29.07.04 · psychology / social-psychology

Prejudice and discrimination: implicit bias, stereotype threat, contact hypothesis

stub3 tiersLean: nonepending prereqs

Anchor (Master): Allport, G. W. — The Nature of Prejudice (1954)

Intuition Beginner

Prejudice is a negative attitude toward a group; discrimination is the negative behavior that flows from it. The two are distinct. Someone can hold a prejudice and never act on it, and discrimination can occur without any conscious prejudice behind it. Social psychologists learned long ago that people can hold prejudices they do not consciously endorse. These are called implicit biases.

The Implicit Association Test (IAT), developed by Anthony Greenwald and Mahzarin Banaji, measures how quickly people associate categories — like Black or White faces with Good or Bad words. Faster pairings reveal automatic preferences. Most people show an implicit preference, even those who explicitly reject prejudice on a written survey.

Claude Steele and Joshua Aronson discovered stereotype threat: when people fear confirming a negative stereotype about their group (like "girls are bad at math"), their performance suffers. The cause is not ability but anxiety and cognitive load — the mind is busy fighting the stereotype instead of doing the task.

The contact hypothesis, proposed by Gordon Allport in 1954, offers a route forward. Prejudice can be reduced when groups interact under the right conditions: equal status, common goals, cooperation, and the support of authorities. Done well, contact turns abstract enemies into real people — the foundation of every school desegregation and peace-building program.

Visual Beginner

Figure: The four core constructs of this unit. The IAT measures the speed of automatic associations, yielding a D score that captures implicit preference. Stereotype threat shows that the salience of a negative group stereotype drains the working memory and raises anxiety needed for a difficult test, erasing the gap when the test is framed as non-diagnostic. Allport's contact hypothesis specifies the four structural conditions under which intergroup contact reduces prejudice. The anatomy triangle fixes the vocabulary: stereotypes are beliefs, prejudice is affect, discrimination is behavior — and the three can come apart.

Worked example Beginner

The Steele-Aronson stereotype threat experiment

In 1995, Claude Steele and Joshua Aronson gave African American and White Stanford students a difficult test built from GRE verbal items. The questions were identical for everyone. The only thing the experimenters varied was what the test was called.

In one condition, the test was described as a diagnostic measure of genuine intellectual ability — a real test of how smart you are. In the other condition, the very same questions were framed as a harmless laboratory puzzle that had nothing to do with ability.

The result turned on race. When the test was framed as diagnostic of ability, African American students performed worse than White students who had matched SAT scores. When the same test was framed as a mere puzzle, the racial gap vanished — the two groups performed identically.

The students were not less able. The stereotype ("African Americans are less intelligent") had become salient in the diagnostic condition, and the effort to push it away consumed the working memory the test demanded. This is stereotype threat. The same mechanism has since been documented in women taking hard math tests, in older adults on memory tests, and in white men told they are being compared with Asian peers.

Check your understanding Beginner

Exercise (easy, multiple choice).

Which of the following best distinguishes a stereotype from prejudice from discrimination?

A. They are three words for the same thing — a negative feeling about a group. B. A stereotype is a belief, prejudice is a feeling, and discrimination is a behavior. C. A stereotype is a behavior, prejudice is a belief, and discrimination is a feeling. D. Stereotypes are conscious, while prejudice and discrimination are always unconscious.

Answer

Option B. The standard "ABC" split is cognition, affect, and behavior. A stereotype is the cognitive component — a belief about a group ("women are emotional"). Prejudice is the affective component — a negative feeling or attitude toward the group. Discrimination is the behavioral component — an action that harms or excludes group members. The three are correlated but dissociable: a person can hold a stereotype without feeling prejudice, and can discriminate without any conscious stereotype at all (as in implicit bias).

Exercise (easy, multiple choice).

According to Allport's contact hypothesis, which set of conditions is most likely to reduce prejudice between two groups?

A. The groups meet once, briefly, in a competitive setting with clear winners and losers. B. The groups interact as equals, on a shared goal, cooperating, with the backing of an authority. C. One group is put in charge of the other so that order is established before contact. D. The groups are kept apart and learn about each other only through news and rumors.

Hint

Allport listed four conditions. Each one removes a specific obstacle to seeing the other group as fellow humans rather than as a category.

Answer

Option B. Allport's four conditions are equal status, common goals, intergroup cooperation, and the support of authorities, law, or custom. Option B names all four. Competition (A) and hierarchy (C) typically inflame prejudice rather than reduce it, and mere separation (D) leaves stereotypes uncorrected. When the four conditions hold, contact generalizes: a positive encounter with one out-group member tends to soften the attitude toward the whole group.

Formal definition Intermediate

The vocabulary of prejudice and discrimination is standardised across the anchor texts ^{[source pending]}. The tripartite model fixes the core terms, and the mechanisms catalogued below are empirically dissociable rather than mere labels.

