[R-sig-ME] How to get dispersion parameter from a binomial mixed model?

Ben Bolker bbolker at gmail.com
Tue Dec 11 22:12:56 CET 2012

 <v_coudrain at ...> writes:

>  I am modelling data with a generalized mixed model with binomial
> error distribution and I am concerned about overdispersion. I know
> that dispersion parameter can be measured as deviance/df, but for
> mixed models (= with random effect), the number of degrees of
> freedom cannot be extracted (I am using the function lmer from
> R). Is there any way to find out if my data are overdispersed?

  Have you searched for "overdispersion" at http://glmm.wikidot.com/faq
... ?  There is no perfect way to estimate the residual degrees of
freedom for a mixed model, it depends considerably on the question,
but since most estimates of overdispersion are approximations anyway
you can use (number of fixed-effect parameters + number of random-effect

More information about the R-sig-mixed-models mailing list