Individual Submission Summary

Direct link:

Four Metrics for Assessing the “Health” of Online Political Discussions

Thu, August 29, 12:00 to 1:30pm, Hilton, Gunston West


Social media companies are facing increasing scrutiny regarding the damaging effects of communication on their platforms. From political polarization to the spread of misinformation, from incivility to hate speech, critics lament the proliferation of serious political harms. However, to date, both the scholarly community and the companies themselves have lacked tools for systematically measuring and assessing whether and to what extent harmful dynamics are unfolding across conversations on social media. This paper reports on four new metrics developed for Twitter: mutual recognition, diversity of perspectives, incivility, and intolerance. Using a mixture of automated and semi-automated computational techniques, we devise measures that allow a broad-based view of what perspectives are being shared on a given political topic on Twitter, how those perspectives are placed in dialogue with one another, and the extent to which anti-normative (uncivil) and intolerant messages impact conversations. The project also involves a series of Twitter user experiments designed to provide context for and interpret the metrics—that is, to help us understand what actually constitutes a “healthy conversation”. We present results from an examination of tweets addressing immigration in the United States.


©2020 All Academic, Inc.   |   Privacy Policy