Prejudice is the term for showing uninformed opinion or assumptions toward an individual or group based on categories of characteristics. It refers to a prejudgment formed before adequate evidence is gathered, typically about people sorted by race, sex, age, religion, nationality, or similar social categories. The word names the attitude itself, not necessarily an action. Understanding what prejudice is, how it forms, and how it differs from related terms matters for anyone trying to recognize it in themselves or others.
Prejudice earns its name from the Latin praejudicium, meaning a judgment made in advance. That etymology captures the core mechanism: an evaluation is rendered before the relevant facts about a person are known. The attitude can be positive or negative in principle, but in behavioral science the term almost always refers to negative or hostile preconceptions.
The mechanism behind prejudgment
Prejudice operates as a cognitive shortcut layered onto social categorization. The human brain sorts people into groups automatically, often within milliseconds of seeing a face. Once a person is filed into a category, stored beliefs about that category get applied to the individual. This is efficient thinking, and that efficiency is exactly the problem: it substitutes group-level generalizations for person-level evidence.
Two psychological processes drive the shortcut. The first is social categorization, which psychologists describe as the tendency to sort people into in-groups and out-groups. The second is out-group homogeneity, the perception that members of other groups are more alike than they actually are. Together these create the conditions for stereotyping, which supplies the content that prejudice then evaluates.
Stereotypes are the cognitive component, prejudice is the affective or evaluative component, and discrimination is the behavioral component. The three often travel together but are separable. A person can hold a prejudiced attitude and never act on it, or can discriminate without conscious hostility. This separation is why measuring prejudice requires attitudes and behavior to be assessed independently.
Neuroscience adds a layer. Brain imaging work has shown that social categories are processed rapidly and can activate learned associations before deliberate reasoning engages. That does not excuse prejudice. It explains why implicit associations can persist even in people who consciously reject group-based bias, and why unlearning them requires more than good intentions.
How much prejudice is measurable and how it is quantified
Quantifying prejudice is difficult because self-report is unreliable. People underreport attitudes they know are socially unacceptable, a pattern researchers call social desirability bias. To work around this, scientists use indirect measures such as the Implicit Association Test, which records response latencies when pairing social groups with positive or negative words. Faster pairings that match cultural stereotypes indicate stronger automatic associations.
Indirect measures have real limits. Test-retest reliability on the Implicit Association Test is modest, and scores fluctuate across occasions. A single test result cannot diagnose an individual as prejudiced. Researchers treat it as a population-level instrument and a rough indicator, not a clinical tool. This caution is one of the most frequently misunderstood points in the field.
Explicit measures capture different information. Surveys that ask directly about group attitudes, hiring preferences, or comfort with neighbors produce stable group averages. Large national survey programs have tracked such attitudes for decades, letting researchers observe direction and pace of change. Exact percentages shift year to year, so only the general pattern is safe to state: overtly expressed prejudice on many measures has declined in the United States since the mid-twentieth century, while indirect measures show smaller and more variable change.
The gap between explicit and implicit scores is itself informative. It suggests that some prejudice has moved from openly stated to privately held or automatic. Field experiments, where matched résumés differing only in a name are sent to employers, provide behavioral evidence. These audit studies consistently find callback differences by perceived group membership, with magnitudes that vary by study and labor market.
Edge cases that complicate the definition
The definition hinges on two elements: the judgment is uninformed, and it is based on category membership. Edge cases arise when one element is present without the other. A well-informed generalization about a group is still a generalization, but calling it prejudice requires that it be applied to individuals without individual evidence.
Context matters enormously. Statistical discrimination describes situations where a person uses group averages because individual information is costly to obtain. Economists distinguish this from taste-based prejudice, which reflects dislike of a group regardless of information. The distinction matters for remedies: better information addresses statistical discrimination, while contact and norm change address taste-based prejudice.
Power and salience change the stakes. Sociologists note that prejudice operates differently when directed from a dominant group toward a marginalized one versus the reverse. Group-based attitudes exist in every direction, but their consequences differ with social position, institutional power, and history. A definition that ignores this asymmetry explains less than one that includes it.
