[BioC] Can I do model comparison with type III ANOVA?

[James W. MacDonald]

Hi Yonggan,

This has nothing to do with Bioconductor, so you should ask elsewhere. 
Note that type III models have been beat to death on R-help, so a simple 
search of that list's archive would be likely to answer your question.

Best,

Jim



On 6/19/2012 11:52 AM, Yonggan Wu wrote:
> Hi,
>
> I just found out type III ANOVA won't give me model comparison result,
> am I not suppose to do model comparison with type III ANOVA?
>
> Say we have two models
> m1=lm(data~factor1+factor2,data=table)
> m2=lm(data~factor1*factor2,data=table)
>
> Type I ANOVA comparison
> anv1=anova(m1,m2)
>
> Type III ANOVA Comparison, from car package
> anv2=Anova(m1,m2,type="III")
>
> anv1 give me correct result
> anv2's result similar to Anova(m1, type="III")
>
> Can anyone please explain why?
>
> Thanks,
> Yonggan
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor

-- 
James W. MacDonald, M.S.
Biostatistician
University of Washington
Environmental and Occupational Health Sciences
4225 Roosevelt Way NE, # 100
Seattle WA 98105-6099