
/parameters/1/standard_errortolerance10.27887500138544810.2788748625406331.39E-77passed
/parameters/1/lowertolerance49.6667406087010949.666740827514312.19E-78passed
/parameters/1/uppertolerance90.56785632196208390.567855971446923.51E-78passed
/percentiles/0/estimatetolerance-203.73762776602547-203.73762782352625.75E-89passed
/percentiles/0/standard_errortolerance51.71368354267659851.713682557329349.85E-77passed
/percentiles/0/lowertolerance-305.09458501757331-305.09458314382891.87E-68passed
/percentiles/0/uppertolerance-102.38067051447764-102.38067250322351.99E-67passed
/percentiles/1/estimatetolerance-46.140815605182851-46.1408156204921061.53E-89passed
/percentiles/1/standard_errortolerance29.20382605760228529.203825328520627.29E-77passed
/percentiles/1/lowertolerance-103.37926288885566-103.379261475191131.41E-67passed
/percentiles/1/uppertolerance11.09763167848996811.0976302342069181.44E-66passed
/percentiles/2/estimatetolerance80.20684633894566780.206846357461881.85E-89passed
/percentiles/2/standard_errortolerance15.33774847577549515.3377478960031235.80E-77passed
/percentiles/2/lowertolerance50.14541172249159550.1454128773407751.15E-67passed
/percentiles/2/uppertolerance110.26828095539975110.2682798375831.12E-67passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

226  life_case("STAT-SURV-LIFE-004", "Smallest-extreme-value life distribution with interval records",
227            "sev", 0.0, 0.95, c(1, 10, 50), c(LIFE_NOTES,
228              "The SEV family has real support: survreg dist = 'extreme' on the unshifted records."), "axis_coverage")

Whole script generators/r/survival_life_families.R


Weibull life regression on a numeric stress predictor with predictions

Case STAT-SURV-LIFEREG-001 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_LIFE_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact24240.00E+2passed
/counts/missing_countexact000.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact24240.00E+2passed
/counts/exacttolerance1616.00.00E+2passed
/counts/righttolerance55.00.00E+2passed
/counts/lefttolerance22.00.00E+2passed
/counts/intervaltolerance55.00.00E+2passed
/coefficients/0/estimatetolerance4.3264784645809384.3264784641495084.31E-1010passed
/coefficients/0/standard_errortolerance0.415033725372613680.41503372179885023.57E-98passed
/coefficients/0/z_valuetolerance10.42440216321375610.4244022519364698.87E-88passed
/coefficients/0/lowertolerance3.51302731048112763.51302731705414536.57E-98passed
/coefficients/0/uppertolerance5.13992961868074795.139929611244877.44E-98passed
/coefficients/1/estimatetolerance0.00127470974349601320.0012747097609005091.74E-117passed
/coefficients/1/standard_errortolerance0.00329185460958848420.0032918545854506562.41E-118passed
/coefficients/1/z_valuetolerance0.387231483365957360.387231491492507968.13E-97passed
/coefficients/1/lowertolerance-0.0051772067336395746-0.0051772066689258076.47E-117passed
/coefficients/1/uppertolerance0.00772662622063160190.0077266261907268252.99E-118passed
/dispersion/estimatetolerance0.670426035866262260.67042603729086081.42E-98passed
/dispersion/standard_errortolerance0.110998258380247860.110998257337744031.04E-98passed
/dispersion/z_valuetolerance-2.4150326324810378-2.415032647460521.50E-88passed
/dispersion/lowertolerance0.484643887652566720.48464389049363012.84E-98passed
/dispersion/uppertolerance0.927425437560803980.92742543606548221.50E-98passed
/shape/estimatetolerance1.49158885022699471.49158884705749473.17E-98passed
/shape/standard_errortolerance0.246953065270901470.246953061901991343.37E-97passed
/shape/z_valuetolerance2.41503263248103782.415032647460521.50E-88passed
/shape/lowertolerance1.07825379755603051.0782537992945381.74E-98passed
/shape/uppertolerance2.06337070471191762.0633706926160941.21E-88passed
/log_likelihoodtolerance-104.94585463698002-104.945854636980023.96E-1516passed
/predictions/0/estimatetolerance18.53666008682253318.5366600452088234.16E-88passed
/predictions/0/standard_errortolerance6.12422440112280696.1242242480092551.53E-77passed
/predictions/0/lowertolerance9.70080389923570699.7008040204059911.21E-77passed
/predictions/0/uppertolerance35.42054563142661335.420545029963486.01E-77passed
/predictions/1/estimatetolerance65.54490233615572965.544902364916552.88E-89passed
/predictions/1/standard_errortolerance13.02845904040664813.0284586878602693.53E-77passed
/predictions/1/lowertolerance44.39604744403214944.39604793912814.95E-77passed
/predictions/1/uppertolerance96.76839425113492296.768393256915969.94E-77passed
/predictions/2/estimatetolerance20.26671894071543220.2667189199090922.08E-88passed
/predictions/2/standard_errortolerance6.52808669297255326.5280865113498341.82E-77passed
/predictions/2/lowertolerance10.77963395532852610.7796341266137171.71E-77passed
/predictions/2/uppertolerance38.10332505946725938.1033243757816.84E-77passed
/predictions/3/estimatetolerance71.66232252313017271.662322641882581.19E-78passed
/predictions/3/standard_errortolerance13.24614446229879913.2461439363314375.26E-77passed
/predictions/3/lowertolerance49.88314498184474449.883145812032748.30E-77passed
/predictions/3/uppertolerance102.95037474638418102.950373374216981.37E-67passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

722  regression_case("STAT-SURV-LIFEREG-001",
723                  "Weibull life regression on a numeric stress predictor with predictions",
724                  "weibull", 0.0, "stress", character(0), "none", NULL, "celsius",
725                  c("stress=80", "stress=150"), c(10, 50), 0.95, REG_NOTES, "nominal")

Whole script generators/r/survival_life_families.R


Lognormal life regression with a sum-coded three-level factor

Case STAT-SURV-LIFEREG-002 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_LIFE_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact24240.00E+2passed
/counts/missing_countexact000.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact24240.00E+2passed
/counts/exacttolerance1616.00.00E+2passed
/counts/righttolerance55.00.00E+2passed
/counts/lefttolerance22.00.00E+2passed
/counts/intervaltolerance55.00.00E+2passed
/coefficients/0/estimatetolerance4.1376906365104964.1376906365289151.84E-1111passed
/coefficients/0/standard_errortolerance0.13128665602480030.131286653317088052.71E-97passed
/coefficients/0/z_valuetolerance31.51645994950833431.516460599657626.50E-77passed
/coefficients/0/lowertolerance3.88037351905118883.8803735243766265.33E-98passed
/coefficients/0/uppertolerance4.39500775396980274.3950077486812045.29E-98passed
/coefficients/1/estimatetolerance0.196060908899182260.19606090890538736.21E-1210passed
/coefficients/1/standard_errortolerance0.175087786330607620.17508777655658579.77E-97passed
/coefficients/1/z_valuetolerance1.11978632552342841.11978638806932026.25E-87passed
/coefficients/1/lowertolerance-0.147104846441653-0.147104827278717041.92E-86passed
/coefficients/1/uppertolerance0.539226664240017530.53922664508949161.92E-87passed
/coefficients/2/estimatetolerance-0.021682242701141122-0.0216822427070456145.90E-129passed
/coefficients/2/standard_errortolerance0.178373251926470880.178373247602873824.32E-97passed
/coefficients/2/z_valuetolerance-0.12155546006460087-0.121555463044090962.98E-97passed
/coefficients/2/lowertolerance-0.37128739228231378-0.37128738381412388.47E-97passed
/coefficients/2/uppertolerance0.327922906880031560.32792289840003268.48E-97passed
/dispersion/estimatetolerance0.650593601803531980.65059360181595961.24E-1110passed
/dispersion/standard_errortolerance0.10561820276475520.105618201450371781.31E-97passed
/dispersion/z_valuetolerance-2.6479406852749996-2.64794071816065653.29E-87passed
/dispersion/lowertolerance0.473287063302128220.47328706518811211.89E-98passed
/dispersion/uppertolerance0.894324116434790930.89432411290519913.53E-98passed
/log_likelihoodtolerance-101.13057111966242-101.130571119662421.88E-1617passed
/predictions/0/estimatetolerance26.63534637752990326.6353463774390249.09E-1111passed
/predictions/0/standard_errortolerance6.54724861384927786.5472485473931426.65E-87passed
/predictions/0/lowertolerance16.45220249900130916.4522025793723588.04E-88passed
/predictions/0/uppertolerance43.12138005194503943.12137984099752.11E-78passed
/predictions/1/estimatetolerance61.31401176929870961.314011770066027.67E-1010passed
/predictions/1/standard_errortolerance12.84997351619103212.8499732299434852.86E-77passed
/predictions/1/lowertolerance40.6600517440813540.6600521168464163.73E-78passed
/predictions/1/uppertolerance92.45949963142716492.459498786086948.45E-78passed
/predictions/2/estimatetolerance33.11486570398324133.114865704271292.88E-1011passed
/predictions/2/standard_errortolerance7.77615037744020037.7761501558860982.22E-77passed
/predictions/2/lowertolerance20.89973736250599920.8997376368315032.74E-77passed
/predictions/2/uppertolerance52.46928760741888552.469286919630986.88E-77passed
/predictions/3/estimatetolerance76.22973010126362776.229730103140771.88E-910passed
/predictions/3/standard_errortolerance15.32928748388167915.329286941213615.43E-77passed
/predictions/3/lowertolerance51.39885689107067951.398857609987577.19E-77passed
/predictions/3/uppertolerance113.05643943846182113.056437862706991.58E-67passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

726  regression_case("STAT-SURV-LIFEREG-002",
727                  "Lognormal life regression with a sum-coded three-level factor",
728                  "lognormal", 0.0, character(0), "temp3", "none", NULL, "celsius",
729                  c("temp3=low", "temp3=high"), c(10, 50), 0.95,
730                  c(REG_NOTES,
731                    "contr.sum on sorted levels high, low, mid: the engine's grp[sum:high] and grp[sum:low] are R's temp31 and temp32 (mid is the -1 reference)."), "axis_coverage")

Whole script generators/r/survival_life_families.R


SEV life regression with the Arrhenius temperature preset at 90%

Case STAT-SURV-LIFEREG-003 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_LIFE_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact24240.00E+2passed
/counts/missing_countexact000.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact24240.00E+2passed
/counts/exacttolerance1616.00.00E+2passed
/counts/righttolerance55.00.00E+2passed
/counts/lefttolerance22.00.00E+2passed
/counts/intervaltolerance55.00.00E+2passed
/coefficients/0/estimatetolerance735.39201635573295735.39201571789426.38E-79passed
/coefficients/0/standard_errortolerance383.36146908118963383.361456297507741.28E-57passed
/coefficients/0/z_valuetolerance1.91827315905863531.91827322136212126.23E-87passed
/coefficients/0/lowertolerance104.81851350409374104.818533893540232.04E-56passed
/coefficients/0/uppertolerance1365.96551920737221365.96549754224822.17E-57passed
/coefficients/1/estimatetolerance-210907.61355408074-210907.61337272131.81E-49passed
/coefficients/1/standard_errortolerance127639.44945493984127639.445133076664.32E-37passed
/coefficients/1/z_valuetolerance-1.6523701289430648-1.65237018347133515.45E-87passed
/coefficients/1/lowertolerance-420855.82493212755-420855.81764193597.29E-37passed
/coefficients/1/uppertolerance-959.4021760339383-959.40910350671046.93E-35passed
/dispersion/estimatetolerance61.18666049033382861.186660560284557.00E-88passed
/dispersion/standard_errortolerance10.03456181009900510.0345618762870116.62E-88passed
/dispersion/z_valuetolerance25.08506041421760825.0850602844055571.30E-78passed
/dispersion/lowertolerance46.72010392512397246.720103909814951.53E-89passed
/dispersion/uppertolerance80.13268609075434880.132686300232762.09E-78passed
/log_likelihoodtolerance-115.74183856168878-115.74183856168881.46E-1415passed
/predictions/0/estimatetolerance-65.219236899632335-65.219237124842042.25E-78passed
/predictions/0/standard_errortolerance30.64278283103697130.6427825883060122.43E-78passed
/predictions/0/lowertolerance-115.62212937914978-115.62212920510261.74E-78passed
/predictions/0/uppertolerance-14.816344420114895-14.8163450445814756.24E-77passed
/predictions/1/estimatetolerance123.50488480048381123.504884791030129.45E-910passed
/predictions/1/standard_errortolerance24.21971763812614524.2197169756562586.62E-77passed
/predictions/1/lowertolerance83.66699439967149583.666995479883781.08E-67passed
/predictions/1/uppertolerance163.34277520129612163.342774102176461.10E-68passed
/predictions/2/estimatetolerance-8.0956874278047621-8.0956877021351372.74E-77passed
/predictions/2/standard_errortolerance34.38120710636105334.381206956856291.50E-78passed
/predictions/2/lowertolerance-64.647740635672449-64.647740664089382.84E-89passed
/predictions/2/uppertolerance48.45636578006292448.4563652598191155.20E-77passed
/predictions/3/estimatetolerance180.62843427231138180.628434213737025.86E-89passed
/predictions/3/standard_errortolerance20.5939275721784320.593927293774882.78E-77passed
/predictions/3/lowertolerance146.75443781203779146.75443821139653.99E-78passed
/predictions/3/uppertolerance214.50243073258497214.502430216077565.17E-78passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

732  regression_case("STAT-SURV-LIFEREG-003",
733                  "SEV life regression with the Arrhenius temperature preset at 90%",
734                  "sev", 0.0, "temp_c", character(0), "arrhenius", "temp_c", "celsius",
735                  c("temp_c=45", "temp_c=75"), c(10, 90), 0.90,
736                  c(REG_NOTES,
737                    "The Arrhenius preset maps Celsius to 1/(temp_c + 273.15); predictions are entered in Celsius and transformed identically.",
738                    "The Arrhenius coefficient is about -2.1e5 on the 1/K covariate and its z is -1.65, so each Wald bound is the small difference of two O(2e5) quantities: last-digit round-off in the estimate or its SE (each within 1e-6 relative of itself) moves the bound by up to a few 1e-3 absolute. The coefficient interval outputs therefore carry a 5e-2 absolute tolerance alongside 1e-6 relative; estimate, SE and z stay at the default."),
739                  "axis_coverage", ci_atol = "5e-2")

Whole script generators/r/survival_life_families.R


Actuarial life table with 2-unit intervals to 10

Case STAT-SURV-LIFETAB-001 · status: passed

Reference

Evidence type
R
Source
Base R arithmetic of the actuarial, Gehan and grouped PH formulas
Citation
Cutler and Ederer (1958) J Chronic Dis 8:699-712; Greenwood (1926); Lee and Desu (1972) Computer Programs in Biomedicine 2:315-321; Gehan (1965) Biometrika 52:203-223; Prentice and Gloeckler (1978) Biometrics 34:57-67; R stats: pchisq, glm (cloglog cross-check), anova(test = 'Rao').
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_life_tables.R · sha256 e7bfcfb4270e21ec
Dataset
reference_data/datasets/SURV_LIFETAB.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/analyzed_countexact30300.00E+2passed
/strata/0/subjectsexact30300.00E+2passed
/strata/0/eventsexact21210.00E+2passed
/strata/0/withdrawnexact990.00E+2passed
/strata/0/median/statusexactestimatedestimatedpassed
/strata/0/median/valuetolerance6.36220498323581836.3622049832358181.03E-1717passed
/strata/0/rows/0/enteringexact30300.00E+2passed
/strata/0/rows/0/withdrawnexact110.00E+2passed
/strata/0/rows/0/eventsexact330.00E+2passed
/strata/0/rows/0/at_risktolerance29.529.50.00E+1passed
/strata/0/rows/0/qtolerance0.101694915254237290.10169491525423734.72E-1816passed
/strata/0/rows/0/survival_starttolerance11.00.00E+2passed
/strata/0/rows/0/survival_endtolerance0.898305084745762760.89830508474576287.89E-1918passed
/strata/0/rows/0/survival_end_setolerance0.055648125576628640.055648125576628643.43E-1917passed
/strata/0/rows/0/hazardtolerance0.0535714285714285750.0535714285714285753.94E-1917passed
/strata/0/rows/0/densitytolerance0.0508474576271186470.050847457627118653.61E-1917passed
/strata/0/rows/0/hazard_setolerance0.0308850645937889450.0308850645937889453.64E-1916passed
/strata/0/rows/0/density_setolerance0.0278240627883143170.0278240627883143172.98E-1916passed
/strata/0/rows/1/enteringexact26260.00E+2passed
/strata/0/rows/1/withdrawnexact110.00E+2passed
/strata/0/rows/1/eventsexact550.00E+2passed
/strata/0/rows/1/at_risktolerance25.525.50.00E+1passed
/strata/0/rows/1/qtolerance0.196078431372549020.196078431372549023.11E-1816passed
/strata/0/rows/1/survival_starttolerance0.898305084745762760.89830508474576287.89E-1918passed
/strata/0/rows/1/survival_start_setolerance0.055648125576628640.055648125576628643.43E-1917passed
/strata/0/rows/1/survival_endtolerance0.722166832834828960.7221668328348294.49E-1817passed
/strata/0/rows/1/survival_end_setolerance0.08360417070625370.08360417070625371.59E-1917passed
/strata/0/rows/1/hazardtolerance0.108695652173913040.108695652173913044.61E-1917passed
/strata/0/rows/1/densitytolerance0.088069125955466940.088069125955466942.19E-1917passed
/strata/0/rows/1/hazard_setolerance0.0483221618034363750.0483221618034363753.49E-1917passed
/strata/0/rows/1/density_setolerance0.0357328318195030620.035732831819503062.54E-1917passed
/strata/0/rows/2/enteringexact20200.00E+2passed
/strata/0/rows/2/withdrawnexact220.00E+2passed
/strata/0/rows/2/eventsexact550.00E+2passed
/strata/0/rows/2/at_risktolerance1919.00.00E+2passed
/strata/0/rows/2/qtolerance0.263157894736842090.26315789473684216.55E-1917passed
/strata/0/rows/2/survival_starttolerance0.722166832834828960.7221668328348294.49E-1817passed
/strata/0/rows/2/survival_start_setolerance0.08360417070625370.08360417070625371.59E-1917passed
/strata/0/rows/2/survival_endtolerance0.532122929457242490.53212292945724251.62E-1817passed
/strata/0/rows/2/survival_end_setolerance0.0954850642507308890.095485064250730891.20E-1917passed
/strata/0/rows/2/hazardtolerance0.151515151515151490.15151515151515151.60E-1816passed
/strata/0/rows/2/densitytolerance0.095021951688793280.095021951688793283.11E-1917passed
/strata/0/rows/2/hazard_setolerance0.066977346294987930.066977346294987931.99E-2018passed
/strata/0/rows/2/density_setolerance0.0381002203628511560.0381002203628511561.37E-2018passed
/strata/0/rows/3/enteringexact13130.00E+2passed
/strata/0/rows/3/withdrawnexact220.00E+2passed
/strata/0/rows/3/eventsexact440.00E+2passed
/strata/0/rows/3/at_risktolerance1212.00.00E+2passed
/strata/0/rows/3/qtolerance0.333333333333333310.33333333333333334.83E-1816passed
/strata/0/rows/3/survival_starttolerance0.532122929457242490.53212292945724251.62E-1817passed
/strata/0/rows/3/survival_start_setolerance0.0954850642507308890.095485064250730891.20E-1917passed
/strata/0/rows/3/survival_endtolerance0.354748619638161680.35474861963816172.57E-1817passed
/strata/0/rows/3/survival_end_setolerance0.0964146477204681360.096414647720468143.06E-1917passed
/strata/0/rows/3/hazardtolerance0.199999999999999980.199999999999999983.35E-1816passed
/strata/0/rows/3/densitytolerance0.0886871549095404050.08868715490954044.79E-1917passed
/strata/0/rows/3/hazard_setolerance0.0979795897113271140.097979589711327111.43E-1917passed
/strata/0/rows/3/density_setolerance0.039549499986535550.039549499986535554.55E-1916passed
/strata/0/rows/4/enteringexact770.00E+2passed
/strata/0/rows/4/withdrawnexact220.00E+2passed
/strata/0/rows/4/eventsexact330.00E+2passed
/strata/0/rows/4/at_risktolerance66.00.00E+2passed
/strata/0/rows/4/qtolerance0.50.50.00E+1passed
/strata/0/rows/4/survival_starttolerance0.354748619638161680.35474861963816172.57E-1817passed
/strata/0/rows/4/survival_start_setolerance0.0964146477204681360.096414647720468143.06E-1917passed
/strata/0/rows/4/survival_endtolerance0.177374309819080840.177374309819080841.29E-1817passed
/strata/0/rows/4/survival_end_setolerance0.0869916875592032650.086991687559203271.23E-1917passed
/strata/0/rows/4/hazardtolerance0.333333333333333310.33333333333333334.83E-1816passed
/strata/0/rows/4/densitytolerance0.0886871549095404190.088687154909540423.56E-1917passed
/strata/0/rows/4/hazard_setolerance0.181443684650605790.181443684650605794.56E-1816passed
/strata/0/rows/4/density_setolerance0.0434958437796016330.043495843779601634.38E-1916passed
/strata/0/rows/5/enteringexact220.00E+2passed
/strata/0/rows/5/withdrawnexact110.00E+2passed
/strata/0/rows/5/eventsexact110.00E+2passed
/strata/0/rows/5/at_risktolerance1.51.50.00E+1passed
/strata/0/rows/5/qtolerance0.666666666666666630.66666666666666663.41E-1918passed
/strata/0/rows/5/survival_starttolerance0.177374309819080840.177374309819080841.29E-1817passed
/strata/0/rows/5/survival_start_setolerance0.0869916875592032650.086991687559203271.23E-1917passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

217  edges1 <- seq(0, 10, by = 2)
218  event <- d$status == 1L
219  rows1 <- life_table(d$time, event, edges1)
220  stopifnot(length(rows1) == 6L)
221  outputs <- c(
222    list("/analyzed_count" = exact(nrow(d), "30 rows analysed"),
223         "/strata/0/subjects" = exact(nrow(d), "30 subjects"),
224         "/strata/0/events" = exact(as.integer(sum(event)), "events"),
225         "/strata/0/withdrawn" = exact(as.integer(sum(!event)), "withdrawals"),
226         "/strata/0/median/status" = exact("estimated", "the median lies in a finite interval"),
227         "/strata/0/median/value" = tol(actuarial_median(rows1), "actuarial median by linear interpolation")),
228    row_outputs(rows1, "/strata/0", 1:6)
229  )
230  emit("STAT-SURV-LIFETAB-001", "Actuarial life table with 2-unit intervals to 10", "nominal",
231       "SURV_LIFETAB", list(time = "time", status = "status", event_value = "1", censored_value = "0",
232                            interval_width = 2, last_boundary = 10),
233       outputs,
234       list("Edges 0, 2, ..., 10 and the open interval [10, inf): findInterval places an event at exactly 2.0 and 4.0 in the interval that starts there.",
235            "The open terminal row reports entering, withdrawn, events, n - w/2, q and the survival at its start only."))

Whole script generators/r/survival_life_tables.R


Unequal boundaries with two strata and the grouped Gehan test

Case STAT-SURV-LIFETAB-002 · status: passed

Reference

Evidence type
R
Source
Base R arithmetic of the actuarial, Gehan and grouped PH formulas
Citation
Cutler and Ederer (1958) J Chronic Dis 8:699-712; Greenwood (1926); Lee and Desu (1972) Computer Programs in Biomedicine 2:315-321; Gehan (1965) Biometrika 52:203-223; Prentice and Gloeckler (1978) Biometrics 34:57-67; R stats: pchisq, glm (cloglog cross-check), anova(test = 'Rao').
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_life_tables.R · sha256 e7bfcfb4270e21ec
Dataset
reference_data/datasets/SURV_LIFETAB.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/strata/[stratum=A]/subjectsexact15150.00E+2passed
/strata/[stratum=A]/eventsexact11110.00E+2passed
/strata/[stratum=A]/median/valuetolerance5.33804347826086995.338043478260871.04E-1717passed
/strata/[stratum=A]/rows/0/enteringexact15150.00E+2passed
/strata/[stratum=A]/rows/0/withdrawnexact000.00E+2passed
/strata/[stratum=A]/rows/0/eventsexact110.00E+2passed
/strata/[stratum=A]/rows/0/at_risktolerance1515.00.00E+2passed
/strata/[stratum=A]/rows/0/qtolerance0.0666666666666666660.066666666666666672.59E-1917passed
/strata/[stratum=A]/rows/0/survival_starttolerance11.00.00E+2passed
/strata/[stratum=A]/rows/0/survival_endtolerance0.933333333333333350.93333333333333331.86E-1817passed
/strata/[stratum=A]/rows/0/survival_end_setolerance0.0644061188719530640.064406118871953063.80E-1917passed
/strata/[stratum=A]/rows/0/hazardtolerance0.0689655172413793090.068965517241379313.88E-1917passed
/strata/[stratum=A]/rows/0/densitytolerance0.0666666666666666660.066666666666666672.59E-1917passed
/strata/[stratum=A]/rows/0/hazard_setolerance0.068924502954527650.068924502954527653.28E-2018passed
/strata/[stratum=A]/rows/0/density_setolerance0.0644061188719530640.064406118871953063.80E-1917passed
/strata/[stratum=A]/rows/1/enteringexact14140.00E+2passed
/strata/[stratum=A]/rows/1/withdrawnexact110.00E+2passed
/strata/[stratum=A]/rows/1/eventsexact220.00E+2passed
/strata/[stratum=A]/rows/1/at_risktolerance13.513.50.00E+1passed
/strata/[stratum=A]/rows/1/qtolerance0.148148148148148140.148148148148148147.57E-2018passed
/strata/[stratum=A]/rows/1/survival_starttolerance0.933333333333333350.93333333333333331.86E-1817passed
/strata/[stratum=A]/rows/1/survival_start_setolerance0.0644061188719530640.064406118871953063.80E-1917passed
/strata/[stratum=A]/rows/1/survival_endtolerance0.795061728395061730.79506172839506171.33E-1817passed
/strata/[stratum=A]/rows/1/survival_end_setolerance0.105609676026385140.105609676026385143.62E-1816passed
/strata/[stratum=A]/rows/1/hazardtolerance0.0800000000000000020.083.35E-1917passed
/strata/[stratum=A]/rows/1/densitytolerance0.0691358024691357960.06913580246913581.44E-1917passed
/strata/[stratum=A]/rows/1/hazard_setolerance0.0563872325974595080.056387232597459511.12E-1917passed
/strata/[stratum=A]/rows/1/density_setolerance0.0453716223399453640.0453716223399453641.34E-1917passed
/strata/[stratum=A]/rows/2/enteringexact11110.00E+2passed
/strata/[stratum=A]/rows/2/withdrawnexact110.00E+2passed
/strata/[stratum=A]/rows/2/eventsexact550.00E+2passed
/strata/[stratum=A]/rows/2/at_risktolerance10.510.50.00E+1passed
/strata/[stratum=A]/rows/2/qtolerance0.476190476190476160.476190476190476164.04E-1817passed
/strata/[stratum=A]/rows/2/survival_starttolerance0.795061728395061730.79506172839506171.33E-1817passed
/strata/[stratum=A]/rows/2/survival_start_setolerance0.105609676026385140.105609676026385143.62E-1816passed
/strata/[stratum=A]/rows/2/survival_endtolerance0.416460905349794240.416460905349794241.47E-1817passed
/strata/[stratum=A]/rows/2/survival_end_setolerance0.134449447533464630.134449447533464633.29E-1816passed
/strata/[stratum=A]/rows/2/hazardtolerance0.208333333333333340.208333333333333342.59E-1816passed
/strata/[stratum=A]/rows/2/densitytolerance0.126200274348422490.12620027434842253.52E-1816passed
/strata/[stratum=A]/rows/2/hazard_setolerance0.0885033618123251540.088503361812325152.01E-1917passed
/strata/[stratum=A]/rows/2/density_setolerance0.0441531922610342780.044153192261034283.54E-1917passed
/strata/[stratum=A]/rows/3/enteringexact550.00E+2passed
/strata/[stratum=A]/rows/3/withdrawnexact220.00E+2passed
/strata/[stratum=A]/rows/3/eventsexact330.00E+2passed
/strata/[stratum=A]/rows/3/at_risktolerance44.00.00E+2passed
/strata/[stratum=A]/rows/3/qtolerance0.750.750.00E+0passed
/strata/[stratum=A]/rows/3/survival_starttolerance0.416460905349794240.416460905349794241.47E-1817passed
/strata/[stratum=A]/rows/3/survival_start_setolerance0.134449447533464630.134449447533464633.29E-1816passed
/strata/[stratum=A]/rows/3/survival_endtolerance0.104115226337448560.104115226337448563.68E-1917passed
/strata/[stratum=A]/rows/3/survival_end_setolerance0.0962277306074084540.096227730607408453.00E-1917passed
/strata/[stratum=A]/rows/3/hazardtolerance0.200000000000000010.21.10E-1817passed
/strata/[stratum=A]/rows/3/densitytolerance0.052057613168724280.052057613168724281.84E-1917passed
/strata/[stratum=A]/rows/3/hazard_setolerance0.0923760430703401140.092376043070340113.61E-1917passed
/strata/[stratum=A]/rows/3/density_setolerance0.0225450802152776310.022545080215277631.19E-1917passed
/strata/[stratum=B]/subjectsexact15150.00E+2passed
/strata/[stratum=B]/eventsexact10100.00E+2passed
/strata/[stratum=B]/median/valuetolerance7.80847058823529327.8084705882352933.62E-1717passed
/strata/[stratum=B]/rows/0/enteringexact15150.00E+2passed
/strata/[stratum=B]/rows/0/withdrawnexact000.00E+2passed
/strata/[stratum=B]/rows/0/eventsexact000.00E+2passed
/strata/[stratum=B]/rows/0/at_risktolerance1515.00.00E+2passed
/strata/[stratum=B]/rows/0/qtolerance00.00.00E+2passed
/strata/[stratum=B]/rows/0/survival_starttolerance11.00.00E+2passed
/strata/[stratum=B]/rows/0/survival_endtolerance11.00.00E+2passed
/strata/[stratum=B]/rows/0/survival_end_setolerance00.00.00E+2passed
/strata/[stratum=B]/rows/0/hazardtolerance00.00.00E+2passed
/strata/[stratum=B]/rows/0/densitytolerance00.00.00E+2passed
/strata/[stratum=B]/rows/1/enteringexact15150.00E+2passed
/strata/[stratum=B]/rows/1/withdrawnexact110.00E+2passed
/strata/[stratum=B]/rows/1/eventsexact220.00E+2passed
/strata/[stratum=B]/rows/1/at_risktolerance14.514.50.00E+1passed
/strata/[stratum=B]/rows/1/qtolerance0.137931034482758620.137931034482758621.22E-1817passed
/strata/[stratum=B]/rows/1/survival_starttolerance11.00.00E+2passed
/strata/[stratum=B]/rows/1/survival_endtolerance0.862068965517241330.86206896551724134.29E-1817passed
/strata/[stratum=B]/rows/1/survival_end_setolerance0.0905561605929121050.09055616059291214.37E-2018passed
/strata/[stratum=B]/rows/1/hazardtolerance0.074074074074074070.074074074074074073.79E-2018passed
/strata/[stratum=B]/rows/1/densitytolerance0.0689655172413793090.068965517241379313.88E-1917passed
/strata/[stratum=B]/rows/1/hazard_setolerance0.0522343834421392860.0522343834421392861.97E-1917passed
/strata/[stratum=B]/rows/1/density_setolerance0.0452780802964560460.0452780802964560464.61E-1916passed
/strata/[stratum=B]/rows/2/enteringexact12120.00E+2passed
/strata/[stratum=B]/rows/2/withdrawnexact110.00E+2passed
/strata/[stratum=B]/rows/2/eventsexact330.00E+2passed
/strata/[stratum=B]/rows/2/at_risktolerance11.511.50.00E+1passed
/strata/[stratum=B]/rows/2/qtolerance0.26086956521739130.26086956521739132.89E-1816passed
/strata/[stratum=B]/rows/2/survival_starttolerance0.862068965517241330.86206896551724134.29E-1817passed
/strata/[stratum=B]/rows/2/survival_start_setolerance0.0905561605929121050.09055616059291214.37E-2018passed
/strata/[stratum=B]/rows/2/survival_endtolerance0.637181409295352270.63718140929535234.57E-1817passed
/strata/[stratum=B]/rows/2/survival_end_setolerance0.1301551545127580.1301551545127582.15E-1816passed
/strata/[stratum=B]/rows/2/hazardtolerance0.099999999999999990.099999999999999991.67E-1816passed
/strata/[stratum=B]/rows/2/densitytolerance0.0749625187406296740.074962518740629674.63E-1917passed
/strata/[stratum=B]/rows/2/hazard_setolerance0.0570818126318123620.057081812631812362.00E-2018passed
/strata/[stratum=B]/rows/2/density_setolerance0.038032771098544450.038032771098544453.17E-1917passed
/strata/[stratum=B]/rows/3/enteringexact880.00E+2passed
/strata/[stratum=B]/rows/3/withdrawnexact220.00E+2passed
/strata/[stratum=B]/rows/3/eventsexact550.00E+2passed
/strata/[stratum=B]/rows/3/at_risktolerance77.00.00E+2passed
/strata/[stratum=B]/rows/3/qtolerance0.71428571428571430.71428571428571431.57E-1817passed
/strata/[stratum=B]/rows/3/survival_starttolerance0.637181409295352270.63718140929535234.57E-1817passed
/strata/[stratum=B]/rows/3/survival_start_setolerance0.1301551545127580.1301551545127582.15E-1816passed
/strata/[stratum=B]/rows/3/survival_endtolerance0.18205183122724350.18205183122724359.11E-1917passed
/strata/[stratum=B]/rows/3/survival_end_setolerance0.114976631957287250.114976631957287254.72E-1816passed
/strata/[stratum=B]/rows/3/hazardtolerance0.18518518518518520.18518518518518522.66E-1816passed
/strata/[stratum=B]/rows/3/densitytolerance0.0758549296780181290.075854929678018132.47E-1917passed
/strata/[stratum=B]/rows/3/hazard_setolerance0.0688609075336687750.068860907533668771.84E-1917passed
/strata/[stratum=B]/rows/3/density_setolerance0.0238512648172498340.0238512648172498342.71E-1916passed
/strata/[stratum=B]/rows/4/enteringexact110.00E+2passed
/strata/[stratum=B]/rows/4/withdrawnexact110.00E+2passed
/strata/[stratum=B]/rows/4/eventsexact000.00E+2passed
/strata/[stratum=B]/rows/4/at_risktolerance0.50.50.00E+1passed
/strata/[stratum=B]/rows/4/qtolerance00.00.00E+2passed
/strata/[stratum=B]/rows/4/survival_starttolerance0.18205183122724350.18205183122724359.11E-1917passed
/strata/[stratum=B]/rows/4/survival_start_setolerance0.114976631957287250.114976631957287254.72E-1816passed
/comparison/statistictolerance1.01410874704491731.01410874704491733.08E-1716passed
/comparison/dfexact110.00E+2passed
/comparison/p_valuetail_probability0.313920518565545150.31392051856554222.94E-1514passed
/comparison/strata/[stratum=A]/score_sumtolerance-43-43.00.00E+2passed
/comparison/strata/[stratum=B]/mean_scoretolerance2.86666666666666672.86666666666666673.73E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

