Hi list, I'm performing cortical thickness among 3 groups by using GLM. I created a group descriptor file with 3 groups and 2 covariates (age and mean thickness). I performed 3 comparisons A-B, A-C, B-C. Now, for one of 3 groups (group C) I have: groups C=30 subjects; I'd like to divide it in some pairs. First pair: X (n=15) vs Y (n=15). Second pair: Z (n=15) vs W (n=15). Between first and second pair there isn't interactions. X or Y subgroup can contain also subjects included in Z or W. Z or W subgroup can contain also subjects included in X or Y. Design: group A group B subgroup X vs subgroup Ygroup C ------< subgroup Z vs subgroup W

I'm thinking to perform qdec for X-Y and Z-W. Is it correct? What do you advise me? Thanks,

Stefano