Intersectionality is a further edge case. A person belongs to multiple categories at once, and assumptions about a Black woman with a disability are not simply the sum of assumptions about each category. Prejudice targeted at combined identities can take distinct forms that single-category measures miss. This is a caution against treating group attitudes as additive.
One counterintuitive finding deserves mention. Contact between groups reduces prejudice more reliably under specific conditions than simple exposure does. The classic conditions include equal status, shared goals, cooperation, and institutional support. Mere proximity without these conditions can intensify hostility. Diversity training that ignores these conditions tends to produce weak or short-lived effects.
Common misconceptions
The most common misconception is that prejudice equals racism or sexism. Prejudice is the umbrella term for prejudgment across any category of characteristics, including age, religion, weight, disability, nationality, and language. Racism and sexism are specific forms that also involve systemic power. Using the terms interchangeably blurs distinctions that matter for measurement and remedy.
A second misconception is that prejudice requires conscious awareness. The distinction between explicit and implicit prejudice shows that automatic associations can operate outside deliberate control. This does not remove responsibility, but it changes what interventions look like. Willpower alone is a poor tool against automatic associations; structured habit change and environmental design work better.
A third misconception treats prejudice as fixed. Attitudes are not immutable. Meta-analytic evidence shows that well-designed contact interventions and certain educational approaches produce measurable reductions, though effect sizes are modest and durability varies. Setting expectations too high sets interventions up for false claims of failure.
RED FLAGS when evaluating claims about prejudice:
A single test score presented as proof that a specific person is prejudiced. Headline percentages with no sample size, date, or population described. Claims that prejudice is extinct, or that it explains every group disparity. Interventions billed as eliminating bias with no outcome data. Definitions that collapse prejudice, stereotyping, and discrimination into one word. Statistics older than a decade presented as current.
Distinguishing prejudice from related terms
The vocabulary around group attitudes is dense, and each term names a different component. Getting the boundaries right is the difference between precise communication and muddy argument.
| Term | What it names | Key mechanism | Example of form |
|---|---|---|---|
| Prejudice | Evaluative attitude toward a group or its members | Prejudgment before evidence | Assuming an older applicant is slow with technology |
| Stereotype | Generalized belief about group traits | Cognitive categorization | Believing a nationality is uniformly frugal |
| Discrimination | Unequal treatment of individuals | Behavior driven by category | Rejecting a résumé because of a name |
| Implicit bias | Automatic association outside awareness | Fast, learned memory links | Faster pairing of a group with negative words |
| Explicit bias | Consciously endorsed attitude | Deliberate evaluation | Stating that a group is less capable |
| Microaggression | Brief, everyday slight or insult | Ambiguous, often unintentional signaling | Repeatedly asking where someone "really" is from |
The table shows why the terms cannot be swapped freely. Prejudice is the attitude; discrimination is the act. Implicit bias describes the automatic pathway, while explicit bias describes the endorsed one. Microaggressions sit at the behavioral edge, where intent and impact frequently diverge.
A related distinction separates prejudice from institutional or structural disadvantage. Structures can produce unequal outcomes even when no individual holds a hostile attitude. A hiring pipeline that relies solely on personal networks can disadvantage outsiders without anyone forming a prejudiced judgment. Structural analysis and attitude measurement answer different questions and should not be conflated.
How prejudice develops and changes over time
Prejudice is not present at birth in any adult form. It emerges through a developmental sequence, shaped by family, peers, media, and direct experience. The timeline below traces the general course documented in developmental and social psychology research. Ages are approximate ranges, not sharp thresholds.
| Stage | Approximate age or window | What develops | Research note |
|---|---|---|---|
| Category detection | Infancy to age 2 | Noticing perceptual differences such as skin tone and language | Preference for familiar faces appears early |
| Group labeling | Ages 3 to 5 | Naming social groups and attaching simple evaluative labels | Labels can carry valence before explanation |
| In-group preference | Ages 4 to 7 | Favoring familiar groups over unfamiliar ones | Not identical to out-group hostility |
| Stereotype absorption | Ages 7 to 12 | Learning cultural stereotypes from media and adults | Children can report stereotypes without endorsing them |
| Attitude consolidation | Adolescence | Attitudes stabilize as identity and peer norms sharpen | Peer norms exert strong pressure |
| Adult stability with plasticity | Adulthood | Attitudes become stable but remain modifiable | Change is possible but slower |
The sequence matters for intervention design. Programs aimed at young children emphasize cross-group play and equal-status cooperation. Programs for adolescents address peer norms and media literacy. Adult interventions lean on structured contact, perspective-taking, and accountability systems.
Two threshold points deserve emphasis. First, in-group preference in young children is not the same as out-group hatred, and treating them as identical overstates the developmental picture. Second, adolescence is a window where peer norms can entrench or loosen attitudes, which is why school climate interventions target that age band. The general pattern is well documented; precise ages vary across studies and cultures.
Practical actions with concrete thresholds
Prejudice cannot be eliminated by a single insight, but its expression can be reduced through structured practice and environmental design. The steps below are general education, not individualized instructions, and thresholds are conventional rules of thumb rather than clinical cutoffs.
Audit decisions at scale. Review hiring, admissions, and evaluation outcomes by group annually. Set a review trigger when callback or selection gaps exceed roughly 10 percentage points across at least two consecutive cycles, then examine the process steps rather than the people. Use structured scoring. Replace unstructured interviews with fixed rubrics scored against defined criteria. Structured interviews show better predictive validity and reduce the influence of first impressions, which form within seconds of meeting someone. Design structured contact. Aim for sustained cooperative contact with equal status and shared goals, repeated across at least several sessions rather than a single event. One-off diversity trainings produce weak, short-lived effects compared with ongoing programs. Slow high-stakes decisions. Introduce a short delay and a second reviewer for consequential judgments such as hiring or disciplinary action. Brief delays and independent ratings reduce reliance on quick categorization. Make category-blind first passes. In evaluation of written work, hide names and identifying details during the first review round. This removes the category cue before the assessor has formed a judgment. Track norms, not just attitudes. Measure what people believe their peers approve of, because perceived norms predict behavior strongly. If repeated surveys of organizational climate show perceived tolerance of biased remarks, address the norm directly through leadership statements and enforcement.
Reconciling thresholds: several rules of thumb circulate for when a group gap merits review. Some organizations use a 5-point trigger, others 10 to 20 points. These are administrative conventions, not scientific constants. The defensible rule is to review any persistent gap that survives statistical noise and process controls, and to look at the mechanism before blaming individuals. A gap is a signal to investigate, not a verdict.
Practical caveats matter. Mandatory training framed as accusation can provoke backlash and worsen attitudes. Voluntary, skills-based programs with clear organizational support perform better. Measurement should focus on outcomes such as selection rates and evaluation consistency rather than on test scores, which are noisy at the individual level.
Limits of the evidence and definitional boundaries
The science of prejudice has real constraints. Self-report is biased, implicit measures are imperfect, and field experiments capture behavior in specific settings that may not generalize. Effect sizes for interventions are often modest, and long-term durability is understudied. Anyone citing a single figure as definitive is overreaching.
Definitional boundaries also blur at the edges. The core answer to the question is clear: prejudice is the term for uninformed opinion or assumptions toward an individual or group based on categories of characteristics. But prejudice shades into stereotyping, discrimination, and structural disadvantage along gradients rather than sharp lines. Free speech, legal, and cultural contexts shape how the term is applied. Precision means naming which component is being discussed rather than letting the umbrella term do all the work.
This article is general educational information about the psychology of group attitudes. It is not a diagnosis, treatment plan, legal opinion, or advice about any specific person or situation. Anyone facing discrimination, harassment, or a mental health concern should consult a qualified local professional or appropriate authority, and anyone in immediate danger should contact local emergency services.