Testing for multicollinearity in fixed effects model

[Janik Kaden]

Hello all,
I am running a meta analysis on WTP for wind energy. I am using Stata 16. I use the command meta regress with option fixed. Now I need to check for multicollinearity. For regular OLS with vce(cluster var) and aweight[inverse standard error for precision] I use the vif command. Unfortunately in the case of meta regress vif is just possible with the uncentered option, but I don't know if this is appropriate or how to interpret the results. Does anybody have a suggestion how I could test for multicollinearity in this case? I would be very grateful.

[Carlo Lazzaro]

Janik:
as oftentimes advised on this list, it is probably better to forget (quasi-extreme) multicollinearity altogether.
Just inspect the 95% CIs and check whether they look in line with the literature in your research field.

[Janik Kaden]

Thanks for your reply. I saw that this was discussed before and that it was often suggested to do what you proposed. I just wondered if there is a possibility, because my advisor for my thesis told me to check for multicollinearity.

[Carlo Lazzaro]

Janik:
I'm not aware of any test for testing multicollinearity after -meta regress-.
Perhaps you can consider running:

estat vce, corr