With the rapid development of information technologies, online communities are becoming increasingly prevalent in the modern society. The level of user engagement is critical for the success of online communities. In the two studies reported in this thesis, I investigate various antecedents of user engagement such as user motivation, platform design/policy, and content characteristics. In the first study, we examine how the platform policy of showing likers’ identity affects users’ giving of “likes” and the underlying mechanism. We rely on an exogenous policy change that happened on an online community douban.com, where likers’ usernames were not shown before the change but publicly visible after the change. Our analyses suggest that users tend to “like” fewer, shorter and more emotiona...[ Read more ]

With the rapid development of information technologies, online communities are becoming increasingly prevalent in the modern society. The level of user engagement is critical for the success of online communities. In the two studies reported in this thesis, I investigate various antecedents of user engagement such as user motivation, platform design/policy, and content characteristics. In the first study, we examine how the platform policy of showing likers’ identity affects users’ giving of “likes” and the underlying mechanism. We rely on an exogenous policy change that happened on an online community douban.com, where likers’ usernames were not shown before the change but publicly visible after the change. Our analyses suggest that users tend to “like” fewer, shorter and more emotional articles after the policy change, which is consistent with their self-presentation motivation. This study provides insights into user behavior and motivation in online communities and has practical implications for platforms and content creators. In the second study, we investigate how content characteristics influence viewers’ engagement with the content on community-based question answering (CQA) platforms. Specifically, we examine the effect of emotional intensity of an answer on the number of votes, comments and rewards that the answer receives from viewers and how this effect is moderated by answer length. Analyses of data from zhihu.com, a CQA platform in China, suggest that the effect of emotional intensity on viewer engagement is negative for short answers but positive for long answers. This study resolves the inconsistency in previous findings on the impact of emotional expression in user-generated content. It also provides useful implications for platforms and content creators.