Hi, I am running several regression specifications using the package ivreghdfe (basically the same as reghdfe, but for when you need an instrument). My sample size is larger than 300,000; and given the iterative process that ivreghdfe uses for estimation, using something like DFBETA (postestimation command for leaving-one-out) seems computationally unfeasible. Are there any other ways I can find influential observations in my data?
-
Login or Register
- Log in with