The ABC components

Prejudice has three dissociable components, conventionally labelled the ABCs ^{[source pending]}. The affective component is prejudice proper — a negative (or, less often, positive) attitude or feeling toward a group. The cognitive component is the stereotype — a generalized belief about the characteristics of group members. The behavioral component is discrimination — unequal treatment of group members in hiring, housing, policing, education, or everyday interaction. The three correlate but can come apart: a person can hold the stereotype without the affect, feel the affect without acting on it, or discriminate (as under institutional pressure) without any personal stereotype at all.

The Implicit Association Test (Greenwald and Banaji, 1995)

Anthony Greenwald and Mahzarin Banaji introduced the construct of implicit social cognition — attitudes and stereotypes that operate outside conscious awareness and control ^{[source pending]}. The Implicit Association Test (IAT) measures the strength of these associations through response latency. In the race IAT, participants sort Black and White faces and pleasant and unpleasant words. In the congruent block (e.g., White + pleasant on one key, Black + unpleasant on the other) most respondents are faster than in the incongruent block (the pairings are reversed). The IAT effect is the response-time difference between blocks, conventionally scored as a D measure analogous to a d-prime. The implicit–explicit correlation is typically modest (around r = 0.24 for the race IAT), which is the empirical basis for treating implicit and explicit attitude as distinct constructs.

Stereotype threat (Steele and Aronson, 1995)

Stereotype threat is the threat that one's behavior will confirm a negative stereotype about one's group, and the resulting pressure that impairs performance on tasks where the stereotype is relevant ^{[source pending]}. The mechanism is not a deficit in ability. It is a disruption of performance: the salience of the stereotype raises anxiety, loads working memory, and triggers self-monitoring ("am I confirming it?") that competes with the task. Steele and Aronson showed that African American students performed worse than equally able White students when a difficult verbal test was framed as diagnostic of ability, but performed identically when the same test was framed as a non-diagnostic puzzle. The effect generalizes across groups: women and mathematics ^{[source pending]}, older adults and memory, white men and athletic performance against a supposed Asian comparison group, and low-socioeconomic-status students and intelligence tests.

Disidentification (Major and Schmader) is the long-term protective response: a student whose group is chronically threatened in a domain (e.g., mathematics) may come to de-identify with that domain, ceasing to care about it as a basis of self-worth. Threat thus explains both underperformance and the slow erosion of domain identification in stigmatized groups.

Stereotype lift (Walton and Cohen)

Stereotype lift is the mirror effect: a group that is not the target of a negative stereotype can perform better when a downward comparison to a stereotyped out-group is made salient ^{[source pending]}. Walton and Cohen estimated the effect at roughly one quarter of a standard deviation. Where stereotype threat depresses the threatened group, stereotype lift can inflate the non-threatened comparison group — meaning the observed gap between groups is the aggregate of two forces, not one.

The contact hypothesis (Allport, 1954)

Gordon Allport proposed that intergroup contact, under the right conditions, reduces prejudice ^{[source pending]}. Allport's four conditions are:

Condition	What it requires	What it removes
Equal status	Participants interact as peers in the situation	Hierarchical framing that confirms the stereotype
Common goals	Both groups seek the same outcome	Competition that pits groups against each other
Intergroup cooperation	The groups work together, not in parallel	The mere co-presence that allows avoidance
Support of authorities	Law, custom, or leadership endorse the contact	Anxiety that contact is illicit or risky

Thomas Pettigrew and Linda Tropp's meta-analysis of 515 studies found an average effect of r = −0.21 between contact and prejudice, with the effect holding — though more weakly — even when Allport's conditions are not fully met ^{[source pending]}. Contact works in part by reducing intergroup anxiety, increasing empathy and perspective-taking, and generating out-group friendships that generalize to the category.

Modern forms of prejudice

Old-fashioned, overt ("redneck") racism has declined sharply in surveyed countries since the mid-twentieth century, but prejudice has not disappeared — it has changed form ^{[source pending]}. Several modern variants are distinguished:

Symbolic (modern) racism — the fusion of antiblack affect with non-race values such as individualism, self-reliance, and the work ethic. The symbolic racist denies prejudice but opposes race-targeted policy on ostensibly principled grounds.
Aversive racism — sincere egalitarian self-images coexisting with unconscious negative associations. The aversive racist discriminates not openly but in situations where the behavior can be justified on non-racial grounds.
Ambivalent sexism (Glick and Fiske) — a distinction between hostile sexism (overtly antagonistic attitudes toward women) and benevolent sexism (subjectively positive but patronizing attitudes that frame women as pure, delicate, and in need of protection). Both predict inequality; benevolent sexism is the harder one to recognize because it feels kind ^{[source pending]}.

