Relaunched: 4 March 2018
The Robber’s Cave study is on a par with Stanley Milgrim’s ‘Obedience Experiments’ and Philip Zimbardo’s infamous prison study at Stanford University (Craig Haney, Curtis Bank & Philip Zimbardo, 1973), both for its sheer audaciousness and what it tells us about situational pressures to produce normative influence.
Muzafer Sherif had been a growing force in the development of Social Psychology ever since his ‘autokinetic effect’ experiments in 1935 had developed the concept of conformity that would come to be known as informational influence. In fact, Sherif could be considered one of the founders of Social Psychology. His work was also highly thought of by interactionist sociologists, becoming the first psychologist to receive the Cooley-Mead Award for contributions to Social Psychology from the American Sociological Association.
By the end of the 1940s his interest in understanding social processes, particularly social norms and social conflict had led him to conceive of developing a field experiment in which pubescent boys would be nurtured into forming 2 distinctive teams with strong group identities to see how conflict between the 2 groups could be exacerbated and then reduced. This would be the basis of the famous and challenging Robber’s Cave study of 1954 (Muzafer Sherif et al, 1961).
Several promising pilot studies were carried out in the early 1950s before a full study took place in the Summer of 1953 in New York state. Sherif was assisted by his wife and collaborator, Carolyn Wood Sherif, and 3 Native American doctoral students under Muzafer’s supervision at the University of Oklahoma: O J Harvey (Choctaw), William Hood (Cherokee) and Jack White (Kiowa). However, the 1953 study was never properly documented because, quite simply, it failed. In 2008 Alex Haslam & Steve Reicher, based on the diary and notes of Herbert Kelman, then acting as an advisor to Sherif, drew attention to the 1953 study. They revealed that the boys “refused to be divided and, together, they resisted attempts by the experimenters to set them against each other.” Reicher (2012) confirms that this was an attempt by Sherif to stage the full experiment a year prior to the documented version but which had been abandoned. Don Beck, future co-developer of Spiral Dynamics and then assisting with the final project documentation in 1961, states (2012) that the reason the 1953 project didn’t work was that Sherif et al didn’t put in enough focus on building the group identities of the 2 sets of boys. They were bussed into the camp on the same vehicle and the attempt to split them into 2 groups at that stage simply didn’t work.
Joseph Trimble. a doctoral student under first Muzafer and then Hood, explains (2018) how the team recovered from the 1953 failure: “After the 1953 New York state summer camp study failed, O J and B Jack White sat with Muzafer Sherif at the former campsite and talked about doing another study. All three agreed they wanted to try it one more time but both O J and B Jack said they wouldn’t be involved unless it ran more smoothly, ie: if Muzafer was in charge. So O J took charge of the Robbers Cave study.”
The video below (copyright © 2015 Cummings Center/Archives of the History of American Psychology) provides an overview of the 1954 study and some of the thinking that led-up to it.
Phase 1: in-group formation
Harvey used his Native American heritage to get permission from the Choctaw Nation Tribal Council to use the land in Robber’s Cave State Park for the study. (The ‘Robber’s Cave’ had been the hideaway of Jesse James.) Permission was granted and spiritually blessed by the tribal elders.
In the Summer of 1954 22 11-12-year-old boys took part in a 2-week summer camp. The boys were screened to ensure they were well-adjusted – no neurotic tendencies and no record of past disturbances in behaviour – and came from a similar background – white, Protestant, stable 2-parent middle class families from Oklahoma. None of the boys knew each other, coming from different schools and neighbourhoods. As part of the matching process, the boys were rated (including IQ) by teachers. The boys were assigned to one of 2 groups. The matched pairs groups themselves were very similar. They were matched as closely possible on criteria such as height and weight, athleticism and popularity outside of camp, previous camp experience and musicality. They were then, as individual groups, picked up by bus on successive days and transported to Robber’s Cave.
The researchers acted as camp counsellors. A nominal fee was charged to parents for the camp; but they were asked not to visit on the pretext that it might make the boys homesick.
