Dear all,
I have a doubt about how to test interaction in my full factorial
second level design.
I performed a 2 x 2 x 2 ANOVA, and I' m interested in some interactions.
Let's call my factor A B and C. So I’ve got A1, A2, B1, B2, C1 and C2.
If I want to know the effect of interactions between factor A and B, I
need to know the voxels significantly related to A1B1, A2B1, A1B2 and
A2B2.
Let's set the contrast with the weight:
A1B1= (1 -1 1 -1) - (1 -1 1 -1) = 1 -1 1 -1 -1 1 -1 1
A2B1=-(1 -1 1 -1) (1 -1 1 -1) = -1 1 -1 1 1 -1 1 -1
A1B2= (-1 1 -1 1) - (-1 1 -1 1) = -1 1 -1 1 1 -1 1 -1
A2B2= - (-1 1 -1 1) (-1 1 -1 1) = 1 -1 1 -1 -1 1 -1 1
As you can see the contrast vector for A2B1 is equal to the contrast
vector for A1B2, as for the contrast vectors for A1B1 and A2B2...so of
course I gain the same activations.
But what I' m looking for is to disentangle the activation related to
MATERIAL1 when is performed PROCESS 1 (A1B1), to MATERIAL2 when is
performed PROCESS1 (A2B1), to MATERIAL 1 when is performed PROCESS 2
(A1B2) and MATERIAL2 when is performed PROCESS 2 (A2B2).
Somehow, I’d like to do a post-hoc. Is there any way to do so?
Thank for your attention and sorry for the length of the mail
--
Federico Nemmi
Dipartimento di Psicologia,
Università La Sapienza, Roma
|