Finally, two ideologies about how to manage diversity shape both policy and individual behavior. The colorblind ideology holds that group differences should be ignored ("I don't see race"). The multicultural ideology holds that group differences should be recognized and valued. Empirically, colorblind approaches can backfire — they make stereotype-relevant differences harder to discuss and can mask inequity — whereas multicultural framing tends to reduce threat and improve out-group attitudes, especially for members of minority groups.

Key experiment Intermediate

The Steele-Aronson paradigm in detail

The 1995 Steele-Aronson study is the canonical demonstration of stereotype threat because it isolates the mechanism with unusual clarity ^{[source pending]}. African American and White Stanford undergraduates, matched on SAT scores, were given thirty difficult items drawn from the GRE Advanced Psychology exam (items chosen to be answerable by bright non-majors). The manipulation was a one-paragraph difference in the test's cover story.

Condition	Framing	Prediction
Diagnostic	"a genuine test of your verbal abilities"	Stereotype about Black intellectual ability becomes salient
Non-diagnostic	"a laboratory tool for studying problem solving"	No stereotype relevance
Non-diagnostic + challenge	same as above but the race of the participant is made salient	Tests whether mere salience is enough

The result was specific. In the diagnostic condition, African American participants performed worse than White participants of equal ability. In the non-diagnostic condition, the racial gap disappeared. The third condition showed that merely making race salient — without calling the test diagnostic — was enough to depress performance. Two follow-up measures located the mechanism: African American participants in the diagnostic condition completed fewer items after a stalling point (suggesting effort withdrawal), and a word-completion task showed that race- and ability-related stereotype words were more accessible in their minds. Threat was not a mood effect; it was a cognitive-load effect driven by the activation and suppression of the stereotype.

The Implicit Association Test: methodology and the D score

The IAT's design rests on a simple premise: people are faster to respond when conceptually associated categories share a response key ^{[source pending]}. A standard five-block procedure first practices single-category sorting (faces, then words), then practices combined sorting in one pairing, then reverses the pairings. The crucial comparison is response latency in the two combined blocks. The D score (Greenwald, Nosek, and Banaji, 2003) normalizes the latency difference by the participant's own response variability, yielding a measure roughly analogous to Cohen's d.

Pairing	Typical finding
White + pleasant / Black + unpleasant	Faster (congruent with cultural associations)
Black + pleasant / White + unpleasant	Slower
D = difference, normalized	~40% of respondents show a "moderate" or "strong" pro-White preference

The interpretation of the D score is contested (see the master tier): it correlates only modestly with explicit attitudes and predicts behavior weakly at the individual level, yet it has been administered tens of millions of times and is the most widely used implicit measure in social psychology.

The Pettigrew-Tropp meta-analysis: does contact work?

Allport's contact hypothesis could have been wrong — competition between groups might inflame rather than reduce prejudice, as Sherif's Robbers Cave study initially showed. The decisive empirical test is Pettigrew and Tropp's (2006) meta-analysis of 515 independent studies, covering over 250,000 participants across 38 nations ^{[source pending]}.

The central finding is an average correlation of r = −0.21 between contact and prejudice — a small but consistent negative relationship, meaning more contact is associated with less prejudice. Three refinements matter. First, the effect is stronger when Allport's four conditions are present, but it remains significant even when they are absent, suggesting contact has a baseline benefit beyond the optimal conditions. Second, the effect generalizes from the specific out-group member contacted to the out-group category as a whole. Third, longitudinal and experimental designs reduce (though do not eliminate) the reverse-causation worry — that low-prejudice people seek contact, rather than contact lowering prejudice. The meta-analysis established contact as one of the most robust prejudice-reduction interventions in social psychology.

Exercises Intermediate

Exercise 1 (medium, multiple choice).

A company gives its hiring managers an Implicit Association Test and finds that most show a moderate implicit preference against older job candidates. The managers all score as strongly non-ageist on an explicit survey. Which conclusion is best supported by the IAT literature?

A. The implicit scores are invalid because they disagree with the explicit survey. B. The managers hold explicit egalitarian attitudes but also automatic negative associations that they do not consciously endorse. C. The IAT proves the managers will discriminate against older candidates if given the chance. D. The disagreement means at least one of the two measures is random noise.

Hint

The implicit–explicit correlation is typically modest. What does a divergence between the two measures tell you about the structure of the attitudes, not about the truth of either measure?

Answer

Option B. The IAT was designed to measure associations that self-report cannot reach, and the modest implicit–explicit correlation is exactly what one expects when people hold explicit egalitarian values alongside automatic associations picked up from a culture that stereotypes older workers. Option A treats disagreement as invalidity, but the two measures are meant to capture different constructs. Option C overclaims: at the individual level, the IAT predicts behavior only weakly (see the Oswald critique in the master tier), so a score does not prove a given manager will discriminate. Option D treats the disagreement as noise when it is the phenomenon the construct was built to track.

