Dear Laura,
I'm more familiar with linear regression in the second stage, than
ordered probit. For OLS in the second stage you can calculate rho by
dividing lamdba (which is provided by the coefficient on your inverse
mills term) by sigma (the standard error of the residual, which Stata
can provide with: predict mysigma, stdr). When it comes to ordered
probit, I am not sure what the equivalent for sigma would be. Also,
the entries for both -heckman- and -heckprob- in the Stata manuals
note that Stata calculates atanh rho, rather than rho directly. Others
on the list will no doubt have an answer to this.
As for the standard errors, you might consider bootstrapping the whole
procedure.
Kind regards
Ian
Friday, August 5, 2005, 7:14:59 AM, you wrote:
ls> Dear Ian,
ls> Thanks a lot for your help and sorry for my late replay (was out of town).
ls> I tried to compute the two-step procedure you suggested for a binary probit
ls> and then compare results with heckprob command. Estimations seem reasonable
ls> although with the two-step procedure I am loosing some observations with
ls> respect to the ml estimation. At this stage I don’t think I fully
ls> understand how the heckprob command works and how the missing observations
ls> of the latent probit equation are treated (?)
ls> However, in this two step procedure I guess I also need to compute new
ls> standard error and rho once I introduce the inverse mills ratio. Do you
ls> think this is correct?
ls> Many thanks again for your precious help.
ls> Best wishes
ls> Laura
*
* 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/