Abstract: Compensation and revenge are the most basic behavioral reaction people employed in the context of interpersonal transgression. In most cases, offenders compensate the victim out of guilt and restore their relationship; the victim, motivated by either the intrinsic of justice/equity or emotions such as anger, will also retaliate the offender to make up their lost. Previous studies in the two areas focus on compensation and revenge respectively, without addressing them in the same context. When the offender has an amount to compensate, the victim also has one to revenge (retaliate). The comparison of the two amounts may start a new perspective to analysis the interpersonal transgression, increasing our prediction about the following interactions. By setting the offender and victim in the same scenario or real unintentional transgression, the present study discovered an over-compensation effect that offenders usually offer more compensation than the amount victims want to revenge (retaliate). Experiment 1 was based on a scenario story where one person’s unintentional mistake cost his (no gender implication in Chinese) a partner a great lost in reward of money. At the end of the story, there was an distribution task by which the story character can compensate or retaliated their partners (the more one distribute to himself, the less to his partner, so does the opposite). Participants were randomly divided into 3 groups (Offender vs. Victim vs. Control). The offender group was instructed to think and behave on the behalf of the offender character, while victim group on the behalf of the victim. The control group read a story of no transgression. The score offender group left for the victim was recorded as the amount of compensation while the score victim group left for the victim character was the amount of revenge. Experiment 2 was also a scenario study similar to Experiment 1, except the story was replaced to an unintentional body harm and participants compensate/revenge with voucher. Experiment 3 applied the paradigm of dot estimation task developed by Nelissen and Zeelenberg (2009) to creat an unintentional transgression. To set offender and victim together, Experiment 2 replaced the fictitious victims with real ones. Participants were randomly divided into 3 groups (Offender vs. Victim vs. Control). Victims suffered a money lost due to offenders’ poor performance (previous manipulated). At the end of the experiment, participants were informed to distribute 20 points in total (concerned their final reward) without being known. The more they left for themselves, the less their partner. In this way, they could compensate or revenge on their partners. The points offender left to the victim was recorded as compensation, while the point victim left for themselves was recorded as revenge. Experiment 4 manipulated a transgression of unintentional body harm. Participants randomly assigned as the offender was require to answer 10 questions with 4 absurd options (5 false, 5 correct; previous manipulated) in 15sec. Victim participants were hit in palm for each wrong answer. In the control group, hitting palm was replace with origami. In the end, all the participants were asked to distribute 15 ￥. The money offender left to the victim was recorded as compensation, while money victim left for themselves was revenge. In Experiments 1-4, the points/vouchers/money of compensation and revenge were both higher than the control condition (control group distribute equally), indicating that offenders compensated the victims and victims revenge/retaliated the offenders. Moreover, the compensation is slightly but statistically significantly higher than the revenge. In Experiment 1, the points offender distributed to the victim (M=14.32, SD=2.00) was significantly higher than the amount victim left to themselves (M=13.14, SD=1.70), t(72)=2.75, p<0.05. In Experiment 2, the voucher offender compensated (M=12.77, SD=2.23) was significantly more than the amount victim retaliate (M=10.14, SD=1.36), t(65)=5.72, p<0.01. In Experiment 3, the points offender left to the victim (M=12.55, SD=1.76) exceed the amount victim wanted (M=11.82, SD=1.40), t(146)=2.79, p<0.01. In Experiment 4, the money offenders compensated the victim (M=10.92, SD=2.25) is significant higher than the amount victims left to themselves (M=9.03, SD=1.25), t(120)=5.71, p<0.01. Given Experiment 1-2 are based on participants’ judgments on scenario story, their result only proves that the over-compensation effect SHOULD exist, while Experiment 3-4 further answered that over-compensation TRULY exist. The gender difference observed in Experiment 4 was also discussed. Over-compensation effect offered a new insight about the extent difference between compensation and revenge. Although the result of Experiment 1-4 gave a robust phenomenal support for this effect in the context of unintentional transgression, further researches about its inner mechanism and influence on the relationship are still in need. The closeness of relationship and relational utility also have a great impact on social interaction. Factor related should be concerned.

Key words: compensation, revenge/retaliation, unintentional transgression, equity, over-compensation effect

摘要： 个体的补偿和报复行为是人际损害领域的研究重点。本研究通过将损害者和受害者置于同一情境，揭示了非故意损害中的过度补偿效应。实验包括虚拟和真实两种非故意损害情境，损害方式包括经济损失和肉体伤害。组间变量是损害者、受害者和中立者三种角色，因变量是被试在分配任务中留给搭档或自己的分数或金钱。结果表明，在阅读或经历了相同的非故意损害事件后，损害者补偿给受害者的分数和金钱会显著高于受害者留给自己的数量，即过度补偿效应。

关键词: 补偿, 报复, 非故意侵害, 公平, 过度补偿效应