Abstract: There has been much research concerning how to promote cooperation and solve the free-rider problem. A common solution to the free-rider problem is to install a leader. But there is a widespread evidence denying the effectiveness of leadership that power can lead people to place greater importance on their own interests and subordinate those of other, and the leader in positions of power can have a substantial negative impact on the common good by acting solely in their own self-interest. However, there is no unanimous conclusion towards the issue whether the effect of power on teenagers’ cooperation is same as adult. Moreover, one’s own and another’s power are ubiquitous in the real world, it is important to explore how one’s own power interacts with another’s power to influence cooperation. In the current study, we mainly examined the effect of self-power and other- power for junior school students’ cooperation. Last but not least, power is a relative construct, it is extremely common for one’s power to change from one situation to the next. Therefore, we further explored stability of power between power and cooperation. Two experiments were conducted in the laboratory with E-Prime 2. Experiment 1 was organized into a 3×3 mixed design, with the first factor referring to the one’s own power (self-power: high vs. low vs. control) as a between-participant factor, the second factor referring to the partner’s power (other-power: high vs. low vs. unknown) as a within-participant factor. Participants consisted of 79 junior school students, and power was manipulated by role playing. The result showed that high power participants were more likely to offer less resource than low power participants. In order to make sure whether the effects from Experiment 1 would remain across different contexts (i.e. when power changes), Experiment 2 was organized into a 2×2 within-participant design, with the first factor referring to the one’s own power (self-power: high vs. low), the second factor referring to the partner’s power (other-power: high vs. low). Participants consisted of 52 junior school students, and power was manipulated within-subjects using the same task, with rounds exercise assignments of interleaved between formal experimental sessions. The result showed that high power participants were more likely to offer more resource than low power participants. Moreover, participants were more likely to offer more resource to low power partner than high power partner, the other-power effect was particularly robust for high power participants. The results are as follows: (1) High power junior school students showed less cooperative behavior than low power students when power is stable, and high power junior school students showed more cooperative behavior than low power students when power is unstable. (2) Participants were more likely to cooperate with low power partner than high power partner. This effect was particularly robust for low power participants. Findings from the two experiments supported the effect of power on cooperation in junior school students and suggested implication for team management. Future research on power using the social dilemmas could explore the effect of social distance, take partner dyad as the unit of analysis which is based on the actor-partner interdependence model (APIM), and test the effect in other social dilemmas.

Key words: power, public goods dilemma, cooperation, junior school students

摘要： 本研究分别操纵被试稳定的权力（实验1）和不稳定的权力（实验2），并与不同权力个体进行公共物品困境任务，以考察了自我权力和他人权力对初中生合作行为的交互作用及其权力稳定性在其中的作用。结果发现，在权力稳定情境下，高权力的初中生合作水平低于低权力组和控制组，而在权力不稳定的情境下，高权力的初中生合作水平高于低权力组。无论权力稳定与否，与低权力同伴合作时，低权力者的合作水平均显著高于与高权力同伴合作时。

关键词: 权力, 公共物品困境, 合作, 初中生