The non-parametric alternative to Repeated Measures ANOVA for comparing three or more related groups.
The Friedman Test is used to determine if there are any statistically significant differences between the distributions of three or more related samples. It's essentially a ranked-based version of a repeated-measures ANOVA, making it suitable for non-normal data.
Use this test when you have one group that has been measured on three or more different occasions or under three or more different conditions. It's ideal when the assumptions for a repeated measures ANOVA (like normality) are not met.
This test is ideal for analyzing how the same subjects (e.g., algorithms, portfolios) perform across multiple different, related conditions.
Example: A quant team wants to compare the performance of five trading algorithms across three different market volatility regimes (Low, Medium, and High). For each regime, they rank the algorithms from 1 (best) to 5 (worst). The Friedman test is used to determine if there is a significant difference in the algorithms' median ranks across the different volatility conditions.