Abstract

Are journalists and Facebook users equally interested in political news? Introducing the conceptualization and measurement of the “news sharing gap”, this study compares the sharing of political news by Facebook users to the production of political news by news media organizations. To paint a broad picture of these differences, we compare the news sharing gap (a) across election and routine periods and (b) across eight countries: Australia, Austria, Brazil, Germany, the Netherlands, Romania, Spain, and the United Kingdom. Analyzing 265,714 articles shared over 12 million times on Facebook, findings show that elections are broadly linked to increases in political news publication, but even larger increases in political news sharing. The study reveals how, overall, political news is shared more often than news publication patterns would suggest, proposing higher political interest by Facebook users than previously thought. In most cases, political news sharing far outpaces political news production in the form of a “negative” news sharing gap, with the relative demand for political news (in the form of news sharing) being higher than the supply. Lastly, building upon previous work, we propose and validate a distant supervised machine learning method for multilingual, large-scale identification of political news across distinct languages, contexts and time periods.

摘要

记者和 Facebook 用户对政治新闻是否同样感兴趣？引入“新闻共享差距”的概念化和测量，本研究将 Facebook 用户对政治新闻的共享与新闻媒体组织对政治新闻的制作进行了比较。为了全面了解这些差异，我们比较了 (a) 选举和常规期间以及 (b) 八个国家/地区的新闻分享差距：澳大利亚、奥地利、巴西、德国、荷兰、罗马尼亚、西班牙和英国. 对在 Facebook 上分享超过 1200 万次的 265,714 篇文章的分析结果表明，选举与政治新闻发布的增加广泛相关，但政治新闻分享的增加幅度更大。该研究揭示了总体而言，政治新闻的分享频率高于新闻发布模式所暗示的程度，提出 Facebook 用户比以前想象的更高的政治兴趣。在大多数情况下，政治新闻分享以“负面”新闻分享差距的形式远远超过政治新闻生产，对政治新闻的相对需求（以新闻分享的形式）高于供应。最后，在以往工作的基础上，我们提出并验证了一种远程监督机器学习方法，用于跨不同语言、上下文和时间段的多语言、大规模识别政治新闻。