Kruskal-Wallis test is primarily used for which scenario?

• A. Two independent samples
• B. Three or more independent samples
• C. A correlation between two variables
• D. Comparing means of two dependent samples

Answer: (B)

The test is used when you have three or more independent groups and you want to compare their central tendency without assuming a normal distribution. It is the nonparametric counterpart to one-way ANOVA, ranking all observations across groups and then evaluating whether the overall rank patterns differ enough to suggest that at least one group median is different. This approach is appropriate when data are not normally distributed or are ordinal, and you have more than two groups to compare. It isn’t used for two independent samples (that would be a Mann-Whitney U test), nor for assessing a relationship between two variables (that would involve correlation methods like Spearman or Pearson), nor for comparing means of two dependent (paired) samples (which uses a paired test such as the Wilcoxon).