Ingroup and Intergroup Relations[1]

Muzafer Sherif and Carolyn W. Sherif

In Chapter 18, we learned that the mere presence of another person watching our actions, or performing the same task beside us, or agreeing to a request, does affect our own experience and behavior. The differential effects of others on our behavior may be manifested in increased self-consciousness or greater absorption in the task, by greater clumsiness or more skill, by a greater inclination to be agreeable or increased obstinacy, depending upon the context of the task, where we are, and the other people involved.

When the other people in a social situation are individuals whom we like or dislike, admire or detest, our behavior in their company becomes more predictable. We are likely to be agreeable in the presence of a friend with whom we expect to have a good time. If we know that he is an excellent tennis player, our own game may suffer as we anticipate his smashing serves, with personal qualms about being able to return them. Such established expectations of another person produce a pattern in the behaviors among individuals.

In group relations, individuals do have mutual expectations of each other. As we have seen, individuals come together because they face some common problem or have some motives that can be handled in concert with others. It may be simply the problem of living in the same dormitory, or the need to have some people who understand one, or the desire for a good time, or the wish to attain social distinction, or to engage in community activities, or to earn a living. Whatever the common problem or motive, if it is conducive to continuing interaction among the individuals, the relationships among them over a period of time assume a pattern, which may be more or less stable. The pattern of behavior among the individuals is referred to as group organization or structure. The existence of such a pattern means that each individual, in some degree, regulates his

( 542) behavior in terms of the expectations he has of the others and the expectations they have of him.

Over a period of time, the individuals develop preferred ways of doing things, common conceptions about the proper and desirable way to treat each other and outsiders, as well as generally agreed on conceptions of what is desired in the way of possessions or personal achievement. They share secrets in common, develop favorite catchwords and jokes, and come to think of themselves as "we"-as "insiders," distinct from those who are not "in." The set of shared agreements on matters of some consequence to members are the norms of the group.

Thus, the earmarks of a human group are the regulation of behavior among a number of individuals in terms of a system of mutual expectations and common agreements on what is desirable, right, and proper in matters of consequence to them.

In this chapter, we discuss the formation and functioning of group structures and, then, the formation of group norms. We will find how useful it is to have a definition of the human group. We need some yardstick to decide when a collection of individuals together becomes a group. The definition of a group directs us to certain aspects of behavior that permit the analysis of group structure and norms and their formation and change. Later in the chapter, the topics include leadership, intergroup relations, and reference groups.


Social relations among individuals may be patterned in many different ways. Any collection of individuals we might pick at random would vary in intelligence, skills, personal appearance, and temperament. If the individuals know each other at all, it is highly probable that some will he liked more than others. The particular aspect of behavior significant in group functioning, however, is not specific individual characteristics, nor is it identical to mutual liking or popularity.

The single, most significant dimension in group structures is the relative effectiveness of individuals in initiating and controlling activities of importance to them all. Effective initiative in group situations reflects the relative power of the members, derived from personal and material resources at their disposal, to determine what will and what will not happen in the group. By observation of interaction in group situations, we find that individuals do differ with respect to the frequency of initiating activities, of making their words "count" in reaching a decision, and in their power to correct other individuals. The frequency with which a given individual successfully initiates and controls interaction in a group defines his status in the group. The pattern of statuses in the group is the most significant aspect of group structure.

Another important way that group interaction is patterned is the extent to which the various members like and prefer each other. Members may also differ in their skills in various activities of importance to the group and in material resources they can provide (for example, money, a car, a place to get together). They may differ in personal styles of behavior, such that one is habitually gay, another gimlet, and a third bumbling and awkward. On the basis of repeated interactions over a period of time in activities of mutual importance, the individuals in a group come to have mutual expectations of each other in any or all these respects. The particular cluster of expectations by others for a given individual in any way pertinent to their activities defines the member's role in the group.

For example, an individual may be intermediate in the status scheme of a group, but he may also be among the most liked companions. He may be known, as well, as a steady fellow to be counted on in a pinch, and a gay contributor to the general merriment at a party. Thus, the roles of members incorporate not only their status in the group, but the particular abilities and personal characteristics they bring into the interaction process as well. Later in the chapter, we discuss one particular status and role in some detail, namely the leader position at the pinnacle

( 543) of the group structure with its associated leadership role.

Equipped with definitions of status and role, we can now examine a concrete study of a group by W. F. Whyte, Street Corner Boys, then continue with experimental findings on group formation.

Street Corner Boys

The context for Whyte's study of Street Corner Boys (1943) was the slum of a large eastern city. This particular area housed a large number of foreign- and American-born Italians. Sociologists have long observed that, although such areas appear disorderly and disorganized, they are particularly fertile grounds for informally organized groups (cf. Thrasher, 192i; Clifford Shaw, 1929). This observation becomes meaningful in terms of the lack of stable social ties with conventional society and the common problems faced by residents.

Whyte moved into this neighborhood and became acquainted as a participant observer with a group of young adult men, most of whom were out of work and were looked down on even in their lowly neighborhood as "little guys." They did not feel comfortable even in the settlement house intended to be a center for residents. They felt that most people who went there felt superior to them. Despite the casual nature of their association while "hanging" on a corner, Whyte found by observing and talking with them over a period of time that their relationships were organized and patterned, even though the individuals involved were not always aware of the fact. It is frequently found in studies of informal groups that some or all of the members feel that there is no differentiation among them. They may say, "We're not a group; we're just a bunch of friends."

Whyte did find a pattern in the associations of the Street Corner Boys. He observed who associated with whom, how often, and with what effect. He noticed who made suggestions, and toward whom the communication flowed. Figure 19-1 illustrates the hierarchical structure of

Figure 19-1. The "Nortons" in Spring and Summer, 1937.

Here is shown the status heirarchy of the informal group studied by Whyte. The members are arranged in terms of their power in the group, and lines are drawn to show the channels through which influence flowed from the leader (Doc) down to the lowliest member. (Whyte, 1943).

Figure 19-1. The "Nortons" in Spring and Summer, 1937

(544) statuses among the Street Corner Boys (the Nortons). He arranged the members in terms of their power in the group and drew lines among them to show the channels through which influence flowed from the leader (Doc) on down to the lowliest member.

Whyte noted that one of the most important characteristics for leadership in this group was toughness. This quality was prized among them and reflects, in part, the context of their setting. They were young males in a neighborhood where a person had to learn to take care of himself.

Whyte reported several instances of the generalization or transfer of general expectations for individuals in different statuses to specific activities. For example, one night Tony began acting up in the presence of Doc, the leader. Doc thought he should cut Tony down to size, but remembered Tony's skill as a boxer and went home to bed. Doc could not sleep. He kept thinking that as a person in his position, he should have stopped Tony. So he arose, dressed, and returned, saving to Tony, "Say that again." Despite the fact that Tony could have beaten Doc (even according to Doc), Tony backed down. Doc's prestige as leader was too much for him to challenge. This is an example of the effect of categorization in interpersonal relationships, mentioned in Chapter 18.

Finally, Whyte describes in some detail how the Street Corner Boys were constantly affected by their neighborhood, as well as themselves contributing to its character by their presence on the corner. When Doc entered a political campaign for a minor office, they backed him to the hilt. He seemed to have a good chance because there were so many candidates in the race. But Doc had troubles. A man in his position should have been able to offer a few treats, to buy tickets to raffles, and so on. He was broke. So, without explaining to the boys, he withdrew from the race. It appeared to them that he must have "sold out" to one of the other candidates, which was unforgivable in view of all of their efforts. From that time, Doc slipped from his high status in the group.

Methods for Studying Group Structure

The accounts of group structure by researchers like Whyte, who spent long periods of intimate contact with members. of a group, are based on a composite of observed incidents of association patterns and effective influence among the members Such composite accounts have face validity reflected in a variety of human episodes.

However, as noted in Chapter 18, there is always the danger that a researcher in the laboratory or in field situations may be selective in what he observes, and that another observer might select differentº actions for emphasis. The safeguard against such research bias is the use of a combination of independent methods for collecting data on which conclusions are based. The problem is not unique to social research.. The student of animal behavior finds that different conclusions may be drawn about learning, depending on the measures of behavior used in an experiment (for ex ample, number of trials to criterion, number of trials to mastery, and trials to extinction). Applied to the study of group structure, the lesson is to ensure the reliability and validity of conclusions by using as many of the following available methods as possible.

Observation is and will continue to be the basic method in the study of human groups. By far the most comprehensive technique consists of complete transcription (by tape recording or stenographic means) of all that transpires during interaction among individuals. Later, however, independent raters or categorizers must: process the transcriptions in terms of the . behaviors of research interest (for example,- makes suggestion," "requests information," "follows suggestion"). The alternatives to complete transcription are to take selected time samples (for example, every; 5 minutes) or to have trained raters classify the behaviors as they occur. Techniques' and categories for analysis of interaction episodes have been developed, notably by Bales (1950).

The major drawback to the most comprehensive methods for observation is that the observer must have the individuals

( 545) in an area of restricted movement (usually one room). Ordinarily, it is not feasible to disguise from them the fact that their every act and word is being studied (barring the use of portable recorders with hidden microphones). If the interest of research is in those forms of human association directly related to individuals' intimate concerns, the restriction of movement and self-awareness of the individuals are barriers to securing valid measures of behavior. Therefore, the complete transcription or obvious rating of behavior is appropriate only for groups whose words and deeds are primarily for public consumption.

Observations can be made without arousing the individuals concern over how they are behaving, if the observer is trained to focus his attention on one aspect of the interaction process at a time, changing his interest as warranted by the purpose of the study. For example, he could first observe who associates with whom, and with what frequency. Then he might shift specifically to observation of who initiates activities, who makes suggestions, whose suggestions are followed over a period of time, and whose words are seldom heeded. After each period of observation, he would then rate the members in terms of the behavior of interest, for example, effective initiative (Sherif and Sherif, 1964).

The possibility remains that the trained single observer is highly selective in what he sees and remembers. Therefore, other methods are warranted to check his possible bias. One such method is to have another observer who is not informed about the group, but must arrive at his own ratings. If his assessment of the status of members agrees with that of the single observer (is significantly correlated with them), we may conclude that both are observing the same aspects of the interaction and reliably so.

Another check on the observations is the "sociometric" interview, pioneered by J. L. Moreno (1934). Depending on the context, sociometric choices from the group members themselves may be obtained in an informal interview or by paper and pencil. The individual is asked whom he prefers to sit by, or work with, or play with. His answers vary with the nature of the question (the criterion), as well as the number of choices he is asked to make and his confidence in the test situation. In the usual sociometric studies, most of the criterion questions concern interpersonal preferences. In studying group formation and intergroup relations, the criteria include queries about each member's observation of effective initiative in group activities ("Who gets things started among you?" "Who sees that things get done, once they have started?") In support of earlier research, it has been found that the individual's preferences for association are not always the same as his perception of the power structure (status) in his group. To the extent that the group is a fairly stable one, their choices of status are closely related to observers' ratings of status. When they are, they support the validity of conclusions drawn from the observers' ratings.


Now, we shall trace the formation of group structure from scratch, starting with individuals who have no interpersonal ties with one another at the beginning. In three different experiments directed by M. Sherif (Sherif and Sherif, 1953; Sherif, White and Harvey, 1955; Sherif et. al., 1961), 12-year-old boys were brought to a summer camp site that had been arranged especially for study purposes. As far as the boys themselves were concerned, they were simply attending a summer camp. However, the camp counsellors, junior counsellors, handymen, and others were all researchers with a definite plan for study. They observed the boys, rated their behavior, and arranged special test situations without the boys knowing that their behavior was being studied.

The figures in this section are based on what happened in all three experiments. For purposes of discussion, however, the concrete findings are selected from the first experiment conducted in Connecticut in 1949. Later in the chapter, results from

(546) the second and third experiments are presented in discussions of expectations among members and of intergroup relations.

You will remember that one essential condition for group formation is a common problem or motive among the members. In the experiments in summer camps, the general common motive was to have a good time in activities that the boys themselves chose as being most attractive to them individually. The prediction was that, over a period of time while interacting toward a series of goals in such activities, the boys would produce a group organization with hierarchical statuses and roles, and would develop common norms in matters of consequence to them, that would be the basis for the individual attitudes of the members.

In order to test this prediction, it was necessary to eliminate the possibility that groups would form among these unacquainted individuals on some basis other than their interactions toward common goals. For example, if there had been differences among them in age, religion, school grade, social class, or ethnic origin, groups might have formed among them along these lines. If any of them had been handicapped or personally disturbed because of unstable situations at home or school, it could have affected the groups that might develop.

Therefore, the subjects were carefully selected from different neighborhoods through interviews with adults and psychological tests. Twenty-four boys were selected who were as similar as possible in educational, religious, ethnic, and family backgrounds. They were all from settled, Protestant families with about average income. All were normal in intelligence and adjusted in home and school.

One additional possibility remained. The boys might form groups, not because of the experimental situations, but because they found common interests not related to the camp, and simply took a liking to one another's personal manner. To eliminate this possibility as much as possible, the first 3 days of the experiment consisted of campwide activities involving all 24 boys, with complete freedom for them to choose their own bunks, seats, buddies, teams, and so on. After 3 days, the observers found that some boys were associating more than others. Small friendship clusters of two, three or four boys were developing simply on the basis of personal preferences. At this time, sociometric choices (as described in the last section) were obtained on the pretext of getting the boys' suggestions about their favorite activities and to improve the camp. The sociometric choices supported the observations of budding friendship clusters.

Now the researchers were in a position to start the study of group formation around common goals planned for the experiment. The first step was to divide the 24 boys into two units, putting each boy in the bunch containing the fewest of his friendship choices on the sociometric test. In each unit, only 35 per cent of the friendship choices were for boys in the same unit, while the remaining 65 per cent were for boys put in the other bunch. In addition, the two units were matched as much as possible in terms of the individuals' size, strength, ability in games, intelligence, and personality test results.

When the separation into two units was announced, it was coupled with the direction for each bunch to choose a bunkhouse and the suggestion that each depart immediately for their own hike and cookout —activities the boys had chosen as most attractive to them. This announcement assuaged some unhappy feelings among boys who were separated from newly-found friends. From this time on, the two bunches of boys were on separate schedules, slept in different bunkhouses, and ate at different times. They were able to choose their own activities, ensuring that there would he strong motivation common to them. However, the timing and facilities in these activities were planned by the researchers so that each bunch of boys faced a series of problems they could solve only by getting together. For example, when they wanted to cook out, they were provided food in bulk form (meat, bread, drink mix). They had to build a fire, divide and apportion the food, and so on.

Now let us see what happened among the boys over a period of time. Figure 19-4 gives a diagram of the two bunches of boys


Figure 19-2. Activities that Require Interdependent Action and Cooperation Lead to Formation of a Group Structure

 Figure 19-2. Activities that Require Interdependent Action and Cooperation Lead to Formation of a Group Structure

In the group formation stage, each collection of boys faced a series of problems that could only be solved through cooperative effort. For example, on "cook-outs" they were provided with food in bulk form, which had to be divided and prepared. Above, we see one of the boys cutting and distributing meat. Below, another boy takes charge of distribution watermelon. (Sherif and Sherif, 1956).

 Figure 19-2. Activities that Require Interdependent Action and Cooperation Lead to Formation of a Group Structure


Figure 19-3. Continued Interdependent Activities Lead to a Stabilization of Status Positions

 Figure 19-3. Continued Interdependent Activities Lead to a Stabilization of Status Positions

Members of both groups engaged in a number of activities of their own choosing, thus ensuring that there would be strong motivation common to them all. Above, We see members of one group carrying their canoe to a hide-out. Below, Members of the other group are on a hike (Sherif and Sherif, 1956).

 Figure 19-3. Continued Interdependent Activities Lead to a Stabilization of Status Positions


Figure 19-4. The Formation of Status Structure in Two Groups over a Period of Time
 Figure 19-4
During early interaction among group members, observers could not agree on any pattern of relationship. As time went on, however, various status positions began to emerge. Top and bottom statuses were the first to stabilize. With further interaction, the middle, or intermediate, status positions stabilized. (After Sherif and Sherif, 1956).

when the study of group formation started (Time a). The circles represent individual boys. Their random locations indicate that, at this time, observers could not agree from one situation to the next on any patterning of relationships among them. Instead, they observed temporary patterns in different activities. For example, in the matter of cooking an outdoor meal, one boy who knew something about outdoor cooking took charge, directing some boys to prepare the fire, asking for help in making hamburgers, and dividing the watermelon himself. But when the boys turned to swimming, or hiking, or ball, another fellow who was particularly apt at organizing and performing such activities was more prominent. Still another boy was the one who took center stage in the bunkhouse at night, when the time was ripe for jokes and pranks. This was a "togetherness situation."

Figure 19-4 represents the same groups after they had been interacting for several days in a variety of common activities. The boys represented by triangles are those whose position in the interaction was stabilized at the time (Time b), as shown by agreement in the observers' ratings. Note that the top status position and the bottom positions are represented as stabilizing first. Both in the experiments and in groups in their natural environments, the leader position typically emerges first, along with positions at the lowest level. Although, as we shall see, the leader status is certainly subject to change in any group, its stabilization indicates that, for the time being, one individual has begun to coordinate and initiate plans in a variety of activities.

(550) Lower status positions stabilize as it becomes evident to other members that one or two persons are not doing their bit, or are usually in the way, or horsing around just when there is crucial business to be done. They may be less interested or less skilled than others.

At Time j, (Fig. 19-4), observers are able to agree on the positions of most of the members over time and in different activities, except those in the middle of the pattern. Again, this finding is rather typical. At the middle level of status, more jockeying for position usually occurs, as the individuals align themselves with those in higher status or team together in some activities.

At Time N, the status structure is represented as stable for the time being. A single observer rates the position of individuals in the same way from day to day. An independent observer who does not know the group rates the pattern in very much the same way. When sociometric questions are posed to the members themselves, their perception of the group structure coincides with the observers, to the extent that the criterion questions reflect the status or power dimension (for example, "Who gets things started?" "Who gets things done once a plan is adopted?").

