New mathematical model-based research suggests that exposing US social media users to viewpoints different than their own furthers political polarization
Mathematicians are teaming up with political scientists to create models of how social media divides us, and results suggest at least … Tweets: @mims , @ingridlunden , @mims , @somospostpc , @rjcc , @jeffjarvis , @garyrosenwsj , @jeffbercovici , @kevinokeefe , @saraheneedleman , @mdudas , and @jessefelder Tweets: Christopher Mims / @mims : Sometimes you research and write something, and it changes you. This is such a piece. If you care about the growing ideological chasm in our country, here's an introduction to what's driving it. (🧵 on polarization in the U.S.) https://www.wsj.com/... Ingrid / @ingridlunden : In other words, no matter how you look at it, social media has had a dire impact on political discourse, and arguably any issue that people feel strongly about. https://twitter.com/... Christopher Mims / @mims : This is especially important: The proposed solution to “filter bubbles” and echo chambers is exposing people to more views from outside their POV. But this new research suggests the way social media does this has the *opposite* of the intended effect. https://www.wsj.com/... @somospostpc : I have a mute list as long as a bad day. I really don't want political views in here, youtube or wherever. more powerful mute options (be it regex for nerds, or ML-powered topics for the rest) could be the next social platform revolution https://twitter.com/... Richard Lawler / @rjcc : intentionally pretending that the world is left/center/right means you have to do a lot of silly contortions to avoid obvious explanations. https://twitter.com/... Jeff Jarvis / @jeffjarvis : Who is more than anyone else responsible for the polarizing of America? THE OWNER OF THIS GODDAMNED NEWSPAPER. Not a mention of Fox, of course. Here's research on that: https://www.amazon.com/... And the myth of filter bubbles: https://www.amazon.com/... https://www.wsj.com/... Gary Rosen / @garyrosenwsj : “If I'm only seeing things that are good for my own side and really crazy from the other, maybe I should look for something slightly toward the center.” https://www.wsj.com/... via @WSJ Jeff Bercovici / @jeffbercovici : Social media polarizes by showing the most engaging content, meaning the most extreme, which repels rather than persuades. All platforms need to do to fix this is reverse their algorithms to favor the most boring, least profitable content. https://www.wsj.com/... Kevin O'Keefe / @kevinokeefe : Americans are more polarized than ever and a growing body of research suggests that social media is accelerating the trend, and many political scientists worry it's tearing our country apart. https://www.wsj.com/... by @WSJ's @mims https://twitter.com/... Sarah E. Needleman / @saraheneedleman : Americans don't disagree with each other any more than they used to—but they do feel more polarized. @mims explains the role of social media. https://www.wsj.com/... @mdudas : “Many Americans hate each other more than ever, but they don't disagree with each other any more than they used to.” https://www.wsj.com/... via @mims Jesse Felder / @jessefelder : 'Because social media and Balkanized TV networks tend to highlight content with the biggest emotional punch—if it's outrageous, it's contagious—when we're exposed to a differing view, it often takes an extreme form, one that seems personally noxious.' https://www.wsj.com/...
Context & Ripple Effects
This WSJ piece lands at the peak of the research wave on social media and polarization: mathematicians modeling opinion dynamics alongside political scientists, finding that cross-viewpoint exposure can backfire rather than heal divides. It sits between two empirical data points in the same week's arc — a study of ~200K US Internet users showing Facebook users read more polarized news the longer they stay on the service, and earlier reporting that Zuckerberg and others had shelved damning internal research into Facebook's polarizing effect.
The model-based framing matters because it gives platforms a moving target: if the mechanism is contested — some researchers later argue fears are overstated, per the New Yorker's survey of the debate — companies can plausibly dispute causal blame, as Facebook's own internal doc co-written by Chris Cox equipping employees with rebuttals shows.
First-order effects
- Platforms like Facebook face renewed pressure to defend their recommendation feeds, right as empirical work ties time-on-service to more polarized news diets with sharply asymmetric effects across political groups.
- Twitter's practice of inserting tweets from unfollowed accounts into feeds — which has inadvertently amplified extremist rhetoric — becomes harder to justify under models where forced viewpoint exposure worsens division.
Second-order effects
- Facebook's response shifts from research to messaging: internal documents arming employees against the polarization narrative signal that reputational management substitutes for product changes when the science is contestable.
- Academic entrepreneurs see an opening for intervention tools, a lane Stanford later occupies with an LLM-based system that downranks antagonistic X posts to reduce partisan animosity.
Third-order effects
- If the pattern holds, feed design moves from maximizing engagement to actively managing intergroup contact — algorithmic curation becoming a de facto instrument of social cohesion policy, with platforms, not regulators, deciding what 'healthy' discourse looks like.
The trend: Social media polarization research is shifting from measuring whether platforms divide us to engineering feeds that manage antagonism directly.