As I understand the literature on the -heckman- procedure, it is customary
to have different sets of covariates for the two equations in the two-step
model, i.e. the selection equation should have an additional covariate over
the outcome equation. See "MUS" http://www.stata.com/bookstore/mus.html,
16.6.5, page 546.
HTH
Martin
_______________________
----- Original Message -----
From: "Johanna Avato" <[email protected]>
To: <[email protected]>
Sent: Wednesday, January 07, 2009 8:57 PM
Subject: st: Heckman specification
Does anyone know whether it is mandatory to have the same variables in
both equations in the heckman two-step procedure (plus the iv in the
selection equation) or is it possible to drop one variable from the
selection equation only? I have a model where one variable causes major
collinearity problems in the selection equation.
THanks.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
