Hi all,
I have a statistics question. I'm using the Glm module to build a design matrix for my DTI data. There are significant between-group differences in both age and years of education in my subject population (patients and controls). However, I found that these two variables are highly correlated (r=-.49, n=30, p<.01), with greater age associated with lower education. Do I need to add both factors as EVs in the Glm, or does controlling for one take care of both?
Best,
Kamy Wakim
|