Group Polarization in a Model of Information Aggregation

Abstract

Experiments identify the empirical regularity that groups tend to make decisions that are more extreme, but in the same direction as the tendency of individual members of the group. We present a model of information aggregation consistent with these findings. We assume individuals and groups are rational decision makers facing monotone statistical decision problems where groups and individuals have common preferences, but groups have superior information. We provide conditions under which the distribution of the optimal actions of the group is more variable than the distribution of actions taken by individuals. (JEL D71, D83)