[R-sig-ME] log of offset for Poisson?
bbolker at gmail.com
Fri Apr 22 02:32:26 CEST 2011
On 11-04-21 08:25 PM, benton at stat.berkeley.edu wrote:
> I am pretty sure that I need to take the log of the offset variable when
> fitting a Poisson model using glmer, but I was hoping someone might be
> kind enough to confirm this.
Yes. Offsets are added on the scale of the linear predictor.
When in doubt about this sort of thing, make up a little simulated
example for yourself and see whether you recover the right answer.
More information about the R-sig-mixed-models