Exercise 2 (medium, short answer).

A high school with a persistent racial achievement gap redesigns its orientation so that incoming students from the stereotyped group complete a brief values-affirmation writing exercise. Explain, in 3–4 sentences, why this intervention is expected to narrow the gap, naming the construct the exercise targets.

Hint

The exercise is not teaching content. It asks students to write about values that matter to them. Which psychological threat does reaffirming one's core values buffer against, and through what resource?

Answer

The intervention targets stereotype threat. By having students write about personally important values, the self-affirmation exercise bolsters their overall sense of self-integrity, so that a difficult test or a stereotype-relevant situation no longer feels like a referendum on their group or their worth. With the threat buffered, the working memory and attention that threat would have consumed are returned to the academic task. Cohen, Garcia, Apel, and Master (2006) found that a single 15-minute exercise raised the African American students' GPA by roughly 0.3 points over the term, with effects still visible years later — consistent with a self-reinforcing cycle in which early success reduces later threat.

Exercise 3 (medium, multiple choice).

A benevolent-sexist statement such as "Women are too precious to be exposed to the harshness of politics" is best described as:

A. Hostile sexism, because it excludes women from power. B. Aversive racism, because it masks a negative attitude with a positive surface. C. Benevolent sexism, because it is subjectively positive yet restricts women to a pedestal. D. Old-fashioned prejudice, because it invokes an explicit gender hierarchy.

Answer

Option C. Glick and Fiske's benevolent sexism consists of subjectively favorable attitudes — protection, idealization, affection — that nonetheless confine women to restricted, dependency-framed roles and rationalize inequality as care. The statement restricts women's access to power while presenting itself as a compliment. Hostile sexism (A) is openly antagonistic, which this is not. Aversive racism (B) concerns race, not gender. Old-fashioned prejudice (D) invokes an explicit and openly endorsed hierarchy; benevolent sexism is insidious precisely because the speaker experiences it as virtue.

Exercise 4 (hard, application).

Two neighboring ethnic groups share a river whose water both depend on. Tensions are rising. Using Allport's contact hypothesis and the concept of a superordinate goal (Sherif), design a contact-based intervention and specify which of Allport's four conditions each design element is meant to satisfy.

Hint

A superordinate goal is one that neither group can achieve alone and that both need. Build the intervention around such a goal and then check each of Allport's conditions off explicitly.

Answer

Convene a joint water-management committee with equal numbers of members from each group, charged by a regional authority with maintaining the river's flow and quality. The committee works on a single shared problem — keeping the river clean and adequately supplied — that neither group can solve on its own: this is the superordinate goal, which supplies Allport's common goals condition. Seat members as equals with rotating chairs and identical voting weight to satisfy equal status, and require that they work side by side on field tasks (sampling, repairs) rather than merely meeting, satisfying intergroup cooperation. Finally, the committee is established and backed by the regional authority's mandate and funding, satisfying support of authorities. Sherif's Robbers Cave logic predicts that a superordinate goal that genuinely requires both groups will, over time, dissolve the prior competitive framing, and Allport's conditions ensure the contact occurs in the structure most likely to reduce prejudice rather than inflame it.

Advanced results Master

The IAT predictive-validity debate

The most consequential methodological controversy in implicit-bias research concerns how well the IAT predicts behavior. Frederick Oswald and colleagues (2013) conducted a meta-analysis of race-based IAT studies and concluded that the IAT predicts discrimination only weakly — a mean r of about 0.14, insufficient to support strong claims about an individual's likelihood of discriminating ^{[source pending]}. Greenwald, Banaji, and colleagues rebutted that the analysis used a constrained outcome criterion and that implicit measures predict aggregate disparities more reliably than individual acts ^{[source pending]}. Helen Meagher and Ugandan Brunton's (2019) reanalysis and related large-scale work found that the IAT does not robustly predict race-based behavioral disparities at the level its strongest claims require ^{[source pending]}.

Three points survive the debate. First, weak individual-level prediction is not the same as irrelevance: a measure can fail to identify which individual will discriminate while still indexing a population-level construct that helps explain group disparities. Second, test-retest reliability of the IAT is moderate (typically r in the 0.5–0.6 range), which mechanically caps any correlation with downstream behavior. Third, the IAT is malleable: exposure to counter-stereotypic exemplars (for instance, admired Black figures and disliked White figures) temporarily reduces the measured preference, demonstrating that the score reflects something responsive to recent context rather than a fixed trait ^{[source pending]}. Whether the measured association represents a personal attitude or the cultural environment the respondent has internalized remains open: Olson and Fazio argued that the IAT partly reflects extra-personal associations absorbed from the culture rather than personally endorsed evaluations ^{[source pending]}.