249  grouped <- edges2[findInterval(d$time, edges2)]
250  cmp <- gehan(grouped, event, d$group)
251  outputs[["/comparison/statistic"]] <- tol(cmp$statistic, "Lee-Desu chi-square on grouped times")
252  outputs[["/comparison/df"]] <- exact(as.integer(cmp$df), "strata minus one")
253  outputs[["/comparison/p_value"]] <- prob(cmp$p, "Lee-Desu chi-square tail")
254  outputs[["/comparison/strata/[stratum=A]/score_sum"]] <- tol(cmp$sums[["A"]], "Gehan score sum of stratum A")
255  outputs[["/comparison/strata/[stratum=B]/mean_score"]] <- tol(cmp$means[["B"]], "mean Gehan score of stratum B")
256  emit("STAT-SURV-LIFETAB-002", "Unequal boundaries with two strata and the grouped Gehan test", "axis_coverage",
257       "SURV_LIFETAB", list(time = "time", status = "status", event_value = "1", censored_value = "0",
258                            boundaries = list(0, 1, 3, 6, 12), strata = "group"),
259       outputs,
260       list("Explicit boundaries 0, 1, 3, 6, 12; each stratum gets its own table and median.",
261            "Gehan scores by the pairwise definition on interval starts (open interval -> 12): case a scores +1 against an uncensored b with an earlier grouped time, or an uncensored b at the same grouped time when a is censored; -1 symmetrically; the scores sum to zero (asserted)."))

Whole script generators/r/survival_life_tables.R


Scheduled visits with text status values and the grouped PH score test

Case STAT-SURV-LIFETAB-003 · status: passed

Reference

Evidence type
R
Source
Base R arithmetic of the actuarial, Gehan and grouped PH formulas
Citation
Cutler and Ederer (1958) J Chronic Dis 8:699-712; Greenwood (1926); Lee and Desu (1972) Computer Programs in Biomedicine 2:315-321; Gehan (1965) Biometrika 52:203-223; Prentice and Gloeckler (1978) Biometrics 34:57-67; R stats: pchisq, glm (cloglog cross-check), anova(test = 'Rao').
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_life_tables.R · sha256 e7bfcfb4270e21ec
Dataset
reference_data/datasets/SURV_LIFETAB_VISITS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/strata/[stratum=A]/rows/0/enteringexact15150.00E+2passed
/strata/[stratum=A]/rows/0/withdrawnexact000.00E+2passed
/strata/[stratum=A]/rows/0/eventsexact000.00E+2passed
/strata/[stratum=A]/rows/0/at_risktolerance1515.00.00E+2passed
/strata/[stratum=A]/rows/0/qtolerance00.00.00E+2passed
/strata/[stratum=A]/rows/0/survival_starttolerance11.00.00E+2passed
/strata/[stratum=A]/rows/0/survival_endtolerance11.00.00E+2passed
/strata/[stratum=A]/rows/0/survival_end_setolerance00.00.00E+2passed
/strata/[stratum=A]/rows/0/hazardtolerance00.00.00E+2passed
/strata/[stratum=A]/rows/0/densitytolerance00.00.00E+2passed
/strata/[stratum=A]/rows/1/enteringexact15150.00E+2passed
/strata/[stratum=A]/rows/1/withdrawnexact000.00E+2passed
/strata/[stratum=A]/rows/1/eventsexact220.00E+2passed
/strata/[stratum=A]/rows/1/at_risktolerance1515.00.00E+2passed
/strata/[stratum=A]/rows/1/qtolerance0.133333333333333330.133333333333333331.48E-1816passed
/strata/[stratum=A]/rows/1/survival_starttolerance11.00.00E+2passed
/strata/[stratum=A]/rows/1/survival_endtolerance0.86666666666666670.86666666666666673.73E-1817passed
/strata/[stratum=A]/rows/1/survival_end_setolerance0.0877707451472511110.087770745147251113.18E-1917passed
/strata/[stratum=A]/rows/1/hazardtolerance0.142857142857142850.142857142857142857.87E-1917passed
/strata/[stratum=A]/rows/1/densitytolerance0.133333333333333330.133333333333333331.48E-1816passed
/strata/[stratum=A]/rows/1/hazard_setolerance0.10075723294965050.10075723294965057.46E-1917passed
/strata/[stratum=A]/rows/1/density_setolerance0.0877707451472511110.087770745147251113.18E-1917passed
/strata/[stratum=A]/rows/2/enteringexact13130.00E+2passed
/strata/[stratum=A]/rows/2/withdrawnexact110.00E+2passed
/strata/[stratum=A]/rows/2/eventsexact220.00E+2passed
/strata/[stratum=A]/rows/2/at_risktolerance12.512.50.00E+1passed
/strata/[stratum=A]/rows/2/qtolerance0.160.163.33E-1816passed
/strata/[stratum=A]/rows/2/survival_starttolerance0.86666666666666670.86666666666666673.73E-1817passed
/strata/[stratum=A]/rows/2/survival_start_setolerance0.0877707451472511110.087770745147251113.18E-1917passed
/strata/[stratum=A]/rows/2/survival_endtolerance0.727999999999999980.7284.60E-1918passed
/strata/[stratum=A]/rows/2/survival_end_setolerance0.116239752236487480.116239752236487484.12E-1816passed
/strata/[stratum=A]/rows/2/hazardtolerance0.173913043478260890.17391304347826092.49E-1816passed
/strata/[stratum=A]/rows/2/densitytolerance0.138666666666666690.13866666666666671.94E-1816passed
/strata/[stratum=A]/rows/2/hazard_setolerance0.122509275962341420.122509275962341424.30E-1816passed
/strata/[stratum=A]/rows/2/density_setolerance0.0909569210202361980.09095692102023622.61E-1917passed
/strata/[stratum=B]/rows/0/withdrawnexact110.00E+2passed
/grouped_ph/statistictolerance0.189497371186181180.189497371186181186.30E-1917passed
/grouped_ph/dfexact110.00E+2passed
/grouped_ph/p_valuetail_probability0.663335280692431310.66333528069243133.91E-1817passed
/grouped_ph/groups/[group=A]/scoretolerance0.920232917276492810.92023291727649282.24E-1817passed
/grouped_ph/groups/[group=B]/lost_before_first_visitexact110.00E+2passed
/grouped_ph/information/0/0tolerance4.46881461594100584.4688146159410061.89E-1717passed
/grouped_ph/zero_information_intervalsexact000.00E+2passed
/grouped_ph/intervals/0/at_risk/1exact14140.00E+2passed
/grouped_ph/intervals/2/events/0exact220.00E+2passed
/comparison/statistictolerance0.500133292320311670.50013329232031173.15E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

264  edges3 <- 0:6
265  eventv <- dv$status == "event"
266  outputs <- list()
267  rows_a <- life_table(dv$time[dv$group == "A"], eventv[dv$group == "A"], edges3)
268  outputs <- c(outputs, row_outputs(rows_a, "/strata/[stratum=A]", 1:3))
269  outputs[["/strata/[stratum=B]/rows/0/withdrawn"]] <- exact(1L, "the subject censored at the origin is withdrawn in [0, 1)")
270  ph <- grouped_ph(dv$time, eventv, dv$group, length(edges3) - 1)
271  outputs[["/grouped_ph/statistic"]] <- tol(ph$statistic, "Prentice-Gloeckler score statistic U' I^-1 U")
272  outputs[["/grouped_ph/df"]] <- exact(as.integer(ph$df), "groups minus one")
273  outputs[["/grouped_ph/p_value"]] <- prob(ph$p, "grouped PH chi-square tail")
274  outputs[["/grouped_ph/groups/[group=A]/score"]] <- tol(ph$scores[1], "score of group A")
275  outputs[["/grouped_ph/groups/[group=B]/lost_before_first_visit"]] <- exact(1L, "one subject of B censored at the origin")
276  outputs[["/grouped_ph/information/0/0"]] <- tol(ph$information[1, 1], "information of the first group coordinate")
277  outputs[["/grouped_ph/zero_information_intervals"]] <- exact(0L, "every visit interval has 0 < q < 1")
278  outputs[["/grouped_ph/intervals/0/at_risk/1"]] <- exact(as.integer(ph$r[2, 1]), "B at risk in (0, 1]: 14 (the origin-censored subject is not)")
279  outputs[["/grouped_ph/intervals/2/events/0"]] <- exact(as.integer(ph$d[1, 3]), "A events detected at visit 3")
280  cmp3 <- gehan(edges3[findInterval(dv$time, edges3)], eventv, dv$group)
281  outputs[["/comparison/statistic"]] <- tol(cmp3$statistic, "Lee-Desu chi-square on the visit times")
282  emit("STAT-SURV-LIFETAB-003", "Scheduled visits with text status values and the grouped PH score test", "axis_coverage",
283       "SURV_LIFETAB_VISITS", list(time = "time", status = "status", event_value = "event",
284                                   censored_value = "censored", boundaries = as.list(as.numeric(edges3)),
285                                   strata = "group", observation_model = "scheduled_visits",
286                                   grouped_ph_test = TRUE),
287       outputs,
288       list("Status values are the texts 'event' and 'censored'.",
289            "Grouped PH score test (docs/manual/chapters/107-grouped-ph.tex): r[g, j] subjects of group g with visit index m >= j, d[g, j] events detected at visit j; under H0 q_j = d_j/r_j, a_j = -log(1 - q_j); U[g] = sum_j (a_j/q_j)(d[g, j] - r[g, j] q_j), I[g, h] = sum_j a_j^2 (1 - q_j)/q_j (delta r[g, j] - r[g, j] r[h, j]/r_j); Q = U' I^-1 U on the first G - 1 coordinates.",
290            "The by-hand statistic is cross-checked in the generator against the expected-information score test of the group term computed from the null complementary log-log binomial GLM fit (one intercept per visit interval; glm fitted values, mu.eta and model matrix) to 1e-9 relative, and against anova(test = 'Rao') to 1e-3; the generator stops if either differs more."))

Whole script generators/r/survival_life_tables.R


Survival STAT-SURV-LOGRANK-001

Case STAT-SURV-LOGRANK-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/survdiff/coxph and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, MASS 7.3.65, jsonlite 2.0.0
Reference script
generators/r/survival_core.R · sha256 1ed8945e6864eb84
Dataset
reference_data/datasets/HARVEST_GEHAN_BOTH.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/log_rank/statistictolerance16.7929409892165416.792940989216544.51E-1616passed
/log_rank/observedsequence[9, 21][9, 21]passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

48  group_fit <- survival::survdiff(survival::Surv(time,status)~arm,data=both)
49  emit("STAT-SURV-LOGRANK-001","tabelz.executor.survival","survivalMenuItem",
50       "HARVEST_GEHAN_BOTH",c(config,list(group_column="arm")),
51       list("/log_rank/statistic"=tol(group_fit$chisq,"R two-group log-rank chi-square"),
52            "/log_rank/observed"=seqexact(group_fit$obs,"R observed failures by arm")),
53       c("R survdiff rho=0 gives the log-rank statistic; checks V14 Gehan grouping."))

Whole script generators/r/survival_core.R


Kaplan-Meier life estimation with frequencies at 95%

Case STAT-SURV-NP-001 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_KM_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact22220.00E+2passed
/counts/missing_countexact000.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact22220.00E+2passed
/counts/exacttolerance1414.00.00E+2passed
/counts/righttolerance1111.00.00E+2passed
/counts/lefttolerance00.00.00E+2passed
/counts/intervaltolerance00.00.00E+2passed
/estimate/kaplan_meier/0/timetolerance28.528.50.00E+1passed
/estimate/kaplan_meier/0/at_risktolerance2525.00.00E+2passed
/estimate/kaplan_meier/0/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/0/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/0/survivaltolerance0.959999999999999960.964.47E-1817passed
/estimate/kaplan_meier/0/lowertolerance0.748385024194757940.74838502419475791.81E-1817passed
/estimate/kaplan_meier/0/uppertolerance0.994266950748837750.99426695074883782.27E-1817passed
/estimate/kaplan_meier/1/timetolerance31.531.50.00E+1passed
/estimate/kaplan_meier/1/at_risktolerance2424.00.00E+2passed
/estimate/kaplan_meier/1/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/1/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/1/survivaltolerance0.920000000000000040.923.20E-2019passed
/estimate/kaplan_meier/1/lowertolerance0.716390337258063450.71639033725806342.34E-1817passed
/estimate/kaplan_meier/1/uppertolerance0.979370585428049560.97937058542804964.04E-1817passed
/estimate/kaplan_meier/2/timetolerance50.20000000000000350.21.58E-1617passed
/estimate/kaplan_meier/2/at_risktolerance2323.00.00E+2passed
/estimate/kaplan_meier/2/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/2/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/2/survivaltolerance0.880000000000000120.88000000000000014.54E-1817passed
/estimate/kaplan_meier/2/lowertolerance0.672561452863359220.67256145286335921.51E-1817passed
/estimate/kaplan_meier/2/uppertolerance0.959639832332032540.95963983233203251.47E-1817passed
/estimate/kaplan_meier/3/timetolerance51.60000000000000151.64.21E-1617passed
/estimate/kaplan_meier/3/at_risktolerance2222.00.00E+2passed
/estimate/kaplan_meier/3/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/3/censoredtolerance11.00.00E+2passed
/estimate/kaplan_meier/3/survivaltolerance0.840000000000000190.84000000000000029.58E-1917passed
/estimate/kaplan_meier/3/lowertolerance0.628056081556177980.62805608155617791.10E-1615passed
/estimate/kaplan_meier/3/uppertolerance0.936733240251340080.93673324025134012.42E-1817passed
/estimate/kaplan_meier/4/timetolerance57.10000000000000157.14.21E-1617passed
/estimate/kaplan_meier/4/at_risktolerance1919.00.00E+2passed
/estimate/kaplan_meier/4/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/4/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/4/survivaltolerance0.795789473684210710.79578947368421073.07E-1817passed
/estimate/kaplan_meier/4/lowertolerance0.576216419680682710.57621641968068261.09E-1615passed
/estimate/kaplan_meier/4/uppertolerance0.909694433563694840.90969443356369482.45E-1817passed
/estimate/kaplan_meier/5/timetolerance73.79999999999999773.81.58E-1617passed
/estimate/kaplan_meier/5/at_risktolerance1616.00.00E+2passed
/estimate/kaplan_meier/5/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/5/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/5/survivaltolerance0.746052631578947520.74605263157894752.68E-1817passed
/estimate/kaplan_meier/5/lowertolerance0.517121503019855710.51712150301985575.48E-1917passed
/estimate/kaplan_meier/5/uppertolerance0.877971626704724060.87797162670472411.87E-1817passed
/estimate/kaplan_meier/6/timetolerance77.59999999999999477.63.16E-1617passed
/estimate/kaplan_meier/6/at_risktolerance1515.00.00E+2passed
/estimate/kaplan_meier/6/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/6/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/6/survivaltolerance0.696315789473684330.69631578947368432.30E-1817passed
/estimate/kaplan_meier/6/lowertolerance0.462752282807844690.46275228280784461.06E-1615passed
/estimate/kaplan_meier/6/uppertolerance0.843650064472524660.84365006447252481.14E-1615passed
/estimate/kaplan_meier/7/timetolerance83.40000000000000683.43.16E-1617passed
/estimate/kaplan_meier/7/at_risktolerance1414.00.00E+2passed
/estimate/kaplan_meier/7/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/7/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/7/survivaltolerance0.646578947368421140.64657894736842111.92E-1817passed
/estimate/kaplan_meier/7/lowertolerance0.411842372997761440.411842372997761441.14E-1817passed
/estimate/kaplan_meier/7/uppertolerance0.807070150005442690.80707015000544281.09E-1615passed
/estimate/kaplan_meier/8/timetolerance97.70000000000000397.71.58E-1617passed
/estimate/kaplan_meier/8/at_risktolerance1313.00.00E+2passed
/estimate/kaplan_meier/8/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/8/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/8/survivaltolerance0.596842105263158060.59684210526315812.56E-1817passed
/estimate/kaplan_meier/8/lowertolerance0.363745381469426750.363745381469426753.98E-1816passed
/estimate/kaplan_meier/8/uppertolerance0.768445668129463690.76844566812946381.09E-1615passed
/estimate/kaplan_meier/9/timetolerance116.5116.50.00E+1passed
/estimate/kaplan_meier/9/at_risktolerance1212.00.00E+2passed
/estimate/kaplan_meier/9/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/9/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/9/survivaltolerance0.547105263157894870.54710526315789492.17E-1817passed
/estimate/kaplan_meier/9/lowertolerance0.318105659698038060.318105659698038063.14E-1817passed
/estimate/kaplan_meier/9/uppertolerance0.727908810527364560.72790881052736461.86E-1817passed
/estimate/kaplan_meier/10/timetolerance135.30000000000001135.31.37E-1516passed
/estimate/kaplan_meier/10/at_risktolerance1010.00.00E+2passed
/estimate/kaplan_meier/10/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/10/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/10/survivaltolerance0.492394736842105380.49239473684210544.96E-1816passed
/estimate/kaplan_meier/10/lowertolerance0.268293697599542910.26829369759954292.97E-1816passed
/estimate/kaplan_meier/10/uppertolerance0.682832770050260770.68283277005026091.14E-1615passed
/estimate/kaplan_meier/11/timetolerance147.69999999999999147.71.37E-1517passed
/estimate/kaplan_meier/11/at_risktolerance44.00.00E+2passed
/estimate/kaplan_meier/11/failurestolerance22.00.00E+2passed
/estimate/kaplan_meier/11/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/11/survivaltolerance0.246197368421052690.24619736842105272.48E-1816passed
/estimate/kaplan_meier/11/lowertolerance0.049042715446879180.049042715446879126.23E-1714passed
/estimate/kaplan_meier/11/uppertolerance0.521223602752182380.52122360275218251.11E-1615passed
/estimate/kaplan_meier/12/timetolerance151.69999999999999151.71.37E-1517passed
/estimate/kaplan_meier/12/at_risktolerance22.00.00E+2passed
/estimate/kaplan_meier/12/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/12/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/12/survivaltolerance0.123098684210526350.123098684210526353.76E-1816passed
/estimate/kaplan_meier/12/lowertolerance0.00792890731453564470.0079289073145356368.65E-1814passed
/estimate/kaplan_meier/12/uppertolerance0.403676844459238630.40367684445923875.18E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

327  km_case("STAT-SURV-NP-001", "Kaplan-Meier life estimation with frequencies at 95%",
328          0.95, NP_NOTES, "nominal")

Whole script generators/r/survival_life_families.R


Turnbull nonparametric life estimation from interval data

Case STAT-SURV-NP-002 · status: passed

Reference

Evidence type
R
Source
Retained R reference with independently reviewed analytical corrections
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
not recorded
Reference script
generators/analytic/reviewed_reference_corrections.py · sha256 49e3645836bb3ac3
Dataset
reference_data/datasets/SURV_INTERVAL_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact12120.00E+2passed
/counts/missing_countexact000.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact12120.00E+2passed
/counts/exacttolerance77.00.00E+2passed
/counts/righttolerance22.00.00E+2passed
/counts/lefttolerance11.00.00E+2passed
/counts/intervaltolerance55.00.00E+2passed
/estimate/turnbull/0/lowtolerance88.00.00E+2passed
/estimate/turnbull/0/hightolerance88.00.00E+2passed
/estimate/turnbull/0/masstolerance0.31759546816668068690605508875124825097062544269256495171520.317595468166680662.57E-1716passed
/estimate/turnbull/0/survival_beforetolerance1.00000000000000000000000000000000000000000000000000000000001.00.00E-67passed
/estimate/turnbull/0/survival_aftertolerance0.68240453183331931309394491124875174902937455730743504828470.68240453183331948.13E-1715passed
/estimate/turnbull/1/lowtolerance1212.00.00E+2passed
/estimate/turnbull/1/hightolerance1414.00.00E+2passed
/estimate/turnbull/1/masstolerance00.00.00E+2passed
/estimate/turnbull/1/survival_beforetolerance0.68240453183331931309394491124875174902937455730743504828470.68240453183331948.13E-1715passed
/estimate/turnbull/1/survival_aftertolerance0.68240453183331931309394491124875174902937455730743504828470.68240453183331948.13E-1715passed
/estimate/turnbull/2/lowtolerance1515.00.00E+2passed
/estimate/turnbull/2/hightolerance1515.00.00E+2passed
/estimate/turnbull/2/masstolerance0.09814239699997195952122315583083683139208244794820342990270.098142396999972384.16E-1614passed
/estimate/turnbull/2/survival_beforetolerance0.68240453183331931309394491124875174902937455730743504828470.68240453183331948.13E-1715passed
/estimate/turnbull/2/survival_aftertolerance0.58426213483334735357272175541791491763729210935923161838190.5842621348333473.63E-1615passed
/estimate/turnbull/3/lowtolerance2424.00.00E+2passed
/estimate/turnbull/3/hightolerance2424.00.00E+2passed
/estimate/turnbull/3/masstolerance0.31759546816668068690605508875124825097062544269256495171520.317595468166679161.52E-1514passed
/estimate/turnbull/3/survival_beforetolerance0.58426213483334735357272175541791491763729210935923161838190.5842621348333473.63E-1615passed
/estimate/turnbull/3/survival_aftertolerance0.26666666666666666666666666666666666666666666666666666666660.266666666666667831.16E-1514passed
/estimate/turnbull/4/lowtolerance2828.00.00E+2passed
/estimate/turnbull/4/hightolerance3030.00.00E+2passed
/estimate/turnbull/4/masstolerance00.00.00E+2passed
/estimate/turnbull/4/survival_beforetolerance0.26666666666666666666666666666666666666666666666666666666660.266666666666667831.16E-1514passed
/estimate/turnbull/4/survival_aftertolerance0.26666666666666666666666666666666666666666666666666666666660.266666666666667831.16E-1514passed
/estimate/turnbull/5/lowtolerance3030.00.00E+2passed
/estimate/turnbull/5/hightolerance4040.00.00E+2passed
/estimate/turnbull/5/masstolerance00.00.00E+2passed
/estimate/turnbull/5/survival_beforetolerance0.26666666666666666666666666666666666666666666666666666666660.266666666666667831.16E-1514passed
/estimate/turnbull/5/survival_aftertolerance0.26666666666666666666666666666666666666666666666666666666660.266666666666667831.16E-1514passed
/estimate/turnbull/6/lowtolerance4040.00.00E+2passed
/estimate/turnbull/6/hightolerance5050.00.00E+2passed
/estimate/turnbull/6/masstolerance0.13333333333333333333333333333333333333333333333333333333330.133333333333334441.11E-1514passed
/estimate/turnbull/6/survival_beforetolerance0.26666666666666666666666666666666666666666666666666666666660.266666666666667831.16E-1514passed
/estimate/turnbull/6/survival_aftertolerance0.13333333333333333333333333333333333333333333333333333333330.13333333333333345.37E-1715passed
/estimate/turnbull/7/lowtolerance5555.00.00E+2passed
/estimate/turnbull/7/hightolerance5555.00.00E+2passed
/estimate/turnbull/7/masstolerance0.13333333333333333333333333333333333333333333333333333333330.133333333333333471.37E-1614passed
/estimate/turnbull/7/survival_beforetolerance0.13333333333333333333333333333333333333333333333333333333330.13333333333333345.37E-1715passed
/estimate/turnbull/7/survival_aftertolerance00.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

39  def corrected_reference(root: Path, case_id: str) -> dict[str, Any]:
40      revisions = json.loads((root / REVISION_PATH).read_bytes())
41      proof = _read_bound(root, revisions["proof"])
42      revision = revisions["cases"][case_id]
43      original = _read_bound(root, revision["original_expected"])
44      data = (root / revision["csv"]["path"]).read_bytes()
45      if hashlib.sha256(data).hexdigest() != revision["csv"]["sha256"]:
46          raise ValueError("Analytical correction applies only to the reviewed dataset")
47      result = copy.deepcopy(original)
48      entries = [
49          item
50          for item in proof["entries"]
51          if item["case_id"] == case_id
52          and item["disposition"] == "approved_mathematical_zero_reference_correction"
53      ]
54      if not entries:
55          raise ValueError("No approved correction for this case")
56      for item in entries:
57          if result["outputs"][item["path"]] != item["retained_reference"]:
58              raise ValueError("Analytical correction preimage differs")
59          result["outputs"][item["path"]].update(
60              value="0",
61              rationale=correction_rationale(item["proof"]),
62          )
63      if case_id == "STAT-SURV-NP-002":
64          solution = proof["proofs"]["turnbull_exact_kkt"]
65          with localcontext() as context:
66              context.prec = 70
67              sqrt5 = Decimal(5).sqrt()

