On 11/19/2012 03:04 AM, Mahinda Yogarajah wrote:
Dear Experts,
I have a two groups - a control group and patient group. The patient group can also be divided into 4 subgroups (but not control gp). Both groups are evenly split between male and female. So far I have compared the patient and control groups as a whole (age and icv as continuous/nuisance variables) but now i want to compare each of the subgroups to controls. I have tried doing this in qdec very crudely by running 4 analyses in qdec where I exclude all the subgroup cases execpt for those belonging to the subgroup I am looking at. My questions are
- Is this valid (I would correct for the multiple tests) ? How does qdec cope with exclusions given that variables (age) have been demeaned across a whole group.
whether it is valid or not depends on what you are trying to conclude. If you want to compare across the tests, I would do the demeaning once across all subjects and not individually for each test.
- Is there a better way of comparing my subgroups to controls ?
Not that I can think of
Thanks.
Mahinda
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer