Dear Yoram,
> I have a group of patients with 2 conditions for each patient. A)
> before treatment B) after treatment. For each patient I also have a
> psichiatric test : s
>
> I want to see if the effect of the treatment positively correlates
> with the test.
> I am using a condition and covariate effect and use the following 2
> contrasts:
>
> condition1 condition2 score
> 1 -1 1
> - 1 1 1
>
> Is my contrast O.K?
No. You are looking for a condition times score interaction. Your model
only inlcudes a main effect of score. You have to mulitply your score by
+1 for A and -1 for B scans and use a contrast
> condition1 condition2 score x condition
> 0 0 1
> 0 0 -1
With best wishes - Karl
|