Accuracy, confidence and consensus in bayesian hypothesis inference
Roger J. Owenα · Communication in Statistics- Theory and Methods · 1985
Interest centres on a group of statisticians , each supplied with the same n sample datapoint sandmaking formal Bayesian inference with a common likelihood function but differing prior knowledge and utility functions. Definitions are proposed which quantify, in a commensurable way, the inference processes of “accuracy”, “confidence” and “consensus” for the case of hypothesis inference with a fixed sample size n. The general significance of comparing the three quantifiers is considered. As n increases the asymptotic behaviour of the quantifiers is evaluated and it is found that the three rates of convergence are of the same order as a function of n. The results are interpreted and some of their implications are discussed.