[Paper Review] Political Discourse on Social Media: Echo Chambers, Gatekeepers, and the Price of Bipartisanship

Kiran Garimella, Gianmarco De Francisci Morales+2 more|arXiv (Cornell University)|Jan 5, 2018

Social Media and Politics31 references80 citations

TL;DR

The paper defines production and consumption polarity to quantify political echo chambers on Twitter, shows strong alignment between what users produce and consume, introduces gatekeepers, and predicts partisans and gatekeepers from content and network features.

ABSTRACT

Echo chambers, i.e., situations where one is exposed only to opinions that agree with their own, are an increasing concern for the political discourse in many democratic countries. This paper studies the phenomenon of political echo chambers on social media. We identify the two components in the phenomenon: the opinion that is shared ('echo'), and the place that allows its exposure ('chamber' --- the social network), and examine closely at how these two components interact. We define a production and consumption measure for social-media users, which captures the political leaning of the content shared and received by them. By comparing the two, we find that Twitter users are, to a large degree, exposed to political opinions that agree with their own. We also find that users who try to bridge the echo chambers, by sharing content with diverse leaning, have to pay a 'price of bipartisanship' in terms of their network centrality and content appreciation. In addition, we study the role of 'gatekeepers', users who consume content with diverse leaning but produce partisan content (with a single-sided leaning), in the formation of echo chambers. Finally, we apply these findings to the task of predicting partisans and gatekeepers from social and content features. While partisan users turn out relatively easy to identify, gatekeepers prove to be more challenging.

Motivation & Objective

• Motivate understanding of how political opinions propagate on social media and how echo chambers form in Twitter networks.
• Define joint content-network measures (production/consumption polarity) to quantify echo chambers.
• Investigate differences between partisan, bipartisan, and gatekeeper users in terms of network position and content shared.
• Explore the relationship between content production/consumption and network structure to explain echo-chamber dynamics.
• Assess whether user types (partisans, gatekeepers) can be predicted from social and content features.

Proposed method

• Define production polarity p(u) as the average political leaning of tweets with news links posted by user u.
• Define consumption polarity c(u) as the average political leaning of tweets with news links received by user u from followed accounts.
• Define delta-partisan and delta-gatekeeper thresholds to classify users on polarity axes.
• Use PageRank and clustering coefficient to characterize network centrality and community embedding.
• Utilize a dataset collection of 2.5B tweets plus multiple political/non-political datasets to compute the measures.
• Apply statistical comparisons (Welch’s t-test) to assess differences between partisan vs bipartisan and gatekeeper vs non-gatekeeper groups.

Experimental results

Research questions

• RQ1 Do production and consumption polarities align to indicate echo chambers on Twitter?
• RQ2 What network and content characteristics distinguish partisan, bipartisan, and gatekeeper users?
• RQ3 Do gatekeepers play a distinct role in cross-cutting information flow within echo chambers?
• RQ4 Can partisan and gatekeeper status be predicted from users’ tweet content and network features?

Key findings

• Production and consumption polarities are highly correlated in political datasets, indicating prevalent echo chambers on Twitter.
• Partisan users tend to have higher PageRank and clustering coefficients, and their content receives more engagement.
• Gatekeepers occupy high-centrality positions but have lower clustering, suggesting cross-community reach without deep embedding.
• Bipartisan behavior is associated with broader exposure and engagement, implying a “price of bipartisanship” in network positioning and content reception.
• Prediction models can identify partisans relatively well, while gatekeepers are more challenging to classify.
• Across datasets, partisan users show more polarized own polarity and higher centrality than bipartisans; gatekeepers show high centrality with diverse consumption but focused production.