Call +1 352-376-0505 | Mid-Term Signals

OZEAN MEDIA | Strategic Public Affairs

We win the outside game with data-driven advocacy.

The Identity Trap: Why We Ignore Information From Political Opponents

TL;DR

Once politics becomes part of people’s identity, people will literally pay to avoid and discount information from opponents, a bias driven by goup loyalty, not the content itself. Also: Go Gators, Auburn isn’t cool.

Introduction

There is not enough money in the world to make me sit through my brother rambling about how great his alma mater, Auburn, “is.”

I am a Florida Gator through and through. In fact, if given the choice, I’d probably pay good money just to guarantee I never have to hear him yell “War Eagle” at Thanksgiving dinner again or any dinner for that matter. (and don’t get me started on this “school” not being able to decide Tiger vs a Hawk / Eagle; nor their ugly shade of orange! AND do not step to either of us with a “Roll Tide!”, you just sound silly.)

We expect that kind of irrational tribalism in sports. But here is the finding that should stop every political argument at your next family dinner: people will literally hand over their own cash to avoid hearing a basic fact from the other political party.  Not a lie. Not spin. An objective fact!

That is the core finding of a new peer-reviewed study in The Economic Journal. It turns out we aren’t failing to persuade each other because the facts are complicated. We’re failing because the messenger’s team jersey matters more than the message itself.

The researchers call this a microfoundation of political polarization. That phrase does a lot of work, so it is worth unpacking before anything else. A microfoundation is an individual level, psychological explanation for a large scale social pattern. Political scientists have spent decades documenting that polarization is real and growing. This study does something different: it goes looking for the small, private decisions inside individual heads that add up to that big public pattern. The pattern here is a bias in favor of your own side and against the other, built into how people choose what to read and how they weigh what they read once they have read it.

Citation and Links

Methodology

The researchers designed a two-wave study tracking more than 1,000 U.S. adults across two distinct settings: real-world political forecasting and an abstract, non-political puzzle.

Wave 1: Politics and Policy

Conducted the week before the 2020 presidential election, participants were incentivized with real cash to predict future U.S. unemployment rates and public health system rankings 10 months down the road.

Before updating their forecasts, participants entered two testing phases:

  • The Demand Test (What will you pay to avoid?): Participants were given a $3 cash endowment and could use a slider to shift the odds of receiving a news summary from a nonpartisan outlet versus a partisan one (The New York Times, Fox News, Nature, Reuters, etc.). Crucially, half the participants saw the outlet logos (Labeled), while the other half saw the exact same reporting with all outlet branding stripped away (Unlabeled).
  • The Processing Test (How do you weigh what you read?): Every participant was given two articles—one left-leaning and one right-leaning—covering identical facts. They rated each article’s credibility and updated their policy forecasts.

Wave 2: Stripping Out Politics (The Marble Experiment)

Why switch to marbles?

I’ll admit, this took me a minute or two to wrap my head around. What do marbles in an urn have to do with political polarization? Economists, right?

Here’s the logic: In the real world, if you ignore an article from the opposing party, it’s hard to tell why. Do you think the journalist is lying? Do you hate the policy they’re promoting? Or do you just hate the team jersey on the masthead?

To isolate the jersey from the message, researchers had to remove politics entirely.

  • The Setup: 740 returning participants played a simple probability game for cash: predict whether an urn holds mostly green or mostly yellow marbles. By default, participants had a 50/50 shot of either drawing a marble themselves or seeing a past guess from someone labeled only as a Democrat or a Republican.
  • The Demand Test: Participants could spend from a cash endowment to increase their chances of drawing their own marble rather than relying on another participant’s guess.
  • The Debiasing Treatment: To test whether people simply thought the other party was bad at math, researchers truthfully told half the participants that Democrats and Republicans in Wave 1 were identically accurate (both 67% correct). This completely eliminated any practical reason to prefer one party’s guess over the other.
  • The Processing Test: Participants received a private marble draw alongside two prior guesses from Democrats, Republicans, or anonymous players, allowing researchers to measure the exact mathematical weight people gave to their own eyes versus an ingroup or outgroup peer.

Here is the bottom line: A green marble has no tax policy. A yellow marble doesn’t vote. And both parties were proven to be equally smart at guessing. If someone still discounts or pays money to avoid a guess from the other party, it has nothing to do with policy, facts, or competence. It demonstrates people will reject truth purely because of who is speaking.

Results and Findings

The data revealed consistent patterns across both the political arena and the abstract marble game:

  • People pay real money to avoid the opposing team. In the baseline political test, participants spent 14.9% of their cash endowment specifically to avoid reading an article from an outgroup news outlet in favor of a nonpartisan source. Almost 40% of participants spent their own money to keep the other side off their screen.
  • People treat nonpartisan sources like their own side. Participants showed virtually no preference between an ingroup outlet and a nonpartisan outlet (spending a negligible 2.8% of their endowment). The hostility wasn’t an obsession with staying inside an echo chamber; it was an active aversion to the outgroup.
  • Hiding the source logo cuts avoidance in half. Stripping the news outlet labels reduced outgroup avoidance by 50% (dropping spending from 14.9% down to 7.4% of the endowment).
  • Priors are intensely sticky. Exposing people to balanced reporting from both sides did not close the partisan divide. Initial beliefs accounted for 62% to 80% of participants’ final updated forecasts. Instead of changing their minds, participants discounted the outgroup article’s quality, rating ingroup reporting significantly more reliable and outgroup reporting less reliable.
  • The bias survives even when the topic has zero politics. In the marble experiment, participants still spent nearly twice as much of their endowment to avoid guesses from political opponents compared to their own party (14.9% vs. 8.2%). In processing the evidence, participants placed 33% more weight on a guess from an ingroup member than an identical guess from an outgroup member, and 40% more weight on their own marble draw.

