[R] Overdispersed poisson - negative observation

Peter Dalgaard BSA p.dalgaard at biostat.ku.dk
Thu Jan 16 17:38:03 CET 2003


"vito muggeo" <vito.muggeo at giustizia.it> writes:

> Roughtly speaking, your code is correct, but you can not fit your data with
> negative value:
> 
> > glm.qD93 <- glm(paym ~ alpha + beta, family=quasipoisson())
> Error in eval(expr, envir, enclos) : Negative values not allowed for the
> quasiPoisson family
> 
> Anyway, in general, quasi-likelihood is a way to handle (low) overdispersion
> in count data: for modest or large amount of overdispersion you could use
> negative binomial models, namely glm.nb() in MASS or gnlr() in the Jim
> Lindsey packages.

But they certainly wouldn't want to see negative responses either...

You should (I think) be able to get away with simply deleting the
check inside the quasipoisson family function.

        -p 

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907




More information about the R-help mailing list