Robbers Cave Experiment

The Robbers Cave experiment, conducted by Muzafer Sherif in the 1950s, studied intergroup conflict and cooperation among 22 boys in Oklahoma. Initially separated into two groups, they developed group identities. Introducing competitive tasks led to hostility between groups. Later, cooperative tasks reduced this conflict, highlighting the role of shared goals in resolving group tensions.

The Robbers Cave experiment is one of the most cited studies in social psychology. It has lasted decades. As social psychologist Roger Brown (1986) put it, the study is “the most successful field experiment ever conducted on intergroup conflict”.

Muzafer Sherif, Carolyn Sherif, and others in the 1950s conducted the study.

Muzafer Sherif argued that intergroup conflict (i.e., conflict between groups) occurs when two groups compete for limited resources.

The hypotheses tested were:

1. When individuals who don’t know each other are brought together to interact in group activities to achieve common goals, they produce a group structure with hierarchical statuses and roles.
2. Once formed, two in-groups are brought into a functional relationship under conditions of competition, and group frustration, attitudes, and appropriate hostile actions about the out-group and its members will arise; these will be standardized and shared in varying degrees by group members.

Study Procedure

The field experiment involved two groups of 11- to 12-year-old boys at Robbers Cave State Park, Oklahoma, America.

The twenty-two boys in the study were unknown to each other and all from white middle-class backgrounds.

They all shared a Protestant, two-parent background and were judged well-adjusted, with no history of behavioural problems.

This matching was deliberate. If hostility still emerged among such ordinary, well-adjusted boys, it could not be blamed on disturbed personalities or existing prejudice.

The researchers randomly divided the boys into two groups, with efforts being made to balance the groups’ physical, mental, and social talents.

Neither group was aware of the other’s existence.

None of the boys knew they were taking part in a psychological study. They believed they were simply at summer camp. The researchers posed as camp counsellors, a janitor and a caretaker to observe them unnoticed.

Individual groups were picked up by bus on successive days in the summer of 1954. Each group was transported to a 200-acre Boy Scouts of America camp at Robbers Cave State Park, Oklahoma.

Phase 1: In-group Formation (5-6 Days)

The members of each group got to know one other, social norms developed, leadership and group structure emerged.

At the camp, the two groups stayed separate from each other. Staff encouraged each group to bond through shared goals that required cooperative planning and teamwork.

During this first phase, the groups did not know of the other group’s existence.

The boys grew attached to their own group fast. Within the first week they built their own cultures and norms, hiking, swimming, and working together daily.

The boys chose names for their groups, The Eagles and The Rattlers, and stenciled them onto shirts and flags.

Toward the end of this first week, the groups discovered each other’s existence. An “us and them” language appeared at once. Before any competition had even begun, both sides were pressing staff to let them challenge the outsiders.

Phase 2: Group Conflict (4-5 Days)

The now-formed groups came into contact with each other, competing in games and challenges, and competing for control of territory.

Sherif now arranged the “competition stage” where friction between the groups was to occur over the next 4-5 days.

In this phase it was intended to bring the two groups into competition with each other in conditions that would create frustration between them.

Competitive Activities and Rising Tension

A series of competitive activities (e.g. baseball, tug-of-war etc.) were arranged with a trophy being awarded on the basis of accumulated team score.

There were also individual prizes for the winning group, such as a medal and a multi-bladed pocket knife. No consolation prizes were given to the “losers.”

The Rattlers’ reaction to the informal announcement of a series of contests was absolute confidence in their victory!

They spent the day talking about the contests and making improvements on the ball field, which they took over as their own. They even spoke of putting a “Keep Off” sign there!

They ended up putting their Rattler flag on the pitch. At this time, several Rattlers made threatening remarks about what they would do if anybody from The Eagles bothered their flag.

Situations were also devised whereby one group gained at the expense of the other. For example, one group was delayed getting to a picnic and when they arrived the other group had eaten their food.

Escalation and Intergroup Hostility

At first, this prejudice was only verbally expressed, such as through taunting or name-calling.

As the competition wore on, this expression took a more direct route. The Eagles burned the Rattlers’ flag.

The Eagles ultimately won the tournament, helped along by discreet encouragement from the counsellors running the games. Losing stung the Rattlers hard.

