Leveraging User-Generated Social Media Content to Understand Misinformation in American Politics

Social Data Research Fellowship

Abstract

While many claim that social media websites facilitate the spread of fake news, current research overlooks the role that content written by social media users plays in misinforming the public. Facebook posts and tweets can contain false information without being subject to rigorous fact-checking. This means that individuals may be exposed to inaccurate information from their online friends without even knowing that they should question it. In this project, I propose to study the extent to which user-generated content on social media platforms distorts information reported by mainstream news outlets. Using a variety of natural language processing methods, I will examine the text of social media posts made by news outlets, as well as the text of comments on those posts, between 2016 and 2020. In an effort to understand which types of news content are more likely to become distorted in the comments, I will conduct an experiment designed to capture the causal effect of the initial news source and the salience of the topic on information distortion. This project stands to change the way we think about misinformation on social media and point toward the need for a different toolkit for trying to combat the problem.

Research Team

Principal Investigator

Taylor Carlson

Assistant Professor, Washington University in St. Louis

  • Bio ▾

    Taylor Carlson is an assistant professor in the political science department at Washington University in St. Louis. She received her PhD in political science from the University of California, San Diego in 2019 and her BA in public policy and psychology from the College of William & Mary in 2014. Her work focuses on understanding the content and consequences of political conversations. Of particular interest is understanding the ways in which individuals distort information they consume in the news as they discuss it with other people, both online and face-to-face. She uses a variety of methodological approaches, including novel experimental designs, surveys, and text analysis of data generated from experiments and on social media. Her work has been published in several peer-reviewed journals, including the Journal of Politics and American Political Science Review. Her first book, coauthored with Marisa Abrajano and Lisa García Bedolla, was published with Oxford University Press in 2020. Her second book, coauthored with Jaime Settle, is under advanced contract with Cambridge University Press. She is currently working on a third book manuscript based on her dissertation research, which won the Jean Fort Dissertation Prize at UC San Diego in 2019.

Back to top