st: Sartori selection model - a good choice?

[Jørgen Carling <Jorgen@prio.no>]

Dear all,

Sartori’s selection model (“sartsel”, Sartori 2003) appears to be exactly what I need: a model with a binary outcome and identical varlists in the selection equation and the outcome equation. However, I am concerned that it is poorly documented and rarely used. The very short help file says nothing about weight options, for instance (which I need). Sartori’s article has been cited >200 times, but searching for use of her Stata model yields next to nothing.

The closest alternative I have been able to find is “biprobit” with the “partial” option (seemingly unrelated bivariate probit regression as a partial observability model), but this comes with a warning in about poor convergence properties. I would like not to use “heckprob” or other models that require different varlists for the two equations.

So, by questions are as follows:
1)	Is Sartori’s “sartsel” more widely used than it seems?
2)	Does anyone have experience using “sartsel” with weighting?
3)	Is “biprobit, partial” a sensible well-documented alternative? Or are there other possibilities?

I am not a hard-core number-cruncher, so any pointers to instructive examples wold be very welcome. Selection models with a single varlist for both equations are discussed briefly in Baum (2006), but other than that, I have found nothing in the Stata literature. Am I overlooking anything?

Many thanks for your consideration.
Jørgen

References:
Baum, CF. (2006) An Introduction to Modern Econometrics Using Stata, Stata Press
Sartori, A. (2003) An Estimator for Some Binary‐Outcome Selection Models Without Exclusion Restrictions, Political Analysis, 11(2): 111-138.



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/