When voters believe their political rivals see them as full human beings rather than mindless brutes, they become more willing to negotiate, compromise, and work across party lines. New research published in Personality and Social Psychology Bulletin suggests this effect works mainly by reducing feelings of disgust toward the opposing party, and that it is especially relevant when people feel political divisions threaten the fabric of society.
In many Western democracies, political disagreements have increasingly become personal and moral clashes. Psychologists call this affective polarization, the tendency of partisans to distrust and dislike members of the opposing party, not merely disagree with their policies. Rather than seeing opponents as fellow citizens, partisans often view them as morally corrupt or even less than human.
Part of this hostility may stem from meta-perceptions, or beliefs about how one’s own group is viewed by outsiders. Partisans often experience meta-dehumanization, the belief that the other party sees them as subhuman, irrational, or animalistic. A 2020 study of American political partisans found that Democrats and Republicans both substantially overestimated how much the other side disliked and dehumanized them, and these exaggerated beliefs predicted stronger support for spiteful actions that harm democratic norms.
Feeling dehumanized can set off a cycle of retaliation. A 2016 series of studies on intergroup conflict indicated that people who feel dehumanized by a rival group tend to dehumanize that group in return and support aggressive policies against it.
The reverse belief, known as meta-humanization, may help repair relations between groups. In a 2021 study of Muslims and non-Muslims in Canada and the United Kingdom, the same two researchers behind the new paper found that perceiving the other group as seeing one’s own group as fully human reduced prejudice and encouraged friendly contact, even when tensions were high. But a 2024 study in post-conflict Kosovo found that feeling humanized by former adversaries did not reliably make people more willing to negotiate with them, suggesting the effect may depend on how immediate and threatening the division feels.
Less is known about how feeling humanized changes people’s emotions toward their opponents. Moral emotions shape how people treat other groups. Anger typically stems from perceived injustice and sparks confrontation, disgust is tied to a sense of moral contamination and motivates avoidance, and empathy encourages perspective-taking and care.
To examine which of these emotions explains the benefits of meta-humanization, and how they interact with fears that society is fracturing, the researchers conducted five experiments with a total of 1,541 participants in the United Kingdom, the United States, and Canada. The work was led by Melissa Pavetich, an assistant professor at Heriot-Watt University in Dubai and a member of the SOCIAL Impact Lab, together with Sofia Stathi of the University of Greenwich.
“A lot of research on polarization focuses on what partisans think of each other,” Pavetich told PsyPost. “We were interested in a different question: what people believe the other side thinks of them.”
“We wanted to know whether that belief matters in politics, and if so, why,” she said. “In particular, we wanted to identify the emotional route that links it to whether people are willing to reconcile.”
Across all five experiments, the researchers statistically controlled for frequency of out-partisan contact on a 1-to-5 scale, accounting for how often participants already encountered members of the rival party in everyday life. The first study, conducted in February 2021 via the online platform Prolific, included 353 British adults (average age 36, 61.1 percent female, 79.9 percent White), split between Conservative Party supporters (52.0 percent) and Labour Party supporters (48.0 percent).
Participants were randomly assigned to read fictional social media posts, styled as messages from Twitter (now X), supposedly written by supporters of the rival party. In the meta-dehumanization condition, the posts referred to the participant’s party as “animals and brutes lacking in self-control,” whereas in the meta-humanization condition, the posts described them as “sophisticated.”
After reading the posts, the British participants rated their feelings and attitudes on 7-point scales, where higher numbers meant stronger agreement or intensity. Both Labour and Conservative supporters who read the humanizing posts reported lower anger, lower disgust, weaker intentions to hurt or humiliate opponents, and greater comfort with close social contact, such as having a doctor from the opposing party.
Among Labour supporters, those in the meta-humanization group reported lower disgust (an average of 2.14 vs. 3.21 in the meta-dehumanization group), lower anger (1.94 vs. 2.61), lower hostility (2.05 vs. 2.55), and less desire for social distance (3.50 vs. 3.75). Conservative supporters showed a similar pattern when comparing the humanizing and dehumanizing conditions for disgust (1.68 vs. 3.09), hostility (1.82 vs. 2.32), and social distance (2.77 vs. 3.05), alongside a pronounced drop in anger (1.64 vs. 2.77).