Then the next day, the Rattlers ransacked the Eagles’ cabin, overturned beds, and stole private property. The groups became so aggressive with each other that the researchers had to physically separate them.

Phase 3: Conflict Resolution (6-7 Days)

Sherif and colleagues tried various means of reducing the animosity and low-level violence between the groups.

The Robbers Cave experiments showed that superordinate goals worked far better than other strategies like communication or simple contact. A superordinate goal is one so large that it takes more than one group to achieve it.

A number of improvised reconciliatory opportunities were tried too, including a bean-collecting contest, a film screening, and a Fourth-of-July firecracker show. None of it lessened the tension much.

Sherif et al. concluded that such contrived contact opportunities were not going to lessen tensions between the groups.

Staff then introduced a series of superordinate-goal scenarios that both groups had to notice. Each posed a problem neither group could solve with its own resources, however hard the boys tried.

These scenarios played out at a new location. Staff hoped a new site would help the boys forget old grievances from Robbers Cave.

Other problem-solving superordinate goals followed. These included sharing a tug-of-war rope, and both groups “accidentally” finding a stuck-in-a-rut truck carrying food for both camps.

In the event, the joint pursuit of these goals lessened the intergroup conflict.

At breakfast and lunch on the last day of camp, the seating arrangements were considerably mixed up insofar as group membership was concerned.

Friendship choices shifted markedly toward the former enemy group. By the end of camp, roughly two-thirds of friendship choices crossed group lines. The boys also asked to travel home together on a single bus, rather than travelling separately.

Realistic Conflict Theory

Realistic conflict theory posits intergroup hostility and conflict arise when groups compete for limited resources. It emphasizes that competition over scarce resources (material goods, power, or social status) can lead to prejudice, discrimination, and animosity between groups.

The theory was solidified by the classic Robbers Cave experiment conducted by Muzafer Sherif in the 1950s.

Three key points from RCT are:

1. Resource Scarcity and Competition: When groups perceive that they compete for limited resources, hostility can arise.
2. Formation of Ingroup and Outgroup Dynamics: Through competition, groups develop a strong sense of “us” (ingroup) versus “them” (outgroup). This distinction can lead to negative stereotyping and increased animosity.
3. Superordinate Goals: Intergroup hostility can be reduced when conflicting groups collaborate on goals that neither group can achieve on its own. These goals supersede their smaller individual goals and encourage cooperation.

Realistic Conflict Theory suggests that competition over limited resources can drive intergroup hostility, but collaborative efforts towards shared objectives can help mitigate this conflict.

Real-World Examples

The theory’s three stages map onto everyday situations, not just a summer camp.

Stage 1 is group formation.

It appears wherever a group boundary already exists, such as rival football fans, competing sales teams in one firm, or neighbouring ethnic communities. A salient “us and them” line supplies the raw material for conflict later.

Stage 2 is competition.

It appears whenever two groups chase one scarce, exclusive prize, such as two towns competing for the same factory jobs in a recession. Sherif’s own example was unemployment: rising joblessness tends to track rising prejudice toward groups competing for scarce work.

Stage 3 is superordinate goals.

It appears whenever a shared threat forces old rivals to cooperate: a natural disaster, a common external threat, or a merged team that can only hit its target together. Interdependence toward that shared goal is what the theory says erodes the hostility competition created.

Critical Evaluation

The Robbers Cave study captured genuine, unscripted intergroup behaviour in a real summer camp. This gives it strong ecological validity: the extent to which findings reflect real-world behaviour rather than an artificial, lab-like setup.

Real-world patterns fit its central claim too. Periods of high unemployment often coincide with rising hostility toward outgroups competing for scarce jobs.

Even so, the study faces several important criticisms:

1. Artificial Groups and Competition: the rivalry was manufactured for the study, unlike entrenched real-world conflicts.
2. Ethical Concerns: the boys were deceived and exposed to real harm without informed consent.
3. Small, Unrepresentative Sample: only 22 white, middle-class, Protestant boys took part.
4. The Social Identity Theory Challenge: mere categorisation into groups, with no competition at all, can also produce hostility.
5. The Tyerman and Spencer Critique: familiar, cooperatively led groups stayed friendly under direct competition.
6. Researcher Intervention: counsellors secretly helped decide the tournament and staged later crises, so the “spontaneous” hostility and its resolution were partly engineered.

