Aggregation of expert opinions

Abstract

Summarizes research which is relevant to the basic question of how people aggregate a variety of expert opinions to generate their own judgment and make decisions. An introductory section discusses terminology and some general points in this line of research. A series of general questions are used to frame the larger literature. The first question concerns who is doing the aggregation: a normative model, an individual decision-maker or a group. The second question concerns the form of the information and the response that the decision-maker generates: binary, probabilities or a quantity estimate. The third question concerns the nature of the events that are relevant to aggregation: is the underlying uncertainty aleatory or epistemic, and is the task intellective or judgmental? The fourth question looks at characteristics of the information pattern and information sources, and includes characteristics of the experts and issues of redundancy in information. The fifth and final question reviews the combination rule being utilized. In the final section, we summarize some preliminary results from a new study by Budescu and Rantilla (1998), focusing on the determinants of confidence in aggregated expert opinions. These results are used as an illustration of the usefulness of our proposed general framework.