Skip to content

Sample selection using multiple logit or similar

1 message · Andrew Perrin

#
Greetings-

I am seeking to fit a model using Heckman-style selection; however the 
wrinkle is that the selection is into multiple categories, not a binary 
in/out. In this case, selection is into the type of higher-education 
institution a student attended; the goal is to estimate post-graduation 
outcomes while correcting for selection into universities.

There is an old Stata file, svyselmlog, that may or may not solve this 
problem in stata. Any guidance or pointers toward solving it in R would 
be most welcome.

Thanks-
Andy Perrin