[R] anova on binomial LMER objects

Robert Bagchi bagchi.r at gmail.com
Wed Sep 28 13:13:56 CEST 2005

Hi Patrick

thanks for your advice. I have now tried glmmPQL, and it worked fine - 
I'm getting consistent results between plots and models fitted by 
glmmPQL. Plus it allows predict() and resid() which is another advantage 
over lmer at present.

quick question though: why does one need to use PQL for binomial models? 
Is there a good reference for this?

A few of my colleagues have also had similar problems, so I'm copying 
this message on to R-help as it might be useful there.

Many thanks

Patrick A. Jansen wrote:

> Hi dr Bacghi,
> I ran into exactly the same problem with lmer models that had an 
> rbind() response variable, as you posted to the R-list.
> The sums of squares produced by anova() seem wrong. They are almost 
> identical to the mean squares, and hence F-values approach 1.
> Since you need PQL for binomial models anyway, you might want to use 
> GlmmPQL instead. Seems to work fine with anova().
> Best regards,
> Patrick Jansen
> *dr Patrick A. Jansen*
> University of Groningen
> Community and Conservation Ecology group
> E P.A.Jansen at .rug.nl
> W _www.rug.nl/fwn/onderzoek/programmas/biologie/cocon_ 
> <http://www.rug.nl/fwn/onderzoek/programmas/biologie/cocon>
> c/o:
> Instituto Smithsonian de Investigaciones Tropicales
> Att. Patrick A. Jansen – Gamboa
> Apartado 0843-03092, Balboa, Ancón, Panamá, República de Panamá
> or:
> Smithsonian Tropical Research Insititute
> Att: Patrick A. Jansen – Gamboa
> Unit 0948, APO AA, 34002-0948, U.S.A.
> T +507-212-8904 (office) / +507-6516-2008 (cell)

More information about the R-help mailing list