Individual Decisions in Group Settings: Experiments in the Laboratory and Field

This dissertation features laboratory and �eld experiments that examine individual decision-making when the individual is a member of a group. The �rst two chapters study a donation setting where the decisions of other group members directly a�ect the payo�s of the individual making a decision. The third chapter studies a sequential recommendation setting where the decisions of other group members may indirectly a�ect the individual by providing payo�-relevant information. The �rst two chapters address the research questions: �In a donor's decision to give to a charitable organization, are beliefs about the behavior of peer donors important? If so, how do those beliefs impact the decision to give?� The �rst chapter presents the results of a �eld experiment whose treatments vary the extent to which an individual's donation impacts the amount of donation matching money given to the charity. Results suggest that beliefs about peer donors' likelihood of contributing are an important factor in the decision to give charitably, and that fundraisers may generate better donation outcomes with more innovative structuring of matching money. To provide an additional and more controlled test of these results, the second chapter presents the �ndings of a laboratory experiment that replicates the donation decision of the �eld experiment while also eliciting donor beliefs. The laboratory results provide further support for the �ndings of the �eld experiment: beliefs about peers appear to matter, and donation outcomes vary substantially across treatments. The third chapter addresses the research question: �In sequential recommendation settings, does the institution used to gather recommendations a�ect information suppression by individual agents?� In this chapter's laboratory experiment, subjects receive private information of heterogeneous quality, and treatments vary the institution used to gather recommendations from subjects. Results suggest that information iii suppression is indeed a�ected by the recommendation-gathering institution. An institution that arranges subjects in increasing order of information quality results in less information suppression than institutions that arrange subjects either in decreasing order of information quality, or in an order determined endogenously by subjects. This �nding yields a policy-relevant conclusion that recommendation-gathering institutions should be designed carefully in order to minimize information suppression.

Note: Abstract extracted from PDF text