Article Summary
Introduction
While the American media landscape was once dominated by a small number of non-partisan news sources, it now includes a wide range of sources, many of which reflect partisan viewpoints. This study first asks whether Americans engage in selective exposure – the practice of choosing to obtain their news from like-minded sources – through a combination of tracking individuals’ news browsing data administering two surveys, all during the 2016 election. It then investigates whether selective exposure exacerbates polarization. This speaks directly to the consequences for American politics and civil society of major shifts in the American media landscape and individual news consumption patterns.
Analytical Approach
The researchers employed a two-wave panel survey, with participants opting in to having their web browsing behavior tracked between the two waves. A crowdsourced set of coders then analyzed the news websites which participants visited, recording the topic of each article and whether it was more favorable toward either political party. The authors then compared behavior in their 2016 sample to two previous studies conducted in 2009 and 2013. Finally, they examined whether the topic (campaign events, policy, strategy, or candidate scandals) and slant of the articles (whether the article was neutral or favored Democrats or Republicans, slightly or strongly) affected selective exposure.
Main Findings
Most strikingly, using the same measure of partisan preference for like-minded news sources that a 2009 study employed, the authors find that selective exposure in 2016 had grown threefold. Using a similar metric, they then compare to 2013 behavior, and find that partisan segregation in news sources in 2016 was twice as high. They argue that these increases are due both to increasing national mass polarization and the specific, highly charged context of a presidential election year. Additionally, they examine whether selective exposure varies across the topic of the news stories, finding that the partisan divide expands from .21 for all news visits to .33 for election related topics. Contrary to predictions, there was no significant difference in selective exposure regarding candidate scandals compared to other election topics. Finally, they find greater selective exposure for articles with a high partisan slant (.42) compared to moderate partisan slant (.39) or neutral (.33).
They then examined whether news exposure over the course of the browser tracking predicted feelings toward the presidential candidates. While they expected that exposure to more partisan news might exacerbate polarization in candidate assessments, they found no evidence of this trend. Rather, they see a relationship between news diet and candidate preference in the wave one survey (participants who prefer Democratic-leaning sources also prefer Clinton, while those who prefer Republican-leaning sources prefer Trump), and this relationship persists, but does not exacerbate at the wave two survey.
Implications
This research illustrates the growth of partisan selective exposure in news sources. While there were still a number of non-partisan sources among the most visited sites, there was a notable growth in viewing of partisan outlets, especially among Republicans. This study also suggests that people may opt into partisan news coverage more specifically regarding elections. As campaigns become longer, this could exacerbate the tendency to choose like-minded sources. Interestingly, they did not find evidence that exposure to more partisan news sources over the period analyzed (Aug-Oct 2016) led to more polarized attitudes toward the candidates, though they raised the possibility that candidate attitudes had mostly crystallized before they began tracking web browsing in August.
Questions left unanswered
The authors collected a remarkable data set and very effectively illustrated the prevalence of selective exposure to partisan news media, which has increased significantly, though has not reached extreme levels. However, important questions remain about the timing and directionality of the relationship between partisan news selection and polarized attitudes toward candidates and parties. Does polarization lead to increased selection of partisan news sources, or does selection of partisan news sources lead to increased polarization? If instead, polarization and news silos reinforce one another, then it will be important to focus on where in that cycle it is possible to intervene.
Methods and Analysis
Was the study and its analyses pre-registered?: No
Did the study rely on proxy variables to measure polarization?: Yes
Feeling thermometer, trait ratings, and emotional reaction to presidential candidates
Were standard p-value thresholds used (p<.05 or 95% Confidence Intervals that don’t overlap zero)?: Yes
- Largest p-value presented as significant: 0.05
Were correlational results interpreted with causal language?: No
Limitations / Weaknesses
Concerns about selection bias, reporting bias, small N, sample quality, design choices, etc. The main concern here is that of the 9,760 included in the initial survey, only 1,076 consented to the web browser tracking and completed the second survey. This attrition would be difficult to avoid, given the high privacy concerns with regard to browser tracking, and authors perform a number of robustness checks. Additionally, given the length and intense media coverage of the 2016 general and primary elections, it is possible that behavior during the specific period of tracking used in this survey varies significantly from other portions of a campaign (or, non-campaign periods).
Open Data & Analyses
Does the article make the replication data publicly available?: No
Does the article make the replication analysis scripts publicly available?: No
Article Citation
Peterson, E., Goel, S., & Iyengar, S. (2021). Partisan selective exposure in online news consumption: Evidence from the 2016 presidential campaign. Political Science Research and Methods, 9(2), 242–258. https://doi.org/10.1017/psrm.2019.55
Bibtex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
@article{Peterson2021,
author = {Peterson, Erik and Goel, Sharad and Iyengar, Shanto},
doi = {10.1017/psrm.2019.55},
file = {:Users/xanni/Documents/Readings for Mendeley/div-class-title-partisan-selective-exposure-in-online-news-consumption-evidence-from-the-2016-presidential-campaign-div.pdf:pdf},
issn = {20498489},
journal = {Political Science Research and Methods},
keywords = {American politics,mass media and political communication},
number = {2},
pages = {242--258},
title = ,
volume = {9},
year = {2021}
}