st: instrumental variable for binary outcome panel data

[Michael Weinerman <michael.weinerman@gmail.com>]

Stata 10 appears to have binary outcome panel data commands and
instrumental variable commands, but not the two combined. I have tried
storing the residuals and comparing a "manual" 2nd stage estimate with
a continuous IV against the xtivreg estimates as a check on my
instrumentation, but the estimates are quite different... I used the
'first' command to make sure I was storing the results from the proper
regression, but I must not be storing the proper data. Suggestions?

Another question while I'm here. Is glamm still the best way to employ
ordered logit/probit to panel data? Most of the posts I've found seem
rather dated, but glamm seemed to be the consensus. I have a
trichotomous variable I would like to employ to compare to the
dichotomous one above.

Any help is greatly appreciated.

Best,
Michael
*
*   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/