Now the problem is the inconsistency between your model specification
-bsVars 'Group+Age+Gender' \
-wsVars Task \
-qVars Age \
and the following three post hoc tests
-gltLabel 4 CON-PDL -gltCode 4 'Group : 1*CON -1*PDL Gender : 1*F -1*M' \
-gltLabel 5 CON-Atypical -gltCode 5 'Group : 1*CON -1*Atypical Gender : 1*F -1*M' \
-gltLabel 6 PDL-Atypical -gltCode 6 'Group : 1*PDL -1*Atypical Gender : 1*F -1*M' \
Your model assumes that no interactions exist between Group and Gender, but the above three post tests are asking 3dMVM to provide you the interaction effects! It seems that you want a different model:
-bsVars 'Group*Gender+Age' \
-wsVars Task \
-qVars Age \
This new model still assumes that the same Age effect across groups and genders. A even more generic model would be
-bsVars 'Group*Gender*Age' \
-wsVars Task \
-qVars Age \
Gang
Edited 1 time(s). Last edit at 02/27/2015 02:13PM by Gang.