Prejudice is a preemptive judgment about a person based on their group membership rather than anything they have actually done. It runs deeper than simple ignorance or dislike. Research across psychology, neuroscience, and evolutionary biology points to prejudice as a product of cognitive shortcuts that evolved to flag potential threats, then got amplified and distorted by culture, language, and social structure. Understanding where it comes from, how it operates beneath conscious awareness, and what genuinely reduces it turns out to be considerably more complicated than most people assume.
Why Humans Developed Prejudice in the First Place
One of the more unsettling findings in psychology is that prejudice appears to have evolutionary roots. Researchers working on what is called the “behavioral immune system” have found that humans evolved psychological mechanisms to detect and avoid potential sources of disease. These mechanisms promoted disease avoidance but also produced an overgeneralized wariness toward anyone who looked unfamiliar or different, even when those people posed no actual health risk.1PubMed. Immunizing against prejudice: effects of disease protection on attitudes toward out-groups The same system that once helped our ancestors steer clear of infection also nudged them toward suspicion of outsiders, shaping social behavior ranging from mate preferences to conformity to intergroup hostility.2PubMed Central. The behavioural immune system and the psychology of human sociality
A broader framework sees prejudice as one product of multiple threat-management systems shaped by natural selection. Under this view, different prejudices map onto different ancestral threats: fears about physical safety, concerns about resource competition, anxiety about disease, and worries about group cohesion. Each type of prejudice, whether directed at strangers, competitors, or people who simply look or behave differently, can be traced to a distinct adaptive problem our ancestors faced.3Current Opinion in Psychology. An Evolutionary Threat-Management Approach to Prejudices This does not mean prejudice is inevitable or acceptable. It means the tendency is baked into our cognitive architecture in ways that require active effort to override.
What Happens in the Brain
Brain-imaging studies have repeatedly found that the amygdala, a region associated with detecting threats and processing emotional significance, responds differently to faces of people from racial outgroups compared to ingroup faces. In one early study, participants showed greater amygdala activation to outgroup faces, but only after repeated exposure. Their amygdala habituated to ingroup faces over time while sustaining its response to outgroup faces, suggesting the brain was continuing to flag the unfamiliar faces as socially significant.4PubMed. Differential response in the human amygdala to racial outgroup vs ingroup face stimuli
The interpretation of this amygdala activity has shifted over the years. Early accounts treated it as evidence of a hardwired racial bias. A more careful review of the neuroimaging literature argues that the pattern is better explained by culturally learned associations between certain groups and potential threat, rather than by some innate outgroup aversion.5PubMed Central. A review of neuroimaging studies of race-related prejudice: does amygdala response reflect threat? In other words, the amygdala is not a “racism center.” It is reacting to threat signals that culture has attached to certain faces.
Supporting this idea, a study of a patient with bilateral amygdala damage found that she still showed implicit racial bias on a standard test, just like control subjects. If the amygdala were the sole driver of implicit bias, destroying it should have eliminated the effect. It did not. All subjects, including the patient, showed a pro-Black stance on explicit self-report measures but a negative automatic evaluation of Black faces on the implicit test.6PubMed. Intact performance on an indirect measure of race bias following amygdala damage Prejudice, it turns out, is distributed across multiple brain systems and can persist even when one important node is removed.
How Early Prejudice Takes Root
The building blocks of prejudice appear astonishingly early in life. Infants as young as three months can visually discriminate faces by race, and by six months they can categorize them. Between ages six and eight, children actively sort others into racial groups.7PubMed Central. Children’s Racial Categorization in Context This is not the same as prejudice in the adult sense, but it provides the perceptual scaffolding on which prejudice can later be built. Children are natural categorizers. They look for patterns, and when the world around them treats race, gender, or other group markers as meaningful categories, children absorb those signals quickly.
Developmental researchers have argued that because stereotyping and prejudice emerge so early, understanding the causal mechanisms during childhood is essential to any serious effort at prevention.8Current Directions in Psychological Science. Developmental Intergroup Theory By the time adults sit through a workplace training session, the cognitive habits underlying prejudice have had decades to solidify.
The minimal-group effect shows just how little it takes to trigger bias in children. When kids are assigned to arbitrary groups with no history, no status difference, and no competition, they still show ingroup preferences on both explicit and implicit measures. They allocate more resources to their own group, attribute better behavior to ingroup members, and expect more reciprocity from them.9PubMed Central. Consequences of “minimal” group affiliations in children Group identity alone, even an identity assigned five minutes earlier, is enough to produce favoritism.
The Gap Between What People Say and What They Do
One reason prejudice is so persistent is that much of it operates below conscious awareness. Research distinguishes between explicit prejudice, which is what people openly report about their attitudes, and implicit prejudice, which is measured through reaction-time tasks and other indirect methods. The two do not always line up. A person can sincerely believe they hold no bias while still showing automatic negative associations on an implicit test.
