In sociology, the concept "group" implies more than simply an aggregate of individuals. Additional elements involved are (1) structure—interaction patterned in terms of statuses and roles, (2) history—some frequency and regularity of interaction over time, (3) interdependence—some degree of members' mutual reliance on each other for needed or valued material and nonmaterial resources, and (4) common identity—grounded in shared meanings, values, experiences, and goals. Frequently there is a group product, not necessarily of a material nature, which is the outcome or consequence of collective effort and interaction.
These elements are dimensional in that groups possess and manifest them to a greater or lesser degree. At one extreme, family groups typically have well-established and enduring structures, share extensive histories, encompass a wide range of activities, exert a broad scope of influence, and provide the basis of individual identity. At the other extreme, ad hoc work groups (and groups studied in laboratory experiments) may be assembled to perform specific tasks of very limited duration with little or no relevance for or influence on the members outside a clearly defined situation and range of activity. McGrath (1984) developed a comprehensive typology of groups in terms of origin, scope of activity, task, duration, and interaction.
Groups are regarded as small if meaningful and direct face-to-face interaction can take place among all members. The number of members usually is thought of as ranging from two to twenty, with three to seven common in many laboratory studies of groups.
PRIMARY AND SECONDARY GROUPS
Cooley (1909) identified a fundamental type of small group that is characterized by intimate association and cooperation, which he regarded as the basic building block of society. Cooley called groups of this sort "primary groups" and held them to be forms of association found everywhere. Primary groups work on the individual to form and develop the social nature of the person. "This nature consists of certain primary social sentiments and attitudes, such as consciousness of one's self in relation to others, love of approbation, resentment of censure, emulation, and a sense of social right and wrong formed by the standards of a group" (1909, p. 32).
Membership and participation in primary groups are valued and rewarding for their own sake. The groups typically are long-lasting. Members interact as "whole persons" rather than merely in terms of specialized, partial roles. Primary groups are basic sources of socioemotional support and gratification, and participation in them is considered essential for a person's psychological and emotional well-being. Some (the family, the neighborhood peer group) are also primary in the sense that they are settings for early childhood socialization and personality development.
In contrast are groups formed and maintained to accomplish a task, to which people belong for extrinsic purposes (because they are paid or to achieve an external goal). These "secondary groups" are characterized by limited, instrumental relationships. They may be relatively short-term, and their range of activity is restricted. Affective ties and other "irrational" personal influences are intended to be minimized or eliminated.
It has been widely observed, however, that primary relationships develop pervasively within secondary groups and organizations. In a synthesis of observations and research findings, Homans (1950) attempted to identify universal variables of group behavior. He sought to develop a general theoretical scheme that would permit an understanding of groups as diverse as an industrial work unit, a street-corner gang, and a Polynesian family. Homans approached the small group as a system in which activity, interaction, and sentiment are interrelated. He concluded that interaction among group members increases their liking for one another and that they tend to express their friendship in an increasing range of activities and to interact more frequently. Affective elements emerge in virtually all ongoing groups and may enhance or interfere with the purposes for which a group was established. Soldiers are motivated to fight and workers are motivated to increase or restrict work output by loyalty to their friends and the norms of the immediate group.
BASES AND DEVELOPMENT OF SMALL GROUP RESEARCH
Sociological interest in small groups has several bases, including (1) the perception of small groups as fundemental, universal social units on which all larger organizational structures depend, (2) a concern with the description and understanding of particular small groups both for their own importance and as a source of observations from which hypotheses and general theories can be developed, and (3) the usefulness of the laboratory group as a research context in which to study the characteristics of the group as the unit of interest and as a setting for the investigation of social influence on individual cognition and behavior.
Foundations for small group research may be seen in nineteenth-century sociological thought, such as Emile Durkheim's analyses of the development of social structures, specialization and task differentiation, and the bases of social cohesion and Georg Simmel's work on the importance of group size and coalition formation. Early in the twentieth century Charles H. Cooley and George Herbert Mead stressed the social construction of the self through interaction within immediate group settings.