Artificial Groups and Competition

Middle-class boys randomly assigned to two summer-camp teams are not rival inner-city gangs, feuding ethnic groups, or opposing football supporters.

Real rivalries run deeper. Those real-world conflicts carry histories, ideologies and material stakes that a two-week summer camp cannot reproduce.

The two groups were strangers before the study began. Their competition was staged entirely by the research team.

It did not arise from any genuine dispute over land, jobs or status. The boys had no grudge to begin with.

Researchers built one from scratch, on a tight schedule of games, trophies and prizes.

Critics argue this artificiality matters. The hostility Sherif recorded may reflect the transitory, invented nature of the experimental groups, not a general law about how real communities compete.

Ethical Concerns

The boys were deceived about the true purpose of the camp. They did not know they were research participants.

Neither they nor their parents could give fully informed consent to having hostility deliberately provoked in them.

Researchers manufactured real conflict: fights, raids, thefts, humiliation, flag-burning.

This exposed children to genuine physical and psychological harm, with no adequate protection built into the design.

The camp staff had to physically separate the boys more than once when fighting broke out between the groups.

This was not incidental to the design. Provoking real hostility was necessary to produce the very conflict the researchers wanted to study.

The conflict was resolved by the end, and the boys reportedly left on good terms. Still, the distress was real while it lasted.

Provoking it in children is difficult to justify by modern standards.

Small, Unrepresentative Sample

Only 22 boys took part, all of a single age, sex, ethnicity, religion and social class.

They were white, Protestant, middle-class 11- to 12-year-olds. Girls, adults, and children from different backgrounds were excluded entirely.

Findings from such a narrow sample may not generalise. The results may say more about this one group of boys than about intergroup conflict in general.

Wetherell (1982) found that Polynesian children in New Zealand were far more generous and welcoming to outgroups than Sherif’s American boys.

Culture shaped the outcome. This reflects cooperative cultural norms, not a fixed human tendency toward hostility.

Intergroup hostility is not inevitable. It appears to be moderated by culture, which the homogeneous Robbers Cave sample could never reveal on its own.

The Social Identity Theory Challenge

Tajfel and Turner’s (1979) Social Identity Theory mounts the most serious theoretical challenge to Sherif’s account. The theory holds that mere categorisation into groups, with no competition at all, is enough to trigger bias.

Using the minimal group paradigm, Tajfel’s team assigned strangers to groups on a trivial, arbitrary basis. There was no competition, no interaction, no shared history.

That alone was enough to produce ingroup favouritism.

No prize was at stake. If people favour their own group with nothing to compete over, competition cannot be a necessary cause of bias.

On this reading, Robbers Cave may over-credit competition. The hostility may instead flow from the simpler act of dividing boys into the Rattlers and the Eagles.

Categorisation alone may have done most of the work, before the first tournament game was even played.

The two theories are not fully exclusive. Sherif’s competition and Tajfel’s categorisation are now often taught as complementary sources of prejudice, rather than as rival explanations.

The Tyerman and Spencer Critique

Tyerman and Spencer (1983) ran a pointed field test with English boy scouts.

Unlike Sherif’s boys, these scouts already knew one another well and had a leader who actively encouraged cooperation.

Competition between four patrols in their camp stayed friendly throughout.

Solidarity held steady. It produced no rise in ingroup favouritism, and the boys’ sense of belonging to the whole camp remained intact.

The researchers concluded that Sherif’s hostility reflected the transitory, artificial nature of his experimental groups, not competition as such.

That is a strong claim.

Prior friendship, familiarity and cooperative leadership were all deliberately excluded from Robbers Cave.

Those factors may be decisive in whether competition turns into conflict at all. This directly challenges Sherif’s claim that competition alone is a sufficient condition for hostility.

Researcher Intervention

Sherif’s team did not just observe the hostility.

Counsellors running the tournament quietly helped the Eagles win. Staff later staged both the water-supply failure and the stalled truck, rather than waiting for real crises to appear.

This matters because the study’s authority rests on the hostility looking spontaneous. Staff helped decide who won, and manufactured the “problems” that supposedly forced cooperation.

The experimenters were not neutral observers. They were, to some degree, co-authors of both the conflict and its resolution.

