Which test is used for two independent samples of ordinal data?

The Mann Whitney U test is designed specifically for comparing two independent samples when the data is ordinal or when the assumptions required for parametric tests, such as normality, are not met. This non-parametric test evaluates whether the distribution of values in the two groups is significantly different.

By using ranks instead of raw data, the Mann Whitney U test effectively handles ordinal data, making it suitable for situations where the data does not have a true zero or where the intervals between values cannot be assumed to be equal. This is particularly pertinent when comparing different groups, as it helps to determine if one group tends, on average, to have higher or lower values than the other without making stringent assumptions about the underlying distribution.

In contrast, the Chi square test is more suitable for categorical data, while the Wilcoxon signed rank test is applicable for paired samples rather than independent samples. ANOVA is a parametric test used for comparing means among three or more groups, which does not accommodate ordinal data directly. Thus, the Mann Whitney U test is the most appropriate choice for the question regarding two independent samples of ordinal data.