Whole script generators/analytic/reviewed_reference_corrections.py


Kaplan-Meier life estimation with a tied failure and censoring at 90%

Case STAT-SURV-NP-003 · status: passed

Reference

Evidence type
R
Source
R survival package: survreg, survfit, predict.survreg
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley; Kalbfleisch and Prentice (2002).
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_life_families.R · sha256 5e01f9659bbd2ca2
Dataset
reference_data/datasets/SURV_KM_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/counts/source_countexact22220.00E+2passed
/counts/missing_countexact000.00E+2passed
/counts/zero_frequency_rowsexact000.00E+2passed
/counts/rows_usedexact22220.00E+2passed
/counts/exacttolerance1414.00.00E+2passed
/counts/righttolerance1111.00.00E+2passed
/counts/lefttolerance00.00.00E+2passed
/counts/intervaltolerance00.00.00E+2passed
/estimate/kaplan_meier/0/timetolerance28.528.50.00E+1passed
/estimate/kaplan_meier/0/at_risktolerance2525.00.00E+2passed
/estimate/kaplan_meier/0/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/0/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/0/survivaltolerance0.959999999999999960.964.47E-1817passed
/estimate/kaplan_meier/0/lowertolerance0.809375467730280680.80937546773028061.07E-1615passed
/estimate/kaplan_meier/0/uppertolerance0.992151550005184180.99215155000518423.96E-1918passed
/estimate/kaplan_meier/1/timetolerance31.531.50.00E+1passed
/estimate/kaplan_meier/1/at_risktolerance2424.00.00E+2passed
/estimate/kaplan_meier/1/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/1/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/1/survivaltolerance0.920000000000000040.923.20E-2019passed
/estimate/kaplan_meier/1/lowertolerance0.765754174304709380.76575417430470931.13E-1615passed
/estimate/kaplan_meier/1/uppertolerance0.974286732282393930.97428673228239392.27E-1918passed
/estimate/kaplan_meier/2/timetolerance50.20000000000000350.21.58E-1617passed
/estimate/kaplan_meier/2/at_risktolerance2323.00.00E+2passed
/estimate/kaplan_meier/2/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/2/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/2/survivaltolerance0.880000000000000120.88000000000000014.54E-1817passed
/estimate/kaplan_meier/2/lowertolerance0.718463558228600530.71846355822860054.68E-1918passed
/estimate/kaplan_meier/2/uppertolerance0.951778051854959340.95177805185495936.27E-1918passed
/estimate/kaplan_meier/3/timetolerance51.60000000000000151.64.21E-1617passed
/estimate/kaplan_meier/3/at_risktolerance2222.00.00E+2passed
/estimate/kaplan_meier/3/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/3/censoredtolerance11.00.00E+2passed
/estimate/kaplan_meier/3/survivaltolerance0.840000000000000190.84000000000000029.58E-1917passed
/estimate/kaplan_meier/3/lowertolerance0.672169350488438950.67216935048843881.06E-1615passed
/estimate/kaplan_meier/3/uppertolerance0.926329911985810380.92632991198581043.61E-1817passed
/estimate/kaplan_meier/4/timetolerance57.10000000000000157.14.21E-1617passed
/estimate/kaplan_meier/4/at_risktolerance1919.00.00E+2passed
/estimate/kaplan_meier/4/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/4/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/4/survivaltolerance0.795789473684210710.79578947368421073.07E-1817passed
/estimate/kaplan_meier/4/lowertolerance0.619734740364458190.61973474036445811.09E-1615passed
/estimate/kaplan_meier/4/uppertolerance0.896686510664494120.89668651066449411.43E-1817passed
/estimate/kaplan_meier/5/timetolerance73.79999999999999773.81.58E-1617passed
/estimate/kaplan_meier/5/at_risktolerance1616.00.00E+2passed
/estimate/kaplan_meier/5/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/5/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/5/survivaltolerance0.746052631578947520.74605263157894752.68E-1817passed
/estimate/kaplan_meier/5/lowertolerance0.560558730780753360.56055873078075321.15E-1615passed
/estimate/kaplan_meier/5/uppertolerance0.862193488341218560.86219348834121871.12E-1615passed
/estimate/kaplan_meier/6/timetolerance77.59999999999999477.63.16E-1617passed
/estimate/kaplan_meier/6/at_risktolerance1515.00.00E+2passed
/estimate/kaplan_meier/6/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/6/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/6/survivaltolerance0.696315789473684330.69631578947368432.30E-1817passed
/estimate/kaplan_meier/6/lowertolerance0.505394367129400220.50539436712940011.07E-1615passed
/estimate/kaplan_meier/6/uppertolerance0.82532460437598920.82532460437598931.12E-1615passed
/estimate/kaplan_meier/7/timetolerance83.40000000000000683.43.16E-1617passed
/estimate/kaplan_meier/7/at_risktolerance1414.00.00E+2passed
/estimate/kaplan_meier/7/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/7/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/7/survivaltolerance0.646578947368421140.64657894736842111.92E-1817passed
/estimate/kaplan_meier/7/lowertolerance0.4532133833695840.45321338336958391.11E-1615passed
/estimate/kaplan_meier/7/uppertolerance0.786415011649540860.7864150116495411.14E-1615passed
/estimate/kaplan_meier/8/timetolerance97.70000000000000397.71.58E-1617passed
/estimate/kaplan_meier/8/at_risktolerance1313.00.00E+2passed
/estimate/kaplan_meier/8/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/8/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/8/survivaltolerance0.596842105263158060.59684210526315812.56E-1817passed
/estimate/kaplan_meier/8/lowertolerance0.403477168155532750.403477168155532754.93E-1917passed
/estimate/kaplan_meier/8/uppertolerance0.745673052578477160.74567305257847722.06E-1817passed
/estimate/kaplan_meier/9/timetolerance116.5116.50.00E+1passed
/estimate/kaplan_meier/9/at_risktolerance1212.00.00E+2passed
/estimate/kaplan_meier/9/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/9/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/9/survivaltolerance0.547105263157894870.54710526315789492.17E-1817passed
/estimate/kaplan_meier/9/lowertolerance0.355884093286293110.355884093286293065.27E-1715passed
/estimate/kaplan_meier/9/uppertolerance0.703227230563131630.70322723056313171.14E-1615passed
/estimate/kaplan_meier/10/timetolerance135.30000000000001135.31.37E-1516passed
/estimate/kaplan_meier/10/at_risktolerance1010.00.00E+2passed
/estimate/kaplan_meier/10/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/10/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/10/survivaltolerance0.492394736842105380.49239473684210544.96E-1816passed
/estimate/kaplan_meier/10/lowertolerance0.303903471308042290.303903471308042245.10E-1715passed
/estimate/kaplan_meier/10/uppertolerance0.656111185963813020.65611118596381311.06E-1615passed
/estimate/kaplan_meier/11/timetolerance147.69999999999999147.71.37E-1517passed
/estimate/kaplan_meier/11/at_risktolerance44.00.00E+2passed
/estimate/kaplan_meier/11/failurestolerance22.00.00E+2passed
/estimate/kaplan_meier/11/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/11/survivaltolerance0.246197368421052690.24619736842105272.48E-1816passed
/estimate/kaplan_meier/11/lowertolerance0.0695500652609831740.069550065260983125.57E-1715passed
/estimate/kaplan_meier/11/uppertolerance0.478562408365463880.4785624083654641.07E-1615passed
/estimate/kaplan_meier/12/timetolerance151.69999999999999151.71.37E-1517passed
/estimate/kaplan_meier/12/at_risktolerance22.00.00E+2passed
/estimate/kaplan_meier/12/failurestolerance11.00.00E+2passed
/estimate/kaplan_meier/12/censoredtolerance00.00.00E+2passed
/estimate/kaplan_meier/12/survivaltolerance0.123098684210526350.123098684210526353.76E-1816passed
/estimate/kaplan_meier/12/lowertolerance0.0145775976941404320.014577597694140421.21E-1715passed
/estimate/kaplan_meier/12/uppertolerance0.354236575184391220.35423657518439135.80E-1715passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

329  km_case("STAT-SURV-NP-003", "Kaplan-Meier life estimation with a tied failure and censoring at 90%",
330          0.90, c(NP_NOTES, "Rows 3 and 9 share a time; the failure precedes the censoring in both implementations."), "axis_coverage")

Whole script generators/r/survival_life_families.R


Invalid survival request STAT-SURV-RELIABILITY-INVALID-001

Case STAT-SURV-RELIABILITY-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid survival request STAT-SURV-RELIABILITY-INVALID-002

Case STAT-SURV-RELIABILITY-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Repairable-system MCF with a constant-rate HPP overlay

Case STAT-SURV-REPAIR-001 · status: passed

Reference

Evidence type
R
Source
base R MCF and Poisson rate algebra
Citation
Turnbull (1976); Nelson (1988); Crow (1975); Meeker, Hahn and Escobar (1998); Therneau T (2024), A Package for Survival Analysis in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_repairable.R · sha256 1f118d2994e4269d
Dataset
reference_data/datasets/SURV_REPAIR_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/unit_countexact550.00E+2passed
/zero_event_unit_countexact110.00E+2passed
/event_countexact12120.00E+2passed
/distinct_event_age_countexact10100.00E+2passed
/total_exposuretolerance5858.00.00E+2passed
/earliest_entrytolerance00.00.00E+2passed
/latest_endtolerance1212.00.00E+2passed
/event_source_countexact12120.00E+2passed
/event_excluded_countexact000.00E+2passed
/window_source_countexact550.00E+2passed
/window_excluded_countexact000.00E+2passed
/mcf_steps/0/segmentexact000.00E+2passed
/mcf_steps/0/agetolerance1.51.50.00E+1passed
/mcf_steps/0/eventsexact110.00E+2passed
/mcf_steps/0/at_riskexact440.00E+2passed
/mcf_steps/0/incrementtolerance0.250.250.00E+0passed
/mcf_steps/0/mcftolerance0.250.250.00E+0passed
/mcf_steps/0/variancetolerance0.058593750.058593750.00E-6passed
/mcf_steps/0/standard_errortolerance0.242061459137963560.242061459137963563.84E-1816passed
/mcf_steps/0/lowertolerance0.0374772450465100590.0374772450465100451.39E-1715passed
/mcf_steps/0/uppertolerance1.66767861198004731.6676786119800486.65E-1615passed
/mcf_steps/1/segmentexact000.00E+2passed
/mcf_steps/1/agetolerance2.52.50.00E+1passed
/mcf_steps/1/eventsexact110.00E+2passed
/mcf_steps/1/at_riskexact550.00E+2passed
/mcf_steps/1/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/1/mcftolerance0.450000000000000010.451.10E-1817passed
/mcf_steps/1/variancetolerance0.0673437499999999940.067343751.06E-1917passed
/mcf_steps/1/standard_errortolerance0.259506743650333660.259506743650333662.15E-2019passed
/mcf_steps/1/lowertolerance0.145324956347510970.145324956347510942.82E-1715passed
/mcf_steps/1/uppertolerance1.39342894083359071.39342894083359071.23E-1818passed
/mcf_steps/2/segmentexact000.00E+2passed
/mcf_steps/2/agetolerance44.00.00E+2passed
/mcf_steps/2/eventsexact220.00E+2passed
/mcf_steps/2/at_riskexact550.00E+2passed
/mcf_steps/2/incrementtolerance0.400000000000000020.42.20E-1817passed
/mcf_steps/2/mcftolerance0.850000000000000090.85000000000000011.18E-1817passed
/mcf_steps/2/variancetolerance0.374843750000000030.374843750000000031.09E-1817passed
/mcf_steps/2/standard_errortolerance0.612244844812922850.61224484481292294.30E-1918passed
/mcf_steps/2/lowertolerance0.207161298089520410.207161298089520271.36E-1615passed
/mcf_steps/2/uppertolerance3.48762054815753773.487620548157542.27E-1515passed
/mcf_steps/3/segmentexact000.00E+2passed
/mcf_steps/3/agetolerance55.00.00E+2passed
/mcf_steps/3/eventsexact110.00E+2passed
/mcf_steps/3/at_riskexact550.00E+2passed
/mcf_steps/3/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/3/mcftolerance1.051.054.44E-1716passed
/mcf_steps/3/variancetolerance0.354843750000000070.354843755.67E-1715passed
/mcf_steps/3/standard_errortolerance0.595687627872192870.59568762787219281.10E-1615passed
/mcf_steps/3/lowertolerance0.345369768837833440.345369768837833444.83E-1816passed
/mcf_steps/3/uppertolerance3.19223076098960233.19223076098960234.50E-1716passed
/mcf_steps/4/segmentexact000.00E+2passed
/mcf_steps/4/agetolerance66.00.00E+2passed
/mcf_steps/4/eventsexact220.00E+2passed
/mcf_steps/4/at_riskexact550.00E+2passed
/mcf_steps/4/incrementtolerance0.400000000000000020.42.20E-1817passed
/mcf_steps/4/mcftolerance1.45000000000000021.45000000000000022.24E-1716passed
/mcf_steps/4/variancetolerance0.292343750000000010.292343753.32E-1816passed
/mcf_steps/4/standard_errortolerance0.54068821884705420.54068821884705422.78E-1817passed
/mcf_steps/4/lowertolerance0.698175810568838040.69817581056883791.11E-1615passed
/mcf_steps/4/uppertolerance3.01141914138645193.0114191413864539.36E-1615passed
/mcf_steps/5/segmentexact000.00E+2passed
/mcf_steps/5/agetolerance6.56.50.00E+1passed
/mcf_steps/5/eventsexact110.00E+2passed
/mcf_steps/5/at_riskexact550.00E+2passed
/mcf_steps/5/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/5/mcftolerance1.65000000000000011.65000000000000013.32E-1716passed
/mcf_steps/5/variancetolerance0.381093750000000010.381093751.12E-1817passed
/mcf_steps/5/standard_errortolerance0.617327911243287340.61732791124328731.67E-1817passed
/mcf_steps/5/lowertolerance0.792532338597347290.79253233859734712.26E-1615passed
/mcf_steps/5/uppertolerance3.43519105456110513.4351910545611068.61E-1615passed
/mcf_steps/6/segmentexact000.00E+2passed
/mcf_steps/6/agetolerance77.00.00E+2passed
/mcf_steps/6/eventsexact110.00E+2passed
/mcf_steps/6/at_riskexact550.00E+2passed
/mcf_steps/6/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/6/mcftolerance1.85000000000000011.851.12E-1717passed
/mcf_steps/6/variancetolerance0.349843750000000060.349843755.11E-1715passed
/mcf_steps/6/standard_errortolerance0.591475908216049010.5914759082160493.32E-1817passed
/mcf_steps/6/lowertolerance0.988617682200297110.98861768220029692.22E-1615passed
/mcf_steps/6/uppertolerance3.4619044971791143.46190449717911444.15E-1615passed
/mcf_steps/7/segmentexact000.00E+2passed
/mcf_steps/7/agetolerance9.59.50.00E+1passed
/mcf_steps/7/eventsexact110.00E+2passed
/mcf_steps/7/at_riskexact550.00E+2passed
/mcf_steps/7/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/7/mcftolerance2.05000000000000032.05000000000000033.35E-1716passed
/mcf_steps/7/variancetolerance0.523593750000000080.523593751.11E-1615passed
/mcf_steps/7/standard_errortolerance0.723597781920315210.72359778192031521.62E-1817passed
/mcf_steps/7/lowertolerance1.02636397086145961.02636397086145961.53E-1716passed
/mcf_steps/7/uppertolerance4.09455136706787394.0945513670678742.60E-1717passed
/mcf_steps/8/segmentexact000.00E+2passed
/mcf_steps/8/agetolerance10.510.50.00E+1passed
/mcf_steps/8/eventsexact110.00E+2passed
/mcf_steps/8/at_riskexact550.00E+2passed
/mcf_steps/8/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/8/mcftolerance2.25000000000000042.25000000000000044.41E-1716passed
/mcf_steps/8/variancetolerance0.483593750000000040.483593750000000044.41E-1817passed
/mcf_steps/8/standard_errortolerance0.695409052285056230.69540905228505621.63E-1817passed
/mcf_steps/8/lowertolerance1.22772491657407761.22772491657407732.59E-1615passed
/mcf_steps/8/uppertolerance4.12348070130133594.1234807013013379.15E-1615passed
/mcf_steps/9/segmentexact000.00E+2passed
/mcf_steps/9/agetolerance1111.00.00E+2passed
/mcf_steps/9/eventsexact110.00E+2passed
/mcf_steps/9/at_riskexact550.00E+2passed
/mcf_steps/9/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/9/mcftolerance2.45000000000000062.45000000000000062.17E-1717passed
/mcf_steps/9/variancetolerance0.592343750000000060.592343751.13E-1615passed
/mcf_steps/9/standard_errortolerance0.769638713943107230.76963871394310711.08E-1615passed
/mcf_steps/9/lowertolerance1.32364377240133371.32364377240133371.53E-1716passed
/mcf_steps/9/uppertolerance4.53483038650978014.534830386509782.02E-1818passed
/mcf_evaluations/0/agetolerance33.00.00E+2passed
/mcf_evaluations/0/mcftolerance0.450000000000000010.451.10E-1817passed
/mcf_evaluations/0/standard_errortolerance0.259506743650333660.259506743650333662.15E-2019passed
/mcf_evaluations/0/lowertolerance0.145324956347510970.145324956347510942.82E-1715passed
/mcf_evaluations/0/uppertolerance1.39342894083359071.39342894083359071.23E-1818passed
/mcf_evaluations/1/agetolerance66.00.00E+2passed
/mcf_evaluations/1/mcftolerance1.45000000000000021.45000000000000022.24E-1716passed
/mcf_evaluations/1/standard_errortolerance0.54068821884705420.54068821884705422.78E-1817passed
/mcf_evaluations/1/lowertolerance0.698175810568838040.69817581056883791.11E-1615passed
/mcf_evaluations/1/uppertolerance3.01141914138645193.0114191413864539.36E-1615passed
/mcf_evaluations/2/agetolerance1212.00.00E+2passed
/mcf_evaluations/2/mcftolerance2.45000000000000062.45000000000000062.17E-1717passed
/mcf_evaluations/2/standard_errortolerance0.769638713943107230.76963871394310711.08E-1615passed
/mcf_evaluations/2/lowertolerance1.32364377240133371.32364377240133371.53E-1716passed
/mcf_evaluations/2/uppertolerance4.53483038650978014.534830386509782.02E-1818passed
/hpp/ratetolerance0.206896551724137930.206896551724137931.84E-1817passed
/hpp/rate_lowertolerance0.106906467391762360.10690646739176244.53E-1715passed
/hpp/rate_uppertolerance0.361406638761671620.361406638761671731.08E-1615passed
/hpp/eventsexact12120.00E+2passed
/hpp/exposuretolerance5858.00.00E+2passed
/hpp/log_likelihoodtolerance-30.906436329101027-30.9064363291010273.28E-1616passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

177  emit("STAT-SURV-REPAIR-001", "Repairable-system MCF with a constant-rate HPP overlay",
178       REPAIR, "repairableMenuItem", REPAIR_REQ, "SURV_REPAIR_B9",
179       c(REPAIR_TABS, REPAIR_BASE,
180         list(models = list("mcf", "hpp"), evaluation_ages = list(3, 6, 12),
181              confidence_level = 0.95)),
182       c(list("/unit_count" = exact(5, "units"),
183              "/zero_event_unit_count" = exact(1, "u5 has no events"),
184              "/event_count" = exact(nrow(rp_events), "events"),
185              "/distinct_event_age_count" = exact(length(mref95$steps), "distinct event ages"),
186              "/total_exposure" = tol(sum(rp_windows$end - rp_windows$start), "exposure"),
187              "/earliest_entry" = tol(min(rp_windows$start), "earliest entry"),
188              "/latest_end" = tol(max(rp_windows$end), "latest end"),
189              "/event_source_count" = exact(nrow(rp_events), "source events"),
190              "/event_excluded_count" = exact(0, "none excluded"),
191              "/window_source_count" = exact(5, "source windows"),
192              "/window_excluded_count" = exact(0, "none excluded")),
193         mcf_outputs(mref95, c(3, 6, 12), 0.95),
194         hpp_outputs(0.95)),
195       REPAIR_NOTES, "base R MCF and Poisson rate algebra")

Whole script generators/r/survival_repairable.R


Repairable-system MCF with a power-law NHPP fit and forecast

Case STAT-SURV-REPAIR-002 · status: passed

Reference

Evidence type
R
Source
base R Crow-AMSAA profile likelihood and MCF algebra
Citation
Turnbull (1976); Nelson (1988); Crow (1975); Meeker, Hahn and Escobar (1998); Therneau T (2024), A Package for Survival Analysis in R.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_repairable.R · sha256 1f118d2994e4269d
Dataset
reference_data/datasets/SURV_REPAIR_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/unit_countexact550.00E+2passed
/event_countexact12120.00E+2passed
/mcf_steps/0/segmentexact000.00E+2passed
/mcf_steps/0/agetolerance1.51.50.00E+1passed
/mcf_steps/0/eventsexact110.00E+2passed
/mcf_steps/0/at_riskexact440.00E+2passed
/mcf_steps/0/incrementtolerance0.250.250.00E+0passed
/mcf_steps/0/mcftolerance0.250.250.00E+0passed
/mcf_steps/0/variancetolerance0.058593750.058593750.00E-6passed
/mcf_steps/0/standard_errortolerance0.242061459137963560.242061459137963563.84E-1816passed
/mcf_steps/0/lowertolerance0.0508478704127639860.050847870412763953.46E-1715passed
/mcf_steps/0/uppertolerance1.22915668822801011.2291566882280118.41E-1615passed
/mcf_steps/1/segmentexact000.00E+2passed
/mcf_steps/1/agetolerance2.52.50.00E+1passed
/mcf_steps/1/eventsexact110.00E+2passed
/mcf_steps/1/at_riskexact550.00E+2passed
/mcf_steps/1/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/1/mcftolerance0.450000000000000010.451.10E-1817passed
/mcf_steps/1/variancetolerance0.0673437499999999940.067343751.06E-1917passed
/mcf_steps/1/standard_errortolerance0.259506743650333660.259506743650333662.15E-2019passed
/mcf_steps/1/lowertolerance0.174284788184219890.174284788184219771.16E-1615passed
/mcf_steps/1/uppertolerance1.16189141984070641.1618914198407076.96E-1615passed
/mcf_steps/2/segmentexact000.00E+2passed
/mcf_steps/2/agetolerance44.00.00E+2passed
/mcf_steps/2/eventsexact220.00E+2passed
/mcf_steps/2/at_riskexact550.00E+2passed
/mcf_steps/2/incrementtolerance0.400000000000000020.42.20E-1817passed
/mcf_steps/2/mcftolerance0.850000000000000090.85000000000000011.18E-1817passed
/mcf_steps/2/variancetolerance0.374843750000000030.374843750000000031.09E-1817passed
/mcf_steps/2/standard_errortolerance0.612244844812922850.61224484481292294.30E-1918passed
/mcf_steps/2/lowertolerance0.259944444586431340.259944444586431231.06E-1615passed
/mcf_steps/2/uppertolerance2.77944004977482572.7794400497748271.36E-1515passed
/mcf_steps/3/segmentexact000.00E+2passed
/mcf_steps/3/agetolerance55.00.00E+2passed
/mcf_steps/3/eventsexact110.00E+2passed
/mcf_steps/3/at_riskexact550.00E+2passed
/mcf_steps/3/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/3/mcftolerance1.051.054.44E-1716passed
/mcf_steps/3/variancetolerance0.354843750000000070.354843755.67E-1715passed
/mcf_steps/3/standard_errortolerance0.595687627872192870.59568762787219281.10E-1615passed
/mcf_steps/3/lowertolerance0.412973957223783110.412973957223783055.97E-1715passed
/mcf_steps/3/uppertolerance2.66965986768646382.66965986768646424.39E-1615passed
/mcf_steps/4/segmentexact000.00E+2passed
/mcf_steps/4/agetolerance66.00.00E+2passed
/mcf_steps/4/eventsexact220.00E+2passed
/mcf_steps/4/at_riskexact550.00E+2passed
/mcf_steps/4/incrementtolerance0.400000000000000020.42.20E-1817passed
/mcf_steps/4/mcftolerance1.45000000000000021.45000000000000022.24E-1716passed
/mcf_steps/4/variancetolerance0.292343750000000010.292343753.32E-1816passed
/mcf_steps/4/standard_errortolerance0.54068821884705420.54068821884705422.78E-1817passed
/mcf_steps/4/lowertolerance0.785226298167816080.7852262981678161.07E-1615passed
/mcf_steps/4/uppertolerance2.67757206413718542.67757206413718638.92E-1615passed
/mcf_steps/5/segmentexact000.00E+2passed
/mcf_steps/5/agetolerance6.56.50.00E+1passed
/mcf_steps/5/eventsexact110.00E+2passed
/mcf_steps/5/at_riskexact550.00E+2passed
/mcf_steps/5/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/5/mcftolerance1.65000000000000011.65000000000000013.32E-1716passed
/mcf_steps/5/variancetolerance0.381093750000000010.381093751.12E-1817passed
/mcf_steps/5/standard_errortolerance0.617327911243287340.61732791124328731.67E-1817passed
/mcf_steps/5/lowertolerance0.891698534839130750.89169853483913052.20E-1615passed
/mcf_steps/5/uppertolerance3.05316190801094053.0531619080109413.95E-1615passed
/mcf_steps/6/segmentexact000.00E+2passed
/mcf_steps/6/agetolerance77.00.00E+2passed
/mcf_steps/6/eventsexact110.00E+2passed
/mcf_steps/6/at_riskexact550.00E+2passed
/mcf_steps/6/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/6/mcftolerance1.85000000000000011.851.12E-1717passed
/mcf_steps/6/variancetolerance0.349843750000000060.349843755.11E-1715passed
/mcf_steps/6/standard_errortolerance0.591475908216049010.5914759082160493.32E-1817passed
/mcf_steps/6/lowertolerance1.09340693903253631.0934069390325362.48E-1615passed
/mcf_steps/6/uppertolerance3.13012463870796553.1301246387079664.78E-1615passed
/mcf_steps/7/segmentexact000.00E+2passed
/mcf_steps/7/agetolerance9.59.50.00E+1passed
/mcf_steps/7/eventsexact110.00E+2passed
/mcf_steps/7/at_riskexact550.00E+2passed
/mcf_steps/7/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/7/mcftolerance2.05000000000000032.05000000000000033.35E-1716passed
/mcf_steps/7/variancetolerance0.523593750000000080.523593751.11E-1615passed
/mcf_steps/7/standard_errortolerance0.723597781920315210.72359778192031521.62E-1817passed
/mcf_steps/7/lowertolerance1.14711301120217631.14711301120217611.75E-1615passed
/mcf_steps/7/uppertolerance3.66354488089693443.66354488089693539.07E-1615passed
/mcf_steps/8/segmentexact000.00E+2passed
/mcf_steps/8/agetolerance10.510.50.00E+1passed
/mcf_steps/8/eventsexact110.00E+2passed
/mcf_steps/8/at_riskexact550.00E+2passed
/mcf_steps/8/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/8/mcftolerance2.25000000000000042.25000000000000044.41E-1716passed
/mcf_steps/8/variancetolerance0.483593750000000040.483593750000000044.41E-1817passed
/mcf_steps/8/standard_errortolerance0.695409052285056230.69540905228505621.63E-1817passed
/mcf_steps/8/lowertolerance1.35331098456719131.35331098456719111.78E-1615passed
/mcf_steps/8/uppertolerance3.7408253222884043.74082532228840541.36E-1515passed
/mcf_steps/9/segmentexact000.00E+2passed
/mcf_steps/9/agetolerance1111.00.00E+2passed
/mcf_steps/9/eventsexact110.00E+2passed
/mcf_steps/9/at_riskexact550.00E+2passed
/mcf_steps/9/incrementtolerance0.200000000000000010.21.10E-1817passed
/mcf_steps/9/mcftolerance2.45000000000000062.45000000000000062.17E-1717passed
/mcf_steps/9/variancetolerance0.592343750000000060.592343751.13E-1615passed
/mcf_steps/9/standard_errortolerance0.769638713943107230.76963871394310711.08E-1615passed
/mcf_steps/9/lowertolerance1.4613732615000051.46137326150000482.06E-1615passed
/mcf_steps/9/uppertolerance4.10743795451603164.1074379545160328.41E-1615passed
/mcf_evaluations/0/agetolerance44.00.00E+2passed
/mcf_evaluations/0/mcftolerance0.850000000000000090.85000000000000011.18E-1817passed
/mcf_evaluations/0/standard_errortolerance0.612244844812922850.61224484481292294.30E-1918passed
/mcf_evaluations/0/lowertolerance0.259944444586431340.259944444586431231.06E-1615passed
/mcf_evaluations/0/uppertolerance2.77944004977482572.7794400497748271.36E-1515passed
/mcf_evaluations/1/agetolerance88.00.00E+2passed
/mcf_evaluations/1/mcftolerance1.85000000000000011.851.12E-1717passed
/mcf_evaluations/1/standard_errortolerance0.591475908216049010.5914759082160493.32E-1817passed
/mcf_evaluations/1/lowertolerance1.09340693903253631.0934069390325362.48E-1615passed
/mcf_evaluations/1/uppertolerance3.13012463870796553.1301246387079664.78E-1615passed
/power_law/log_scaletolerance-2.0074001125122911-2.0074001122286332.84E-109passed
/power_law/shapetolerance1.17016436353475741.17016436342264171.12E-1010passed
/power_law/log_likelihoodtolerance-30.781818147938132-30.781818147938147.42E-1515passed
/power_law/log_scale_standard_errortolerance0.95252082631408330.95252079602850483.03E-87passed
/power_law/log_shape_standard_errortolerance0.306604517481343540.306604508871016378.61E-97passed
/power_law/covariance_log_scale_log_shape/0/0tolerance0.907295924562063980.90729586686677645.77E-87passed
/power_law/covariance_log_scale_log_shape/0/1tolerance-0.27831223393582605-0.27831221604493781.79E-87passed
/power_law/covariance_log_scale_log_shape/1/1tolerance0.0940063301399675090.094006324860037155.28E-97passed
/power_law/shape_profile_lowertolerance0.675003173903701010.67500317390349172.09E-1312passed
/power_law/shape_profile_uppertolerance1.86023210957707021.86023210957706974.52E-1615passed
/power_law/lrt_statistictolerance0.249236362325788720.249236362325778061.07E-1413passed
/power_law/lrt_p_valuetail_probability0.617613292153301960.61761329215330937.33E-1513passed
/power_law/expected_counts/0/estimatetolerance0.680309216678741560.68030921676597938.72E-119passed
/power_law/expected_counts/0/lowertolerance0.298059273098018720.29805927968796066.59E-97passed
/power_law/expected_counts/0/uppertolerance1.55278051069339231.55278047676042473.39E-87passed
/power_law/expected_counts/1/estimatetolerance0.413047987308021780.413047987311285173.26E-1211passed
/power_law/expected_counts/1/lowertolerance0.256445960450911390.25644595930673881.14E-98passed
/power_law/expected_counts/1/uppertolerance0.665281057729374740.66528106070813962.98E-98passed
/power_law/expected_counts/2/estimatetolerance1.53094972116226251.53094972123960627.73E-1110passed
/power_law/expected_counts/2/lowertolerance0.888402195268732990.88840219931198014.04E-98passed
/power_law/expected_counts/2/uppertolerance2.63822743933881252.63822742759842431.17E-88passed
/power_law/expected_counts/3/estimatetolerance0.456799407241855530.456799407215204572.67E-1110passed
/power_law/expected_counts/3/lowertolerance0.259500005945535520.25950000426812381.68E-98passed
/power_law/expected_counts/3/uppertolerance0.80410671936672640.80410672447065595.10E-98passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

299  emit("STAT-SURV-REPAIR-002", "Repairable-system MCF with a power-law NHPP fit and forecast",
300       REPAIR, "repairableMenuItem", REPAIR_REQ, "SURV_REPAIR_B9",
301       c(REPAIR_TABS, REPAIR_BASE,
302         list(models = list("mcf", "power_law"), evaluation_ages = list(4, 8),
303              forecast_horizon = 2, confidence_level = 0.90)),
304       c(list("/unit_count" = exact(5, "units"),
305              "/event_count" = exact(nrow(rp_events), "events")),
306         mcf_outputs(mref90, c(4, 8), 0.90),
307         power_outputs),
308       c(REPAIR_NOTES,
309         paste0(MANUAL, ":34-41: Lambda(t) = e^a t^b fitted by profile maximum likelihood (a = log(N / S(b)), S(b) = sum (end^b - entry^b)); profile-likelihood interval for b with cutoff chi2_1(level)/2; likelihood-ratio test of b = 1."),
310         "Iterative searches (optimize, uniroot to 1e-12) and numeric second derivatives with step 1e-4: outputs allow 1e-7 absolute or 1e-6 relative."),
311       "base R Crow-AMSAA profile likelihood and MCF algebra")

Whole script generators/r/survival_repairable.R


Survival STAT-SURV-RISK-001

Case STAT-SURV-RISK-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/survdiff/coxph and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, MASS 7.3.65, jsonlite 2.0.0
Reference script
generators/r/survival_core.R · sha256 1ed8945e6864eb84
Dataset
reference_data/datasets/HARVEST_GEHAN_MP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0exact21210.00E+2passed
/1exact17170.00E+2passed
/2exact15150.00E+2passed
/3exact660.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

38  ticks <- c(6,7,10,23)
39  risk <- vapply(ticks,function(t) sum(mp$time>=t),0L)
40  emit("STAT-SURV-RISK-001","tabelz.analysis.survival_nonparametric:number_at_risk",
41       "survivalMenuItem","HARVEST_GEHAN_MP",c(config,list(times=as.list(ticks))),
42       list("/0"=exact(risk[1],"At risk at time 6"),
43            "/1"=exact(risk[2],"At risk at time 7"),
44            "/2"=exact(risk[3],"At risk at time 10"),
45            "/3"=exact(risk[4],"At risk at time 23")),
46       c("number_at_risk is an in-process payload step after an engine survival fit; ties remain at risk at t."))

