[Limdep Nlogit List] Ordered Probit Selection Model: Can sampling weights be used?

james benson jbenson at ssc.wisc.edu
Thu Jan 21 04:18:32 EST 2010

Does anyone know if it is possible to employ sampling (probability) 
weights with either the selection model, the outcome model, or both, 
when implementing the ordered probit sample-selection modeling strategy?

I am using the two-step Heckman-style ordered probit selection model 
with a continuous outcome, according to section E31.4.6 of the 
Economometric Modeling Guide (LIMDEP 9). I am using a national database 
that provides (and recommends) sampling weights, and the weights do make 
a difference in my results when I run the selection and outcome models 
separately. I need to use the weights if possible.

If it is possible to implement sampling weights in this modeling 
strategy, I would greatly appreciate knowing what syntax to use!


More information about the Limdep mailing list