Software:

TLNise: TLNise stands for Two-Level Normal independent sampling estimation, and enables inference about the parameters of a 2-level Normal hierarchical model based on independent draws from their exact posterior distributions. This is in contrast to MCMC methods, which produces correlated draws from a Markov chain that converges to the exact distribution.

rcwish: The SPlus function rcwish generates random draws from the Wishart and related distributions.

To use TLNise in R, go to http://cran.r-project.org/web/packages/tlnise/index.html and download the tlnise R package. This does not require a Fortran compiler to use.

norm.hm: The S-Plus function norm.hm performs empirical Bayes inference for a two-level normal hierarchical model with univariate outcomes. The file normhm.txt contains the S-Plus source code, which must be read into S-Plus. The file normhm.pdf is a README file.