Default Bayes Factors for Nonnested Hypothesis Testing

Bayesian hypothesis testing for psychologists: A tutorial on the Savage–Dickey method

“…In order to increase the robustness of Bayesian hypothesis testing to the vagueness of the prior, several procedures have been proposed, including the local Bayes factor (Smith & Spiegelhalter, 1980), the intrinsic Bayes factor (Berger & Mortera, 1999;Berger & Pericchi, 1996), the fractional Bayes factor (O'Hagan, 1995), and the partial Bayes factor (O'Hagan, 1995; for a summary see Gill, 2002, chap. 7).…”

Objective Bayesian multiple hypothesis testing for the shape parameter of generalized exponential distribution

“…This fact does not a surprising result. Berger and Mortera (1999) studied that the FBF gives the considerable bias in the direction of one of the hypotheses in non-symmetric situations. And so the FBF can not be used for the hypotheses testing in the case of non-symmetric circumstances.…”

Incidental Haptic Sensations May Not Influence Social Judgments: A Purely Confirmatory Replication Attempt of Study 1 by Ackerman, Nocera, & Bargh (2010)

“…Finally, the experiment would also stop on October 1st, 2013. Even though we preregistered a sampling plan, note that the sampling plan is irrelevant for Bayesian inference (e.g., Berger & Mortera, 1999;Berger & Wolpert, 1988;Rouder, 2014); as summarized by Edwards et al (1963, p. 193): "the rules governing when data collection stops are irrelevant to data interpretation. It is entirely appropriate to collect data until a point has been proven or disproven, or until the data collector runs out of time, money, or patience".…”

“…Specifically, we planned to use default Bayes factors for unpaired, one-sided t-tests as outlined in Rouder et al (2009) and Wetzels et al (2009), that is, a folded Cauchy distribution with a mode at 0 and a scale of 1 (see also Jeffreys, 1961, and. We would monitor the Bayes factor as the data come in, and report the results as a function of the number of participants, using sequential analysis plot (see Figure 2 in Wagenmakers et al, 2012; see also Berger & Mortera, 1999, Table 1). This study was planned in 2013, and since then Morey and Rouder (2015) have advocated a default Cauchy scale of r = 0.707.…”

“…The weight of the evidence provided by the data is given by a quantity known as the Bayes factor (BF; e.g., Berger & Mortera, 1999;Edwards, Lindman, & Savage, 1963;Rouder, Speckman, Sun, Morey, & Iverson, 2009;Wagenmakers, 2007;Wagenmakers et al, 2012;Wetzels, Raaijmakers, Jakab, & Wagenmakers, 2009 …”