Re: st: 'xtgee' for logistic regression

["Nora Ghaemi" <nghaemi@hotmail.com>]

sarah,

no i haven't time-series (panel) data.
i have just variables like in a cross-sectional (or case-control) study, 
with only one measurement for each variable, and i am trying to repeat the 
kind of logistic analysis (taking clustering into account) that other people 
apparently performed using a gee model with cross-sectional data.

using the cluster option of logistic are gee estimations used at all in the 
stata ado file?

my question is the following: is it possible at all to rearrange 
cross-sectional data so that gee-glm models can be used? or this is an 
impossible operation to do? may be duplicating the one-time measurements so 
to have kind of longitudinal data, or in some other way?
in sum, is there any possible (even a bit artificial) way to use gee-glm 
models for one-time measurements data?

thanks for your help.

nora







> From: "Sarah A. Mustillo" <smustillo@psych.duhs.duke.edu>
> Reply-To: statalist@hsphsun2.harvard.edu
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: 'xtgee' for logistic regression
> Date: Mon, 19 Jul 2004 11:09:47 -0400
>
> Nora -
>
> Do you have time-series data?  Just from what you've said, I'm guessing a 
> GEE model might not be your best option for what you are trying to do.
>
> The logistic command does have a cluster option.  So, you could do:
>
> logistic dep_var indep_var, cluster(famvar)
>
> If this doesn't help, write back with more details about what you are 
> trying to model.
>
> Sarah
>
> Nora Ghaemi wrote:
>
>
> > dear statalisters,
> >
> > i have to run a logistic regression that has to take into account 
> > clustering of subjects into families.
> > the command 'logistic' works but does not account for clustering, so the 
> > standard errors are smaller that they should be.
> >
> > i am trying to use on the same dataset the command xtgee:
> >
> > xtgee  dep_var indep_var, family(binomial) link(logit) i(id)
> >
> >
> > but invariably i get this response from stata:
> >
> > estimates diverging (absolute correlation > 1)
> > r(430);
> >
> >
> > should I rearrange the data somehow in order to run �xtgee�? it was my 
> > understanding that xtgee does not need to �tsset�
> >
> > can anyone help me with this issue?
> >
> > thanks a lot!
> >
> > nora
> >
> > _________________________________________________________________
> > Planning a family vacation? Check out the MSN Family Travel guide! 
> > http://dollar.msn.com
> >
> > *
> > *   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/
> --
> Sarah A. Mustillo, Ph.D
> Department of Psychiatry and Behavioral Sciences
> Duke University School of Medicine
> Box 3454
> Durham NC 27710
>
> 919 687-4686 x231
> *
> *   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/
_________________________________________________________________
Discover the best of the best at MSN Luxury Living. http://lexus.msn.com/

*
*   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/