What it does
G*Power is a desktop tool for statistical power analysis across several test families, with graphical output and effect-size calculations.
G*Power analyzes statistical power across t, F, chi-square, z and selected exact tests. Choose a test family/model and analysis type, supply its parameters and calculate. In the version 3.1 manual, a-priori analysis solves sample size from effect size, significance and desired power; post-hoc and sensitivity analyses solve different unknowns. Parameter plots help compare assumptions, and the Windows 3.1.9.7 note requires Calculate after changing main-window inputs to update the plot. The university makes it free, including commercial use, while restricting redistribution. It states normal use runs locally, apart from Windows runtime retrieval when needed. The listed Mac 3.1.9.6 build is Intel-compiled.
Strengths
- Power analysis across multiple statistical test families
- Effect-size calculations and graphical output
Best for
- Planning a study with explicit statistical assumptions
Sources
- G*Power (Oct 26, 2026)
- G*Power 3.1 Manual (Oct 26, 2026)