Dear all,
I have 2 groups (A and B), subdivided in 2 subgroups (A1, A2, B1, B2). I used a 2x2 between-groups ANOVA to estimate the main effect of group, the main effect of subtype, and their interactions. Now I want to estimate post-hoc differences between my subgroups (2-sample t-tests).
Is it better to do so within the ANOVA model, or to create separate 2 sample t-tests?
I used a FWE-correction at the cluster level. Can someone confirm me that (independently of the above option chosen) I will have to correct for the number of test on my (cluster) statistics (in this case, only report cluster with a significance of p<.008)?
Thanks for your answer!
Best,
Yann
|