How to choose effect-size inputs for a defensible power analysis
Use prior evidence, smallest effects of interest, design constraints, and sensitivity analysis instead of an unexplained conventional value.
Use prior evidence, smallest effects of interest, design constraints, and sensitivity analysis instead of an unexplained conventional value.
Start with the decision the study must support
Power depends on an estimand, design, analysis model, error criterion, and target operating characteristic. Naming a test without those inputs is not enough.
Build a range of justified inputs
Use relevant prior studies cautiously, measurement reliability, plausible heterogeneity, and the smallest effect that would matter. Record why each source is transferable or not.
Use sensitivity as the main communication tool
Show how required sample size changes across plausible effects, attrition, clustering, and variance assumptions. This makes uncertainty visible before recruitment.
Failure modes to check
- Choosing a method because it is familiar rather than because it matches the estimand and data structure.
- Treating assumptions as a generic pass/fail checklist instead of evidence about a specific model.
- Changing the workflow after seeing results without recording the decision or its effect on uncertainty.
- Reporting software output without tracing the claim to final code, diagnostics, and limitations.
Continue the decision path
Explore the related statistical consulting service, or discuss the decision in your own study.
Related decision guides
- Reliability is not validity: a measurement decision guide
- EFA vs CFA: choose from the research decision, not the software menu
Methodological note
This guide provides general educational information. The appropriate decision depends on the study question, design, data, evidence, ethics, and institutional requirements.