st: gllamm conditional random intercept multinomial logit model

[kasakawa@ualberta.ca]

Hello,

I am trying to estimate a multilevel multinomial logit model using gllamm.

I have a model with a random intercept, which is a function of level-2  
covariates.  The gllamm manual has an example of a random intercept,  
but it seems that the example is an unconditional model (i.e. no  
level-2 covariates in the intercept).

I tried an unconditional random intercept model as following and it  
seems to be working:

-------
eq a2: a2
eq a3: a3
eq a4: a4
gllamm Y X1 X2, i(group) fam(bimon) link(mlogit) pwe(normw) eqs(a2 a3  
a4) nrf(3) adapt nip(4)
------

where Y is a dependent variable that consists of 4 groups. X1 & X2 are  
level-1 independent variables.

I am wondering if anyone can give me some advice on how to incorporate  
level-2 ccovariates in the above random intercept model so that I can  
get three sets of level-2 coefficients on intercepts (group 4 vs group  
1; group 3 vs group 1; group 2 vs group 1).

Also in the gllamm manual, an option "expand(   )" was used - is this  
option necessary for multinomial models?

Regards,
Keiko




*
*   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/