Tsz Keung Wong, Samuel Pawel, Jorge Tendeiro • PsyArXiv
Finding: BayesPower is an R package with Shiny interface that enables researchers to calculate sample sizes, statistical power, and Bayes factors for various statistical tests without simulation, supporting both traditional hypothesis testing and equivalence testing.
Why it matters: This tool directly supports behavioral science research design by making Bayesian statistical planning more accessible, which is crucial for studies on decision-making, cognitive biases, and other behavioral phenomena.
Method: Developed analytical (non-simulation) methods for Bayesian power analysis across multiple statistical test types, implemented in user-friendly software with reproducible code generation.
Jargon: Bayes factor - statistical measure comparing evidence for competing hypotheses; interval hypothesis - testing whether an effect falls within a specific range rather than being exactly zero.
⚠️ No control group mentioned | Single study (not replicated)