Because the first experiment did not include a baseline control group or a measure of empathy, the authors designed Study 2a to add both elements while testing how moral emotions connect to broader political reconciliation and whether perceived threat to social cohesion alters the process. Threat to social cohesion refers to the worry that political divisions are eroding shared community values, mutual trust, and collective belonging. The remaining four experiments were preregistered on the Open Science Framework, with materials and data posted publicly.
Conducted one week before the 2024 U.S. presidential election, Study 2a recruited 301 American Republicans (average age 47, 63.8 percent female, 83.1 percent White) through Prolific. Participants first rated their perceived threat to social cohesion on a 7-point scale, reacting to statements such as “The values in our society have gone seriously off track,” and were then randomly assigned to read humanizing X posts from Democrats, dehumanizing X posts from Democrats, or no posts at all in a control group.
They next rated their anger, disgust, and empathy toward Democrats on 7-point scales, and evaluated how civilized and evolved they considered Democrats to be on a slider from 0 to 10. Finally, they completed 7-point scales measuring affective polarization and four conciliatory attitudes: support for cross-party negotiation, willingness to compromise, support for bipartisan teamwork, and openness to future contact with Democrats.
Republicans who read humanizing posts felt much more humanized by Democrats (an average of 5.31 out of 7) than those in the control group (3.91) or the dehumanizing group (1.52). Statistical modeling indicated that feeling humanized by Democrats lowered Republicans’ disgust, which in turn led them to view Democrats as more human and civilized, predicting lower affective polarization and higher support for negotiation, compromise, teamwork, and contact. This indirect pathway held across all outcomes under moderate perceived threat to social cohesion, compared to both the meta-dehumanization and control groups, and extended to high threat for several outcomes. The pathways through anger and empathy were not statistically significant.
Study 2b mirrored this setup in the same pre-election context in the United States, examining 302 American Democrats (average age 46, 66.6 percent female, 79.9 percent White) reading fictional posts attributed to Republicans. Democrats in the meta-humanization group scored an average of 5.73 out of 7 on feeling humanized by Republicans, compared to 3.43 in the control group and 1.70 in the meta-dehumanization group.
Just as with Republicans, disgust was the sole emotion that linked meta-humanization to viewing the opposing party as human and adopting more conciliatory attitudes. This chain reaction through reduced disgust held most consistently under moderate perceived threat to social cohesion, where it predicted lower affective polarization and greater support for negotiation, compromise, and contact compared to both other conditions (and teamwork compared to the control group), while extending to low threat for several conciliatory outcomes when compared to the meta-dehumanization condition and to high threat when compared to the control condition. The pathways through anger and empathy were not statistically significant.
To test whether these patterns held outside the intense atmosphere of a U.S. presidential election, the researchers replicated the experiment in Canada using Qualtrics Panel. Study 3a surveyed 310 supporters of the Conservative Party of Canada (average age 46, 55.4 percent male, 73.1 percent White), evaluating their reactions to supporters of the rival Liberal Party. Canadian Conservatives who read humanizing posts felt more humanized by Liberals (an average of 4.55 out of 7) than those in the control group (3.81) or the dehumanizing group (2.92).
Consistent with the American findings, feeling humanized reduced disgust and increased how human Canadian Conservatives viewed Liberals to be, which in turn predicted more conciliatory attitudes. This indirect effect emerged under moderate threat to social cohesion for conciliatory outcomes such as compromise and teamwork and extended across all outcomes, including reduced affective polarization, when perceived threat was high. The pathways via anger and empathy were not statistically significant.
Study 3b completed the series by surveying 300 supporters of the Liberal Party of Canada (average age 49, 51.5 percent male, 71.1 percent White) regarding their attitudes toward Canadian Conservatives. Liberals in the meta-humanization condition reported feeling more humanized by Conservatives (an average of 4.64 out of 7) than those in the control group (3.91) or the meta-dehumanization group (2.64).
