Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Evaluating the Relationship Between News Source Sharing and Political Beliefs

Sofía M del Pozo, Sebastián Pinto, Matteo Serafino, Federico Moss, Tomás Cicchini, Hernán A Makse, Pablo Balenzuela

TL;DR

Despite a noticeable polarization based on political affiliation observed in a retweet network analysis, center-left users showed more diverse media sharing preferences, particularly concerning the MSI, a measure that captures bias in media outlets and user preferences within a hidden space.

Abstract

In an era marked by an abundance of news sources, access to information significantly influences public opinion. Notably, the bias of news sources often serves as an indicator of individuals' political leanings. This study explores this hypothesis by examining the news sharing behavior of politically active social media users, whose political ideologies were identified in a previous study. Using correspondence analysis, we estimate the Media Sharing Index (MSI), a measure that captures bias in media outlets and user preferences within a hidden space. During Argentina's 2019 election on Twitter, we observed a predictable pattern: center-right individuals predominantly shared media from center-right biased outlets. However, it is noteworthy that those with center-left inclinations displayed a more diverse media consumption, which is a significant finding. Despite a noticeable polarization based on political affiliation observed in a retweet network analysis, center-left users showed more diverse media sharing preferences, particularly concerning the MSI. Although these findings are specific to Argentina, the developed methodology can be applied in other countries to assess the correlation between users' political leanings and the media they share.

Evaluating the Relationship Between News Source Sharing and Political Beliefs

TL;DR

Despite a noticeable polarization based on political affiliation observed in a retweet network analysis, center-left users showed more diverse media sharing preferences, particularly concerning the MSI, a measure that captures bias in media outlets and user preferences within a hidden space.

Abstract

In an era marked by an abundance of news sources, access to information significantly influences public opinion. Notably, the bias of news sources often serves as an indicator of individuals' political leanings. This study explores this hypothesis by examining the news sharing behavior of politically active social media users, whose political ideologies were identified in a previous study. Using correspondence analysis, we estimate the Media Sharing Index (MSI), a measure that captures bias in media outlets and user preferences within a hidden space. During Argentina's 2019 election on Twitter, we observed a predictable pattern: center-right individuals predominantly shared media from center-right biased outlets. However, it is noteworthy that those with center-left inclinations displayed a more diverse media consumption, which is a significant finding. Despite a noticeable polarization based on political affiliation observed in a retweet network analysis, center-left users showed more diverse media sharing preferences, particularly concerning the MSI. Although these findings are specific to Argentina, the developed methodology can be applied in other countries to assess the correlation between users' political leanings and the media they share.
Paper Structure (13 sections, 6 equations, 4 figures)

This paper contains 13 sections, 6 equations, 4 figures.

Table of Contents

  1. Introduction
  2. Background
  3. Argentinian Context
  4. Data and Methods
  5. Data description
  6. Methods
  7. Correspondence Analysis
  8. Results
  9. Media Sharing Index
  10. Media sharing and political preferences
  11. Retweet user networks
  12. Discussion and Conclusions
  13. Original dataset

Figures (4)

  • Figure 1: Methodology pipeline.I. Raw data. Example of original data from Twitter (now X), with a tweet sharing an URL to a news article at the top and another one with a political hashtag at the bottom. II. User ideology. Hashtags were used to train a logistic regression model to classify tweets as supportive of either candidate. Users are assigned to the candidate for whom they demonstrate the highest number of supportive tweets (further details in zhou2021). III. User-media. The news URLs in the tweets are used to extract the media outlet. A bipartite network of users and media outlets is then created. For example, user i is linked to the media outlet El Destape because this user shared a news article from this media outlet.
  • Figure 2: Probability density of the Media Sharing Index (MSI). This graph shows the probability density of the MSI for users and the 12 main media outlets. Radio Mitre is excluded from this display due to its status as a positive MSI outlier. Histograms have been smoothed using a Gaussian kernel with a bandwidth of $0.15$. The grey lines indicate the positions of the media outlets.
  • Figure 3: Joint Probability Density of MSI and IV. This figure illustrates the relationship between the media sharing behavior of users and their ideological leaning.
  • Figure 4: Retweet network. The two largest communities detected by the Louvain algorithm are displayed (account for 75% of the entire network). Histograms illustrate the distribution of the MSI and IV for each community. Based on these distributions, the red community can be associated with a center-right political leaning, while the blue community can be associated with a center-left leaning.