A Fistful of Votes: Understanding the Voting System in Social Question and Answer Websites

Recently, there is an increasing trend that people start sharing their comments in online question and answer (QA) websites. For such websites, the ranking of answers is usually determined based on several factors like the received upvotes and downvotes, the publishing time, and the reputation of users, such that a highly ranked answer has much more chances to be propagated over the Internet. Due to the popularity of these QA websites, some public relationship companies try to cheat users by promoting or blocking certain answers. In this paper, we take Quora and Zhihu as two case studies to understand the impact of follower numbers, the publishing time, the upvotes and downvotes on the ranking of answers, and to investigate how these websites are vulnerable to voting spammers. Meanwhile, we propose a scheme to estimate the list of downvotes, which is usually hidden behind such websites.