Detecting mediation in multiple regression: Please help

[Salma Aboelmaaty]

Dear All

If I have a multiple regression model with two indepdent variables and I added a third one. This third one naturally caused co-efficients of other two to drop but it affected one variable more than the other and it also decreased its significance (it was still significant but it dropped from 0.00 to 0.007 or 0.012) .. Does that suggest a partial mediating role of this added variable?

Salma

[Øyvind Snilsberg]

not necessarily. it's more likely due to multicollinearity and fewer degrees of freedom. you might want to check out https://data.library.virginia.edu/in...tion-analysis/.

[Salma Aboelmaaty]

I checked for multicollinearity by doing VIF and its was 1.27 .. If there is no multicollinearity, does this interpretation count as a possibility?