Read this topic

Ordinary language uses "prejudice" as one loose word for anything from a private dislike to a violent act. Gordon Allport, a psychologist, argued this loose language hides three genuinely different things, and the difference matters for how you spot it, measure it, and intervene on it.

Allport separated stereotyping (a cognitive generalisation about a group, meaning a belief held in the mind), prejudice (an affective, or emotional, antipathy based on a faulty and inflexible generalisation about a group), and discrimination (the behavioural enactment of bias, meaning acting on the belief or feeling). For example, imagine a person who privately believes a stereotype about a group but never says or does anything about it, next to another person who openly refuses to hire someone from that group. The first shows stereotyping without visible discrimination; the second shows discrimination in action.

The exam trap is assuming these three components always travel together. They don't: a person can hold a stereotype without feeling prejudice, and can feel prejudice without ever discriminating. Exam questions often describe a scenario and ask you to name which of the three components is being tested, so keeping the cognitive, affective, and behavioural layers distinct is the whole skill.

One classic explanation is the frustration-aggression hypothesis, which holds that when people are frustrated and cannot direct their aggression at the true source, they displace it onto a safer target instead. This displacement process is called scapegoating: blaming a substitute out-group (a group the person does not belong to) for problems actually caused by something else entirely. Suppose a factory worker is angry at an employer he cannot safely confront, and instead directs hostility toward a minority group of co-workers: that redirection is scapegoating, not a rational response to anything the co-workers actually did.

Muzafer Sherif's Robbers Cave experiment, a famous field study of two groups of boys at a summer camp, found that competition over scarce resources actively generates hostility between groups, worsening prejudice rather than merely reflecting it. Hostility could be reduced, but only by introducing a superordinate goal (a shared aim that neither group could achieve alone, requiring joint effort), simply placing the groups near each other did nothing on its own. Allport built on this with his contact hypothesis: intergroup contact (members of different groups interacting) reduces prejudice, but only when four specific conditions are all present together: equal status between the groups, common goals, genuine cooperation, and institutional or authority support for the contact. He also proposed a five-step scale of prejudiced action, from antilocution (negative talk about a group, such as jokes or slurs, without direct contact) through avoidance, discrimination, physical attack, up to extermination, showing that everyday attitudes and extreme violence sit on the same underlying continuum.

Two further theories are commonly confused with Allport's ideas and with each other. Social identity theory explains in-group bias (favouring one's own group) through categorisation and the need to protect self-esteem, and does not require any real competition between groups. Realistic conflict theory, by contrast, specifically requires actual or perceived competition over resources, exactly as Sherif demonstrated at Robbers Cave. Stereotype threat is a further distinct concept: an individual's anxiety about confirming a negative stereotype about their own group, rather than a mechanism by which one group displaces aggression onto another. The safest way to tell them apart: check whether real competition for a resource is present (realistic conflict theory), whether self-esteem and categorisation alone would explain it (social identity theory), or whether the anxiety belongs to the person's own performance rather than their attitude toward another group (stereotype threat).

Because prejudice has cognitive, affective, and behavioural layers, no single question captures it. There is no diagnostic criterion in the psychiatric-disorder sense and no single screening test that diagnoses prejudice. In practice, researchers and clinicians use several different tools depending on which layer they're probing: explicit attitude scales and self-report questionnaires for the affective component; the Implicit Association Test (a computer-based reaction-time task measuring automatic, less-conscious associations between a group and positive or negative concepts) for attitudes people may not consciously endorse or admit to; social distance measures, which ask how close a relationship (neighbour, colleague, family member by marriage) a person would accept with an out-group member, as a proxy for underlying antipathy; and direct observation or documented incidents for the behavioural, discriminatory layer. Consider a person who scores low on an explicit attitude questionnaire but shows a strong automatic negative association on an implicit measure: that gap is itself a recognised and testable finding, illustrating why relying on self-report alone can miss real prejudice.

Knowing Allport's four contact conditions is only useful if you can apply them. Real-world interventions built on this theory deliberately engineer all four conditions rather than hoping mere proximity will help. The jigsaw classroom, a cooperative-learning technique developed by Elliot Aronson, is a well-known example: it assigns each student a unique piece of information needed by the whole group, forcing genuine interdependence and cooperation toward a shared academic goal, with a teacher providing institutional backing, directly recreating Allport's four conditions in a school setting. WHO mental-health human-rights guidance and current anti-stigma toolkits treat stigma and discrimination as practical barriers to care, so management means anti-discrimination policy, social-contact intervention design, and supported reporting routes, not medication or individual psychotherapy. Diversity and anti-bias training programmes in workplaces attempt the same thing on a larger scale, though they only work reliably when leadership visibly backs the initiative (institutional support) and when the activities require actual joint effort toward a shared outcome, not just shared physical space.

The classic trap: a scenario mentions "contact" between two groups and invites you to assume prejudice will fall. Check whether all four of Allport's conditions: equal status, common goals, cooperation, and institutional support, are actually present. If even one is missing, contact alone would not be expected to reduce prejudice under Allport's model.

Prejudice has three separable components (stereotyping, prejudice, discrimination), is measured through a mix of self-report, implicit, and behavioural tools, and is reduced not by mere contact but only when Allport's four conditions: equal status, common goals, cooperation, and institutional support, are deliberately engineered together.