[R] glmmML vs. lmer - fitting overdispersed Poisson outcome.
Martin Henry H. Stevens
HStevens at muohio.edu
Wed Oct 17 12:40:46 CEST 2007
I don't understand how you can estimate a separate id variance other
than the residual variance, with one observation per subject. I would
think the proper model would be
glm(y ~ 1, family='poisson)
On Oct 15, 2007, at 12:05 PM, Sam Field wrote:
> I have count data with one observation per subject. I would like
> to fit
> a glmm to these data in order to account for overdispersion in the
> outcome. The lmer() function does not appear to be able to handle data
> that have only one observation per-cluster id, even though separate
> variance components for the count and normal portions of the
> outcome are
> identifiable. The glmmML() function does not seem to have problems
> this (as the code below illustrates).
> u <- rnorm(100)
> y <- rpois(100,exp(u))
> id <- seq(1:100)
> summary(glmmML(y~1,cluster=as.factor(id),family =poisson))
> I would like to use lmer() rather then glmmML(). Does anybody know
> of a
> way of getting the lmer() function to work with these kinds of data?
> Samuel H. Field
> Division of Internal Medicine - University of Pennsylvania
> CHERP - Philadelphia VA Medical Center
> 3900 Woodland Ave (9 East)
> Philadelphia, PA 19104
> (215) 823-5800 EXT. 6155 (Office)
> (215) 823-6330 (Fax)
> R-help at r-project.org mailing list
> PLEASE do read the posting guide http://www.R-project.org/posting-
> and provide commented, minimal, self-contained, reproducible code.
Dr. Hank Stevens, Associate Professor
338 Pearson Hall
Oxford, OH 45056
Office: (513) 529-4206
Lab: (513) 529-4262
FAX: (513) 529-4243
"E Pluribus Unum"
If you send an attachment, please try to send it in a format anyone
can read, such as PDF, text, Open Document Format, HTML, or RTF.
Please try not to send me MS Word or PowerPoint attachments-
Why? See: http://www.gnu.org/philosophy/no-word-attachments.html
More information about the R-help