Improving Group Decision-Making for Contentious Topics
Abstract
Group decision-making is a central feature of civilian and military organizations as well as democratic societies. Groups offer thepotential for substantial gains in decision-making quality relative to individual decision-makers: The ability of collective intelligence and the wisdom of crowds to produce accurate judgments by integrating private information has been widely documented. But groups are not always more accurate than individuals, and a large literature has emerged exploring the conditions under which groups perform better or worse than individuals. While this body of work has produced a number of important insights, it has almost entirely overlooked a central element often present in real-world group decision-making: pressure to abide by social norms. Group dynamics are governed by social norms, informal rules regarding acceptable versus unacceptable behavior that are enforced by social sanctions. In the context of information aggregation and group decision-making, social norms can have deleterious consequences. When trying to make accurate assessments, it is essential for the group to integrate the private information held by each group member. But, in thecontext of contentious (e.g. ideologically-relevant) issues, individuals may withhold relevant private information (e.g. assessments) that questions the prevailing wisdom out of fear of being sanctioned. Thus, dissenting opinions are often silenced (e.g. by self-censorship) due to fear of sanctions. This may stifle discussion and undermine group performance, potentially even leading groups to perform more poorly than individuals. Here we will use online experiments in which Americans from around the country interact anonymously via the internet to investigate this issue. First, we propose a systematic investigation of the impact of contentiousness on group decision-making (Task 1). To do so, we will develop two experimental paradigms, using fundamentally different types of collective intelligence tasks, in which contentiousness (as well as task complexity) can be systematically and independently varied. Oncewe have identified regions in the [contentiousness, complexity] space in which social norms impair group decision-making, we will test a variety of interventions aimed at counteracting that impairment (Task 2). In addition to improving communication in the group decision-making stage, our interventions also aim to improve baseline performance in the individual decision-making stage - thereby increasing the proportion of participants who enter the group interaction with the correct answer and thus reducing the social pressure to report the incorrect answer.The research we have proposed will have important scientific and practical benefits. From a scientific perspective, we will substantially advance research on collective intelligence by integrating social norms into that literature. We believe that in addition to generating new insights itself, the work we propose here - andthe paradigms we will develop - willstimulate a great deal of subsequent research by other groups interested in the wisdom of crowds, politically motivated reasoning, and social norms. From a practical perspective, interventions we identify to combat social pressure can be used to improve group decision-making in a way variety of contentious settings. Approved for Public Release
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jun 09, 2021
- Source ID
- N000142112573
Entities
People
- David G. Rand
Organizations
- Massachusetts Institute of Technology
- Office of Naval Research
- United States Navy