Skip to main content


A social scientist probes social media–induced polarization

Breaking the Social Media Prism: How to Make Our Platforms Less Polarizing

Chris Bail
Princeton University Press
240 pp.
Purchase this item now

The media writ large, and social media in particular, have a lot to answer for with regard to the polarization currently plaguing society. But the influence of so-called “echo chambers” on ideology is much more complicated than it first appears. In Breaking the Social Media Prism: How to Make Our Platforms Less Polarizing, computational social scientist Chris Bail has done the hard work of breaking down ideological polarization, misinformation, and the impact that other people’s content has on our social media experience.

The book brings the reader through a carefully constructed set of experiments, featuring bots, secret pairings between people on opposite ends of the political spectrum, and insightful personal stories, designed to identify promising interventions that might repair our fractured politics, as well as those that probably are not worth the effort of implementing. Bail does a masterful job weaving together insights drawn from people’s backgrounds, politics, and online reactions to help readers understand how individuals across the ideological spectrum experience social media platforms. The experiments answer many of the questions a curious reader may wonder about, often with surprising results.

The people and data science in this book are compelling enough to stand on their own, but Bail has also done excellent work making the book immediately relevant. It speaks directly to the present moment, with stories about Trump-style politics, the lead-up to the 2020 US election, the COVID-19 pandemic, and ongoing international political issues. The reader never feels as though she has strayed into the theoretical because all of the
book’s context comes directly from these strange times.

Bail argues that the structure of social media is designed to amplify extreme voices and make moderates nearly invisible, shifting the discourse. Algorithms then push the more extreme discourse to like-minded users. But is this design responsible for our polarization? If we were exposed to more diverse views, would we all shift back toward the middle?

Bail’s careful research suggests that is not the case. Through compelling personal profiles of a left-leaning moderate and a strong Trump supporter who participated in a long-term experiment to evaluate such an intervention, he shows that exposure to content from the other side of the ideological spectrum actually pushed the subjects further toward their existing positions: The liberal-leaning moderate adopted more left-wing ideas and embraced Democratic policy positions more strongly, and the Trump supporter ramped up her attacks on liberals.

Bail argues that platforms make it easy to misread the social environment. Rather than reflecting a world full of mostly well-meaning, moderate people, they twist reality, refracting our view (hence the eponymous “prism”) and distorting our understanding of ourselves and others. He further argues that people have a natural desire to experiment with their self-presentation and to receive feedback that helps hone and refine their self-expression and identity. Social media platforms satisfy this desire, providing immediate reward—or punishment—in the form of likes, shares, and comments. Together, these factors can refine and polarize our perception of ourselves without us even knowing that it is happening.

Social media also magnifies extremists, making political polarization appear worse than it really is. However, this perception makes people, especially moderates, less likely to engage and thus less likely to
be seen. The appearance of polarization feeds itself.

The book culminates in an experiment around a simulated redesign of a social media platform that brings anonymous users together in direct conversation. The result of chatting one-on-one with someone from the other side of the ideological divide? People became considerably less polarized. Bail hopes that this suggests that there are platform changes that can be made to make people more moderate. His experiments are solid, and I believe his results, although as a social media researcher, I am not quite as optimistic as he is about the success this would have at scale. (That said, I hope he is right.)

Breaking the Social Media Prism answers important questions about the origins of our current political environment and suggests how existing platforms and reward systems might be redesigned to make things better. Bail’s scientific conclusions are refreshing in a space dominated by informed speculation, and the book offers hope that data-driven solutions can bring us back from the brink.

About the author

The reviewer is at the Social Intelligence Lab, University of Maryland, College Park, MD 20742, USA.