A polemic for Bayesian statistics

Abstract

In the early part of the 20th century, forefathers of current statistical methodology were largely Bayesians. However, by the mid-1930's the Bayesian method fell into disfavor for many, and frequentist statistics became popular. Seventy years later the frequentist method continues to dominate. My purpose here is to compare the Bayesian and frequentist approaches. I argue for Bayesian statistics claiming the following: 1) Bayesian methods solve a wider variety of problems; and 2) It is sometimes difficult to interpret frequentist results. © 2008 Springer-Verlag Berlin Heidelberg.