[Paper Review] Partisan Sharing: Facebook Evidence and Societal Consequences

Jisun An, Daniele Quercia+1 more|arXiv (Cornell University)|Sep 25, 2014

Social Media and PoliticsSocial Sciences21 citations

TL;DR

This study investigates partisan sharing on Facebook and Twitter, demonstrating that users predominantly share news aligned with their political views, even when exposed to diverse content. Using large-scale social media data and a controlled experiment, it finds partisan sharing distorts political perception but also correlates with higher political engagement and knowledge.

ABSTRACT

The hypothesis of selective exposure assumes that people seek out information that supports their views and eschew information that conflicts with their beliefs, and that has negative consequences on our society. Few researchers have recently found counter evidence of selective exposure in social media: users are exposed to politically diverse articles. No work has looked at what happens after exposure, particularly how individuals react to such exposure, though. Users might well be exposed to diverse articles but share only the partisan ones. To test this, we study partisan sharing on Facebook: the tendency for users to predominantly share like-minded news articles and avoid conflicting ones. We verified four main hypotheses. That is, whether partisan sharing: 1) exists at all; 2) changes across individuals (e.g., depending on their interest in politics); 3) changes over time (e.g., around elections); and 4) changes depending on perceived importance of topics. We indeed find strong evidence for partisan sharing. To test whether it has any consequence in the real world, we built a web application for BBC viewers of a popular political program, resulting in a controlled experiment involving more than 70 individuals. Based on what they share and on survey data, we find that partisan sharing has negative consequences: distorted perception of reality. However, we do also find positive aspects of partisan sharing: it is associated with people who are more knowledgeable about politics and engage more with it as they are more likely to vote in the general elections.

Motivation & Objective

• To investigate whether partisan sharing—sharing only like-minded political news—exists on social media platforms such as Facebook and Twitter.
• To examine how partisan sharing varies across individuals based on political interest and partisanship.
• To analyze temporal changes in partisan sharing, particularly around elections and during primary seasons.
• To assess the societal consequences of partisan sharing, including effects on political perception and civic engagement.
• To validate findings across platforms (Facebook and Twitter) and in controlled experimental settings to ensure external validity.

Proposed method

• Collected and analyzed 44,999 news articles from 37 U.S. news outlets shared by 12,495 Facebook users over an extended period.
• Developed a partisan score to quantify the political slant of news sources and users’ sharing behavior based on source affiliation.
• Conducted a controlled web-based experiment with UK BBC viewers of the political program 'Question Time' to study sharing behavior on Twitter.
• Combined survey data from participants with their actual sharing behavior to assess political perception and knowledge.
• Used statistical modeling to test hypotheses about partisan sharing across demographics, time, and topic importance.
• Validated findings across two platforms (Facebook and Twitter) to ensure consistency and external validity of results.

Experimental results

Research questions

• RQ1 Does partisan sharing exist on Facebook and Twitter, where users are exposed to diverse political content?
• RQ2 How does partisan sharing vary across individuals based on their political interest and partisanship?
• RQ3 How does partisan sharing change over time, particularly during election periods and primary elections?
• RQ4 What are the societal consequences of partisan sharing, such as distorted political perception or increased civic engagement?
• RQ5 To what extent does partisan sharing correlate with political knowledge and voting behavior?

Key findings

• Partisan sharing is strongly present: only 33% of U.S. Facebook news readers in the sample can be classified as politically moderate.
• Partisan sharing is specific to political news; it disappears for non-political content such as entertainment.
• Individuals who engage in partisan sharing are more politically interested and more likely to vote in general elections.
• Partisanship levels among Facebook users in a U.S. state remain stable over time but increase during primary elections before returning to baseline.
• Users perceive news outlets as biased based on their own political leanings, regardless of the outlet’s actual slant.
• The study’s findings are consistent across both Facebook and Twitter, supporting the external validity of the results.