Hello, I am trying to understand canonical correlation analysis. I am confused. Sources say that, use canonic analysis especially when there is correlation between dependent variables; on the other side, they say avoid multicollinarity in both sets. I cant interpret this situation. If correlation exists between dependent variables, doesn't that imply multicollinarity? I also wonder using factors(obtained froma factor analysis)as independet variables create a problem in canonical correlation analysis? I will be very glad for help. Thanks a lot.