Competition is ubiquitous in markets and organizations. Prior research has hinted at the power of rivalry relationships, based in historical interactions rather than current levels of competition, to drive behavior and performance in competitive settings. However, empirical evidence comes almost exclusively from sports contexts, leaving the role of rivalry in knowledge work poorly understood. In this paper, we conceptualize competition and rivalry as peer effects, and investigate the causal effects of rivalry on performance of knowledge workers. We use data from computer programming contests involving over 10.6 million dyadic competitive encounters across 63,220 software developers (‘coders’). We find that coders who are randomly assigned to compete against others with whom they share a history consistent with rivalry exhibit higher performance, controlling for current levels of competition. This provides causal evidence for the importance of competitive histories, and suggests that rivalry applies to knowledge work as well as physical tasks. We also find evidence that rivalry especially benefits the performance of highly skilled individuals. This work contributes to research on rivalry and competition more generally, as well as to the literature on knowledge work and the drivers of performance in technical settings.
Chris is Associate Professor for Information Systems at the D’Amore McKim School of Business. He employs business analytics and data science to investigate research questions about group-decision making, network science, and social media, and develops novel computational approaches to study collective intelligence mechanisms.