Search
Program Calendar
Browse By Day
Browse By Time
Browse By Person
Browse By Session Type
Browse By Topic
Personal Schedule
Sign In
Deadlines
Policies
Updating Your Submission
Requesting AV
Presentation Tips
Request a Visa Letter
FAQs
X (Twitter)
Search Tips
Annual Meeting App (Available July)
About Annual Meeting
Large groups of individuals can generate highly accurate aggregate beliefs, a phenomenon known as the wisdom of crowds. At the same time, social influence can lead to herding effects that produce fads, bubbles, groupthink, and information cascades. Under what conditions will social influence preserve the wisdom of crowds? Using a controlled experiment, we show that the structure of communication networks plays a major role in determining whether information exchange leads to social herding or improved collective beliefs. In highly centralized networks, central individuals have a disproportionate influence on the group. These opinion leaders can undermine the wisdom of crowds, leading to less collective accuracy. In contrast, in decentralized networks, where individuals are more equally connected, individuals are drawn toward the aggregate beliefs of the group, improving the wisdom of crowds. Our results provide novel empirical support for theories describing the conditions under which social influence leads to herding and the conditions under which large networks can harness the wisdom of crowds.