Power Analysis & Sample Size Advisor

Sample size is one of the most consequential decisions in any empirical study. Too small, and your study lacks the statistical power to detect a real effect — producing a null result that proves nothing. Too large, and you waste resources and, in clinical research, expose more participants than necessary to experimental conditions. Getting it right requires explicit power analysis grounded in defensible assumptions. This AI assistant makes that process accessible and rigorous.

The assistant guides researchers, grant writers, and study planners through power analysis for a wide range of study designs and statistical tests. It covers the foundational logic of statistical power — the relationship between alpha, beta, effect size, and sample size — in plain language, so you understand the calculation rather than just running a formula. It helps you make defensible assumptions about expected effect sizes, drawing on published benchmarks, pilot data, or minimum clinically important differences.

For specific designs, the assistant calculates and explains sample size requirements for two-sample t-tests, one-way and factorial ANOVA, chi-square tests of association, correlation and regression analyses (including multiple regression with covariates), paired and within-subject designs, survival analysis based on expected event rates, and multilevel studies with nested structures. It explains design efficiency concepts such as the benefit of within-subject designs and the cost of clustering in hierarchical data.

Beyond calculation, the assistant helps you communicate your power analysis in grant applications and ethics submissions — explaining your effect size source, alpha and power conventions, dropout assumptions, and any adjustments for multiple comparisons. It also helps you interpret an existing study's post-hoc power retrospectively, with appropriate caveats about why post-hoc power is often misunderstood.

This assistant is ideal for academic researchers preparing grant applications, graduate students designing dissertations, IRB submitters, and clinical trial teams establishing protocol sample size justifications.