statistical-testing

What problem does it solve?

This Skill helps you avoid incorrect statistical choices by selecting the appropriate hypothesis test based on your data’s structure and assumptions, then executing it with key supporting outputs like effect sizes and (optional) power-friendly metrics.

Core Features & Use Cases

• Automatic test selection via decision logic: Chooses parametric vs non-parametric and the correct test for independent vs paired data, including multi-group cases.
• Assumption checking (normality): Runs Shapiro-Wilk, Kolmogorov-Smirnov, and D’Agostino-Pearson and uses a consensus rule to decide whether data are approximately normal.
• Effect size support: Computes effect-size measures such as Cohen’s d, eta-squared, and Cramér’s V (and rank-based alternatives where relevant).
• Multiple comparison correction: Provides p-value correction using common methods like FDR (Benjamini-Hochberg) when doing post-hoc testing.
• Power analysis: Estimates required sample sizes for two-sample t-tests to plan study design.

Quick Start

Use the statistical-testing skill to compare two groups from a dataset while automatically checking normality and returning the selected test name, statistic, p-value, effect size, and significance decision.