In the 1930s and 1940s, Jacob L. Moreno developed a systematic approach to the understanding and charting of group structure and Muzafer Sherif conducted key studies of group influence and conformity. William Foote Whyte's field study of a street-corner gang demonstrated the existence and importance of group norms and structure in an urban milieu generally thought to lack social organization. Of major importance was Kurt Lewin's work, which provided direction and inspiration for the postwar generation of social psychologists. Lewin combined principles of Gestalt psychology and concepts from the physical sciences to develop field theory in social psychology as a basis for the study of group dynamics. Interested in both theoretical and applied aspects of group interaction, in 1945 he established the first organization devoted to research on group dynamics. The widely utilized sensitivity-training group method originated serendipitously in sessions Lewin organized in 1946.
The period from the end of World War II to the early 1960s produced burgeoning activity in small group research (Hare et al. 1965). The pervasiveness of Lewin's ideas was evident in the growth of group dynamics as an area of research and theoretical development. Cartwright and Zander's important compilation, Group Dynamics (1968), first published in 1953, presented a theoretical overview and numerous influential studies of cohesiveness, group pressures and standards, individual motives and group goals, leadership and group performance, and the structural properties of groups.
Substantial work with a different orientation reflected concerns with functional needs that groups must meet in order to survive and with the relationship of those functions to dimensions of interpersonal behavior and personality traits. At the same time, influences from anthropology, economics, and behavioral psychology were being melded in a view of social interaction as an exchange of resources, a perspective applied to the analysis of interdependence, cooperation and competition, and interpersonal relationship (Homans 1950, 1974; Thibaut and Kelley 1959). During those years small group research shared the methodological advances that were occurring throughout the social sciences, developing an increasing sophistication in research design, measurement, and analysis. The excitement, optimism, and productivity of the field led some to define social psychology as the study of small groups.
Small group research since the 1960s has not been as prominent, prolific, or influential as it was during the immediate postwar years, when social psychology was virtually dominated by the small groups "movement" (Borgatta 1981). The production of studies is steady, if moderate compared to the enthusiasm of the peak period, and some significant attempts have been made to organize and integrate the diverse body of work and theory that has accumulated (Hare 1982; McGrath 1984; Foschi and Lawler 1994). There is renewed interest in conceptualizing groups as entities with distinctive properties that cannot be understood in terms of reductionist individual psychology (Turner 1987). Many aspects and procedures of group process and dynamics are commonly utilized in applied settings, while practical concerns with group productivity, efficiency, and success are widespread (Hare et al. 1992; Forsyth 1999).
APPROACHES TO SMALL GROUP RESEARCH
Small group studies are characterized by a wide variety of research techniques and theoretical and practical concerns. Research methods vary in regard to the types of groups and circumstances studied—whether "natural" or contrived for research purposes—and in the intrusiveness of research procedures. Some investigators are concerned with properties of the group itself as the unit of interest, while others use the small group setting as context for explorating individual behavior. Although laboratory studies have predominated, the research techniques employed include direct observation of groups in natural as well as controlled settings; the use of structured observational systems to code communication or other aspects of behavior; the use of checklists, questionnaires, or interviews to elicit ratings, choices, opinions, or attitudes from group members; and field experimentation.
Laboratory studies have marked advantages in terms of the control and manipulation of variables in the precision of observation and measurement. The procedures employed normally permit replication of observation under controlled conditions. The experimental method is regarded as superior to others for rigorously testing causal hypotheses. Fundamental technical issues are whether relevant variables can be brought into laboratory situations and whether a meaningful range of variation can be achieved.
Criticisms of laboratory research center on the artificiality of the setting and the short-term nature of most studies. Representativeness of subject groups and thus the generalization of the findings also are questioned. Concerns for protecting the rights and well-being of human subjects have led to procedural safeguards that now inhibit or prevent practices that were typical of some well-known earlier studies.
The technical advantages of laboratory procedures, the desire to emulate the natural sciences in developing theory based on experimental evidence, and the compatibility of laboratory methods with the academic environment within which most researchers work all have contributed to the proliferation of laboratory studies that constitute much of small group research.
Direct observation of group behavior under basically uncontrolled ("natural") conditions may be coupled with the investigator's more or less active participation in the affairs of the group. Such research can employ structured systems for coding behavior and interaction patterns that are used by uninvolved "objective" observers, as when a children's play group is studied by adults. A more informal ethnographic approach was employed by Goffman (1964) in collecting the information that illustrated his characterization of human interaction as an elaborate sequence of symbolic presentations of self and groups as collaborating teams of performers.
