Re: [R] glm poisson and quasipoisson

I am sure, that this is not a pure Poisson! Huge overdispersion! You get inflated confidence intervals! (although, the point estimates of the regression coefficients stay the same) Try to look for the causes of overdispersion! It may be geteroscedastisity? What is the nature of the response, is

Re: [R] glm poisson and quasipoisson

On Thu, Jan 31, 2013 at 2:13 PM, Wim Kreinen wkrei...@gmail.com wrote: Hello, I have a question about modelling via glm. I think you are way off track. Either the data, glm, or both, are not what you think they are. I have a dataset skn300.tab - structure(list(n = 1:97, freq = c(0L, 0L,

[R] glm poisson and quasipoisson

Hello, I have a question about modelling via glm. I have a dataset (see dput) that looks like as if it where poisson distributed (actually I would appreciate that) but it isnt because mean unequals var. mean (x) [1] 901.7827 var (x) [1] 132439.3 Anyway, I tried to model it via poisson and