The Robbers Cave Experiment: How Groups Learn to Hate Each Other

In the summer of 1954, twenty-two eleven-year-old boys arrived at a camp in Robbers Cave State Park, Oklahoma, not knowing they were subjects in one of social psychology's most consequential studies. Within days they had formed tribes, within a week they were raiding each other's cabins, burning flags, and refusing to eat together. It took three weeks to reverse even part of the damage. Muzafer Sherif's experiment showed that hostility between groups does not require ancient grievances, prior hatred, or personality differences. Competition for limited resources is sufficient.

The Question Sherif Was Asking

The 1950s were a decade acutely focused on the origins of intergroup hostility. The Holocaust had ended less than ten years earlier. Racial segregation was openly practiced across the American South. The Cold War was hardening. Researchers wanted to know: what is the minimum condition for group hatred to emerge?

Two competing theories dominated the field. The authoritarian personality framework, proposed by Theodor Adorno and colleagues in 1950, suggested that prejudice was rooted in individual psychology — people with rigid, hierarchical personalities were prone to in-group/out-group thinking. Prejudice, by this account, was a pathology of certain individuals.

Muzafer Sherif, a Turkish-American social psychologist at the University of Oklahoma, disagreed. He argued that intergroup hostility was not a disease of individuals but a product of social structure. Specifically, he proposed that competition over scarce resources or incompatible goals was sufficient to generate hostility between groups, regardless of the personalities involved. This became known as realistic conflict theory (RCT).

Realistic conflict theory — the proposition that competition between groups for genuinely limited resources generates prejudice, discrimination, and hostility, and that those responses diminish when groups are united by shared goals that require cooperation.

To test this, Sherif needed a controlled environment where he could create groups, manufacture competition, and then observe what happened. A summer camp was his answer.

Source: Sherif, M., Harvey, O. J., White, B. J., Hood, W. R., & Sherif, C. W. (1961). Intergroup Conflict and Cooperation: The Robbers Cave Experiment. University of Oklahoma Book Exchange.

The Participants and the Setup

The 22 boys selected for the study were carefully screened to be as homogenous as possible. All were white, middle-class, Protestant, from stable two-parent households, academically average, and had no prior acquaintance with one another. None had behavioral or psychological problems. The researchers selected for normality deliberately: if hostility emerged in this group, it could not be attributed to prior disadvantage, individual pathology, or demographic difference.

The camp was structured in three phases, designed to flow naturally from one to the next.

Phase	Name	Duration	Goal
Phase 1	In-Group Formation	~1 week	Allow each group to bond independently
Phase 2	Intergroup Friction	~1 week	Generate competition and hostility between groups
Phase 3	Conflict Reduction	~1 week	Test whether cooperation on shared goals reduces hostility

The two groups were kept entirely separate during Phase 1. They were unaware of each other's existence. Each group hiked, swam, cooked, and solved problems together. Each group chose a name: one group became the Rattlers, the other the Eagles. Both groups developed internal hierarchies, norms, and group identities within days.

The camp staff — all of whom were Sherif's research team in disguise — observed continuously, noting group structure, leadership, language, and behavior. None of the boys or their parents knew they were participants in a psychology study.

Source: Sherif, M., Harvey, O. J., White, B. J., Hood, W. R., & Sherif, C. W. (1961). Intergroup Conflict and Cooperation: The Robbers Cave Experiment. University of Oklahoma Book Exchange.

Phase 2: Competition and the Rapid Emergence of Hatred

When the two groups were introduced to each other, Sherif and his team immediately organized a tournament. The events included baseball, tug-of-war, a treasure hunt, and tent-pitching competitions. Prizes — penknives, medals, and camping gear — would go only to the winning group. These were genuinely valued prizes, and only one group could have them.

The results of Phase 2 were stark and fast.

Within hours of the first competitive event, derogatory names had appeared. The Rattlers called the Eagles "sissies" and "stinkers." The Eagles called the Rattlers "braggers" and "cheats." By the second day, the Eagles burned the Rattlers' flag. The Rattlers retaliated by raiding the Eagles' cabin, overturning beds, tearing mosquito nets, and stealing a pair of jeans which they turned into a makeshift Rattlers flag.

