Dear Karina,
SPM is there to help you specifying the necessary contrasts. Among other answers, there is this one in the mailing list archives: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=spm;82979a6.1110. There, Will Penny explains how to derive contrast weights for a full factorial ANOVA model. In general, you will have to add weights on your subject or group columns as well - (+/-)1/2 for the group columns, since you have 2 groups, (+/-)1/n1 for each subject in group 1 and (+/-)1/n2 for each subject in group 2. However, since the interaction contrasts are based on differences, the effects of the group and subject constants will cancel each other in some of the contrasts.
Hope this helps,
Volkmar
|