hi there,

i am running a 2x2x3 repeated measures anova with a covariate

i have an interaction between two within-subjects variables, and between a within subjects variable and the covaraite - how do i post-hoc test this to see what is going on? i am analysing using spss, but when i click on "compare means (with bonferroni correction)" it does not give me pairwise comparisons for any interaction effects??

i have tried running contrasts in spss but i don't really understand them as for one of my main effects the contrasts state that i have differences between Level 1 and Level 3 (p<.05) and between Level 2 and Level 3 (p<0.05) but when i look at the bonferroni corrected paired means comparison, it says the Level 2 vs Level 3 comparison is not sig (p=.7) but that the level 1 and level 3 is, at a "higher" sig level (p<0.001) than the contrast states... so i'm not really sure what that means... why would contrasts and pairwise comparisons be different??

thanks for any advice