Reka,
gllamm really only estimates random effects
models.
The reason you get the same estimates with
and without the random effect is that
the random effect varies between units (id),
not between alternatives or categories
(within units). It will therefore
cancel out in the multinomial logit
probabilities:
exp(x_ib_a + u_i)/ sum_b(exp(x_ib_b + u_i))
=
exp(x_ib_a)/ sum_b(exp(x_ib_b))
where _i is a subscript for units
and _a and _b are subscripts for alternatives
or categories.
If you want to include random effects in a
multinomial logit model, you need to expand
the data first, use the expanded() option
and use dummy variables for the alternatives
in the equations for the random effects,
see Sections 9.1 and 9.3 of the gllamm manual
available at
http://www.iop.kcl.ac.uk/IoP/Departments/BioComp/programs/gllamm.html
Best wishes,
Sophia
At 07:00 AM 5/7/2003 -0500, you wrote:
>Dear Statalist,
>
>I want to estimate a fixed effects model using GLLAMM. I am wondering
>if the code below is correct:
>
>gen cons=1
>
>eq fe_c: cons
>
>gllamm Y X1 X2..., i(id) eqs(cons) link(mlogit) adapt
>
>I get the same estimates with or without eqs(cons) which leads me to
>believe that I am doing something wrong.
>
>Thanks for the help in advance,
>
>Reka
>
>Reka Sundaram-Stukel
>Graduate student
>320 Taylor Hall
>Department of Agriculture And Applied Economics
>Madison, WI-53706
>
>
>*
>* 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/
>
*
* 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/