Participant observation is a procedure in which the researcher acts as part of a (usually natural) group to understand a situation from within, as members of the group define and experience it. Group members may know that the observer is an outsider who is there for his or her own purposes or may be led or allowed to believe that the observer is simply another "genuine" group member. In either case the observer's status influences and constrains both the kinds and amount of information available and the opportunities for recording information. The observer also has some influence on the situations and processes being studied, thus producing outcomes different from those which would have occurred in his or her absence. The use of multiple observers increases opportunities for observation while also increasing the effect of the research on the group's behavior (Festinger et al. 1964). For these reasons, reliability and validity are particularly problematic issues in using this technique.
Participant observation is regarded as useful primarily for descriptive and exploratory research and for generating or illustrating, as opposed to testing, theory. It is favored by those who want to understand the meanings of situations and actions generated and maintained by groups in their natural, everyday environments.
An important naturalistic study was conducted in the late 1930s by Whyte (1955), who studied a street-corner gang as a participant observer over a period of three and a half years. (The appendix to his monograph provides an informative discussion of practical and ethical issues in participant observation.) Whyte gained access to the gang through his association with its leader, and his view is from the top of the social structure. He described the recurrent patterns of relationships among members, group values and codes of behavior, the existence of implicit exchange relationships, territorial behavior, and the nature and functions of gang leadership. His observations of the ways in which members' social rankings in the group affected their performance in athletic competition suggested a program of experimental studies of diffuse status characteristics: an exploration of the manner in which "logically" irrelevant social rank affects the amount of influence an individual has on others in activities ranging from pedestrian behavior to the making of perceptual judgments.
Sociometry, a seminal form of network analysis developed by Moreno (1953), is a technique for eliciting and representing the patterns and structure of choices and liking among group members. While the most common procedure is for researchers to ask group members who they like, dislike, would prefer to work with, or would like to "be like," ratings also can be based on direct observations of members' behavior. The information can be represented as a sociogram showing individuals as circles and choices as arrows between the circles: The diagram depicts group structure in terms of affective relations. Indices of liking or disliking can be computed for each member, and ratings can be organized in a matrix format. The density and patterning of choices may be taken as indicators of group cohesiveness. In practical applications, sociometric data are used to restructure groups on the basis of members' mutual choices.
Interaction Process Research. A prominent research concern has been the description and analysis of group interaction processes, focusing primarily on communication. The approaches employed have ranged from purely formal examination of the amount of communication sent and received by each member of the group to extremely detailed analyses of linguistic and paralinguisitic material, including posture, gestures, and inflection.
The widely used system for Interaction Process Analysis (IPA) developed by Bales (1950, 1970) involves a set of twelve categories for coding units (acts) of communication. The categories reflect Bales's conclusion that all groups confront two domains of concerns: instrumental concerns related to the task the group must accomplish and expressive concerns associated with the socioemotional needs and interrelationships of the group members. Both sets of concerns operate continuously and must be dealt with if a group is to succeed and survive, and there is a virtually constant conflict between them. The set of categories is used by observers to code types of active and passive task-related acts and positive and negative socioemotional acts, as they are generated by group members in the course of interaction.
Numerous studies using the IPA system have sought to document the patterns or "phase movements" of instrumental and expressive communication as groups try to establish the equilibrium necessary to operate. Interaction process scores have been related to personality characteristics and to peer assessments and self-assessments (Borgatta 1962). Attention also has been paid to the roles of particular group members in exercising task leadership or socioemotional leadership.
The division of group leadership into instrumental and expressive functions proved compatible with accepted notions of "typical" male and female personal attributes and with a conceptualization of the family (at least in the Western world) as a small group with the father as task leader and the mother as socioemotional specialist. However, recent research comparing "natural" families with ad hoc laboratory groups indicates that the instrumental versus expressive specialization found in the laboratory seldom holds for groups in natural settings. There is greater diversity of behavior and less gender-linked stereotypical conduct in longerlasting groups that cover a greater scope of activities (McGrath 1984).
The IPA system has been criticized on both theoretical and operational grounds, and numerous revisions and alternatives have been proposed. Bales and his colleagues developed an elaborated observational system, SYMLOG (Bales and Cohen 1979), that models personal space in three dimensions: dominant-submissive, friendly-unfriendly, and instrumentally controlled–emotionally expressive. Group interaction is observed and members' behaviors are coded on each dimension by outside observers or by the group members themselves. On the basis of combinations of multiple observations, each individual is located within the three-dimensional space and the positions of all group members are charted. The resulting diagram and indices based on the scores indicate the degree to which members are perceived as acting in a similar fashion. Interest in the SYMLOG technique is substantial, and it is utilized in many studies of group structure and performance.