The process of group formation traced in the figures is intended to be representative. For example, the sequence may be much faster or much slower (in the experiments, it lasted from 5 days to 1 week). Internal frictions among members, an outside threat or emergency, striking changes in the location and facilities, changed activities of interest to members — all of these and other conditions could hasten or retard the emergence of a stable pattern or change it, once it has taken shape. Nevertheless, the sequence in general is typical of group formation among normal individuals.

The particular groups in the 1949 experiment in the summer camp adopted the names of "Bull Dogs" and "Red Devils." Figures 19-5 and 19-6 present the pattern of personal friendship choices within each of the groups when the status structure was stabilized. At that time, it could be concluded that groups had indeed formed on the basis of the series of common problems planned in the experiment. The individuals are arranged in terms of the number o friendship choices each received from his fellows. Choices that were mutual are represented by solid black lines and un reciprocated choices by dash lines. These: lines permit us to specify some of the ways the leaders and their lieutenants wielded influence and power in their groups.

For example, in the Bull Dog sociogram, C was the effective leader of the group and the most popular, acknowledged as such by all members. Note that he and E, who was his chief lieutenant, mutually chose each other as friends. C also reciprocated the choice from H, who was less 3 popular. H is an example of a discrepancy I between status (effective initiative) and j popularity. Even though only four other â members chose him, H was close to C, the leader, who delegated authority to H whenever the group played baseball or other games. H was the "athletic captain." His status was higher than his popularity.

Note, in Figure 19-6 that the Red Devil group presents a different pattern. The most popular member (L) was almost completely subordinate to S, who was the actual " leader of the group. In fact, S was the only person chosen by L as a friend. S, on the other hand, exercised effective initiative through his mutual ties with L, M, and B. These boys, in turn, saw to it that S's suggestions were attended to and translated into action.

Now, we can evaluate the original hypothesis of the study and see how the process of group formation affected the personal preferences of individual members. As predicted, status and role organization a did arise. Each group developed preferred ways of doing things, which differed from ï the ways prevalent in the other group. (For example, each group had its own way of braiding lanyards and preparing meals.) The proper ways of behaving with other members were agreed upon, and there were s correctives (sanctions) for the individual I who got out of line. For example, in the Bull Dog group, a member who cut up, or failed to pitch in when he was needed, was told by the leader to remove a certain


Figure 19-5. Pattern of Friendship Choices in the "Bull Dog" Group at the End of Stage 2.
 Figure 19-5
Mutual friendship choices are shown by solid lines in this sociometric diagram, while one-way choices are shown by dotted lines. (After Sherif and Sherif, 1953).

number of rocks from their swimming hole. The improvement of the swimming place had been a group effort that actually succeeded in raising the water level. Bull Dogs thought that this means of punishment was, therefore, not only beneficial but fairly administered by their leader.

What of the members' interpersonal relations? When it all started, you remember, only 35 per cent of their friendship choices were for other fellows in their own unit. Table 19-1 compares the friendship choices at the time the two units were separated and at the end of the period of group formation. It is important to emphasize that, on both occasions, the boys were asked for their friendship choices in the entire camp. Nevertheless, when the groups had taken shape, 95 per cent of the Red Devils and 87.7 per cent of the Bull Dogs selected fellow members as their best friends. This reversal of the direction of friendship choices attests to the great significance for the individual's personal preferences of interaction with others in a group organization.


Figure 19-6. Pattern of Friendship Choices in the "Red Devil" Group at the End of Stage 2
 Figure 19-6.
Note the difference in group structure shown in the sociometric diagram here, as compared with the structure of the Bull Dog group, shown in Figure 19-5. (After Sherif and Sherif, 1953).

Table 19-1. Choices of friends in experimental formation of groups 



At Time of Division into Two Units Future Red Devils 31.5% 64.9%
Future Bull Dogs 35.0% 65.0%
After Group Formation Red Devils 95.0% 5.0%
Bull Dogs 87.7% 12.3%

(553) Differential Effects of Group Structure in Judgment of Performance

The differential effects of group membership for individual experience and behavior are, as w' have said, referable in part to ties with other people who expect certain things of us and of whom w', in turn, have fairly stable expectations. The stabilization of a group structure implies, on the part of the individual member, categorizations of the positions and capacities of the others. If this is so, the level of the individual's expectations for the others should vary in terms of their positions in the status hierarchy.

In a second experiment on group formation and intergroup relations carried out in upper New York state in 1953, Sherif, Whit', and Harvey (1955) studied the judgments by the members of two groups of the performance by their fellow members in throwing a ball at a target. In this experiment, the two groups had formed with some stability by the end of a week. On' group had already adopted a name, calling itself the Panthers. The other group, which later took the name of Pythons, was not as clearly structured at the time.

Just before the two groups were to participate in a softball game, a practice session in ball-throwing was suggested to each group. Each individual took turns throwing a ball at a large target board 25 times. Both h' and his fellow group members estimated his performance each time, a procedure for which judgments were secured under the pretext of making the practice "more fun."

In Chapter 14, it was emphasized that the individual's motives and attitudes may b' reflected in his perceptions of a situation if the situation is unstructured-that is, if it contains many alternatives in the way it can b' experienced. This principle was utilized, for the purpose of the experiment was to see if the individual's established expectations (attitudes) for his fellow members would h' reflected in his judgment of their performance.

Accordingly, the target board was covered with a plain blue denim cloth. When the ball was thrown at it, it rebounded quickly, leaving no trace of where it had hit. However, under the cloth the board was wired so that the researchers behind it could determine exactly where the ball hit. A light signal showed the score mad', in terms of distance from the center of the target.

When every individual had made judgments of his own performance and that of each other member in his group, it was found that judgments of an individual's performance by his fellow members were significantly related to his status in the group. Performance of members with high status was overestimated by others. The performance by members with low status tended to b' underestimated. The difference between judgment and actual performance by middle status members was intermediate to that for top and lower status's. The extent of the relationship between judgment and status was greatest in the more stable Panther group.

In the less stable Python group, the judgments an individual made of his own performance were not significantly related to those mad' by his fellow members for his performance. However, in the Panther group, the individual member's estimate of his own performance did hear some relationship to the judgments of his fellows. In other words, this experiment shows that expectations for behavior in relevant group activities become stabilized as group structure takes shape. To the extent that the structure is stable, the individual comes to assess his own performance more and more in line with the expectations of his peers. This is an example of the "social" (others' expectations for us) becoming "personal" (my expectations for myself). You will recall the experience of Frank (Street Corner Boys), who actually could not bowl well with the Nortons because they did not expect him to excel.

Expectations for behavior of leaders and followers with differing statuses are discussed on page 560. First, however, we need to specify another property of groups: the norms for behavior shared by members.



As a group structure takes shape, members of a group come to prefer certain distinctive ways of going about their important activities. They may adopt a name for themselves. They set up standards for the ways a member should and should not behave toward his fellows and toward outsiders. Such behaviors reflect the formation of group norms.

Throughout this book, reference has also been made to common conceptions prevailing in a given culture of the male and female roles, of a good parent or a good son or daughter, and of groups of people as superior or inferior on the basis of skin color or ethnic origin. These are examples of social norms maintained by most members of a certain culture, or by sizeable proportions of people within it. At some time during the history of the society, these cultural norms were formed through the interaction of individuals in human groups. By understanding the formation of norms in small groups, we can learn something about the rise of cultural norms and their effects on behavior.

Norm, or social norm, is a general term used to refer to any and all products of social interaction that are shared by participants and that regulate their individual behaviors. Unlike the term norm in, say, study of child development, a social norm does not, necessarily, refer to the average behavior observed in a group. In fact many norms embody what is expected, or even ideal, by those who share them.

Typically, the behavior valued in a norm is not a single way of acting or a single deed. Instead, there is usually a range of behaviors that are considered socially acceptable or desirable, and a range of behaviors that are deemed objectionable. A norm is a concept used by members of a social unit (including groups) in evaluating other people, events, and social objects of consequence to them. We may define a norm as an evaluative scale (or yardstick) defining for individual members of a social unit a latitude of acceptance and a latitude of rejection, which regulate their behaviors in a matter of consequence to them. Not all social behavior is regulated by clear-cut norms, especially in times of rapid change.

As an example, consider the norm regulating behavior toward the Hag. The ideal and most acceptable actions for a man when the flag goes by is to stand, remove his hat, and place it over his heart (or salute). If the scene is a busy street, it is acceptable simply to stand quietly, particularly if one is a passerby. These actions are within the latitude of acceptance, as are other signs of respect. Neglecting to pay any attention to the passing flag may be frowned on by others. Acts of disrespect (throwing mud, sticking out one's tongue, and the like) occasion severe censure and even punishment. They are definitely in the latitude of rejection. This example indicates ways in which we can determine the existence of a norm.

First, a norm can be detected by observing similarities in the behaviors of members in a group (or in a cultural unit) that are not found in very young children or in other groups (or cultures). For example, in the experiments on group formation, one group of boys formed the norm of being very tough and manly. The latitude of acceptance included not crying when hurt, ignoring scratches and bumps, venturing into the woods or rough activities with no sign of fear, and the use of profanity. Members of the other group, which formed separately in the same experiment, not only did not value toughness in this sense, but took an oath not to swear and held regular prayer meetings. In this group, if a boy hurt his leg, it was perfectly acceptable for him to attend to it and show concern.

