For comparing more than two independent groups, which nonparametric test is used?

For comparing more than two independent groups, which nonparametric test is used?

• A. One-way ANOVA
• B. Mann-Whitney U test
• C. Kruskal-Wallis ANOVA by ranks
• D. Friedman test

Answer: (C)

When comparing more than two independent groups without assuming normal distribution, you use a nonparametric approach that relies on ranks across all observations. The Kruskal-Wallis test does exactly this: it ranks all data from every group together and then tests whether those ranks differ among the groups. It’s the nonparametric counterpart to one-way ANOVA, which tests differences in means under normality assumptions. Because it doesn’t require normality, it’s appropriate for ordinal data or continuous data with skew, outliers, or small samples. If the test detects differences, it indicates that at least one group tends to have a different distribution of values, though it won’t specify which groups differ—that’s where post hoc tests like Dunn’s come in. The other options either assume normality and compare means (parametric one-way ANOVA), are limited to two groups (Mann-Whitney U), or apply to related samples or repeated measures (Friedman).

When comparing more than two independent groups without assuming normal distribution, you use a nonparametric approach that relies on ranks across all observations. The Kruskal-Wallis test does exactly this: it ranks all data from every group together and then tests whether those ranks differ among the groups. It’s the nonparametric counterpart to one-way ANOVA, which tests differences in means under normality assumptions. Because it doesn’t require normality, it’s appropriate for ordinal data or continuous data with skew, outliers, or small samples. If the test detects differences, it indicates that at least one group tends to have a different distribution of values, though it won’t specify which groups differ—that’s where post hoc tests like Dunn’s come in. The other options either assume normality and compare means (parametric one-way ANOVA), are limited to two groups (Mann-Whitney U), or apply to related samples or repeated measures (Friedman).