It turns out that what I thought were constraints in
the model were actually normalizations performed after
estimation. So my original question doesn't make any
sense. Sorry for the confusing post.
Owen
--- Owen Haaga <[email protected]> wrote:
> Dear Statalisters,
> I am trying to write a program which generalizes the
> censored ordered probit model (inspired by Edoardo
> Masset's earlier post). Can anyone tell me how the
> cutpoints on the regular oprobit are constrained? I
> have seen the ordered probit model used with either
> the first or the "middle" cutpoint constrained to be
> zero. I am currently constraining the cutpoints to
> be
> centered around zero (ie. the lowest cutpoint is -1
> times the highest cutpoint), but I would like to be
> able to use the results from a regular oprobit as
> the
> intial values for my model. I can rescale the cut
> points from oprobit's e(b) to center around zero
> (maintaining them the same "probability" apart), but
> I'm still working on rescaling the coefficients. It
> would be a lot easier if I knew how the cutpoints in
> the oprobit were constrained originally, and maybe
> in
> that case I could use the same method in my own ml
> estimation, for consistency.
> Any advice would be greatly appreciated. I'm afraid
> I
> can't make head or tail of the built-in ado files
> for
> oprobit.
>
> Thank you,
> Owen
> *
> * 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/
*
* 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/
