eprobit and margins

[Asal Pilehvari]

Hi,

I am going to use eprobit for my purpose of regressing a binary variable(y) on another endogenous binary variable(x). The related instrument is called z. The Stata command that I use is as follows:

eprobit y x, endogenous(x= z, probit)

I am interested in marginal effects. However, when I use margins cmd, it gives me an error that "e(sample) does not identify the estimation sample".
I was wondering if anybody know what might be the problem.


Thanks.

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions. Knowing exactly your code and being able to replicate the problem will help us help you.

[Asal Pilehvari]

Thank you very much for notifying me about this.

This is my code:

local dvars "age gender Asian white hgheduc married hghincome"
eprobit dic_physhlth retir `dvars', endogenous(retir= Age_fr) vce(cluster id)
margins , dydx(*) post

And this is the error when I execute it:
"e(sample) does not identify the estimation sample"

Where dic_physhlth is a binary variable of health status, retir is a binary variable for whether individual is retired or not.