I'd
like to create a design matrix for a group analysis outside of the DODS
and DOSS models. I understand that in order to do this the -X flag must
be used. However, I have been unable to find examples of how to do this.
I
am hoping to reveal a difference in thickness or gyrification amongst a
clinical population. The data set contains two factors: diagnosis, and
study site and one covariate: age. Diagnosis has two levels: controls,
and patients. Study site has four levels, one level for each location
the data has been collected from.
What I would ideally like to do is:
1) Take into account offset differences amongst diagnosis and study site.
2) Allowing a difference in age slope amongst the diagnosis levels
3) Modeling the age slope as the same for the study site levels
My FSGD file is designed as follows
Class SITE 1-Control
Class SITE 1-PATIENT
Class SITE 2-Control
Class SITE 2-PATIENT
Class SITE 3-Control
Class SITE 3-PATIENT
Class SITE 4-Control
Class SITE 4-PATIENT
Variables age_at_scan
study site levels = 1,2,3 and 4 diagnosis levels = PATIENT and Control