Cheating, incentives, and money manipulation
通过掷骰子任务研究不同激励方案下的诚实报告行为，发现无金钱激励时无显著作弊，有金钱损益时出现作弊，但金钱操纵（如自行取款）能减少作弊，尤其在损失情境下。
Abstract We use different incentive schemes to study truth-telling in a die-roll task when people are asked to reveal the number rolled privately. We find no significant evidence of cheating when there are no financial incentives associated with the reports, but do find evidence of such when the reports determine financial gains or losses (in different treatments). We find no evidence of loss aversion in the standard case in which subjects receive their earnings in a sealed envelope at the end of the session. When subjects manipulate the possible earnings, we find evidence of less cheating, particularly in the loss setting; in fact, there is no significant difference in behavior between the non-incentivized case and the loss setting with money manipulation. We interpret our findings in terms of the moral cost of cheating and differences in the perceived trust and beliefs in the gain and the loss frames.