[R-sig-ME] Partialing out the effect of a covariable in a Poisson GLMM

[Ben Bolker]

 <v_coudrain at ...> writes:

> 
> Dear all,
> I recenly asked a question about the use of an offset in a Poisson GLMM 
> and it came out that an ofset would be
> reasonable if the estimate of the covariable to be put 
> as an offset is about 1. However if this estimate is not close 
> to one, does anybody know about a partial linear
> regression equivalent for a Poisson GLMM? Or what 
> are your method to control for the effect of a covariable in a Poisson GLMM?

   Don't know exactly what you want to do, but presumably fitting
a "null model" with the covariate and then building more complex models
on top of this -- adding the other covariates -- and doing sequential testing
(anova(full_model,model_with_only_first_covariate)).  The linear models
idea of partialing out one variable and running a model on the residuals
doesn't work as well, unfortunately.

  Ben Bolker