The Robbers Cave Experiment. Muzafer Sherif
the organism, effects of past experience, and so on. The boundary between the two is the skin of the individual—the skin being on the side of the organism.
It is possible, therefore, to set up situations in which the appraisal or evaluation of a social situation will be reflected in the judgments and perceptions of the individual. In short, under appropriate and relevant conditions, the way the individual sizes up a situation in terms of the whole person he or she is at the time can be tapped through apparently simple perceptual and judgmental reactions.
An additional principle should be clearly stated because of certain conceptions in psychology which imply that perception is almost an altogether arbitrary, subjective affair. If external stimulus situations are well structured in definite objects, forms, persons, and groupings, perception will, on the whole, correspond closely to the stimulus structure. This is not to say that functionally related internal factors do not play a part in the perception of structured situations. That some well-structured situations are singled out by the individual as “figure” rather than others indicates that they do. Such facts are referred to under the concept of perceptual selectivity.
If, on the other hand, the external field is vague, unstructured—in short, allows for alternatives—to that extent the relative weight of internal factors (motives, attitudes) and social factors (suggestion, etc.) will increase. It is for this reason that the exhortations of the demagogue are relatively more effective in situations and circumstances of uncertainty. Since perceptions and judgments are jointly determined by external and internal factors, it is possible to vary the relative weights of these factors in differing combinations, giving rise to corresponding judgmental and perceptual variations. This has been done in various experiments. In a study carried out as part of our research program at the University of Oklahoma, James Thrasher varied the stimulus structure and the nature of interpersonal relations of subjects (strangers and friends) to determine the reciprocal effects of these variations on judgmental reactions. It was found that as the stimulus situation becomes more unstructured, the correspondence between stimulus values and judgment values decreases and the influence of social factors (established friendship ties in this case) increases (Thrasher 1954).
Following the implications of these observations, it is plausible to say that behavior revealing discriminations, perceptions, and evaluations of individuals participating in the interaction process as group members will be determined not only by whatever motivational components and unique personality characteristics each member brings in, nor solely by the properties of external stimulus conditions (social or otherwise). Rather, it will arise as influenced, modified, and even transformed by these features and by the special properties of the interaction process, in which a developing or established state of reciprocities plays no small part. Interaction processes are not voids.
The starting point in our program of research was the experimental production of group norms and their effects on perception and judgment (Sherif 1936). This stems from our concern for experimental verification of one essential feature of any group—a set of norms (feature 4 in the list of small group features above). Groups are not transitory affairs. Regulation of behavior in them is not determined by the immediate social atmosphere alone.
Especially suggestive in the formulation of the problem was F. Thrasher’s observation on small groups that behavior of individual members is regulated in a binding way (both through inner attachment and, in cases of deviation, through correctives applied) by a code or set of norms. Equally provocative in this formulation was Emile Durkheim’s Elementary Forms of Religion, which pointed strongly to the rise of representations collectives in interaction situations and their effect in regulating the experience and outlook of the individual.
After thus delineating the problem, the next step was to devise an experimental situation that lacked objective anchorages or standards (i.e., was vague or unstructured) in order to maximize the effects of the social interaction process. When individuals face such an unstructured stimulus situation they vary markedly in their reactions. However, such marked individual variations will not be found if the stimulus is a definite, structured object like a circle or a human hand. Individuals will agree, on the whole, when they face a circle or a normal hand even if they are 5,000 miles apart and members of different cultures. The fact of objective determination of perception and judgment and the ineffectiveness of social influences (suggestion, etc.) in relation to structured stimuli were clearly noted, in several contexts, in the original report of this experiment. A later publication, in order to stress cases of objective determination of psychological processes, devoted a chapter to the effects of technology and its decisive weight in determining social norms and practices, with numerous illustrations from various parts of the world. Among them, our study conducted in the early 1940s of five Turkish villages with varying degrees of exposure to modern technology dealt specifically with the compelling effects of such differential exposure on judgmental, perceptual, and other psychological processes.6
The experimental situation chosen for the study of norm formation was the autokinetic situation (the apparent movement of a point of light in a lightproof room lacking visible anchorages). The dimension chosen was the extent of movement. As this study is reported in detail in various places, we shall give only the bare essentials.
After first establishing that the judgment of the extent of movement for given brief exposures varies markedly from individual to individual, the experiment brings individuals to the situation to make their judgments together. If, during the course of their participation, their judgments converge within a certain range and toward some modal point, we can say they are converging to a common norm in their judgments of that particular situation. It is possible, however, that this convergence may be due to immediate social pressure to adjust to the judgments spoken aloud by the other participants in the situation. Therefore, going a step further, if it is shown that the individual maintains this common range and modal point in a subsequent session on a different day when alone, then we can say that the common range and modal point have become the individual’s own.
The results substantiate these hunches. When individuals face the same unstable, unstructured situation for the first time together with other participants, a range of judgment and a norm within it are established that are peculiar to that group. After the group range and norm are established, an individual participant facing the same situation alone makes judgments preponderantly in terms of the range and norm brought from the group situation. But convergence of judgments is not as marked as this when individuals first go through individual sessions and then participate in group sessions.
When the individual gives judgments repeatedly in the alone situation, the judgments are distributed within a range and around a modal point peculiar to the individual. This finding has important theoretical implications. The underlying psychological principle, in individual and group situations, is the same, namely, that there is a tendency to reach a standard in either case. Here we part company with Durkheim and other sociologists who maintained a dichotomy between individual and social psychology, restricting the appearance of emergent properties to group situations alone. In both cases, there are emergent properties. In the individual sessions they arise within the more limited frame of reference consisting of the unstructured stimulus situation and special psychological characteristics and states of the individual; whereas in togetherness situations the norm is the product of all of these features within the particular interaction situation. The norm that emerges in group situations is not an average of individual norms. It is an emergent product that cannot be simply extrapolated from individual situations; the properties of the unique interaction process have to be brought into the picture. Therefore, the fact remains that group norms are the products of an interaction process. In the last analysis, no interaction in groups means no standardized and shared norms.
In a subsequent unit, it was found that a characteristic mode of reaction in a given unstructured situation can be produced through the introduction of a prescribed range and norm (Sherif 1937). When one subject is instructed to distribute his or her judgments within a prescribed range and around a modal point, which vary for each naive subject, the preponderant number of judgments by the naive subjects come to fall within the prescribed range and around the modal point introduced for them, a tendency that continues in subsequent alone sessions. This trend is accentuated if the cooperating