[R-sig-eco] binomial regression with non integers

[Pierre THIRIET]

Hi Valérie,

If I well understood, you want to test differences in the ratio 
(multiple-site turnover component/multiple-site total dissimilarity) 
between groups of sites. I think that beta.sample() function could help 
you in comparing between groups of sites their multiple-site total 
dissimilarity and the respective contributions of the 2 components.  
With this function, you can get for each group of sites the boostraped 
confidence intervals of the 3 values(2 components and total 
dissimilarity). You could also plot the distributions of resampled 
dissimilarities .This is what Baselga, A. and C. D. L. Orme (2012) did 
in their exemple (see pasted code below).
I imagine that if you are interested in the ratio (turnover/total 
dissimilarity), this is because total multiple-site dissimilarities are 
really different between groups of sites. You could compute for each 
group (and so compare among groups) the bootstraped confidence interval 
of this ratio by combining the use of boot() and beta.multi(). However, 
the ecological meaning of this comparison of ratio is confused for me. 
What do you want to test?
If you are interested in caracterizing differences in assemblage 
composition between group of sites, I am wondering if another approach 
like dbRDA (with group as constraint) on each of the 3 dissimilarity 
matrices returned by beta.pair(feeded by the whole site x species matrix 
including every groups of sites) might give you better insights about 
how are differences between groups (turnover and/or nestedness). 
However, I read in Baselga and Orme 2012 that  multiple-site measures 
are different than averaged pairwise dissimilarities, this is why I am 
not sure on the suitability of this variation partionning approach. Does 
any one know if it is allowed and make sense?

Cheers,
Pierre


Baselga, A. and C. D. L. Orme (2012). "betapart: an R package for the 
study of beta diversity." Methods in Ecology and Evolution 3(5): 808-812.

library(betapart)
data(ceram.s)
data(ceram.n)
# get betapart objects
ceram.s.core <- betapart.core(ceram.s)
ceram.n.core <- betapart.core(ceram.n)
# multiple site measures
ceram.s.multi <- beta.multi(ceram.s.core)
ceram.n.multi <- beta.multi(ceram.n.core)
# sampling across equal sites
ceram.s.samp <- beta.sample(ceram.s.core,
sites=10, samples=100)
ceram.n.samp <- beta.sample(ceram.n.core,
sites=10, samples=100)
# plotting the distributions of components
dist.s <- ceram.s.samp$sampled.values
dist.n <- ceram.n.samp$sampled.values
plot(density(dist.s$beta.SOR),
xlim=c(0,0Æ8), ylim=c(0, 19), xlab='Beta
diversity', main='', lwd=3)
lines(density(dist.s$beta.SNE), lty=1, lwd=2)
lines(density(dist.s$beta.SIM), lty=2, lwd=2)
lines(density(dist.n$beta.SOR), col='grey60',
lwd=3)
lines(density(dist.n$beta.SNE), col='grey60',
lty=1, lwd=2)
lines(density(dist.n$beta.SIM), col='grey60',
lty=2, lwd=2)



Le 15/02/2013 13:37, v_coudrain at voila.fr a écrit :
> Dear all,
>
> I am investigating diversity in different sites. I partitioned my measure of diversity into to additive components (Baselga 2012) and get for each site a value of overal
> diversity change (between 0 and 1) and a value for each additive component, such that for each site beta1+beta2=beta_total. I would like to make a regression
> model to test if the proportion of diversity due to beta1 (beta1/beta_total) is signifcantly different according to an explanatory factor. If beta1 had been an integer
> value, I would have used a binomial model. However, since beta1 is not an integer I don't think that I am allowed to use the formel
> glm(cbind(beta1,beta2)~x,family=binomial)? What alternative method could I use?
> I hope that my question is not too confuse.
> Thank you very much.
> Valérie
> ___________________________________________________________
> C'est l'année du Serpent ! Connaissez-vous votre signe astral chinois ? Découvrez-le ici http://astrocenter.voila.fr/voila/Presentation.aspx?product=StEdCH2K2&Af=-3000
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology