Dishonest behavior is influenced by situational factors, such as the reward size and externalities, and personal factors, including the participant’s gender and age. Field experiments exploring dishonesty are relatively scarce compared to the experimental lab literature. To address this gap, we conducted a new two-phase experiment using a sample of subjects from a previous study, where their reasoning ability, altruism, and trustworthiness were assessed. This new study draws upon data from 180 participants involved in a two-phase experiment: the laboratory phase and the field phase. In the first phase, participants were invited to our laboratory to complete four personality questionnaires (the HEXACO-100 and the Dark Triad including the Machiavelism Personality Scale, the Narcissistic Personality Inventory and the Levenson Self-Report Psychopathy Scale) and a sociodemographic questionnaire. Subjects were informed that they would receive €10 for their participation, paid via bank transfer. In the second phase, subjects were randomly assigned to one of three groups: a control group, where participants were deliberately underpaid by €5, and two treatment groups, where participants were overpaid by either €5 or €15. To ensure that all participants were aware of the erroneous payment, we asked them to inform us via email about the date they received the money.

Once participants notified us of the date they received the transfer, confirming their awareness of the payment error, we categorized them based on their response to this fact. Those who reported the payment error were classified as “claimers,” while those who did not were labeled as “non-claimers”. At this stage, economic dishonesty can be detected in individuals who were either low or highly overpaid and chose not to report the payment error. Our experimental evidence showed that while nearly all underpaid subjects claimed the error amount, almost three-quarters of the overpaid subjects did not report and return the extra money received. This dishonesty rate is the highest observed in any field study where well-aware subjects must decide whether to keep the extra money.

After controlling for potential covariates, including socio-economics demographics, self-reported personality traits, cognitive ability, and measures of altruism and trustworthiness revealed in experimental games, it was observed that highly overpaid subjects were nearly 15 percentage points more likely to act honestly than those slightly overpaid, highlighting the role of temptation in shaping ethical behavior. Our estimates also reveal that factors such as age, living standards, prior participation in experiments, and reasoning ability significantly influence the likelihood of behaving honestly. Specifically, older individuals, those with higher living standards, and those with higher reasoning abilities are more likely to act honestly, suggesting a positive relationship between these factors and ethical behavior. In contrast, the impact of prior participation in experiments has the opposite effect. Participants with more experience in experiments were less likely to behave honestly, indicating that familiarity with experimental settings might reduce the pressure to conform to ethical standards. These results highlight the complex interplay of personal characteristics in shaping behavior.