Ambiguity aversion, group size, and deliberation: Costly information and decision accuracy

Abstract

Earlier work has shown that in a “costly attention” setup – where group members incur costs to pay attention to informative signals - smaller group sizes generally result in more accurate decisions, contrary to the Condorcet jury theorem. This result obtains whether group members deliberate or whether they vote secretly without deliberations. Moreover, when group sizes are small, barring deliberations leads to more accurate decision-making, while allowing deliberations is optimal in larger groups. In the current paper, I analyze a costly attention setup where some group members are ambiguity averse (ambiguity exists about the other group members’ likelihood of paying attention) and find that groups larger than a threshold make the best decisions. Thus, this supports Condorcet's conclusion about the benefits of a larger group, but under very different conditions. Implications for whether to allow or bar deliberations in these large committees are more nuanced and depend on whether attentive members may be sure of making the correct judgment. The results yield policy implications regarding optimal committee size and on whether committees should allow or bar deliberations.