Normally it should be:
a <- mcmc.normal(rnorm(10),mu=0,tau=rep(c(1,3),each=5))
However, what I need is a group-wise tau.y, which is in turn an mcmc.gamma object. Say
tau.y <- mcmc.gamma(c(sd(y_Grp1),sd(y_Grp2,na.rm=TRUE)),alpha=0.1,beta=0.1)
Will this work as expected?
y.lik <- mcmc.normal(y,mu=y.hat,tau=rep(tau.y,each=5),observed=TRUE)
PS: I found rcppbugs recently and it seems really fast and useful.