Whole script generators/r/survival_core.R


Survival boundary STAT-SURV-RISK-ALL-CENS-001

Case STAT-SURV-RISK-ALL-CENS-001 · status: passed

Reference

Evidence type
R
Source
R survival and analytic boundary
Citation
R survival::survfit/coxph/survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_boundaries.R · sha256 528007972452fb15
Dataset
reference_data/datasets/SURV_ALL_CENS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0exact550.00E+2passed
/1exact110.00E+2passed
/2exact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

60  emit("STAT-SURV-RISK-ALL-CENS-001","tabelz.analysis.survival_nonparametric:number_at_risk",
61       "survivalMenuItem","SURV_ALL_CENS",c(cfg,list(times=list(1,5,6))),
62       list("/0"=exact(5L,"Five at risk at first time"),
63            "/1"=exact(1L,"One censored row still observed at last time"),
64            "/2"=exact(0L,"None beyond last time")),
65       c("Number at risk counts censoring rows through their recorded time."))

Whole script generators/r/survival_boundaries.R


Hard-numeric survival STAT-SURV-RISK-HARD-001

Case STAT-SURV-RISK-HARD-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_hard.R · sha256 bc20ccd17540f4e7
Dataset
reference_data/datasets/SURV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0exact40400.00E+2passed
/1exact31310.00E+2passed
/2exact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

55  ticks <- c(1e6+137,1e6+1370,1e6+5500)
56  risk <- vapply(ticks,function(t) sum(d$time>=t),0L)
57  emit("STAT-SURV-RISK-HARD-001","tabelz.analysis.survival_nonparametric:number_at_risk",
58       "survivalMenuItem","SURV_HARD",c(cfg,list(times=as.list(ticks))),
59       list("/0"=exact(risk[1],"Risk at first large time"),
60            "/1"=exact(risk[2],"Risk at interior large time"),
61            "/2"=exact(risk[3],"Risk near largest time")),
62       c("At-risk comparisons use exact stored double times, not rounded display labels."))

Whole script generators/r/survival_hard.R


Invalid survival request STAT-SURV-RISK-INVALID-001

Case STAT-SURV-RISK-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/HARVEST_GEHAN_MP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid survival request STAT-SURV-RISK-INVALID-002

Case STAT-SURV-RISK-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/HARVEST_GEHAN_MP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Small-sample survival boundary STAT-SURV-RISK-MIN-001

Case STAT-SURV-RISK-MIN-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::survfit/coxph and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_minimum.R · sha256 2997d37929a6dae8
Dataset
reference_data/datasets/SURV_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0exact110.00E+2passed
/1exact110.00E+2passed
/2exact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

36  emit("STAT-SURV-RISK-MIN-001","tabelz.analysis.survival_nonparametric:number_at_risk",
37       "survivalMenuItem","SURV_MIN",1:4,c(cfg,list(times=list(0,1,2))),
38       list("/0"=exact(1L,"One row at risk before time 1"),
39            "/1"=exact(1L,"Event row is still at risk at time 1"),
40            "/2"=exact(0L,"No observation at or after time 2")),
41       c("At-risk count includes a failure at the requested instant."))

Whole script generators/r/survival_minimum.R


Survival population STAT-SURV-RISK-POP-001

Case STAT-SURV-RISK-POP-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::Surv/survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_population.R · sha256 3b3b4629e272fcd9
Dataset
reference_data/datasets/SURV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0exact29290.00E+2passed
/1exact23230.00E+2passed
/2exact10100.00E+2passed
/3exact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

56  ticks <- c(7,10,15,20)
57  risk <- vapply(ticks,function(t) sum(kmdata$time>=t),0L)
58  emit("STAT-SURV-RISK-POP-001","tabelz.analysis.survival_nonparametric:number_at_risk",
59       "survivalMenuItem","SURV_POP",7L,c(cfg,list(times=as.list(ticks))),
60       list("/0"=exact(risk[1],"Risk at 7"),"/1"=exact(risk[2],"Risk at 10"),
61            "/2"=exact(risk[3],"Risk at 15"),"/3"=exact(risk[4],"Risk at 20")),
62       c("Number at risk counts rows still observed at or beyond each query time."))

Whole script generators/r/survival_population.R


Survival STAT-SURV-STABILITY-001

Case STAT-SURV-STABILITY-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/survdiff/coxph and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, MASS 7.3.65, jsonlite 2.0.0
Reference script
generators/r/survival_core.R · sha256 1ed8945e6864eb84
Dataset
reference_data/datasets/SURV_STABILITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/estimatetolerance100.00660035196987100.006600351969843.30E-1415passed
/coefficients/1/estimatetolerance-0.37077323672499457-0.370773236724998073.50E-1514passed
/nexact30300.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

75  j <- seq_len(30L)
76  stability <- data.frame(time=as.double(rep(0:9,3L)),
77                          response=100-0.37*rep(0:9,3L)+0.08*sin(j*1.31),
78                          batch=rep(c("A","B","C"),each=10L))
79  export_dataset(stability,"SURV_STABILITY","Constructed linear shelf-life measurements",
80                 "R stats::lm pooled regression",
81                 "Constructed validation data; no third-party license.",
82                 c(time="Float64",response="Float64",batch="String"))
83  s <- read_dataset("SURV_STABILITY")
84  lmfit <- stats::lm(response~time,data=s)
85  emit("STAT-SURV-STABILITY-001","tabelz.analysis.product_stability:compute_product_stability",
86       "productStabilityMenuItem","SURV_STABILITY",
87       list(time_column="time",response_column="response",model_type="pooled",lsl=95),
88       list("/coefficients/0/estimate"=tol(coef(lmfit)[[1]],"Pooled OLS intercept"),
89            "/coefficients/1/estimate"=tol(coef(lmfit)[[2]],"Pooled OLS time slope"),
90            "/n"=exact(nrow(s),"All complete shelf-life measurements")),
91       c("Pooled OLS uses all batches without batch terms; lower spec limit is 95 response units."))

Whole script generators/r/survival_core.R


Hard-numeric survival STAT-SURV-STABILITY-HARD-001

Case STAT-SURV-STABILITY-HARD-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_hard.R · sha256 bc20ccd17540f4e7
Dataset
reference_data/datasets/SURV_STABILITY_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/1/estimatetolerance-0.030000796960761165-0.0300007969606335131.28E-1311passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

91  fit_stable <- stats::lm(response~time,data=s)
92  emit("STAT-SURV-STABILITY-HARD-001","tabelz.analysis.product_stability:compute_product_stability",
93       "productStabilityMenuItem","SURV_STABILITY_HARD",
94       list(time_column="time",response_column="response",lsl=95),
95       list("/coefficients/1/estimate"=tol(coef(fit_stable)[[2]],
96                                            "Large-offset time OLS slope","1e-9")),
97       c("Time is near 1e6 with unit increments; QR regression resolves a small slope."))

Whole script generators/r/survival_hard.R


Survival option STAT-SURV-STABILITY-INTERCEPTS-001

Case STAT-SURV-STABILITY-INTERCEPTS-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_options.R · sha256 95af0dbeed660def
Dataset
reference_data/datasets/SURV_STABILITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/3/estimatetolerance-0.37077323672499568-0.37077323672497232.34E-1413passed
/nexact30300.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

69  s <- read_dataset("SURV_STABILITY")
70  intercepts <- stats::lm(response~0+batch+time,data=s)
71  emit("STAT-SURV-STABILITY-INTERCEPTS-001",
72       "tabelz.analysis.product_stability:compute_product_stability",
73       "productStabilityMenuItem","SURV_STABILITY",
74       list(time_column="time",response_column="response",batch_column="batch",
75            model_type="separate_intercepts_pooled_slope",confidence_level=0.9,lsl=95),
76       list("/coefficients/3/estimate"=tol(coef(intercepts)[[4]],"Pooled time slope","1e-9"),
77            "/n"=exact(nrow(s),"All stability measurements")),
78       c("R 0+batch+time has separate batch intercepts and one pooled slope."))

Whole script generators/r/survival_options.R


Invalid survival request STAT-SURV-STABILITY-INVALID-001

Case STAT-SURV-STABILITY-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/SURV_STABILITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid survival request STAT-SURV-STABILITY-INVALID-002

Case STAT-SURV-STABILITY-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/SURV_STABILITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Small-sample survival boundary STAT-SURV-STABILITY-MIN-001

Case STAT-SURV-STABILITY-MIN-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::survfit/coxph and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_minimum.R · sha256 2997d37929a6dae8
Dataset
reference_data/datasets/SURV_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/nexact330.00E+2passed
/coefficients/1/estimatetolerance-0.50500000000000511-0.50500000000000262.55E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

71  emit("STAT-SURV-STABILITY-MIN-001","tabelz.analysis.product_stability:compute_product_stability",
72       "productStabilityMenuItem","SURV_MIN",3:4,
73       list(time_column="time",response_column="response",lsl=95),
74       list("/n"=exact(3L,"Three rows leave one residual degree of freedom"),
75            "/coefficients/1/estimate"=tol(coef(stats::lm(response~time,data=d[1:3,]))[[2]],
76                                               "Three-row OLS slope","1e-9")),
77       c("Three distinct times are the smallest two-parameter OLS sample with residual df."))

Whole script generators/r/survival_minimum.R


Survival boundary STAT-SURV-STABILITY-OUTSIDE-001

Case STAT-SURV-STABILITY-OUTSIDE-001 · status: passed

Reference

Evidence type
R
Source
R survival and analytic boundary
Citation
R survival::survfit/coxph/survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_boundaries.R · sha256 528007972452fb15
Dataset
reference_data/datasets/SURV_STABILITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/overall_shelf_life_statusexactout_of_specification_at_zeroout_of_specification_at_zeropassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

113  emit("STAT-SURV-STABILITY-OUTSIDE-001",
114       "tabelz.analysis.product_stability:compute_product_stability",
115       "productStabilityMenuItem","SURV_STABILITY",
116       list(time_column="time",response_column="response",lsl=1000),
117       list("/overall_shelf_life_status"=exact("out_of_specification_at_zero",
118         "One-sided lower mean bound is already below LSL=1000 at time zero")),
119       c("R pooled lm intercept is near 100, so its one-sided lower 95% bound is certainly below 1000."))

Whole script generators/r/survival_boundaries.R


Survival population STAT-SURV-STABILITY-POP-001

Case STAT-SURV-STABILITY-POP-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::Surv/survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_population.R · sha256 3b3b4629e272fcd9
Dataset
reference_data/datasets/SURV_STABILITY_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact30300.00E+2passed
/excluded_countexact110.00E+2passed
/missing_countexact220.00E+2passed
/coefficients/1/estimatetolerance-0.37251041269502028-0.37251041269501966.70E-1614passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

106  stability_eligible <- s[-10L,,drop=FALSE]
107  stability_complete <- stability_eligible[stats::complete.cases(stability_eligible[c("time","response")]),,drop=FALSE]
108  lmfit <- stats::lm(response~time,data=stability_complete)
109  emit("STAT-SURV-STABILITY-POP-001","tabelz.analysis.product_stability:compute_product_stability",
110       "productStabilityMenuItem","SURV_STABILITY_POP",9L,
111       list(time_column="time",response_column="response",lsl=95),
112       list("/source_count"=exact(nrow(s),"CSV population"),
113            "/excluded_count"=exact(1L,"One source row excluded"),
114            "/missing_count"=exact(2L,"Missing response and time"),
115            "/coefficients/1/estimate"=tol(coef(lmfit)[[2]],"Complete-case pooled slope")),
116       c("Pooled OLS excludes missing time/response and declared source exclusion."))

Whole script generators/r/survival_population.R


Survival option STAT-SURV-STABILITY-SLOPES-001

Case STAT-SURV-STABILITY-SLOPES-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_options.R · sha256 95af0dbeed660def
Dataset
reference_data/datasets/SURV_STABILITY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/3/estimatetolerance-0.37447507711523798-0.37447507711520393.41E-1413passed
/nexact30300.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

80  slopes <- stats::lm(response~0+batch+batch:time,data=s)
81  emit("STAT-SURV-STABILITY-SLOPES-001",
82       "tabelz.analysis.product_stability:compute_product_stability",
83       "productStabilityMenuItem","SURV_STABILITY",
84       list(time_column="time",response_column="response",batch_column="batch",
85            model_type="separate_intercepts_separate_slopes",lsl=95),
86       list("/coefficients/3/estimate"=tol(coef(slopes)[[4]],"Batch A time slope","1e-9"),
87            "/n"=exact(nrow(s),"All stability measurements")),
88       c("R 0+batch+batch:time has separate intercepts and slopes per batch."))

Whole script generators/r/survival_options.R


Weibull warranty prediction from a calendar triangle with future shipments and a limit

Case STAT-SURV-WARRPRED-001 · status: passed

Reference

Evidence type
R
Source
R survival package survreg on interval2 records with frequency weights; independent base-R implementations of the warranty triangle conversion and the expected-count algebra
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_warranty.R · sha256 4e304617e4b0f92b
Dataset
reference_data/datasets/SURV_WARR_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/log_likelihoodtolerance-336.06856173438467-336.06856173438465.39E-1415passed
/field_expectedtolerance160.70966849733594160.709666975610621.52E-68passed
/future_expectedtolerance149.12458269274785149.124581798438168.94E-78passed
/expectedtolerance309.83425119008382309.83424877404882.42E-68passed
/eligible_unitstolerance15471547.00.00E+2passed
/lowertolerance217.2083633699134217.20836764166554.27E-67passed
/uppertolerance429.21036995102338429.21035680771271.31E-57passed
/cohorts/0/survivorstolerance107107.00.00E+2passed
/cohorts/0/agetolerance66.00.00E+2passed
/cohorts/0/eligibleexactTrueTruepassed
/cohorts/0/expectedtolerance22.08715415145826122.0871539181316662.33E-77passed
/cohorts/1/survivorstolerance120120.00.00E+2passed
/cohorts/1/agetolerance55.00.00E+2passed
/cohorts/1/eligibleexactTrueTruepassed
/cohorts/1/expectedtolerance25.10094672057580625.100946464047162.57E-77passed
/cohorts/2/survivorstolerance135135.00.00E+2passed
/cohorts/2/agetolerance44.00.00E+2passed
/cohorts/2/eligibleexactTrueTruepassed
/cohorts/2/expectedtolerance28.66277722926251328.662776947466262.82E-78passed
/cohorts/3/survivorstolerance144144.00.00E+2passed
/cohorts/3/agetolerance33.00.00E+2passed
/cohorts/3/eligibleexactTrueTruepassed
/cohorts/3/expectedtolerance31.10284887944267531.1028485876308662.92E-78passed
/cohorts/4/survivorstolerance134134.00.00E+2passed
/cohorts/4/agetolerance22.00.00E+2passed
/cohorts/4/eligibleexactTrueTruepassed
/cohorts/4/expectedtolerance29.54140961211292929.5414093509153842.61E-78passed
/cohorts/5/survivorstolerance107107.00.00E+2passed
/cohorts/5/agetolerance11.00.00E+2passed
/cohorts/5/eligibleexactTrueTruepassed
/cohorts/5/expectedtolerance24.21453190448376724.214531707419261.97E-78passed
/future/0/unitstolerance500500.00.00E+2passed
/future/0/timetolerance1.51.50.00E+1passed
/future/0/expectedtolerance107.27425698755201107.274256296299756.91E-78passed
/future/1/unitstolerance300300.00.00E+2passed
/future/1/timetolerance66.00.00E+2passed
/future/1/expectedtolerance41.85032570519583641.8503255021383962.03E-78passed
/periods/0/horizontolerance11.00.00E+2passed
/periods/0/expectedtolerance17.01604323691496417.016043110377271.27E-78passed
/periods/0/incrementtolerance17.01604323691496417.016043110377271.27E-78passed
/periods/0/lowertolerance12.27533477386265512.2753349372717141.63E-77passed
/periods/0/uppertolerance23.52897176996170423.5289711147235186.55E-77passed
/periods/1/horizontolerance22.00.00E+2passed
/periods/1/expectedtolerance42.01487979400724342.0148795698166352.24E-78passed
/periods/1/incrementtolerance24.99883655709227924.9988364594393649.77E-88passed
/periods/1/lowertolerance32.62781352638503332.627813642157151.16E-78passed
/periods/1/uppertolerance53.98256464376776653.982563885213467.59E-77passed
/periods/2/horizontolerance33.00.00E+2passed
/periods/2/expectedtolerance70.74261556181259970.74261516515853.97E-78passed
/periods/2/incrementtolerance28.72773576780535628.7277355953418621.72E-78passed
/periods/2/lowertolerance54.53855368434999254.538553953253582.69E-78passed
/periods/2/uppertolerance91.39211473533433191.392113291769751.44E-67passed
/periods/3/horizontolerance44.00.00E+2passed
/periods/3/expectedtolerance97.02558242146889797.025581826463885.95E-78passed
/periods/3/incrementtolerance26.28296685965629826.282966661305381.98E-78passed
/periods/3/lowertolerance73.53478492084693173.534785499913315.79E-78passed
/periods/3/uppertolerance127.20707633887618127.207073863736452.48E-67passed
/periods/4/horizontolerance55.00.00E+2passed
/periods/4/expectedtolerance121.72310443282765121.723103631174648.02E-78passed
/periods/4/incrementtolerance24.69752201135875724.697521804710762.07E-78passed
/periods/4/lowertolerance90.67244801641314290.67244897949449.63E-77passed
/periods/4/uppertolerance161.91806401061035161.918060340192973.67E-67passed
/periods/5/horizontolerance66.00.00E+2passed
/periods/5/expectedtolerance145.18513883164255145.185137820963321.01E-68passed
/periods/5/incrementtolerance23.46203439881489323.462034189788682.09E-78passed
/periods/5/lowertolerance106.42948055672369106.429481943508121.39E-67passed
/periods/5/uppertolerance195.6328112284948195.63280627973424.95E-67passed
/periods/6/horizontolerance77.00.00E+2passed
/periods/6/expectedtolerance177.31589739311673177.315896186705741.21E-68passed
/periods/6/incrementtolerance32.13075856147418132.130758365742421.96E-78passed
/periods/6/lowertolerance130.88273750697246130.882739126126751.62E-67passed
/periods/6/uppertolerance237.45984745800803237.45984165212755.81E-67passed
/periods/7/horizontolerance88.00.00E+2passed
/periods/7/expectedtolerance206.38349851442604206.3834970687721.45E-68passed
/periods/7/incrementtolerance29.06760112130930929.0676008820662452.39E-78passed
/periods/7/lowertolerance150.90568933890032150.90569142023382.08E-67passed
/periods/7/uppertolerance278.19264827475939278.192641099119277.18E-67passed
/periods/8/horizontolerance99.00.00E+2passed
/periods/8/expectedtolerance233.90119248906103233.90119079845631.69E-68passed
/periods/8/incrementtolerance27.51769397463499427.517693729684312.45E-78passed
/periods/8/lowertolerance169.16469953534235169.164702136297142.60E-67passed
/periods/8/uppertolerance317.69900623431431317.69899757670378.66E-67passed
/periods/9/horizontolerance1010.00.00E+2passed
/periods/9/expectedtolerance260.2047251931865260.2047232577161.94E-68passed
/periods/9/incrementtolerance26.30353270412547326.3035324592597142.45E-78passed
/periods/9/lowertolerance186.14695874110816186.146961886754043.15E-67passed
/periods/9/uppertolerance356.01797951030528356.01796934309391.02E-57passed
/periods/10/horizontolerance1111.00.00E+2passed
/periods/10/expectedtolerance285.47549821117252285.47549603343782.18E-68passed
/periods/10/incrementtolerance25.27077301798601625.270772775721812.42E-78passed
/periods/10/lowertolerance202.10667683287139202.106680535637963.70E-67passed
/periods/10/uppertolerance393.18008089268551393.18006922614341.17E-57passed
/periods/11/horizontolerance1212.00.00E+2passed
/periods/11/expectedtolerance309.83425119008382309.83424877404882.42E-68passed
/periods/11/incrementtolerance24.35875297891129824.3587527406109532.38E-78passed
/periods/11/lowertolerance217.2083633699134217.20836764166554.27E-67passed
/periods/11/uppertolerance429.21036995102338429.21035680771271.31E-57passed
/parameters/0/estimatetolerance0.848085165893830320.84808516222519543.67E-98passed
/parameters/0/standard_errortolerance0.108501928413844650.108501918695707459.72E-97passed
/parameters/0/lowertolerance0.659992251527137250.65999226277906421.13E-87passed
/parameters/0/uppertolerance1.08978317085528991.08978314284770342.80E-87passed
/parameters/1/estimatetolerance56.0809706860391856.0809715744482068.88E-77passed
/parameters/1/standard_errortolerance21.00792345080549221.007922078182891.37E-67passed
/parameters/1/lowertolerance26.91259327075908926.9125953011523472.03E-67passed
/parameters/1/uppertolerance116.86258702187403116.86258190786245.11E-67passed
/parameters/2/estimatetolerance00.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

435  triangle_records <- data.frame(
436    start = vapply(conv_calendar$records, function(x) x$start, 0),
437    end = vapply(conv_calendar$records, function(x) if (is.na(x$end)) NA_real_ else x$end, 0),
438    frequency = vapply(conv_calendar$records, function(x) x$frequency, 0L)
439  )
440  triangle_survivors <- vapply(conv_calendar$cohorts, function(x) x$survivors, 0)
441  triangle_ages <- vapply(conv_calendar$cohorts, function(x) x$age, 0)
442  future_units_1 <- c(500, 300)
443  future_times_1 <- c(1.5, 6)
444  pred_weibull <- warranty_prediction_reference(
445    triangle_records, "weibull", triangle_survivors, triangle_ages,
446    future_units_1, future_times_1, 12, 24, 0.95, 1
447  )
448  emit("STAT-SURV-WARRPRED-001",
449       "Weibull warranty prediction from a calendar triangle with future shipments and a limit",
450       PRED, "warrantyPredictionMenuItem", PRED_REQ, "SURV_WARR_B9",
451       list(source = "triangle", shipped = "shipped", returns = as.list(RCOLS),
452            return_axis = "calendar", period_width = 1, period_unit = "month",
453            last_period = "incomplete", family = "weibull", horizon = 12,
454            warranty_limit = 24, future_shipments = list("500@1.5", "300@6"),
455            unit_cost = 3.2, confidence_level = 0.95),
456       prediction_outputs(pred_weibull, "weibull", triangle_survivors, triangle_ages,
457                          future_units_1, future_times_1, 0.95),
458       c(PREDICTION_NOTES,
459         "The exposure cohorts are the converted shipment rows with survivors; the cap min(a_j+h, L) binds through the 24-month warranty duration and the future shipments contribute d F(min(h-s, L))."),
460       "nominal")

Whole script generators/r/survival_warranty.R


Lognormal warranty prediction from a censoring table without a warranty limit

Case STAT-SURV-WARRPRED-002 · status: passed

Reference

Evidence type
R
Source
R survival package survreg on interval2 records with frequency weights; independent base-R implementations of the warranty triangle conversion and the expected-count algebra
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_warranty.R · sha256 4e304617e4b0f92b
Dataset
reference_data/datasets/SURV_CENSOR_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/log_likelihoodtolerance-118.13684011875651-118.136840118756482.86E-1415passed
/field_expectedtolerance19.05499301818587919.0549929345785438.36E-88passed
/future_expectedtolerance1.12344188955647991.12344189730085557.74E-98passed
/expectedtolerance20.17843490774236120.17843483187947.59E-88passed
/eligible_unitstolerance540540.00.00E+2passed
/lowertolerance14.8678336638295114.8678336905292672.67E-88passed
/uppertolerance27.2873482323012227.2873479820839652.50E-78passed
/cohorts/0/survivorstolerance4040.00.00E+2passed
/cohorts/0/agetolerance1212.00.00E+2passed
/cohorts/0/eligibleexactTrueTruepassed
/cohorts/0/expectedtolerance4.17479520872492944.17479519404247951.47E-88passed
/cohorts/1/survivorstolerance3030.00.00E+2passed
/cohorts/1/agetolerance1818.00.00E+2passed
/cohorts/1/eligibleexactTrueTruepassed
/cohorts/1/expectedtolerance4.03161786433389854.0316178469295191.74E-88passed
/cohorts/2/survivorstolerance4545.00.00E+2passed
/cohorts/2/agetolerance2424.00.00E+2passed
/cohorts/2/eligibleexactTrueTruepassed
/cohorts/2/expectedtolerance6.82285476813729646.8228547362058433.19E-88passed
/cohorts/3/survivorstolerance2525.00.00E+2passed
/cohorts/3/agetolerance3030.00.00E+2passed
/cohorts/3/eligibleexactTrueTruepassed
/cohorts/3/expectedtolerance4.02572517698975574.0257251574007021.96E-88passed
/future/0/unitstolerance400400.00.00E+2passed
/future/0/timetolerance2.52.50.00E+1passed
/future/0/expectedtolerance1.12344188955647991.12344189730085557.74E-98passed
/periods/0/horizontolerance11.00.00E+2passed
/periods/0/expectedtolerance2.28014616781870982.2801461578528629.97E-98passed
/periods/0/incrementtolerance2.28014616781870982.2801461578528629.97E-98passed
/periods/0/lowertolerance1.60570001394953431.605700017908633.96E-98passed
/periods/0/uppertolerance3.2312672957576173.2312672598332533.59E-87passed
/periods/1/horizontolerance22.00.00E+2passed
/periods/1/expectedtolerance4.60681509380546754.606815073609042.02E-88passed
/periods/1/incrementtolerance2.32666892598675772.32666891575617771.02E-88passed
/periods/1/lowertolerance3.23892271060303213.2389227181465347.54E-98passed
/periods/1/uppertolerance6.52484666594457796.5248465947134127.12E-87passed
/periods/2/horizontolerance33.00.00E+2passed
/periods/2/expectedtolerance6.97055614893315496.9705561183067293.06E-88passed
/periods/2/incrementtolerance2.36374105512768742.3637410446976891.04E-88passed
/periods/2/lowertolerance4.90382197131104654.90382198417810861.29E-88passed
/periods/2/uppertolerance9.89222058573613529.8922204735719231.12E-77passed
/periods/3/horizontolerance44.00.00E+2passed
/periods/3/expectedtolerance9.36399876225091539.3639987211042384.11E-88passed
/periods/3/incrementtolerance2.39344261331776042.3934426027975091.05E-88passed
/periods/3/lowertolerance6.58623484597374196.5862348648981181.89E-88passed
/periods/3/uppertolerance13.28411703647512413.2841168829029091.54E-77passed
/periods/4/horizontolerance55.00.00E+2passed
/periods/4/expectedtolerance11.80285775845149811.8028577071560315.13E-88passed
/periods/4/incrementtolerance2.43885899620058312.4388589860517931.01E-88passed
/periods/4/lowertolerance8.33163502009662878.331635042862752.28E-88passed
/periods/4/uppertolerance16.67494325216305716.6749430637115561.88E-77passed
/periods/5/horizontolerance66.00.00E+2passed
/periods/5/expectedtolerance14.35316543227750914.3531653718177676.05E-88passed
/periods/5/incrementtolerance2.55030767382601022.55030766466173559.16E-98passed
/periods/5/lowertolerance10.23808687682345910.2380869056959532.89E-88passed
/periods/5/uppertolerance20.05962181952445920.0596215968435772.23E-77passed
/periods/6/horizontolerance77.00.00E+2passed
/periods/6/expectedtolerance17.11269654121729617.112696472705446.85E-88passed
/periods/6/incrementtolerance2.75953110893978782.7595311008876748.05E-98passed
/periods/6/lowertolerance12.39993322704180512.3999332571105083.01E-88passed
/periods/6/uppertolerance23.53670489155090123.5367046495350462.42E-77passed
/periods/7/horizontolerance88.00.00E+2passed
/periods/7/expectedtolerance20.17843490774236120.17843483187947.59E-88passed
/periods/7/incrementtolerance3.06573836652506423.06573835917395957.35E-98passed
/periods/7/lowertolerance14.8678336638295114.8678336905292672.67E-88passed
/periods/7/uppertolerance27.2873482323012227.2873479820839652.50E-78passed
/parameters/0/estimatetolerance3.81895404398836653.81895404709238573.10E-99passed
/parameters/0/standard_errortolerance0.151309005711001620.15130900279202212.92E-97passed
/parameters/0/lowertolerance3.57007287715420453.5700728850595187.91E-98passed
/parameters/0/uppertolerance4.0678352108225284.0678352091252541.70E-99passed
/parameters/1/estimatetolerance0.763436519566730530.76343652130667941.74E-98passed
/parameters/1/standard_errortolerance0.111584927200787250.111584924608461082.59E-97passed
/parameters/1/lowertolerance0.600291828934540470.60029183398437315.05E-98passed
/parameters/1/uppertolerance0.970919961450481210.97091995770846423.74E-98passed
/parameters/2/estimatetolerance00.00.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

464  censor <- data.frame(
465    start = c(8.3, 6, 12, 15.2, 12, 18, 18, 21.5, 24, 30, 24, 27.8),
466    end = c(8.3, 9, NA_real_, 15.2, 18, NA_real_, 24, 21.5, NA_real_, NA_real_, 30, 27.8),
467    freq = c(3L, 4L, 40L, 2L, 5L, 30L, 6L, 1L, 45L, 25L, 5L, 2L)
468  )
469  export_dataset(censor, "SURV_CENSOR_B9",
470                 "Constructed warranty censoring table with exact, interval and right-censored records (B9)",
471                 "Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley",
472                 "Constructed validation data; no third-party license.",
473                 c(start = "Float64", end = "Float64", freq = "Int64"))
474  cn <- read_dataset("SURV_CENSOR_B9")
475  right <- is.na(cn$end)
476  censor_ages <- sort(unique(cn$start[right]))
477  censor_survivors <- vapply(censor_ages, function(a) sum(cn$freq[right & cn$start == a]), 0)
478  censor_records <- data.frame(start = cn$start, end = cn$end, frequency = cn$freq)
479  future_units_2 <- 400
480  future_times_2 <- 2.5
481  pred_lognormal <- warranty_prediction_reference(
482    censor_records, "lognormal", censor_survivors, censor_ages,
483    future_units_2, future_times_2, 8, NULL, 0.90, 1
484  )
485  emit("STAT-SURV-WARRPRED-002",
486       "Lognormal warranty prediction from a censoring table without a warranty limit",
487       PRED, "warrantyPredictionMenuItem", PRED_REQ, "SURV_CENSOR_B9",
488       list(source = "censoring_table", start = "start", end = "end", frequency = "freq",
489            family = "lognormal", horizon = 8, future_shipments = list("400@2.5"),
490            confidence_level = 0.90, period_width = 1, period_unit = "month"),
491       prediction_outputs(pred_lognormal, "lognormal", censor_survivors, censor_ages,
492                          future_units_2, future_times_2, 0.90),
493       c(PREDICTION_NOTES,
494         "The censoring-table source reads start, end and frequency columns directly; the risk-set cohorts are the distinct right-censoring ages with their surviving weights; no warranty limit means the cap is infinite."),
495       "axis_coverage")

