Dear Bugs-list,
I have to calculate an intraclass correlation coefficient between
judges in a complicated concordance pattern: two continuous
measurements on right and two on left in several subjects by several
judges. A naive model includes only main effects:
for(s in 1:Subject){for(c in 1:2){for(j in 1:Judge){ for(r in 1:Repetition){
y[s,c,j,r] ~ dnorm(mu[s,c,j,r], tau.w)
mu[s,c,j,r] < - theta + subj[s] + side[c] + judge[j] + repet[r]}}}}
theta ~ dnorm(0.0,.00001)
for(s in 1:Subject){sujet[s] ~ dnorm(0.0, tauS)}
for(c in 1:2){side[c] ~ dnorm(0.0,tauC)}
for(j in 1:Judge){judge[j] ~ dnorm(0.0,tauJ)}
for(r in 1:Repet){repet[r] ~ dnorm(0.0,tauR)}
tauS ~ dgamma(0.00001,.00001)
tauC ~ dgamma(0.00001,.00001)
tauJ ~ dgamma(0.00001,.00001)
tauR ~ dgamma(0.00001,.00001)
tau.w ~ dgamma(0.00001,.00001)
sigma.w < - 1/tau.w
sigmaS < - 1/tauS
sigmaC < - 1/tauC
sigmaJ < - 1/tauJ
sigmaR < - 1/tauR
icc < - (sigmaS + sigmaC + sigmaR)/(sigmaS + sigmaC + sigmaJ + sigmaR
+ sigma.w)
But how reflect the nesting design between the variables (repetition
in side in subject)? Can I use some interactions term, eg a [r,c] and
assign a multinormal prior but then how to calculate the ICC. Can I
use "fixed" effects?
Thanks in advance. With all my best regards and wishes for the new year,
Erik Sauleau
-------------------------------------------------------------------
This list is for discussion of modelling issues and the BUGS software.
For help with crashes and error messages, first mail [log in to unmask]
To mail the BUGS list, mail to [log in to unmask]
Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html
Please do not mail attachments to the list.
To leave the BUGS list, send LEAVE BUGS to [log in to unmask]
If this fails, mail [log in to unmask], NOT the whole list
|