Second, the existence of a norm is revealed by rewards and punishments (sanctions) for certain kinds of behavior, with the support and approval of other members. Behavior within the latitude of acceptance, especially that approaching the ideal, calls forth praise, appreciation, even gifts from other members. Actions within the latitude of rejection arouse some kind of social pressure from other members, such as reprimands, frowns, or threats. The observation of socially supported sanctions is among the strongest evidence for a norm. The correctives may vary in severity

(555) from "cold shouldering" or a "silent treatment" designed to bring the individual into line, to verbal chastisement, physical punishment, and even death (for example, in a criminal gang).

The third line of evidence for the existence of a norm is the occurrence, over a period of time, of increasing similarity or convergence in the behaviors of interacting individuals who initially behaved in different ways. We shall examine this process in the next section through experiments on norm formation.

Norm Formation

Some norms doubtless form as individuals interact while dealing with a particular range of well structured stimuli. For example, individuals brought into the laboratory to judge a clear-cut series of weights or lines may differ quite a bit in their judgments at first. After some experience with the stimuli, they do, rather quickly, come to agree on their judgments (Tresselt and Volkmann, 1942). Their consensus, in turn, is closely related to the actual values of the weights or lines. In such cases, the resulting norm reflects the compelling differences in the particular set of stimuli they all face, and their interaction plays a relatively minor role in the result.

Many of our conceptions of what is good and bad, and desirable or undesirable, reflect our common exposure to a particular range of stimulation. For example, our assessments of how much income is desirable and how much is inadequate, our appraisals of how fast one should travel and what speed is excessive, and our notions of an acceptable automobile would differ considerably if I had dwelt in a poverty-stricken peasant village and you grew up in a wealthy neighborhood of a modern city. In a country where most women are under 5 ft. 3 in. in height, a sturdy girl of 5 ft. 10 in. would be appraised as much too tall.

However, many situations and events of great significance to individuals in their social life are not as well structured as people's heights, the prevailing standard of living!, or the contrasting speeds of a donkey and a car. For this reason, the experiments on norm formation in the laboratory (M. Sherif, 1936) utilize a highly unstructured situation -judgment of extent of autokinetic movement. This experiment was summarized in Chapter 14; therefore, the procedures need not be repeated. You will recall that autokinetic movement is the apparent movement of a stationary pinpoint of light in a dark room.

By studying the judgments of the distance the light appeared to move, Sherif was able to draw the following conclusions about norm formation in situations that lack objective structure:

1. When an individual repeatedly faces an unstructured stimulus by himself, he develops, in time, a pattern in his judgments of it. He does not give a hodgepodge of different estimates, but comes to distribute them within a specific range and around a mode, both of which are distinctive to him. His individual norm becomes increasingly stable over a period of time, provided no outside event disrupts it. This finding shows that norm formation does not occur solely as a result of exposure to social situations, but is basic in individual psychological functioning as well. Some early writers had treated the individual as a passive prey to social influences and regarded the patterning of his social behavior as the sole result of society's influence on him.

2. When an individual has stabilized an individual norm and is placed in a togetherness situation with others whose individual norms differ from his, over a period of time the individuals' judgments converge toward a common range and mode. This is a social norm formed through individuals' interactions while making judgments aloud. When the individual faces the situation for the first time together with others, the various individuals' judgments approach agreement rapidly, and the social norm continues throughout their experiences together. The social norm need not be a "leveling" of their individual judgments toward some "happy medium." In some cases, the social norm was greater than any of the previous individual norms, and in other cases smaller. Thus, even


Figure 19-7. Norm Formation In Groups of Two Subjects
 Figure 19-7. Norm Formation In Groups of Two Subjects
In this figure are shown medians of judgment in the autokinetic situation. When individual sessions came first (left, above), divergent norms were established. However, subjects who had established an individual norm first, tended to evolve a common norm when in the group situation on later trials. When group sessions came first and individual sessions later (right, above), the group norm continued to influence judgments when the subject served alone. (After Sherif and Sherif, 1956).


Figure 19-8. Norm Formation In Groups of Three Subjects
 Figure 19-8.
In this figure, we see much the same results with three subjects as was shown with two subjects in Figure 19-7. Individuals who established independent norms alone evolved a group norm when others are present. In addition, those who participate in the development of a group norm first are influenced by that norm when later serving alone. (After Sherif and Sherif, 1956).

(558) though these individuals were not related with any strong personal ties, interacting together in an unstructured situation where they were uncertain concerning the "proper'' way of responding resulted in consensus on the "proper way" for that set of individuals. One of the essential properties of a group, as defined in these chapters, had been formed in the laboratory.

3. Finally, after the individual had participated with others in the formation of a norm, he was asked to make judgments by himself. Even when alone, he continued to adhere to the norm established in his particular social situation. This finding is of utmost psychological importance. It shows us that a social product of human interaction (the norm) can become the individual's own personal standard for judgment. Usually, the individuals were not aware that their judgments had been affected by others. His fellow judges, in turn, had made no deliberate effort to affect him (contrary to studies of suggestion or persuasion). Their spoken judgments reciprocally affected each other. This is another example of the difficulty of conceiving of the human individual as an isolated being with characteristic behaviors that appear regardless of that individual's interactions with his fellows. More appropriately, the individual is conceived as a part of the patterns of social interaction in which he moves, reciprocally contributing to the patterns and being affected by them.

In the experimental studies of group formation summarized previously, we saw that norms were formed as the boys interacted in highly attractive activities toward common goals. In fact, this process proceeded simultaneously with the formation of a status-and-role structure among them. Members who did not pitch in to do their part in ways that most of the others thought "proper," or members who behaved in some way not fitting for a Red Devil or a Bull Dog were subjected to correctives from other members (for example, ridicule or scorn). It will be recalled that the Bull Dogs had an established sanction for violators of group custom—removing an appropriate number of stones from their swimming place.

Still, as the groups formed, the observance of norms by the members (behaving within the latitude of acceptance) was most frequently not accompanied by direct social pressure or threat of sanctions from the others. Voluntarily, most members did conform to the group norms most of the time and did their best to be "good" Bull Dogs or Red Devils. The formation of norms during group interaction is accompanied by the formation of attitudes by each individual member. As a result, the behavior of every individual is regulated within the bounds of acceptability, because he himself has formed an attitude defining what is acceptable for him.

In fact, the degree of consensus among individual members of a group on the correctness of the norms they share and their voluntary actions that uphold these norms are among the best indicators of the stability of a group. Thus, the attitudes of individuals belonging to stable and cohesive groups do reflect the norms of their groups, particularly when these individuals had a part in the formation of these norms.

As noted in Chapter 3, much of the socialization process consists of learning the norms of one's groups and of the society in which one lives. But simply learning what is acceptable and desirable in one's social surroundings is not sufficient. Socialization means that the child takes the social definitions of what he is and how he should act as his own personal conceptions. In other words, they do become a basis for his social motives, as discussed in Chapter 6, giving direction to his behavior toward some objectives and away from others.

Leadership in Groups

A leader is an individual who has the highest position in a group in terms of status (power or effective initiative). Leadership refers to his role in the group-the particular skills and qualities he uses, his popularity with other members and so on. These definitions mean that many individuals who are called leaders in real life are not, in a technical sense. The student who is called a campus leader merely because

(559) he gets high grades, the movie star who is called a leader in the entertainment world merely because teenagers copy her styles, the citizen who is called a leader merely because he is prominent in the community-all these may be influential; however, each one is a leader in a technical sense only if he is also the most powerful single individual in a social organization of which he is a member.

Older studies of leadership frequently failed to define leaders as members of organizations, but concentrated on attempting to find what personal qualities and traits distinguished leaders from those who were not leaders. After many years of research, it was evident that leaders differed from each other in about as many ways as other individuals differ (cf. Gibb, 1954; Jenkins, 1947; Stogdill, 1948). When it was possible to compare leaders with other members of their own organizations, it was usually found that in some respect leaders were different or superior to other members. However, the variety of characteristics and skills was bewildering, and varied in terms of the purposes and activities of the groups in question. To take extreme cases, the leader of an athletic group might be a better athlete than others, though not more sociable or intelligent; the leader of a friendship clique may be more sociable, though not as good an athlete as others; the leader of a criminal gang may be more intelligent and less easily frightened, though physically smaller and weaker than others.

Interested for practical reasons in methods for selecting leaders in military units, the military in several countries began studies of so-called leaderless groups. The procedure was to bring several potential candidates in a togetherness situation, to present a variety of tasks for them to perform together, and to see who would show most initiative in their efforts. Since there were no established relations among them, it was found that different individuals frequently come to the fore in different tasks. For example, if the task concerned transporting a delicate instrument over a stream, an individual with some experience, or physical or mechanical ability, probably would take initiative in planning and executing the maneuver. If the task were to discuss some problem and arrive at a policy decision, a person with verbal ability and persuasive manner probably would be prominent.

