Peer-to-peer interactions play an important role in the sharing economy and more specifically within Internet social networking sites (SNSs). However, little is known about the nature of these interactions and the role of reciprocity in the peer-to-peer exchanges facilitated by SNSs. We use the context of the Chinese Weibo SNS to explore this issue. Our research develops a conceptual model and hypotheses to examine the effects of reputation and emotion on reciprocal behaviours in the Chinese Weibo SNS. We conduct a study to test the influence of reputation on reciprocity and emotion in these peer-to-peer interactions. Our study shows that bridging behaviours, which create social status across heterogeneous groups, generate increased reciprocation as a direct effect. A second indicator of reputation, bonding behaviours that promote social resources within homogenous groups, do not have significant direct effects on intention to reciprocate. Both bridging and bonding behaviors create direct effects on evoked emotions, and emotions in turn have significant effects on the intention to reciprocate. We found that evoked emotions are an important mediator between bridging and bonding actions by one party in the sharing economy and the resulting intention to reciprocate by another party. Our findings contribute to developing a theoretical understanding of the role of reciprocity in peer-to-peer interactions in the sharing economy and more specifically in virtual environments. We show the functioning and appeal in the Chinese Weibo SNS depend heavily on reciprocal user behaviors including the exchange of emotional support, information and knowledge. Our findings show peer-to-peer interaction in SNSs create emotional responses, and these emotions mediate reciprocation. Actions by users with stronger reputations and higher status have stronger effects, which is consistent with the importance of social standing in China for fostering reciprocity in non-virtual business interactions. This suggests that offline habits are being emulated on the Internet.