Hi,
I have a quick question regarding study design:
I am trying to do an unpaired t-test (using randomise) between 30 subjects with a disease and 17 controls. To introduce the nuisance covariate of disease duration, I tried to create a third column in the design.mat file with de-meaned disease duration values (e.g. -1, +3, -2, +1, etc). However, because the controls are disease-free, they were all given a value of zero for "disease duration". Running this analysis gave an abnormal coloured area across the whole brain.
As an illustrative example, here is the kind of design.mat I used (e.g. for 5 controls vs. 5 subjects), with the third column being disease duration:
1 0 0
1 0 0
1 0 0
1 0 0
1 0 0
0 1 -1
0 1 2
0 1 -3
0 1 2
0 1 5
Is there a better way to design the matrix?
Thanks,
|