The Wilcoxon Signed-Rank Test is used for which type of groups?

The Wilcoxon Signed-Rank Test is used for which type of groups?

• A. Dependent Groups
• B. Independent Groups
• C. One-way ANOVA
• D. Kruskal-Wallis

Answer: (A)

Two related or paired samples are what the Wilcoxon Signed-Rank Test is built for. Think of measurements taken on the same subjects under two conditions, or on matched pairs. The test asks whether the typical difference between each pair is zero, by looking at the signs and ranks of those differences. Because it uses ranks of the differences rather than the raw values, it doesn’t require the differences to be normally distributed, making it a nonparametric alternative to the paired t-test. It works with ordinal data as well as continuous data, provided you have paired observations and the differences are symmetrically distributed around zero. If the groups were independent rather than paired, a different test would be appropriate—the Wilcoxon rank-sum (Mann-Whitney) test handles independent samples. For more than two independent groups, Kruskal-Wallis is used, and for a parametric counterpart with two related samples you’d use the paired t-test.

Two related or paired samples are what the Wilcoxon Signed-Rank Test is built for. Think of measurements taken on the same subjects under two conditions, or on matched pairs. The test asks whether the typical difference between each pair is zero, by looking at the signs and ranks of those differences. Because it uses ranks of the differences rather than the raw values, it doesn’t require the differences to be normally distributed, making it a nonparametric alternative to the paired t-test. It works with ordinal data as well as continuous data, provided you have paired observations and the differences are symmetrically distributed around zero.

If the groups were independent rather than paired, a different test would be appropriate—the Wilcoxon rank-sum (Mann-Whitney) test handles independent samples. For more than two independent groups, Kruskal-Wallis is used, and for a parametric counterpart with two related samples you’d use the paired t-test.