Hi experts,
I've run ICA on a sample of 60 (20 controls, 20 group 1, 20 group 2). I'm aware that the groups should be balanced for DR (and I suppose for ANOVAs generally). Should I be worried about bias (against the controls) with this?
I need the 3 groups to demonstrate disease subtype specific differences, otherwise I'd just run two separate ICAs prior to the dual regression
Does this sound alright?
All the best,
Jonathan
|