For my mediation model I am running the Sobel, Aroian, and Goodman Test in State. My question is, are you supposed to include your covariates in the equation?

sgmediation y, mv(x) iv(x)

or

sgmediation y, mv(x) iv(x) cv(c1 c2 c3...)

I am asking because when I do not include my covariates, the tests reveal my mediation to be insignificant. However, whilst including my covariates the tests are significant..

Similar questions and discussions