Doing Bayesian Data Analysis: An Introduction
- John Kruschke, Indiana University
Bayesian data analysis is rapidly supplanting traditional statistical methods because it provides richer inferences from empirical observations, without having to resort to ill-defined probability values in hypothesis tests. This workshop introduces participants to modern Bayesian methods. We will begin with the basic ideas of probability and Bayes' rule. After that, we move on to cover probability distributions, grid approximation, Markov chain Monte Carlo methods, and Bayesian approaches to some specific statistical models (e.g., the multiple linear regression model, ANOVA, contingency table analysis, hierarchical models). Along the way, we will consider additional topics, including null hypothesis significance testing, Bayesian model comparison, Bayesian assessment of null values, and statistical power. Upon completion of this workshop, participants should be able to incorporate Bayesian tools into their own research projects and data analyses.
Prerequisites: It would be helpful if participants have had previous exposure to the basic concepts of multiple regression and ANOVA. And, they should be comfortable with summation and integral notation. No matrix algebra will be used in the workshop. Data analysis for this class will be carried out in the R statistical software environment, but no prior experience with R is necessary.
You are encouraged to bring a notebook computer to the course, but it is not required. If you bring a computer, then you will need to install some free software to run the data analyses. Please install the software before arriving at the course. The programs are being updated, so please check here a week before the course to be sure you have the most recent programs. For complete installation instructions, please refer to this page.
Fee: Members = $1600; Non-members = $3000
Location: ICPSR -- Ann Arbor, MI
Date(s): June 20 - June 23
Time: 9:00 AM - 5:00 PM