[R-sig-ME] contrasts for generalized linear model

[espesser]

  Dear Wade,

If I understand well your data, year (in fact,  TSB) is the (main ?)  
fixed effect, and is not a random term.
I suppose there are levels enough for pop.

I think your model would be something like  :

lmer(flowering ~ TSB +(1|pop)   ,family=binomial, data=....)


I would use successive difference contrast for the 3 levels factor TSB.
see contr.sdif()  in library(MASS).
Regards
Robert

Le 25/05/2011 17:15, Wade Wall a écrit :
> Hi all,
>
> This question has perhaps been answered before, but I can't find a reference
> to it.  I have a data set that consists of binomial data (plant flowering)
> and time since the plant was last burned (1, 2, or 3 years ago).  There
> appears to be a main effect ( flowering is reduced in year 1), but I want to
> compare years 2 and 3 to see if flowering is decreasing by year three.  The
> model that I included random effects of Year (year data collected) and
> population (multiple pops).
>
> The data set basically looks like this:
>
> Pop    Year    TSB     Flowering
> 1         2008     1             1
> 1         2008     1             0
> . . . . .
> 24       2010     3              0
> 24       2010     3              1
>
> What is the best way to set up contrasts for binomial data in R?  Is it
> basically the same as with an ANOVA?
>
> Thanks for any information.
>
> Wade
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models


-- 
Robert Espesser
CNRS UMR 6057 - Université de Provence
5 Avenue Pasteur
13100 AIX-EN-PROVENCE

Tel: +33 (0)442 95 36 26