The non-parametric alternative to the Paired T-Test, used for 'before and after' analysis on non-normal data.
This test is the non-parametric version of the Paired T-Test. It's designed for comparing two related measurements from the same subject. Instead of using raw data, it ranks the differences between pairs to see if there's a significant change.
Use this for "before and after" scenarios when your data is **not normally distributed**. It's perfect for measuring the impact of an intervention on the same group of subjects, like testing if a new risk model reduced portfolio drawdown.
This test is ideal for seeing if a change had a consistent effect across a group, even when the outcomes are skewed.
Example: A risk management team implements a new model for 10 of their portfolios. They record the maximum drawdown of each portfolio for a month before the change and a month after. Since drawdown data is often skewed (many small values, few large ones), they use the Wilcoxon Signed-Rank Test to see if the new model led to a statistically significant reduction in drawdown.