Muzafer Sherif, the Robbers Cave experiment, and the study that came first. · published 10 September 2026
Robbers Cave was the second attempt at the study — the 1953 Middle Grove camp failed because the boys already knew each other, worked out the adults were staging the sabotage, and turned on the researchers instead, so the famous design quietly made sure the two groups met as strangers.
These are the working notes, unedited. Everything the video
states as fact comes from here, and nothing else may be stated as fact.
The sections on limits and on what the finding does not show are
part of the brief, not an afterthought — see
the method.
The famous one: Sherif and colleagues, Robbers Cave State Park, Oklahoma, summer 1954. Twenty-two boys aged eleven and twelve, all white, Protestant, middle-class, from Oklahoma City, none of whom knew each other beforehand. They were brought to the camp in two separate buses and kept apart for the first week, long enough for each group to name itself — the Rattlers and the Eagles — and make its own flag.
Stage two put the groups against each other in a tournament: baseball, tug of war, tent-pitching, cabin inspection, with a trophy, medals and pocket knives for the winners and nothing for the losers. Hostility escalated fast. The Eagles burned the Rattlers’ flag; the Rattlers raided the Eagles’ cabin, overturned beds and took private property. Staff had to physically separate them.
Stage three removed the hostility again by giving both groups a problem neither could solve alone. The camp’s water supply failed and the boys spent about three quarters of an hour finding and clearing the blockage together. A truck carrying food would not start, and they pulled it with the same rope they had used for tug of war. They pooled money to hire a film. By the end, most of the boys’ friendship choices crossed group lines, and they asked to go home on one bus rather than two.
What it establishes
Competition over something only one group can have is enough to produce hostility between groups with no prior grievance, and contact alone does not undo it — a shared goal that needs both sides does. This is the founding evidence for realistic conflict theory.
The part usually left out
Robbers Cave was the third camp study, not the first, and the one before it failed. In 1953, at Middle Grove in New York state, Sherif ran essentially the same design and could not make it work. The boys had been allowed to mix before they were split, and had friendships across the two groups. When the staff tried to start the friction stage the boys would not take it up. One of Sherif’s assistants pulled down a tent belonging to one group and trampled their belongings, expecting them to blame the other group; the boys worked out that the adults were behind it. They turned on the researchers rather than on each other. Sherif ended the study and did not publish it. It was reconstructed from his own field notes and from interviews with surviving participants by Gina Perry, in “The Lost Boys” (2018).
At Robbers Cave the following summer, the boys were kept from meeting until each group had already formed. Several of the tournament events — cabin inspection, tent-pitching — were scored by the staff rather than won outright, which gave the researchers room to keep the contest close. The water failure and the stalled truck in stage three were staged by the experimenters; Sherif’s own published account says so.
Limits
Twenty-two boys, one camp, one summer, a sample chosen to be as similar to each other as possible. The 1953 study is known mainly through Perry’s archival work and the participants’ memories decades later, so it is evidence that the design failed once, not a measured result with a number attached. Sherif’s stage-three manipulations were disclosed in his own writing and are not a hidden scandal — the concealed part is the earlier failure and the lengths taken to avoid repeating it. Perry’s reading of Sherif’s motives is her interpretation, well sourced but contested.
What this does not show
It does not mean the Robbers Cave findings are fabricated or that realistic conflict theory is wrong; group conflict over contested resources has been demonstrated many times since, and the superordinate-goal result has held up better than most findings of its era. Nor does it mean the boys were actors. What it means is narrower and more interesting: the experiment needed particular conditions to produce its effect, and one of those conditions was that the two groups were strangers.