Using Distraction to Reveal Hidden Ethnic Bias in Economics Games

• In Event: Innovations in Experimental Design

Abstract

Behavioral economics games such as the Dictator Game (DG) have been widely used to quantify ethnic or other between-group biases. However, in numerous contexts where ethnicity is thought to be extremely salient, the DG has failed to detect any evidence of ethnic bias. One compelling hypothesis is that economics games like the DG are explicit measures that are susceptible to self-monitoring and social desirability concerns—specifically, participants may alter behavior to fit social norms that discourage ethnic bias or discrimination, even in cases where implicit ethnic bias is present. Drawing on dual process models from psychology, we propose an experimental design to test this hypothesis. We randomly induce cognitive load through a concurrently played distraction task in a within-subjects experimental design. If self-monitoring explains the apparent lack of ethnic bias under DG, then evidence of bias should emerge in the DG played concurrently with the distraction task. We deploy this design among 450 Kikuyu and Luo participants in Nairobi, Kenya during summer 2022; previous DG studies among these groups failed to produce evidence of ethnic bias, even where other evidence suggests such bias exists. We first confirm that the DG in the control condition fails to produce evidence of ethnic bias. We then analyze whether the concurrent distraction task generates evidence of ethnic bias within the DG. This study promotes better understanding of commonly used measurement tools.

Author

• Daniel N. Posner, University of California, Los Angeles