[R-sig-ME] glmm.admb {glmmADMB} vs glmer {lme4}

[Douglas Bates]

On Wed, Oct 21, 2009 at 4:24 AM, Raldo Kruger <raldo.kruger at gmail.com> wrote:
> Hi mixed modelers,
>
> I'm using glmm.admb and glmer functions to fit mixed models on the same data
> (downloadable here
> <http://www.castafile.com/get/03a1935310acf4a7271b09cbfbf477d1>) and trying
> to assess which model provides the better fit.
>
> m1<-glmm.admb(Counts~T*Year+B*Year+P*Year, random=Site, group="Year",
> data=ex1o, family="nbinom", zeroInflation=TRUE)
>
> m2<-glmer(Counts~T*Year+B*Year+P*Year*(1|Site), data=ex1o,
> family=quasipoisson)

Should that formula be Counts~T*Year+B*Year+P*Year + (1|Site)?  I
don't know what a term like P*Year*(1|Site) in the formula passed to
glmer would do but it should throw an error.

>
> My questions are:
>
> 1) How can I extract the AIC values from m1?
>
> 2) Are the AIC values comparable between the two models (i.e. can I compare
> them for model selection)?
>
> 3) For m2, the true estimates for the fixed effects can be calculated by
> exp(returned estimate). Is this true for m1 too, or does the negative
> binomial distribution require a different conversion?
>
> 4) What is the difference between the 'random' and the 'group' argument in
> the glmm.admb function (I've read the documentation but it's still unclear)?
>
> Much appreciated,
>
> Raldo
> MSc student
> University of Cape Town
>
>        [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>