Conclusion
This paper analyzes a rich database of 4,279 Jeopardy! episodes with 8,169 contestants to study whether the gender of one’s opponents affects behavior in a highly competitive situation with large stakes. As contestants are unable to choose the gender of their opponents, Jeopardy! provides an attractive field setting to explore such dynamics, with individual outcomes being readily available for the researcher. We study four distinct scenarios: The likelihood (i) to win an episode, (ii) to respond to a clue, (iii) to respond correctly to a clue, and (iv) the wagering decisions in Daily Double clues. Contrary to existing studies, we find that a woman is more likely to win and competes more aggressively when paired against males. Further, the otherwise robust gender gap in risktaking disappears once a woman competes in an all-male field of competitors. These results are robust to the inclusion of a rich list of potentially confounding variables and player-fixed effects, which allow us to control for any unobservable differences on the individual level. Further, these results are unlikely to be driven by a strategic consideration of women performing more aggressively because of a potential under-performance in particularly high-stakes clues. From the male perspective, we find performance indicators to be less responsive to their opponents’ gender, but a notable heterogeneity emerges for wagering decisions. In particular, a man wagers significantly less when competing against women. We discuss potential policy consequences in Section 6, but advise caution in the interpretation and generalization of our results. At the least, we hope that this study stimulates further research on the role of the opponents’ gender in influencing performance and risk-taking in high-pressure situations.