Among Canadian Liberals who reported low to moderate threat to social cohesion, feeling humanized by Conservatives again reduced disgust and increased outgroup humanization, predicting lower affective polarization and stronger support for negotiation, compromise, teamwork, and contact compared to both other conditions. Uniquely in this fifth experiment, heightened empathy also served as a parallel bridge toward reconciliation alongside reduced disgust, though only when comparing the meta-humanization group to the control group under low to moderate threat.
Taken together, the five experiments suggest that learning one’s political group is respected as human promotes compromise mainly by disarming disgust rather than by soothing anger or sparking empathy. Because disgust functions as a psychological barrier that motivates people to avoid perceived moral contamination, lowering disgust removes a major obstacle to treating political opponents as fellow citizens worthy of dialogue.
“How we think the other side sees us shapes how we treat them,” Pavetich explained. Across the five studies, she said, “when people feel seen as human, they feel less moral disgust toward the other side, and that makes room for reconciliation.”
“I expected empathy to do much of the work, since it’s so often discussed as the path to bridging divides,” Pavetich noted. “Instead, disgust stood out as the main pathway. This suggests that reducing a negative moral reaction may matter more than building positive feelings.”
The researchers also noted a difference in how perceived threat operated across political camps: the pathway through reduced disgust was most consistent at moderate threat across groups, extending to high threat among conservative participants and to low threat among liberal and Democratic participants. The authors suggest that because conservative voters are often particularly attuned to threat cues, seeing goodwill from the opposing party when threat is elevated may be especially reassuring to them or create a psychological tension that resolves by warming toward the outgroup.
As a competing explanation, the authors note that low threat may simply allow people to feel safe and open enough to benefit from humanizing messages, whereas high threat can push people into a defensive posture that blocks reconciliation.
“We were also interested to find that these processes were strongest when people saw a moderate level of threat to social cohesion,” Pavetich said. “Feeling some stake in the conflict seems to make how you’re perceived matter more.”
“The effects were small to moderate,” Pavetich added. “Effects like this won’t erase partisan hostility on their own. They were consistent across countries and outcomes, though, and they come from people’s beliefs about the other side.”
“That matters because those beliefs are often inaccurate and can potentially be corrected,” she said. “Small shifts in how people think they’re seen could add up across many everyday political interactions.”
As with all research, there are some caveats to consider when interpreting these findings. Although the researchers experimentally manipulated the social media posts that participants read, participants’ emotional reactions, their humanization of the opposing party, and their willingness to compromise were all measured at the same time through survey questions, meaning the study cannot prove that a drop in disgust directly causes someone to become more open to political negotiation.
“These findings don’t mean people should simply be told ‘the other side sees you as human’ whether or not it’s true,” Pavetich cautioned. “Meta-perceptions work best when they reflect genuine signals from the other group.”
“Perceived threat was measured as an existing attitude rather than manipulated experimentally, so we’re more cautious about interpreting its role,” she added. “Finally, humanization was measured with a widely used but debated scale, the Ascent of Man scale. Future work using a range of measures would strengthen the conclusions.”
In addition, the experiments took place in three Western nations where politics largely revolves around two dominant rival parties, meaning the findings might not apply to countries with multi-party coalitions or entrenched violent conflicts where partisans have little desire to reconcile. Future work could test whether combining humanizing messages with face-to-face collaborative problem-solving helps partisans build lasting trust across the political aisle.
“One direction I’m especially keen to pursue is the role of disgust itself,” Pavetich said. “Disgust is closely tied to how people judge right and wrong, so I want to understand why it emerged as the key pathway here. Does it reflect seeing the other side as morally contaminated or as violating core values? And can that response be reduced directly?”
“Alongside this, I’m examining how intergroup contact interacts with meta-humanization,” she added. “We’re testing whether the quality and type of contact, both face-to-face and through media, strengthens the benefits of feeling humanized. I also plan to extend this work beyond Western two-party systems to multi-party contexts and regions that are underrepresented in this research.”
“I’d love readers to remember that polarization isn’t only about disagreement over issues,” Pavetich concluded. “It’s also about feeling looked down on. Signs that we see the other side as fully human, even when we disagree strongly, may do more to lower the temperature than winning the argument.”
The study, “Bridging the Political Divide: Meta-Humanization and Moral Emotions Shape Partisan Reconciliation,” was authored by Melissa Pavetich and Sofia Stathi.