The understanding of what holds a social unit together, a central issue in sociology, also has been central in small group analysis. Cohesion—the sum of the forces that bind members to the group—was viewed by Lewin and other Gestaltists as a property or characteristic of the group itself, a sort of force field analogous to a magnetic or gravitational field. However, the assessment of cohesion usually depends on observations of the attitudes and behaviors of the individual group members: their self-reported attraction to the group, their feeling of being accepted by the group, similarity in expressions of sentiment, how regularly they attend group meetings, how prompt or tardy they are, or how responsible they are in performing actions that benefit the group. Members also may asked to describe the unity of the group (Evans and Jarvis 1986; Bollen and Hoyle 1990). Although Steiner (1972) suggested that "A true test of a group's cohesion would entail observation of its members' reaction to disruptive influences," he rejected this procedure on technical and ethical grounds (1972, p. 161).
The bases of cohesion include (1) rewards available within and through the group, (2) the congruence between individual goals and group goals, (3) the attraction and/or liking of members for each other, (4) the importance of the group as a source or ground of the individual's identity and self-perception and his or her internalization of group culture and values, and (5) in psychoanalytic group theory, the members' identification with and attraction to the group leader and "the alignment between particular individual superego formation and its corresponding punitive group structure" (Kellerman 1981, p. 11).
Although high cohesiveness often is taken as indicating a "healthy" group, its effect is to heighten members' susceptibility to influences in the group. Thus, group productivity, for example, may be increased or decreased depending on the nature of the predominant influences. A positive association between group cohesion and performance was found by Evans and Dion (1991) in a review of previous studies, but the relationship was modest.
Major importance in the study of cohesiveness has been placed on interpersonal attraction and interdependence, emphasizing the exchange of emotional and affective resources. Work by Tajfel (1981) and Turner (1987) supports, alternatively, an emphasis on social identity and self-categorization. A concept of cohesion based on interpersonal liking that is not mediated by shared group membership and depersonalized attraction to the group is held to be inadequate. Group membership and the resultant self-categorization occur prior to interaction and the emergence of interdependence, cooperation, influence, and cohesion. Hogg (1987, 1992) advocates research that will produce answers about group solidarity and social identity rather than about interpersonal relationships.
Self-categorization in a most elemental form has been demonstrated in "minimal group" experiments (Tajfel 1981). Subjects are divided into two groups, sometimes presumably on the basis of an arbitrary and unimportant criterion and sometimes in an obviously random manner. The participants do not interact within or between groups during the experiment. Given the task of dividing a sum of money between two persons about whom they know nothing except their group membership, subjects show a marked bias in favor of members of their own group.
Interdependence in a most elemental form has been realized in experiments with the "minimal social situation" (Sidowski 1957). Two subjects, each of whom controls resources that may reward or punish the other and each of whom depends primarily on the other's behavior as a source of reward or punishment, learn to exchange rewards despite being completely unaware of the nature of the situation.
Thibaut and Kelley (1959) identified two criteria individuals use in evaluating the rewards available within a particular situation: a usual, expected level of reward to which the person feels entitled, called the "comparison level," and the person's perceived best level of reward available outside the situation, called the "comparison level for alternatives." An individual's satisfaction with his or her group membership and participation depends on the relationship of rewards available within the group to his or her comparison level, while the likelihood that one will stay in or leave a group depends on the comparison level for alternatives.
Although the value and availability of rewards are usually emphasized in assessing the attractiveness of a group, Leon Festinger has pointed out the persistence of loyalty to "lost causes" and the effect that insufficient reward, or even aversive experiences, can have in strengthening members' positive attitudes. In one experiment (Aronson and Mills 1959), potential group members who were subjected to a severe initiation expressed greater liking for the group than did those who had a mild initiation. And while an equitable and balanced exchange of rewarding outcomes is considered important in sustaining interpersonal relationships and participants' satisfaction with them, Kelley and Thibaut (1978) noted that problematic situations provide particular opportunities. Attributions about a partner's personality and motivations and self-presentations that encode messages of commitment and concern for the other person are facilitated when behavior cannot be explained simply in terms of "rational" self-interest. Such attributions and encodings strengthen affective ties and promote interdependence of the characteristics and attitudes displayed in the relationship.
