st: Stata codes for average partial effects for correlated random effect probit model.

[natasha agarwal <agarwana2@googlemail.com>]

Dear All,

I want to estimate a correlated random effect probit model given by
Woolridge (2002). I find it very easy to compute the model as I
introduce the time averages of each variable in the specification and
go ahead with the same. However, I find it very difficult to compute
the average partial effects of the same.

I have tried programming it but I am not very sure whether it is
correct or not. If anybody has used it before, any help will be
appreciated. My codes look as follows:

program probit_export, rclass
xtprobit expdum lnk alnk, re
predict xdh, xb
replace xdh = xdh/sqrt(1 + e(sigma_u)^2)
gen pe1 = normal(xdh)
summarize pe1, meanonly
return scalar ape1=r(mean)
drop xdh pe1
end

many thanks in advance
Natasha
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/