Measuring preferences for competition

Abstract

Recent research has found that competitive behavior measured in experiments strongly predicts individual differences in educational and labor market outcomes. However, there is no consensus on the underlying factors behind competitive behavior in these experiments. Are participants who compete more capable, more confident, and more tolerant of risk, or are they competing because they enjoy competition per se? In this study, we present an experiment designed to measure individuals’ preferences for competition. Compared to previous work, our experiment rules out risk preferences by design, measures beliefs more precisely, and allows us to measure the magnitude of preferences for competition. In addition, we collect multiple decisions per participant, which lets us evaluate the impact of noisy decision-making. We find strong evidence that many individuals possess preferences for competition. Most participants are either reliably competition-seeking or competition-averse, and their choices are highly consistent with expected utility maximization. We also find that preferences for competition depend on the number of competitors but not on the participants’ gender.