Social Facilitation and Inhibition. In a study credited as the first social psychological experiment (1897), Triplett measured the average time his subjects took to wind 150 turns on a fishing reel, working both alone and in competition with one another. Subjects working in competition wound the reels faster than did those working alone. Numerous subsequent experiments (including some with nonhuman subjects) have supported and modified these results. It was found that the mere presence of other persons (as observers or coactors, whether or not they were competitors) facilitated well-learned responses but that the presence of others interfered with the acquisition of new responses. This "audience effect" thus facilitates performance but inhibits learning. Various explanations of social facilitation and inhibition have been proposed, generally incorporating the idea that the presence of others increases motivational arousal. Such arousal is a basic feature of the group environment (Zajonc 1966).
Conformity. Similarities of values, attitudes, beliefs, perceptions, and behavior are a ubiquitous and virtually defining feature of group existence. These similarities can facilitate coordination of goal-directed activity, motivate the members, provide sources of psychological security and emotional reward, reinforce members' identification with the group, and increase cohesiveness. They also may prevent reasoned consideration of alternatives to group decisions and the potential consequences of group actions, reduce flexibility in adapting to new circumstances, and inhibit change in general. Closed circles of conformity in cohesive groups that are isolated from dissenting view-points, producing "groupthink" ( Janis 1982), have been implicated in producing military blunders, fascist atrocities, government scandals, and space shuttle disasters. Conformity (to modeled indifference or uncertainty) is a factor in the failure of bystanders to help others in emergencies.
The amount of conformity in a group may be seen as a characteristic of the collectivity. Experimental studies, however, usually have been concerned with effects on the individual. Considered from this viewpoint, conformity is defined as a change in an individual's attitudes, beliefs, or behavior in the direction of a group norm. It is an example of social control resulting from peer influence (as distinct from, for example, obedience to a constituted authority) (Milgram 1974). Two types of conformity have been identified: belief (or informational) conformity and behavioral (or normative) conformity. Both types are increased by strong group cohesiveness.
Belief conformity involves an internalized and lasting change grounded in an individual's dependence on social sources of information and guidance. Once they are internalized, the group's standards and perceptions are constantly carried with the individual and constitute an ongoing element of social control.
Sherif (1935) asked individual subjects to judge the apparent movement of a pinpoint of light in an otherwise totally dark room. Under these conditions the light, which in fact was stationary, appeared to most people to move. Different individuals perceived different amounts of movement. Assembled in small groups viewing the light together, the subjects began to agree on the amount of movement they perceived: A group norm emerged in an ambiguous situation. After the group interaction, subjects were asked to view the light, again in isolation. They continued to see the amount of movement agreed on by the group rather than the amount they originally perceived individually. The group's perceptions apparently had been internalized.
The strength of belief conformity varies with the ambiguity and unfamiliarity of the situation, the individual's trust in the credibility of the group, the individual's attraction to and identification with the group, and the individual's prior experience and confidence.
Behavioral conformity is grounded in the potential rewards and punishments dispensed by the group and in the individual's previous experience with the consequences of conformity and nonconformity. The consequences of agreeing with others' judgments and opinions, emulating others' behaviors, and following the customs of a group are usually pleasant, while disagreement and deviancy generally lead to unpleasant effects. Group members who hold deviant opinions typically receive, at first, greater than normal amounts of communication in an attempt to influence them to conform. If these efforts fail, they are likely to be isolated or rejected, depending on the severity of the deviance. Monitoring of behavior is necessary if reward or punishment is to depend on its occurrence; thus this type of influence is effective only if and when an individual's actions are known to the group.
Experiments conducted by Asch (1951) demonstrated behavioral conformity. The subjects engaged in a perceptual estimation task that required them to pick out lines of the same length printed on boards that were presented side by side. The boards were presented in pairs, and the judgment of each pair constituted one experimental trial. In a typical experiment there was only one genuine subject; the other participants were employed by Asch, and their judgments were prearranged. After a number of trials in which correct judgments were given, the confederate "subjects" began stating unanimous wrong judgments. The genuine subjects conformed to a substantial extent by expressing judgments that agreed with those of the group. When removed from the group or allowed to state judgments in private, the real subjects did not persist in making these errors. Their conformity occurred only when it was witnessed by the other group members.
