How language use on Facebook drives affective polarization
| dc.contributor.advisor | Stroud, Natalie Jomini | |
| dc.contributor.committeeMember | Hart, Roderick P. | |
| dc.contributor.committeeMember | Jarvis, Sharon | |
| dc.contributor.committeeMember | Lukito, Josephine | |
| dc.creator | Kim, Yujin (Ph. D. in communication studies) | |
| dc.creator.orcid | 0000-0001-6607-2312 | |
| dc.date.accessioned | 2022-06-24T18:39:58Z | |
| dc.date.available | 2022-06-24T18:39:58Z | |
| dc.date.created | 2022-05 | |
| dc.date.issued | 2022-05-06 | |
| dc.date.submitted | May 2022 | |
| dc.date.updated | 2022-06-24T18:39:59Z | |
| dc.description.abstract | Social media has changed people’s experience with political language. Social media platforms have become places where political tolerance is rarely preserved and expressions of high levels of intense dislike proliferate among the American public toward the opposing political party. Focusing on political content on Facebook, I argue that partisan language, when combined with a highly charged affective aspect, exacerbates affective polarization. I further propose that the type of language that increases affective polarization is promoted by social media algorithms, hence making partisan content more visible. To examine these dynamics, I provide both longitudinal and cross-sectional evidence of how partisan language, combined with affect, influences users’ political attitudes, leading to increases in affective polarization in society-at-large. At the aggregate level, I look at the effects of affective political language on engagement metrics and also examine time-series analysis to see how partisan affective language correlates with feelings toward the political parties. Using a computational social scientific lens, I apply natural language processing techniques such as sentiment analysis, incivility identification, and the determination of partisan in-/out-group entities to identify affective language in Facebook hyper-partisan pages. I use data consisting of 14 million posts from 1,493 hyper-partisan pages on Facebook between January 2008 and mid-November 2020 were as well as aggregated public opinion surveys corresponding to those timelines. At an individual level of analysis, I test the effects of partisan affective language on polarization using a mock Facebook setting. The survey experiment provides causal evidence about the relationship between language and polarization. Results confirm two routes to polarization: negative and uncivil language attracts more engagement leading to more visibility and such language corresponds with how people feel about the political parties. This dissertation further emphasizes how incivility and negativity related to polarization differently and how the presence of political targets influences affective polarization. Partisan differences in the mechanisms of engagement with political posts and affective polarization are discussed. The dissertation overall shows that political language can shape affective polarization, a process that is facilitated by social media. | |
| dc.description.department | Communication Studies | |
| dc.format.mimetype | application/pdf | |
| dc.identifier.uri | https://hdl.handle.net/2152/114766 | |
| dc.identifier.uri | http://dx.doi.org/10.26153/tsw/41669 | |
| dc.language.iso | en | |
| dc.subject | Affective polarization | |
| dc.subject | Social media | |
| dc.subject | ||
| dc.subject | Language | |
| dc.subject | Hyper-partisan content | |
| dc.title | How language use on Facebook drives affective polarization | |
| dc.type | Thesis | |
| dc.type.material | text | |
| thesis.degree.department | Communication Studies | |
| thesis.degree.discipline | Communication Studies | |
| thesis.degree.grantor | The University of Texas at Austin | |
| thesis.degree.level | Doctoral | |
| thesis.degree.name | Doctor of Philosophy |
Access full-text files
Original bundle
1 - 1 of 1