Diogo B. Provete
2010-Nov-25 11:53 UTC
[R] Generalized linear models with categorical data
An embedded and charset-unspecified text was scrubbed... Name: n?o dispon?vel URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20101125/6e4fdccc/attachment.pl>
Hi: On Thu, Nov 25, 2010 at 3:53 AM, Diogo B. Provete <dbprovete@gmail.com>wrote:> I got a question about using a GLZ with categorical x categorical data. > Below there is a data set I want to know the influence of treatments (CONT, > and LPS2H LPS24H) on the categories of pigmentation of the right testis of > an amphibian. From these data, I used the function glm with binomial family > (logit). But in the result (see below) is not possible to know the > influence > of the three treatments on the categories, only the treatments as a whole. > Could someone help me? >You have three response categories, which makes the response multinomial, not binomial. with(td, table(Tratamento, Categoria)) Categoria Tratamento Cat.1 Cat.2 Cat.3 CONT 0 4 6 LPS24h 5 5 0 LPS2h 0 3 7 Why is it that for each treatment, the responses only fall into two of the three categories? Is this by happenstance or by design? Methinks some information is lacking at the moment...but I'm pretty sure the binomial model is not correct in its current manifestation. Dennis Thanks in advance> > > > td=read.table(file.choose(), h=T) > > td > Tratamento Categoria > 1 CONT Cat.2 > 2 CONT Cat.2 > 3 CONT Cat.2 > 4 CONT Cat.2 > 5 CONT Cat.3 > 6 CONT Cat.3 > 7 CONT Cat.3 > 8 CONT Cat.3 > 9 CONT Cat.3 > 10 CONT Cat.3 > 11 LPS2h Cat.2 > 12 LPS2h Cat.2 > 13 LPS2h Cat.2 > 14 LPS2h Cat.3 > 15 LPS2h Cat.3 > 16 LPS2h Cat.3 > 17 LPS2h Cat.3 > 18 LPS2h Cat.3 > 19 LPS2h Cat.3 > 20 LPS2h Cat.3 > 21 LPS24h Cat.1 > 22 LPS24h Cat.1 > 23 LPS24h Cat.1 > 24 LPS24h Cat.1 > 25 LPS24h Cat.1 > 26 LPS24h Cat.2 > 27 LPS24h Cat.2 > 28 LPS24h Cat.2 > 29 LPS24h Cat.2 > 30 LPS24h Cat.2 > > > mod.test=glm(Categoria~Tratamento, family=binomial(logit), data=td) > > > mod.test > > Call: glm(formula = Categoria ~ Tratamento, family = binomial(logit), > data = td) > > Coefficients: > (Intercept) TratamentoLPS24h TratamentoLPS2h > 2.057e+01 -2.057e+01 1.558e-08 > > Degrees of Freedom: 29 Total (i.e. Null); 27 Residual > Null Deviance: 27.03 > Residual Deviance: 13.86 AIC: 19.86 > > > anova(mod.test, test="Chisq") > Analysis of Deviance Table > > Model: binomial, link: logit > > Response: Categoria > > Terms added sequentially (first to last) > > > Df Deviance Resid. Df Resid. Dev P(>|Chi|) > NULL 29 27.034 > Tratamento 2 13.171 27 13.863 0.001380 ** > --- > Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 > > > > -- > Atenciosamente, > *Diogo Borges Provete* > > =============================> Biólogo > Mestre em Biologia Animal (UNESP) > Laboratório de Ecologia Animal > Departamento de Zoologia e Botânica > Instituto de Biociências, Letras e Ciências Exatas > Universidade Estadual Paulista - UNESP > São José do Rio Preto-SP > Brazil > > Rua Cristóvão Colombo, 2265 > Jardim Nazareth - 15054-000 > > *Skype*: diogoprovete > *MSN*: diogoprov@yahoo.com.br > > *Personal web page <https://sites.google.com/site/diogoprovetepage/>* > > Traduza conosco: > <<<American Journal Experts>>> <http://www.journalexperts.com/br/> > <<<D-Lang Soluções linguisticas< > http://www.d-lang.com.br/site/sitept/index.htm> > >>> > > =============================> > > > > -- > Atenciosamente, > *Diogo Borges Provete* > > =============================> Biólogo > Mestre em Biologia Animal (UNESP) > Laboratório de Ecologia Animal > Departamento de Zoologia e Botânica > Instituto de Biociências, Letras e Ciências Exatas > Universidade Estadual Paulista - UNESP > São José do Rio Preto-SP > Brazil > > Rua Cristóvão Colombo, 2265 > Jardim Nazareth - 15054-000 > > *Skype*: diogoprovete > *MSN*: diogoprov@yahoo.com.br > > *Personal web page <https://sites.google.com/site/diogoprovetepage/>* > > Traduza conosco: > <<<American Journal Experts>>> <http://www.journalexperts.com/br/> > <<<D-Lang Soluções linguisticas< > http://www.d-lang.com.br/site/sitept/index.htm> > >>> > > =============================> > [[alternative HTML version deleted]] > > > ______________________________________________ > R-help@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide > http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > >[[alternative HTML version deleted]]