Whole script generators/r/survival_warranty.R


Warranty triangle conversion on the calendar axis with the final period incomplete

Case STAT-SURV-WARRPREP-001 · status: passed

Reference

Evidence type
R
Source
R survival package survreg on interval2 records with frequency weights; independent base-R implementations of the warranty triangle conversion and the expected-count algebra
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_warranty.R · sha256 4e304617e4b0f92b
Dataset
reference_data/datasets/SURV_WARR_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/conversion/as_of_periodexact660.00E+2passed
/conversion/shipped_totalexact8208200.00E+2passed
/conversion/failures_usedexact73730.00E+2passed
/conversion/survivors_totalexact7477470.00E+2passed
/conversion/excluded_failuresexact770.00E+2passed
/conversion/excluded_shipmentsexact000.00E+2passed
/conversion/source_countexact660.00E+2passed
/conversion/excluded_row_countexact000.00E+2passed
/conversion/cohorts/0/shipment_periodexact110.00E+2passed
/conversion/cohorts/0/shippedexact1201200.00E+2passed
/conversion/cohorts/0/failuresexact13130.00E+2passed
/conversion/cohorts/0/survivorsexact1071070.00E+2passed
/conversion/cohorts/0/agetolerance66.00.00E+2passed
/conversion/cohorts/0/excluded_failuresexact110.00E+2passed
/conversion/cohorts/1/shipment_periodexact220.00E+2passed
/conversion/cohorts/1/shippedexact1351350.00E+2passed
/conversion/cohorts/1/failuresexact15150.00E+2passed
/conversion/cohorts/1/survivorsexact1201200.00E+2passed
/conversion/cohorts/1/agetolerance55.00.00E+2passed
/conversion/cohorts/1/excluded_failuresexact110.00E+2passed
/conversion/cohorts/2/shipment_periodexact330.00E+2passed
/conversion/cohorts/2/shippedexact1501500.00E+2passed
/conversion/cohorts/2/failuresexact15150.00E+2passed
/conversion/cohorts/2/survivorsexact1351350.00E+2passed
/conversion/cohorts/2/agetolerance44.00.00E+2passed
/conversion/cohorts/2/excluded_failuresexact220.00E+2passed
/conversion/cohorts/3/shipment_periodexact440.00E+2passed
/conversion/cohorts/3/shippedexact1601600.00E+2passed
/conversion/cohorts/3/failuresexact16160.00E+2passed
/conversion/cohorts/3/survivorsexact1441440.00E+2passed
/conversion/cohorts/3/agetolerance33.00.00E+2passed
/conversion/cohorts/3/excluded_failuresexact000.00E+2passed
/conversion/cohorts/4/shipment_periodexact550.00E+2passed
/conversion/cohorts/4/shippedexact1451450.00E+2passed
/conversion/cohorts/4/failuresexact11110.00E+2passed
/conversion/cohorts/4/survivorsexact1341340.00E+2passed
/conversion/cohorts/4/agetolerance22.00.00E+2passed
/conversion/cohorts/4/excluded_failuresexact110.00E+2passed
/conversion/cohorts/5/shipment_periodexact660.00E+2passed
/conversion/cohorts/5/shippedexact1101100.00E+2passed
/conversion/cohorts/5/failuresexact330.00E+2passed
/conversion/cohorts/5/survivorsexact1071070.00E+2passed
/conversion/cohorts/5/agetolerance11.00.00E+2passed
/conversion/cohorts/5/excluded_failuresexact220.00E+2passed
/conversion/records/0/starttolerance00.00.00E+2passed
/conversion/records/0/endtolerance11.00.00E+2passed
/conversion/records/0/frequencyexact220.00E+2passed
/conversion/records/1/starttolerance11.00.00E+2passed
/conversion/records/1/endtolerance22.00.00E+2passed
/conversion/records/1/frequencyexact330.00E+2passed
/conversion/records/2/starttolerance22.00.00E+2passed
/conversion/records/2/endtolerance33.00.00E+2passed
/conversion/records/2/frequencyexact440.00E+2passed
/conversion/records/3/starttolerance33.00.00E+2passed
/conversion/records/3/endtolerance44.00.00E+2passed
/conversion/records/3/frequencyexact220.00E+2passed
/conversion/records/4/starttolerance44.00.00E+2passed
/conversion/records/4/endtolerance55.00.00E+2passed
/conversion/records/4/frequencyexact110.00E+2passed
/conversion/records/5/starttolerance55.00.00E+2passed
/conversion/records/5/endtolerance66.00.00E+2passed
/conversion/records/5/frequencyexact110.00E+2passed
/conversion/records/6/starttolerance66.00.00E+2passed
/conversion/records/6/frequencyexact1071070.00E+2passed
/conversion/records/7/starttolerance00.00.00E+2passed
/conversion/records/7/endtolerance11.00.00E+2passed
/conversion/records/7/frequencyexact330.00E+2passed
/conversion/records/8/starttolerance11.00.00E+2passed
/conversion/records/8/endtolerance22.00.00E+2passed
/conversion/records/8/frequencyexact550.00E+2passed
/conversion/records/9/starttolerance22.00.00E+2passed
/conversion/records/9/endtolerance33.00.00E+2passed
/conversion/records/9/frequencyexact330.00E+2passed
/conversion/records/10/starttolerance33.00.00E+2passed
/conversion/records/10/endtolerance44.00.00E+2passed
/conversion/records/10/frequencyexact220.00E+2passed
/conversion/records/11/starttolerance44.00.00E+2passed
/conversion/records/11/endtolerance55.00.00E+2passed
/conversion/records/11/frequencyexact220.00E+2passed
/conversion/records/12/starttolerance55.00.00E+2passed
/conversion/records/12/frequencyexact1201200.00E+2passed
/conversion/records/13/starttolerance00.00.00E+2passed
/conversion/records/13/endtolerance11.00.00E+2passed
/conversion/records/13/frequencyexact440.00E+2passed
/conversion/records/14/starttolerance11.00.00E+2passed
/conversion/records/14/endtolerance22.00.00E+2passed
/conversion/records/14/frequencyexact660.00E+2passed
/conversion/records/15/starttolerance22.00.00E+2passed
/conversion/records/15/endtolerance33.00.00E+2passed
/conversion/records/15/frequencyexact330.00E+2passed
/conversion/records/16/starttolerance33.00.00E+2passed
/conversion/records/16/endtolerance44.00.00E+2passed
/conversion/records/16/frequencyexact220.00E+2passed
/conversion/records/17/starttolerance44.00.00E+2passed
/conversion/records/17/frequencyexact1351350.00E+2passed
/conversion/records/18/starttolerance00.00.00E+2passed
/conversion/records/18/endtolerance11.00.00E+2passed
/conversion/records/18/frequencyexact550.00E+2passed
/conversion/records/19/starttolerance11.00.00E+2passed
/conversion/records/19/endtolerance22.00.00E+2passed
/conversion/records/19/frequencyexact770.00E+2passed
/conversion/records/20/starttolerance22.00.00E+2passed
/conversion/records/20/endtolerance33.00.00E+2passed
/conversion/records/20/frequencyexact440.00E+2passed
/conversion/records/21/starttolerance33.00.00E+2passed
/conversion/records/21/frequencyexact1441440.00E+2passed
/conversion/records/22/starttolerance00.00.00E+2passed
/conversion/records/22/endtolerance11.00.00E+2passed
/conversion/records/22/frequencyexact660.00E+2passed
/conversion/records/23/starttolerance11.00.00E+2passed
/conversion/records/23/endtolerance22.00.00E+2passed
/conversion/records/23/frequencyexact550.00E+2passed
/conversion/records/24/starttolerance22.00.00E+2passed
/conversion/records/24/frequencyexact1341340.00E+2passed
/conversion/records/25/starttolerance00.00.00E+2passed
/conversion/records/25/endtolerance11.00.00E+2passed
/conversion/records/25/frequencyexact330.00E+2passed
/conversion/records/26/starttolerance11.00.00E+2passed
/conversion/records/26/frequencyexact1071070.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

382  emit("STAT-SURV-WARRPREP-001",
383       "Warranty triangle conversion on the calendar axis with the final period incomplete",
384       PREP, "warrantyPreprocessMenuItem", PREP_REQ, "SURV_WARR_B9", PREP_CALENDAR,
385       conversion_outputs(conv_calendar),
386       c(CONVERSION_NOTES,
387         "The default 'incomplete' setting excludes calendar period 7: its 7 returns are reported as excluded failures, and units unreturned at the end of period 6 stay survivors right censored at (6 - i + 1) months."),
388       "nominal")

Whole script generators/r/survival_warranty.R


Warranty triangle conversion on the age axis with the final period complete

Case STAT-SURV-WARRPREP-002 · status: passed

Reference

Evidence type
R
Source
R survival package survreg on interval2 records with frequency weights; independent base-R implementations of the warranty triangle conversion and the expected-count algebra
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_warranty.R · sha256 4e304617e4b0f92b
Dataset
reference_data/datasets/SURV_WARR_AGE_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/conversion/as_of_periodexact660.00E+2passed
/conversion/shipped_totalexact104010400.00E+2passed
/conversion/failures_usedexact78780.00E+2passed
/conversion/survivors_totalexact9629620.00E+2passed
/conversion/excluded_failuresexact000.00E+2passed
/conversion/excluded_shipmentsexact000.00E+2passed
/conversion/source_countexact660.00E+2passed
/conversion/excluded_row_countexact000.00E+2passed
/conversion/cohorts/0/shipment_periodexact110.00E+2passed
/conversion/cohorts/0/shippedexact2002000.00E+2passed
/conversion/cohorts/0/failuresexact20200.00E+2passed
/conversion/cohorts/0/survivorsexact1801800.00E+2passed
/conversion/cohorts/0/agetolerance66.00.00E+2passed
/conversion/cohorts/0/excluded_failuresexact000.00E+2passed
/conversion/cohorts/1/shipment_periodexact220.00E+2passed
/conversion/cohorts/1/shippedexact1801800.00E+2passed
/conversion/cohorts/1/failuresexact21210.00E+2passed
/conversion/cohorts/1/survivorsexact1591590.00E+2passed
/conversion/cohorts/1/agetolerance55.00.00E+2passed
/conversion/cohorts/1/excluded_failuresexact000.00E+2passed
/conversion/cohorts/2/shipment_periodexact330.00E+2passed
/conversion/cohorts/2/shippedexact1901900.00E+2passed
/conversion/cohorts/2/failuresexact17170.00E+2passed
/conversion/cohorts/2/survivorsexact1731730.00E+2passed
/conversion/cohorts/2/agetolerance44.00.00E+2passed
/conversion/cohorts/2/excluded_failuresexact000.00E+2passed
/conversion/cohorts/3/shipment_periodexact440.00E+2passed
/conversion/cohorts/3/shippedexact1701700.00E+2passed
/conversion/cohorts/3/failuresexact12120.00E+2passed
/conversion/cohorts/3/survivorsexact1581580.00E+2passed
/conversion/cohorts/3/agetolerance33.00.00E+2passed
/conversion/cohorts/3/excluded_failuresexact000.00E+2passed
/conversion/cohorts/4/shipment_periodexact550.00E+2passed
/conversion/cohorts/4/shippedexact1601600.00E+2passed
/conversion/cohorts/4/failuresexact660.00E+2passed
/conversion/cohorts/4/survivorsexact1541540.00E+2passed
/conversion/cohorts/4/agetolerance22.00.00E+2passed
/conversion/cohorts/4/excluded_failuresexact000.00E+2passed
/conversion/cohorts/5/shipment_periodexact660.00E+2passed
/conversion/cohorts/5/shippedexact1401400.00E+2passed
/conversion/cohorts/5/failuresexact220.00E+2passed
/conversion/cohorts/5/survivorsexact1381380.00E+2passed
/conversion/cohorts/5/agetolerance11.00.00E+2passed
/conversion/cohorts/5/excluded_failuresexact000.00E+2passed
/conversion/records/0/starttolerance00.00.00E+2passed
/conversion/records/0/endtolerance11.00.00E+2passed
/conversion/records/0/frequencyexact440.00E+2passed
/conversion/records/1/starttolerance11.00.00E+2passed
/conversion/records/1/endtolerance22.00.00E+2passed
/conversion/records/1/frequencyexact660.00E+2passed
/conversion/records/2/starttolerance22.00.00E+2passed
/conversion/records/2/endtolerance33.00.00E+2passed
/conversion/records/2/frequencyexact550.00E+2passed
/conversion/records/3/starttolerance33.00.00E+2passed
/conversion/records/3/endtolerance44.00.00E+2passed
/conversion/records/3/frequencyexact330.00E+2passed
/conversion/records/4/starttolerance44.00.00E+2passed
/conversion/records/4/endtolerance55.00.00E+2passed
/conversion/records/4/frequencyexact220.00E+2passed
/conversion/records/5/starttolerance66.00.00E+2passed
/conversion/records/5/frequencyexact1801800.00E+2passed
/conversion/records/6/starttolerance00.00.00E+2passed
/conversion/records/6/endtolerance11.00.00E+2passed
/conversion/records/6/frequencyexact550.00E+2passed
/conversion/records/7/starttolerance11.00.00E+2passed
/conversion/records/7/endtolerance22.00.00E+2passed
/conversion/records/7/frequencyexact440.00E+2passed
/conversion/records/8/starttolerance22.00.00E+2passed
/conversion/records/8/endtolerance33.00.00E+2passed
/conversion/records/8/frequencyexact660.00E+2passed
/conversion/records/9/starttolerance33.00.00E+2passed
/conversion/records/9/endtolerance44.00.00E+2passed
/conversion/records/9/frequencyexact330.00E+2passed
/conversion/records/10/starttolerance44.00.00E+2passed
/conversion/records/10/endtolerance55.00.00E+2passed
/conversion/records/10/frequencyexact330.00E+2passed
/conversion/records/11/starttolerance55.00.00E+2passed
/conversion/records/11/frequencyexact1591590.00E+2passed
/conversion/records/12/starttolerance00.00.00E+2passed
/conversion/records/12/endtolerance11.00.00E+2passed
/conversion/records/12/frequencyexact660.00E+2passed
/conversion/records/13/starttolerance11.00.00E+2passed
/conversion/records/13/endtolerance22.00.00E+2passed
/conversion/records/13/frequencyexact550.00E+2passed
/conversion/records/14/starttolerance22.00.00E+2passed
/conversion/records/14/endtolerance33.00.00E+2passed
/conversion/records/14/frequencyexact440.00E+2passed
/conversion/records/15/starttolerance33.00.00E+2passed
/conversion/records/15/endtolerance44.00.00E+2passed
/conversion/records/15/frequencyexact220.00E+2passed
/conversion/records/16/starttolerance44.00.00E+2passed
/conversion/records/16/frequencyexact1731730.00E+2passed
/conversion/records/17/starttolerance00.00.00E+2passed
/conversion/records/17/endtolerance11.00.00E+2passed
/conversion/records/17/frequencyexact330.00E+2passed
/conversion/records/18/starttolerance11.00.00E+2passed
/conversion/records/18/endtolerance22.00.00E+2passed
/conversion/records/18/frequencyexact550.00E+2passed
/conversion/records/19/starttolerance22.00.00E+2passed
/conversion/records/19/endtolerance33.00.00E+2passed
/conversion/records/19/frequencyexact440.00E+2passed
/conversion/records/20/starttolerance33.00.00E+2passed
/conversion/records/20/frequencyexact1581580.00E+2passed
/conversion/records/21/starttolerance00.00.00E+2passed
/conversion/records/21/endtolerance11.00.00E+2passed
/conversion/records/21/frequencyexact440.00E+2passed
/conversion/records/22/starttolerance11.00.00E+2passed
/conversion/records/22/endtolerance22.00.00E+2passed
/conversion/records/22/frequencyexact220.00E+2passed
/conversion/records/23/starttolerance22.00.00E+2passed
/conversion/records/23/frequencyexact1541540.00E+2passed
/conversion/records/24/starttolerance00.00.00E+2passed
/conversion/records/24/endtolerance11.00.00E+2passed
/conversion/records/24/frequencyexact220.00E+2passed
/conversion/records/25/starttolerance11.00.00E+2passed
/conversion/records/25/frequencyexact1381380.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

400  ACOLS <- paste0("r", 1:5)
401  warr_age <- data.frame(
402    label = c("2026-01", "2026-02", "2026-03", "2026-04", "2026-05", "2026-06"),
403    shipped = c(200L, 180L, 190L, 170L, 160L, 140L),
404    r1 = c(4L, 5L, 6L, 3L, 4L, 2L),
405    r2 = c(6L, 4L, 5L, 5L, 2L, NA_integer_),
406    r3 = c(5L, 6L, 4L, 4L, NA_integer_, NA_integer_),
407    r4 = c(3L, 3L, 2L, NA_integer_, NA_integer_, NA_integer_),
408    r5 = c(2L, 3L, NA_integer_, NA_integer_, NA_integer_, NA_integer_)
409  )
410  export_dataset(warr_age, "SURV_WARR_AGE_B9",
411                 "Constructed warranty shipment-by-return triangle, age return periods (B9)",
412                 "Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley",
413                 "Constructed validation data; no third-party license.",
414                 c(label = "String", shipped = "Int64",
415                   setNames(rep("Int64", length(ACOLS)), ACOLS)))
416  wa <- read_dataset("SURV_WARR_AGE_B9")
417  conv_age <- convert_reference(wa, ACOLS, "age", 1, "complete")
418  stopifnot(
419    conv_age$failures_used + conv_age$survivors_total == conv_age$shipped_total,
420    conv_age$excluded_failures == 0, conv_age$excluded_shipments == 0
421  )
422  emit("STAT-SURV-WARRPREP-002",
423       "Warranty triangle conversion on the age axis with the final period complete",
424       PREP, "warrantyPreprocessMenuItem", PREP_REQ, "SURV_WARR_AGE_B9",
425       list(shipped = "shipped", returns = as.list(ACOLS), return_axis = "age",
426            period_width = 1, period_unit = "month", last_period = "complete",
427            shipment_label = "label"),
428       conversion_outputs(conv_age),
429       c(CONVERSION_NOTES,
430         "Age-axis reading: column k counts returns in age period k (calendar period i + k - 1), so the last calendar period equals the number of shipment rows; the 'complete' setting includes it as a full period of exposure, and blank cells lie after the last calendar period."),
431       "axis_coverage")

Whole script generators/r/survival_warranty.R


Derived warranty censoring table with interval records and right-censored survivors

Case STAT-SURV-WARRTAB-001 · status: passed

Reference

Evidence type
R
Source
R survival package survreg on interval2 records with frequency weights; independent base-R implementations of the warranty triangle conversion and the expected-count algebra
Citation
Therneau T (2024). A Package for Survival Analysis in R. R package version 3.6; Meeker, Hahn and Escobar (1998) Statistical Methods for Reliability Data, Wiley.
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, survival 3.8.6, jsonlite 2.0.0
Reference script
generators/r/survival_warranty.R · sha256 4e304617e4b0f92b
Dataset
reference_data/datasets/SURV_WARR_B9.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/0/starttolerance00.00.00E+2passed
/0/endtolerance11.00.00E+2passed
/0/frequencyexact220.00E+2passed
/1/starttolerance11.00.00E+2passed
/1/endtolerance22.00.00E+2passed
/1/frequencyexact330.00E+2passed
/2/starttolerance22.00.00E+2passed
/2/endtolerance33.00.00E+2passed
/2/frequencyexact440.00E+2passed
/3/starttolerance33.00.00E+2passed
/3/endtolerance44.00.00E+2passed
/3/frequencyexact220.00E+2passed
/4/starttolerance44.00.00E+2passed
/4/endtolerance55.00.00E+2passed
/4/frequencyexact110.00E+2passed
/5/starttolerance55.00.00E+2passed
/5/endtolerance66.00.00E+2passed
/5/frequencyexact110.00E+2passed
/6/starttolerance66.00.00E+2passed
/6/frequencyexact1071070.00E+2passed
/7/starttolerance00.00.00E+2passed
/7/endtolerance11.00.00E+2passed
/7/frequencyexact330.00E+2passed
/8/starttolerance11.00.00E+2passed
/8/endtolerance22.00.00E+2passed
/8/frequencyexact550.00E+2passed
/9/starttolerance22.00.00E+2passed
/9/endtolerance33.00.00E+2passed
/9/frequencyexact330.00E+2passed
/10/starttolerance33.00.00E+2passed
/10/endtolerance44.00.00E+2passed
/10/frequencyexact220.00E+2passed
/11/starttolerance44.00.00E+2passed
/11/endtolerance55.00.00E+2passed
/11/frequencyexact220.00E+2passed
/12/starttolerance55.00.00E+2passed
/12/frequencyexact1201200.00E+2passed
/13/starttolerance00.00.00E+2passed
/13/endtolerance11.00.00E+2passed
/13/frequencyexact440.00E+2passed
/14/starttolerance11.00.00E+2passed
/14/endtolerance22.00.00E+2passed
/14/frequencyexact660.00E+2passed
/15/starttolerance22.00.00E+2passed
/15/endtolerance33.00.00E+2passed
/15/frequencyexact330.00E+2passed
/16/starttolerance33.00.00E+2passed
/16/endtolerance44.00.00E+2passed
/16/frequencyexact220.00E+2passed
/17/starttolerance44.00.00E+2passed
/17/frequencyexact1351350.00E+2passed
/18/starttolerance00.00.00E+2passed
/18/endtolerance11.00.00E+2passed
/18/frequencyexact550.00E+2passed
/19/starttolerance11.00.00E+2passed
/19/endtolerance22.00.00E+2passed
/19/frequencyexact770.00E+2passed
/20/starttolerance22.00.00E+2passed
/20/endtolerance33.00.00E+2passed
/20/frequencyexact440.00E+2passed
/21/starttolerance33.00.00E+2passed
/21/frequencyexact1441440.00E+2passed
/22/starttolerance00.00.00E+2passed
/22/endtolerance11.00.00E+2passed
/22/frequencyexact660.00E+2passed
/23/starttolerance11.00.00E+2passed
/23/endtolerance22.00.00E+2passed
/23/frequencyexact550.00E+2passed
/24/starttolerance22.00.00E+2passed
/24/frequencyexact1341340.00E+2passed
/25/starttolerance00.00.00E+2passed
/25/endtolerance11.00.00E+2passed
/25/frequencyexact330.00E+2passed
/26/starttolerance11.00.00E+2passed
/26/frequencyexact1071070.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

390  emit("STAT-SURV-WARRTAB-001",
391       "Derived warranty censoring table with interval records and right-censored survivors",
392       TABLE, "warrantyPreprocessMenuItem", PREP_REQ, "SURV_WARR_B9", PREP_CALENDAR,
393       table_outputs(conv_calendar),
394       c(CONVERSION_NOTES,
395         "The derived frequency table keeps one row per cohort and period with a frequency and never expands units into rows (manual :155-157); the result is a frame of row dicts, so numeric fields are compared positionally, the null end of right-censored rows and the string cohort and censoring columns are skipped."),
396       "nominal")

Whole script generators/r/survival_warranty.R


Parametric survival STAT-SURV-WEIBULL-AFT-001

Case STAT-SURV-WEIBULL-AFT-001 · status: passed

Reference

Evidence type
R
Source
R survival::survreg
Citation
R survival::survreg and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_parametric.R · sha256 7c99eb709767fd04
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/0/estimatetolerance2.43827905966047712.43827905996057133.00E-109passed
/coefficients/1/estimatetolerance0.00715852277534005050.0071585227798142084.47E-129passed
/coefficients/2/estimatetolerance3.86955713024814243.86955712887317161.37E-99passed
/analyzed_countexact32320.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

74  emit("STAT-SURV-WEIBULL-AFT-001","tabelz.analysis.weibull_aft:compute_weibull_aft",
75       "weibullAftMenuItem","SURV_CORE",
76       list(time_column="time",event_column="status",stress_column="stress",
77            prediction_stresses=list(50)),
78       list("/coefficients/0/estimate"=tol(coef(fit_aft)[[1]],"Log scale intercept"),
79            "/coefficients/1/estimate"=tol(coef(fit_aft)[[2]],"Log scale stress slope"),
80            "/coefficients/2/estimate"=tol(1/fit_aft$scale,"Weibull shape=1/survreg scale"),
81            "/analyzed_count"=exact(nrow(d),"All complete rows")),
82       c("R survreg Weibull scale is inverse shape; Tabelz k=1/scale and beta0 is log scale.",
83         "R maxiter=200; no predictor standardization."))

Whole script generators/r/survival_parametric.R


Survival boundary STAT-SURV-WEIBULL-CONSTANT-001

Case STAT-SURV-WEIBULL-CONSTANT-001 · status: passed

Reference

Evidence type
R
Source
R survival and analytic boundary
Citation
R survival::survfit/coxph/survreg documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_boundaries.R · sha256 528007972452fb15
Dataset
reference_data/datasets/SURV_CONST_STRESS.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

84  emit("STAT-SURV-WEIBULL-CONSTANT-001","tabelz.analysis.weibull_aft:compute_weibull_aft",
85       "weibullAftMenuItem","SURV_CONST_STRESS",
86       list(time_column="time",event_column="status",stress_column="stress"),list(),
87       c("Constant stress makes the Weibull AFT stress coefficient unidentified."),
88       error=list(error_class="ValueError",message_contains="stress covariate is constant"))

Whole script generators/r/survival_boundaries.R


Hard-numeric survival STAT-SURV-WEIBULL-HARD-001

Case STAT-SURV-WEIBULL-HARD-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats
Citation
R survival::survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_hard.R · sha256 bc20ccd17540f4e7
Dataset
reference_data/datasets/SURV_HARD.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/coefficients/2/estimatetolerance2523.58273055680592523.5827307557521.99E-710passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

85  emit("STAT-SURV-WEIBULL-HARD-001","tabelz.analysis.weibull_aft:compute_weibull_aft",
86       "weibullAftMenuItem","SURV_HARD",
87       list(time_column="time",event_column="status",stress_column="stress",
88            prediction_stresses=list(use_stress)),
89       list("/coefficients/2/estimate"=tol(1/fit_alt$scale,"Large-offset Weibull shape=1/survreg scale")),
90       c("The independent reference uses the same censored Weibull likelihood and converted shape."))

Whole script generators/r/survival_hard.R


Invalid survival request STAT-SURV-WEIBULL-INVALID-001

Case STAT-SURV-WEIBULL-INVALID-001 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Invalid survival request STAT-SURV-WEIBULL-INVALID-002

Case STAT-SURV-WEIBULL-INVALID-002 · status: passed

Reference

Evidence type
Analytical
Source
Strict request-contract rejection
Citation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · jsonlite 2.0.0
Reference script
none (values transcribed from the source)
Dataset
reference_data/datasets/SURV_CORE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention


Small-sample survival boundary STAT-SURV-WEIBULL-MIN-001

Case STAT-SURV-WEIBULL-MIN-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::survfit/coxph and Surv(type=interval2) documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_minimum.R · sha256 2997d37929a6dae8
Dataset
reference_data/datasets/SURV_MIN.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/failure_countexact330.00E+2passed
/analyzed_countexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

65  emit("STAT-SURV-WEIBULL-MIN-001","tabelz.analysis.weibull_aft:compute_weibull_aft",
66       "weibullAftMenuItem","SURV_MIN",integer(),
67       list(time_column="time",event_column="status",stress_column="stress"),
68       list("/failure_count"=exact(3L,"Minimum three observed failures"),
69            "/analyzed_count"=exact(5L,"Five complete rows")),
70       c("Three failures and varying stress meet Weibull AFT sample condition."))

Whole script generators/r/survival_minimum.R


Survival population STAT-SURV-WEIBULL-POP-001

Case STAT-SURV-WEIBULL-POP-001 · status: passed

Reference

Evidence type
R
Source
R survival/stats and analytic counts
Citation
R survival::Surv/survfit/coxph/survreg and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · survival 3.8.6, stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/survival_population.R · sha256 3b3b4629e272fcd9
Dataset
reference_data/datasets/SURV_POP.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/source_countexact32320.00E+2passed
/excluded_countexact110.00E+2passed
/missing_response_countexact220.00E+2passed
/incomplete_covariate_countexact110.00E+2passed
/analyzed_countexact28280.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 96  emit("STAT-SURV-WEIBULL-POP-001","tabelz.analysis.weibull_aft:compute_weibull_aft",
 97       "weibullAftMenuItem","SURV_POP",7L,
 98       list(time_column="time",event_column="status",stress_column="stress",prediction_stresses=list(50)),
 99       list("/source_count"=exact(nrow(d),"CSV population"),
100            "/excluded_count"=exact(1L,"One stable identity excluded"),
101            "/missing_response_count"=exact(2L,"Missing time and status"),
102            "/incomplete_covariate_count"=exact(1L,"Missing stress"),
103            "/analyzed_count"=exact(nrow(altdata),"Complete AFT rows")),
104       c("This checks Weibull AFT population partition; seeded ML coefficients are compared in a separate case."))

Whole script generators/r/survival_population.R


Biased sample ACF and lag-one PACF

Case STAT-TS-ACF-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::acf, stats::pacf and stats::diff documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, urca 1.3.4
Reference script
generators/r/timeseries_correlations.R · sha256 806e71b5ec8f6997
Dataset
reference_data/datasets/TS_CORRELATIONS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sample_countexact36360.00E+2passed
/correlations/acf/0tolerance11.00.00E+2passed
/correlations/acf/1tolerance0.896038737170922860.89603873717092299.97E-1917passed
/correlations/acf/2tolerance0.757123274564197810.75712327456419813.35E-1615passed
/correlations/pacf/1tolerance0.896038737170922860.89603873717092271.10E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

166  emit("STAT-TS-ACF-001", "Biased sample ACF and lag-one PACF", "timeSeriesAcfPacfMenuItem",
167       list(response_column = "value", task = "correlations", maximum_lag = 5L),
168       list("/sample_count" = exact(length(series), "All 36 values are finite."),
169            "/correlations/acf/0" = tol(acf[1], "Lag-zero ACF equals one; owner 1e-9 tier."),
170            "/correlations/acf/1" = tol(acf[2], "Biased lag-one ACF divides by n; owner 1e-9 tier."),
171            "/correlations/acf/2" = tol(acf[3], "Biased lag-two ACF divides by n; owner 1e-9 tier."),
172            "/correlations/pacf/1" = tol(pacf[1], "Yule-Walker lag-one PACF; owner 1e-9 tier.")),
173       c("Manual chapter 12 ACF/PACF task: autocorrelation uses a fixed n denominator.",
174         "R acf(type=correlation) uses the biased n denominator and R pacf uses Yule-Walker."))

