<>
I don't see why you come to that conclusion. Austin pointed out that
your estimates are inconsistent, so you need an estimator that makes
them consistent. Once you do so, there is nothing wrong with the
marginal effects generated from that model for dy/dZ, as long as you
have a consistent estimate of the coefficient on X*Z.
Kit Baum, Boston College Economics and DIW Berlin
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html
On Jan 16, 2009, at 02:33 , Conner wrote:
Thanks, that is the same conclusion I came to, but it didn't
seem like it could be right. It implies that even in the
ideal setting where we can exogenously vary explanatory
variables of interest, we are still unable to consistently
estimate marginal effects if there are any interactions
between our exogenous and endogenous variables. How
depressing!
*
* 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/