Neuroimaging adds a convergent but contested data stream. Elizabeth Phelps and colleagues found that amygdala activation in response to out-group faces correlates with IAT scores, suggesting a shared implicit-evaluation substrate ^{[source pending]}. The amygdala responds to novelty and ambiguity as well as to threat, however, so the activation is open to multiple readings — cultural familiarity, not personal animus, may drive part of the signal.

Stereotype threat: ecological-validity questions

Steele and Aronson's laboratory finding has been replicated many times, but its real-world magnitude is disputed. Roland Fryer (2010) and analyses of PISA data found smaller or inconsistent threat effects in large administrative datasets than in controlled laboratory studies ^{[source pending]}. Paul Sackett, Chaitra Hardison, and Mary Cullen (2012) argued that the literature overstates the effect because many studies report the gap after threat is induced without comparing it to the full no-threat baseline, and that threat accounts for only a small fraction of observed standardized-test gaps ^{[source pending]}. Colleen Ganley and colleagues (2013) found no stereotype-threat effect on large-scale mathematics assessments administered under realistic conditions ^{[source pending]}.

On the other side, Schmader, Johns, and Forbes's (2008) meta-analytic review found a reliable negative effect of stereotype threat on performance across studies, and proposed an integrated mechanistic account ^{[source pending]}. The resolution is that stereotype threat is a real and replicable laboratory phenomenon whose contribution to field achievement gaps is bounded — it is one of several forces, not the single cause. The mechanistic account matters here: threat operates through physiological stress (elevated cortisol), working-memory depletion, mind-wandering toward the threatening stereotype, and the paradoxical cost of suppression — the active attempt not to think about the stereotype consumes the same cognitive resources the task demands, a variant of the ironic-process effect. These mechanisms explain why the effect is strongest on difficult, working-memory-intensive tasks and weaker or absent on easy ones.

Interventions: self-affirmation, mindset, and belonging

If stereotype threat depresses performance through threats to self-integrity, then protecting self-integrity should restore performance. Geoffrey Cohen, Jennifer Garcia, Nancy Apel, and Allison Master (2006) had African American middle-school students complete a 15-minute values-affirmation exercise — writing about two or three values (e.g., relationships, creativity) that mattered most to them — early in the term ^{[source pending]}. The exercise raised the African American students' GPA by roughly 0.3 points and closed roughly 40% of the racial achievement gap, with effects still visible two years later. The intervention is striking because it teaches no academic content; it changes the psychological frame in which the content is encountered.

Several related interventions target the same threat system through different routes:

Growth mindset training reframes ability as malleable, which undercuts the implication that a stereotype-relevant failure diagnoses a fixed group trait. Mindset interventions disproportionately benefit stereotyped groups, especially when combined with a supportive teacher message.
Sense-of-belonging interventions (Walton and Cohen) reframe early social adversity as normal and transient rather than as evidence that "people like me don't belong here," and have produced long-lasting gains for first-generation and minority students.
Synergistic effects: combining values-affirmation, growth-mindset, and belonging interventions tends to outperform any single intervention, because each defuses a distinct facet of the threat.

The methodological caution is that these interventions show large effects in some contexts and null effects in others, and recent large-scale replications have produced mixed results. The construct — that psychological threat is a tractable component of group-level underperformance — is the durable contribution; the size and durability of any specific exercise remain under active investigation.

Realistic group conflict, social dominance, and system justification

Three structural theories situate prejudice in the material and ideological organization of societies rather than in individual attitudes.

Realistic group conflict theory (revived by Lawrence Bobo) holds that prejudice arises from real competition over scarce material resources — jobs, housing, political power ^{[source pending]}. When groups compete, prejudice is a rationalization of self-interest; when interests align, prejudice recedes. The theory explains why prejudice rises during economic contraction and why it tracks the competition structure rather than the contact structure alone.

Social dominance theory (Jim Sidanius and Felicia Pratto) posits that societies produce and maintain group-based hierarchy through three routes: aggregated individual discrimination, aggregated institutional discrimination, and behavioral asymmetry (rights and duties distributed unequally across groups) ^{[source pending]}. A core construct is social dominance orientation (SDO) — an individual difference in preference for group-based hierarchy — which predicts a wide range of policy attitudes independent of specific prejudice.

System justification theory (John Jost and Mahzarin Banaji) identifies a counterintuitive motive: people (including members of disadvantaged groups) often engage in system justification, accepting and rationalizing the status quo even when it harms them, because certainty and perceived legitimacy reduce uncertainty and threat ^{[source pending]}. The theory explains why inequality can be stable even without coercion: the belief that the system is fair is itself a stabilizing force, and it is strongest exactly where one might expect resistance.

