Message-ID: <x2k73xqjf1.fsf@biostat.ku.dk>
Date: 2004-01-12T15:24:34Z
From: Peter Dalgaard
Subject: GLMM (lme4) vs. glmmPQL output
In-Reply-To: <JLEPLGAANFCEAEDCAGJNIEDLCHAA.dieter.menne@menne-biomed.de>
"Dieter Menne" <dieter.menne at menne-biomed.de> writes:
> I have compared glmmPQL, glmmML, geese and GLMM, results and code see below.
> I am aware that glmmPQL uses another method to handle the problem, and
> geese (geepack) has considerable different assumptions, but the
> results are very similar. On the other hand, I had expected that glmmML
> results, if reasonable at all, should be close to GLMM. Yet they are not,
> but rather come close to the other three.
I suspect that a small simulation study would be enlightening. Given
the experimental status of lme4, I wouldn't feel too sure that there
is agreement between theory and implementation. There might be a bug
there, or maybe all the other methods make essentially the same
(large) error. In either case, I'd certainly like to know the reason.
--
O__ ---- Peter Dalgaard Blegdamsvej 3
c/ /'_ --- Dept. of Biostatistics 2200 Cph. N
(*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk) FAX: (+45) 35327907