Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: gllamm option geqs


From   "Liberini, Federica" <F.Liberini@warwick.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: gllamm option geqs
Date   Tue, 31 May 2011 17:48:11 +0100

I have a question on gllamm. 

Say a have a simple random parameter model of the form

y_it = c + beta_i x_it + u_it

with 

beta_i = b + a z_i + v_i

where z_i is a time-invariant dummy for individual i.

What would be the difference between 

1. generating the interaction x_it*z_i and estimating the model with
gllamm specifying just "eqs eq1: x_it" before the gllamm command, and
then "nfr(eq1)" among the gllamm options and the interaction term among
the regressors 

and 

2. specifying "geqs geq1: x_it z_i" before the gllamm command and then
just the gllamm command with nfr(geq1) but no interaction term among
regressors?

Rabe-Hesketh and Skrondal have an example in their book of a similar
model (in paragraph 4.9), but they estimate it with xtmixed and using
the interactions following the reduced form of their model. I can't use
xtmixed, as I am estimating a random-coefficient dynamic panel probit.
But I was just wondering whether the geqs is appropriate (as it seems
from its definition).

Many thanks, 
Best

Federica

*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index