This has been an excellent course. Professor Jackman keeps a brisk pace through the materials, but also gave time generously to respond to questions. — participant from the U.S.
This course provides participants with a practical introduction to Bayesian statistical modeling and inference, with a particular emphasis on applications across the social sciences. It begins with a brief discussion of how Bayesian statistical inference differs from classical or frequentist inference in the context of simple, familiar statistical procedures and models, such as the inference for proportions and regression. Following these initial considerations, the bulk of the course focuses on applied, simulation-based Bayesian inference, using the free software packages R and JAGS.
As this course presumes a basic background in statistics, participants without any prior statistical knowledge might want to consider taking another quantitative methods course instead.
This course was offered in 2015.
This course provides participants with a practical introduction to Bayesian statistical modeling and inference, with an emphasis on applications in the social sciences. Beginning slowly, the course discusses how Bayesian statistical inference differs from classical or frequentist inference. It examines these differences in the context of such statistical procedures and models as one- and two-sample t-tests, the analysis of a two-by-two tables, one-way ANOVA, and regression.
Following these initial considerations, the bulk of the course emphasizes applied, simulation-based Bayesian inference. It highlights how Bayesian approaches have become particularly attractive for more complex models thanks to the computing power that is available to researchers today. The course specifically focuses on a set of algorithms known as Markov chain Monte Carlo (MCMC) algorithms that allow social scientists to tackle classes of problems that used to fall in the 'too hard' basket.
The course examines how MCMC algorithms make Bayesian inference feasible, discusses their strengths and weaknesses, and explains some of the pitfalls to avoid when deploying MCMC algorithms. Some of the applications that are considered include:
To make Bayesian analysis a truly integral part of the participants' statistical computing toolkit, the course supplements theoretical discussions with many examples and the active use of the free general purpose Bayesian analysis program JAGS and various R packages.
The course presumes a working knowledge of statistics. Familiarity with probability theory and measurement models would also be helpful, but is not formally required. Participants without any prior knowledge of statistics should consider a more basic quantitative methods course.
Participants are expected to bring a WiFi-enabled laptop computer. Access to data, temporary licenses for the course software, and installation support will be provided by the Methods School.
Jackman, Simon. 2009. Bayesian Analysis for the Social Sciences. Chichester: John Wiley & Sons.
Gelman, Andrew, John B. Carlin, Hal S. Stern, and Donald B. Rubin. 2013. Bayesian Data Analysis. 3rd edition. Boca Raton, FL: Chapman & Hall.
Gelman, Andrew, and Jennifer Hill. 2007. Data Analysis Using Regression and Multilevel/Hierarchical Models. New York, NY: Cambridge University Press.
Gill, Jeff. 2007. Bayesian Methods: A Social and Behavioral Sciences Approach. 3rd edition. Boca Raton, FL: Chapman & Hall.
Greenberg, Edward. 2008. Introduction to Bayesian Econometrics. 2nd edition. New York, NY: Springer.
Koop, Gary. 2003. Bayesian Econometrics. Chichester: John Wiley & Sons.
Lancaster, Tony. 2004. An Introduction to Modern Bayesian Econometrics. Malden, MA: Blackwell.