I have a pre-post desing intervention with 2 interventin groups and 1 control group. I carried out a 3X2 repeated measured ANOVA, but just the interaction was significant so I did the simple main effects to analyze the interaction and I get good results. But I wonder if it is better to use a ANCOVA with the pretest measures as a covariate. I tried but I get better results with the simple main effect.
I think the ANCOVA will be appropriate only if the pre-test measures between the three groups were significantly different. If not I think the repeated measures ANOVA is fine.
Thanks for your reply, but one of the assumptions in ANCOVA is that it can’t be significantive difference between the groups in the covariate, so if I had significantive difference, I couldnt use it.