maggy yan <kiotoqq <at> gmail.com> writes: > > I read something on http://glmm.wikidot.com/faq, under "How can I deal with > overdispersion in GLMMs?": > > library(lme4) ## 1.0-4set.seed(101) > d <- data.frame(y=rpois(1000,lambda=3),x=runif(1000), > f=factor(sample(1:10,size=1000,replace=TRUE))) > m1 <- glmer(y~x+(1|f),data=d,family=poisson) > overdisp_fun(m1) ## chisq ratio rdf p > ## 1026.7780815 1.0298677 997.0000000 0.2497659 > library(glmmADMB) ## 0.7.7 > m2 <- glmmadmb(y~x+(1|f),data=d,family="poisson") > overdisp_fun(m2) ## chisq ratio rdf p > ## 1026.7585031 1.0298480 997.0000000 0.2499024 > > In both case, the chisq is > rdf, does it mean there is over dispersion? > > thanks for any help >
Off-topic here, but: the residual deviance is greater than the residual degrees of freedom, but only a little bit (3%). So, technically, there is overdispersion here, but not more than expected if the underlying data generating process was not overdispersed (p-value = 0.25). Which is a good thing because the data are generated from a Poisson distribution, so the null hypothesis is actually true. ______________________________________________ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.