A new study published in Political Psychology suggests that in the United States, political identity is the strongest driver of how people view others in everyday social situations—outweighing other key social identities like race, religion, class, or education. Even in community settings unrelated to politics, partisanship appears to be the most powerful predictor of interpersonal preferences, revealing the extent to which polarization has moved beyond the political sphere and into social life.
The study examined how political identity compares to other social traits in predicting affective polarization, a term used to describe the growing dislike between members of different political parties. The researchers sought to understand whether this polarization is driven mainly by political affiliation or by other traits commonly associated with partisanship, such as ideology, race, and religion. They also wanted to assess whether individuals with “cross-cutting” identities—those who do not align neatly with party stereotypes—are less likely to display polarized attitudes.
“It’s well known that Americans are politically polarized,” said study author Julie Norman, an associate professor at University College London. “It’s also clear that Americans are increasingly living in different social bubbles that reflect not only political differences, but also other types of divides, including preferences related to class, race, religion, education attainment, and geography. We wanted to better understand what was most driving this division in American social life. Is it pure politics, or is it one (or more) of these other factors?”
To address these questions, the researchers conducted a large-scale conjoint survey experiment using a nationally representative sample of 1,330 participants in the United States. Respondents were repeatedly shown side-by-side profiles of two hypothetical people with randomly assigned traits, including political party, ideology, race, gender, religion, education level, income, and geographic region. After viewing each pair, participants were asked to rate how favorably they viewed each individual and to make choices about who they would prefer as a friend, neighbor, or family member by marriage. These social distance questions allowed the researchers to examine both abstract attitudes and more concrete social preferences.
The results showed that political party identity had the strongest and most consistent effect on participants’ responses. People expressed more favorable views of profiles that matched their own party affiliation and showed significantly less favorable attitudes toward those affiliated with the opposing party. This effect was especially strong in general favorability ratings but still present in social distance decisions, such as choosing a preferred neighbor or in-law.
In fact, political identity was a more powerful predictor of social preferences than race, religion, or gender. While religious identity did show some effects—particularly negative biases toward Muslims and atheists—the impact of party identification was even stronger. For example, Democrats consistently rated Democratic profiles more positively than independents or Republicans and showed pronounced dislike for Republican profiles. Republicans similarly favored members of their own party, although the degree of out-group dislike was less intense than among Democrats.
“Overall, our findings indicate that political partisanship is the primary driver of social divides in America today, outweighing race and ethnicity, class, level of education, geographic locale, and other demographic traits and identities that are so often subject to social biases,” Norman told PsyPost. “Many of these traits and identities are indeed intertwined with partisan identity, yet even when we untangle those overlapping identities, partisanship remains the strongest and most significant indicator of social preferences.”
One of the study’s more nuanced findings is that out-group animosity tended to be stronger than in-group favoritism when participants rated profiles in terms of favorability. In other words, people disliked the opposing party more than they liked their own. However, when it came to the interpersonal social distance questions—such as selecting a friend or neighbor—the pattern was more mixed. A deeper analysis using Bayesian modeling techniques found that in-group preference was sometimes more significant than out-group animosity in those behavioral measures, suggesting that while political bias is present in social decision-making, it may not always translate into outright hostility.
The researchers also explored whether individuals or profiles with cross-cutting identities—those who defy partisan stereotypes would generate less polarized reactions. Contrary to expectations, they found no evidence that these identities reduced polarization. Respondents were not more accepting of people with diverse or atypical combinations of traits, nor were individuals with cross-cutting identities themselves less polarized in their responses.
“We were especially interested in the impact of ‘cross-cutting identities,’ or profiles that do not fully conform to partisan stereotypes, such as Black Republicans, or evangelical Christian Democrats. Contrary to our assumptions, cross-cutting identities do not appear to dampen social polarization,” Norman explained. “People’s basic preference for someone with the same political affiliation as them applied even to seemingly ‘atypical’ Democrats or Republicans. And people who themselves had cross-cutting identities were still more likely to prefer their political group over the other.”
Another key finding was that while both ideology and partisanship influenced social preferences, political identity typically had a stronger effect than ideological labels such as “liberal” or “conservative.” This was especially true among Democrats and liberals, who showed more consistent alignment with their party. Among Republicans, conservative identity played a somewhat stronger role than party affiliation, suggesting a slightly different dynamic.
The study also uncovered differences in how each party responds to race and religion. Democrats rated Black profiles more favorably than other racial groups, while Republicans did not display a statistically significant preference or aversion based on race. For religion, Republicans rated Jewish, Muslim, and atheist profiles significantly lower than comparable Christian profiles. Democrats also rated atheists somewhat less favorably, but were more neutral toward Muslim and Jewish individuals.
“Aside from partisanship, religion emerged as the other key trait that guided people’s choices, but religion’s influence also depended on the respondent’s political affiliation,” Norman said. “Republicans viewed Jewish, Muslim, and atheist profiles more negatively than otherwise identical Christian profiles, and Democrats also viewed atheists more negatively than Christian profiles.”
“Regarding how partisan respondents view racial identity, Democrats viewed Black profiles more favorably than Republicans did. But importantly, this was because Democrats were uniquely positively inclined to Black profiles, not that Republicans viewed Black profiles especially negatively. Interestingly, White participants did not view Black profiles any differently than non-White participants did. So, the racial prejudices that some might expect in these situations were superseded by people’s political allegiances.”
Taken together, these findings show how strongly political identity now influences Americans’ everyday social preferences, often eclipsing other traditional social divisions. The study provides evidence that polarization in the United States is not just about policy or ideology—it’s about who people feel comfortable interacting with in their communities.
At the same time, the findings also suggest that affective polarization may be more limited in certain real-world contexts. While abstract ratings of favorability showed strong partisan bias, behavioral measures of social distance produced more ambiguous results. People may still be willing to befriend or live next to someone from the opposing party, even if they express negative attitudes in more general terms.
The study does have limitations. It relied on hypothetical profiles and survey responses, which may not fully capture how people behave in real life. Social desirability bias—participants modifying their responses to appear more tolerant—was minimized through the survey design, but it cannot be entirely ruled out. Additionally, while the researchers expected cross-cutting identities to play a moderating role, they found no such effect, raising questions about how to measure the influence of complex, overlapping identities in studies of polarization.
Future research may focus more directly on real-world behavior and explore how people navigate political differences in specific social settings. It may also examine how individuals define their identities and which ones are most meaningful to them in different contexts. More work is also needed to explore the role of elite messaging, media framing, and social networks in reinforcing or mitigating partisan divisions.
“While we focus on the identities motivating polarization, recent scholarship has shown that polarization between parties is not inevitable, but rather reflects an interplay between supply and demand factors that interact and reinforce each other,” Norman explained. “That is, it is not just our identities that drive polarization, but rather external cues from politicians, political parties, interest groups, media, social media, etc that activate those identities. In future studies, we’d be interested in building on emerging scholarship to investigate what kinds of messaging or primes helps counter polarization or mitigate social divides.”
The study, “Why can’t we be friends? Untangling conjoined polarization in America,” was authored by Julie M. Norman and Beniamino Green.