The staff deliberately allowed the raids and confrontations to continue without intervention. The hostility escalated:

• Food fights broke out in the dining hall when the groups were forced to eat together
• The Rattlers put rocks in their socks for use as weapons
• Boys who had expressed willingness to befriend members of the other group were called traitors by their own group
• Sherif himself had to physically intervene to prevent a planned assault

The speed of this deterioration is the finding. These were not troubled boys. They were normal, carefully screened, middle-class eleven-year-olds. Within a week of structured competition, they were displaying the behavioral signatures of genuine intergroup hostility: dehumanizing language, property destruction, in-group solidarity, out-group contempt, and readiness for physical violence.

Source: Perry, G. (2018). The Lost Boys: Inside Muzafer Sherif's Robbers Cave Experiment. Scribe Publications.

How Groups Transform Individuals

One of the subtler findings of Phase 2 was how group membership changed individual boys' perceptions. In one exercise, both groups were shown a jar of beans and asked to estimate how many each group had counted correctly. Boys consistently rated their own group's estimates as more accurate, even when both groups had performed identically. The in-group favoritism operated on cognition, not just behavior.

This is the social identity mechanism that social psychologist Henri Tajfel would later formalize in his work on social identity theory. People derive self-esteem partly from their group memberships. To feel good about themselves, they need to feel good about their group, which requires the group to compare favorably with out-groups. Competition supplies the contrast, and the mind does the rest.

In-group favoritism — the tendency to evaluate members of one's own group more positively than members of out-groups, often without awareness that one is doing so.

The analogy here is a sports fan watching a referee's call. The same action — a tackle, a charge, a foul — is evaluated as fair or unfair depending almost entirely on which team committed it. The cognition follows the identity, not the other way around. Sherif's boys were doing exactly this with bean counts and swimming scores.

Source: Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The Social Psychology of Intergroup Relations (pp. 33–47). Brooks/Cole.

Phase 3: The Superordinate Goals Solution

If competition over resources created the hostility, Sherif predicted that cooperation toward shared goals — goals that neither group could achieve alone — should reduce it. He called these superordinate goals: objectives that are highly desirable to both groups but require their combined effort to achieve.

Superordinate goal — a shared objective that requires the cooperative effort of two or more groups and cannot be achieved by any single group acting alone.

The researchers staged a series of situations requiring inter-group cooperation. In one, the camp's water supply was sabotaged (actually by the research team), and both groups had to work together to locate and fix the problem. In another, the camp's truck broke down while on a trip; both groups had to pull it up a hill with a rope used earlier in tug-of-war competition. In a third, both groups pooled money to rent a movie. Each scenario was designed so that neither group could succeed without the other.

The results were measurable. After each cooperative task, the boys' sociometric ratings — their willingness to name out-group members as friends — shifted incrementally upward. The change was not immediate or dramatic. A single shared task did not dissolve the hostility. But across multiple superordinate goal experiences, the hostility declined. By the end of Phase 3, boys who had been enemies were sharing meals, sitting together on the bus home, and in some cases forming cross-group friendships.

Crucially, contact alone was not enough. Simply being placed in the same space — eating in the same dining hall, watching the same movie — did not reduce hostility. The contact had to be structured around a shared, meaningful goal.

Source: Sherif, M., Harvey, O. J., White, B. J., Hood, W. R., & Sherif, C. W. (1961). Intergroup Conflict and Cooperation: The Robbers Cave Experiment. University of Oklahoma Book Exchange.

The Failed Experiments Nobody Talks About

Gina Perry's 2018 book The Lost Boys revealed, through archival research and interviews with surviving participants, that Robbers Cave was not Sherif's first attempt to run this study. It was his third.

The first attempt, in 1949 at a camp in Connecticut, failed when the boys recognized what was happening and refused to behave as the theory predicted. Some befriended members of the out-group despite the competition structure. Others formed a coalition against the adult experimenters, whom they correctly perceived as manipulative.

The second attempt, in 1953 at Middle Grove, New York, also failed. The researchers had allowed the two groups to form friendships before separating them. When competition was introduced, the boys directed their hostility at the adult staff rather than at each other. The hoped-for intergroup conflict did not materialize. Sherif reportedly shut this study down before completion and barely mentioned it in his subsequent publications.

The 1954 Robbers Cave study — the third attempt, the one that worked — was the one that became famous. The failures, which could have complicated or falsified the theory, were buried.

Perry also documented that the adult staff were not passive observers. They actively instigated confrontations, spread rumors between the groups, and manipulated situations to produce the outcomes the theory predicted. One researcher posed as a camp caretaker and fed false information to the boys. Perry describes the experiment not as naturalistic observation but as a "choreographed enactment."