Importantly, implicit bias is not just an abstract laboratory artifact. In one study, participants who showed stronger negative associations with Black people on an implicit test also had more negative social interactions with a Black experimenter, and their explicit self-reports reflected more prejudice as well.10Journal of Experimental Social Psychology. Relations among the Implicit Association Test, Discriminatory Behavior, and Explicit Measures of Racial Attitudes The implicit associations predicted real behavior in a real social situation.
Research has also shown that implicit prejudice is not one thing. Different tasks capture different processes. Some measure how quickly a prejudiced association is activated when a person sees a member of a target group. Others measure how automatically that association is applied to a judgment. A factor analysis of multiple measures found that these represent distinct components: automatic activation and automatic application, along with a separate explicit factor.11Review of General Psychology. Implicit and Explicit Components of Prejudice Prejudice is not a single dial that can be turned down. It is a set of interlocking processes, each of which may need to be addressed differently.
How Threat Fuels Prejudice
Feeling threatened is one of the most reliable accelerants of prejudice. Intergroup threat theory distinguishes between two types: realistic threats, which involve competition over tangible resources like jobs, housing, or safety, and symbolic threats, which involve perceived challenges to a group’s values, beliefs, or worldview. Both types increase prejudice, but they do not always work the same way. In experiments testing both, the level of religiosity moderated the relationship differently for each type, and gender interacted only with symbolic threat.12PubMed Central. Testing Intergroup Threat Theory: Realistic and Symbolic Threats, Religiosity and Gender as Predictors of Prejudice
When researchers looked specifically at what drives outright hate and aggression rather than garden-variety dislike, symbolic threats turned out to be the stronger predictor. Across multiple studies, hate was better predicted by the sense that an outgroup challenges your group’s way of life than by fears about economic competition.13Journal of Experimental Social Psychology. A threat-based hate model: How symbolic and realistic threats underlie hate and aggression This is worth sitting with. The things that push prejudice toward its most dangerous forms are not mainly about material scarcity. They are about identity and meaning, about feeling that who you are is under siege.
How Language Keeps Prejudice Alive
Prejudice reproduces itself partly through the way people talk. Researchers identified a pattern called the linguistic intergroup bias: people tend to describe desirable behaviors by ingroup members and undesirable behaviors by outgroup members in more abstract terms, while describing undesirable ingroup behaviors and desirable outgroup behaviors in more concrete, situation-specific terms.14PubMed. Language use in intergroup contexts: the linguistic intergroup bias
What this means in practice is subtle but powerful. If someone from your group does something generous, you might say “she is a generous person” (abstract, trait-level). If someone from an outgroup does the same thing, you are more likely to say “she handed him money” (concrete, one-time event). And the reverse applies for negative actions. The effect is that good ingroup behavior gets framed as a stable characteristic while good outgroup behavior gets treated as a fluke, and bad outgroup behavior becomes a trait while bad ingroup behavior becomes a one-off. Over time, this linguistic pattern reinforces stereotypes even among speakers who have no conscious intent to do so.
What Prejudice Does to Its Targets
Prejudice does not just reflect how the world is perceived. It reshapes the performance and well-being of those it targets. Stereotype threat, the phenomenon where awareness of a negative stereotype about your group impairs your performance, is one of the most studied examples. African Americans in a stereotype-threat condition performed significantly worse on cognitive tests compared to those in a non-threat condition.15PubMed Central. Effects of stereotype threat, perceived discrimination, and examiner race on neuropsychological performance: simple as black and white? The gap widened further when participants who reported high levels of perceived discrimination were tested by an examiner of a different race.
The mechanism is not mysterious. Stereotype threat disrupts performance through a combination of a physiological stress response that impairs the prefrontal cortex, active self-monitoring that diverts attention from the task, and efforts to suppress negative thoughts and emotions. These three processes consume the executive resources needed to perform well.16PubMed Central. An integrated process model of stereotype threat effects on performance The result is that prejudice can produce the very deficits it claims to be observing. A test designed to measure ability ends up partly measuring the psychological burden of being stereotyped.
When Prejudice Escalates to Dehumanization
At its most extreme, prejudice slides into dehumanization, where members of an outgroup are seen as less than fully human. Research has found that imagining outgroup members as having fundamentally different interpretations of shared concepts, like what it means to be a member of a nation, increases blatant dehumanization.17PubMed Central. Imagined otherness fuels blatant dehumanization of outgroups The more alien you perceive someone’s inner world to be, the easier it becomes to strip them of human qualities in your mind.
Dehumanization is not just an attitude problem. A meta-analysis of the relationship between dehumanization and aggression found a moderate overall correlation, with blatant dehumanization showing a stronger link to aggression than subtle forms.18Aggression and Violent Behavior. Dehumanization and aggression: A meta-analysis The connection was most robust when dehumanization involved both animalistic (treating people as animal-like) and mechanistic (treating people as machine-like or emotionless) dimensions. Dehumanization makes aggression easier to justify, and when both forms are at play, the brakes come off most completely.
