Interactive resource

Clinical Statistics Workbench

Plan analyses, calculate core epidemiologic measures, audit deidentified data, and draft reproducible methods for clinical quality and research projects.

Educational and planning support. Verify results in validated statistical software. Never enter identifiable patient or confidential institutional data.

Clinical statistics workbench

Statistical Test Selector

Use this as a planning aid, then confirm the choice against your protocol, estimand, sample size, and model assumptions.

Recommended starting point

Welch two-sample t-test

Compares two independent means without requiring equal variances.

  • Report an effect estimate and 95% confidence interval, not only a p-value.
  • Check missing data, independence, and whether adjustment is required.
  • For repeated measurements with more than two time points, consider a mixed-effects model.