If parttime is the indicator for whether an observation selects, then it
only takes on the value 1 in the regression for the selected subsample. SO
it can not be included. And it wouldn't really make sense even if it were.
Perhaps parttime is not dichotomous? Though Stata treats all values
greater than zero as positive outcomes in the selection equation.
On Wed, 18 Jul 2007, Ying Zhou wrote:
Date: Wed, 18 Jul 2007 12:47:32 +0100
From: Ying Zhou <ying.zhou@nuffield.ox.ac.uk>
Reply-To: statalist@hsphsun2.harvard.edu
To: statalist@hsphsun2.harvard.edu
Subject: st: Heckman selection model, collinearity problem
Hi there
I am trying to do a Heckman selection model, looking at the effect of
part-time contract on wages, while controlling for selection into
part-time work. The syntax I use looks something like this heckman
logwage educ age parttime, select (parttime = married children educ age)
But the output says that parttime is dropped (from the main equation) due
to collinearity. I need to keep the part-time variable in the main
equation in order to look at its effect on wages after controlling for
selection into part-time work. Does anyone know how to solve this
problem?
Thanks,
Ying
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
- Fred
Note updated contact info (effective August 01, 2007):
< ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> Frederick J. Boehmke
< Associate Professor of Political Science
> University of Iowa
< 341 Schaeffer Hall
> Iowa City, IA 52242
<
> Office: 361 Schaeffer
< Phone: 319 335 2342
> Fax: 319 335 3400
<
> frederick-boehmke@uiowa.edu
< http://www.fredboehmke.net
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/