You can get Stata ado files that can accommodate two-way clustering of
standard errors with linear regression and logit models from either
of these websites
http://www.kellogg.northwestern.edu/faculty/petersen/htm/papers/se/se_programming.htm
http://personal.anderson.ucla.edu/judson.caskey/data.html
Padmakumar
On Tue, Dec 8, 2009 at 11:29 AM, Stas Kolenikov <[email protected]> wrote:
> On Tue, Dec 8, 2009 at 10:47 AM, Marina Richardson <[email protected]> wrote:
>> I am performing Forward and Backward Stepwise Binary Logistic Regression Models for
>> which I would like to account for clustering at two levels. The option for clustering at one
>> level is simply cluster(PATID) however I would like to also control for clustering at DOCID.
>> Does anyone know if this is possible and what the syntax is for the command? Thank you.
>
> You only need to cluster at the highest level. If you want something
> more sophisticated, you could try -xtmelogit-, although it would be
> considerably slower due to numeric integration it performs (and you
> probably want something fast enough for your model selection
> purposes). I hope you've read Harrell's (2001) Regression Modeling
> Strategies tips and know what you are doing with all of this stepping
> around.
>
>
> --
> 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/
>
*
* 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/
