Message-ID: <20110615155331.82691fhmh2aklbvo@webmail.ics.forth.gr>
Date: 2011-06-15T12:53:31Z
From: vlagani at ics.forth.gr
Subject: Jointly test multiple terms in a MCMCglmm model to be zero
Hello listmembers,
I am interested in testing whether multiple coefficients of a MCMCglmm
model are all zero. In particular, I have a categorical term in the
model with four different values, and I want to check if this term is
significant or not (that means, I want to check if *any* of the three
coefficients encoding this term is statistically different from zero).
Please accept my apologizes if this question has already been
answered, but I spent two days looking for an answer across mailing
lists, MCMCglmm help pages and vignettes, and I have not found
anything (maybe I am just not so good in searching!)
So, here a small example code:
#loading the libraries
require(MCMCglmm)
require(nlme)
require(lme4)
#setting the contrasts
contrasts(BodyWeight$Diet) = contr.sum(3)
#fitting the models: I am interested in the Diet term, that is a categorical
#variable with 3 values
model.lme <- lme(weight ~ Time * Diet, data = BodyWeight, random= ~ Time)
model.lmer <- lmer(weight ~ Time * Diet +(Time|Rat) , data = BodyWeight)
model.mcmc <- MCMCglmm(weight ~ Time * Diet, data = BodyWeight,
random= ~ Time, verbose=FALSE)
#checking the significance of the terms: the anova function will provide a
#p-value for the Diet term in the lme model, and an F value for the lme4 model
#How to obtain a similar statistic/pvalue for the MCMC model?
anova(model.lme, type = 'm')
anova(model.lmer, type = 'm')
Thanks in advance,
Vincenzo