Which nonparametric test would you use for comparing two independent groups when the outcome is ordinal or not normally distributed?

• A. Mann-Whitney U test.
• B. Independent samples t-test.
• C. Wilcoxon signed-rank test.
• D. Chi-square test.

Answer: (A)

When comparing two independent groups where the outcome is ordinal or not normally distributed, you want a method that doesn’t assume normality and can handle ranks. The Mann-Whitney U test does exactly this: it ranks all observations across both groups and compares the sums of those ranks between the groups. If one group tends to have higher (or lower) values, the rank sums differ in a way that the test detects, indicating a difference between groups. This makes it the appropriate nonparametric alternative to the independent-samples t-test for this situation.

The independent-samples t-test, by contrast, assumes normality and compares means, which isn’t suitable for non-normal or ordinal data. The Wilcoxon signed-rank test is designed for paired or matched samples, not two independent groups. The Chi-square test handles categorical data frequencies, not ordinal or continuous outcomes.