The research methods used were:-
- Observation – a participant observer was allocated to each group for 12 hours per day
- Sociometric analysis – patterns such as those in friendship groups were noted and studied
- Experiment – eg: the boys had to collect beans and estimate how many each boy had collected
- Tape recordings – adjectives and phrases used to refer to their own group members and the members of the other group
Each group, initially unaware of the other’s presence, had their own cabin and were independent, camping out, cooking, improving swimming places, carrying canoes over rough terrain to water and playing various games. They were assigned activities that held a common appeal for group members and that depended on the collective effort of the group as a whole – such as a treasure hunt with a $10 reward that the group could spend as it wanted to.
Each group soon developed a distinctive set of ideas and rules about how to behave. In one group it became the norm – the shared meme – to act tough, swear a lot and not complain about small injuries. The other group swam in the nude and made any expression of homesickness taboo. Each group was tasked with coming up with a name for itself – thus, ‘Rattlers’ and ‘Eagles’ respectively – and a flag for their group. The researchers gave the 2 groups caps and t-shirts with their group names on to increase this sense of group identity. They became cohesive groups, with low-ranking and high-ranking members, reflecting the PURPLE vMEME’s working to create a tribal structure to meet its safety-in-belonging needs.
After a week the groups were made aware of each other. The researchers observed that in-group/out-group terms began to be used. When they watched a film together, they sat in their own distinct groups. The 2 groups were clearly discriminating against each other.
Phase 2: friction
The 2 groups wanted to play each other at baseball which enabled the researchers to introduce a competition: a grand tournament comprising 10 sporting events, plus cabin cleanliness awards and acting events. The boys were told that the best performing group in the tournament would receive a trophy, 4-bladed knives and medals. There were to be no prizes for the losers. The 2 groups were made to eat together in a common dining hall where the tournament’s grand prizes were on display for all to see.
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 to such an extent that they 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. This reflects how important the land in its tribe’s possession is to the PURPLE vMEME.
Even before the tournament began, the groups were insulting each other – eg: “Ladies, first” – singing offensive songs about each other and refusing to ear together. They were even physical fights between members of the 2 groups! Soon epithets such as ‘sneaks’, ‘cheats’, bums’, ‘cowards’ and ‘stinkers’ were being used in reference to members of the other group. (Terms like ‘friendly’, ‘tough’ and ‘brave’ were used for their own group members.) The Rattlers, in particular, became concerned about encroachment on what they considered their territory – eg: “They had better not be in our swimming hole.”
The researchers manipulated the points so they could control the competition.When the Rattlers won a tug of war competition, the Eagles responded by burning their flag, with the group’s leader proclaiming: “You can tell those guys I did it … I’ll fight ’em!” The Rattlers retaliated by raiding the Eagles camp (amid scuffles!) and damaging their property – overturning beds and ripping out mosquito netting. They stole one boy’s jeans and a stack of comic books. The Eagles were incensed. When the Rattlers were eating dinner, they returned the raid, bringing with them sticks and bats to wreak maximum havoc. They then filled their socks with stones to use as weapons, on the chance that the Rattlers would soon plan a counter-raid of their own. The researchers had to intervene to calm things down.
With some ‘help’ from the researchers, the Eagles won – but their prizes, when awarded, were stolen by the Rattlers.
Now nearly at the end of the second week, the 2 sides met for a fight. However, the researchers again intervened, forcing both sides to withdraw.
The researchers mow instigated a 2-day cooling off period. In this time the boys were asked to list features of the 2 groups. The boys tended to characterise their own in-group in very favourable terms – eg: ‘friendly’, ‘tough’ and ‘brave’ – while the other, out-group was characterised in very unfavourable terms such as ‘sneaky’, ‘bums’ and ‘cowards’.
Other evidence of in-group bias included, during the bean collecting task, members of one team consistently overestimating the numbers of beans collected by boys in their team and consistently underestimating the amount collected by the other team.
Phase 3: Reducing inter-group hostility
The researchers now moved to on to work at reducing hostility between the 2 groups. Their prime strategy for this was to replace the competitive goals with goals that could only be achieved by members of the 2 groups co-operating together.
First the researchers tried simply letting the 2 groups interact on an equal footing in the hope simply associating with each other would, over time, repair the breach. Though outings were planned, movies to be watched together and meals served at the same time, the Rattlers and the Eagles refused to associate. The closest they came to interacting was throwing food and papers – in equal proportion to flying epithets – at one another in the dining hall.
