The Power Analysis Toolkit: GUI Software, Code or Simulation?

Software choice should follow the study design

Different test definitions, tails and effect metrics can produce different samples. Reproducibility matters more than brand.

Practical framework

• G*Power is accessible for standard tests.

• R supports auditable code and simulation.

• PASS offers a broad commercial library.

• Complex designs may require simulation.

Reporting

Report version, test family, tails, effect definition and all inputs.

Common mistakes

• Accepting default effects.

• Choosing one-sided tests opportunistically.

• Reporting screenshots without inputs.

References

• G*Power 3 paper

• FDA Statistical Principles for Clinical Trials

Accessed 20 June 2026. General methodological information.