[R] accounting for overdispersion in poisson distribution with lmer procedure

[Aurelie bocher]

Hello,

I am analysing counts data with a mixed model using lmer procedure. I 
therefore use the quasipoisson distribution but I'm not sure if this is 
sufficient to account for overdispersion. Actually the results are not very 
different to what I get when specifying a poisson distribution although my 
data are clearly overdispersed.

this my model:
>model <- lmer(NB ~ T + volume + (G|col) , data=immuno15, 
>family=quasipoisson)
where
- NB is bacterial count in haemolymph of ants 10h after bacteria 
injection(measure of immunocompetence)
- T is the treatment applied to ants
- volume is the volume of haemolymph sample from ants
- col is the colony of origin of ants, which is a random factor

I know it's possible to get a estimator of overdispersion with:
>sum.model <- summary(model)
>sum.model at sigma

but I don't know how to interpret this scale parameter.

Does anyone could help me?
thanks a lot,
Aurelie Bocher

_________________________________________________________________

mobile comme sur PC ! http://mobile.live.fr/messenger/bouygues/