[Limdep Nlogit List] Sample selection question

[Paul J. Wolfson]

I have an unbalanced panel that I want to analyze via a sample selection model.

As is most common, I have two equations.  The dependent variable of the
selection equation, z, is the number of observations in the unbalanced panel
for each individual.  The dependent variable in the regression equation, y, is
a poisson variable that measures something about their experience each time
they are observed shopping.  The reason for considering this model is the
suspicion that the number of times an individual chooses to go shopping is
related to the measurement on their shopping experience. 

My inclination is to run:
1) an ordered probit for the selection equation;
2) calculate the inverse mills ratio for each individual from the corresponding
residual;
3) replicate it to the appropriate number of observations for that individual
so that it lines up with that person's data, plug it into a poisson regression.

Steps 1 and 2 are straightforward.
I don't know limdep well enough to perform 3.  How would I do that?

Also, while I'm asking for help, is this a reasonable approach or does it
strike readers as flawed?

Thank you,




-------------------------------------------------------
Paul Wolfson

Research Computing
Tuck School @ Dartmouth College

voice: 603-646-0519
fax:     603-646-1308