JAGSmeans.R

#There isn't really an easy way to do this, unfortunately.

#In your stated example, the populations are ordered, so you should be able to calculate the mean in JAGS like this:

Pop1.mean <- mean(data[1:3,1:10])
Pop2.mean <- mean(data[4:6,1:10])

#but maybe your actual dataset isn't organized that way, and even if it was you'd have to 
#manually go through the dataset to figure out the cutoffs.

#I've also done a hacky workaround as follows. Define new vectors, one for each population. 
#For the pop1 vector, the value is 1 if the individual is from population 1, 0 otherwise. 
#For pop2 vector, the value is 1 if an individual is from population 2, 0 otherwise, and so on. 
#Feed these vectors in as data. You'll also need to know the number of individuals in each population n1 and n2.

pop1 <- c(1,1,1,0,0,0)
pop2 <- c(0,0,0,1,1,1)

n1 <- n2 <- 6

nindividuals <- 6
nmeasurements <- 10

#Then, in your JAGS model:

#Make a new set of nested loops or stick in existing one, it shouldn't matter
for (i in 1:nindividuals){
  for (j in 1:nmeasurements){
    pop1mean.hold[i,j] <- pop1[i]*data[i,j] #note that this will be the data value if in pop1, 0 otherwise
    pop2mean.hold[i,j] <- pop2[i]*data[i,j] #and so on for other populations
  }
}

#Then, outside loop, manually calculate the mean:
pop1mean <- sum(pop1mean.hold[])/(n1*nmeasurements) #all non pop 1 are zeroes and so are ignored
pop2mean <- sum(pop2mean.hold[])/(n2*nmeasurements) #note that you need to include the empty brackets in sum or JAGS won't like it