Intersectionality

Kimberlé Crenshaw's intersectionality framework argues that race, gender, class, sexuality, and other categories do not act additively but interact, producing experiences of advantage and disadvantage that cannot be predicted from any single axis ^{[source pending]}. The experience of a Black woman, for instance, is not "the prejudice against Black men plus the prejudice against white women"; it is a distinct configuration shaped by the simultaneous operation of both categories. The framework began in legal scholarship — Crenshaw showed how anti-discrimination law, structured around single-axis claims, failed to capture the discrimination faced by Black women — and has become a general analytic principle in social psychology. Its methodological implication is that research designs that average over a category (e.g., pooling all women, or all Black participants) can mask effects that appear only at the intersection.

Intergroup conflict resolution: Robbers Cave, jigsaw, and recategorization

Three interventions translate the contact hypothesis into practice.

The Robbers Cave experiment (Muzafer Sherif and colleagues, 1954) brought two groups of boys to a summer camp, engineered intergroup friction through competition, and then reversed it through superordinate goals — problems (a stuck truck, a broken water supply) that required both groups to cooperate ^{[source pending]}. Sherif showed that prejudice could be produced and then reduced under controlled conditions, and that the operative variable was the structure of goals, not the personality of the boys. The study is the empirical anchor of realistic group conflict theory and of the superordinate-goal principle in the contact literature.

The jigsaw classroom (Elliot Aronson and colleagues, 1971), designed in the aftermath of school desegregation in Austin, Texas, structures lessons so that each student in a mixed group becomes the expert on one piece of the material and must teach it to the others to complete the assignment ^{[source pending]}. The design forces interdependence: no single student can succeed alone, and the out-group member becomes an indispensable resource. Evaluations show reduced prejudice, increased empathy, and improved academic outcomes for minority students, though the effect depends on faithful implementation.

The Common Ingroup Identity Model (Samuel Gaertner and John Dovidio) generalizes the principle: prejudice is reduced when former out-group members are recategorized as members of a shared, more inclusive in-group ^{[source pending]}. The move from "us and them" to a superordinate "we" does not require erasing subgroup identities; a dual-identity model (we are all students at this school, and we each retain our ethnic group) works as well as or better than a single common identity, because it preserves the valued subgroup while reaping the benefits of the inclusive category. The model integrates Allport's contact hypothesis with social identity theory and is among the most influential bridges between the two traditions.

Connections Master

Social psychology: groups, prejudice 29.07.01 is the direct prerequisite and parent unit. Social identity theory, the minimal group paradigm, and the fundamental attribution error introduced there are specialised here into the mechanisms of prejudice and discrimination. The parent unit supplies the framing (groups, influence); this unit supplies the application (prejudice, its reduction).
Group dynamics 29.07.03 pending connects through deindividuation, conformity, and the bystander effect — all of which shape how discrimination is produced and tolerated in groups. Sherif's Robbers Cave, catalogued in this unit, is also a foundational group-dynamics study, and the jigsaw classroom draws directly on the contact and cooperation principles developed in the group-dynamics literature.
Attitudes and persuasion 29.07.02 pending connects through the implicit–explicit attitude distinction. The elaboration likelihood model explains how persuasive messages can change explicit attitudes while leaving implicit associations intact, and vice versa — which is why implicit-bias retraining interventions have proven hard to durably shift.
Developmental psychology and moral development 29.06.03 pending (pending) connects through the acquisition of stereotypes and prejudice in childhood. Children show race-based preferences well before they can articulate them, and the trajectory of moral reasoning (Kohlberg, Gilligan) intersects with the development of tolerance and empathy that the contact hypothesis relies on.
Cognitive development 29.06.02 pending (pending) supplies the categorization processes that make stereotyping cognitively cheap and hard to suppress. The same essentialist reasoning that helps children learn natural kinds also underwrites the reification of social categories.
Neuroscience [29.02.NN] (pending) provides the amygdala and prefrontal-circuit substrate for implicit evaluation and its regulation, and the working-memory and stress-hormone systems through which stereotype threat operates.
Cross-cultural psychology 29.12.01 connects through the variability of prejudice forms across cultures, the WEIRD critique of IAT and threat research, and the cultural shaping of whether race, ethnicity, caste, or religion is the dominant intergroup axis.
Decision-making and judgment 29.05.02 pending connects through the heuristic and biased bases of stereotyping (representativeness, availability) and the suppression-as-ironic-process account of stereotype threat.

Historical & philosophical context Master

Gordon Allport's The Nature of Prejudice (1954) is the founding text of the modern study of intergroup attitudes, and the contact hypothesis it contains is among the most consequential ideas in twentieth-century social science. Allport wrote in the immediate aftermath of the Second World War and the Holocaust, and in the early years of the American civil-rights movement, at a moment when the question of whether prejudice could be reduced was not academic but political. The Supreme Court's Brown v. Board of Education decision, handed down the same year, rested in part on social-science evidence (including Kenneth and Mamie Clark's doll studies) that segregation harmed Black children's self-concept. Allport's hypothesis — that contact, under the right conditions, reduces prejudice — supplied the psychological theory behind the policy of school desegregation, and its empirical track record (Pettigrew and Tropp) has sustained that link for seventy years.

The notion that people hold attitudes they cannot or will not report has a longer prehistory. The psychoanalytic tradition (Freud, Adorno's Authoritarian Personality) treated prejudice as a projection of unconscious conflict. The cognitive revolution reframed the unconscious not as a reservoir of repressed impulse but as a system of automatic associations operating outside awareness and control. Greenwald and Banaji's 1995 paper gave this reframing its construct — implicit social cognition — and the IAT gave it a measurement instrument. The shift was consequential: it moved implicit bias from speculation into a measurable quantity, which is why the IAT became both widely adopted and intensely contested. The measurement made the claim testable, and the tests produced the predictive-validity debate catalogued above.

Claude Steele's development of stereotype threat theory in the early 1990s answered a different puzzle. Standard explanations of the Black–White test-score gap attributed it to either ability differences (a claim lacking scientific support and freighted with prejudice) or to socioeconomic disadvantage (real but incomplete, since the gap persisted within socioeconomic levels). Steele and Aronson proposed a third factor: the situation of taking a test on which one's group is stereotyped. The proposal reframed a portion of the achievement gap as a situational pressure that could be experimentally turned on and off — and, crucially, intervened upon. The subsequent wave of values-affirmation, growth-mindset, and belonging interventions, many showing real if uneven effects, descends directly from this reframing.

The structural theories — realistic group conflict, social dominance, system justification — pushed the analysis back out from the individual mind to the social structure. They share a refusal to locate the cause of inequality solely in the heads of prejudiced individuals, and a corresponding insistence that the material and ideological organization of societies generates and maintains the attitudes that individuals carry. System justification theory, in particular, poses a sharp challenge to naive assumptions about self-interest: people, including the disadvantaged, often work to uphold systems that harm them, because the perceived legitimacy of the system is itself a psychological good. Intersectionality extends the structural critique methodologically, insisting that the single-axis research design is itself a political artifact that hides the experience of those at the crossings of multiple axes.

The deepest philosophical issue the literature raises is the relationship between implicit measures and moral responsibility. If a person sincerely and explicitly rejects prejudice, yet a measurable implicit association influences their behavior, in what sense is that behavior theirs, and to what does the moral evaluation attach? The question is live in ethics and in law. A parallel issue concerns the relationship between the individual and the structural: if prejudice is maintained partly by the system rather than by individuals, then interventions aimed only at changing individual hearts (diversity training, implicit-bias workshops) may be necessary but insufficient, and may even substitute for the structural change that the theory predicts is required. The empirical record on brief diversity trainings — generally small, short-lived effects — is consistent with this prediction, and is one reason the field has shifted toward structural and situational interventions (contact under Allport's conditions, jigsaw classrooms, values-affirmation) alongside the individual ones.

Bibliography Master

Allport, G. W. — The Nature of Prejudice (Addison-Wesley, 1954).
Greenwald, A. G. & Banaji, M. R. — "Implicit social cognition: Attitudes, self-esteem, and stereotypes", Psychological Review 102 (1), 4–27 (1995).
Greenwald, A. G., Nosek, B. A. & Banaji, M. R. — "Understanding and using the Implicit Association Test: I. An improved scoring algorithm", Journal of Personality and Social Psychology 85 (2), 197–216 (2003).
Steele, C. M. & Aronson, J. — "Stereotype threat and the intellectual test performance of African Americans", Journal of Personality and Social Psychology 69 (5), 797–811 (1995).
Spencer, S. J., Steele, C. M. & Quinn, D. M. — "Stereotype threat and women's math performance", Journal of Experimental Social Psychology 35 (1), 4–28 (1999).
Major, B. & Schmader, T. — "Coping with stigma through psychological disengagement", in Swim, J. K. & Stangor, C. (eds.), Prejudice: The Target's Perspective (Academic Press, 1998), pp. 219–241.
Walton, G. M. & Cohen, G. L. — "A question of belonging: Race, social fit, and achievement", Journal of Personality and Social Psychology 92 (1), 82–96 (2007); see also Walton, G. M. & Cohen, G. L. — "Stereotype lift", Journal of Experimental Social Psychology 39 (5), 456–467 (2003).
Pettigrew, T. F. & Tropp, L. R. — "A meta-analytic test of intergroup contact theory", Journal of Personality and Social Psychology 90 (5), 751–783 (2006).
McConahay, J. B. — "Modern racism, ambivalence, and the Modern Racism Scale", in Dovidio, J. F. & Gaertner, S. L. (eds.), Prejudice, Discrimination, and Racism (Academic Press, 1986), pp. 91–125.
Gaertner, S. L. & Dovidio, J. F. — "The aversive form of racism", in Dovidio, J. F. & Gaertner, S. L. (eds.), Prejudice, Discrimination, and Racism (Academic Press, 1986), pp. 61–89.
Glick, P. & Fiske, S. T. — "The Ambivalent Sexism Inventory: Differentiating hostile and benevolent sexism", Journal of Personality and Social Psychology 70 (3), 491–512 (1996).
Oswald, F. L., Mitchell, G., Blanton, H., Jaccard, J. & Tetlock, P. E. — "Predicting ethnic and racial discrimination: A meta-analysis of IAT studies", Journal of Personality and Social Psychology 105 (2), 171–192 (2013).
Greenwald, A. G., Banaji, M. R. & Nosek, B. A. — "Statistically small effects of the Implicit Association Test can have societally large effects", Journal of Personality and Social Psychology 108 (4), 553–561 (2015).
Meagher, H. & Brunton, U. — reanalysis and critique of IAT predictive validity, in Implicit Bias and Performance Disparities (2019); see also Forscher, P. S. et al. — "A meta-analysis of change in implicit bias", Personality and Social Psychology Bulletin 45 (9), 1301–1322 (2019).
Dasgupta, N. & Greenwald, A. G. — "On the malleability of automatic attitudes: Combating automatic prejudice with images of admired and disliked individuals", Journal of Personality and Social Psychology 81 (5), 800–814 (2001).
Olson, M. A. & Fazio, R. H. — "Relations between implicit measures of prejudice: What are we measuring?", Psychological Science 14 (6), 636–639 (2003).
Phelps, E. A., O'Connor, K. J., Cunningham, W. A., Funayama, E. S., Gatenby, J. C., Gore, J. C. & Banaji, M. R. — "Performance on indirect measures of race evaluation predicts amygdala activation", Journal of Cognitive Neuroscience 12 (5), 729–738 (2000).
Schmader, T., Johns, M. & Forbes, C. — "An integrated process model of stereotype threat effects on performance", Psychological Review 115 (2), 336–356 (2008).
Fryer, R. G., Levitt, S. D. & List, J. A. — "Exploring the impact of the Tuskegee Syphilis Study on medical-care decisions and African-American life outcomes" and related field work on stereotype threat (NBER Working Papers, 2010–2012).
Sackett, P. R., Hardison, C. M. & Cullen, M. J. — "On interpreting stereotype threat as accounting for African American–White differences on cognitive tests", American Psychologist 59 (1), 7–13 (2004); extended in Sackett et al. (2012).
Ganley, C. M., Mingle, L. A., Ryan, A. M., Ryan, K., Vasilyeva, M. & Perry, M. — "An examination of stereotype threat effects on girls' mathematics performance", Developmental Psychology 49 (10), 1886–1897 (2013).
Cohen, G. L., Garcia, J., Apel, N. & Master, A. — "Reducing the racial achievement gap: A social-psychological intervention", Science 313 (5791), 1307–1310 (2006).
Walton, G. M. & Cohen, G. L. — "A brief social-belonging intervention improves academic and health outcomes of minority students", Science 331 (6023), 1447–1451 (2011).
Bobo, L. — "Group conflict, prejudice, and the paradox of contemporary racial attitudes", in Katz, P. A. & Taylor, D. A. (eds.), Eliminating Racism (Plenum, 1988), pp. 85–114.
Sidanius, J. & Pratto, F. — Social Dominance: An Intergroup Theory of Social Hierarchy and Oppression (Cambridge University Press, 1999).
Jost, J. T. & Banaji, M. R. — "The role of stereotyping in system-justification and the production of false consciousness", British Journal of Social Psychology 33 (1), 1–27 (1994).
Crenshaw, K. — "Demarginalizing the intersection of race and sex: A Black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics", University of Chicago Legal Forum 1989 (1), 139–167 (1989).
Sherif, M., Harvey, O. J., White, B. J., Hood, W. R. & Sherif, C. W. — Intergroup Conflict and Cooperation: The Robbers Cave Experiment (University of Oklahoma Book Exchange, 1961).
Aronson, E. & Gonzalez, A. — "The jigsaw classroom: A new start for schools in crisis", in The Jigsaw Approach to Education (1971ff.); see Aronson, E. — The Jigsaw Classroom (Sage, 2nd ed., 2010).
Gaertner, S. L. & Dovidio, J. F. — Reducing Intergroup Bias: The Common Ingroup Identity Model (Psychology Press, 2000).
Tajfel, H. & Turner, J. C. — "An integrative theory of intergroup conflict", in Austin, W. G. & Worchel, S. (eds.), The Social Psychology of Intergroup Relations (Brooks/Cole, 1979), pp. 33–47.
Myers, D. G. & DeWall, C. N. — Psychology, 13th ed. (Worth, 2021), Ch. 12.
Gleitman, H., Gross, J. & Reisberg, D. — Psychology, 8th ed. (Norton, 2011), Ch. 11.