If, yes, will you please share with me?

It will be very helpful for me.

Thanks. ]]>

It is important to distinguish between y_i ~ Normal(μ, σ^2) and y_i ~ Normal(μ_i, σ^2). The former is your normal distribution, since there is just a single mean μ. The latter is a distribution with variable mean μ_i, and if you generate data from the latter distribution, and plot their histogram, you won't see the familiar bell-shaped normal curve. But if you fit the y_i ~ Normal(μ_i, σ^2) model to the data, then you will see the curve in the residuals. ]]>

Can you elaborate on the below please?

You write:

There is no assumption about the distribution of the predictor xi or the response variable yi.

But you also mentioned - yi∼Normal(μ^i,σ2) which says that yi follows a normal distribution. I am sure I am interpreting something incorrectly. Would be great if you could let me know what I am missing here.

Thank you for this post!

]]>49: parallel.bugs <- function(chain, x.data, params, seeds)

66: working.directory=sub.folder, bugs.seed=seeds[chain])

74: sfLapply(1:3, fun=parallel.bugs, x.data=x.data, params=params, seeds=c(1,2,3))

I too found the same posteriors for the three chains. As you suggested, I tried to use the following in above code:

sfClusterSetupRNGstream(123)

sfClusterEval(rnorm(1))

I could not run the model by including the two lines in the parallel.bugs function. Can you suggest me how to avoid same starting values.

]]>