Facebook News Exposure and Acceptance in the 2016 Election

• In Event: A Matter of Trust: Media Credibility, Platforms, and Context

Abstract

This study examines how recent changes in the political and media landscape may have engaged (and enraged) partisans in the United States. In particular, it examines how users process election news distributed to them through social media, and whether such news amplifies partisan polarization, hatred and contempt. 

While recent scholarship has done an excellent job using large-scale monitoring of news sharing on social media (see Grinberg et al. (2019) and Guess et al. (2019) for examinations of user sharing behavior on Twitter and Facebook, respectively), this study uses a combination of surveys and Facebook scraping to better understand not only the information shared with users through social media, but also their individual-level characteristic and their evaluations of and reactions to that information. 

While many have observed the increasing polarization of American politics in government and the electorate, others have noted that this polarization seems to extend well past traditional politics. Iyengar, Sood, and Lelkes (2012) documented the increasing dislike (or even hatred) Americans feel toward their partisan opponents and suggested campaign communication might play a role in this polarization. In this study, we test whether exposure to and trust in highly-partisan or fake news organizations predicts greater levels of affective polarization, controlling for other partisan, attitudinal, and psychological variables. We hypothesize that exposure to lower-quality, one-sided information about politics should increase affective polarization, while exposure to higher-quality, balanced political information should decrease it.

To provide a measure of news exposure, we asked a convenience sample of more than 200 politically diverse undergraduates across two universities to search their own Facebook news feed for Trump- and Clinton-related stories appearing prior to the 2016 election (10 mentioning each candidate, for a total of 20 per participant). We then used a variety of indexes and ratings sites, as well as independent coding by trained research assistants, to characterize the relative prevalence of "fake news" providers and partisan news in their feed, as well as their subjective reactions to such stories. Those subjective reactions are part of a larger bank of survey questions measuring participants’ affective polarization toward the opposing party using a variety of instruments, as well as their partisanship, ideology, and a variety of demographic and psychological controls. Finally, using a series of embedded split-ballot surveys, we test for the implications of extreme partisanship on participants’ attitudes toward justice, fairness, and political violence.

Authors

• Tim Groeling, University of California-Los Angeles

• Johanna Dunaway, Texas A&M University