Un/Masking News In Cyberspace: Examining Censorship Patterns Of News Portal Sites In China

Internet censorship has been attracting much attention from scholars and institutes, and several aspects have been studied, including IP blocking, keywords filtering and deletion in social media. However, although news is frequently subject to censorship, censorial practices in news websites have never been comprehensively described or quantified. In this paper, we present the first empirical study to systematically examine news deletion on major news portals inside China, focusing on NetEase and Sina Beijing. Sina North America, which was found not to delete news, was used as a control. To automatically detect and archive news deletions, we developed a computer program and from October 30, 2011 to September 27, 2012 crawled nearly two thousand news articles deleted from the two domestic portals and none from the oversea one. Many of deleted news stories overlap across websites, which implies systematic control of content, the quintessential component of censorship, as opposed to editorial operations. Moreover, statistical analyses have revealed common patterns that are associated with higher likelihood of news deletion, and they are negative news, domestic news, news occurring in Beijing and/or having a nationwide impact, topical news covering politics, food and drugs, foreign affairs and military. Indeed, these patterns are well aligned with the goals of censorship (Peleg, 1993) and further suggest censorship over journalistic management. Surprisingly though, party organs, those news organizations tightly affiliated with the Chinese Communist Party, published a considerable amount of deleted news and did not give up the role of whistleblowers completely to commercial news media. Also distinct from prior studies, Tibet and other outlying areas were not predictors of news deletion, while they were for social media censorship (Bamman, O’Connor, Smith, 2012). Quite the opposite, news originating from Beijing and surrounding provinces was more likely to be deleted.