In real life, groups do have leaders in the technical sense, even when the members do not identify one person as the leader. To the extent that a group structure has taken shape, one person may ordinarily be singled out as more effective than other members in initiating and controlling the activities of others. There are several reasons for this aside from the particular capabilities of the individuals.

First, groups in real life ordinarily specialize in certain types of activities. They interact in matters related to the problems or common motives that bring them together. While the number and kind of activities may be quite large, they infrequently extend beyond those in which the leader has some skill and interest, particularly if the leader attained his position during the interaction of the present membership.

Second, once leaders and other high status members achieve their positions, they are able to discourage activities in which they do not excel or lack interest, or activities that they believe would be harmful to group goals and maintenance. For example, Whyte observed that the top status members in the Street Corner Boys successfully vetoed activities and, in one instance, diverted most of the members from associating with a girls' club, when it seemed that their own group was suffering from excessive interest in the girls.

Finally, part of the leadership function in most groups is the delegation of authority to other members. Thus, a leader who does not excel in sports or party-planning delegates duties related to these activities to other members more able in these respects. In order to determine whether there is a leader in a group or whether a different patterning occurs in certain activities, a group must be studied over a period of time in a variety of situations. In studies of groups of adolescents in their natural surroundings, the present authors have

( 560) found that informal groups typically do have a leader, whose absence or disapproval exerts considerable effect on other members (Sherif and Sherif, 1964).

The leader is the single most powerful member of a group, but as a member, he is not immune to the opinions of others and to the group norms. The power of a group on a leader and the importance of its norms ,were strikingly shown in a study of children's groups by Merei (1949). From preschool classrooms, units of three or four children were chosen who had not been particularly outstanding in taking initiative in their classrooms. Each unit of children was taken regularly to a special playroom and allowed to play together until observers began to note the formation of certain norms and a pattern to their interaction. For example, they developed certain genes that they played regularly in distinctive ways, with one child doing one thing and a second another in traditional fashion. Into each of these incipient little groups, they introduced a new child. The new child had been selected because lie (or she) was slightly older than the others, and had shown considerable initiative as a leader in his own classroom. The question was: What would happen in the group when the new child appeared?

Merei found that in every group, the new child was at first absorbed by the group. He (or she) quickly found that if lie did not follow the rules of the games and the little customs, no one paid attention to him. Some children tried to suggest new games or suggest new ways of playing the old, but at first, none was successful. Over a period of time, several things happened, depending on the group and the new child. Some of the newcomers completely adopted the norms of the groups, and never tried to exert initiative themselves. Other newcomers adopted the norms of the group, but took initiative by coordinating the already established scheme of things. For example, the newcomer would give orders for activities that were traditional, or make suggestions about the timing of things already engaged in before. Still others adopted the role of guardian of the toys, never interfering in what was done with them, but determining how they would be distributed.

The only newcomers in the Merei study who succeeded in changing the customs (norms) of the group were those children who first adopted the norms, and then introduced small variations in them. For example, one group had the game of climbing up a wardrobe and swinging on the doors. One successful newcomer suggested that they line up on one side, climb one at a time, and swing down the other side, permitting a steady succession of participants. Such variations in the customs, in turn, were followed by the introduction of new activities and changes in customs.

In summary, the leader does have power in the group, but the other members and their joint norms have power over him.

Member Expectations and Group Norms. In summarizing the study of Street Corner Boys, Doers downfall as leader of the Norton's was related. When a leader "lets down" the other members in a matter of major importance to them-fails to meet their expectations-he is likely to lose status severely.

What do members expect of their leader? The answer varies in terms of the group's major concerns and activities. But one conclusion is warranted by research. Whatever the group's major activities and goals, in these matters the leader is expected to go out of his way to contribute to their success, to coordinate activities to promote their success, and in his own behavior to set an example of what it means lo be a good member in these matters. In most groups, the leader is expected to exemplify the interests of other members in relation to outsiders-to promote their interests as members and to avoid putting them or group secrets into any jeopardy. Thus, in matters related to the major goals and maintenance of the group, the latitude of acceptance for the leader's behavior is more limited than for other members. Their expectations of him are high.

On the other hand, in strictly in-group relations and matters of less importance, the leader frequently has greater leeway than others. In the Sherifs' studies of natural groups in their settings, leaders of adoles-

(561) -cent groups were observed to violate minor customs of their groups with impunity, whereas lower status members were corrected for similar actions. In one group, the leader made flagrant fouls in ball games with other members of his group. Yet, he reprimanded a lower status boy for similar fouls, and almost prevented him from playing in" a game with another group, on the grounds that the offender would give their Group a bad reputation. Though he fouled while playing with fellow members, the leader strictly observed the rules in intergroup competition to ensure the good name of his group.

The solidarity or cohesiveness of a group is, of course, related to the intensity of the motives that brought the members together. The leader has a considerable part in promoting solidarity, since the coordination of activities and member relations is a function of the leadership role. High solidarity results over a period of time, when most members of a group interact toward goals of some importance to the group, and consistently meet or weed each others' expectations. The general good feelings and esprit de corps that result are usually called high morale. We refer to solidarity and morale later in the chapter when discussing intergroup relations.

Like the leader position, other parts of a group organization can be affected and changed when members fail to meet established expectations for their status and for the norms of a group. Changes in status structure occur most frequently when a member fails to meet expectations, or when the external situation and major activities of the group change.

Leadership Styles. The particular manner and procedure that a leader adopts also varies with the purposes of the group and its major norms, in addition to his idiosyncrasies and persuasive skills. It also varies with the structure or organization of the group. By reference to Figures 19-5 and 19-6, you might recall the differences in the organization of the Bull Dogs and Red Devils. The two leaders also differed in style, or leadership roles. C was a quieter boy, congenial and friendly. S was a larger, handsome blond with an air of adventure. Both leaders made friendship choices of only two or three other boys in their groups, each of which vas reciprocated. Each worked with and through their mutual friends in initiating and controlling interaction and activities. There the similarities in role behavior diminished.

C, leader of the Bull Dog group, was highly skilled in bringing all members, even those in the lowest status positions, into group activities in some capacity. He frequently delegated the immediate conduct of an activity to his lieutenants, and he encouraged and helped others. The Bull Dog group was well coordinated and its members, most of the time, were quite satisfied with their achievements.

S, leader of the Red Devil group, was exclusive in his associations. He was admired by all the members for his stature, skills, and daring; but his exclusive attachments to L, B, and M were at times resented by other members. Typically, S might tell one of his pals (especially L, who was extremely popular) what was going to happen, or give an order to a low status member. If things went wrong, he was inclined to blame the others rather than take responsibility himself. With his prestige, he was often joined by other members in casting blame, with the result that bickering was frequent in the group. Probably as a result of the bickering in the lower ranks, the Red Devils were less cohesive as a group.

The first studies of leadership styles were performed in the late 1930's by Lewin, Lippitt, and White (1952). These early experiments on social interaction dealt with the treatment by adult supervisors of boys in handicraft clubs initiated for purposes of the research. Therefore, the generalizations drawn from them are probably more appropriate for adult-supervised classroom or recreational situations than for other groups.

In the experiment, the role assumed by the adult supervisor was varied, as well as the particular adults who filled the roles. In one style, the adult laid down all the policies, procedures, and techniques for performing the handicrafts, assigning the boys to tasks and partners (authoritarian

( 562) style). In another, he encouraged the boys to discuss how to go about the tasks, giving encouragement and information when needed, suggesting alternative technical procedures, but allowing the boys to choose work partners and divide responsibilities (democratic style). In a third, he made it clear what materials were available and his willingness to give advice if needed, but otherwise took very little initiative or made few suggestions (laissez-faire or hands-off style).

The effects of the adult supervisor's style were marked in terms of what happened in the clubs. Lippitt and White (1963) reported: "Expressions of irritability and aggressiveness toward fellow members occurred more frequently in both the authoritarian ... and laissez-faire situation than in the democratic social climates... . There were more requests for attention and approval from fellow club members to each other in the democratic and laissez-faire situations than in the two authoritarian climates. . . . It is interesting to find nearly as high a level of interpersonal friendliness in the authoritarian situations as in the democratic and laissez-faire atmospheres. ...Intermember suggestions for group action and group policy were significantly lower in both types of autocracy than in the laissez-faire and democratic atmosphere."

The researchers also reported significant differences in reactions to the different supervisory styles according to which style was experienced first, and the personal history of the boys (an army officer's son preferred the authoritarian supervisor). In some groups, there were consistent differences in behavior that seemed unrelated to the supervisor's style. One group that started with a hands-off supervisor contained three boys who were already friends and who engaged in horseplay and were rowdy throughout subsequent changes in supervisory style. Such factors are doubtless among the reasons that subsequent studies, most of them attempting to duplicate the research in classroom situations, have produced rather confusing and contradictory conclusions. In reviewing 49 experiments comparing authoritarian and democratic styles (which may have varied from study to study), Anderson (1959) concluded, "studies based on two a priori styles have not led to consistent or easily interpretable results."

