Thesis defended by Gwen-Jiro Clochard on July, 5th 2022
Experimental economics offers the possibility to study the behavior of economic agents in ideal conditions, as it allows for the complete control of the environment of participants. This thesis presents results from several field experiments, as well as their contributions to the literature. This manuscript is composed of five chapters.
The first three chapters contribute to the literature on the contact theory, a methodology used to improve intergroup relations. Chapters 1 and 2 present results from field experiments, in France and in Senegal, while chapter 3 presents a meta-analysis of the experimental literature on the topic. The last two chapters present results of various experiments. Chapter 4 investigates the influence of providing rankings on performance, depending on the nature of the task to be performed. Chapter 5 talks about genetic adaptation to fishing risk in Senegal.
Chapter 1 discusses a field experiment investigating the effects of a brief and controlled discussion on trust between disadvantaged students in Paris’ suburbs, and police officers. The contact protocol is adapted from the social psychology literature and aims at quickly creating friendships. Results indicate that the contact intervention increases trust of students toward the specific police officers met, but it does not increase trust in the police in general.
Chapter 2 analyzes an experimental protocol closely related to that of chapter 1, but applied in the context of inter-ethnic relations in Senegal. The short contact is effective at increasing inter-ethnic trust with the specific individuals met, but does not increase altruism toward the outgroup in general. Moreover, the effect at the individual level evaporates within one month of the intervention. Machine learning techniques enable to discover that contact is particularly effective for older and less educated participants.
Chapter 3 proposes a meta-analysis of the rapidly growing experimental literature on the contact hypothesis. Based on 62 outcomes from 37 research papers, the analysis revealed three main results. First, there exists a large heterogeneity in the definition of a “contact”, making difficult the comparison of papers. Second, on average, contact interventions are effective, by significantly improving the perception of the outgroup. Third, machine learning algorithms enable the study of determinants of the most effective contact interventions.
Chapter 4 investigates the effects of providing rankings on performance, depending on the nature of the task. The analysis is performed by analyzing the existing empirical literature, proposing a novel theoretical framework and the empirical analysis of two experiments. Results indicate that the main moderating factor of the effects of relative performance feedback is the existence of room for technological improvement.
Chapter 5 investigates the interplay between the effects of the dopamine receptor gene D4 and background risk on risk-aversion for fishermen in Northern Senegal. The analysis consists in the genetic and behavioral comparison of two populations differing in their main professional occupation. Results indicate that the 7R allele is associated with an increased risk-tolerance, through an additive, rather than dominance, effect. Moreover, the effects of the 7R allele are not driven by the background risk, as we do not observe heterogeneous effects depending on the level of background risk.
