R help - Jan 2010 - Quasi-Poisson regression - using parameter estimates for QAICc

Quasi-Poisson regression - using parameter estimates for QAICc

Hello,
I am using lmer (package lme4), for a GLMM, where I am modeling overdispered
data with 1 random effect and several fixed effects.

I want to use QAICc for my model selection, however I have 2 concerns

1) I don't know how to properly estimate the overdispersion parameter
(c_hat), which is needed to calculate QAICc.  

I believe this is done via the deviance provided and the DF (in my case I
have 31 obs. but only 22 are unique, therefore 22 minus the # of parameters
in model).  Is this correct (deviance/df)? 

If the overdispersion parameter is supposed to be a parameter in the model,
why isn't this included in the output, like in glm?

2) The model is fit  via Laplace approximation.  Are the logLik provided in
the output suitable for calculating the QAICc?

Thanks,
Dave
-- 
View this message in context:
http://n4.nabble.com/Quasi-Poisson-regression-using-parameter-estimates-for-QAICc-tp1015848p1015848.html
Sent from the R help mailing list archive at Nabble.com.