---

name: statistical-tests-and-diagnostics description: "Use statsmodels statistical tests, diagnostics, ANOVA, contrasts, influence, multiple testing, contingency tables, power, effect-size, mediation, treatment, and meta-analysis utilities." disable-model-invocation: true metadata: disco-role: operating license: BSD 3-Clause

Statistical tests and diagnostics

Use this sub-skill when the task is selecting or running statistical tests, model diagnostics, ANOVA/contrasts, residual checks, influence/outlier analysis, contingency tables, multiple comparisons, power/effect size, mediation, treatment-effect utilities, or meta-analysis.

Workflow

1. Identify the design: one sample, two sample, paired/grouped, regression residuals, categorical table, time-series residuals, multiple hypotheses, or power calculation.
2. Check assumptions and input shapes before selecting a test: independence, normality, equal variance, stationarity, group sizes, missing data, and fitted result class.
3. Run the focused function and preserve the statistic, p-value, degrees of freedom, correction method, and null/alternative interpretation.
4. Route model fitting back to the owning model sub-skill if the user has not yet fit the model.

Read or run

• Read references/api-reference.md for major modules and functions.
• Read references/workflows.md for residual diagnostics, ANOVA/contrasts, multiple testing, contingency tables, and power recipes.
• Read references/troubleshooting.md for test selection, missing assumptions, singular designs, and multiple-testing interpretation.
• Run scripts/smoke_stats_diagnostics.py for a tiny OLS diagnostics and multiple-testing smoke check.

Cross-routes

• Use linear-and-formula-models or discrete-and-count-models to fit the model before diagnostics.
• Use time-series-analysis when the diagnostic is a time-series model choice or forecasting problem.
• Use datasets-results-graphics for diagnostic plots and table export.