On Thu, Mar 7, 2013 at 6:36 PM, Peter Maclean <pmaclean2...@yahoo.com>wrote:
> Is there a R package that use sampling weights in multilevel modeling? The
> survey package does not handle multilevel modeling and the weight option in
> lmer and nlmer functions from lme4 (used for multilevel modeling) is for
> weighted least squares estimation.
> Suggestion from one with experience in this subjet (including creating
> weights from strata and sampling unit variables) will be helpful. For
> example if analyzing data clustered in schools, how to use
> student's sampling weight or school sampling weight or both?
>
There is not.
What you need depends on whether you are interested in the variance
components for their own sake or just in valid standard error estimation
for a regression model.
If the latter, you can use svyglm in the survey package, which will give
you correct standard errors for a linear model whose coefficients have the
same interpretation as those from lmer.
If you need to estimate variance components, you currently need to use some
other software. Mixed-model estimation based on composite likelihood is on
my list of things to do, but not terribly high up.
-thomas
--
Thomas Lumley
Professor of Biostatistics
University of Auckland
[[alternative HTML version deleted]]
______________________________________________
R-help@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.