Where Prejudice Enters Institutions
Individual attitudes and institutional structures interact in ways that can be difficult to untangle. When the people who staff decision-making roles carry biases, and when institutional incentives align with those biases, the result is systemic discrimination that can persist even without any single actor intending it. A study of South Korea’s asylum process found that falsification of Arabic-speaking asylum seekers’ interview records was rooted not in individual misconduct but in systemic flaws designed to maintain low refugee recognition rates and procedural efficiency.19Journal of Refugee Studies. Gatekeeping in asylum interviews: institutional bias and narrative inequality in South Korea’s refugee screening process The prejudice did not need to be conscious or deliberate to produce devastating outcomes. The system was structured so that bias was the default, and departing from it required active resistance.
Similarly, the way social identities intersect adds layers of complexity. The experience of a Black woman facing prejudice is not simply the sum of anti-Black bias plus anti-woman bias. The intersection creates emergent properties, distinct stereotypes and distinct patterns of discrimination that would not be predicted from either identity alone.20Social and Personality Psychology Compass. Stereotyping, prejudice, and discrimination at the intersection of race and gender: An intersectional theory primer Research that treats identities as independent variables misses these interaction effects, leading to interventions that may work for some group members but not others.
What Actually Reduces Prejudice
The most robustly supported strategy for reducing prejudice is intergroup contact. Positive, meaningful interaction between members of different groups consistently lowers prejudice. A landmark meta-analysis of over 500 studies found that contact was associated with prejudice reduction in the vast majority of cases, and the effects tended to generalize beyond the specific individuals involved to shape broader intergroup attitudes. Both face-to-face contact and online contact contributed to lower prejudice, with face-to-face contact showing the stronger relationship.21PLOS ONE. The contact hypothesis and the virtual revolution: Does face-to-face interaction remain central to improving intergroup relations?
Even indirect or “vicarious” contact, such as watching television shows that depict positive intergroup interactions, can reduce both explicit and implicit prejudice. In one experiment, viewers who watched shows featuring intergroup contact reported lower prejudice and greater perceived warmth toward Black individuals compared to a control group. The effects were mediated by increased empathy and reduced intergroup anxiety.22Race and Social Problems. Vicarious Intergroup Contact Via Television Shows to Enhance Empathy and Reduce Explicit and Implicit Prejudice Toward Black Individuals This finding matters because direct contact is not always possible or logistically easy, and media representation can reach enormous audiences.
Perspective-taking, the act of actively imagining the world from another person’s viewpoint, is another intervention with promise. When participants adopted a perspective-taking mindset, their sensorimotor resonance, a neural marker of empathic mirroring, increased for outgroup members specifically. The effect generalized: people who had taken the perspective of one outgroup member later showed more empathic neural responses to other outgroup members they had not encountered before.23PubMed. Perspective taking reduces group biases in sensorimotor resonance A video game designed around active perspective-taking found that gameplay combined with high empathy for the character produced the lowest levels of implicit bias. Empathy alone was not enough, and gameplay alone was not enough. The combination mattered.24PubMed Central. “Fair Play”: A Videogame Designed to Address Implicit Race Bias Through Active Perspective Taking
Why Standard Diversity Training Often Falls Short
Given how deeply rooted and multifaceted prejudice is, it should not be surprising that quick-fix interventions struggle. A comprehensive review spanning hundreds of articles on diversity training found that the evidence does not support decisive conclusions about best practices. Many evaluations of training programs rely on proxy measures, like short-term attitude shifts measured by survey, that are far removed from the real-world behavioral outcomes the programs are supposed to produce. The enthusiasm for diversity training, and the money spent on it, has outpaced the evidence that it works.25PubMed Central. Diversity Training Goals, Limitations, and Promise: A Review of the Multidisciplinary Literature
This does not mean all training is useless, but it does mean that the most common corporate approaches, a one-off session, a set of slides, a couple of exercises, are unlikely to dislodge cognitive habits that have been forming since infancy. The interventions that show the most promise in the research literature tend to be sustained, experiential, and emotionally engaging. They involve actual contact or meaningful perspective-taking, not lectures about the importance of inclusion. The gap between what research supports and what organizations typically deploy remains wide, partly because effective interventions are harder to package, scale, and mandate than a two-hour seminar.
There is also an underappreciated risk: poorly designed training can backfire. When people feel coerced into attending or perceive the training as accusatory, they sometimes become more defensive and more entrenched in their existing attitudes. The research base increasingly suggests that voluntary programs grounded in genuine engagement outperform mandatory ones that treat prejudice as a knowledge deficit to be corrected through information alone.