Fisher, Bayes, and predictive Bayesian inference

Sandy Zabell (Northwestern University)

Sandy Zabell, Northwestern University

Title: Fisher, Bayes, and predictive Bayesian inference.

Abstract: R. A. Fisher is usually perceived to have been a staunch critic of the Bayesian approach to statistics, yet his last book (Statistical Methods and Scientific Inference, 1956) is much closer in spirit to the the Bayesian approach than the frequentist theories of Neyman and Pearson. This mismatch between perception and reality is best understood as an evolution in Fisher’s views over the course of his life. In my talk I will discuss Fisher’s initial and harsh criticism of “inverse probability”, his subsequent advocacy of fiducial inference starting in 1930, and his admiration for Bayes expressed in his 1956 book. Several of the examples Fisher discusses there are best understood when viewed against the backdrop of earlier controversies and antagonisms