Dear SPM experts,
I'm using SPM to run a between-group comparison of VBM grey matter images from a data cohort, with very unbalanced group size (60 subjects vs. 1000 subjects). I wonder whether I should worry about this unbalance and account for it by using undersampling (will loose power though).
For now I'm involving all subjects in a multiple regression design, the first regressor being the group membership (1s and 0s); the other regressors include gender and acquisition sites (regressors of no interest).
Could you kindly let me know whether my procedure is correct, and is the unbalanced group size a problem?
Best wishes,
Zuo
-----------------------------------------------------
MRC SGDP Centre
Institute of Psychiatry, Psychology & Neuroscience
King's College London
|