Description:
Academic honesty is crucial for scientific advancement, yet replication crises and misconduct scandals are omnipresent. We provide evidence on scientists' truth-telling from two incentivized coin-tossing experiments with more than 1,300 scientists. Experiment I, with predominantly European and North-American scientists, shows that fewer scientists over-report winning tosses when their professional identity is salient. The global Experiment II yields heterogeneous effects. We replicate Experiment I's effect for North-American scientists, but find the opposite for Southern European and East-Asian scientists. Over-reporting correlates with publication metrics and country-level measures of academic and field-experimental dishonesty, suggesting that country-level honesty norms also guide truth-telling by scientists.