R help - Oct 2013 - cluster option in stata for random intercept model in the R language?

Dear R-list,

I am currently working on a dataset with a colleague who uses stata.
We fit a random intercept model to the data (decisions clustered in
participants) and get closely the same results in stata (using xtreg re) and R
(using the lme4 or multilevel package).


Now in stata, there is an additional option for the regression to control for
clustering; the vce(cluster clustvar) option, which changes the standard errors
quite a bit.
(see
http://www.stata.com/support/faqs/statistics/standard-errors-and-vce-cluster-option/
or http://www.stata.com/manuals13/xtxtreg.pdf).

Unfortunately I don't understand what this 'correction' does and why
it yields different results.
First I thought it would control for autocorrelations over time (decisions), but
if I model this directly with a random-intercept random-slope model, I don't
get nearly the same results.

Can someone help me understand what stata is doing here?
And what would be the equivalent in R to get similar results?


thanks!

On Oct 15, 2013, at 3:32 AM, Martin Batholdy wrote:
> Dear R-list,
>
> I am currently working on a dataset with a colleague who uses stata.
> We fit a random intercept model to the data (decisions clustered in  
> participants) and get closely the same results in stata (using xtreg  
> re) and R (using the lme4 or multilevel package).
>
>
> Now in stata, there is an additional option for the regression to  
> control for clustering; the vce(cluster clustvar) option, which  
> changes the standard errors quite a bit.
> (see
http://www.stata.com/support/faqs/statistics/standard-errors-and-vce-cluster-option/
>  or http://www.stata.com/manuals13/xtxtreg.pdf).
>
> Unfortunately I don't understand what this 'correction' does
and why
> it yields different results.
> First I thought it would control for autocorrelations over time  
> (decisions), but if I model this directly with a random-intercept  
> random-slope model, I don't get nearly the same results.
>
> Can someone help me understand what stata is doing here?
> And what would be the equivalent in R to get similar results?
http://www.stata.com/statalist/

> _______
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
David Winsemius, MD
Alameda, CA, USA