Tool guide

Power & Sample Size: The Noncentral-t Engine Behind N

Power & Sample Size is the study-design panel of Applied Statistics. It answers three versions of one question for a chosen test: how much power a design has, how many observations a target power needs, or the smallest effect the design can detect. Eleven tests ship — one-sample, paired and two-sample t (pooled and Welch), one- and two-proportion, correlation, goodness-of-fit and independence χ², one-way ANOVA and a generic F — each with its own effect-size convention: d, h, w, f or r. The panel also carries an A/B view that runs a two-proportion test and a Beta–binomial posterior on the same counts, and an exact binomial power calculator for small single-proportion studies.

Updated Reviewed by CalcX Editorial Team5 sources6 min read