Randomly assigned to Team Spain, people still believed biased facts about it

Published in Psychological Science, Association for Psychological Science

Why do people on opposite sides of the political aisle look at the same set of facts and reach different conclusions? The standard explanation is straightforward: they consume different media, encounter different information, and build different mental models of reality. Fix the information problem -- get everyone reading the same reliable sources -- and the bias should shrink.

A study published in Psychological Science suggests that explanation is, at best, incomplete. The bias appears even when there is no information gap to blame.

Can you create partisan bias from nothing?

Tyler Hubeny, a social psychologist at the University of Texas at Austin, and his colleagues wanted to isolate the mechanism behind partisan bias. The challenge with studying real political partisanship -- Democrats versus Republicans, Leave versus Remain -- is that participants arrive with years of accumulated knowledge, media exposure, and emotional investment. Untangling whether someone rejects a fact because they never encountered it or because they are motivated to disbelieve it is nearly impossible in that context.

So the researchers removed real politics entirely. They recruited more than 600 American citizens through an online portal and randomly assigned them to one of three groups: Team Spain, Team Greece, or No Team. The assignment mechanism was deliberately absurd -- a personality quiz styled after the kind of thing you might find on a social media site. Answer a few questions, get sorted into your team. No prior knowledge, no media ecosystem, no history of allegiance. Just an arbitrary label.

The phony quiz and the real bias

After receiving their team assignment, participants evaluated a series of factual statements. Some statements favored Spain (for example, claims about Nobel Prize winners or economic indicators), while others favored Greece. The researchers measured two things using signal detection theory, a statistical framework borrowed from perceptual psychology.

First, truth sensitivity -- how well participants could actually distinguish true statements from false ones. Second, acceptance threshold -- how willing they were to accept or reject statements in general, and whether that willingness shifted depending on whether the statement flattered their assigned team.

The results were clear. Team assignment did not change how accurately people could tell true from false. Their perceptual ability was intact. What shifted was their threshold for acceptance. Participants were more willing to accept statements that favored their randomly assigned team and more inclined to reject statements that did not.

This is worth pausing on. These people had no reason to care about Spain or Greece. They had no lifetime of cable news reinforcing their position. They had no social identity wrapped up in their team membership. They had been sorted by a quiz they knew was trivial. And still, the mere act of being labeled shifted how they processed factual claims.

Motivation over information

The finding supports what psychologists call the motivated reasoning account of partisan bias. In this framework, people do not simply arrive at biased conclusions because they lack good information. They arrive at biased conclusions because their identity -- even an arbitrary, freshly minted one -- creates a motivation to see the world in ways that validate their group membership.

The competing explanation, the differential knowledge account, holds that partisan bias is fundamentally an information problem. People in different political camps consume different media, develop different factual beliefs, and therefore reach different conclusions. This account suggests that better information environments -- fact-checking, media literacy, balanced news consumption -- should reduce bias.

Hubeny's experiment was designed to make the differential knowledge explanation impossible. By randomly assigning teams and using topics participants had no prior investment in, the researchers ensured that everyone started from the same informational baseline. Any bias that emerged could not be attributed to differences in what people knew. It had to come from somewhere else.

And it did.

Implications for fighting misinformation

The practical consequences are uncomfortable. Most misinformation interventions -- fact-checking labels, accuracy prompts, media literacy programs -- operate on the assumption that the problem is informational. People believe false things because they encounter false things and lack the tools or context to evaluate them properly. Provide better information, better tools, better context, and belief accuracy should improve.

This study does not invalidate that approach entirely. Information quality clearly matters. But it suggests that even a perfect information environment would not eliminate partisan bias, because part of the bias has nothing to do with information at all. It arises from the basic human tendency to evaluate claims through the lens of group identity.

Hubeny acknowledged in discussing the work that interventions targeting motivated reasoning are far less developed than those targeting information deficits. The field has robust tools for improving information quality. It has far fewer tools for addressing the psychological motivation to reject inconvenient truths.

What the study cannot tell us

The experiment used arbitrary teams and low-stakes factual claims about foreign countries. Real political partisanship involves deeply held values, moral commitments, social networks, economic interests, and years of reinforcement. Whether the motivated bias observed with Team Spain scales to the intensity seen in actual political divisions is an assumption the study supports but does not prove.

The sample -- 600-plus Americans recruited online -- is also not representative of broader populations. Online research panels tend to skew younger and more educated than the general population, and the dynamics of motivated reasoning may differ across cultures with different political structures.

The study also measured stated beliefs in a survey context, not real-world behavior. Saying you believe a flattering statement about your team in a psychology experiment is a low-cost act. Whether the same motivational bias drives high-stakes decisions -- voting, vaccination, financial choices -- requires different kinds of evidence.

Still, the core finding is hard to explain away. Strip away every informational difference, give people a team identity they know is meaningless, and they still evaluate facts through a partisan lens. That does not make the information problem irrelevant. It means the information problem is not the whole story.