How to detect group differences between groups with low group size like for instant n=8 vs n=50? Are there any statistics measures I could use like looking at interaction between group and gender, or should I use doss? Is it feasible to design a custom based design matrix?

Knut J