Updating Consistency in Bayes Factors

Schwaferts, Patrick (10. Februar 2021): Updating Consistency in Bayes Factors. Department of Statistics: Technical Reports, Nr. 236

Abstract

When it comes to extracting information from data by means of Bayes rule, it should not matter if all available data are considered at once or if Bayesian updating is performed subsequently with partitions of the data. This property is called updating consistency. However, in the context of Bayes factors, a prominent Bayesian tool that is used for comparing hypotheses, some researchers illustrated that updating consistency might not be given. Therefore, this technical report addresses the updating consistency of Bayes factors and shows its existence. In that, it serves as mathematical basis for the evaluation of the origin of putative updating inconsistencies. In addition, results about updating mixture priors are brought into the terminology commonly employed in the context of Bayes factors, as these were used in the elaboration about updating consistency. The depicted results imply that a necessary condition for updating consistency is to consider and report not only the Bayes factor value alone but also the posterior distributions as outcome of the analysis.

• Dokumententyp: Paper
• Keywords: Bayesian Statistics; Bayes Factor; Sequential Updating; Updating Consistency; Mixture Prior; Spike-and-Slab Prior
• Fakultät: Mathematik, Informatik und Statistik > Statistik > Technische Reports
• Themengebiete: 300 Sozialwissenschaften > 310 Statistiken
• Sprache: Englisch

Literaturliste

M. Gönen, W. O. Johnson, Y. Lu, and P. H. Westfall. The Bayesian two-sample t test. The American Statistician, 59(3):252–257, 2005. R. E. Kass and A. E. Raftery. Bayes factors. Journal of the American Statistical Association, 90(430):773–795, 1995. T. J. Mitchell and J. J. Beauchamp. Bayesian variable selection in linear regression. Jounalof the American Statistical Association, 83(404):1023–1032, 1988. R. D. Morey, J.-W. Romeijn, and J. N. Rouder. The philosophy of Bayes factors and the quantification of statistical evidence. Journal of Mathematical Psychology, 72:6–18, 2016. J. N. Rouder and R. D. Morey. A Bayes factor meta-analysis of Bem’s ESP claim. Psychonomic Bulletin & Review, 18(4):682–689, 2011. J. N. Rouder, P. L. Speckman, D. Sun, R. D. Morey, and G. Iverson. Bayesian t tests for accepting and rejecting the null hypothesis. Psychonomic Bulletin & Review, 16(2): 225–237, 2009. J. N. Rouder, J. M. Haaf, and F. Aust. From theories to models to predictions: A Bayesian model comparison approach. Communication Monographs, 85(1):41–56, 2018a. J. N. Rouder, J. M. Haaf, and J. Vandekerckhove. Bayesian inference for psychology, part iv: Parameter estimation and Bayes factors. Psychonomic Bulletin & Review, 25(1): 102–113, 2018b. B. RÜger. Test-und Sch¨atztheorie: Band I: Grundlagen. De Gruyter Oldenbourg, 1998. W. Vanpaemel. Prior sensitivity in theory testing: An apologia for the Bayes factor. Journal of Mathematical Psychology, 54(6):491–498, 2010. W. Vanpaemel and M. D. Lee. Using priors to formalize theory: Optimal attention and the generalized context model. Psychonomic Bulletin & Review, 19(6):1047–1056, 2012.