Opinion Averaging Versus Argument Exchange

Ulrike Hahn, Leon Assaad, Jason W. Burton

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearch

32 Downloads (Pure)

Abstract

Opinion averaging is a common means of judgment aggregation that is employed in the service of crowd wisdom effects. In this paper, we use simulations with agent-based models to highlight contexts in which opinion averaging leads to poor outcomes. Specifically, we illustrate the conditions under which the optimal posterior prescribed by a normative model of Bayesian argument exchange diverges from the mean belief that would be arrived at via simple averaging. The theoretical and practical implications of this are discussed.
Original languageEnglish
Title of host publication46th Annual Meeting of the Cognitive Science Society (CogSci 2024)
Number of pages7
PublisherCognitive Science Society
Publication date2024
Pages4554-4560
Publication statusPublished - 2024
SeriesProceedings of the Annual Meeting of the Cognitive Science Society
Volume46
ISSN1069-7977

Keywords

  • Opinion dynamics
  • Averaging
  • Argument
  • Wisdom of the crowd
  • Agent-based simulation

Cite this