Source: Perry, G. (2018). The Lost Boys: Inside Muzafer Sherif's Robbers Cave Experiment. Scribe Publications.

What the Ethics Look Like in Retrospect

The Robbers Cave study involved deliberate deception of both the boys and their parents. The parents were told their sons were attending a regular summer camp. None of the participants knew they were in a psychological study. The hostility that was deliberately manufactured produced real psychological distress and real property destruction and came close to producing physical violence.

The boys were debriefed only at the end — and only partially. Gina Perry reported that when she contacted surviving participants decades later, most had not known until she told them that they had been subjects in a psychology study. Several expressed genuine disturbance at learning the truth.

By contemporary standards, the study would not survive ethical review. It involved deception of minors, no genuine informed consent, deliberate induction of distress, and no adequate debriefing. The fact that its findings are important does not rehabilitate the methods.

The treatment of the failed 1953 study raises a separate problem: selective publication. Sherif did not fabricate data, but he did suppress a study that produced results inconsistent with his theory. That kind of publication bias distorts the scientific record in ways that take decades to correct.

Source: Perry, G. (2018). The Lost Boys: Inside Muzafer Sherif's Robbers Cave Experiment. Scribe Publications.

What Survived the Critique

The ethical problems and the revelation of staff interference do not entirely invalidate the theory that Robbers Cave was designed to test. Realistic conflict theory has accumulated substantial support from other research streams over the decades.

Field studies of naturally occurring intergroup conflict consistently show that resource competition correlates with intergroup hostility. Laboratory minimal group paradigm experiments — in which participants are randomly assigned to meaningless groups and immediately begin to favor in-group members — suggest that even trivial group distinctions are sufficient to activate in-group favoritism, which is consistent with the broader framework if not with Sherif's specific causal story.

The superordinate goals finding — that cooperation toward shared objectives reduces intergroup hostility — has been tested in real-world contexts ranging from school desegregation programs to workplace diversity interventions. The results are mixed: superordinate goals help, but only when the cooperation is structured, equal-status, and successful. Contact under conditions of failure or unequal status can make things worse.

Source: Pettigrew, T. F., & Tropp, L. R. (2006). A meta-analytic test of intergroup contact theory. Journal of Personality and Social Psychology, 90(5), 751–783.

The Real-World Implications

The most important application of Robbers Cave's framework is in understanding how ordinary situations can generate extraordinary hostility. Sherif showed that you do not need to start with prejudiced people to end up with intergroup conflict. You need structure: groups with distinct identities, competition for resources only one side can win, and enough time.

This applies to workplace dynamics, political polarization, ethnic conflict, and international relations. When two departments compete for the same budget allocation, they often develop the same mutual contempt the Rattlers and Eagles displayed — name-calling, attribution of bad faith, refusal to cooperate on anything, even when cooperation would serve both groups better than competition.

The superordinate goal solution also has direct applications. Researchers studying political polarization have found that shared threats — climate change, pandemic response, economic crisis — can temporarily reduce in-group/out-group hostility by creating exactly the kind of superordinate goal structure Sherif described. The challenge is that the cooperation must actually succeed and must be sustained long enough to change attitudes, not just behavior.

Source: Gaertner, S. L., & Dovidio, J. F. (2000). Reducing Intergroup Bias: The Common Ingroup Identity Model. Psychology Press.

Gina Perry's Reappraisal: The Failed First Experiment and Researcher Influence

The Robbers Cave study presented itself as naturalistic observation — researchers watching boys behave freely under structured conditions. Gina Perry's archival investigation, published in The Lost Boys (2018), documented something different: a study shaped at nearly every stage by researchers who already knew what they wanted to find.

Perry located records of not one but two earlier attempts to run the study before the 1954 Oklahoma camp. The first, in 1949 at a camp in Connecticut, collapsed when boys from both groups befriended each other and collectively turned on the adult researchers, whom they correctly perceived as manipulative and controlling. Sherif moved on, redesigning the study. The second, the 1953 Middle Grove, New York camp, broke down for similar reasons: Sherif allowed the boys to form social bonds before separation, and when competition was subsequently introduced, the boys' hostility flowed toward the research staff rather than toward each other. The predicted intergroup conflict never emerged. Sherif closed the study early. Neither failure received meaningful treatment in his published work.

Publication suppression — the practice of not publishing or reporting studies that fail to confirm a hypothesis, which distorts the scientific record by making a theory appear more robustly supported than the full data set would suggest.

