I have a simple 'split plot' design; 2 groups (headache sufferers vs
Controls), 2 conditions (pre vs post intervention), 1 DV (pain). RMANOVA
gives grp, time, and grp*time, with the interaction not significant.
However, t-tests show the groups differ at both pre and post treatment,
and split file matched samples t-tests show that DV changed across time in
the headache but not control group. the t-tests were done independant to
the RMANOVA because io dont know how to do them as a post-hoc procedure
within RMANOVA. How do i do it?