In order to learn more about leadership styles, researchers must also consider factors such as the following: (1) the extent to which a group structure has stabilized, its pattern and solidarity; (2) the tasks being performed by the group, especially in terms of the common motives that bring the members together and the individual's abilities to perform the tasks; (3) the norms of the group (for example, whether defiance and horseplay, or serious concentration on the task at hand are acceptable modes of behavior); and (4) the personal backgrounds of members as these relate to the structure, norms, and the activities of the group.



Groups in actual social life are seldom isolated from other groups of people. In the following section, we summarize experiments on intergroup relations that were concerned with the rise of hostile attitudes, unfavorable stereotypes, and their change.

Intergroup relations is the term that refers to the state of affairs between two or more groups, whether these are reciprocally friendly or conflicting. When individuals from one reference group interact with those from another group in terms of their group identifications, we have an instance of intergroup behavior. If the individuals have established attitudes toward the other group, it is convenient to refer to them as intergroup attitudes, although this does not imply they are formed in different ways than any other attitude.

We have already seen that during the process of group formation, the members delineate themselves as an ingroup, referring to themselves as "we," and to nonmembers as "they" (outgroups). Individual members tend to endow their own group and its members with positive and praiseworthy qualities. These qualities are products of participating in group activities toward common goals. The "good"

(563) members, who constitute the majority of a group as long as group solidarity is maintained, adhere to the norms and endorse these qualities, which reflect the group's particular kind of ethnocentrism.

Similarly, while interacting with fellow members, group members come to attribute certain qualities to other groups and their members. Their views of other groups may be favorable or unfavorable, depending on the positive or negative state of their relations with the other group. If relations between their groups are harmonious and complementary, most individuals in the groups have favorable intergroup attitudes. If the two groups are in conflict, if one suppresses or impedes the other, members within the group have unfriendly attitudes and unfavorable stereotypes of the other group.

Earlier in this chapter, the process of group formation in three experiments in summer camps was described. The major aim of these experiments was to study intergroup relations and, in particular, to produce hostile attitudes and unfavorable stereotypes between members of two groups, in order to assess methods for changing them in a favorable direction. This was another reason why such care was taken to select individuals who were unacquainted with one another.

It was necessary to eliminate any prior bases for friendly or hostile relations between groups once they had formed in the experiment. By selecting boys similar in background and characteristics (healthy, average or above in intelligence, well adjusted members of stable middle-class, Protestant families, and adjusted in school, free of signs of personal disturbance or home conditions conducive to them), explanations of their behavior in the experiment could not rely on pre-existing differences in these respects.

You will recall that the boys in these experiments were not aware they were being studied and observed for research purposes. Activities in the camp were chosen on the basis of the boys' expressed interests before and during the experiment. As much as possible, within the experimental design, they chose their own activities and made their own schedules. At every step, a combination of methods was used to ensure the reliability and validity of major findings and conclusions.

The experiments started with the stage of group formation, for it was necessary to have groups in order to study intergroup behavior. As predicted, two groups did form in each experiment. During this stage, the two groups had no contacts with each other. In fact, in the 1954 experiment, the boys assigned to each unit came to camp on separate buses and did not know of the presence of another unit in camp until the second stage of the experiment. The 1954 experiment was carried out at Robbers Cave in Oklahoma, an area containing a hideout reputed to have been used by Jesse James and Belle Starr. These groups called themselves the "Rattlers" and the "Eagles."

Intergroup Conflict in Stage 2. The prediction to be tested was that intergroup conflict and hostility, along with unfavorable intergroup attitudes, would develop even among similar, normal individuals when their group goals were conflicting, such that one group could achieve its aims only at the expense of the other group. As American boys, these boys had an abiding interest in competitive sports. This cultural value provided a natural means of creating an experimental condition with mutually exclusive goals. However, the boys also upheld the cultural norm for good sportsmanship. Adherence to this value would have to dissipate if the experimental plan was to be effective.

Since the boys wanted to compete with another team, the staff complied with their wishes by arranging a tournament of events and supplied handsome prizes (for example, a trophy to the winning group and knives to each member). Various sports (baseball, touch football, tug of war), games (a treasure hunt), and camp skills were scheduled. The latter were included so that the cumulative scores could be manipulated somewhat, to keep each group's score within sight of the other's.

The tournaments, in each experiment, started in a spirit of good sportsmanship, the winners of each even living a cheer for the losers and vice versa ("Two, four, six,

(564) eight, who do we appreciate, Bull Dogs, Bull Dogs, Bull Dogs!"). As event followed event, with the score see-sawing back and forth, sportsmanlike feelings began to fade (The cheer became: "Two, four, six, eight, who do we appreci-hate....").

The boys in the Red Devils and Bull Dogs (1949 study) turned against buddies they had chosen as "best friends" when they first arrived. The rival groups raided each other's cabins, made threatening posters, and collected secret hoards of green apples "in case they were needed" as ammunition.

In the Robbers Cave study, the Eagles, after defeat in a baseball game, burned a banner left behind by the Rattlers. The next morning the Rattlers seized the Eagles' flag. From that time, name-calling, scufles, and raids on each other's property were ordinary events. As the tournament progressed, each group began to explain its losses in terms of the other's bad character or events over which they had no control. ("The ground was against us." "They must have greased the rope." "They're cheats." "They're sneaky." "Stinkers!")

The results of the stage of intergroup conflict may be summarized as follows:

1. When the two groups, whose members were not unfriendly initially, came together in a series of activities where one could achieve its goal only at the expense of the other group, in time, the rivalry changed to open conflict between them.

2. Members of each group developed hostile and unfriendly attitudes toward the other group and its members. When asked to rate the members of the other groups on several traits (such as "brave," "sneaky," "cheats"), the boys in each group gave extremely negative ratings ("all of them are cheats"), indicating the rudiments of intergroup stereotypes.

3. By the end of the period, social distance had developed between the two groups, such that neither wanted to have anything to do with the other, nor to be in any activities at the same time and place.

4. The effects of intergroup conflict within each of the groups were as follows:

A. Especially in the early days. when the group's main activities changed from peaceful pursuit of outdoor camping activities to intergroup competition, then to hostile rivalry, there were changes in the status structure of several groups in these experiments. For example, the leader of one had been very skillful at coordinating activities before, but shrank from intergroup encounters. I-le walked away from one of the contests and avoided physical conflict. He was rapidly replaced by another boy who rallied the group. In another group, a large boy who had been "put in his place" because lie roughed up smaller boys, became a hero in intergroup conflict for his prowess in raids and scuffles.

B. Each group had new norms added to their repertory. These norms largely pertained to treatment of the other group, forbidding expression of kind thoughts about the other, ruling out contact between them, and regulating how one should behave in their presence.

C. The net effect of intergroup conflict was to increase solidarity, morale, and cooperativeness among members within each group. While rating the other group negatively, each group rated his own in a strongly positive way. The heightening of cooperativeness and democratic planning within each group did not, however, transfer to relations with the other group.

In Chapter 14, a part of the Robbers Cave study was reviewed, showing how members of the Rattlers and Eagles glorified their own groups, as reflected in their great overestimations of their own performance, and their tendency to underplay the achievements of the other group.

These results show that intergroup relations and intergroup behavior of members are not simply a transfer of the relations and norms within the group to outsiders. Intergroup relations have "a life of their own," in the sense that they are not entirely dependent on relations within groups. They do affect relations within the groups.

Intergroup Cooperation in Stage 3. The purpose in producing intergroup conflict, hostile attitudes, and stereotypes was to investigate their reduction and ways of producing cooperation between hostile groups. Various measures could have been

(565) tried. In the 1949 study, a common enemy was introduced in the form of a team from outside the camp playing against a campwide team. While conflict between the camp groups was reduced as a result, the logical outcome, had the common enemy been continued, would have been intergroup conflict on an even larger scale.

Another possibility would have been an information campaign. However, a large body of research has shown the limited effectiveness of information that runs contrary_ to the main concerns of a group. In the Robbers Cave study, a Sunday sermon by a visiting preacher on brotherhood was listened to attentively by the boys, who then proceeded to return to their vituperative discussions about each other.

A third possible method would have been individual competitions cutting across group lines, for example, in individual sports. However, our interest was in producing intergroup cooperation, rather than simply breaking up the groups as this practice tends to do.

A fourth possibility would have been meetings between the leaders of the two groups. There were several difficulties and drawbacks in this. In the first place, the groups did not want to have anything to do with one another. In the second place, in this state of conflict, the leaders would have had to go beyond the limits for good members of their own groups, in order to have effected any change in the situation. In the 1949 study, one boy who took it upon himself to make a "peace mission" to the other group was greeted with a hail of green apples. Although he was high in the status hierarchy of his own group, he was strongly reprimanded by the other members when he returned for doing such a foolish thing.

The experimental measures actually used in the 1954 Robbers Cave study were planned to clarify conditions in which contacts as equals between hostile groups may be effective in reducing conflict between them and changing member attitudes. These measures were contact in activities extremely pleasant to each group but not involving interdependence between them, and contact in situations where each group strongly desires to reach a goal that can only be achieved through the efforts and resources of both groups-superordinate goals.

