Essays on the effects of communication on group decision making

Jingjing Zhang, Purdue University

Abstract

This dissertation consists of three independent essays that examine the effects of pre-play within-group text-form communication on group decision making in a variety of contexts, ranging from strategic interactions to intellective tasks. The first essay examines the effects of communication on asymmetric sized group contests when there are strategic interactions both between and within groups. In the first experiment, we construct a laboratory contest based on the participation game by Palfrey and Rosenthal (1983) in which group members make a binary contribution decision on whether or not to contribute and the group with more contributors wins the contest. When communication is not allowed the small group has more than a 10% chance of winning the game, in spite of its size disadvantage. Once group members can communicate, groups coordinate better and free-riding incentives are eliminated. The small group rarely contributes and the large group is usually able to win. Content analysis of group communication demonstrates that the most effective communication strategy for the large group is to explicitly designate specific non-contributors following a rotation scheme. In a second experiment, we allow group members to make a continuous contribution decision on how much to contribute and the all-pay auction contest success function is used to determine the winning group. Besides documenting similar effects of communication on group decision making, we provide insights for developing a new theory involving correlated equilibrium to explain the experimental data. The second essay investigates whether small groups are better than individuals in handling the winner’s curse. In an experimental setting, we find that groups of three members, who reach a unanimous bid via communication, place better bids than individuals. In the group treatment there is significant learning in several measures of performance and the group treatment outperforms the individual treatment. Neither risk attitude nor “truth wins” norm can explain the superiority of groups. When there is disagreement in a group, what prevails is usually not the best opinion but in 75% of the cases is the proposal of the median member. Learning form observing views of other people does significantly increase the number of optimal bids and significantly reduces the number of winner’s curse bids placed by an individual. The third essay explores whether and how communication influences the behavior of decision makers who are left out of communication. To study this problem we introduce a novel three-person trust game where the first player can send any part of his endowment to the second player. The amount sent gets tripled. The second player then decides how much to send to the third player. The amount is again tripled, and the third player then finally decides the allocation among the three players. Our baseline treatment with no communication shows that the first and second players send significant amounts and the third player reciprocates. When we add communication between the second and third players, the amount sent and returned between these two increases. Content analysis reveals that the only messages that significantly increase cooperation between the second and third players are ones that demonstrate the third player’s trustworthiness. The interesting finding is that there are external effects of communication: the first player who is outside communication receives higher amounts and thus sends 60% more than the no communication treatment. As a result, efficiency increases from 48% to 73%.

Degree

Ph.D.

Advisors

Cason, Purdue University.

Subject Area

Economics|Communication

COinS