[R] help with contrasts for a binomial 3-way GLM

From: Susana Zuloaga <szuloaga_aguilar_at_hotmail.com>
Date: Fri, 31 Oct 2008 12:21:28 -0600


Hi
 I am a new user the R and I am very grateful for all your help but.......

I have a problem and I can't resolve yet. I am trying to get the contrasts for a binomial 3-way GLM (T= 4 temperature, t= 2 time and c= 2 substrate levels, plus treatment control) in total they are 17 treatments.   

I have tried with the glht but this function only work for 1-way GLM,  

acacia<-cbind(g,N-g)

> acacia1<-glm(acacia~tratamiento,binomial)
 

> acacia2<-glht(acacia1, linfct = mcp(tratamiento = "Tukey"))

> summary(acacia2)
 

         Simultaneous Tests for General Linear Hypotheses  

Multiple Comparisons of Means: Tukey Contrasts    

Fit: glm(formula = acacia ~ tratamiento, family = binomial)  

Linear Hypotheses:

            Estimate Std. Error z value p value

B - A == 0 2.9073 1.0395 2.797 0.2417

C - A == 0 4.0860 1.0227 3.995 <0.01 **

D - A == 0 4.8458 1.0207 4.748 <0.01 ***

E - A == 0 3.0445 1.0362 2.938 0.1727

F - A == 0 3.3399 1.0306 3.241 0.0744 .........etc.  

If the GLM is with interactions don't work  

acacia3<-glm(acacia~T*t*c,binomial)

> glht(acacia3, linfct = mcp(T = "Tukey"))

Error in modelparm.default(model, ...) :

dimensions of coefficients and covariance matrix don't match

Warning message:

In mcp2matrix(model, linfct = linfct) :

covariate interactions found -- default contrast might be inappropriate    

Other function could be simtes but this was deprecated from package multcomp  

There are others options to get comparisons contrasts for model GLM with interactions? or this test don't exist ?  

I hope help me  

        [[alternative HTML version deleted]]



R-help_at_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. Received on Fri 31 Oct 2008 - 18:31:04 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Fri 31 Oct 2008 - 19:31:08 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive