Political content and news are polarized but other content is not in YouTube watch histories

Authors

  • Magdalena Wojcieszak UC Davis, U of Amsterdam
  • Rong-Ching (Anna) Chang
  • Ericka Menchen-Trevino

DOI:

https://doi.org/10.51685/jqd.2023.018

Keywords:

polarization,, news, YouTube, echo chambers, partisanship, exposure, selectivity, computational methods

Abstract

Research on ideological biases and polarization on social media platforms primarily focuses on news and political content. Non-political content, which is vastly more popular, is often overlooked. Because partisanship is correlated with citizens’ non-political attitudes and non-political content can carry political cues, we explore whether ideological biases and partisan segregation extend to users’ non-political exposures online. We focus on YouTube, one of the most popular platforms. We rely online data from American adults (N = 2,237). From over 129 million visits to over 37 million URLs, we analyze 1,037,392 visits to YouTube videos from 1,874 participants. We identify YouTube channels of 942 news domains, utilize a BERT-based classifier to identify political videos outside news channels, and estimate the ideology of all the videos in our data. We compare ideological biases in exposure to (a) news, (b) political, and (c) non-political content. We examine both exposure congeniality (i.e., are users consuming like-minded content?) and polarization (i.e. are there overlaps between Democrats and Republicans in the content they consume?). We find substantial congeniality in the consumption of news and political videos, especially among Republicans, and high levels of polarization in this exposure (i.e., limited overlaps between Democrats and Republicans). We also show that both exposure congeniality and polarization are significantly lower for non-political content, in that non-political videos are less likely to be ideologically like-minded and both Democrats and Republicans consume similar non-political content. Theoretical and practical implications of these findings are discussed.

Downloads

Published

2023-11-10

Issue

Section

Articles

How to Cite

Wojcieszak, M., Chang, R.-C. (Anna), & Menchen-Trevino , E. (2023). Political content and news are polarized but other content is not in YouTube watch histories. Journal of Quantitative Description: Digital Media , 3. https://doi.org/10.51685/jqd.2023.018