[R] incorrect MCMC CIs in pvals.fnc (languageR) ?

Mike Allerhand michael.allerhand at ed.ac.uk
Fri May 11 15:16:25 CEST 2007


library(lme4)
library(coda)
library(languageR)

fit = lmer(Reaction~Days + (1|Subject) + (0+Days|Subject), 
data=sleepstudy)

pvals.fnc(fit)$random

# compare with...

samp = mcmcsamp(fit, n=10000, trans=FALSE)
HPDinterval(samp)
densityplot(samp, plot=F)

# 'pvals.fnc' reports sigma instead of sigma^2, but it looks like the
# Sbjc.(In) and Sbjc.Days are also sqrt compared with the corresponding
# results from HPDinterval and densityplot. Is that intended?


best wishes,  Mike Allerhand



More information about the R-help mailing list