On Wed, 9 Jul 2014 18:25:07 +0200, Kate SPM <[log in to unmask]> wrote:
>Hi SPM list,
>
>I am doing a VBM analysis to look at grey matter differences in two
>different groups.
>
>To test the interaction between group and a certain genotype (dichotomized)
>I created two different models.
>
>1. T-test group 1 vs group 2 with covariates gene coded 0 and 1
>(interaction with factor 1, center on factor 1 mean) and age (no
>interaction). Giving me the following columns:
>group1 group2 gene*group1 gene*group2 Age
>
>2. Full Factorial model with two factors (group and genotype), each having
>two levels. Adding age as a covariate. This gives me the following columns.
>
>group1gene1 group1gene2 group2gene1 group2gene2
>
>When testing the main effect of group in both models I get different
>results from the following contrasts:
>model 1: 1 -1 0 0 0
>model 2: 1 1 -1 -1 0
Your contrasts are testing the main effect "group" but not the interaction. The correct contrasts for interaction might be:
model 1: 0 0 1 -1 or 0 0 -1 1
model 2: 1 -1 -1 1 or -1 1 1 -1
Regards,
Christian
>
>I would have expected them to give the same results. What am I missing?
>Hope someone can explain.
>
|