Accession Number : ADP005296

Title :   Group Decision Analysis and its Application to Combining Opinions,

Corporate Author : BRITISH COLUMBIA UNIV VANCOUVER

Personal Author(s) : Zidek,James V.

Report Date : 16 MAR 1987

Pagination or Media Count : 31

Abstract : This paper is primarily about multi-Bayesian statistical decision theory. Thus the author is deliberately expanding this discussion beyond the realm of merely aggregating opinions. There are a number of reasons for doing so. There may well be situations involving groups of individuals wherein the real goal is merely that of combining degrees of belief concerning statements (hypotheses) about the world. However, opinions in the usual sense of that word would generally incorporate preferences as well. There is even some reason to doubt that degrees of beliefs and preferences (utilizes) can be separately elicited even in a normative setting (Dr. H. Rubin, personal communication). In short, there is some doubt as to whether the aggregation of degrees of belief is a meaningful undertaking. In any case, it would seem that in a significant number if not the majority of situations involving groups of individuals, a joint decision is called for. This joint decision could itself be an aggregate of the individual expressions of degrees of belief. But this recasts the problem into the context of group decision making and redirects our interest towards this latter area.

Descriptors :   *STATISTICAL DECISION THEORY, BAYES THEOREM, DECISION MAKING, STATISTICAL ANALYSIS, ESTIMATES, WORKSHOPS

Subject Categories : Statistics and Probability

Distribution Statement : APPROVED FOR PUBLIC RELEASE