Whole script generators/r/timeseries_correlations.R


ARMA(1,0) with a constant

Case STAT-TS-ARIMA-001 · status: passed

Reference

Evidence type
R
Source
R stats::arima
Citation
R stats::arima and stats::predict.Arima documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_MODELS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/forecast_rows/0/forecasttolerance10.33730744978985610.3373071404408033.09E-77passed
/forecast_rows/0/lowertolerance8.65982100367315878.6598206921924373.11E-77passed
/forecast_rows/0/uppertolerance12.01479389590655312.014793588689173.07E-77passed
/forecast_rows/1/forecasttolerance10.05947771824398110.0594774159406873.02E-77passed
/forecast_rows/1/lowertolerance8.22365460224096988.223654455301581.47E-77passed
/forecast_rows/1/uppertolerance11.89530083424699211.8953003765797964.58E-77passed
/forecast_rows/2/forecasttolerance9.93594879334517789.9359485584617742.35E-77passed
/forecast_rows/2/lowertolerance8.0704147087446118.0704146880322762.07E-88passed
/forecast_rows/2/uppertolerance11.80148287794574511.8014824288912734.49E-77passed
/parameters/1/termexactar.L1ar.L1passed
/parameters/1/estimatetolerance0.444620970590417770.444620739199785732.31E-76passed
/parameters/2/termexactsigma2sigma2passed
/parameters/2/estimatetolerance0.732523998889767050.73252400075147991.86E-98passed
/parameters/0/termexactinterceptinterceptpassed
/parameters/0/estimatetolerance5.46329404662685745.46329627382267452.23E-66passed
/diagnostics/ljung_box_model_dfexact110.00E+2passed
/convergedexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

174  arima_case("STAT-TS-ARIMA-001", "ARMA(1,0) with a constant", "ar", c(1, 0, 0), "constant",
175             arima(d$ar, order = c(1, 0, 0), method = "ML", optim.control = tight_optim),
176             "R parameterises the constant as the mean; the product labels the equation constant 'intercept'.")

Whole script generators/r/timeseries_models.R


ARIMA(1,1,0) without a deterministic term

Case STAT-TS-ARIMA-002 · status: passed

Reference

Evidence type
R
Source
Closed-form exact AR(1) profile likelihood of the differenced series in R
Citation
Box, Jenkins and Reinsel, Time Series Analysis, exact AR(1) likelihood; statistician ruling on issue #10
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_MODELS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/forecast_rows/0/forecasttolerance43.06766522329891743.0676652230660652.33E-1011passed
/forecast_rows/0/lowertolerance40.92595132141254840.9259513211640352.49E-1011passed
/forecast_rows/0/uppertolerance45.20937912518528645.2093791249680952.17E-1011passed
/forecast_rows/1/forecasttolerance43.06778923758066243.067789237355552.25E-1011passed
/forecast_rows/1/lowertolerance40.06399586681535440.063995867346045.31E-1010passed
/forecast_rows/1/uppertolerance46.0715826083459746.071582607365069.81E-1010passed
/forecast_rows/2/forecasttolerance43.06778717784550343.067787177620192.25E-1011passed
/forecast_rows/2/lowertolerance39.39885003476469239.39885003576511.00E-910passed
/forecast_rows/2/uppertolerance46.73672432092631446.736724319475281.45E-910passed
/parameters/0/termexactar.L1ar.L1passed
/parameters/0/estimatetolerance-0.016608854514315676-0.0166088550322454655.18E-107passed
/parameters/1/termexactsigma2sigma2passed
/parameters/1/estimatetolerance1.19406159264893691.19406159266641.75E-1110passed
/diagnostics/ljung_box_model_dfexact110.00E+2passed
/convergedexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

203    out[["/parameters/0/term"]] <- exact("ar.L1", "The AR coefficient comes first without a constant.")
204    out[["/parameters/0/estimate"]] <- fit6(rw_phi, "Exact ML AR coefficient of the differenced series.")
205    out[["/parameters/1/term"]] <- exact("sigma2", "Innovation variance last.")
206    out[["/parameters/1/estimate"]] <- fit6(rw_sigma2, "Exact ML innovation variance of the differenced series.")
207    out[["/diagnostics/ljung_box_model_df"]] <- exact(1L, "Chapter 12: subtract the fitted AR and MA terms.")
208    out[["/converged"]] <- exact(TRUE, "The optimizer converges on these data.")
209    deferred <- list(defer("/parameters/0/standard_error", "The manual does not state how parameter standard errors are estimated (observed Hessian or outer product of gradients)."))
210    emit("STAT-TS-ARIMA-002", "timeSeriesArimaMenuItem", "ARIMA(1,1,0) without a deterministic term", "TS_MODELS_01",
211         list(method = "forecast", response_column = "rw", model = "arima", arima_order = list(1L, 1L, 0L),
212              deterministic_term = "none", horizon = 3L), out,
213         c("Exact diffuse initialization of the integrated state (statistician ruling, issue #10): for complete data the likelihood of the differenced series.",
214           "Certified by the closed-form profile likelihood of AR(1) on diff(rw), solved to machine precision; iterative-fit tier 1e-6.",
215           "R stats::arima (approximate diffuse, kappa = 1e6) gave phi -0.016609067949116645, the reference before this method change.",
216           "One ordinary difference and no drift."),
217         "Closed-form exact AR(1) profile likelihood of the differenced series in R",
218         "Box, Jenkins and Reinsel, Time Series Analysis, exact AR(1) likelihood; statistician ruling on issue #10", deferred = deferred)

Whole script generators/r/timeseries_models.R


First ordinary difference of the guided workflow

Case STAT-TS-DIFF-001 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::acf, stats::pacf and stats::diff documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, urca 1.3.4
Reference script
generators/r/timeseries_correlations.R · sha256 806e71b5ec8f6997
Dataset
reference_data/datasets/TS_CORRELATIONS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/sample_countexact35350.00E+2passed
/points/0/transformedtolerance0.0999354330033099810.099935433003309983.59E-1917passed
/points/1/transformedtolerance0.264119688719411270.264119688719411271.90E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

180  first <- diff(series, differences = 1L)
181  emit("STAT-TS-DIFF-001", "First ordinary difference of the guided workflow", "timeSeriesStationarityMenuItem",
182       list(response_column = "value", task = "difference", difference_order = 2L,
183            maximum_lag = 5L),
184       list("/sample_count" = exact(length(first), "One leading row is unavailable after first differencing."),
185            "/points/0/transformed" = tol(first[1], "x2 - x1; owner 1e-9 tier."),
186            "/points/1/transformed" = tol(first[2], "x3 - x2; owner 1e-9 tier.")),
187       c("Manual chapter 12 guided differencing retains leading unavailable values as missing in a created column.",
188         "Transformation arithmetic only: the reference differences once because the guided workflow selects order one on this series (selection checked by STAT-TS-DIFF-002 on a well-conditioned dataset; this deterministic construction makes the ADF regression near-singular, so lag counts and t-values here are not reproducibly defined across independent least-squares implementations)."),
189       kind = "row_alignment", version = "2.0.0")

Whole script generators/r/timeseries_correlations.R


Guided selection of the first stationary difference order

Case STAT-TS-DIFF-002 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::acf, stats::pacf and stats::diff documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, urca 1.3.4
Reference script
generators/r/timeseries_correlations.R · sha256 806e71b5ec8f6997
Dataset
reference_data/datasets/TS_DIFF_GUIDED_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/recommended_difference_orderexact110.00E+2passed
/difference_orderexact110.00E+2passed
/sample_countexact59590.00E+2passed
/stationarity/assessmentexactdifference_candidatedifference_candidatepassed
/points/0/transformedtolerance-2.3107595600425452-2.31075956004254523.53E-1716passed
/differencing_stages/0/sample_countexact60600.00E+2passed
/differencing_stages/0/stationarity/assessmentexactdifference_candidatedifference_candidatepassed
/differencing_stages/0/stationarity/adf_statistictolerance-1.3980353474388854-1.39803534743888544.37E-1716passed
/differencing_stages/0/stationarity/adf_lagsexact110.00E+2passed
/differencing_stages/0/stationarity/kpss_statistictolerance0.938716023171423550.93871602317142382.25E-1615passed
/differencing_stages/0/stationarity/kpss_lagsexact440.00E+2passed
/differencing_stages/1/sample_countexact59590.00E+2passed
/differencing_stages/1/stationarity/assessmentexactstationarystationarypassed
/differencing_stages/1/stationarity/adf_statistictolerance-9.2680344058689013-9.2680344058689053.51E-1515passed
/differencing_stages/1/stationarity/adf_lagsexact000.00E+2passed
/differencing_stages/1/stationarity/kpss_statistictolerance0.12420026529096750.124200265290967455.13E-1715passed
/differencing_stages/1/stationarity/kpss_lagsexact220.00E+2passed
/differencing_stages/2/sample_countexact58580.00E+2passed
/differencing_stages/2/stationarity/assessmentexactstationarystationarypassed
/differencing_stages/2/stationarity/adf_statistictolerance-4.4997292587027369-4.4997292587027425.31E-1514passed
/differencing_stages/2/stationarity/adf_lagsexact660.00E+2passed
/differencing_stages/2/stationarity/kpss_statistictolerance0.123663525340192440.12366352534019222.36E-1614passed
/differencing_stages/2/stationarity/kpss_lagsexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

177  # stage-selection oracle (STAT-TS-DIFF-002). The old version of this case

Whole script generators/r/timeseries_correlations.R


Original series selected: no differencing recommended

Case STAT-TS-DIFF-003 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::acf, stats::pacf and stats::diff documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, urca 1.3.4
Reference script
generators/r/timeseries_correlations.R · sha256 806e71b5ec8f6997
Dataset
reference_data/datasets/TS_DIFF_STATIONARY_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/recommended_difference_orderexact000.00E+2passed
/difference_orderexact000.00E+2passed
/sample_countexact60600.00E+2passed
/stationarity/assessmentexactstationarystationarypassed
/points/0/transformedtolerance7.65122694130849557.65122694130849554.50E-1717passed
/differencing_stages/0/sample_countexact60600.00E+2passed
/differencing_stages/0/stationarity/assessmentexactstationarystationarypassed
/differencing_stages/0/stationarity/adf_statistictolerance-4.2581178142613352-4.2581178142613349.11E-1615passed
/differencing_stages/0/stationarity/adf_lagsexact000.00E+2passed
/differencing_stages/0/stationarity/kpss_statistictolerance0.223511258199238980.223511258199238983.14E-1816passed
/differencing_stages/0/stationarity/kpss_lagsexact440.00E+2passed
/differencing_stages/1/sample_countexact59590.00E+2passed
/differencing_stages/1/stationarity/assessmentexactstationarystationarypassed
/differencing_stages/1/stationarity/adf_statistictolerance-5.215336522986993-5.21533652298699754.46E-1515passed
/differencing_stages/1/stationarity/adf_lagsexact220.00E+2passed
/differencing_stages/1/stationarity/kpss_statistictolerance0.115642753210503370.115642753210503446.50E-1715passed
/differencing_stages/1/stationarity/kpss_lagsexact110.00E+2passed
/differencing_stages/2/sample_countexact58580.00E+2passed
/differencing_stages/2/stationarity/assessmentexactstationarystationarypassed
/differencing_stages/2/stationarity/adf_statistictolerance-5.7917001119297398-5.7917001119297551.51E-1414passed
/differencing_stages/2/stationarity/adf_lagsexact550.00E+2passed
/differencing_stages/2/stationarity/kpss_statistictolerance0.193684093510601040.193684093510600762.82E-1614passed
/differencing_stages/2/stationarity/kpss_lagsexact770.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

226  dataset_id <- "TS_DIFF_STATIONARY_01"
227  set.seed(7L)
228  stationary_raw <- data.frame(value = as.numeric(stats::arima.sim(list(ar = 0.55), n = 60L, sd = 1.8)))
229  export_dataset(stationary_raw, dataset_id, "Constructed stationary AR(1) series",
230                 "R stats::arima.sim with seed 7; manual chapter 12 guided differencing",
231                 "Constructed validation data; no third-party license.", c(value = "Float64"))
232  stationary_series <- read_dataset(dataset_id)$value
233  stages_stationary <- list(
234    stage_outputs("/differencing_stages/0", stationary_series),
235    stage_outputs("/differencing_stages/1", diff(stationary_series, differences = 1L)),
236    stage_outputs("/differencing_stages/2", diff(stationary_series, differences = 2L))
237  )
238  selected_stationary <- Position(function(item) item$diagnostics$assessment == "stationary", stages_stationary) - 1L
239  outputs_stationary <- list(
240    "/recommended_difference_order" = exact(selected_stationary, "The original series is the first tested stationary stage."),
241    "/difference_order" = exact(selected_stationary, "Order zero keeps the original values as the transformed column."),
242    "/sample_count" = exact(length(stationary_series) - selected_stationary, "No observations are lost at order zero."),
243    "/stationarity/assessment" = label(stages_stationary[[1]]$diagnostics$assessment, "Source-level combined decision."),
244    "/points/0/transformed" = tol(stationary_series[1], "At order zero the transformed column equals the original series; owner 1e-9 tier.")
245  )
246  for (item in stages_stationary) outputs_stationary <- c(outputs_stationary, item$outputs)
247  emit("STAT-TS-DIFF-003", "Original series selected: no differencing recommended",
248       "timeSeriesStationarityMenuItem",
249       list(response_column = "value", task = "stationarity", difference_order = 2L, maximum_lag = 5L),
250       outputs_stationary,
251       c(selection_notes,
252         "Constructed with R seed 7: an AR(1) series with modest persistence is level-stationary at order zero."),
253       kind = "boundary")

Whole script generators/r/timeseries_correlations.R


No eligible differencing stage

Case STAT-TS-DIFF-004 · status: passed

Reference

Evidence type
R
Source
R stats
Citation
R stats::acf, stats::pacf and stats::diff documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, urca 1.3.4
Reference script
generators/r/timeseries_correlations.R · sha256 806e71b5ec8f6997
Dataset
reference_data/datasets/TS_DIFF_NONE_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/recommended_difference_orderabsentNonepassed
/difference_orderexact000.00E+2passed
/sample_countexact60600.00E+2passed
/stationarity/assessmentexactdifference_candidatedifference_candidatepassed
/points/0/transformedtolerance2.28724716134052392.2872471613405244.00E-1716passed
/differencing_stages/0/sample_countexact60600.00E+2passed
/differencing_stages/0/stationarity/assessmentexactdifference_candidatedifference_candidatepassed
/differencing_stages/0/stationarity/adf_statistictolerance0.874998616572586040.8749986165686693.92E-1211passed
/differencing_stages/0/stationarity/adf_lagsexact220.00E+2passed
/differencing_stages/0/stationarity/kpss_statistictolerance1.17408025838627421.17408025838627441.89E-1615passed
/differencing_stages/0/stationarity/kpss_lagsexact440.00E+2passed
/differencing_stages/1/sample_countexact59590.00E+2passed
/differencing_stages/1/stationarity/assessmentexactdifference_candidatedifference_candidatepassed
/differencing_stages/1/stationarity/adf_statistictolerance-0.272760508348389-0.272760508348283971.05E-1312passed
/differencing_stages/1/stationarity/adf_lagsexact110.00E+2passed
/differencing_stages/1/stationarity/kpss_statistictolerance1.26584488325249421.26584488325249422.40E-1716passed
/differencing_stages/1/stationarity/kpss_lagsexact440.00E+2passed
/differencing_stages/2/sample_countexact58580.00E+2passed
/differencing_stages/2/stationarity/assessmentexactdifference_candidatedifference_candidatepassed
/differencing_stages/2/stationarity/adf_statistictolerance-1.7700395407584297-1.77003954075844261.29E-1414passed
/differencing_stages/2/stationarity/adf_lagsexact000.00E+2passed
/differencing_stages/2/stationarity/kpss_statistictolerance0.855141627351893920.85514162735189391.83E-1817passed
/differencing_stages/2/stationarity/kpss_lagsexact440.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

257  dataset_id <- "TS_DIFF_NONE_01"
258  set.seed(7L)
259  none_raw <- data.frame(value = cumsum(cumsum(cumsum(stats::rnorm(60L, sd = 1.0)))))
260  export_dataset(none_raw, dataset_id,
261                 "Constructed triple-integrated series; no tested order is stationary",
262                 "R stats::rnorm with seed 7; manual chapter 12 guided differencing",
263                 "Constructed validation data; no third-party license.", c(value = "Float64"))
264  none_series <- read_dataset(dataset_id)$value
265  stages_none <- list(
266    stage_outputs("/differencing_stages/0", none_series),
267    stage_outputs("/differencing_stages/1", diff(none_series, differences = 1L)),
268    stage_outputs("/differencing_stages/2", diff(none_series, differences = 2L))
269  )
270  eligible_none <- Filter(function(item) item$diagnostics$assessment == "stationary", stages_none)
271  stopifnot(length(eligible_none) == 0L)
272  outputs_none <- list(
273    "/recommended_difference_order" = absent("No order from zero through the configured maximum met the combined criteria."),
274    "/difference_order" = exact(0L, "Without a recommended order the workflow displays order zero."),
275    "/sample_count" = exact(length(none_series), "The transformed column is the original series at order zero."),
276    "/stationarity/assessment" = label(stages_none[[1]]$diagnostics$assessment, "Source-level combined decision."),
277    "/points/0/transformed" = tol(none_series[1], "At order zero the transformed column equals the original series; owner 1e-9 tier.")
278  )
279  for (item in stages_none) outputs_none <- c(outputs_none, item$outputs)
280  emit("STAT-TS-DIFF-004", "No eligible differencing stage", "timeSeriesStationarityMenuItem",
281       list(response_column = "value", task = "difference", difference_order = 2L, maximum_lag = 5L),
282       outputs_none,
283       c(selection_notes,
284         "Constructed with R seed 7: a triple-integrated random walk stays non-stationary through ordinary difference order two."),
285       kind = "boundary")

Whole script generators/r/timeseries_correlations.R


Regression with AR(1) errors on a future predictor path

Case STAT-TS-DYNREG-001 · status: passed

Reference

Evidence type
R
Source
R stats::arima with xreg
Citation
R stats::arima documentation (xreg)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_MODELS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/parameters/1/estimatetolerance1.52311142096500831.52311141797910742.99E-98passed
/parameters/2/estimatetolerance0.531513185560356090.53151318190244793.66E-98passed
/parameters/0/estimatetolerance0.655400664114721860.65540068048745011.64E-87passed
/conditional_on_predictorsexactTrueTruepassed
/forecast_rows/0/forecasttolerance6.38930924717120076.3893092675244132.04E-88passed
/forecast_rows/0/uppertolerance8.45518682177676698.4551868472232032.54E-88passed
/forecast_rows/1/forecasttolerance8.42874466715207948.4287446846755871.75E-88passed
/forecast_rows/1/uppertolerance10.76830543056782110.7683054503126151.97E-88passed
/forecast_rows/2/forecasttolerance10.22628910161527610.2262891143472391.27E-88passed
/forecast_rows/2/uppertolerance12.63754619960925912.6375462129003951.33E-88passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

253  emit("STAT-TS-DYNREG-001", "timeSeriesDynamicRegressionMenuItem", "Regression with AR(1) errors on a future predictor path",
254       "TS_MODELS_01",
255       list(method = "dynamic_regression", response_column = "dyn", features = list(list(column = "x", kind = "level")),
256            future_scenarios = list(list(column = "x", values = as.list(future_x))), arima_order = list(1L, 0L, 0L),
257            deterministic_term = "constant"),
258       dout, c("Chapter 12: the forecast is conditional on the explicit future predictor path.",
259               "R's arima with xreg is the same regression with ARMA errors; the product's constant sits in the error equation."),
260       "R stats::arima with xreg", "R stats::arima documentation (xreg)",
261       deferred = list(defer("/parameters/0/standard_error", "Standard-error method not stated in the manual.")))

Whole script generators/r/timeseries_models.R


Drift forecast with the manual's interval scale

Case STAT-TS-FC-DRIFT-001 · status: passed

Reference

Evidence type
R
Source
R forecast package
Citation
forecast::naive, snaive, meanf and rwf documentation; Hyndman and Athanasopoulos, Forecasting: Principles and Practice, section 5.5
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_MODELS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/forecast_rows/0/forecasttolerance42.94156752690878442.941567526908786.90E-1515passed
/forecast_rows/0/lowertolerance40.77926213082608340.7792621308260766.68E-1515passed
/forecast_rows/0/uppertolerance45.10387292299148545.103872922991487.12E-1515passed
/forecast_rows/1/forecasttolerance42.80800307345584342.8080030734558366.67E-1515passed
/forecast_rows/1/lowertolerance39.72466374635375539.724663746353756.81E-1515passed
/forecast_rows/1/uppertolerance45.89134240055793245.8913424005579247.54E-1515passed
/forecast_rows/2/forecasttolerance42.67443862000289642.674438620002897.54E-1515passed
/forecast_rows/2/lowertolerance38.86730710922237438.867307109222377.47E-1515passed
/forecast_rows/2/uppertolerance46.48157013078341746.481570130783416.62E-1515passed
/forecast_rows/3absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

135  drift <- (y[T] - y[1]) / (T - 1)
136  bench("STAT-TS-FC-DRIFT-001", "Drift forecast with the manual's interval scale", "drift", "rw", list(),
137        y[T] + drift * 1:3, rwf(y, h = 3, drift = TRUE, level = 95),
138        "Chapter 12: drift fitted values are one-step forecasts plus the drift; interval scale sqrt(h(1 + h/(T-1))).")

Whole script generators/r/timeseries_models.R


Mean forecast

Case STAT-TS-FC-MEAN-001 · status: passed

Reference

Evidence type
R
Source
R forecast package
Citation
forecast::naive, snaive, meanf and rwf documentation; Hyndman and Athanasopoulos, Forecasting: Principles and Practice, section 5.5
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_MODELS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/forecast_rows/0/forecasttolerance47.49927450914190547.4992745091419052.98E-1617passed
/forecast_rows/0/lowertolerance41.84691445511206441.846914455112064.64E-1616passed
/forecast_rows/0/uppertolerance53.15163456317174753.151634563171756.10E-1717passed
/forecast_rows/1/forecasttolerance47.49927450914190547.4992745091419052.98E-1617passed
/forecast_rows/1/lowertolerance41.84691445511206441.846914455112064.64E-1616passed
/forecast_rows/1/uppertolerance53.15163456317174753.151634563171756.10E-1717passed
/forecast_rows/2/forecasttolerance47.49927450914190547.4992745091419052.98E-1617passed
/forecast_rows/2/lowertolerance41.84691445511206441.846914455112064.64E-1616passed
/forecast_rows/2/uppertolerance53.15163456317174753.151634563171756.10E-1717passed
/forecast_rows/3absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

133  bench("STAT-TS-FC-MEAN-001", "Mean forecast", "mean", "rw", list(), rep(mean(y), 3),
134        meanf(y, h = 3, level = 95), "Mean: the sample mean; forecast::meanf uses a t quantile, which the manual does not settle.")

Whole script generators/r/timeseries_models.R


Naive forecast

Case STAT-TS-FC-NAIVE-001 · status: passed

Reference

Evidence type
R
Source
R forecast package
Citation
forecast::naive, snaive, meanf and rwf documentation; Hyndman and Athanasopoulos, Forecasting: Principles and Practice, section 5.5
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_MODELS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/forecast_rows/0/forecasttolerance43.07513198036173243.0751319803617257.03E-1515passed
/forecast_rows/0/lowertolerance40.93311376496359340.9331137649635866.67E-1515passed
/forecast_rows/0/uppertolerance45.21715019575987145.2171501957598647.39E-1515passed
/forecast_rows/1/forecasttolerance43.07513198036173243.0751319803617257.03E-1515passed
/forecast_rows/1/lowertolerance40.04586076929546640.045860769295467.32E-1515passed
/forecast_rows/1/uppertolerance46.10440319142799846.104403191427996.74E-1515passed
/forecast_rows/2/forecasttolerance43.07513198036173243.0751319803617257.03E-1515passed
/forecast_rows/2/lowertolerance39.36504760055413539.365047600554137.12E-1515passed
/forecast_rows/2/uppertolerance46.78521636016932946.785216360169326.94E-1515passed
/forecast_rows/3absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

130  y <- d$rw; T <- length(y)
131  bench("STAT-TS-FC-NAIVE-001", "Naive forecast", "naive", "rw", list(), rep(y[T], 3),
132        naive(y, h = 3, level = 95), "Naive: the last value, interval scale sqrt(h).")

Whole script generators/r/timeseries_models.R


Seasonal naive forecast

Case STAT-TS-FC-SNAIVE-001 · status: passed

Reference

Evidence type
R
Source
R forecast package
Citation
forecast::naive, snaive, meanf and rwf documentation; Hyndman and Athanasopoulos, Forecasting: Principles and Practice, section 5.5
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_MODELS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/forecast_rows/0/forecasttolerance23.1966205019604123.1966205019604073.19E-1515passed
/forecast_rows/0/lowertolerance21.78668483356226921.7866848335622663.40E-1515passed
/forecast_rows/0/uppertolerance24.60655617035855224.6065561703585483.97E-1515passed
/forecast_rows/1/forecasttolerance19.64715575255276619.6471557525527664.72E-1616passed
/forecast_rows/1/lowertolerance18.23722008415462518.2372200841546252.55E-1616passed
/forecast_rows/1/uppertolerance21.05709142095090821.0570914209509083.10E-1616passed
/forecast_rows/2/forecasttolerance21.88365576006345521.8836557600634557.97E-1717passed
/forecast_rows/2/lowertolerance20.47372009166531420.4737200916653142.97E-1616passed
/forecast_rows/2/uppertolerance23.29359142846159623.2935914284615961.38E-1617passed
/forecast_rows/3absentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

139  z4 <- d$s4
140  bench("STAT-TS-FC-SNAIVE-001", "Seasonal naive forecast", "seasonal_naive", "s4", list(season_length = 4L),
141        z4[T - 4 + 1:3], snaive(ts(z4, frequency = 4), h = 3, level = 95),
142        "Seasonal naive: the value one season earlier.")

Whole script generators/r/timeseries_models.R


Gaussian GARCH(1,1) with a constant mean

Case STAT-TS-GARCH-001 · status: passed

Reference

Evidence type
R
Source
Independent R implementation of the arch 8.0.0 GARCH(1,1) objective
Citation
arch GARCH.backcast documentation (Sheppard); Bollerslev (1986)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_LONG_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/parameters/0/estimatetolerance0.0435443572669782650.043544354984444532.28E-97passed
/parameters/1/estimatetolerance0.101877457107107160.101877455361550241.75E-97passed
/parameters/2/estimatetolerance0.165125244921047850.165125238412653476.51E-97passed
/parameters/3/estimatetolerance0.718560688417348330.71856069901583411.06E-87passed
/persistencetolerance0.883685933338396180.88368593742848764.09E-98passed
/conditional_volatility/0tolerance1.00069123944318661.00069124065122871.21E-98passed
/conditional_volatility/299tolerance0.699862425853214450.69986243236628386.51E-98passed
/forecast_volatility/0tolerance0.91700597154067920.91700597053315321.01E-98passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

346  ga <- arch_garch11(l$ret)
347  stopifnot(abs(ga$backcast - 1.0179017970706792) < 1e-12)
348  emit("STAT-TS-GARCH-001", "timeSeriesGarchMenuItem", "Gaussian GARCH(1,1) with a constant mean", "TS_LONG_01",
349       list(method = "advanced", kind = "garch", columns = list("ret"), distribution = "normal", mean = "constant",
350            garch_p = 1L, garch_q = 1L, horizon = 1L),
351       list("/parameters/0/estimate" = fit6(ga$mu, "Constant mean."),
352            "/parameters/1/estimate" = fit6(ga$omega, "omega, in squared return units."),
353            "/parameters/2/estimate" = fit6(ga$alpha, "alpha1."),
354            "/parameters/3/estimate" = fit6(ga$beta, "beta1."),
355            "/persistence" = fit6(ga$alpha + ga$beta, "alpha + beta."),
356            "/conditional_volatility/0" = fit6(ga$volatility[1], "sqrt(omega + (alpha + beta) b): the backcast start."),
357            "/conditional_volatility/299" = fit6(ga$volatility[300], "Recursion at row 300."),
358            "/forecast_volatility/0" = fit6(ga$forecast_vol, "One-step conditional volatility.")),
359       c("Iterative fit, tier 1e-6.",
360         "Statistician ruling, issue #9: the arch weighted backcast is retained; this reference reproduces its objective (rescale, fixed 0.94-weighted backcast over min(75, n) residuals, sigma_1^2 = omega + (alpha + beta) b, Gaussian likelihood) independently in R.",
361         sprintf("Backcast b = %.17g; maximized log L = %.17g.", ga$backcast, ga$loglik),
362         "Until 2026-09-25 the reference was rugarch, whose variance start is the mean squared residual."),
363       "Independent R implementation of the arch 8.0.0 GARCH(1,1) objective",
364       "arch GARCH.backcast documentation (Sheppard); Bollerslev (1986)")

Whole script generators/r/timeseries_models.R


Johansen trace statistics for a cointegrated pair

Case STAT-TS-JOHANSEN-001 · status: passed

Reference

Evidence type
R
Source
R urca package
Citation
urca::ca.jo documentation; Johansen (1988)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_SYSTEM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/johansen_trace/0/null_rankexact000.00E+2passed
/johansen_trace/0/statistictolerance53.77594611818158753.7759461182223864.08E-1112passed
/johansen_trace/1/statistictolerance3.8645957477460183.86459574774613171.14E-1313passed
/johansen_trace/0/reject_5pctexactTrueTruepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

296  cj <- ca.jo(data.frame(c1 = s$c1, c2 = s$c2), type = "trace", ecdet = "none", K = 2, spec = "transitory")
297  trace <- rev(cj@teststat)  # ca.jo lists r <= 1 before r = 0
298  emit("STAT-TS-JOHANSEN-001", "timeSeriesVarVecmMenuItem", "Johansen trace statistics for a cointegrated pair", "TS_SYSTEM_01",
299       list(method = "advanced", kind = "var_vecm", columns = list("c1", "c2"), lag_order = 2L, cointegration_rank = 1L,
300            horizon = 2L, deterministic = "constant"),
301       list("/johansen_trace/0/null_rank" = exact(0L, "Null r = 0 first."),
302            "/johansen_trace/0/statistic" = num(trace[1], "urca::ca.jo trace statistic, r = 0."),
303            "/johansen_trace/1/statistic" = num(trace[2], "urca::ca.jo trace statistic, r <= 1."),
304            "/johansen_trace/0/reject_5pct" = exact(unname(cj@teststat[2] > cj@cval[2, "5pct"]), "The pair is cointegrated.")),
305       c("K = 2 level lags; unrestricted constant outside the cointegration relation (ca.jo ecdet = 'none')."),
306       "R urca package", "urca::ca.jo documentation; Johansen (1988)")

Whole script generators/r/timeseries_models.R


A random walk is a differencing candidate

Case STAT-TS-STATIONARITY-001 · status: passed

Reference

Evidence type
Analytical
Source
Analytic
Citation
Dickey and Fuller (1979); Kwiatkowski et al. (1992)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_STATIONARITY_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stationarity/assessmentexactdifference_candidatedifference_candidatepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

410  emit("STAT-TS-STATIONARITY-001", "timeSeriesStationarityMenuItem", "A random walk is a differencing candidate", "TS_STATIONARITY_01",
411       list(response_column = "walk", task = "stationarity"),
412       list("/stationarity/assessment" = exact("difference_candidate", "ADF cannot reject a unit root and KPSS rejects stationarity for a 600-step driftless random walk, at every ADF lag 0-12 and both KPSS lag rules.")),
413       c("Chapter 12: a differencing candidate is labelled only when ADF does not reject and KPSS rejects at 5 percent.",
414         "Verdict only: the manual does not define the lag selection, so the statistics themselves are not asserted; the verdict is checked to hold at every lag choice.",
415         "Until 2026-09-25 this case used TS_LONG_01's drifting level, whose correct verdict is inconclusive."),
416       "Analytic", "Dickey and Fuller (1979); Kwiatkowski et al. (1992)", oracle_class = "analytic")

