Parametric or non-parametric tests

This Fact Sheet provides practical guidance on choosing between parametric and non-parametric methods for common group comparisons. It explains what “parametric” and “non-parametric” mean, clarifies common misconceptions about normality testing, and outlines assumption checking. It contrasts mean-based inference (e.g., t-tests, ANOVA, linear regression) with rank-based approaches (e.g., Mann–Whitney U, Kruskal–Wallis, Wilcoxon, Friedman), discusses when ordinal outcomes and severe assumption violations warrant non-parametric methods, and highlights robust alternatives (e.g., Welch procedures, robust standard errors, GLMs, mixed-effects models). No new data were collected; ethical approval is not applicable.