Content area
Full Text
Published online: 4 August 2017
© The Author(s) 2017. This article is an open access publication
Abstract Bayesian parameter estimation and Bayesian hypothesis testing present attractive alternatives to classical inference using confidence intervals and p values. In part I of this series we outline ten prominent advantages of the Bayesian approach. Many of these advantages translate to concrete opportunities for pragmatic researchers. For instance, Bayesian hypothesis testing allows researchers to quantify evidence and monitor its progression as data come in, without needing to know the intention with which the data were collected. We end by countering several objections to Bayesian hypothesis testing. Part II of this series discusses JASP, a free and open source software program that makes it easy to conduct Bayesian estimation and testing for a range of popular statistical scenarios (Wagenmakers et al., this issue).
Keywords Hypothesis test * Statistical evidence * Bayes factor * Posterior distribution
Theoretical satisfaction and practical implementation are the twin ideals of coherent statistics. Dennis Lindley, 1980.
The psychology literature is rife with p values. In almost every published research article in psychology, substantive claims are supported by p values, preferably ones smaller than .05. For instance, the December 2014 issue of Psychonomic Bulletin & Review featured 24 empirical brief reports, all of which reported p values. The dominance of the p value statistical framework is so complete that its presence feels almost prescriptive ("every empirical article in psychology shall feature at least one p value."). In Part I of this two-part series we aim to demonstrate that there exists a valid and feasible alternative -Bayesian inference- whose adoption brings considerable benefits, both in theory and in practice.
Based on a superficial assessment, the continued popularity of p values over Bayesian methods may be difficult to understand. The concept of p value null hypothesis statistical testing (NHST) has been repeatedly critiqued on a number of important points (e.g., Edwards, Lindman, & Savage, 1963; Morrison & Henkel, 1970; Mulaik & Steiger, 1997; Wagenmakers, 2007), and few methodologists have sought to defend the practice. One of the critiques is that p values are often misinterpreted as Bayesian posterior probabilities, such that it is all too easy to believe that p < .05 warrants the rejection of the null...