That matrix will test for an effect of group regressing out the effect of age. I would first test for an interaction between group and age with
0 0 0 1 -1 0
0 0 0 1 0 -1
Assuming that there is no effect, then I would re-run with DOSS (different-offset, same-slope) with a contrast like
1 -1 0 0
1 0 -1 0

DOSS forces all the age slopes to be the same which makes differences in the group intercepts more interpretable

On 12/10/18 1:53 PM, Barletta, Valeria wrote:

Dear Freesurfers,

I am trying to run a GLM analysis with 3 groups (1)long disease duration MS, 2)short disease duration MS, 3)healthy controls) and 1 variable (age).

According to the formula provided on freewiki:

N classes (N of variables + 1),

is this contrast matrix correct?

1 -1 0 0 0 0

1 0 -1 0 0 0

thank you,

Valeria

From: Barletta, Valeria
Sent: Friday, December 7, 2018 2:23:10 PM
To: freesurfer@nmr.mgh.harvard.edu
Subject: glm analysis

Dear Freesurfers,

I am trying to run a GLM analysis with 3 groups (1)long disease duration MS, 2)short disease duration MS, 3)healthy controls) and 1 variable (age).

According to the formula provided on freewiki:

N classes (N of variables + 1),

is this contrast matrix correct?

1 -1 0 0 0 0

1 0 -1 0 0 0

thank you,

Valeria
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer