Difficulty: Intermediate | Prerequisites: Familiarity with social identity theory, the IAT and basic experimental design (see Parts 1 and 2).
This unit tackles one of the most socially significant topics in the course: how group-based beliefs (stereotypes), feelings (prejudice) and behaviour (discrimination) arise, persist and cause harm. You will learn how stereotypes are structured along dimensions of warmth and competence, how the mere awareness of a negative stereotype can undermine performance (stereotype threat), and how competition between groups fuels prejudice (realistic conflict theory). The glass ceiling and glass cliff round out the unit by connecting these processes to workplace inequality. If you have followed the earlier material on social identity theory, implicit attitudes and the IAT, this unit will feel like a natural extension.
Stereotypes are cognitive shortcuts that categorise people by group membership along warmth and competence dimensions. Stereotype threat shows that simply being aware of a negative stereotype about your group can impair your performance, even if you do not personally endorse the stereotype. Realistic conflict theory explains prejudice as a product of intergroup competition for limited resources, and concepts like the glass ceiling and glass cliff describe how these processes translate into structural workplace barriers.
Stereotype
A generalised belief about the characteristics of a group of people. Stereotypes can be positive or negative, but even positive stereotypes constrain how individuals are perceived.
The ABC of stereotypes
Stereotypes have three components, mirroring the ABC of attitudes: Affect (prejudice, the emotional reaction), Behaviour (discrimination, the differential treatment) and Cognition (the stereotype itself, the belief about group traits).
Stereotype content model (Fiske et al., 2002)
A framework proposing that stereotypes of social groups cluster along two dimensions: warmth (is this group friendly or threatening?) and competence (is this group capable or ineffective?). Groups fall into four quadrants: high warmth/high competence (in-group, allies), high warmth/low competence (elderly, disabled, often eliciting pity), low warmth/high competence (rich people, Asians in some Western stereotypes, often eliciting envy), and low warmth/low competence (homeless people, drug users, often eliciting contempt).
Stereotype threat
The phenomenon where awareness of a negative stereotype about one's group creates anxiety and cognitive load that impairs performance in the stereotyped domain. The threat does not require believing the stereotype; knowing it exists is enough.
Steele & Aronson (1995)
The foundational stereotype threat study. Black participants performed worse on a difficult verbal test when it was described as diagnostic of intellectual ability (making race salient) than when it was described as a problem-solving exercise. White participants were unaffected by the framing.
Spencer et al. (1999)
Extended stereotype threat to gender and maths. Women performed worse than men on a difficult maths test when told the test produced gender differences, but performed equally when told it did not.
Process model of stereotype threat (Schmader et al., 2008)
A model explaining how stereotype threat impairs performance. It proposes three mechanisms working together: physiological stress, active monitoring of performance (vigilance) and efforts to suppress negative thoughts and emotions. These drain working memory, which is the cognitive resource needed for complex tasks.
Realistic conflict theory
The theory that prejudice arises from competition between groups for limited resources (jobs, land, funding). When groups compete, hostility and negative stereotypes increase; when they cooperate toward shared goals, prejudice decreases.
Robber's Cave study (Sherif, 1961)
A field experiment at a boys' summer camp. Two groups of boys were created, introduced to competition (tug-of-war, baseball), which quickly produced hostility and negative stereotyping. Prejudice was later reduced through superordinate goals that required both groups to cooperate.
In-group favouritism
The tendency to give preferential treatment to members of your own group, even when group membership is arbitrary (as in the minimal group paradigm). This does not require hostility toward the out-group; it is a bias toward your own.
Glass ceiling
An invisible barrier that prevents women and minorities from advancing beyond a certain level in organisations, despite having the qualifications. The barrier is structural and systemic, not a single act of discrimination.
Glass cliff
The phenomenon where women and minorities are more likely to be appointed to leadership positions during times of crisis or decline, when the risk of failure is highest. In simple terms, they get the top job when the organisation is already in trouble.
The ABC framework distinguishes the cognitive component (stereotype: "Group X is Y"), the affective component (prejudice: "I feel negatively toward Group X") and the behavioural component (discrimination: "I treat Group X differently"). These three can operate independently. You can hold a stereotype without feeling prejudice, or feel prejudice without acting on it.
The stereotype content model (Fiske et al., 2002) maps stereotypes onto two dimensions:
Warmth: is the group perceived as friendly, trustworthy, well-intentioned?
Competence: is the group perceived as capable, skilled, intelligent?
The four quadrants produce distinct emotional reactions:
High warmth + high competence: admiration (in-group, close allies).
High warmth + low competence: paternalistic pity (elderly, people with disabilities).
Low warmth + high competence: envious prejudice (rich people, some ethnic stereotypes).
Low warmth + low competence: contemptuous prejudice (homeless people, drug users).
This model is useful because it explains why stereotypes are not all alike. The emotions they trigger, and the discrimination they produce, differ by quadrant.
Steele & Aronson (1995) is the landmark study. Black students took a difficult verbal test. In the "diagnostic" condition (told the test measured intellectual ability), Black students underperformed compared to White students. In the "non-diagnostic" condition (told it was just a problem-solving exercise), the gap disappeared. The mere framing of the test as relevant to the stereotype was enough to impair performance.
Spencer et al. (1999) replicated the pattern with gender and maths. Women underperformed men when told the test showed gender differences, but performed equally when this cue was removed.
Schmader et al. (2008) proposed the process model explaining the mechanism:
Physiological stress: anxiety triggers a stress response.
Performance monitoring: the person becomes hyper-vigilant, constantly checking for errors.
Thought suppression: the person tries to push away stereotype-related thoughts and emotions.
All three drain working memory, the cognitive resource needed for complex reasoning. With less working memory available, performance drops on difficult tasks.
Stereotype threat does not require endorsing the stereotype. You only need to be aware of it and to care about the domain.
Realistic conflict theory (Sherif) proposes that intergroup hostility is a rational (if ugly) response to perceived competition over scarce resources. When groups compete, prejudice rises; when they cooperate, prejudice falls.
The Robber's Cave study (Sherif, 1961) tested this with 22 boys at a summer camp:
Phase 1 (group formation): boys were split into two groups (Eagles, Rattlers) and each developed its own identity.
Phase 2 (competition): the groups competed in games with prizes. Hostility escalated rapidly, with name-calling, cabin raids and refusal to eat together.
Phase 3 (cooperation): researchers introduced superordinate goals (problems that required both groups to work together, such as fixing a broken water supply). Hostility dropped significantly.
The study demonstrates that prejudice can be created and reduced through the structure of intergroup relations.
In-group favouritism: even without competition, people prefer their own group. The minimal group paradigm (Tajfel) showed that assigning people to groups based on trivial criteria (a coin flip) was enough to produce favouritism in resource allocation.
The glass ceiling describes the invisible barrier that stops women and minorities from reaching the highest levels of organisations. It is "glass" because it is not an explicit policy; it operates through accumulated biases, networking exclusion, stereotyping and structural disadvantage.
The glass cliff adds a further layer: when women and minorities do break through, they are disproportionately appointed during crises, when the probability of failure is elevated. If they fail, the failure may reinforce the stereotype that they were not suited for leadership; if they succeed, the credit may go to the recovery rather than their skill.
Students often think stereotype threat means the person believes the stereotype about their group. It does not. The threat arises from awareness of the stereotype, not endorsement of it. High-achieving individuals who care deeply about the domain are often the most affected.
Students sometimes assume realistic conflict theory explains all prejudice. It explains prejudice rooted in resource competition, but prejudice also arises from social identity processes (in-group favouritism) that require no competition at all.
The glass ceiling is sometimes mistaken for overt discrimination. It is subtler than that: it refers to systemic, cumulative barriers rather than a single discriminatory act.
Students occasionally conflate the glass ceiling and the glass cliff. The ceiling is about being blocked from the top; the cliff is about being placed at the top under conditions of high risk.
⚠️ Be able to place social groups in the four quadrants of the stereotype content model and name the associated emotion.
⚠️ Steele & Aronson (1995) is almost certainly on the exam. Know the design, the conditions, the finding and the mechanism.
⚠️ Spencer et al. (1999) extends the pattern to gender and maths. Be ready to compare the two studies.
⚠️ Know the three mechanisms in Schmader et al.'s (2008) process model and how they connect to working memory.
⚠️ The Robber's Cave study is a favourite. Know the three phases and what each demonstrates. Be ready to name one strength (ecological validity, real behaviour observed) and one limitation (small sample of similar boys, demand characteristics, ethical concerns).
⚠️ The exam asks you to critically evaluate studies. For each study, prepare at least one strength and one limitation.
The two dimensions of the stereotype content model are ________ and ________. (Warmth and competence.)
True or False: Stereotype threat only affects people who believe the stereotype about their group. (False. It affects anyone aware of the stereotype who cares about the domain.)
In the Robber's Cave study, what reduced intergroup hostility? (Superordinate goals that required cooperation between the two groups.)
True or False: The glass cliff describes an invisible barrier preventing promotion. (False. That is the glass ceiling. The glass cliff describes being appointed to leadership during a crisis.)
Name the three mechanisms in Schmader et al.'s process model of stereotype threat. (Physiological stress, performance monitoring and thought suppression.)
Q: Using the stereotype content model, explain why elderly people and rich people tend to elicit different emotional reactions despite both being stereotyped.
A: According to Fiske et al. (2002), elderly people are stereotyped as high warmth but low competence, which elicits paternalistic pity ("they are kind but need help"). Rich people are stereotyped as low warmth but high competence, which elicits envy ("they are capable but cold"). The emotional reaction depends on where the group falls on the two dimensions, not simply on whether the stereotype is positive or negative.
Q: Describe the Steele & Aronson (1995) study. What was the IV, the DV and the key finding?
A: The IV was how the test was framed: diagnostic of intellectual ability (stereotype-relevant) or a non-diagnostic problem-solving exercise. The DV was performance on a difficult verbal test. The key finding was that Black participants underperformed in the diagnostic condition compared to the non-diagnostic condition, while White participants' performance was unaffected by the framing. This demonstrated that activating awareness of a racial stereotype impaired performance.
Q: How does the process model (Schmader et al., 2008) explain the performance drop in stereotype threat?
A: The model identifies three processes that operate simultaneously: a physiological stress response that generates anxiety, active performance monitoring (vigilance for errors), and effortful suppression of negative thoughts and emotions related to the stereotype. Each process consumes working memory. Because complex tasks depend on working memory, the combined drain causes performance to decline.
Q: Evaluate the Robber's Cave study. Give one strength and one limitation.
A: Strength: the study had high ecological validity because real intergroup behaviour was observed in a natural setting over several weeks, with actual hostility and cooperation rather than self-report measures. Limitation: the sample was 22 White, middle-class, Protestant boys of similar age, which severely limits generalisability. The findings may not extend to groups that differ in ethnicity, gender, socioeconomic status or age. There are also ethical concerns about deliberately inducing hostility in children.
Q: Explain the difference between the glass ceiling and the glass cliff, and give a real-world example of how they might operate together.
A: The glass ceiling is the invisible barrier that prevents women and minorities from reaching top leadership positions. The glass cliff is the tendency for those who do break through to be appointed during times of organisational crisis. Together: a woman might be blocked from the CEO role for years (glass ceiling), then finally appointed when the company is losing money and the position is seen as a poisoned chalice (glass cliff). If she fails, the failure reinforces the stereotype; if she succeeds, the credit may be attributed to the recovery itself.
Stereotype threat connects to automatic vs controlled thinking (Part 1): the threat operates partly because monitoring and suppression consume the controlled-processing resources that would otherwise be available for the task. In-group favouritism links back to social identity theory (Part 1) and the motivation to maintain positive self-esteem through group membership. The IAT (Part 2) is one of the main tools for measuring the implicit stereotypes discussed here. Realistic conflict theory also connects to persuasion (Part 2): propaganda and political messaging often frame outgroups as competitors for resources, leveraging the same mechanism Sherif demonstrated at Robber's Cave.
Stereotypes, prejudice, discrimination, ABC of stereotypes, affect behaviour cognition, stereotype content model, Fiske, warmth, competence, admiration, pity, envy, contempt, stereotype threat, Steele, Aronson, Spencer, women and maths, gender gap, process model, Schmader, working memory, physiological stress, performance monitoring, thought suppression, realistic conflict theory, Sherif, Robber's Cave, Rattlers, Eagles, superordinate goals, in-group favouritism, in-group bias, minimal group paradigm, Tajfel, glass ceiling, glass cliff, leadership, workplace inequality, women in leadership, crisis leadership, NYU London, FA24 midterm, social psychology