R help - Feb 2008 - Overdispersion, AIC and mixed models

Hi 

I want to do use AIC for model selection on mixed model. However, before
going deeper in the model selection, I want to assess is there is
overdispersion with the full model in order to decide if I should use QAIC
instead of AIC.

Is there a way to compute a chisquare goodness of fit test for mixed model
calculate the variance inflation factor (sensu Burnham and Anderson,
chisquare/DF)?

Or should I use another approach to evaluate overdispersion in mixed model?

Thanks

 

Julien 


Checked by AVG Free Edition. 

10:55

Julien Martin <julien.martin2 <at> usherbrooke.ca> writes:
> 
> Hi 
> 
> I want to do use AIC for model selection on mixed model. However, before
> going deeper in the model selection, I want to assess is there is
> overdispersion with the full model in order to decide if I should use QAIC
> instead of AIC.
> 
> Is there a way to compute a chisquare goodness of fit test for mixed model
> calculate the variance inflation factor (sensu Burnham and Anderson,
> chisquare/DF)?
> 
> Or should I use another approach to evaluate overdispersion in mixed model?
> 
> Thanks
> 
> Julien 
  Julien,
    have you tried the r-sig-mixed mailing list with this yet?
Franklin et al 2002 divide the residual deviance by the residual df
in this context ...


Franklin, A. B., D. R. Anderson, and K. P. Burnham. 2002. Estimation of
long-term trends and variation in avian survival probabilities using random
effects models. Journal of Applied Statistics 29, no. 1:267-287.