Whole script generators/r/timeseries_models.R


White noise is stationary

Case STAT-TS-STATIONARITY-002 · status: passed

Reference

Evidence type
Analytical
Source
Analytic
Citation
Dickey and Fuller (1979); Kwiatkowski et al. (1992)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_STATIONARITY_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/stationarity/assessmentexactstationarystationarypassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

417  emit("STAT-TS-STATIONARITY-002", "timeSeriesStationarityMenuItem", "White noise is stationary", "TS_STATIONARITY_01",
418       list(response_column = "noise", task = "stationarity"),
419       list("/stationarity/assessment" = exact("stationary", "iid noise rejects a unit root and keeps level stationarity at every ADF lag 0-12 and both KPSS lag rules.")),
420       c("Chapter 12 combined criterion; verdict only, checked at every lag choice.",
421         "Until 2026-09-25 this case used GARCH(1,1) returns, which KPSS legitimately rejects."),
422       "Analytic", "Dickey and Fuller (1979); Kwiatkowski et al. (1992)", oracle_class = "analytic")

Whole script generators/r/timeseries_models.R


Linear trend with a two-step extrapolation

Case STAT-TS-TREND-001 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::nls on t = 1..n
Citation
R Core Team, stats package help: lm, nls, summary.lm, vcov
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_trend.R · sha256 d3ad3e09dcb6f002
Dataset
reference_data/datasets/TS_TREND_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/parameters/0/estimatetolerance4.79121739130433974.7912173913043561.59E-1414passed
/parameters/0/standard_errortolerance0.254860419830669980.254860419830669653.31E-1614passed
/parameters/1/estimatetolerance0.810222608695652550.81022260869565141.11E-1514passed
/parameters/1/standard_errortolerance0.0178365138347611540.0178365138347611342.05E-1714passed
/maetolerance0.481407246376812310.48140724637681193.84E-1615passed
/msetolerance0.335373875507247450.33537387550724677.73E-1614passed
/rmsetolerance0.579114734320624440.57911473432062386.70E-1614passed
/mapetolerance3.69320685787535033.6932068578753482.22E-1515passed
/fitted_countexact24240.00E+2passed
/series_lengthexact24240.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact000.00E+2passed
/extrapolation/0/0tolerance2525.00.00E+2passed
/extrapolation/0/1tolerance25.04678260869565425.0467826086956441.05E-1415passed
/extrapolation/1/1tolerance25.85700521739130825.857005217391291.81E-1415passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

24  emit <- function(id, title, config, outputs, notes, kind = "nominal", excluded = integer()) {
25    write_expected(family, id, list(
26      schema_version = "2.0.0", case_id = id, outputs = outputs,
27      expected_error = NULL, deferred_outputs = list(), generator_run = run))
28    write_case(family, id, list(
29      schema_version = "2.0.0", case_id = id, version = "1.0.0", family = family, entry_point = ep,
30      menu_items = list("trendAnalysisMenuItem"), title = title, case_kind = kind,
31      dataset = list(path = "reference_data/datasets/TS_TREND_01.json", sha256 = ""),
32      excluded_rows = as.list(as.integer(excluded)), configuration = config, "repeat" = 1L,
33      expected = list(path = paste0("reference_data/expected/", family, "/", id, ".json"), sha256 = ""),
34      oracle = list(oracle_class = "independent_implementation",
35                    source = "R stats::lm and stats::nls on t = 1..n",
36                    citation = "R Core Team, stats package help: lm, nls, summary.lm, vcov",
37                    generator = list(path = generator, sha256 = ""), independent = TRUE,
38                    convention_notes = as.list(c(notes,
39                      "Expected computed from the re-read TS_TREND_01 CSV with positions t = 1..n in table order.",
40                      "docs/manual/chapters/102-deterministic-trends.tex:3-6: positions t = 1..n in table order; excluded or missing responses retain their positions.",
41                      "docs/manual/chapters/102-deterministic-trends.tex:23-24: MAE, MSE, RMSE and MAPE describe in-sample original-scale errors."))),
42      supported = TRUE, unsupported_reason = NULL, requirement_ids = list("REQ-STAT-TS-TREND-001"),
43      risk_ids = list("RISK-REFERENCE-001"), review_reference = NULL))

Whole script generators/r/timeseries_trend.R


Quadratic trend

Case STAT-TS-TREND-002 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::nls on t = 1..n
Citation
R Core Team, stats package help: lm, nls, summary.lm, vcov
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_trend.R · sha256 d3ad3e09dcb6f002
Dataset
reference_data/datasets/TS_TREND_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/parameters/0/estimatetolerance19.25128112648221919.2512811264821279.24E-1414passed
/parameters/0/standard_errortolerance0.622221464249808950.62222146424980741.55E-1514passed
/parameters/1/estimatetolerance-1.4560429355427191-1.4560429355427081.11E-1414passed
/parameters/1/standard_errortolerance0.114679811788293590.114679811788293342.50E-1614passed
/parameters/2/estimatetolerance0.119607578291273970.119607578291273336.39E-1614passed
/parameters/2/standard_errortolerance0.00445334706012214060.0044533470601221328.64E-1814passed
/maetolerance0.727251812861052340.72725181286105745.11E-1514passed
/msetolerance0.760998406287562880.7609984062875628.90E-1614passed
/rmsetolerance0.872352226046086750.87235222604608634.40E-1615passed
/mapetolerance3.59803986706734723.5980398670673416.18E-1514passed
/fitted_countexact24240.00E+2passed
/series_lengthexact24240.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

100  fit <- stats::lm(y ~ tt + I(tt^2), data = data.frame(y = d$y_quad, tt = tt))
101  cf <- coef(fit); se <- sqrt(diag(vcov(fit)))
102  emit("STAT-TS-TREND-002", "Quadratic trend", list(response = "y_quad", model = "quadratic", horizon = 0L),
103       c(param(0L, cf[[1]], se[[1]], "b0"), param(1L, cf[[2]], se[[2]], "b1"), param(2L, cf[[3]], se[[3]], "b2"),
104         accuracy(d$y_quad, fitted(fit)), counts(nrow(d), nrow(d))),
105       c("lm(y ~ t + I(t^2)): ordinary least squares on 1, t, t^2; standard errors with residual df n - 3.",
106         "docs/manual/chapters/102-deterministic-trends.tex:8: linear and quadratic models use original-scale least squares."),
107       kind = "axis_coverage")

Whole script generators/r/timeseries_trend.R


Exponential growth, back-transformed log-linear fit

Case STAT-TS-TREND-003 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::nls on t = 1..n
Citation
R Core Team, stats package help: lm, nls, summary.lm, vcov
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_trend.R · sha256 d3ad3e09dcb6f002
Dataset
reference_data/datasets/TS_TREND_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/parameters/0/estimatetolerance2.92889388854168552.92889388854168782.25E-1515passed
/parameters/0/standard_errortolerance0.066881881994605930.066881881994606041.11E-1614passed
/parameters/1/estimatetolerance0.0903726735206236980.090372673520623571.25E-1614passed
/parameters/1/standard_errortolerance0.00159813119275152440.00159813119275152641.96E-1814passed
/maetolerance0.443860600908432050.44386060090842635.78E-1513passed
/msetolerance0.365637307542699570.36563730754269167.99E-1513passed
/rmsetolerance0.604679508122029530.60467950812202296.66E-1513passed
/mapetolerance4.36408282200351934.3640828220034922.76E-1414passed
/fitted_countexact24240.00E+2passed
/series_lengthexact24240.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact000.00E+2passed
/extrapolation/2/1tolerance33.605408603831333.6054086038312158.50E-1414passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

110  fit <- stats::lm(ly ~ tt, data = data.frame(ly = log(d$y_exp), tt = tt))
111  cf <- coef(fit); se <- sqrt(diag(vcov(fit)))
112  c_hat <- exp(cf[[1]]); d_hat <- cf[[2]]
113  fitted_exp <- c_hat * exp(d_hat * tt)
114  emit("STAT-TS-TREND-003", "Exponential growth, back-transformed log-linear fit",
115       list(response = "y_exp", model = "exponential", horizon = 3L),
116       c(param(0L, c_hat, c_hat * se[[1]], "c = exp(a)"), param(1L, d_hat, se[[2]], "d = b"),
117         accuracy(d$y_exp, fitted_exp), counts(nrow(d), nrow(d)),
118         list("/extrapolation/2/1" = num(c_hat * exp(d_hat * (nrow(d) + 3)), "c exp(d (n + 3)) from the log-linear fit."))),
119       c("lm(log(y) ~ t); c = exp(a) and d = b; SE(c) = exp(a) SE(a) (delta method) and SE(d) = SE(b); no bias adjustment.",
120         "Accuracy measures use the back-transformed fitted values c exp(d t) on the original scale.",
121         "docs/manual/chapters/102-deterministic-trends.tex:8-10: log-linear exponential growth fits log y = a + b t and displays e^a (e^b)^t, a median trend without bias adjustment."),
122       kind = "axis_coverage")

Whole script generators/r/timeseries_trend.R


Exponential growth, original-scale least squares

Case STAT-TS-TREND-004 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::nls on t = 1..n
Citation
R Core Team, stats package help: lm, nls, summary.lm, vcov
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_trend.R · sha256 d3ad3e09dcb6f002
Dataset
reference_data/datasets/TS_TREND_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/parameters/0/estimatetolerance3.11545636188406633.11545636193527645.12E-1110passed
/parameters/0/standard_errortolerance0.115128417964287240.115128418996017171.03E-98passed
/parameters/1/estimatetolerance0.0866415859812505050.086641585981242288.23E-1513passed
/parameters/1/standard_errortolerance0.00187466346551756350.00187466351068437754.52E-117passed
/maetolerance0.405365838848420260.4053658388503341.91E-1211passed
/msetolerance0.296276885387928120.29627688538792792.18E-1615passed
/rmsetolerance0.544313223601933240.5443132236019332.24E-1615passed
/mapetolerance4.50042622679925944.5004262270101132.11E-1010passed
/fitted_countexact24240.00E+2passed
/series_lengthexact24240.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact000.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

125  start_exp <- list(c = exp(cf[[1]]), d = cf[[2]])
126  fit <- stats::nls(y ~ c * exp(d * tt), data = data.frame(y = d$y_exp, tt = tt), start = start_exp,
127                    control = stats::nls.control(maxiter = 1000, tol = 1e-8, minFactor = 1e-14))
128  cf4 <- coef(fit); se4 <- sqrt(diag(vcov(fit)))
129  emit("STAT-TS-TREND-004", "Exponential growth, original-scale least squares",
130       list(response = "y_exp", model = "exponential_nls", horizon = 0L),
131       c(param(0L, cf4[["c"]], se4[["c"]], "c", "1e-6", "1e-5"), param(1L, cf4[["d"]], se4[["d"]], "d", "1e-6", "1e-5"),
132         accuracy(d$y_exp, fitted(fit), "1e-6"), counts(nrow(d), nrow(d))),
133       c("nls(y ~ c exp(d t)) Gauss-Newton from the log-linear start exp(a), b; tolerance 1e-6 (1e-5 for the standard errors) because both programs stop at an iterative optimum.",
134         "docs/manual/chapters/102-deterministic-trends.tex:11-12: the original-scale exponential minimizes squared response residuals for y = c e^(d t)."),
135       kind = "axis_coverage")

Whole script generators/r/timeseries_trend.R


S-curve three-parameter logistic trend

Case STAT-TS-TREND-005 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::nls on t = 1..n
Citation
R Core Team, stats package help: lm, nls, summary.lm, vcov
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_trend.R · sha256 d3ad3e09dcb6f002
Dataset
reference_data/datasets/TS_TREND_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/parameters/0/estimatetolerance100.38748466626321100.387484669843093.58E-910passed
/parameters/0/standard_errortolerance0.523863684671037830.52386368345833121.21E-98passed
/parameters/1/estimatetolerance19.38894658559878419.388946575725589.87E-99passed
/parameters/1/standard_errortolerance1.07362975747215491.07362969677551016.07E-87passed
/parameters/2/estimatetolerance0.342043451564515490.34204345150402176.05E-119passed
/parameters/2/standard_errortolerance0.00690875793690438110.006908757547067733.90E-107passed
/maetolerance0.892484404061700380.89248440459780255.36E-109passed
/msetolerance1.30587017302107441.30587017302107291.51E-1514passed
/rmsetolerance1.14274676679528381.1427467667952837.13E-1615passed
/mapetolerance2.45606381383364262.45606381387949174.58E-1110passed
/fitted_countexact24240.00E+2passed
/series_lengthexact24240.00E+2passed
/missing_countexact000.00E+2passed
/excluded_countexact000.00E+2passed
/extrapolation/0/1tolerance100.0125813678995100.012581371091343.19E-910passed
/pearl_reed/2/estimatetolerance19.38894658559878419.388946575725589.87E-99passed
/pearl_reed/3/estimatetolerance0.710317339638904530.71031733968187434.30E-1110passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

138  k0 <- 1.05 * max(d$y_logis)
139  z <- log(k0 / d$y_logis - 1)
140  lin <- coef(stats::lm(z ~ tt))
141  start_s <- list(K = k0, A = exp(lin[[1]]), r = -lin[[2]])
142  fit <- stats::nls(y ~ K / (1 + A * exp(-r * tt)), data = data.frame(y = d$y_logis, tt = tt), start = start_s,
143                    control = stats::nls.control(maxiter = 1000, tol = 1e-8, minFactor = 1e-14))
144  cf5 <- coef(fit); se5 <- sqrt(diag(vcov(fit)))
145  relative_k <- se5[["K"]] / cf5[["K"]]
146  if (relative_k > 0.25) stop("S-curve asymptote is weakly determined; choose different data")
147  outputs5 <- c(param(0L, cf5[["K"]], se5[["K"]], "K", "1e-6", "1e-5"),
148                param(1L, cf5[["A"]], se5[["A"]], "A", "1e-6", "1e-5"),
149                param(2L, cf5[["r"]], se5[["r"]], "r", "1e-6", "1e-5"),
150                accuracy(d$y_logis, fitted(fit), "1e-6"), counts(nrow(d), nrow(d)),
151                list("/extrapolation/0/1" = fit6(cf5[["K"]] / (1 + cf5[["A"]] * exp(-cf5[["r"]] * (nrow(d) + 1))),
152                                               "K / (1 + A exp(-r (n + 1))); SE(K)/K below 0.25 so the extrapolation is supplied."),
153                     "/pearl_reed/2/estimate" = fit6(cf5[["A"]], "Pearl-Reed beta1 = A."),
154                     "/pearl_reed/3/estimate" = fit6(exp(-cf5[["r"]]), "Pearl-Reed beta2 = exp(-r).")))
155  emit("STAT-TS-TREND-005", "S-curve three-parameter logistic trend",
156       list(response = "y_logis", model = "scurve", horizon = 1L), outputs5,
157       c("nls(y ~ K / (1 + A exp(-r t))) Gauss-Newton from the linearized start K0 = 1.05 max(y), log(K0/y - 1) ~ t; tolerance 1e-6 (1e-5 for the standard errors).",
158         sprintf("SE(K)/K = %.4g from vcov(nls), below the 0.25 screen, so the one-step extrapolation is expected.", relative_k),
159         "docs/manual/chapters/102-deterministic-trends.tex:15-20: logistic S-curve y = K/(1 + A e^(-rt)); extrapolation withheld when asymptote relative SE exceeds 0.25."),
160       kind = "axis_coverage")

Whole script generators/r/timeseries_trend.R


Linear trend with a missing response and an excluded row

Case STAT-TS-TREND-006 · status: passed

Reference

Evidence type
R
Source
R stats::lm and stats::nls on t = 1..n
Citation
R Core Team, stats package help: lm, nls, summary.lm, vcov
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_trend.R · sha256 d3ad3e09dcb6f002
Dataset
reference_data/datasets/TS_TREND_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/parameters/0/estimatetolerance4.77412275850851044.7741227585085261.60E-1414passed
/parameters/0/standard_errortolerance0.271329770071724210.271329770071723764.46E-1614passed
/parameters/1/estimatetolerance0.809965477981539730.80996547798153961.13E-1615passed
/parameters/1/standard_errortolerance0.0185496913632072390.0185496913632072073.15E-1714passed
/maetolerance0.495117461361881060.49511746136187931.77E-1514passed
/msetolerance0.349678974567782460.34967897456778141.05E-1514passed
/rmsetolerance0.591336600057684960.59133660005768418.90E-1614passed
/mapetolerance3.78818854592183433.78818854592182361.07E-1414passed
/fitted_countexact22220.00E+2passed
/series_lengthexact24240.00E+2passed
/missing_countexact110.00E+2passed
/excluded_countexact110.00E+2passed
/extrapolation/0/0tolerance2525.00.00E+2passed
/extrapolation/0/1tolerance25.02325970804700425.0232597080470141.03E-1415passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

163  keep <- !is.na(d$y_gap)
164  keep[12] <- FALSE  # 0-based row index 11 is excluded by the case
165  fit <- stats::lm(y ~ tt, data = data.frame(y = d$y_gap[keep], tt = tt[keep]))
166  cf <- coef(fit); se <- sqrt(diag(vcov(fit)))
167  emit("STAT-TS-TREND-006", "Linear trend with a missing response and an excluded row",
168       list(response = "y_gap", model = "linear", horizon = 1L),
169       c(param(0L, cf[[1]], se[[1]], "Intercept b0"), param(1L, cf[[2]], se[[2]], "Slope b1"),
170         accuracy(d$y_gap[keep], fitted(fit)), counts(sum(keep), nrow(d), 1L, 1L),
171         list("/extrapolation/0/0" = num(nrow(d) + 1, "Extrapolation starts at n + 1 counting the kept positions."),
172              "/extrapolation/0/1" = num(cf[[1]] + cf[[2]] * (nrow(d) + 1), "b0 + b1 (n + 1)."))),
173       c("lm(y ~ t) on the 22 fitted rows with t at their original positions (row 7 missing, row 12 excluded); the series length stays 24.",
174         "docs/manual/chapters/102-deterministic-trends.tex:4-6: excluded or missing responses retain their positions, so removing a value does not compress elapsed index time."),
175       kind = "axis_coverage", excluded = 11L)

Whole script generators/r/timeseries_trend.R


Rolling-origin naive accuracy

Case STAT-TS-VALIDATE-001 · status: passed

Reference

Evidence type
Analytical
Source
Analytic
Citation
Hyndman and Athanasopoulos, Forecasting: Principles and Practice, section 5.10
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_MODELS_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/accuracy_by_horizon/0/countexact12120.00E+2passed
/accuracy_by_horizon/0/metrics/maetolerance0.794836243508699520.79483624350869954.80E-1817passed
/accuracy_by_horizon/2/metrics/maetolerance1.59099259000726321.59099259000726322.99E-1716passed
/accuracy_by_horizon/1/metrics/rmsetolerance1.33767260825470881.33767260825470885.18E-1817passed
/overall_metrics/maetolerance1.14115219690942231.14115219690942233.27E-1716passed
/final_evaluation_metrics/maetolerance1.99784287415174581.99784287415174332.47E-1514passed
/final_evaluation_metrics/rmsetolerance2.06337474091164722.0633747409116443.12E-1514passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

425  yy <- d$rw
426  origins <- 40:51  # 0-based origins; the last 6 rows are the final evaluation
427  err <- sapply(1:3, function(k) sapply(origins, function(o) yy[o + k] - yy[o]))
428  final_err <- yy[55:60] - yy[54]
429  emit("STAT-TS-VALIDATE-001", "timeSeriesValidateMenuItem", "Rolling-origin naive accuracy", "TS_MODELS_01",
430       list(method = "study", forecast = list(response_column = "rw", model = "naive", horizon = 3L),
431            validation = list(initial_window = 40L, horizon = 3L, final_evaluation = 6L)),
432       list("/accuracy_by_horizon/0/count" = exact(12L, "Origins 40 to 51 fit before the six final rows."),
433            "/accuracy_by_horizon/0/metrics/mae" = num(mean(abs(err[, 1])), "Mean absolute one-step error."),
434            "/accuracy_by_horizon/2/metrics/mae" = num(mean(abs(err[, 3])), "Mean absolute three-step error."),
435            "/accuracy_by_horizon/1/metrics/rmse" = num(sqrt(mean(err[, 2]^2)), "Root mean squared two-step error."),
436            "/overall_metrics/mae" = num(mean(abs(err)), "Every rolling error."),
437            "/final_evaluation_metrics/mae" = num(mean(abs(final_err)), "Six-step forecast from the last training row."),
438            "/final_evaluation_metrics/rmse" = num(sqrt(mean(final_err^2)), "Final evaluation RMSE.")),
439       c("Chapter 12: an expanding window from 40 initial observations, step 1, horizon 3, and an untouched final 6.",
440         "MASE, sMAPE, bias sign and interval coverage depend on conventions the manual leaves open and are not asserted."),
441       "Analytic", "Hyndman and Athanasopoulos, Forecasting: Principles and Practice, section 5.10", oracle_class = "analytic")

Whole script generators/r/timeseries_models.R


Level VAR(1) forecasts and impulse response

Case STAT-TS-VAR-001 · status: passed

Reference

Evidence type
R
Source
R vars package
Citation
vars::VAR, predict.varest and irf documentation (Pfaff 2008)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_SYSTEM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/model_labelexactLevel VAR (1 lag(s))Level VAR (1 lag(s))passed
/forecasts/0/v1tolerance1.64000354994172691.64000354994172654.21E-1615passed
/forecasts/0/v2tolerance0.0195404182104870210.0195404182104875765.55E-1613passed
/forecasts/0/v1_lowertolerance-0.29553969794459101-0.29553969794459198.85E-1614passed
/forecasts/0/v2_uppertolerance1.25156393161187831.25156393161187928.77E-1615passed
/forecasts/1/v1tolerance2.00891740182511042.00891740182510959.28E-1615passed
/forecasts/1/v2tolerance-0.12344868326392272-0.123448683263922116.10E-1614passed
/forecasts/1/v1_lowertolerance-0.18896403853233146-0.188964038532333231.77E-1514passed
/forecasts/1/v2_uppertolerance1.25292992440430311.2529299244043048.52E-1615passed
/impulse_responses/0/shockexactv1v1passed
/impulse_responses/0/responseexactv1v1passed
/impulse_responses/0/valuetolerance0.987540211531250930.98754021153125093.24E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

273  ir <- irf(vfit, impulse = "v1", response = "v1", n.ahead = 2, ortho = TRUE, boot = FALSE)$irf$v1
274  vo[["/impulse_responses/0/shock"]] <- exact("v1", "First impulse.")
275  vo[["/impulse_responses/0/response"]] <- exact("v1", "First response.")
276  vo[["/impulse_responses/0/value"]] <- num(ir[1, 1], "vars::irf orthogonalised response at horizon 0 (Cholesky factor).")
277  emit("STAT-TS-VAR-001", "timeSeriesVarVecmMenuItem", "Level VAR(1) forecasts and impulse response", "TS_SYSTEM_01",
278       list(method = "advanced", kind = "var_vecm", columns = list("v1", "v2"), lag_order = 1L, cointegration_rank = 2L,
279            horizon = 2L, deterministic = "constant"),
280       vo, c("VAR(1) with an unrestricted constant, fitted equation by equation by OLS.",
281             "Interval and impulse scale use the residual covariance; vars divides by T - Kp - 1 as the degrees-of-freedom-adjusted estimate."),
282       "R vars package", "vars::VAR, predict.varest and irf documentation (Pfaff 2008)")

Whole script generators/r/timeseries_models.R


Rank zero: a constant-only VAR in differences

Case STAT-TS-VAR-002 · status: passed

Reference

Evidence type
Analytical
Source
Analytic
Citation
Luetkepohl (2005), New Introduction to Multiple Time Series Analysis, section 6.3
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_SYSTEM_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/forecasts/0/v1tolerance1.24511452937522221.24511452937522224.34E-1716passed
/forecasts/1/v2tolerance0.50045256477853450.50045256477853453.28E-1817passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

284  dv <- apply(cbind(s$v1, s$v2), 2, diff)
285  mu <- colMeans(dv)
286  last <- c(tail(s$v1, 1), tail(s$v2, 1))
287  emit("STAT-TS-VAR-002", "timeSeriesVarVecmMenuItem", "Rank zero: a constant-only VAR in differences", "TS_SYSTEM_01",
288       list(method = "advanced", kind = "var_vecm", columns = list("v1", "v2"), lag_order = 1L, cointegration_rank = 0L,
289            horizon = 2L, deterministic = "constant"),
290       list("/forecasts/0/v1" = num(last[1] + mu[1], "Last level plus the mean difference."),
291            "/forecasts/1/v2" = num(last[2] + 2 * mu[2], "Two steps of the mean difference.")),
292       c("Chapter 12: the lag control counts level lags, so one level lag at rank zero fits zero lags of differences, leaving only the constant."),
293       "Analytic", "Luetkepohl (2005), New Introduction to Multiple Time Series Analysis, section 6.3",
294       kind = "boundary", oracle_class = "analytic")

Whole script generators/r/timeseries_models.R


A delayed copy peaks at its delay

Case STAT-TS-XCORR-001 · status: passed

Reference

Evidence type
R
Source
R stats::cor on explicitly aligned lag pairs
Citation
R stats::cor and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_cross_correlation.R · sha256 024b06115a56c162
Dataset
reference_data/datasets/TS_XCORR_SHIFT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/lag_rows/0/lagexact-8-80.00E+2passed
/lag_rows/0/correlationtolerance0.235064011502449740.235064011502449631.13E-1615passed
/lag_rows/0/overlapexact52520.00E+2passed
/lag_rows/0/inside_bandexactFalseFalsepassed
/lag_rows/1/lagexact-7-70.00E+2passed
/lag_rows/1/correlationtolerance0.3269587394774760.3269587394774761.41E-1817passed
/lag_rows/1/overlapexact53530.00E+2passed
/lag_rows/1/inside_bandexactFalseFalsepassed
/lag_rows/2/lagexact-6-60.00E+2passed
/lag_rows/2/correlationtolerance0.510805910210714090.51080591021071413.03E-1817passed
/lag_rows/2/overlapexact54540.00E+2passed
/lag_rows/2/inside_bandexactFalseFalsepassed
/lag_rows/3/lagexact-5-50.00E+2passed
/lag_rows/3/correlationtolerance0.44912689355475460.44912689355475471.14E-1615passed
/lag_rows/3/overlapexact55550.00E+2passed
/lag_rows/3/inside_bandexactFalseFalsepassed
/lag_rows/4/lagexact-4-40.00E+2passed
/lag_rows/4/correlationtolerance0.356947043341169830.356947043341169834.02E-1816passed
/lag_rows/4/overlapexact56560.00E+2passed
/lag_rows/4/inside_bandexactFalseFalsepassed
/lag_rows/5/lagexact-3-30.00E+2passed
/lag_rows/5/correlationtolerance0.326088558511963330.32608855851196345.89E-1715passed
/lag_rows/5/overlapexact57570.00E+2passed
/lag_rows/5/inside_bandexactFalseFalsepassed
/lag_rows/6/lagexact-2-20.00E+2passed
/lag_rows/6/correlationtolerance0.456197053042572630.456197053042572631.12E-1817passed
/lag_rows/6/overlapexact58580.00E+2passed
/lag_rows/6/inside_bandexactFalseFalsepassed
/lag_rows/7/lagexact-1-10.00E+2passed
/lag_rows/7/correlationtolerance0.540371337568610220.54037133756861021.60E-1817passed
/lag_rows/7/overlapexact59590.00E+2passed
/lag_rows/7/inside_bandexactFalseFalsepassed
/lag_rows/8/lagexact000.00E+2passed
/lag_rows/8/correlationtolerance0.499381878663106850.49938187866310695.40E-1715passed
/lag_rows/8/overlapexact60600.00E+2passed
/lag_rows/8/inside_bandexactFalseFalsepassed
/lag_rows/9/lagexact110.00E+2passed
/lag_rows/9/correlationtolerance0.522027081273694130.52202708127369421.08E-1615passed
/lag_rows/9/overlapexact59590.00E+2passed
/lag_rows/9/inside_bandexactFalseFalsepassed
/lag_rows/10/lagexact220.00E+2passed
/lag_rows/10/correlationtolerance0.640422359736301240.64042235973630111.12E-1615passed
/lag_rows/10/overlapexact58580.00E+2passed
/lag_rows/10/inside_bandexactFalseFalsepassed
/lag_rows/11/lagexact330.00E+2passed
/lag_rows/11/correlationtolerance0.978181897984164880.97818189798416512.25E-1615passed
/lag_rows/11/overlapexact57570.00E+2passed
/lag_rows/11/inside_bandexactFalseFalsepassed
/lag_rows/12/lagexact440.00E+2passed
/lag_rows/12/correlationtolerance0.658989115528283990.6589891155282843.70E-1817passed
/lag_rows/12/overlapexact56560.00E+2passed
/lag_rows/12/inside_bandexactFalseFalsepassed
/lag_rows/13/lagexact550.00E+2passed
/lag_rows/13/correlationtolerance0.555246820828212970.55524682082821291.09E-1615passed
/lag_rows/13/overlapexact55550.00E+2passed
/lag_rows/13/inside_bandexactFalseFalsepassed
/lag_rows/14/lagexact660.00E+2passed
/lag_rows/14/correlationtolerance0.494381025431530230.49438102543153035.55E-1715passed
/lag_rows/14/overlapexact54540.00E+2passed
/lag_rows/14/inside_bandexactFalseFalsepassed
/lag_rows/15/lagexact770.00E+2passed
/lag_rows/15/correlationtolerance0.537176353346769050.5371763533467693.23E-1817passed
/lag_rows/15/overlapexact53530.00E+2passed
/lag_rows/15/inside_bandexactFalseFalsepassed
/lag_rows/16/lagexact880.00E+2passed
/lag_rows/16/correlationtolerance0.493141506439267690.49314150643926774.98E-1816passed
/lag_rows/16/overlapexact52520.00E+2passed
/lag_rows/16/inside_bandexactFalseFalsepassed
/omitted_lag_countexact000.00E+2passed
/peak_lagexact330.00E+2passed
/peak_labelexactx leads y by 3x leads y by 3passed
/max_lag_usedexact880.00E+2passed
/band_basisexactnominalnominalpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

 83  set.seed(20260930)
 84  n <- 60L
 85  x <- round(cumsum(rnorm(n + 3)) * 0.5 + rnorm(n + 3), 6)
 86  y <- round(c(rep(0, 3), x[1:n])[1:n] + rnorm(n, 0, 0.3), 6)
 87  x <- x[1:n]
 88  export_dataset(data.frame(t = seq_len(n), x = x, y = y), "TS_XCORR_SHIFT",
 89                 "Constructed: y(t) = x(t - 3) + N(0, 0.3^2) noise, seed 20260930, rounded to 6 decimals",
 90                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
 91                 c(t = "Int64", x = "Float64", y = "Float64"))
 92  d <- read_dataset("TS_XCORR_SHIFT")
 93  rows <- lag_table(d$x, d$y, 8)
 94  out <- c(lag_outputs(rows), list(
 95    "/peak_label" = exact("x leads y by 3", "A positive peak is labelled 'X leads Y by d' (line 32-34)."),
 96    "/max_lag_used" = exact(8L, "The requested maximum lag."),
 97    "/band_basis" = exact("nominal", "No prewhitening.")))
 98  emit("STAT-TS-XCORR-001", "A delayed copy peaks at its delay", "TS_XCORR_SHIFT",
 99       list(series_x = "x", series_y = "y", max_lag = 8L), out,
100       c("Manual line 81-83: a synthetic shifted pair y(t) = x(t - d) peaks at d with the 'X leads Y' label."))

