GPower

GPower, typically written as G*Power, is a free software tool for statistical power analysis. It is widely used in psychology, social, and biomedical research to calculate the required sample size for a given effect size, alpha level, and desired power, or to determine the achieved power for a given sample size. The program helps researchers plan experiments and evaluate study designs with a focus on planning and interpreting statistical tests.

GPower supports a wide range of tests and analyses. It covers t tests (independent and paired), F tests (ANOVA and related designs), chi-square tests, z tests, and exact tests, as well as regression and correlation analyses. Users can specify effect sizes in common metrics such as Cohen’s d, f, eta-squared, f squared, and correlation r. The software offers several analysis options, including a priori (sample size planning), post hoc (power estimation after data collection), and compromise analyses. Outputs include recommended total sample size or per-group sizes, achieved power given N, and related statistics, with results exportable to common formats.

Accessibility and licensing are notable features: GPower runs on major desktop platforms and is available free for academic use. The interface is graphical and user-friendly, making it suitable for teaching and grant proposals, as well as routine power analyses in research planning.

History and development: GPower was developed by researchers including Faul, Erdfelder, Lang, and Buchner. The major release, G*Power 3, was described in a 2007 paper, and subsequent updates expanded test support and usability. The project is maintained with an official website hosting downloads and documentation.