This document summarizes nonparametric hypothesis testing methods that can be used when sample sizes are small and variables are not normally distributed. It describes the sign test, which tests for differences in paired data by counting how many pairs differ, and the Mann-Whitney U test for unpaired data, which ranks all values together and tests for differences in average ranks between samples. An example compares the income of applicants for fixed-rate versus variable-rate mortgages using the Mann-Whitney U test, finding the probability of difference being due to chance is 18.3%, suggesting variable rate applicants have higher income.