Hey all,

I am doing a research about the effect a dummy has on multiple dependent variables while always controlling for the same control variables. The control variables consist of both variant and invariant variables. The panel data consists of multiple entities with data over multiple years.

When conducting the random effects regressions for the dependent variables and the dummy, all regressions return a significant coefficient for the dummy. However when we add the control variables to these regressions, two of the three regressions become insignificant (both for the dummy variable as for the control variables). One regression, remains significant for the dummy variables and for two out of four control variables. I was thinking about a multicollinearity problem but I don't understand how that is possible because we use the same control variables for all three regressions. Furthermore, I can't seem to find a method to measure for multicollinearity for a random effects model.

Can anybody help me with this?