G*Power for Windows

Snapshot — what the tool does

G*Power is a free program for conducting statistical power calculations and related sample-size planning. It’s built around standard hypothesis-test families and provides both numeric outputs and graphical displays so you can estimate power, compute required sample sizes, or assess sensitivity under a declared model.

Supported test families and modules

The app groups analyses by statistical test type and offers dedicated modules for common families, including:

• F tests (ANOVA, regression-derived tests)
• z tests (large-sample or known-variance scenarios)
• χ² tests (contingency tables, goodness-of-fit)
• t tests (paired, independent, one-sample)
• several exact-test options for small-sample or discrete situations

Each module opens an entry panel where you select one- or two-tailed designs, distribution assumptions, and the parameters required for that test.

Effect-size assistance and entry helpers

Built-in calculators help you determine standardized effect magnitudes and pass them directly into your power model:

• f² and other regression-based effect measures for F-type analyses
• Cohen’s d and equivalents intended for t-type comparisons

Common conventions (small, medium, large) are shown next to entries to guide consistent choices. The software automatically transfers computed effect values into the power-analysis fields, so no manual copy–paste is needed.

How power computations are organized

You can run the usual types of power investigations:

• A priori calculations (determine needed sample size for a target power)
• Post hoc checks (estimate achieved power with fixed n)
• Sensitivity analyses (smallest effect detectable at given power and alpha)
• Criterion or compromise approaches (mixing constraints on alpha and power)

Results are presented in numeric tables that report critical thresholds, noncentrality parameters, and estimated power for the chosen configuration. Outputs are designed to be clear for reporting and interpretation.

Graphing, customization, and export

The plotting module creates visual summaries such as:

• power curves across a range of sample sizes or alpha values
• plots comparing alternative distributional assumptions or effect sizes
• customizable axis ranges and labeling options

Numeric tables and graphs can be exported for inclusion in reports or external documentation. Note that the program does not import raw datasets or perform primary data analyses — it focuses strictly on planning and evaluating power.

Practical notes and limits

• Intended purpose: planning and evaluating statistical power and sample-size decisions, not data processing or hypothesis testing on empirical datasets.
• Visual and numeric outputs facilitate comparison across scenarios, but assumptions must be specified carefully (distribution, tails, effect metric).
• No facility for importing data or performing model fitting directly from a dataset; results are based on user-entered parameters and built-in formulas.

Quick feature summary

• Exportable tables and figures for documentation
• Integrated effect-size calculators that feed into analyses
• Organized, module-based selection of test families