I am trying the compute a robust post-hoc test following a mixed-ANOVA (3 between and 5 with-in factors) with trimmed means using the WSR2 package. I have found a sig. interaction.
I am aware that there is no function in the WSR2 package for simple main effect post-hoc analysis. Therefore I am wondering if it is possible to do the post hoc analysis by computing seperate robust t-tests and adjusting with the bonferroni method? While investigating the with-in factor on every step of the between factor I would use a robust pairwise test and a robust test for unrelated samples in the other case.
Any thoughts on that?
Cheers