A basic chapter introducing psychologists to the world of modern Bayesian statistics. We cut out a lot of the dogma and go into sales pitch mode on the benefits of going Bayesian. If we pique your interest in learning more about what Bayesian tools can offer, then we consider the chapter a success.

Feinberg, F. & Gonzalez, R. (2012). Bayesian modeling for psychologists: An applied approach. Handbook of Research Methods: Vol 2. Research Designs, H. Cooper, Editor-in-Chief, New York: American Psychological Association, 445-464.

doi:10.1037/13620-024 PDF

Abstract

Bayesian methods offer new insight into standard statistical models and provide novel solutions to problems common in psychological research, such as missing data. Appeals for Bayesian methods are often made from a dogmatic, theory-based standpoint concerning the philosophical underpinnings of statistical inference, the role of prior beliefs, claims about how one should update belief given new information, and foundational issues, such as the admissibility of a statistical decision. Although such a rhetorical approach is academically rigorous, it usually is not the kind of argument a practicing researcher wants to read about. Researchers care about analyzing their data in a rigorous manner that leads to clear, defensible conclusions. In this chapter, we address the reader who wants to learn something about what all the Bayesian fuss is about and whether the Bayesian approach offers useful tools to incorporate into one’s data analytic toolbox. We hope this chapter prompts readers to learn more about what Bayesian statistical ideas have to offer in standard data analytic situations. Throughout the chapter, we highlight important details of the Bayesian approach; how it differs from the frequentist approach typically used in psychological research; and most important, where it offers advantages over the methods most commonly used by academic researchers in psychology and cognate disciplines.