PSYC 607 (Inferential Statistics)


Bayesian Statistics

I have taught Inferential Statistics and Regression at the graduate level since 2012.  These are the two courses graduate students in the Psychology department at Texas A&M University are required to take their first year.

Recently I spent a large amount of time learning about Bayesian inference and Bayesian data analysis.  This approach makes much more sense to me, and I gave a short course to graduate students and faculty at Texas A&M in May 2017.  The course was based on John Kruschke’s (excellent) book Doing Bayesian Data Analysis: A Tutorial with R, JAGS, and Stan. The book’s website is here.

I also recommend JASP.  It can basically replace SPSS to perform traditional analyses and it also easily performs Bayesian analyses from the General Linear Model, yielding Bayes Factors that can replace p values.

The course covers Bayesian Inference and includes hands on lectures that cover both the model comparison approach that utilizes Bayes Factors, and the parameter estimation approach that focuses on 95% highest density (or credible) intervals.  I favor reporting Bayes Factors instead of p values and 95% HDIs for parameters instead of confidence intervals.

I have provided the power points for the course lectures to facilitate the field’s switch to Bayesian approaches.  I highly recommend purchasing Kruschke’s book as well.   The JASP tutorial is based on a tutorial E.J. Wagenmakers and others have written and made available on the Open Science Framework here.

Day 1 Lectures

Introduction: Credibility, Models, and, Parameters

Bayes’ Rule

Markov Chain Monte Carlo (MCMC)

Null Hypothesis Significance Testing (NHST)

Bayesian Approaches To (NHST)

Day 2 Tutorials

Doing Bayesian Data Analysis with JASP

Doing Bayesian Data Analysis with R, JAGS

*Mac users had trouble getting the JAGS code to work until they installed the latest version of iquartz (or something like that).