Which test would you use to assess differences when the dependent variable is ordinal?

The Friedman Test is specifically designed for situations where the dependent variable is ordinal and the data are not normally distributed. It is a non-parametric statistical test that is used to detect differences in treatments across multiple test attempts. When you have repeated measures on the same subjects (for example, measuring the same individuals under different conditions), the Friedman Test allows for the assessment of variance among the groups, making it an appropriate choice for analyzing ordinal data that involve matched or related samples.

In contrast, the Kruskal-Wallis Test is also a non-parametric test, but it is more suitable when you are comparing three or more independent groups. Ordinary Least Squares and Linear Regression require interval or ratio-level data for the dependent variable and typically assume normality in the data distribution, making them inappropriate for ordinal data assessments. Thus, the Friedman Test appropriately addresses the need to analyze differences in ordinal dependent variables under the right conditions of repeated measures or related groups.