Sample Size & Power Calculator
Plan an analyzable sample size or calculate achieved (post hoc) power across common mean, proportion, correlation, ANOVA, regression, chi-square, survival, equivalence, diagnostic-accuracy, precision, matched, and cluster-randomized designs. Calculations are deterministic and run locally in your browser.
Calculator scope and limitations
What this calculator does — and does not do
Sample-size calculations are conditional on the design, estimand, assumed effect, variability or event rate, significance level, power target, allocation, and any design effects entered. A precise numerical answer does not remove uncertainty in those assumptions. Use sensitivity analysis and justify assumptions from subject-matter evidence whenever possible.
Post hoc achieved power is provided because it is often requested, but it should not be used as a substitute for confidence intervals, effect estimates, or prospective design reasoning. The generated report explicitly flags this limitation.
Supported design families
- One-sample and paired means; two independent means
- One and two proportions; unmatched case-control; McNemar paired proportions
- Pearson correlation and comparison of two independent correlations
- One-way ANOVA, multiple regression R² increase, and chi-square tests
- Log-rank / proportional-hazards event planning
- Equivalence and non-inferiority approximations for continuous outcomes
- Mean, prevalence, sensitivity, and specificity precision calculations
- Cluster-randomized continuous and binary designs using a design-effect approximation
This tool supports research planning and statistical education. It is not a substitute for a design-specific statistical review when the sampling scheme, missing-data mechanism, multiplicity, interim monitoring, complex covariance structure, competing risks, informative clustering, survey weights, or regulatory requirements materially affect the calculation.