Hello,

I have a mixed model with several independent variables which I also have the correlation matrix for. However, Field (2013) states that merely looking at this matrix is insufficient as it "misses more subtle forms of multicollinearity". Hence, I would like to calculate the VIFs for my variables.

My question is: I used the mixed model because I wanted to control for the feat that people repeatedly interacted within the same groups. If I now were to conduct a multiple regression without controlling for this violation of the assumption of independence, would the results for the VIF analysis still be reliable?

Or is there any other way I could get the VIFs?

Thanks a lot in advance!