Social Media Interaction Does Not Improve Political Knowledge, but It Does Polarize Us

Two new studies say there's no evidence of political learning on social media, but it does increasingly teach us to hate our opponents.

North Americans spend an average of two hours and six minutes daily interacting with various social media platforms, according to the online database Broadband Search. Globally, Facebook's 1.9 billion daily active users average 33 minutes per day, and Twitter's 206 million daily active users troll and doom-scroll for 31 minutes per day.

Two new political science studies investigate how all of this time spent on social media affects our politics. The first asks what, if anything, digital denizens learn about politics, while the second develops a model to explain how social media interactions spark culture wars by sorting people into antagonistic political tribes.

Published in the Journal of Communication, the first study finds that whatever else millions of social media devotees learn from their online activity—catching up with friends and finding new ones, checking out product reviews, and watching cat videos—one thing they do not do is learn about is politics.

There are "no observable political knowledge gains from using Facebook, Twitter,
or SNS [social network sites] in general; when measuring policy-specific, campaign-related, or general political knowledge; in election and routine periods; and when individuals use social media specifically for news or for more general purposes," find Israeli communications researchers Eran Amsalem and Alon Zoizner. They reached this conclusion after parsing data involving more than 440,000 subjects in a pre-registered meta-analysis of 76 different studies on the effect of social media on political knowledge.

One minor exception to this sweeping conclusion is that research using experimental setups finds small but statistically significant knowledge gains from using social media. However, in such experiments, Amsalem and Zoizner note that subjects are given no choice over what political information they see, so they may be artificially induced to pay greater attention to it than they ordinarily would scrolling out in the wild. The experiments also fail to account for information decay, since they test their subjects' political knowledge gains immediately after exposure to it. "Considering these caveats, our conclusion remains that social media contribute little, if at all, to political knowledge," write the authors.

While interacting with social media is not increasing people's knowledge of politics, the second study finds that social media does teach partisans to increasingly dislike their opponents, according to University of Amsterdam digital geographer Petter Törnberg. In his new article in the Proceedings of the National Academy of Sciences, he argues that digital media are driving affective polarization by enhancing the process of partisan sorting. Affective polarization is defined as the difference between positive in-group bias toward the party someone supports and negative out-group bias toward other parties.

Törnberg observes that accumulating data indicate that political polarization in the U.S. and other countries is not growing because people are increasingly isolating themselves with like-minded folks in social media echo chambers that confirm the righteousness of their views. Instead, Törnberg notes that the "empirical literature suggests that digitalization does not appear to lead to a reduction of interaction across political divide, but quite the opposite: it confronts us with diverse individuals, perspectives, and viewpoints, often in contentious ways." So how might social media contribute to the apparent increase in partisan rancor? Törnberg develops a model showing how political disputes, controversies, and debates on nonlocal social media can drive people to adopt increasingly extreme positions.

Back in the good old days, political and other differences of opinion were largely local and neighbors had many cross-cutting issues and commonalities that moderated their views of those who disagreed with them. "Social conflict is sustainable as long as there are multiple and non-overlapping lines of disagreement: we may differ on our views on one issue but agree on another; we may vote differently, but if we support the same football team or go to the same church, there remains space for interpersonal respect," writes Törnberg. "The recent rise in polarization is thus expressive of a gradual breakdown of this cohesive glue, driven by a gradual alignment of social, economic, geographic, and ideological differences and conflicts."

In Törnberg's model, the moderating influences of cross-cutting local social ties have loosened as people spend more time attending to and picking sides in nonlocal partisan Twitter fights. "By connecting individuals with others from outside their local social bubbles, digital media pressure local political cultures to align globally," he argues. "Over time, the system comes to sort on the global scale, with a single political culture becoming system-wide." That single political culture polarizes into two mutually hostile tribes.

Based on the results of his model, Törnberg argues that we need to rethink "digital media as not merely arenas for rational deliberation and political debate but as spaces for social identity formation and for symbolic displays of solidarity with allies and difference from outgroups. Digital media do not isolate us from opposing ideas; au contraire, they throw us into a national political war, in which we are forced to take sides."

The paradoxical conclusion of his analysis is it "suggests that attempts of media platforms to reduce polarization by acting against echo chambers—algorithmically increasing exposure to opposing ideas—may backfire, instead resulting in intensified polarization and conflict." Any similar effort by, say, a federal Disinformation Governance Board would surely inflame cultural warfare even more.

Of course no political science research is dispositive, but the disheartening upshot of these two exploratory studies is that social media users learn essentially nothing about actual political issues—but do learn to hate their political opponents.

For more background on political polarization in the United States, check out my article, "Why Is It So Hard To Admit When You're Wrong?"