Luca Tiberti--
-biprobit- can handle the first case of a binary dep var and
a binary endog var; see
http://www.stata.com/statalist/archive/2007-05/msg00345.html
and I suspect -qvf- (SJ3-4: st0049, st0050, st0051) could also produce
consistent estimates.
You might have to write your own estimator for the second case,
using code from -mlogit- or somesuch in the "second" stage.
Note that requests for help on Stata should be sent to Statalist, not
individuals on the list. Following the FAQ:
"
The Statalist convention is that you reply to the list, not to the
poster. Please do not request a direct personal reply, unless the
question refers to information about (for example) a job or paid
consultancy. Otherwise, you will bother the list with your question,
but the list does not benefit from any answers. If people reply
directly to you, consider posting a summary to the list. Go private,
however, if a thread turns into an arcane or convoluted dialogue
unlikely to be of wide interest, and then summarize to the list.
Continuing or closing a thread you started is important, especially by
answering secondary questions and by reporting what solved your
problem.
"
I have copied my response to Statalist.
On 6/5/07, [email protected] <[email protected]> wrote:
Dear Dr Nichols,
I am a PhD student at the University of Florence. For my thesis I am
studying the effectiveness of a social policy for children on some
outcomes in South Africa. I had contacted you a month ago for the RDD
(that I left because I have not the proper variables).
I am now writing to you again for a suggestion.
In my model of analysis I have a binary variable as outcome and a
binary
(endogenous) variable as regressor, plus many other explanatory
variables.
My question is which is the best command in Stata that takes into
consideration the characteristics of my model. I looked at the ivprobit
command but, from what I have understood, it seems that the explanatory
endogenous variable is treated as continuous and not binary (it is thus
not the proper command). In addition, in a different model using a
discrete variable as outcome (3 values) and the same dummy endogenous
treatment as before, which command can I use?
Thank you very much for any suggestion and I apologize for taking you
again time,
Best regards,
Luca
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
