Hi there,
while implementing Donald's suggestion, I realised that there's another issue I don't fully understand. There are I guess 2 ways to subtract the before and after training for each participant:
1.) from my understanding, since it's a repeated measure, I need to account for that
EV1(scan) EV2(mean)
1 1
-1 1
contrast 1: 1 0
and I'd feed the cope into the group analyses. But what do I do with the RTs? My understanding is that if I were only interested in the interaction, I can do a simple subtraction of the measures and add them as covariate. But if I were also interested in the other effects, I'd need to demean (I hope that's right). How would I do this though, before the subtraction (and across groups) or after subtraction (again across groups). I suppose it's the former (or it's some other routine)...
2.) the other way though may be that I don't have to account for repeated measures, ie:
EV1(scan1) EV2(scan2)
1 0
0 1
contrast 1: 1 -1
in which case I could simply demean the difference RTs (ie RTs-retest minus RTs-test) across groups if needed.
Thank you very much for clarification on this issue.
Cheers,
Torsten
|