Name a nonparametric test for paired samples.

Name a nonparametric test for paired samples.

• A. Wilcoxon signed-rank test
• B. Pearson correlation
• C. ANOVA
• D. Paired t-test

Answer: (A)

When comparing two related measurements without assuming they come from a normal distribution, you use a method that handles paired data in a nonparametric way. The Wilcoxon signed-rank test does exactly that: it looks at the differences within each pair, ignores any pairs with zero difference, ranks the absolute differences, and then examines whether the positive-difference ranks and negative-difference ranks balance as they would if there were no real difference between the conditions. This approach relies on the data being at least ordinal and the differences being symmetrically distributed around zero, not on normality, making it robust to outliers and suitable for small samples or non-normal data. The other options aren’t the best fit here. A paired t-test is the parametric counterpart that requires the differences to be normally distributed. ANOVA is used for comparing means across more than two groups (and the repeated-measures version is still a parametric approach with its own assumptions). Pearson correlation measures linear association between two variables, not a test of whether paired measurements differ.

When comparing two related measurements without assuming they come from a normal distribution, you use a method that handles paired data in a nonparametric way. The Wilcoxon signed-rank test does exactly that: it looks at the differences within each pair, ignores any pairs with zero difference, ranks the absolute differences, and then examines whether the positive-difference ranks and negative-difference ranks balance as they would if there were no real difference between the conditions. This approach relies on the data being at least ordinal and the differences being symmetrically distributed around zero, not on normality, making it robust to outliers and suitable for small samples or non-normal data.

The other options aren’t the best fit here. A paired t-test is the parametric counterpart that requires the differences to be normally distributed. ANOVA is used for comparing means across more than two groups (and the repeated-measures version is still a parametric approach with its own assumptions). Pearson correlation measures linear association between two variables, not a test of whether paired measurements differ.