[R] Offset in GLMM

Douglas Bates bates at stat.wisc.edu
Mon Feb 16 13:58:13 CET 2004

SAULEAU Erik-André <SAULEAUEA at ch-mulhouse.fr> writes:

> Dear R-list,
> I try to adjust GLMM on incidence cancer data. Without random
> effect, in GLM the command is, for example with sex effect,
> glm(Observed~sex+offset(log(Expected)),family=poisson) because the
> observed are Poisson distribued with parameter Expected*incidence
> rate. But know I want to introduce random effect (for example
> spatial effect) and it seems to me that the "offset" does not work
> in
> glmmPQL(Observed~sex+offset(log(Expected)),random=~1|ED,...,family=poisson)?
> I mean it does not work in lme.
> what is the problem with my offset? thank you in advance, with best
> regards,

Did you try function GLMM from package lme4?  I believe that GLMM
allows an offset.

More information about the R-help mailing list