Basically, I am confused how I got an interaction effect for Age * Emotion using 3dMVM but cannot find that effect on any post-hoc tests I've done. In 3dMVM I used the option -wsE2 because Emotion has 3 levels, although now I am seeing that in the 3dMVM help page there is conflicting information on this option. In one place it says:
"This option only makes sense if a within-subject factor has more than
3 level."
In another place it says: "Use the option only if at least one within-subject factor has more than
two levels."
Did I misuse this option and possibly get a spurious interaction effect here?