07-17-2017 07:53 AM - edited 07-17-2017 07:56 AM
I am running a 2SLS estimation with a continuous dependent variable and an ordinal treatment variable taken values 1,2 or 3, which I assume is endogenous. I use a binary instrument and standard LATE assumptions should hold.
My questions are how I interpret coefficients on the treatment variable fro proc syslin and whether it is possible to compute treatment effects for a specific value (say treat=1).
I use SAS EG 7.1 and have used the following code:
Proc syslin data=sample 2SLS; endogenous treatment; instruments Z X; first_stage: model treatment=Z X; second_stage: model Y = treatment+X; run; quit;