Facebook       Blog       Contact Me
Eric Guntermann

Methods


Multilevel Bayesian Modelling

This is an introductory workshop on multilevel models. The idea is to show simple ways to run relatively complex multilevel models using Bayesian methods. The substantive example used will involve fitting models of party preferences along the lines of what van der Eijk et al. (2006) propose, but using a simpler modelling strategy. The focus will be on understanding multilevel models.

Advanced computer or mathematical skills are not necessary, although basic knowledge of regression is strongly recommended. Knowledge of R would be useful. You can learn R by watching the recording of the R workshop I gave in December 2013. Here's the link.

If possible, bring a laptop to the workshop. If you want to follow all the examples, you should install R, RStudio, JAGS, and the R packages listed below. You can also download the R code file, containing code and explanations of everything we cover, and the datasets we will use. You can get the slides below.

You can download R here and RStudio here.

For JAGS, go to this page

You also need to install R2jags, coda, R2WinBUGS, lattice, and rjags, in that order. You should do so in R by typing the following in R:

install.packages("R2jags", dependencies = TRUE)
install.packages(“coda", dependencies = TRUE)
install.packages("R2WinBUGS", dependencies = TRUE)
install.packages(“lattice", dependencies = TRUE)
install.packages(“rjags", dependencies = TRUE)

R code
Dataset
Slides


You can watch a recording of this workshop on Youtube: