Lecture 38—Wednesday, April 18, 2007

What was covered?

Bayesian statistics—why should a frequentist care?

The Gibbs sampler

These are readily obtainable in hierarchical models when viewed from a Bayesian perspective. The basic regression model is a conditional statement, random effects are conditional on the values of parameters of a normal distribution, and all the remaining parameters are conditional on their priors.


i.e., the probability of going from  to  in n steps in the limit depends only on the final state. When a Gibbs sampler is formulated from the likelihood and priors from a Bayesian model, the limiting probability  is in fact the desired posterior probability.

The BUGS program

Course Home Page

Jack Weiss
Phone: (919) 962-5930
E-Mail: jack_weiss@unc.edu
Address: Curriculum in Ecology, Box 3275, University of North Carolina, Chapel Hill, 27516
Copyright © 2007
Last Revised--April 19, 2007
URL: http://www.unc.edu/courses/2007spring/enst/562/001/docs/lectures/lecture38.htm