Using WinBUGS to implement Bayesian inferences of estimation and testing hypotheses, Bayesian Methods for Measures of Agreement presents useful methods for the design and analysis of agreement studies. It focuses on agreement among the various players in the diagnostic process.
The author employs a Bayesian approach to provide statistical inferences based on various models of intra- and interrater agreement. He presents many examples that illustrate the Bayesian mode of reasoning and explains elements of a Bayesian application, including prior information, experimental information, the likelihood function, posterior distribution, and predictive distribution. The appendices provide the necessary theoretical foundation to understand Bayesian methods as well as introduce the fundamentals of programming and executing the WinBUGS software.
Taking a Bayesian approach to inference, this hands-on book explores numerous measures of agreement, including the Kappa coefficient, the G coefficient, and intraclass correlation. With examples throughout and end-of-chapter exercises, it discusses how to successfully design and analyze an agreement study.
Table of Contents
Introduction to Agreement. Bayesian Methods of Agreement for Two Raters. More Than Two Raters. Agreement and Correlated Observations. Modeling Patterns of Agreement. Agreement with Quantitative Scores. Sample Sizes for Agreement Studies. Bayesian Statistics. Appendices.
Lyle D. Broemeling
"This book is a welcome addition to the literature on Bayesian inference as it presents methods for the design and analysis of agreement studies. … The approach presented by the author is novel and the novice will find a helpful introduction to Bayesian inference in an appendix. … The text is readable and will form a valuable reference source. For those unfamiliar with WinBUGS, the author introduces the fundamentals of programming and executing BUGS."
—International Statistical Review, 2010
"This book deals with measures of agreement from a Bayesian perspective, focusing mainly on variants of Cohen’s ¿, but also other measures included in Shoukri (2003) and von Eye and Mun (2005), frequentist texts for which this book is intended to be a Bayesian companion. Dr. Broemeling uses examples throughout the book to illustrate concepts rather than resorting to jargon … This book would be valuable for those using the methods in Shoukri and von Eye and Mun. …"
—Journal of the Royal Statistical Society, Series A, Volume 173, Issue 1, January 2010