st: Generating a confidence interval for marginal effects of a combination of outcomes

[Jed Cohen <jedcohen@vt.edu>]

Hello All,

I am trying to generate a confidence interval for a combination of
outcomes using the prchange command after an oprobit. I use stata
11.2.

I have 4 ordinal outcomes, with prchange I can get the marginal effect
that a discrete change in two dummy variables has on the probablility
of falling into each category, with a confidence interval. What I
would really like is to calculate the marginal effect, or change in
probability of falling into a combination of categories. ie. what is
the change in the probability of falling into categories 1 or 2 after
applying the treatment dummies vis a vis the baseline (without
treatment dummies).
A code sample:
prvalue, x(nimT1=0 nimT2=0 nimT3=0 germany=1) rest(min) save
prvalue, x(nimT1=1 nimT2=0 nimT3=0 germany=1 germanyT1=1) rest(min)  diff

which gives the change in the probability of falling into each
category from the base line to the treatment (nimT1=1, germanyT1=1),
with 95% CI. I can of course combine the point estimates to get an
estimate of the probability of falling into a combination of
categories, and then difference this from the same combination in the
baseline to get this marginal effect, but this does not have a CI
which is crucial in this case.
I believe I could bootstrap this manually, which would take a
looooonnnggg time and maybe not be so accurate, but I was hoping
someone would have a better way.

Thanks!
*
*   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/