Abstract: Gossip is a form of affective information about someone who is not present, which shows a great influence on interpersonal relationships. It was found that negative gossip is positively related with poor relationships within the enterprise. In addition, it also proves that gossip can affect cooperation between listeners. Despite the wealth insights have be gained from previous research, it is still unknown that if the impact of gossip on interpersonal trust can be generalized to similar others. Moreover, gossip has be deem as a putative intrasexual competitions strategy from an evolutionary perspective, which presents significant sexual differences in gossip frequency, content, attitudes, as well as tendency. It remains an open question that whether any sexual differences would be involved in the generalization effect of gossip on interpersonal trust. In the present experiment, the learning–testing paradigm was applied. Thirty one participants were recruited, three of whom were excluded from final analysis for questioning the authenticity of the trust game. All participants were firstly instructed to complete an associated learning task with faces and personal information randomly paired. The type (gossip and real personal information) and valence (positive, neutral, and negative) of the personal information was manipulated. After learning, different impressions of 12 neutral faces were formed. Then, participants joined an online trust game with different partners in a role of investor. Half of the partners’ faces are completely strange, while others’ are 40% similar to the learning faces. The investment data during the trust game were recorded and analyzed. At the end of the experiment, participants were asked to assess the trustworthiness of the learning faces on a 7-point Likert-type scale ranging from 1(extremely untrustworthy) to 7 (extremely trustworthy). It is worth noticing that real personal information were included in the present experiment to elicit specific processing for gossip information. However, the data for the real personal information condition were not present here. In the learning phase, a repeated-measures ANOVA with information type, information valence and facial gender as three within-subject variables on trustworthiness ratings indicated a significant main effect of information valence. The ratings decreased gradually from the faces paired with positive to neutral and then negative information conditions. A significant main effect of facial gender was also found, with slightly higher ratings for female faces than for male faces. Importantly, in the testing phase, a same repeated-measures ANOVA was conducted on the amount of investments which revealed a significant main effect of information valence. The investments decreased gradually from the faces related with positive to neutral and then negative information. Moreover, a significant three-way interaction was also reported. Further analysis indicated that the influence of gossip on interpersonal trust appeared with a sexual bias. For female partners, when their physically similar faces are paired with positive gossip information, investments are higher. Conversely, when their physically similar faces are paired with negative information, the investment are lower, indicating a generalization effect of interpersonal trust for gossip information. However, for male partners the influence of gossip on the investment disappeared. On the contrary, the influence of real information on interpersonal trust was shown for both female and male partner without a sexual bias. In conclusion, the results from the present study show that the influence of gossip on interpersonal trust can be selectively generalized to physical similar strangers. The findings of the current research will contribute to our understandings of gossip, and further shed light on studies of social learning and generalization.

Key words: interpersonal trust, gossip, generalization, sexual bias

摘要： 本研究使用学习-测试范式考察传言对人际信任影响的泛化效应。学习阶段，被试对中性面孔和不同效价的传言或真实信息进行配对学习；测试阶段，被试作为投资者与陌生对家完成信任投资游戏，对家面孔与学习面孔具有40%相似性。结果发现,传言能够影响学习面孔的信任，并调节与学习面孔相似的对家面孔的信任投资；同时，传言对信任投资的影响只发生在女性对家身上。上述发现揭示传言对人际信任能够进行选择性泛化。

关键词: 人际信任, 传言, 泛化, 性别偏向