Reputation and Contribution in Online Question-Answering Communities

Question-Answering (QA) communities have emerged as a rich source of information by allowing users to ask questions, contribute responses, and evaluate responses submitted by others. QA platforms encourage the contribution of useful content by establishing a measure of reputation, based on the volume and quality of each user’s contributions. While previous work has verified the positive relationship of reputation with response volume, its relationships of with other important aspects of contribution have been overlooked. Our work addresses this gap by studying how reputation relates to users’ risk-taking propensity (difficulty of questions tackled), performance (answer quality), and topical interests. First, we introduce a novel technique for measuring user ability and question difficulty and find that users tend to tackle harder questions as their ability grows. Interestingly, we also find that, while reputation is positively associated with risk-taking for users with very low reputation, the association becomes increasingly negative once the user’s reputation surpasses the community’s mean. Our third finding reveals the duality of the association between reputation and performance, which is positive when users tackle questions within their ability and negative when they reach beyond their expertise. Finally, we use a novel method to reveal the diamond-like shape of the correlation pattern between a user’s reputation and interests. We find that low-reputation users focus on a narrow set of introductory topics, medium-tier users have a wider spectrum of interests, and top-tier experts specialize on a small set of advanced topics. Our findings have important implications for online communities and reputation-based systems.