Evidence Pack: Does Cross-Partisan Dialogue Actually Change Minds? What the 'Contact Hypothesis' Data Shows

The United States is witnessing historic levels of political hostility. Americans consistently report hating the opposing party more than they love their own, and growing partisan animosity is increasingly associated with a breakdown in civic trust and a rise in social anxiety.[4]

But a growing body of psychological and political science research suggests this animosity is largely built on a foundation of misperception. We are not as divided as we think, and the cure for our mutual distrust might be simpler—and more scientifically grounded—than expected.[7]

The foundation of this research is the "Contact Hypothesis," a concept first introduced by psychologist Gordon Allport in 1954. Allport theorized that prejudice mostly stems from misinformation and ignorance, and that exposing individuals to members of different social groups can foster familiarity and reduce hostility.[6]

For decades, this theory was primarily applied to racial and ethnic prejudice. In 2006, a landmark meta-analysis of 713 independent samples confirmed that intergroup contact effectively reduces prejudice across a wide variety of marginalized groups. Today, researchers are applying this exact framework to the modern crisis of political polarization.[6][7]

The core problem driving modern political animosity is what researchers call the "basic morality bias." A 2024 study published in the Proceedings of the National Academy of Sciences found that partisans vastly overestimate their opponents' willingness to accept basic moral wrongs, such as theft or cruelty.[4]

This bias leads directly to political dehumanization. Because we mistakenly doubt the basic moral sense of the other side, we view them as fundamentally flawed rather than simply holding different policy preferences.[4]

Correcting this bias is surprisingly straightforward. The same study found that simply providing information that a single political opponent condemns blatant moral wrongs substantially increases a person's willingness to work with the other side and decreases dehumanization.[4]

But taking the next step—actually talking to the other side—is something most people actively avoid. A 2024 paper from the Association for Psychological Science revealed that individuals consistently underestimate the social connection they can make with a stranger who disagrees with them on contentious issues.[1]

Fearing a negative, awkward, or hostile interaction, people shy away from cross-partisan dialogue. Yet, across a series of experiments, researchers found that these conversations are far more gratifying and informative than participants expect, drawing people closer together rather than pushing them apart.[1]

This holds true even in highly contentious political environments. A study published in the Journal of Politics tested cross-partisan conversations between Democrats and Republicans immediately following the contested 2020 presidential election—a period of intense partisan threat.[3]

Participants chatted via text online. Despite the heated context, the researchers found that cross-partisan conversation caused a measurable 6.22-point increase in positive out-party affect and reduced social polarization for at least three days following the interaction.[3]

However, the scientific consensus also points to clear limitations. A 2022 study in Science Advances paired out-partisan strangers for video calls and found that the topic of conversation is the most critical variable in determining success.[2]

When participants were instructed to discuss non-political topics—such as describing their "perfect day"—affective polarization dramatically decreased. But when the conversations were explicitly directed toward areas of political disagreement, the depolarizing effects vanished entirely.[2]

Furthermore, the benefits of these brief interventions are not permanent. The Science Advances study noted that the reduction in partisan animosity decays over the long term, typically fading after three months if the social connection is not maintained.[2]

It is also crucial to distinguish between "affective polarization" (how much we dislike the other side) and "issue polarization" (how much we disagree on policy). Cross-partisan dialogue is highly effective at reducing the former, but it rarely changes the latter.[7]

In the Journal of Politics study, while participants felt warmer toward their political opponents, the conversations did not increase their perceptions of election integrity or alter their core democratic attitudes.[3]

This aligns with broader sociological research. As noted in The Guardian, a sea of evidence demonstrates that arguments and debates alone have almost no meaningful effect on people's political beliefs. Instead, our social relationships and shared experiences carry far more influence.[5]

This brings the science back to Allport's original 1954 framework. Allport stressed that contact only works optimally under four conditions: equal status, common goals, intergroup cooperation, and support from authorities.[6]

When we argue on social media, none of these conditions are met. But when we engage in structured, cooperative dialogue—or simply build friendships outside of political contexts—the contact hypothesis proves remarkably resilient.[7]

Ultimately, the evidence suggests that the internet has not permanently isolated our political views, but it has distorted our perception of each other. By stepping outside our echo chambers and engaging in genuine dialogue, we can dismantle the basic morality bias and rediscover the shared humanity beneath our political divides.[7]