The researchers then arranged for the water supply to break down. (They turned off the valve and then placed 2 large boulders over it, blaming vandals for the problem.) First each group explored the 1.6 km pipeline separately; then they came together at the behest of the researchers and jointly located the source of the problem (a clogged valve). When they restored the water supply, they cheered together. By reducing the situation to a BEIGE survival level, the researchers had ensured the 2 groups would co-operate. However, once the problem had been resolved, the behaviour degenerated again and that evening, another food fight erupted over dinner.
The next tactic was to tell the 2 groups that the camp could not afford to take them to see a film (‘Treasure Island’) most boys had high on their list of preferences. The 2 groups got together and worked out how they could get the money together jointly and see the film. By creating a mutually-owned desire, again the researchers had given the 2 groups common cause.
With each successive task, including preparing and pitching tents together, the antagonism showed signs of mellowing. Familiarity, it seems, was reducing discrimination.
Finally the lorry due to transport their food on an outing to Cedar Lake some distance away wouldn’t start (by arrangement of the researchers) – so the boys got the tug-of-war rope and pulled together to get it to start.
Evaluating the Robber’s Cave study
An in-group preference shown by the boys in each group increased substantially when explicit competition between them was introduced. The introduction of common objectives over a period of days reduced friction equally substantially.
How well these ‘cooling down’ strategies worked was indicated by the boys chosing to travel home on a single bus when offered the opportunity for the 2 groups to travel separately. When a stop was made for refreshments, one group used their last $5 prize money to buy malted milks for all the boys.
Put into a group, the boys developed group identity with group norms, leadership and a status hierarchy.
Competition increased prejudice & discrimination, leading to clear inter-group conflict. This finding led Sherif to develop Realistic Conflict Theory. Working together towards common goals led to much better relations and even something of a superordinate identity.
On a technical level ecological validity was high – it was a field study with the boys in a natural environment.
The study had good internal validity too – the boys were unaware they were being observed.
Population validity could be criticised as the sample did not represent the wider population. Extending this point, Sherif et al’s study can be accused of gender bias – as only boys were involved. Some would argue that, while growing up, girls are rewarded for co-operation while boys are rewarded for competitiveness and that 2 groups of girls might have produced quite different results
Andrew Tyerman & Christopher Spencer (1983) failed to produce the same degree of inter-group conflict in a scout group in the UK. They also found it relatively easy to increase co-operation between 4 different scout patrols – even in the absence of a superordinate goal. Tyerman & Spencer suggested this have been due to the fact that, unlike the boys in Sherif’s study, the scout group already possessed a superordinate goal – ie: being members of the scout movement and subscribing to its values and goals. It also should be noted that a number of the boys already knew each other and were used to competing against each other at their annual camp. They were also encouraged by their leader to see themselves as part of the whole group.
However, Galina Andreeva (1984) found that in-group favouritism and prejudice increased while Russian boys at Pioneer youth camps were engaged in competitive sports. Andreeva also found that prejudice decreased when the boys co-operated when working on agricultural collectives. Her findings strongly support Sherif et al’s findings.
.There are serious ethical issues with the Robber’s Case study. Children were manipulated into developing hostile attitudes towards other children, their parents were not allowed fully-informed consent, the boys were not offered the right to withdraw and they were caused stress (psychological harm). Some boys also experienced pain (physical harm). Sherif and his team justified the study by the sheer amount of insight into the way inter-group tensions develop. They also pointed out that none of the boys suffered any lasting damage, psychological or physical.
In fairness to Sherif et al, the 1954 Robber’s Cave study was the culmination of several lesser summer camp studies done in the preceding years, starting in 1949. Variations included all the boys initially being roomed together and then split up according to friendships made, etc. The 1954 experiment represents the most (startlingly!) successful attempt at inducing inter-group bias and discrimination. However, Sherif et al knew from the earlier studies that the introduction of overriding common objectives would force the two groups to work together and thus reduce prejudice and animosity. (In 1951 Sherif had reduced hostility between 2 groups by introducing a common enemy – effectively foreshadowing Samuel Gaertner et al’s 1993 Common In-group Identity Model)
Although Sherif et al published a substantial account of the study in 1954, the definitive account, which included validation of the data, wasn’t published until 1961.