Abstract

YouTube is increasingly important for American politics. Long dominated by extremists, the demand for political media on YouTube means that there are now “channels” run by people from every point on the ideological spectrum. Increasingly cheap and simple technologies have lowered the barriers to entry, and direct payments from YouTube for popular channels has incentivized some to make careers as political YouTubers, and the 2016 election was a catalyst for many. I have identified 1,433 channels that primarily discuss American politics. My first goal is simply to characterize this population: How many videos will they produce in 2020, what topics are most prominent, what are their ideological leanings, and how popular are they? The primary research question aims to understand how these people decide what kind of media to produce: Are they mostly ideologues, hoping to promote an agenda, or are they profit maximizers who shift their messages to better accommodate their audiences? By examining the full transcript of the videos and all of the comments by viewers, I will test whether demand is driving supply: does the content of the comments left on a channel’s earlier videos predict the content of that channel’s future videos? If extremist content is mostly supplied by ideologues, the implied policy solution for YouTube extremism is to demote, suspend, or ban individuals who espouse extremist views. If that content instead comes from profit maximizers, the solution is to demonetize individual videos that espouse extremist views, eliminating the incentive for anyone to produce those videos.

Principal Investigator

Kevin Munger

Assistant Professor, Penn State University

Bio

Kevin Munger is assistant professor of political science and social data analytics at Penn State University. He received his PhD from New York University in 2018. Munger's research looks at how social media and other contemporary internet technologies have changed political communication. Munger has published research on the subject using a variety of methodologies, including textual analysis, field experiments, longitudinal surveys, quantitative descriptions of social media trace data, and qualitative theory. Munger's research has appeared in leading journals like the American Journal of Political Science, Political Behavior, Political Communication, and Political Science Research & Methods. His present interests include cohort conflict in American politics and developing new methods for social science in a rapidly changing world.