Dear Doug et al.,
I have 2 groups, 2 genders, and 2 scanners (A and B). (I also have
Age as a covariate). In my FSGD they are listed in this order:
ControlFemaleA
ControlFemaleB
ControlMaleA
ControlMaleB
CaseFemaleA
CaseFemaleB
CaseMaleA
CaseMaleB
I used the following contrast to test for a group X scanner
interaction: [1 -1 1 -1 -1 1 -1 1 0]
And I also looked at a group X scanner X gender interaction: [1 -1
-1 1 -1 1 1 -1 0]
For both, there are no significant results in either hemi. Would
this mean that I can instead just have 4 groups (Control/Case
Female/Male), or even just 2 groups (Control/Case)? Is that even
appropriate?
Thanks,
Chris