I just ran a MANCOVA to compare my grouping variable with my dependent variables whilst controlling for my covariates.
I am running a descriptive study, therefore already by looking at descriptive statistics and running a couple of t-test, I came to see that there is really no difference between the two groups.
When running MANCOVA, all my multivariate tests say that nothing is significant. When reading around the limited resources I could find on MANCOVA, I read that if the multivariate tests are not significant, you do not continue to report or search for other results, as finding something significant continues. Is this true?
Also, how do you report non-significant MANCOVA results? I understand that you should report main effects, but I can't seem to find anything regarding how to actually report that in my result section.
I find it really impossible to find anything on MANCOVA, like this doesn't even exist, ha ha.
Best Answer
If you're doing an exploratory study, then looking at individual effects might be reasonable, but in a testing context, if the multivariate tests fail to find results, then looking at individual pieces is indeed widely considered fishing and frowned upon, unless you use p-value adjustments to avoid Type I error level inflation.
Check multivariate analysis textbooks for information on MANCOVA. Using Multivariate Statistics by Barbara Tabachnick and Linda Fidell is popular, and covers MANCOVA.