Statistical methods
Capability organized by research purpose.
Version 5 exposes 36 routed analysis workflows. Availability and validation are stated separately: a method is not described as universally equivalent to another package unless that claim has actually been earned.
Analysis Studio
Direct method access remains available.
Research Guide simplifies entry. Analysis Studio keeps the routed method catalogue visible for researchers who already know the analysis they need.
Method atlas
01 / Descriptive
Descriptive StatisticsFrequency DistributionShapiro–Wilk
02 / Comparative
One-Sample T-TestIndependent T-TestPaired T-TestOne-Way ANOVAWelch ANOVALeveneBrown–ForsytheTwo-Way ANOVAMann–Whitney UWilcoxonKruskal–WallisFriedmanSign Test
03 / Association
Chi-SquareFisher ExactMcNemarPearsonSpearmanKendall tau-bCorrelation Matrix
04 / Regression
Linear RegressionMultiple Linear RegressionLogistic RegressionPoisson Regression
05 / Diagnostic
ROCAUCSensitivitySpecificityThreshold TableYouden Index
06 / Survival
Kaplan–MeierCensoringGrouped SurvivalMedian SurvivalLog-Rank Test
07 / Reliability
Cronbach AlphaItem DiagnosticsCohen's KappaBland–Altman
08 / Multivariate + specialized
PCAK-MeansPower AnalysisTime Series / ACF
Validation
Reference posture
Implemented is not the same as independently reference-validated.
The current regression suite distinguishes methods with direct reference checks from methods whose broader equivalence audit remains pending.
Reference validatedValidatedSpearman, Kendall tau-b, one-sample t, Welch independent t, paired t, one-way ANOVA, Welch ANOVA, Levene, Brown–Forsythe, Mann–Whitney, Wilcoxon, Fisher exact, McNemar, logistic regression, Poisson regression, Bland–Altman, Cronbach alpha, tested AUC path.
Validated approximationApproximationShapiro–Wilk in tested cases; AUC confidence interval path uses a validated numerical approximation posture.
Implemented — audit pendingAudit pendingIncludes descriptive/frequency breadth, Pearson matrix-wide equivalence, two-way ANOVA, Kruskal–Wallis, Friedman, Sign, Chi-square, linear and multiple regression, Cohen kappa, PCA, clustering and Kaplan–Meier broad equivalence suites.
Methods remain inspectable.
Research Guide simplifies entry. Analysis Studio keeps the direct route.