In the APSA Public Scholarship Program, graduate students in political science produce summaries of new research in the American Political Science Review. This piece, written by Sienna Nordquist, covers the new article by Jon Green, Stefan McCabe, Sarah Shugars, Hanyu Chwe, Luke Horgan, Shuyang Cao, and David Lazer, “Curation Bubbles“.
There are many differences between the 20th and 21st centuries: computers which used to be the size of entire buildings now fit in our pockets, skinny jeans have gone out of style only to come back again, and most forms of communication have become instantaneous. Less obvious is how the transition from print to social media has changed how individuals search for information and the number of sources they are exposed to. In their new APSR paper, Green and coauthors develop a new way of thinking about the partisan spread of information on social media sources. They describe what they term as “partisan curation bubbles” as an alternative to the direct consumption of partisan news sources. From this new frame of curation bubbles, Green and his coauthors suggest that it is not the partisan nature of news sources which is driving polarization online, but rather how partisan news stories are shared by specific curators to their (likewise partisan) followers. In tests of the distributions of Twitter and Facebook news sources and post engagement linked individual users to voting records, they confirm that both very ideological and neutral news stories are able to reach “atypical” viewers, but that this sharing comes from partisan curators instead of the amplification of partisan information sources.
In the past, individuals may have chosen their preferred partisan slant to the news (which newspaper they read), but they would have also been exposed to several news topics (like sports, commentary, politics, etc.). However, with the advent of social media, individuals became users who self-selected into the topics they most cared about. By following influencers or specific groups, they became exposed to several sources of information— both neutral and partisan sources—but ones that were curated through the partisans in their social media bubble. From the perspective of the news sources themselves, suddenly there was also a mechanism for reaching an “atypical” audience, or unlikely followers based on their partisan leaning. If content curators of various ideological leanings could be prompted to share their news source, then atypical viewers could be reached.
As an example of how this works in practice, the authors highlight the top 10 Democrat and Republican news headlines from the Wall Street Journal in 2017 and 2018. Importantly, the fact that the Wall Street Journal publishes news stories which reflect good news for Republicans and bad news for Democrats —and vice versa— demonstrates that it is a neutral source of information. Thus, if a neutral source of information reliably shares news stories that are more likely to be shared by Republicans, and also news stories that are more likely to be shared by Democrats, then the platform itself can be connected to many different users through partisan curators.
“Green and his coauthors suggest that it is not the partisan nature of news sources which is driving polarization online, but rather how partisan news stories are shared by specific curators to their (likewise partisan) followers.”Green and his coauthors expect that curation bubbles lead partisan curators to share ideologically different news sources, partisan audiences engage more with partisan news stories, and neutral news sources will reach partisan users. In examining user-level Twitter and Facebook data from 2017 and 2018, the authors are able to match 1.6 million Twitter accounts to voter records and to analyze over 200,000 news stories shared on Facebook from 780 domains. After classifying the ideology of the news article and news sources (domains), they find that there is a difference between the distributions of news sources and news stories shared by curators, suggesting that partisan curators are sharing various neutral and ideological news sources on both social media platforms. Consistent with prior research, they also find that there is more left-leaning than right-leaning political news sharing on Twitter, at least back in 2017-2018.
The authors also demonstrate that partisan news stories have more of a partisan appeal, and thus are more likely to be shared or otherwise engaged with online by users who have the same partisan affiliation, even if they are atypical readers of the news source. Finally, the authors compare the ideological score of the news sources with the difference in the partisanship of individual news stories with the news source’s ideological score. In other words, they examine the relationship between how neutral or extreme a news source is with how distinctly partisan the news stories are in practice. The authors find that this difference is greatest for neutral news sources, thereby confirming that neutral news sources publish stories which are partisan in both directions, which allows them to reach atypical partisan users more effectively than news sources which are highly partisan to begin with.
An exception to the widespread support across measures, types of engagement, and news sources the authors find for their expectations is that partisan sources which do not reflect domestic ideological leanings —like Russia Today— do not similarly reach atypical users. Foreign news sources which are classified as neutral because they share highly partisan information on both sides of the ideological spectrum may therefore be primarily reaching their main users. This suggests that stopping the spread of deliberate misinformation may require limiting social media network engagement with bad actors, whereas minimizing polarization online will require content curators to share differing news stories, instead of differing news sources.
- Sienna Nordquist is a 3rd year PhD Student in Social and Political Science at Bocconi University, Italy. She is also a visiting researcher at the WZB’s Transformations of Democracy Unit in Berlin, Germany. Originally from the US, Sienna was a Robert W. Woodruff Scholar at Emory University, received her Master’s degree from the LSE’s European Institute, and has been a Fellow at the Atlantic Council.
- GREEN, JON, STEFAN MCCABE, SARAH SHUGARS, HANYU CHWE, LUKE HORGAN, SHUYANG CAO, and DAVID LAZER. 2025. “Curation Bubbles.” American Political Science Review, 1–19.
- About the APSA Public Scholarship Program.