[R-sig-ME] weighting observations in a MCMCglmm model?

[Malcolm Fairbrother]

I have actually been wondering just the same thing about MCMCglmm. (This question came up once before: http://permalink.gmane.org/gmane.comp.lang.r.lme4.devel/6095.)

Among other things, weights would allow for ordinal models to be fit MUCH faster. Combinations of covariates and outcomes could each be listed by row, with the row's weight reflecting the number of cases with each unique combination (as Rune suggested for clmm: http://permalink.gmane.org/gmane.comp.lang.r.lme4.devel/8744).

Anybody (probably Jarrod) know? If this isn't possible now, it would be useful inclusion in a future version (though maybe this would be more difficult than I realise).

Cheers,
Malcolm





> Date: Sun, 28 Oct 2012 21:25:58 -0700 (PDT)
> From: Giancarlo Sadoti <gcsadoti at yahoo.com>
> To: "r-sig-mixed-models at r-project.org" <r-sig-mixed-models at r-project.org>
> Subject: [R-sig-ME] weighting observations in a glmmadmb model?
> 
> Greetings list members,
> 
> Is there a way to weight observations in a glmmadmb model?  glm, lmer, etc. allow a "weights= ..." argument, but glmmadmb doesn't and I haven't found a work-around.  Many thanks.
> 
> Giancarlo