Hi FreeSurfer Experts,

 

I had a question regarding contrasts for univariate LME models.

 

I am interested in looking at the interaction of alcohol with change in volume across two time-points. Included in my model are covariates Gender, ICV, and Previous Drinking History.

 

My X matrix includes 7 columns:

Column 1 = 1s

Column 2 = time

Column 3 = alcohol

Column 4 = time*alcohol

Column 5 =  Gender

Column 6 = ICV

Column 7 = Previous Drinking History

 

My contrast is C = [0 0 0 1 0 0 0 ].

 

I have 4 different previous drinking history variables.  However, when I change column 7 with my different previous drinking history variables, the F and P values for the interaction are the same. Am I setting up my contrasts incorrectly?

 

Best,

 

Jessica Hua