[R] test statistic in anova.glm when quasi family is used

Ben Bolker bbolker at gmail.com
Mon Jan 31 17:40:52 CET 2011


Eiiti Kasuya <ekasuscb <at> kyushu-u.org> writes:

> 
> When quasi family (not quasipoisson or quasibinomial) is used in glm,
> what is the appropriate test statistic in anova.glm?
> Help of anova.glm tells “For models with known dispersion (e.g.,
> binomial and Poisson fits) the chi-squared test is most appropriate, and
> for those with dispersion estimated by moments (e.g., gaussian,
> quasibinomial and quasipoisson fits) the F test is most appropriate”. I
> assume that  F is appropriate in the case of quasi (not quasipoisson or
> quasibinomial). Is this correct?
> 
> Ei Kasuya
> 
  Yes.

  References are Venables and Ripley (i.e. MASS) and Crawley's
Statistical Data Analysis book.

  Ben Bolker



More information about the R-help mailing list