Scalable news slant measurement using Twitter

Abstract

Prior research has reported evidence of polarization in the consumption and distribution of political news in online social media. Methods to effectively measure slant of news articles at a large scale are lacking in prior literature. To fill this gap, we propose a method to measure slant of individual news articles by observing their sharing patterns on Twitter. Our method monitors tweets about a news article and a set of landmark Democrat and Republican accounts on Twitter to estimate its slant. Our results show that the slant of news articles estimated by our method matches the slant estimated by crowdsourced workers.