It is self-evident that in society, many if not the most important decisions are made by groups. Group involvement is endemic, whether it be voting among large populations or deliberation by small committees. A standard view of group decision-making is that it facilitates exchanges of ideas and information, which necessarily lead to better outcomes. This view stands in contrast to a striking empirical regularity, dubbed group polarization, whereby members of a deliberating group predictably move toward a more extreme point in the direction indicated by the members' predeliberation tendencies. The aim of this project is to bridge the gap between the standard naive view of group decision-making and the empirical literature.
The project focuses on two distinct processes of group polarization. The first process operates in group decisions over risky prospects, where group members exhibit choice-shifts: They tend to take on more extreme levels of risk (either riskier or more cautious), relative to the levels of risk they would accept as individuals outside the group. The researchers propose an explanation for choice shifts, which is based on a model of individual decision-making. Hence, the framework they offer applies not only to choice shifts observed in groups but also to commonly observed biases in individual decision making. The second process analyzed in this proposal operates in groups that face a threat of disagreement. This threat plays an important role in explaining the fact that preference intensities matter in group interaction, and do so in non-additive ways. The researchers propose a dynamic model of group deliberation where each member wishes to implement a distinct policy, and the favorite policy of each member imposes externalities on the other members. The researchers discuss how this model can accommodate notions of `extremism' and `tolerance', and offer a methodology to study the formal relation between these two concepts. The proposed framework has the potential of allowing the researchers to rank different supermajority rules by the level of `extremism' that they impose in equilibrium.
