Definition · Plain-language
G*Power
G*Power is a free statistical software program used to perform prospective power analyses and calculate required sample sizes across a wide range of experimental designs.
The step most authors miss
Doing CRediT right? Don’t stop at the statement.
A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.
Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.
The role of G*Power in research design
G*Power is a free, widely used software application developed by researchers at Heinrich Heine University Düsseldorf to perform statistical power analyses and calculate required sample sizes. In quantitative research, underpowered studies are a major issue because they fail to detect true experimental effects and contribute to the replication crisis. G*Power solves this problem by helping researchers determine the exact sample size needed before starting data collection. This prospective planning is essential for ethical committee approvals, research proposals, and grant applications. By ensuring that studies have sufficient statistical power, researchers can avoid wasting time, money, and participant resources on inconclusive experiments.
Types of power analysis supported
G*Power supports five major types of power analysis, categorised by the researcher’s goals and design stage. The most common is the a priori (prospective) analysis, which calculates the required sample size based on the chosen significance level, expected effect size, and desired power. Post-hoc analysis calculates statistical power after data collection, using the sample size and observed effect size, although methodologists advise caution with this approach. The programme also offers compromise analysis, which balances alpha and beta error rates, sensitivity analysis to determine the minimum detectable effect size, and criterion analysis to find the necessary alpha level. These varied modes allow investigators to thoroughly evaluate their study designs under different resource constraints, ensuring they choose the most mathematically sound approach before collecting any experimental data.
Supported statistical tests and inputs
G*Power organises analyses into five major test families: exact tests, t-tests, F-tests, chi-square tests, and z-tests. Within these families, researchers can calculate power for simple designs, such as independent samples t-tests, as well as complex setups, such as repeated-measures ANOVA and multiple linear regressions. Users must input parameters like effect size, significance level (alpha), and power (typically 0.80 or 0.95). The software features built-in calculators to help estimate effect sizes from prior descriptive statistics. Available for Windows and macOS, G*Power remains a critical, free tool for planning quantitative academic studies. By utilising these integrated tools, scientists can accurately justify their sample sizes in ethical reviews and grant applications, proving their studies are computationally viable.
Key facts
At a glance
- Availability: completely free to download and use for academic and research purposes.
- Developer: maintained by the Department of Experimental Psychology at Heinrich Heine University Düsseldorf.
- Core function: calculates required sample sizes and statistical power to prevent underpowered research.
- Calculators: includes built-in calculators to help estimate effect sizes from descriptive statistics.
- Visualisation: generates custom power curves showing how sample size varies with power and effect size.
- Cross-platform: available for both Microsoft Windows and macOS operating systems.
Common misconceptions
What people often get wrong
Often heard: G*Power is a general statistical analysis software like SPSS or R.
Actually: G*Power cannot analyse raw datasets or run descriptive statistics. It is a specialised utility designed solely for sample size planning and power calculations.
Often heard: You should always use the default effect size values in G*Power.
Actually: The default effect sizes in G*Power are generic guidelines (small, medium, large). Researchers should base their inputs on prior literature, pilot data, or clinical relevance.
Common questions
FAQ
Why is prospective power analysis required by grant agencies?+
Grant agencies require prospective power analysis to ensure that public or private funds are not wasted on studies that are too small to detect an effect, or too large (which is ethically problematic in human or animal research).
What is a sensitivity analysis in G*Power?+
A sensitivity analysis calculates the minimum effect size that a study could reliably detect given a fixed sample size, alpha level, and power, which is useful when recruitment is limited by budget or rare populations.
Going deeper
