Is there a Mann Whitney test alternative when all variables are categorical?

I have two teams, the continuers and the dropouts, and I want to compare them on gender, education level, and age level. The sample size is small and we need non parametric tests. Which test would you recommend?