1. The Rattlers and Eagles were brought into a series of situations together in which each wanted to participate, though not together. These included eating in the same dining room, attending a movie, and shooting off fireworks (it being July 4). Far from reducing conflict, these situations gave further opportunity for the two groups to embarrass and degrade one another. In the dining hall, they shoved for first place in line, the group losing the contest shouting, "Ladies first!" They threw paper, food, and bad names at each other. An Eagle who bumped into a Rattler was cautioned by his fellow Eagles to "brush the dirt off" his clothes. Since it was apparent that intergroup hostility was increasing, not decreasing, it was concluded that contiguity in activities that were themselves rewarding for each group separately vas not sufficient to change hostile intergroup attitudes.

2. The measure of superordinate goals was suggested by what we found to be a sufficient condition for the appearance of intergroup conflicts. If conflict develops from mutually incompatible ,goals, common goals should promote cooperation. But what kind of common goals? They would have to override each groups interest in maintaining the hostility. We predicted that a series of superordinate goals-that is, goals with compelling appeal for members of each group that neither group can achieve without the other-would, over a period of time, result in cooperation between the groups and a lessening of intergroup hostility.

A series of urgent situations was created that seemed natural in the camp setting. One was an apparent breakdown in the water supply. Water came to camp in pipes from a tank about a mile distant. The flow of water was interrupted to a trickle. The groups were called together and informed of an impending crisis. By nightfall the camp would be without water. Both groups promptly volunteered to search the line for trouble. The two groups divided into teams composed of their own members.

(566) They explored separately, then came together at the water tank, and jointly located the source of the difficulty. While there was rejoicing at this common victory, the groups returned to their old recriminations when they came to camp.

When the boys requested another movie, they were told that the camp could not afford to rent one, but could manage with some contribution from them. The two groups got together, calculated how much each group would have to contribute, chose the film by common vote, and enjoyed the movie together.

Then the two groups were taken on a trip to a lake some miles away. A large truck was to go for food. But the truck would not start (as planned), just when everyone was very hungry and ready to eat. The boys got a rope-the same rope that they had used during the tournament for a tug of war-and pulled together to start the truck.

These cooperative efforts did not immediately dispel their hostility. But gradually, members of the groups became more friendly. They stopped shoving each other in the meal line. They began to sit together at tables. One Rattler whom the Eagles particularly disliked became a good egg. Procedures that "worked" in one joint activity were transferred to others. For example, the idea of taking turns in the dinner line was transferred to a joint campfire, which they decided themselves to hold. They took turns presenting skits and songs, to the mutual enjoyment of all.

At the end of camp, both groups requested ' that they go home together on the same bus. No one paid attention to a few die-hards who muttered "Let's not." On the way home, the bus stopped for refreshments. One group had five dollars won as a prize at camp. They spent this sum on refreshments for both groups, rather than using it so they could have more to eat.

Interviews with the boys confirmed the change in their attitudes. Figure 19-9 shows their choices for best friends in the stage of intergroup conflict and after the series of superordinate goals. Very few choices were for members outside one's own group during conflict, but after cooperation had developed, new friendships involving members of both groups rose significantly.

They were glad to have the opportunity to rate the boys in the other group as to their personal qualities again, some remarking that they had changed their minds since the ratings during the tournament. Figure 19-10 compares the percentage of ratings by each group of the other group, which were completely negative ("all of them are..." sneaks, cheats, etc.) during the tournament, and then after the series of superordinate goals. The ratings of the other group changed from predominantly negative to la gels favorable ratings.

Figure 19-9.  Figure 19-9. Friendship Choices During Intergroup Conflict and Following the Introduction of Superordinate Goals
Left, During conflict between the two groups, in the Robbers Cave experiment, there were few friendship choices between cabins. Right, After cooperation toward common goals had restored good feelings, however, the number of friendship choices between groups rose significantly. (After Sherif and Sherif, 1956).


Figure 19-10. Per Cent of Negative Ratings of Each Group by the Other During Conflict and Following the Introduction of Subordinate Goals Figure 19-10. 
Left, we can see that negative ratings of each group by the other were common during the conflict phase of the experiment. Right, They were less so following the reduction of tension. The graphs show the per cent in each group who thought that all (rather than some or none) of the other group were "cheaters," "sneaks," and so on. (After Sherif and Sherif, 1956).

 The results of this stage led to the following conclusions:

1. Contact between hostile groups as equals in contiguous and pleasant situations does not, in itself, reduce conflict between them.

2. Contact between groups requiring interdependent actions toward superordinate goals is conducive to cooperation between hostile groups, but single episodes of cooperation are not sufficient to reduce established unfavorable attitudes and stereotypes.

3. A series of superordinate goals has a cumulative effect in reducing intergroup hostility. The cumulative effect involves the successful development of procedures for cooperation and their transfer to new situations.

4. Cooperative endeavor toward superordinate goals changes the significance of other measures to reduce hostility between them. Casual social contacts were used now to develop plans, make decisions, and have pleasant exchanges. Information about the other group and its members became interesting and was listened to, rather than ignored or reinterpreted to fit their unfavorable stereotypes. Individual members of each group could participate with each other without being "traitors" to the group. The leaders of each group found that they could take positive steps toward working out procedures for joint endeavors and future contacts.

In other words, various methods used with limited success in reducing intergroup hostility become more effective when groups have started to cooperate together toward goals that are genuinely appealing to all, but that require equal participation and contribution from all.


In modern complex societies such as the United States, individuals seldom belong or relate themselves to just one group. At the same time, an individual may be a member of a family, leader in a friendship group, student at a school, member of a church, member of an ethnic group, or resident of a community, state, region, and country. Each of these social units has certain norms, and the individual may have different status and roles in each. To the

( 568) extent that he relates himself to them psychologically as a member of these various groups, they are his reference groups.

If his ties with each of these groups and their norms are complementary and harmonious, the human individual is capable (if moving from situation to situation, behaving appropriately within the latitude of acceptance and with all experience of personal satisfaction. However, in complex societies in the process of rapid technological, political, and social change, such is frequently not the case.

For example, in Chapter 6, the prevailing social distances between various ethnic and national groups in the country were described. The individual belonging to an ethnic or racial group that has been traditionally discriminated against feels, in the changing scene today, considerable difficulty in relating himself at the same time to his ethnic group and to a larger society that condones discriminatory treatment and prejudice. Thus, especially after the Supreme Court decision of 1954 desegregating public schools and opening possibilities for further opportunities denied the American Negroes, many Negro leaders and eventually millions of Negro citizens began to feel with great urgency the need to adjust the values of their society and their own aspirations into a harmonious pattern.

Conversely, some white Americans found themselves caught in situations of great stress because they were loyal members of communities where discrimination had been the rule, but at the same time considered themselves loyal Americans. Some conceived the social distance scale and discriminatory practice as norms included in that loyalty. As the national government began to support changes in the pattern, many were caught between loyalty to local and regional groups on the one hand, and the changing norms for loyalty to the nation. Campbell and Pettigrew (1959), for example, gave an account based on interviews with the ministers of Little Rock, Arkansas, after the crisis there over desegregation of public schools. Caught between their identifications as Americans and men of God who taught brotherhood on the one hand, and the knowledge of the norms of their congregations and community on the other, the great majority of the ministers kept silent during the strife. Only a few followed the dictates of their conscience as men of God or the clear rulings of the federal government in preference to local commitments.

The period of adolescence in our society, roughly the teenage years, offers a particularly vivid illustration of the conflicting - norms of different reference groups. Some of the problems faced by the youth in our society are built into the very nature of the changing status of the individual from childhood to adulthood. For several years, the youth is literally betwixt and between, neither child nor adult, boy nor man, girl nor woman, neither wholly dependent on his family_ nor wholly independent. At the same time, his changing body brings the experience of new and more powerful urges.

Adults expect at this time that he will be "weaned" psychologically from dependence on adult figures; but they provide few clear-cut formulae for the process. Prompted by newly developing urges and his developing adult body, the youth is more impatient than adults who have something to say about it. The dilemmas he faces lead him into active seeking for social ties with those who can understand him. Often, he finds that those who can understand him are his age-mates. Like any other individual feeling shake in his social ties, lie turns to them to seek ways in concert to assert and satisfy their desires.

In a thesis at the University of Oklahoma, William Prado (1958) showed that a shift in evaluation of parents and age-mates occurred even in upper middle-class families in which the father was the favored parent. He brought boys, their best friends, and their fathers to perform a simple game whose outcome was subject to conjecture (unstructured). He asked the boys to estimate the performance of their friends and of their fathers (in each case, their favorite parent). Preadolescent boys almost unanimously assessed their father's performance as better than their friends. Adolescent boys

(569) just as overwhelmingly appraised their friends as superior, even though their friends' actual achievements were not better than their fathers'.

Recent research on adolescent boys (13 to 18 years) in several large cities of the southwestern United States (Sherif and Sherif, 1964) focused upon the common and conflicting values of their reference groups, as well as the structure of their face-to-face friendship groups. By comparing the attitudes and personal goals of youth living in different areas of these cities, it was found that there were some values common to all of them. For example, youth in low, middle and upper class neighborhoods from Negro and Spanish-speaking groups all desired the good things of life -exemplified by automobiles, television, telephones, clothing, appliances, and so on. However, their self-radius for achievement in other respects differed considerably.