What Perry documented at Middle Grove was not simply a negative result left unpublished. Staff at the 1953 camp had actively tried to provoke hostility between the boys — spreading rumors, encouraging raids, positioning adults to stoke conflict — and the boys still would not cooperate with the experimental script. That finding, that boys with pre-existing cross-group friendships resisted induced conflict even when adult researchers pushed for it, was itself theoretically significant. It implied that prior social bonds are a meaningful buffer against intergroup hostility. Sherif's theory had no place for that implication, and the data went unexamined.

The Robbers Cave study of 1954 succeeded partly because it incorporated lessons from these failures: keep the groups isolated during Phase 1, never let them form friendships across group lines, and give staff more active roles in managing the trajectory of events. Michael Billig, the British social psychologist who first drew scholarly attention to Sherif's unpublished archives, argued that the 1954 study was less an observation of how intergroup conflict naturally unfolds and more a demonstration of what happens when adult authority figures systematically prevent the social processes that ordinarily moderate it.

The implication for replicability is direct. If researcher management was load-bearing — if the result depended on staff actively suppressing cross-group friendships and feeding inter-group grievances — then a replication that permits natural social contact might produce something closer to Middle Grove than Robbers Cave.

Source: Perry, G. (2018). The Lost Boys: Inside Muzafer Sherif's Robbers Cave Experiment. Scribe Publications — Perry conducted interviews with surviving participants and examined Sherif's personal archives, which had not been systematically analyzed in prior scholarship.

Source: Billig, M. (1976). Social Psychology and Intergroup Relations. Academic Press — Billig was among the first researchers to examine Sherif's unpublished records and highlight their implications for evaluating the Robbers Cave findings.

Frequently Asked Questions

Q: How long did Phase 2 take to produce serious hostility between the groups?
A: Less than one week. Within 24 to 48 hours of the first competitive event, derogatory naming, flag-burning, and cabin raids had begun. The speed was a central finding — it suggested that situational structure, not individual pathology, was the primary driver.

Q: Why did mere contact between the groups not reduce hostility in Phase 3?
A: Contact theory research shows that simply placing groups together is insufficient and can even reinforce prejudice when the contact occurs in a competitive or unequal context. The effective ingredient was structured cooperative activity toward a shared goal that required both groups and that both groups successfully achieved.

Q: How does realistic conflict theory differ from social identity theory?
A: Realistic conflict theory holds that actual competition over real resources drives intergroup hostility. Social identity theory, developed by Henri Tajfel and John Turner after Robbers Cave, holds that mere group membership and the drive for positive social identity are sufficient to generate in-group favoritism, even without any competition. Both mechanisms appear to operate; they are complementary rather than contradictory.

Q: What did the failed 1953 study actually show?
A: When boys were allowed to form cross-group friendships before being divided, they directed their hostility toward the adult researchers rather than toward each other. This finding — that prior social bonds can buffer against induced intergroup conflict — is itself important and arguably more optimistic than the Robbers Cave findings.

Q: Were the boys who participated psychologically harmed?
A: According to Gina Perry's interviews, several participants recalled the experience as frightening and disturbing and reported being troubled to learn, decades later, that it had been an experiment. The deliberate manufacture of fear, hostility, and real property destruction constituted genuine psychological harm without informed consent.

Q: How has realistic conflict theory been applied to reduce actual intergroup conflict?
A: Elliot Aronson's Jigsaw Classroom, developed in 1971 to improve interracial relations in newly desegregated Texas schools, used the superordinate goal principle directly. Students in mixed-race groups had to cooperate to complete assignments, with each student holding a unique piece of the required information. Academic performance and cross-race friendships both improved.

Q: Did Sherif's later work acknowledge the problems with the study?
A: Not significantly. Sherif defended the study's findings throughout his career. The 1953 failed study was barely mentioned in his publications and was not publicly analyzed as contradicting evidence until Perry's archival research decades later.

The Takeaway

The Robbers Cave experiment built real intergroup hatred out of nothing — no prior grievances, no personality disorders, no demographic differences — using only competition and group identity. That finding, however imperfect the study that produced it, remains uncomfortable and important. The mechanism Sherif identified operates in school cafeterias, corporate org charts, electoral politics, and between nations. What the experiment also showed, in Phase 3, is that the mechanism runs in reverse: when groups need each other to succeed at something both of them care about, the hatred dissolves faster than it formed. The challenge is engineering those conditions.

---

Read next:

• The Stanford Prison Experiment
• The Milgram Obedience Experiment
• The Blue Eyes Brown Eyes Exercise