How Groups Differ from Individuals in Learning from Experience: Evidence from a Contest Platform

• 成果类型: Article
• 署名作者: He, Tianyu; Minervini, Marco S.; Puranam, Phanish
• 署名单位: National University of Singapore; IE University; INSEAD Business School
• 刊物名称: ORGANIZATION SCIENCE
• ISSN/ISSBN: 1047-7039
• DOI: 10.1287/orsc.2021.15239
• 发表日期: 2024
• 页码: 1512-1534
• 关键词: aggregation; Organizational learning; Groups versus individuals; group learning; teams

摘要

We examine how groups differ from individuals in how they tackle two fundamental trade-offs in learning from experience-namely, between exploration and exploitation and between over- and undergeneralization from noisy data (which is also known as the bias-variance trade-off in the machine learning literature). Using data from an online contest platform (Kaggle) featuring groups and individuals competing on the same learning task, we found that groups, as expected, not only generate a larger aggregate of alternatives but also explore a more diverse range of these alternatives compared with individuals, even when accounting for the greater number of alternatives. However, we also discovered that this abundance of alternatives may make groups struggle more than individuals at generalizing the feedback they receive into a valid understanding of their task environment. Building on these findings, we theorize about the conditions under which groups may achieve better learning outcomes than individuals. Specifically, we propose a self-limiting nature to the group advantage in learning from experience; the group advantage in generating alternatives may result in potential disadvantages in the evaluation and selection of these alternatives.