Definition

Non-parametric test

A test that makes minimal assumptions about the data’s distribution, typically by working on ranks — Mann-Whitney, Wilcoxon, Kruskal–Wallis, Spearman.

Parametric tests (t-test, ANOVA, Pearson) assume roughly normal residuals; their non-parametric twins replace values with ranks and drop that assumption. The price is some statistical power when the data actually are normal, and hypotheses about medians/distributions rather than means.

Use them for small samples with visible skew, ordinal data (single Likert items), and outlier-heavy measures like reaction times. With n ≥ 30 per group, the central limit theorem usually makes the parametric test both valid and more sensitive.