Gina Perry’s (2018) archival research backs this reading. Her account of Sherif’s abandoned 1953 pilot study at Middle Grove shows the same team going to considerable lengths, including damaging boys’ property, to provoke conflict that never appeared.

None of this erases the two headline findings. It does mean the tidy three-stage narrative understates how heavily staff shaped events at Robbers Cave.

Contemporary Research

Scholarship has moved on since 1961.

Gina Perry’s 2018 book The Lost Boys is the most influential modern reappraisal of Robbers Cave.

Drawing on Sherif’s own field notes and interviews with surviving participants, Perry examined an earlier, abandoned 1953 study at Middle Grove, New York.

That study used a similar design.

At Middle Grove, the boys worked out that the adults were engineering their conflict.

They turned their resentment on the researchers instead of on each other, forcing the study to be scrapped.

Perry argues that the “clean” 1954 narrative understates how heavily Sherif’s team stage-managed events at Robbers Cave.

The engineering ran deep.

As Perry’s archival analysis shows, this does not undo the study’s two robust lessons.

Contact alone still failed to reduce hostility, and cooperation toward superordinate goals still succeeded.

It does sharpen the researcher-effects and artificiality criticisms above.

A second, separate question is whether the Stage 3 finding still holds up against the modern evidence base on intergroup contact.

Paluck, Green and Green (2019) tested this by updating an earlier, influential meta-analysis. They kept only the 27 experiments that randomly assigned people to a contact condition and measured prejudice at least a day afterwards, not immediately.

Contact interventions reduced prejudice on average. The effect was notably weaker for ethnic and racial prejudice than for prejudice toward people with disabilities, and the largest, most rigorous studies produced the smallest effects.

That pattern fits publication bias: smaller, noisier studies tend to inflate the average.

Applied back to Robbers Cave, the finding both supports and qualifies the study’s third stage. Cooperation still helps. But the confident claim that contact plus cooperation “always works” outruns what even the best recent studies can show.

Applications

Robbers Cave offers a practical recipe, not just a theory: replace competition with cooperation toward a shared, necessary goal. That recipe has shaped real prejudice-reduction work well beyond the original camp.

The Jigsaw Classroom

Elliot Aronson’s jigsaw classroom is the best-known application.

It was developed to ease ethnic tension in newly desegregated American schools. Pupils work in small, mixed groups, and each pupil learns one unique piece of a lesson.

No pupil can succeed alone. Each has to teach their piece to the others, and no group can finish the task unless every member’s piece gets used.

This makes classmates from different backgrounds genuinely dependent on one another. No single pupil can opt out. It scales Sherif’s superordinate-goal principle down to classroom size (Aronson & Patnoe, 1997).

Studies link the technique to reduced prejudice, better cross-group friendships, and stronger outcomes for lower-achieving pupils. A pupil who once mocked a rival group’s accent now depends on that same voice to pass the test.

Peacebuilding Between Communities

The superordinate-goals finding also underpins peacebuilding work.

Programmes that put formerly opposed communities to work on one shared project follow the same logic Sherif tested at camp. A joint economic venture, a shared environmental problem, or a public-health goal can all serve as the superordinate goal.

No single act of cooperation ends a conflict.

It is the accumulation of shared work, project after project, that slowly erodes old hostility, just as it did for the Eagles and the Rattlers.

Interdependence for a necessary, shared end supersedes the competition that divided the groups in the first place. Two communities rebuilding a shared water system after a disaster, for instance, often end up cooperating on much more than just pipes and pumps.

Cross-Functional Teams at Work

The same principle shapes how organisations design teams and structure incentives.

When departments compete for one fixed pool of budget, bonus, or recognition, Robbers Cave’s Stage 2 structure predicts what happens next. Turf wars follow, along with hoarded information, missed handoffs, and mutual blame whenever targets are missed.

Restructuring around shared goals changes the underlying incentive entirely.

Collective rewards that require inter-departmental cooperation are the Stage 3 corrective. Sales and product teams sharing one bonus pool tied to a single, shared customer-satisfaction score, for example, have a genuine reason to help each other rather than compete for credit.

No team can hit it alone. Otherwise the old rivalry simply continues under a new name, just with different logos on the shirts.

References