Overview
John K. Kruschke is a professor at Indiana University Bloomington (Department of Psychological and Brain Sciences). He is best known for the Bayesian data analysis textbook Doing Bayesian Data Analysis (Academic Press, 2011; 2nd ed. 2015), colloquially called "the puppy book," which is a standard introduction to Bayesian methods for psychologists and social scientists. He advocates highest density intervals (HDI) over null-hypothesis significance testing and promotes Bayesian ordinal regression as the default for Likert and rating-scale data.
Key Contributions / Features
- Metric-vs-ordinal failure modes (Liddell-Kruschke 2019): co-authored the paper establishing that applying t-tests/ANOVA to ordinal data produces Type I errors, Type II errors, and directional inversions; demonstrated via sigmoid-curve analysis and Amazon movie-rating data.
- Bayesian Data Analysis textbook: Doing Bayesian Data Analysis (2015, 2nd ed.); introduces MCMC via JAGS/Stan for a psychology audience; includes chapters on ordinal regression, hierarchical models, and the Bayesian new statistics.
- HDI and ROPE framework: promotes credible intervals (HDI) combined with a Region of Practical Equivalence (ROPE) as a Bayesian alternative to classical hypothesis testing.
Related