Kadane J.B. Robustness of Bayesian Analyses

Elsevier, 1984. — 326 p.Robustness is a fundamental issue for all statistical analyses; in fact it might be argued that robustness is the subject of statistics. In Bayesian statistics, the prior distribution can be seen as weighting the possible values of the parameter by their probability. The studies reported in this volume concern the sensitivity of Bayesian analyses to the inputs: the prior, the likelihood, and, where applicable, the utility function. There are many approaches to Bayesian robustness, some of which are represented, and many of which are mentioned, in this volume. The first paper, by Edwards, Lindman and Savage, is a classic in this area. The paper of Berger, with comments by Brown, Hill, Kadane and Lindley, represents a Bayesian view that draws on some non-Bayesian elements. The brief paper of Kadane and Chuang introduces the Chuang dissertation, which treats robustness as a kind of limiting sensitivity analysis. Finally, the paper of Polasek shows an econometrically-oriented view of Bayesian robustness, with applications.