Hi all,
I've got a question about propensity score matching. It's a bit complicated but I tried to simplify it below:
I have two different regressions I've conducted so far: one fixed effects panel regression (-xtreg-) and one panel probit regression (-xtprobit)
One of the independent variables of the probit model is the residuals of the first regression.
I've been told that the outcomes of these two regressions might be endogenously determined. My questions:
1) Can I detect self selection bias by utilising propensity score matching?
2) If I can, what is the correct pathway to follow?
3) Which part exactly needs to be corrected in case of detecting self selection?
4) If I need to create a matched sample and run the first regression on this matched sample, does this overcome the problem?
Thank you so much.
0 Response to Detecting endogeneity or self selection bias using propensity score matching
Post a Comment