Dear statalisters,
I'm running a probit model that has a regressor that is correlated
with the error term of the latent model. So I could use the ivprobit
command, but the problem is that this regressor also enters the model
with a square term (this is one of my hypothesis) and I'm not sure how
to deal with it.
I searched the statalist archives and found an email
(http://www.statalist.eu/statalist/archive/2003-11/msg00795.html)
where it was recommended to do the following:
ivreg2 y q (x xsquared = z xhatsquared)
xhatsquared: square of the predicted values of a first estimation of x
on q and z
But since this was for a continuous dependent variable and for the
case where y and x were simultaneously determined, I'm not sure if I
could use:
ivprobit y q (x xsquared= z xhatsquared), mle
Would this be correct? And would it also be valid if I used z^2, q^2
and q*z instead of xhatsquared?
Please, any help would be much appreciated.
*
* 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/
