摘要

We study how cheating behavior is affected by incentives. After replicating the finding in the cheating game literature that lying does not increase with incentives, we show that this insensitivity is not a characteristic of the intrinsic lying cost, but rather a result of concern about being exposed as a liar. In a modified “mind” game in which this concern is eliminated, we find that people lie more, and in particular lie more when the incentives to do so increase. Thus, our results show that for many participants, the decision to lie follows a simple cost–benefit analysis: they compare the intrinsic cost of lying with the incentives to lie; once the incentives are higher than the cost, they switch from telling the truth to lying

变量是incentives

复现文献已有的Cheating game结果后，发现说谎没有随着incentives增加，我们发现这不是谎言内在成本（intrinsic lying cost）特征，而是被试担心被暴露成为说谎者的结果，当采用改进过的Mind Game后，这种担忧被消除了。结果发现说谎增加了，尤其是当这样做的动机增加时。因此结果表明，对于许多参与者而言，说谎的决定遵循简单的成本效益分析：他们将说谎的内在成本与说谎的动机进行了比较；一旦激励措施超过成本，它们就会从说实话转变为说谎。

实验综述

原则上可以根据内在撒谎成本（intrinsic lying cost）来区分三种类型的人：

• “ethical type”：intrinsic lying cost → $\infty$
• “finite positive cost type”：intrinsic lying cost → C>0
• “economic type”：intrinsic lying cost → 0

目前研究欺骗的两种实验设计

• Sender&Receiver(Deception Game)
  • strategic
  • may affect the receiver’s beliefs
  • includes social preferences
• non-strategic（cheating game）
  • participant receives private information about the state of the world and reports it to the experimenter
  • no social preferences
  • but only have the aggregate distribution

incentives in deception games（显著） but not in cheating games（不显著）？

一种可能性是，Deception Game中，说谎是游戏规则显示的一种选择，在该规则中，要求发件人决定发送真还是假消息。根据游戏的战略结果，可以很好地定义每个消息的结果。但是在cheating game中并非如此，该游戏要求参与者报告实际结果，而不会被告知说谎的选项。由于在作弊游戏的规则中没有明确提及说谎选项，因此参与者可能会担心说谎的可能后果。

此外，cheating game的参与者可能会质疑自己的掷骰子是否真的是私密的。当激励增加时，人们对说谎时暴露的可能后果的担忧可能会增加。

引入Mind Game验证担心暴露的后果

正如预测那样，Mind Game的结果显示出参与者在不同收入每一个级别上都增加了说谎。

这个结果表明，cheating games中的说谎没有随着incentives显著是因为担心暴露。