Does stata have a package that can handle regression discontinuity designs with binary outcomes (in this case: likelihood of completion of further education)? I want to perform a local polynomial (4th order) model and estimate bandwidth, bias-correction and confidence intervals and have looked into - rdrobust - but it is my impression that it only performs linear probability models that will likely result in biased estimates for binary outcomes. I guess what I am asking for is a RD package with logistic regression? Or an alternative (and equally easy) approach? Or do you suspect that rdrobust will do the job just fine?
Thank you so much!
Thank you so much!
