Abstract: The behavior decisions in social dilemmas are research highlights for the sociology, economics as well as social psychology. Cooperation index (CI) reflexes the benefit of cooperative behavior, and social distance refers to an abstract conception that describes the closeness of emotional relationship among individuals. In social dilemmas, the CI is of rational factor and the social distance belongs to emotional factor. Early theories view that behavior decisions are mostly determined by benefit and payment, which represents rational idea. On the contrary, irrational idea considers that the behavior decisions are not only affected by rational but emotional factors. However, even recent researches have proved that social distance or CI can motivate cooperative behaviors, the influencing mechanism of those factors is still unclear. To be specific, how do CI and social distance affect competitive and cooperative behaviors? Does an interaction exist between these two factors? Accordingly, we hypothesized that there was an interaction between social distance and CI. To address this, the Prisoner’s Dilemma paradigm was employed to discover the mechanism. A 2 (Social distance: close vs. strange) × 9 (CI: 0.1~0.9) within-subjects experiment was designed for testing the hypothesis. In the experiment, 20 pairs of same-gender friends (males and females were also 10 pairs) were recruited as participants with an average age 21.55. All participants were undergraduates or graduates. Participants were directed to play Prisoner’s Dilemma Game with same-gender friends and strangers respectively. In the procedure, different payoff matrices were randomly presented based on 9 levels of CI (0.1~0.9). The participant’s reward was entirely depended on outcomes of the game. The cooperative rate was computed from frequency of cooperation as a measurement in each conditions (In a condition, cooperative rates = 100%* numbers of trials of cooperating/ total of trials). To analyze those effects, we performed repeated-measured ANOVA and t test with gender, CI and Social distance on the cooperative rates. The results of the ANOVA revealed that main effects of CI and social distance were significant (CI: p < .001, ηp2 = .55; Social distance: p < .001, ηp2 = .33), while the interactive effect was insignificant (ps > .2). The results of t tests revealed that, with all CI conditions, the cooperative rates for gaming with friend were significantly higher than that with stranger (ps < .05) (Table 3). To determine how the social distance affected specific behaviors, we compared the cooperative rates in Close and Strange conditions with baseline (50% cooperative rate), and the results revealed competition both in close and strange conditions with CI = 0.1~0.3 (ps < .01). With CI = 0.4~0.6, it was appeared that chance level (no difference with baseline) was found in close condition (ps > .09) but competition was still in strange condition (ps < .001). Then, with CI = 0.7~0.9, cooperation in close condition (ps < .01) and chance level in strange condition were found (ps > .1). (Fig. 1 and Table. 3) For effects by CIs, the Pairwise Comparisons showed that differences among the cooperative rates with CI = 0.1~0.3 were insignificant. However, the cooperative rate with CI = 0.1 was significantly lower than that with CI = 0.4~0.9 (ps < .05). The operative rates with CI = 0.2 and 0.3 were significantly lower than those with CI = 0.5~0.9 (ps < .05). The cooperative rates with CI = 0.4 and 0.5 were significantly lower than those with CI = 0.6~0.9 (ps < .05). The cooperative rate with CI = 0.6 was significantly lower than that with CI = 0.7~0.9 (ps < .001). The cooperative rate with CI = 0.7 was significantly lower than that with CI = 0.9 (ps < .05). The cooperative rates with CI = 0.8 and 0.9 were insignificantly different (p = .56). (Table. 2). Those results suggested that both CI and social distance played positive roles in cooperative behaviors, but those two facilitating effects worked separately. Furthermore, for effect by CI, even cooperative rate was promoted by CI, but did not increasing immediately, which indicted a threshold in the promoting process (Table.3). In this experiment, the threshold was approximate CI = 0.3. For effect by social distance, a closer social distance could enhance individual’s cooperative inclination, which came to motivate more behaviors that are cooperative.

Key words: Cooperation index, Social distance, Prisoner’s dilemma, Cooperative rate

摘要： 利用囚徒困境范式，考察了合作指数CI（0.1~0.9）与社会距离（亲密、陌生）对个体合作行为的影响。结果发现，社会距离和合作指数均会对合作行为产生影响，两者的交互效应不显著；合作率会受到合作指数的促进作用，但不会随合作指数的增高立即增长；在CI = 0.1~0.9这9个水平下，亲密被试间的合作率均显著高于陌生被试间的合作率。这一结果说明，合作指数对合作行为的促进作用存在阈值；社会距离缩小可促进个体的合作倾向增强，并更快地达到合作水平；合作指数与社会距离相互独立地影响合作行为。

关键词: 合作指数, 社会距离, 囚徒困境, 合作率