Whole script generators/r/timeseries_cross_correlation.R


First differences of both series

Case STAT-TS-XCORR-002 · status: passed

Reference

Evidence type
R
Source
R stats::cor on explicitly aligned lag pairs
Citation
R stats::cor and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_cross_correlation.R · sha256 024b06115a56c162
Dataset
reference_data/datasets/TS_XCORR_SHIFT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/lag_rows/0/lagexact-5-50.00E+2passed
/lag_rows/0/correlationtolerance0.0495710400198947150.0495710400198946942.12E-1715passed
/lag_rows/0/overlapexact54540.00E+2passed
/lag_rows/0/inside_bandexactTrueTruepassed
/lag_rows/1/lagexact-4-40.00E+2passed
/lag_rows/1/correlationtolerance-0.04917862293905096-0.049178622939050956.72E-1815passed
/lag_rows/1/overlapexact55550.00E+2passed
/lag_rows/1/inside_bandexactTrueTruepassed
/lag_rows/2/lagexact-3-30.00E+2passed
/lag_rows/2/correlationtolerance-0.24075160154174607-0.240751601541746025.13E-1715passed
/lag_rows/2/overlapexact56560.00E+2passed
/lag_rows/2/inside_bandexactFalseFalsepassed
/lag_rows/3/lagexact-2-20.00E+2passed
/lag_rows/3/correlationtolerance0.0575176086645714030.0575176086645713752.81E-1715passed
/lag_rows/3/overlapexact57570.00E+2passed
/lag_rows/3/inside_bandexactTrueTruepassed
/lag_rows/4/lagexact-1-10.00E+2passed
/lag_rows/4/correlationtolerance0.191557445736856790.191557445736856822.84E-1715passed
/lag_rows/4/overlapexact58580.00E+2passed
/lag_rows/4/inside_bandexactFalseFalsepassed
/lag_rows/5/lagexact000.00E+2passed
/lag_rows/5/correlationtolerance-0.12662860425340947-0.12662860425340952.36E-1715passed
/lag_rows/5/overlapexact59590.00E+2passed
/lag_rows/5/inside_bandexactTrueTruepassed
/lag_rows/6/lagexact110.00E+2passed
/lag_rows/6/correlationtolerance-0.11918104264197839-0.119181042641978335.75E-1715passed
/lag_rows/6/overlapexact58580.00E+2passed
/lag_rows/6/inside_bandexactTrueTruepassed
/lag_rows/7/lagexact220.00E+2passed
/lag_rows/7/correlationtolerance-0.33259030985867177-0.332590309858671835.87E-1715passed
/lag_rows/7/overlapexact57570.00E+2passed
/lag_rows/7/inside_bandexactFalseFalsepassed
/lag_rows/8/lagexact330.00E+2passed
/lag_rows/8/correlationtolerance0.946234750736571820.94623475073657234.41E-1615passed
/lag_rows/8/overlapexact56560.00E+2passed
/lag_rows/8/inside_bandexactFalseFalsepassed
/lag_rows/9/lagexact440.00E+2passed
/lag_rows/9/correlationtolerance-0.29858296864565903-0.29858296864565915.59E-1715passed
/lag_rows/9/overlapexact55550.00E+2passed
/lag_rows/9/inside_bandexactFalseFalsepassed
/lag_rows/10/lagexact550.00E+2passed
/lag_rows/10/correlationtolerance-0.069718968928631053-0.069718968928631032.75E-1715passed
/lag_rows/10/overlapexact54540.00E+2passed
/lag_rows/10/inside_bandexactTrueTruepassed
/omitted_lag_countexact000.00E+2passed
/peak_lagexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

102  emit("STAT-TS-XCORR-002", "First differences of both series", "TS_XCORR_SHIFT",
103       list(series_x = "x", series_y = "y", max_lag = 5L, difference_order_x = 1L, difference_order_y = 1L),
104       lag_outputs(lag_table(difference(d$x, 1), difference(d$y, 1), 5)),
105       c("Both series are differenced; the leading position is unavailable (line 58-60)."), kind = "axis_coverage")

Whole script generators/r/timeseries_cross_correlation.R


Prewhitening with an OLS AR(2) fit

Case STAT-TS-XCORR-003 · status: passed

Reference

Evidence type
R
Source
R stats::cor on explicitly aligned lag pairs
Citation
R stats::cor and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_cross_correlation.R · sha256 024b06115a56c162
Dataset
reference_data/datasets/TS_XCORR_SHIFT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/lag_rows/0/lagexact-5-50.00E+2passed
/lag_rows/0/correlationtolerance0.168849632606778180.16884963260677831.14E-1615passed
/lag_rows/0/overlapexact53530.00E+2passed
/lag_rows/0/inside_bandexactFalseFalsepassed
/lag_rows/1/lagexact-4-40.00E+2passed
/lag_rows/1/correlationtolerance-0.071440913136184378-0.071440913136184241.39E-1614passed
/lag_rows/1/overlapexact54540.00E+2passed
/lag_rows/1/inside_bandexactTrueTruepassed
/lag_rows/2/lagexact-3-30.00E+2passed
/lag_rows/2/correlationtolerance-0.21289627226994903-0.212896272269948958.04E-1715passed
/lag_rows/2/overlapexact55550.00E+2passed
/lag_rows/2/inside_bandexactFalseFalsepassed
/lag_rows/3/lagexact-2-20.00E+2passed
/lag_rows/3/correlationtolerance0.125159242251393190.12515924225139331.09E-1615passed
/lag_rows/3/overlapexact56560.00E+2passed
/lag_rows/3/inside_bandexactTrueTruepassed
/lag_rows/4/lagexact-1-10.00E+2passed
/lag_rows/4/correlationtolerance0.240611282788472210.240611282788472321.15E-1615passed
/lag_rows/4/overlapexact57570.00E+2passed
/lag_rows/4/inside_bandexactFalseFalsepassed
/lag_rows/5/lagexact000.00E+2passed
/lag_rows/5/correlationtolerance-0.050071728571962994-0.0500717285719627862.08E-1614passed
/lag_rows/5/overlapexact58580.00E+2passed
/lag_rows/5/inside_bandexactTrueTruepassed
/lag_rows/6/lagexact110.00E+2passed
/lag_rows/6/correlationtolerance-0.20164258466340879-0.201642584663408735.96E-1715passed
/lag_rows/6/overlapexact57570.00E+2passed
/lag_rows/6/inside_bandexactFalseFalsepassed
/lag_rows/7/lagexact220.00E+2passed
/lag_rows/7/correlationtolerance0.0142914963920898610.0142914963920899195.76E-1714passed
/lag_rows/7/overlapexact56560.00E+2passed
/lag_rows/7/inside_bandexactTrueTruepassed
/lag_rows/8/lagexact330.00E+2passed
/lag_rows/8/correlationtolerance0.951901050377940040.95190105037794043.32E-1615passed
/lag_rows/8/overlapexact55550.00E+2passed
/lag_rows/8/inside_bandexactFalseFalsepassed
/lag_rows/9/lagexact440.00E+2passed
/lag_rows/9/correlationtolerance-0.024753636037684483-0.024753636037684113.75E-1613passed
/lag_rows/9/overlapexact54540.00E+2passed
/lag_rows/9/inside_bandexactTrueTruepassed
/lag_rows/10/lagexact550.00E+2passed
/lag_rows/10/correlationtolerance-0.11329104364519434-0.113291043645194132.12E-1614passed
/lag_rows/10/overlapexact53530.00E+2passed
/lag_rows/10/inside_bandexactTrueTruepassed
/omitted_lag_countexact000.00E+2passed
/peak_lagexact330.00E+2passed
/band_basisexactprewhitenedprewhitenedpassed
/prewhiten_orderexact220.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

107  wx <- ar_residuals(difference(d$x, 0), 2); wy <- ar_residuals(difference(d$y, 0), 2)
108  out <- lag_outputs(lag_table(wx, wy, 5))
109  out[["/band_basis"]] <- exact("prewhitened", "Bands are computed on the AR residuals (line 48-50).")
110  out[["/prewhiten_order"]] <- exact(2L, "The declared AR order.")
111  emit("STAT-TS-XCORR-003", "Prewhitening with an OLS AR(2) fit", "TS_XCORR_SHIFT",
112       list(series_x = "x", series_y = "y", max_lag = 5L, prewhiten = TRUE, prewhiten_order = 2L), out,
113       c("Each series is replaced by the residuals of lm(x_t ~ x_{t-1} + x_{t-2}) with an intercept over its gap-free block.",
114         "The manual says 'a deterministic OLS AR(p) fit' without stating the intercept; this reference includes one."),
115       kind = "axis_coverage")

Whole script generators/r/timeseries_cross_correlation.R


Missing and excluded rows stay as unavailable positions

Case STAT-TS-XCORR-004 · status: passed

Reference

Evidence type
R
Source
R stats::cor on explicitly aligned lag pairs
Citation
R stats::cor and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_cross_correlation.R · sha256 024b06115a56c162
Dataset
reference_data/datasets/TS_XCORR_MISSING.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/lag_rows/0/lagexact-6-60.00E+2passed
/lag_rows/0/correlationtolerance0.471081601468697650.47108160146869765.28E-1715passed
/lag_rows/0/overlapexact48480.00E+2passed
/lag_rows/0/inside_bandexactFalseFalsepassed
/lag_rows/1/lagexact-5-50.00E+2passed
/lag_rows/1/correlationtolerance0.421621556051539730.42162155605153974.63E-1816passed
/lag_rows/1/overlapexact50500.00E+2passed
/lag_rows/1/inside_bandexactFalseFalsepassed
/lag_rows/2/lagexact-4-40.00E+2passed
/lag_rows/2/correlationtolerance0.317822297272127070.3178222972721275.10E-1715passed
/lag_rows/2/overlapexact49490.00E+2passed
/lag_rows/2/inside_bandexactFalseFalsepassed
/lag_rows/3/lagexact-3-30.00E+2passed
/lag_rows/3/correlationtolerance0.304831577744172630.30483157774417265.29E-1715passed
/lag_rows/3/overlapexact50500.00E+2passed
/lag_rows/3/inside_bandexactFalseFalsepassed
/lag_rows/4/lagexact-2-20.00E+2passed
/lag_rows/4/correlationtolerance0.415567350034175720.415567350034175665.57E-1715passed
/lag_rows/4/overlapexact51510.00E+2passed
/lag_rows/4/inside_bandexactFalseFalsepassed
/lag_rows/5/lagexact-1-10.00E+2passed
/lag_rows/5/correlationtolerance0.538669803218762230.53866980321876223.14E-1817passed
/lag_rows/5/overlapexact52520.00E+2passed
/lag_rows/5/inside_bandexactFalseFalsepassed
/lag_rows/6/lagexact000.00E+2passed
/lag_rows/6/correlationtolerance0.531072362868722720.53107236286872281.16E-1615passed
/lag_rows/6/overlapexact55550.00E+2passed
/lag_rows/6/inside_bandexactFalseFalsepassed
/lag_rows/7/lagexact110.00E+2passed
/lag_rows/7/correlationtolerance0.570922540995668190.57092254099566831.14E-1615passed
/lag_rows/7/overlapexact52520.00E+2passed
/lag_rows/7/inside_bandexactFalseFalsepassed
/lag_rows/8/lagexact220.00E+2passed
/lag_rows/8/correlationtolerance0.658826331465650060.65882633146565011.33E-1817passed
/lag_rows/8/overlapexact51510.00E+2passed
/lag_rows/8/inside_bandexactFalseFalsepassed
/lag_rows/9/lagexact330.00E+2passed
/lag_rows/9/correlationtolerance0.978464530539545850.97846453053954553.32E-1615passed
/lag_rows/9/overlapexact50500.00E+2passed
/lag_rows/9/inside_bandexactFalseFalsepassed
/lag_rows/10/lagexact440.00E+2passed
/lag_rows/10/correlationtolerance0.680384079341247230.68038407934124722.35E-1817passed
/lag_rows/10/overlapexact49490.00E+2passed
/lag_rows/10/inside_bandexactFalseFalsepassed
/lag_rows/11/lagexact550.00E+2passed
/lag_rows/11/correlationtolerance0.651240083364866050.65124008336486621.12E-1615passed
/lag_rows/11/overlapexact49490.00E+2passed
/lag_rows/11/inside_bandexactFalseFalsepassed
/lag_rows/12/lagexact660.00E+2passed
/lag_rows/12/correlationtolerance0.541066890325356930.54106689032535696.24E-1917passed
/lag_rows/12/overlapexact48480.00E+2passed
/lag_rows/12/inside_bandexactFalseFalsepassed
/omitted_lag_countexact000.00E+2passed
/peak_lagexact330.00E+2passed
/excluded_countexact220.00E+2passed
/source_countexact60600.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

118  m <- read_dataset("TS_XCORR_SHIFT")
119  m$x[c(10, 31)] <- NA
120  m$y[45] <- NA
121  export_dataset(m, "TS_XCORR_MISSING", "TS_XCORR_SHIFT with x missing at rows 10 and 31 and y at row 45 (1-based)",
122                 "R stats::rnorm documentation", "Constructed validation data; no third-party license.",
123                 c(t = "Int64", x = "Float64", y = "Float64"))
124  mm <- read_dataset("TS_XCORR_MISSING")
125  excluded <- c(4L, 20L)
126  xa <- mm$x; ya <- mm$y; xa[excluded + 1L] <- NA; ya[excluded + 1L] <- NA
127  out <- c(lag_outputs(lag_table(xa, ya, 6)), list(
128    "/excluded_count" = exact(2L, "Two excluded rows."),
129    "/source_count" = exact(n, "Every source row is counted.")))
130  emit("STAT-TS-XCORR-004", "Missing and excluded rows stay as unavailable positions", "TS_XCORR_MISSING",
131       list(series_x = "x", series_y = "y", max_lag = 6L), out,
132       c("Rows 4 and 20 (0-based) are excluded and three values are missing; the series never closes up (line 61-65)."),
133       kind = "missing_excluded", excluded = excluded)

Whole script generators/r/timeseries_cross_correlation.R


Lags with fewer than three pairs are omitted

Case STAT-TS-XCORR-005 · status: passed

Reference

Evidence type
R
Source
R stats::cor on explicitly aligned lag pairs
Citation
R stats::cor and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_cross_correlation.R · sha256 024b06115a56c162
Dataset
reference_data/datasets/TS_XCORR_SHORT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/lag_rows/0/lagexact-3-30.00E+2passed
/lag_rows/0/correlationtolerance0.984155544660445080.98415554466044517.51E-1918passed
/lag_rows/0/overlapexact330.00E+2passed
/lag_rows/0/inside_bandexactFalseFalsepassed
/lag_rows/1/lagexact-2-20.00E+2passed
/lag_rows/1/correlationtolerance-0.0097197063478077599-0.009719706347807751.05E-1714passed
/lag_rows/1/overlapexact440.00E+2passed
/lag_rows/1/inside_bandexactTrueTruepassed
/lag_rows/2/lagexact-1-10.00E+2passed
/lag_rows/2/correlationtolerance0.993851040344959080.99385104034495912.08E-1918passed
/lag_rows/2/overlapexact550.00E+2passed
/lag_rows/2/inside_bandexactFalseFalsepassed
/lag_rows/3/lagexact000.00E+2passed
/lag_rows/3/correlationtolerance0.345438239455572950.3454382394555735.40E-1715passed
/lag_rows/3/overlapexact660.00E+2passed
/lag_rows/3/inside_bandexactTrueTruepassed
/lag_rows/4/lagexact110.00E+2passed
/lag_rows/4/correlationtolerance0.990932917153488970.9909329171534892.02E-1817passed
/lag_rows/4/overlapexact550.00E+2passed
/lag_rows/4/inside_bandexactFalseFalsepassed
/lag_rows/5/lagexact220.00E+2passed
/lag_rows/5/correlationtolerance-0.1921663988801095-0.192166398880109472.64E-1715passed
/lag_rows/5/overlapexact440.00E+2passed
/lag_rows/5/inside_bandexactTrueTruepassed
/lag_rows/6/lagexact330.00E+2passed
/lag_rows/6/correlationtolerance0.995617648003896290.99561764800389633.93E-1817passed
/lag_rows/6/overlapexact330.00E+2passed
/lag_rows/6/inside_bandexactFalseFalsepassed
/omitted_lag_countexact440.00E+2passed
/peak_lagexact330.00E+2passed
/max_lag_usedexact550.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

136  export_dataset(data.frame(x = c(1.2, 3.4, 2.2, 5.1, 4.0, 6.3), y = c(2.0, 1.1, 3.9, 2.7, 5.5, 4.2)),
137                 "TS_XCORR_SHORT", "Constructed: six paired values", "Hand-constructed",
138                 "Constructed validation data; no third-party license.", c(x = "Float64", y = "Float64"))
139  s <- read_dataset("TS_XCORR_SHORT")
140  out <- c(lag_outputs(lag_table(s$x, s$y, 5)),
141           list("/max_lag_used" = exact(5L, "The lag range is capped at n - 1 = 5.")))
142  emit("STAT-TS-XCORR-005", "Lags with fewer than three pairs are omitted", "TS_XCORR_SHORT",
143       list(series_x = "x", series_y = "y", max_lag = 30L), out,
144       c("With n = 6 the lags |k| >= 4 have at most 2 pairs and are omitted (line 35-36); max_lag 30 is reduced to 5."),
145       kind = "boundary")

Whole script generators/r/timeseries_cross_correlation.R


Two rows give an insufficient-data outcome

Case STAT-TS-XCORR-006 · status: passed

Reference

Evidence type
R
Source
R stats::cor on explicitly aligned lag pairs
Citation
R stats::cor and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_cross_correlation.R · sha256 024b06115a56c162
Dataset
reference_data/datasets/TS_XCORR_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statusexactinsufficient_datainsufficient_datapassed
/peak_lagabsentNonepassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

146  export_dataset(data.frame(x = c(1, 2), y = c(3, 1)), "TS_XCORR_TWO", "Constructed: two paired values",
147                 "Hand-constructed", "Constructed validation data; no third-party license.", c(x = "Float64", y = "Float64"))
148  emit("STAT-TS-XCORR-006", "Two rows give an insufficient-data outcome", "TS_XCORR_TWO",
149       list(series_x = "x", series_y = "y", max_lag = 1L),
150       list("/status" = exact("insufficient_data", "No lag reaches 3 pairs (line 77-78)."),
151            "/peak_lag" = list(comparison = "absent", rationale = "No lags are invented.")),
152       c("Two rows give at most two pairs at every lag."), kind = "boundary")

Whole script generators/r/timeseries_cross_correlation.R


A 1e8 offset leaves the correlations unchanged

Case STAT-TS-XCORR-007 · status: passed

Reference

Evidence type
R
Source
R stats::cor on explicitly aligned lag pairs
Citation
R stats::cor and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_cross_correlation.R · sha256 024b06115a56c162
Dataset
reference_data/datasets/TS_XCORR_OFFSET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/lag_rows/0/lagexact-4-40.00E+2passed
/lag_rows/0/correlationtolerance0.35694704328668880.356947043920701036.34E-108passed
/lag_rows/0/overlapexact56560.00E+2passed
/lag_rows/0/inside_bandexactFalseFalsepassed
/lag_rows/1/lagexact-3-30.00E+2passed
/lag_rows/1/correlationtolerance0.3260885589658270.326088559946597369.81E-108passed
/lag_rows/1/overlapexact57570.00E+2passed
/lag_rows/1/inside_bandexactFalseFalsepassed
/lag_rows/2/lagexact-2-20.00E+2passed
/lag_rows/2/correlationtolerance0.456197053385547220.4561970537893894.04E-109passed
/lag_rows/2/overlapexact58580.00E+2passed
/lag_rows/2/inside_bandexactFalseFalsepassed
/lag_rows/3/lagexact-1-10.00E+2passed
/lag_rows/3/correlationtolerance0.540371337670383370.54037133797848563.08E-109passed
/lag_rows/3/overlapexact59590.00E+2passed
/lag_rows/3/inside_bandexactFalseFalsepassed
/lag_rows/4/lagexact000.00E+2passed
/lag_rows/4/correlationtolerance0.499381878965036270.49938187928343333.18E-109passed
/lag_rows/4/overlapexact60600.00E+2passed
/lag_rows/4/inside_bandexactFalseFalsepassed
/lag_rows/5/lagexact110.00E+2passed
/lag_rows/5/correlationtolerance0.52202708171615630.52202708207112493.55E-109passed
/lag_rows/5/overlapexact59590.00E+2passed
/lag_rows/5/inside_bandexactFalseFalsepassed
/lag_rows/6/lagexact220.00E+2passed
/lag_rows/6/correlationtolerance0.64042235999493480.64042236001205221.71E-1110passed
/lag_rows/6/overlapexact58580.00E+2passed
/lag_rows/6/inside_bandexactFalseFalsepassed
/lag_rows/7/lagexact330.00E+2passed
/lag_rows/7/correlationtolerance0.978181898104165670.97818189822415491.20E-109passed
/lag_rows/7/overlapexact57570.00E+2passed
/lag_rows/7/inside_bandexactFalseFalsepassed
/lag_rows/8/lagexact440.00E+2passed
/lag_rows/8/correlationtolerance0.658989115129510310.65898911511518241.43E-1110passed
/lag_rows/8/overlapexact56560.00E+2passed
/lag_rows/8/inside_bandexactFalseFalsepassed
/omitted_lag_countexact000.00E+2passed
/peak_lagexact330.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

155  h <- read_dataset("TS_XCORR_SHIFT")
156  h$x <- h$x + 1e8
157  export_dataset(h, "TS_XCORR_OFFSET", "TS_XCORR_SHIFT with x shifted by 1e8", "R stats::rnorm documentation",
158                 "Constructed validation data; no third-party license.", c(t = "Int64", x = "Float64", y = "Float64"))
159  ho <- read_dataset("TS_XCORR_OFFSET")
160  emit("STAT-TS-XCORR-007", "A 1e8 offset leaves the correlations unchanged", "TS_XCORR_OFFSET",
161       list(series_x = "x", series_y = "y", max_lag = 4L), lag_outputs(lag_table(ho$x, ho$y, 4)),
162       c("Pearson correlation is shift invariant; R centres each lag window before multiplying."), kind = "hard_numeric")

Whole script generators/r/timeseries_cross_correlation.R


A duplicated time key is refused

Case STAT-TS-XCORR-008 · status: passed

Reference

Evidence type
R
Source
R stats::cor on explicitly aligned lag pairs
Citation
R stats::cor and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_cross_correlation.R · sha256 024b06115a56c162
Dataset
reference_data/datasets/TS_XCORR_DUPKEY.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

165  dup <- read_dataset("TS_XCORR_SHIFT")
166  dup$t[12] <- dup$t[11]
167  export_dataset(dup, "TS_XCORR_DUPKEY", "TS_XCORR_SHIFT with a duplicated time key at row 12 (1-based)",
168                 "Hand-constructed", "Constructed validation data; no third-party license.",
169                 c(t = "Int64", x = "Float64", y = "Float64"))
170  emit("STAT-TS-XCORR-008", "A duplicated time key is refused", "TS_XCORR_DUPKEY",
171       list(series_x = "x", series_y = "y", time_column = "t", max_lag = 4L), list(),
172       c("Line 18-21: duplicate keys are rejected; keys are never interpolated."), kind = "invalid_input",
173       error = list(error_class = "ValueError", message_contains = "uplicate"))

Whole script generators/r/timeseries_cross_correlation.R


A maximum lag above 250 is refused

Case STAT-TS-XCORR-009 · status: passed

Reference

Evidence type
R
Source
R stats::cor on explicitly aligned lag pairs
Citation
R stats::cor and stats::lm documentation
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/timeseries_cross_correlation.R · sha256 024b06115a56c162
Dataset
reference_data/datasets/TS_XCORR_SHIFT.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

174  emit("STAT-TS-XCORR-009", "A maximum lag above 250 is refused", "TS_XCORR_SHIFT",
175       list(series_x = "x", series_y = "y", max_lag = 251L), list(),
176       c("Line 22: a fixed ceiling of 250."), kind = "invalid_input",
177       error = list(error_class = "ValueError", message_contains = "250"))

Whole script generators/r/timeseries_cross_correlation.R


Zivot-Andrews with a level and slope break

Case STAT-TS-ZA-001 · status: passed

Reference

Evidence type
R
Source
Independent R implementation of the statsmodels 0.15.0 Zivot-Andrews procedure
Citation
Zivot and Andrews (1992); Baum (2004)
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0, forecast 9.0.2, vars 1.6.1, urca 1.3.4
Reference script
generators/r/timeseries_models.R · sha256 f0a3960b73b330ae
Dataset
reference_data/datasets/TS_LONG_01.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/statistictolerance-8.3538253486038538-8.353825348603265.93E-1313passed
/break_labelexact300300passed
/selected_lagsexact14140.00E+2passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

398  zb <- za_baum(l$za)
399  stopifnot(zb$maxlag == 19L)
400  emit("STAT-TS-ZA-001", "timeSeriesZivotAndrewsMenuItem", "Zivot-Andrews with a level and slope break", "TS_LONG_01",
401       list(method = "advanced", kind = "zivot_andrews", columns = list("za"), break_deterministic = "both"),
402       list("/statistic" = num(zb$statistic, "Minimum t on the lagged level over the candidate breaks."),
403            "/break_label" = exact(as.character(zb$break_label), "The last row before the break dummies switch on (1-based)."),
404            "/selected_lags" = exact(as.integer(zb$lag), "AIC in the no-break ADF regression, maximum lag 19.")),
405       c("Statistician ruling, issue #9: Baum's approximation (lag chosen once by AIC in the no-break constant-and-trend ADF regression, maximum 19 at n = 600, trim 0.15), reproduced independently in R.",
406         "Until 2026-09-25 the reference was urca::ur.za at lag 0, a different lag procedure."),
407       "Independent R implementation of the statsmodels 0.15.0 Zivot-Andrews procedure", "Zivot and Andrews (1992); Baum (2004)")

Whole script generators/r/timeseries_models.R


Specification checks (not independent references)

These cases verify that the product follows its own documented rules; they are not compared with an external result and are not part of the validation matrix or its counts.

A 1e6 offset in both methods

Case STAT-MSA-DEMING-005 · status: passed

Reference

Evidence type
Analytical
Source
R closed-form Deming and specified policy checks
Citation
NCSS Deming Regression chapter 303; Tabelz manual chapter 39 for policy
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_deming.R · sha256 4f9a1dc380b3c156
Dataset
reference_data/datasets/MSA_DEMING_OFFSET.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/slopetolerance1.03096917528950161.03096917528950182.27E-1615passed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

121  h <- read_dataset("MSA_DEMING"); h$method_a <- h$method_a + 1e6; h$method_b <- h$method_b + 1e6
122  export_dataset(h, "MSA_DEMING_OFFSET", "MSA_DEMING with both methods shifted by 1e6", "R stats::rnorm documentation",
123                 "Constructed validation data; no third-party license.", c(method_a = "Float64", method_b = "Float64"))
124  ho <- read_dataset("MSA_DEMING_OFFSET")
125  emit("STAT-MSA-DEMING-005", "A 1e6 offset in both methods", "MSA_DEMING_OFFSET", list(x_column = "method_a", y_column = "method_b"),
126       list("/slope" = tol(deming(ho$method_a, ho$method_b, 1)$b, "Centered sums make the slope offset-invariant.")),
127       c("R centres before forming the sums."), kind = "hard_numeric")

Whole script generators/r/measurement_deming.R


Two pairs are too few

Case STAT-MSA-DEMING-006 · status: passed

Reference

Evidence type
property
Source
R closed-form Deming and specified policy checks
Citation
NCSS Deming Regression chapter 303; Tabelz manual chapter 39 for policy
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_deming.R · sha256 4f9a1dc380b3c156
Dataset
reference_data/datasets/MSA_DEMING_TWO.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
No run results loaded for this case.

Convention

Reference code

The lines of the reference script that produce this case's expected values.

128  export_dataset(data.frame(method_a = c(1, 2), method_b = c(1.1, 2.3)), "MSA_DEMING_TWO", "Constructed: two pairs", "Hand-constructed",
129                 "Constructed validation data; no third-party license.", c(method_a = "Float64", method_b = "Float64"))
130  emit("STAT-MSA-DEMING-006", "Two pairs are too few", "MSA_DEMING_TWO", list(x_column = "method_a", y_column = "method_b"), list(),
131       c("Line 68: at least three complete pairs."), kind = "invalid_input",
132       error = list(error_class = "DemingRegressionExecutionError", message_contains = "three"))

Whole script generators/r/measurement_deming.R


Three pairs keep the fit and withhold the jackknife interval

Case STAT-MSA-DEMING-009 · status: passed

Reference

Evidence type
Analytical
Source
R closed-form Deming and specified policy checks
Citation
NCSS Deming Regression chapter 303; Tabelz manual chapter 39 for policy
Reference software
R version 4.6.1 (2026-06-24 ucrt) · stats 4.6.1, jsonlite 2.0.0
Reference script
generators/r/measurement_deming.R · sha256 4f9a1dc380b3c156
Dataset
reference_data/datasets/MSA_DEMING_THREE.json

Settings compared

Outputs

OutputRuleReferenceHypoNum |difference|DigitsResult
/slopetolerance0.834384977286466880.83438497728646692.10E-1817passed
/jackknife_slope_intervalabsentNonepassed
/interval_unavailable_reasoncontains['at least 4']Jackknife intervals need at least 4 eligible pairs; 3 are avpassed

Convention

Reference code

The lines of the reference script that produce this case's expected values.

171  export_dataset(data.frame(method_a = c(1, 2, 3.5), method_b = c(1.1, 2.3, 3.2)), "MSA_DEMING_THREE", "Constructed: three pairs",
172                 "Hand-constructed", "Constructed validation data; no third-party license.", c(method_a = "Float64", method_b = "Float64"))
173  d3 <- read_dataset("MSA_DEMING_THREE")
174  emit("STAT-MSA-DEMING-009", "Three pairs keep the fit and withhold the jackknife interval", "MSA_DEMING_THREE",
175       list(x_column = "method_a", y_column = "method_b", interval_method = "jackknife_t"),
176       list("/slope" = tol(deming(d3$method_a, d3$method_b, 1)$b, "The point fit needs only three pairs."),
177            "/jackknife_slope_interval" = absent("Jackknife intervals need at least four pairs."),
178            "/interval_unavailable_reason" = list(comparison = "contains", values = list("at least 4"),
179                                                  rationale = "The withheld interval states its reason.")),
180       c("Issue #15 ruling: at least four eligible pairs for leave-one-out intervals; three-pair point fits remain."),
181       kind = "boundary")

Whole script generators/r/measurement_deming.R