David Klinowski<[email protected]>:
You can use a non-ordered model for ordered discrete variables, but
you lose some efficiency, I think. In any case, I think you want
-cmp- (ssc d cmp) and perhaps you would like to come see the tutorial
by its author in DC on July 30:
http://www.stata.com/meeting/dcconf09/abstracts.html#roodman
On Mon, Jul 6, 2009 at 10:14 AM, David Klinowski<[email protected]> wrote:
> Dear members of statalist,
> I am trying to estimate and ordered probit model which contains a
> dummy, endogenous independent variable. I have read previous
> discussions on the topic but none seem to apply to my case: if I
> understand correctly, I cannot use ivprobit because my endogenous
> variable is dichotomous (and its instrument is also a discrete
> variable with six possible values), and I cannot use neither mprobit
> nor biprobit because my probit model is ordered. Can anyone help me or
> correct me?
> Thank you,
> David
*
* 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/
