What you can do in GLLAMM is an exchangeable correlation by specifying
random effects at the cluster level. GLLAMM is a full likelihood
routine rather than a GEE-type one, you need to model everything
explicitly. The identity correlation structure would probably come out
of -ologit, cluster()-.
On Tue, Jun 2, 2009 at 4:30 PM, Shubhabrata Mukherjee
<[email protected]> wrote:
> Hi,
>
> I-emailed statalist about a month back regarding longitudinal ordinal regression in STATA (as I couldn't find anything such as `xtologit').
>
> A few people recommended that I should use GLLAMM or ologit (with `cluster id' option). But there is no option to specify the type of covariance structure (unstructured, ar1 etc.) on either of them if I am not wrong. I think the cluster(id) option for ologit (or i(id) for GLLAMM) are just nesting the observations by their respective ids.
>
> I was wondering if anyone can point out some other models which could take care of this problem (specifying covariance structure in a longitudinal OLR)?
>
> Thanks,
>
> Joey
>
>
>
>
>
> *
> * 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/
>
--
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.
*
* 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/
