Re: st: Panel Multinomial Logistic Model

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Mon, 24/5/10, Lee, Albert wrote:
> I understand that there is not a stata command for
> multinomial logistic model for panel data estimation. 

Such a thing has been implemented in Stata in -gllamm-, see
-findit gllamm- and <www.gllamm.org>, and in Peter Haan and
Arne Uhlendorff (2006) "Estimation of multinomial logit 
models with unobserved heterogeneity using maximum simulated 
likelihood", The Stata Journal, 6(2):229-245.
<http://stata-journal.com/article.html?article=st0104>

> I wonder if the following can be done for a three-outcome
> categorical dependent variable (say, 0, 1, 2):
> 
> 1.  Estimate a panel logit for outcome=1, and predict
> the exp(xb+random effects), 
> 2.  Estimate a panel logit for outcome=2, and predict
> the exp(xg+random effects), and

The problem is that those random effects represent unobserved
variables on the group level, and you would expect that whatever
influences the choice between 0 and 1 will also influence the
choice between 0 and 2, i.e. the two random effects need to be
correlated. In your method the two will be uncorrelated. You
can add this correlation with the two methods, I refered to above.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   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/