For example, in upper rank areas, youth considered graduation from college as the norm-as defining an educated person, as what a "person like me" needs to achieve. There was less consensus in the low rank areas of these large cities about schooling. The expectation for oneself was graduation from high school for about half of the youth, and over 40 per cent expressed the desire to go to college.

Figure 19-11 makes these differences in norms and iii the self-radius of one's aspirations in different levels of society very concrete. It compares the latitudes of acceptance and the goals for weekly income for youth in low, middle, and high rank areas of a southwestern city with many Spanish-speaking residents. The lower bars represent the average conceived as the minimum weekly income necessary to live. The upper bars show the average signifying prosperity ("really well off"), and the triangles show the average level of personal goals.

The letters along the baseline indicate the cultural composition of the samples as A ("Anglo") or L ("Latin"), the middle rank sample of youth being divided accordingly since it was mixed. Comparable data have been collected in other cities. They show that youth in different areas do have differing conceptions of prosperity and of self-achievement. Note that relative to their own conceptions of comfort, the Spanish-speaking youth are most "ambitious." This is especially true when these figures are compared to their own family's income.

Within these areas of large cities, the youth in all areas associate in age-mate groups that ordinarily are well structured little groups composed of from 3 to 15 or so. How important in these different areas are these small, face-to-face groups for the individual members, as compared to school, family, and other reference groups?

A recent study by Walter L. Slocum (1963) of youth from unbroken homes in six rather small communities in the state of Washington provides a good background for the discussion. Comparing the families in terms of family democracy, affection among family members, cooperation among family members, and fairness of discipline, Slocum found that participation in school activities was related to these conditions, although scholastic achievement was only slightly related. On the other hand, he found only weak association between the family norms, as described by these factors, and the youth's activities outside of school. He concluded that the data suggested that peer group standards and expectations may in many cases be more important and more powerful influences than family standards and expectations."

Our studies of natural groups in larger urban areas (Sherif and Sherif, 1964) accentuate this conclusion. Age-mate reference groups were major sources for personal satisfaction, for conceptions of what is proper and not proper and what is to he desired. Frequently, what was proper and what was desired was to engage in activities that proved one as a maturing adult, and many of these conflict with what parents or other adults deem proper for the age level. However, whether or not activities were considered by adults as undesirable, most of the boys were absorbed in their age-mate groups.

In upper and middle class neighborhoods, the youth belonged to a variety of groups in school, church, and community.


Figure 19-11. Reference Scales for Evaluating Income in Areas of Different Social Rank.
Figure 19-11
In this figure, we see median estimates in high (I), middle (II), and low (III)  rank neighborhoods of minimum income for survival (squares), for comfort (circles), and personal goals (triangles). S = bilinguals. (After Sherif and Sherif, 1964).

Usually, there was a core of most frequent associates cutting across these; but each individual did have many reference groups. In lower class neighborhoods, some boys and girls also attended school, belonged to clubs, and were congenial with their families. But there were also youngsters who had unstable family ties, who had dropped out of school, and who were affiliated with no recreational groups. These were the boys most absorbed in their age-mate groups. Not all of them, however, were accepted as members. As noted earlier, to he a good member of a group, one has to live up to expectations of others, to prove oneself trustworthy and reliable, and to abide by the group norms.

Some of the groups engaged in delinquent activities. In delinquent groups, those to whom the group was most important, those most devoted to its norms, those most loyal, responsible, and consistent in their dealings with their fellows-these individuals were most likely to be the worst offenders in terms of the norms of their communities and larger societies. Severely disturbed individuals were not accepted for long in these groups. They soon

(571) proved themselves unsteady, unreliable, and a danger to others, especially since they were more likely to be caught.

Indeed, it may be generally the case that individuals who are caught betwixt and between the norms of conflicting reference groups, or the obligations of their different roles, are personally disturbed. Conversely, in many cases of behavior that deviate from the latitude of acceptance prevailing in society, the individual is acting as a loyal, trustworthy member of a reference group whose norms conflict with those of a larger social unit. Thus, the individual's conformity and his personal stability are better diagnosed, for purposes of change, when related to his reference group ties.


1. The earmarks of a human group are the regulation of behavior among a number of individuals in terms of a system of mutual expectations and common agreements on what is desirable, right, and proper in matters of consequence to them. The definition of a group, encompassing this idea, directs us to certain aspects of behavior that allow analysis of group structure and norms and their formation and change.

2. All social groups have a structure or organization, and the most significant dimension in group structures is the relative effectiveness of individuals in initiating and controlling activities of importance to them all. The frequency with which a given individual successfully initiates and controls interaction in a group defines his status in the group.

3. Group structures may be studied by spending long periods with group members and observing incidents of association patterns and effective influence among the members. Such observation is and will continue to be the basic method in the study of human groups. Other methods, however, are also used to assess structure. The sociometric interview, in which individuals state their preferences for various other group members, is widely used and provides a check on the validity of observational data. Another method of studying group structures is to experimentally study groups in the process of formation.

4. In several experimental studies, Sherif has found that one essential condition for group formation is a common problem or motive among the members that requires their interaction. In these experiments, when groups of boys who were strangers to one another were brought to a camp, their interaction to satisfy common motives resulted in the formation of group structures. Status and role organization arose under these circumstances, and each group developed preferred ways of doing things (norms), and correctives (sanctions) for individuals who got out of line.

5. Once structures emerged in these groups, there were differential effects of the group structures in the judgment of performance. When presented with an unstructured situation in which each member of the group performed and was judged by others, expectations of the individual's performance clearly affected the judgment of his performance by other members of the group.

6. Along with status and role relations that made up the structure of groups, social norms also emerged. Social norms are products of social interaction that are shared by participants and that regulate their behavior. A norm is a concept used by members of a social unit in evaluating other people,

(572) events, and social objects of consequence to them. Norms are evaluative scales that define for group members a latitude of acceptable behavior and a latitude of unacceptable behavior.

7. On the basis of the experimental study of norm formation, a number of generalizations may be made: First, when a person repeatedly faces an unstructured stimulus by himself, he develops a consistent pattern of judgments or a norm relative to it. Second, when an individual has developed an individual norm and is placed in a situation with others who hold different norms, in time they develop some common social norm. And finally, when an individual who has participated with others in the formation of a norm is asked to make judgments alone, lie still continues to adhere to the social norm evolved in interaction with others.

8. A leader is an individual who has the highest position in a group in terns of status (power or effective initiation). Leadership refers to his role in the group-the particular skills and qualities lie uses, his popularity with other members, and so on. All groups have leaders, even when the members do not identify one person as the leader. To the extent that a group structure has taken shape, one person may ordinarily be singled out as more effective than other members in initiating and controlling the activities of others.

9. Groups in actual social life are seldom isolated from other groups of people. Intergroup relations refers to the state of affairs between two or more groups, whether these are reciprocally friendly or conflicting. From his experimental studies of intergroup relations, Sherif has concluded that: (A) when two groups come together in a series of activities where one could achieve its goal only at the expense of the other, open conflict develops; (B) social distance develops in such a situation; (C) members of each group develop hostile and unfriendly attitudes toward members of the other group. In addition, often in such situations, there are changes in the status structures of the groups involved, new norms develop, and increased solidarity, morale, and cooperativeness may develop among members within each group.

10. Intergroup tension and hostility may be reduced by both groups uniting against a common enemy or by breaking up the individual groups. In general, a more desirable method involves the introduction of superordinate goals — goals that are attractive to members of both groups and that can be attained only through intergroup cooperation.


Gibb, C.: Leadership. In Lindzey, G. (ed.): Handbook of Social Psychology. Vol. II, Reading, Mass., Addison-Wesley, 1959. (This is a detailed discussion of many experimental and empirical studies of leadership.)

Lewin, K., Lippitt, R., and White, R.: Au experimental study of leadership and group life. In Swanson, G., Newcomb, T., and Hartley, E. (eds.): Readings in Social Psychology. New York, Holt, 1952. (This is one of the classic experimental studies of leadership.)

Sherif, M., and Sherif, C.: Groups in Harmony and Tension. New York, Harper, 1953. (This is a report of the experimental studies of group formation and intergroup relations referred to in the chapter.)

Sherif, M., and Sherif, C.: Reference Groups: Exploration into Conformity and Deviation of Adolescents. New York, Harper & Row, 1964. (This is a report of an experimental study of reference groups and the implications of such groups for the conformity and deviation of adolescents.)

Sherif, M., Harvey, O. J., White, B., Hood, W., and Sherif, C.: Intergroup Conflict and Cooperation, The Robbers Cave Experiment. Norman, Okla., University Book Exchange, The University of Oklahoma, 1961. (This is a detailed report of the hypotheses tested in, methodology of, and results of the Robbers Cave study.)

Whyte, W.: Street Corner Society. Chicago, Chicago University Press, 1943. (This is one of the best known social-psychological studies of small groups. It is based on Whyte's observations of a Boston street gang.)


  1. This chapter was prepared by Muzafer Sherif and Carolyn Sherif of the Institute of Group Relations, University of Oklahoma.



Valid HTML 4.01 Strict Valid CSS2