I am using -eprobit, entreat()- because my dependent variable and endogenous treatment variables are both binary.
For marginal effects, I use -estat teffects, atet-.
I ran two regressions with different outcome variables (grade in class 5 and grade in class 10). The first model converged, but estat teffects took too long.
The second model did not converge even after 250+ iterations. My sample size is ~10,000. I am including regional fixed effects and clustered SE.
Any ways to speed things up? Or any alternatives? I used -treatreg- earlier, but the model does not allow for binary dependent variable, hence I am exploring -eprobit-.
For marginal effects, I use -estat teffects, atet-.
I ran two regressions with different outcome variables (grade in class 5 and grade in class 10). The first model converged, but estat teffects took too long.
The second model did not converge even after 250+ iterations. My sample size is ~10,000. I am including regional fixed effects and clustered SE.
Any ways to speed things up? Or any alternatives? I used -treatreg- earlier, but the model does not allow for binary dependent variable, hence I am exploring -eprobit-.
