Consumerland revisited
Buck CE, Cavanagh WG, & Litton CD (1996) Bayesian approach to interpreting archaeological data.  Wiley: Chichester. p154-159 

© Andrew Millard 2001 

A simple mixture model.  M is the  number of mugs, N is total  mugs + goblets, p.B is prior probability of being a bar.
alpha[i] and beta[i] summarise the expert's prior judgement about the proportions of mugs at bars (i=1) and restaurants (i=2)

model{
	for ( i in 1:2){
		pM[i] ~ dbeta(alpha[i],beta[i])
	}
	M ~ dbin(pM[shrine],N)
	B ~ dbern(p.B)  # mean of this node is posterior probability of being a bar
	R <- 1 - B           # mean of this node is posterior probability of being a restaurant
	shrine <-  2 - B

}


DATA
list(alpha=c(13,3.6667), beta=c(4,5), M=2, N=5, p.B=0.6667)