[R] Questions about glmms.

[Horacio Montenegro]

	Hi,
	
	I have a couple of questions related to glmm (glmmPQL
in MASS and GLMM in lme4). 

	1) is there some way do obtain the fitted values by
group, similar to:

> predict(dbd.glmmPQL, dbd.cytdens, 
+       type="response", level=0)

	where dbd.glmmPQL is the fitted model and dbd.cytdens
is a data frame with a subset of the factors?
	
	2) when I double-click on a saved workspace (.RData
file) containing lme4 objects there is an error
message and R crashes. Instead, if I load lme4 and
then open the file, it loads smoothly. Is this a bug
or just a feature? I'm working with R 2.0.1, lme4
0.9-2, Matrix 0.95-3, on windows 98. Unfortunately, I
can’t provide the error message, as I don’t have R
installed on this computer, but I can provide it
later, if asked.
	
	3) Has the result of anova(dbd.glmmPQL) or
anova(dbd.GLMM) any meaning? If yes, what is it? I
mean, if I do an anova(glmmPQL.object) or
anova(GLMM.object), I do obtain an anova table that,
for me (a biologist), is easily interpretable. I know
that this question has been asked before, but, so far,
I did not find a comprehensive answer. I've read
Venables & Ripley MASS (2000), but glmmPQL is not
mentioned there, and I do not have access to the later
edition. Besides, in lme4 manual, under GLMM-class
there is an anova method, but there are no further
details explaining the method or its usage.

	cheers, 
	Horacio

--
Horacio Montenegro
PhD student in Genetics
Departamento de Genética e Evolução
UNICAMP
Campinas - SP - Brasil