In my two-way repeated measures ANOVA analysis, I use Language (Frisian, Dutch) and Category (identical-cognates, form-similar cognate, non-cognate) as within subject variables. Sphericity is assumed.
There is a significant interaction between Language and Category for one reading measure in the Test of Within-Subject Effects table. The Multivariate table shows that the effect of Category is significant only in the Frisian version, and not in the Dutch version. However, in LSD post hoc tests, both Frisian and Dutch show significant effects of Category.
Can I report the multivariate table to show that the post-hoc results for Dutch should not be taken into account, since the effect of Category is only significant for Frisian? I have not seen any previous literature using the Multivariate table in their one-way repeated measures report, only the Pairwise comparisons.