Re: st: heckprob and rho

[Antoine Terracol <Antoine.Terracol@univ-paris1.fr>]

Hello,
Since -heckprob- assumes normality of the residuals, you should use 
-probit- instead of -logit-.
Of course the results will not be exactly the same as in -heckprob- 
since rho will be constrained to equal zero.

hope this helps,
Antoine

M. Nizam Khan wrote:


> Hi statausers,
>
> I am trying to estimate "ever visit to health care provider" model with sample 
> selection. My probit equation is visit=a+xb+u1 and selection equation is 
> reported_sickness=c+zb+u2. In both cases, dependent variables are binary (1,0). 
> Variables used for identification are: smoker, ADL, use of hygenic latrine.
> These variables have very low correlation with "VISIT". 
> I used heckprob to estimate my model. What I would like to know: When rho is 
> not significant, should I run a simple logit model and interpret my results 
> based on that? I was expecting that when rho is not significant, coefficents 
> and SE (of visit equation) in the simple logit procedure should not be 
> different from heckprob procedure. In my case, I found it different. Any 
> suggestion on this would be appreciated.
>
> Thnaks.
>
> Nizam  
> *
> *   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/