Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We study an observed cheating game in which senders can earn money by lying to the experimenter, knowing the experimenter will later be able to observe whether they told the truth or lied. We modify the game by matching each sender with a receiver, whose earnings negatively correlate with the earnings of the sender. Our results show that senders lie less when matched with a receiver who loses money if they lie. However, once we increase the stakes by a factor of five, participants lie as much in the two-player game as in the one-player game. That is, an externality reduces lying but only as long as the stakes are low.