Hi,
I am trying to run a Sobel test for mediation on a panel database to test for significance of indirect effects in a mediation. I am using the Baron & Kenny (1986) method.
My regress code looks like
svy: nbreg DV Mediator Predictor
Ive seen the “sgmediation” command created by Phil Ender, but it does not work with -svy-.
Any ideas on how to do this?
Much appreciated.