Deep Dive: The Mechanics of “Source Utility”

Economists traditionally assume people seek out information based on its instrumental value (how much it helps them make an accurate decision) or its content valence (how good the news makes them feel).

This paper isolates a third, more stubborn driver: source utility.

Source utility is an intrinsic preference for, or disgust toward, an information channel based strictly on the social identity attached to it, completely independent of whether the information is accurate or useful.  Think of brand loyalty. Some shoppers buy one brand over an identical competitor for no reason beyond attachment to the label. Here, the labels are political parties.

“People would rather pay to remain in their information bubble than receive free information from a political opponent.”

“Groupish” Reflex vs. “Universalist” Error

To understand who drives this effect, the researchers categorized participants using an initial “minimal group” allocation game and gave them $6 to divide between two anonymous strangers randomly assigned to arbitrary groups (“Circle” vs. “Triangle”):

“Groupish” people consistently favor their own side. “Universalists” do not show that favoritism. This split mattered a great deal. Removing source labels, which lowers how obvious someone’s group identity is, worked mainly on groupish participants. A separate treatment that told people all sources were equally accurate worked only on universalists.

Put plainly: groupish avoidance looks like a gut reaction, and universalist avoidance looks like a correctable mistake about which sources are trustworthy.

This divide explains why a single policy fix never works across the board:

  1. For Universalists, avoidance is an informational calculation. When told that Democrats and Republicans were equally accurate, universalists stopped paying extra to see their own private signal. They treated the problem as an error in perceived competence, and facts corrected it.
  2. For Groupish individuals, avoidance is a tribal reflex. Proving that the other party was just as accurate did not change their behavior at all. The only intervention that moved them was unlabeling or removing the team jersey so that their tribal alarm bells never went off in the first place.

Why It Matters

These findings help explain why political polarization appears so intractable. The problem is not simply a disagreement over facts or policies. It is rooted in a fundamental, identity-driven bias in how people select and weigh information. Simply providing more facts or fact-checking is unlikely to bridge divides if people are already predisposed to ignore or discount the information based on who is delivering it. This dynamic creates and reinforces information bubbles, making constructive public discourse nearly impossible. It also explains the business model for partisan media, which caters to the demand for identity-congruent information.

Critiques and Future Research

The study’s reliance on online experiments presents limitations. Behavior in a controlled experiment may not perfectly mirror real-world information consumption habits. And yes, the financial stakes were small, which might not capture the full cost people would bear in real life.

The interventions tested were effective at changing information selection but had little impact on information processing. Future research could explore interventions that target the cognitive process of belief updating. Studying whether different communication styles or framing techniques can overcome the tendency to discount outgroup information is a logical next step. Exploring these dynamics outside the American political context would also test the universality of the findings.

Practical Implications for Policymakers

  • Reduce the partisan salience of new policies. Frame initiatives around common values, not party affiliation.
  • Use nonpartisan or bipartisan messengers to communicate with the public, if possible. Credibility depends on the messenger, not just the message.
  • Create opportunities for positive intergroup contact to reduce the “us versus them” mentality that fuels source utility bias.
  • If you want to foster a bi-partisan reputation, stake-out issues before they become partisan.

Practical Implications for Public Affairs Professionals

  • Build broad coalitions that include unexpected allies. This can break down partisan framing and increase the perceived neutrality of a message.
  • Lead with data and analysis from respected, nonpartisan sources before introducing partisan validators.
  • For “groupish” audiences, the messenger’s identity is paramount. For “universalist” audiences, demonstrating the quality and accuracy of the information is more effective.

Conclusion

Put the fact-checkers down for a second, and don’t invite them to Thanksgiving dinner. This isn’t a knowledge deficit.

You cannot fact check your way out of a bias that decides, before a single word gets read, whose words are even worth reading.

People in this study would rather stay wrong and stay comfortable than pay nothing at all to hear the other side out, and a fair number of them would pay real money to keep it that way.

Group identity is not a footnote on political polarization. It is the engine. Until we confront the reality that we treat political sources the same way I treat Auburn football, dumping more facts into the arena will just mean louder shouting from further apart.

And as my mom used to tell me: be careful who and what you surround yourself with, because eventually, you become them.

To sum up: Go Gators!

Navigate Your Next Move.

We help organizations win contested political and regulatory fights using campaign-style strategy.

Don’t leave your outcome to chance.

Let’s Talk Strategy

20-minute introductory call.  No obligation.

RELATED INTELLIGENCE