G*Power is a free statistical calculation utility that provides power analysis functions for multiple hypothesis test families. It offers modules for t tests, F tests, χ² tests, z tests, and several exact tests, with structured input fields for specifying effect size, sample parameters, and significance levels.

G*Power includes calculators for deriving effect size values based on common statistical conventions and allows you to generate numerical tables reflecting variations across selected parameters. The software features graphical displays of power curves and related outputs. This uses a priori, post hoc, sensitivity, and criterion analysis, showing results numerically and visually for interpretation.

Graph your power

G*Power organizes power analysis into categorized menus corresponding to statistical test types. Each test selection displays input panels for selecting tails, distribution assumptions, and effect size parameters. You can compute required sample sizes or determine achieved power by entering values into the parameter fields. The software displays results in structured numeric output tables. These outputs include critical values, noncentrality parameters, and projected power estimates associated with the selected model.

The effect size tools provide built-in formulas for deriving standardized effect values based on common statistical frameworks. These calculators vary depending on the test family selected, such as Cohen’s d for t-tests or f² for regression-related F-tests. Effect size conventions are displayed alongside entry fields to support consistent parameter selection. Outputs from effect size calculations feed directly into the power analysis panel without requiring external transfer.

It also produces visual representations of power calculations. The graphing module can display power curves across a parameter range, including alpha level or sample size variations. You can adjust axis limits, specify alternative distributions, and generate plots reflecting the assumptions of different statistical approaches. However, it lacks dataset import or direct data analysis features. Numeric summaries and plots can be exported and saved for external interpretation or documentation.

Expanded statistical coverage

G*Power supports a broader range of statistical procedures than basic power calculations, including analyses for correlations, linear regression, logistic regression, Poisson regression, ANOVA, proportions, variances, and several nonparametric and exact tests. Its official documentation also covers specialized procedures such as tetrachoric correlations and comparisons between dependent or independent correlations, making the software useful for more specialized research designs.

Built for research planning

G*Power can be used at several stages of research planning rather than only after a study has been designed. Researchers can estimate required sample sizes before collecting data, examine achieved power afterward, or perform sensitivity analyses to determine which effect sizes can be detected under particular sample-size conditions. This makes its calculation tools useful for experimental design and statistical planning.

Detailed test-specific controls

The software provides specialized parameters depending on the selected statistical procedure. For example, its regression modules can account for numbers of predictors, changes in R², individual regression coefficients, and different regression models, while other procedures provide controls appropriate to their respective distributions and assumptions. This test-specific structure allows researchers to configure analyses more precisely instead of relying on a single generic power calculator.

Continued development and future support

The current official release remains G*Power 3.1.9.7 for Windows, released in March 2020, with later releases having addressed calculation and interface bugs. The developers note that G*Power 4 is in development, including a native Apple-silicon version intended to provide better support for newer Mac hardware. This suggests the project is being prepared for a newer generation of Apple devices even though the established 3.1 release remains available.

For statistical power

With its extensive collection of test procedures, specialized effect-size calculations, and graphical power analysis, G*Power remains focused specifically on research planning rather than general statistical analysis. Its continued availability for Windows and Mac, together with work toward a newer Apple-silicon version, gives researchers a dedicated environment for estimating power and sample requirements without requiring a full statistical analysis package.

Analyze power

G*Power offers structured power analysis tools for common hypothesis test families. It includes parameter entry panels, effect size computation utilities, numeric result summaries, and plotting features that display power curves. The software focuses on producing power estimates and sample size targets under specified model assumptions. It provides analytical power analysis and visualization, not data analysis. Key features: categorized tests, calculation modules, and outputs.