External Email - Use Caution
Hi all FS users, I have the basic doubt of including age and gender as co-variates for a cortical thickness comparison between 2 groups. I read extensively many mails answering these questions but still could not get a satisfactory answer for the question. As I have 2 discrete variables i.e. Group with 2 levels and gender with 2 levels and one continuous variable of age my contrast matrix should be like this: [0 0 0 0 0 0 0 0] which is [malegroup1 femalegroup1 malegroup2 femalegroup2 age.malegroup1 age.femalegroup1 age.malegroup2 age.femalegroup2] I understood that if I have to test the effect of only the groups my contrast matrix should be: [0.5 0.5 -0.5 -0.5 0 0 0 0] What should be my contrast matrix if I have to include age and gender as covariates respectively? I know gender cannot be included as columns of 1s and 0s by my reading of previous mails. In one of the mail I read that if we want to test for interaction of group and gender (disregarding effect of age) the contrast matrix file should be: [0.5 -0.5 -0.5 0.5 0 0 0 0] But is testing for gender as a covariate same as testing group*gender interaction? I do not feel so. Hence requesting you all to throw some clarity on this. If this is not the way to study gender as a covariate how else should the contrast matrix be written? And regarding age should it be four 0.5s for the last 4 columns of the matrix if we want to study age as a covariate? Kindly let know. Finally I want a contrast matrix that studies the group differences in cortical thickness covarying for the effect of age and gender. Kindly help me.
Thanks, Dr. Rakshathi Basavaraju (MD Psychiatry) Postdoctoral Research Scientist Taub Institute Columbia University Irving Medical Centre 622 West, 168th Street, New York - 10032 E-Mail: rb3419@cumc.columbia.edumailto:rb3419@cumc.columbia.edu http://www.wellcomedbt.org/fellowsprofile/dr-rakshathi-basavaraju-288