Table : TABLE E.5 Nonparametric Statistics and Their Parametric Counterparts Every parametric hypothesis test has at least one nonparametric counterpart. If the data are far from meeting the assumptions for a parametric test or at least one variable is ordinal, we should use the appropriate nonparametric test instead of a parametric test.

Design	Parametric Test	Nonparametric Test
Association between two variables	Pearson correlation coefficient	Spearman rank-order correlation coefficient
Two groups; within-groups design	Paired-samples t test	Wilcoxon signed-rank test
Two groups; between-groups design	Independent-samples t test	Mann–Whitney U test
More than two groups; between-groups design	One-way between-groups ANOVA	Kruskal–Wallis H test