The strength of behavioral conformity varies with the size and unanimity of the group, the importance of the group to the individual, and the disclosure of relevant judgments or behaviors to the group.
Belief and behavioral conformity can be distinguished analytically (and empirically under some laboratory conditions), but in natural situations they operate in conjunction. The group member not only is rewarded for conforming but also depends on others as models for behavior and guides for judgments and opinions. While it is common to think of beliefs and attitudes as existing before the behaviors that reflect them, a large body of research indicates that people come to believe the opinions they express. "Mere" behavioral conformity can lead to internalization.
Conformity effects usually are thought to reflect the majority influence in a group, but evidence shows that a determined minority can prevail. Minority influence seems especially relevant in regard to internalization (Moscovici 1980).
Group Polarization. Early theories of "group contagion" and the madness of crowds notwithstanding, a general assumption has been that conformity processes within a group operate to bring extreme opinions and judgments in toward the center of the range of opinions and judgments. However, a body of research has contradicted the notion that group actions are more moderate than those of individuals.
The experimental procedure called for individual subjects to evaluate each of twelve "choice dilemmas," situations in which a person was asked to choose between a highly desirable risky alternative and a less desirable but certain alternative. The subjects were instructed to indicate for each dilemma, the lowest probability of success they would accept in recommending that the desirable risky alternative be chosen. Probabilities were averaged for each subject over all dilemmas to generate a "riskiness" score for that person. Small groups of subjects were then formed and instructed to discuss each situation, reach a group decision, and indicate the group riskiness score for the dilemma. A group's scores were averaged over the twelve situations, and that value was compared to the mean of the individual scores of the group members.
Initial research that employed the choice dilemmas procedure found a significant "risky shift" in the group decisions compared to the mean of the individual scores. Numerous experiments and further analyses followed that extended and qualified those findings (Cartwright 1973). Certain kinds of choice dilemma scenarios produced risky shifts, while others produced conservative shifts or showed no significant difference. Shifts tended to move in the direction of the initial inclinations of the group: The interaction resulted in a collective outcome more extreme than might have been predicted on the basis of the individual positions, but the individual positions forecast the nature of the shift.
Group polarization, as the effect is now called, has been theoretically interpreted in terms of risk as a cultural value, the persuasive influence of "risky" individuals, and the diffusion of responsibility in group action. However, the effect can be explained as being due to the normative and informational influences involved in conformity processes (McGrath 1984).
GROUP INTERACTION AND PERFORMANCE
Group performance in terms of problem solving, productivity, or effectiveness is a subject of both practical and theoretical concern that has generated numerous studies and a large body of theory. Productivity may refer to the quality of a group product, the efficiency of output per unit time or progress toward a group goal, or the realization of group potential. The establishment of an appropriate basis of evaluation is often problematic, and expected outcomes depend heavily on the type of task undertaken. When groups fall short of what (from some standpoint) it is felt they should accomplish, the failure often is attributed to "process losses" resulting from problems in interaction.
Steiner (1972) distinguished between tasks that require a coordinated division of effort, which he labeled "divisible," and those with a single outcome or product, which he called "unitary." Disjunctive unitary tasks are those which can be accomplished successfully by one individual alone. In such cases the group should be as "good" as the best member. Conjunctive unitary tasks require all the members to contribute successfully; in these tasks, the group can be only as good as the worst member. Tasks in which members' contributions are simply summed to produce the group outcome are called additive, and group performance should depend on the "average" member. Numerous laboratory studies of ad hoc groups performing a wide range of judgment tasks have been conducted. Overall, the results indicate that groups seldom do as well as the best member but usually do better than the average member.
Field studies of industrial workers in natural settings illustrate how influence processes in the group can regulate behavior. Work groups develop norms with respect to what they, not the company, regard as an appropriate day's output. While pay, potential promotion, and retention or termination may be controlled by the employer, the immediate group controls powerful social rewards and sanctions that are brought to bear on a day-to-day basis. Those who exceed the group's production norm ("rate busters") and those who fail to produce an acceptable amount or attain an acceptable standard of quality are subjected to group pressure ranging from "kidding" and mild criticism to serious harassment. Since group cohesiveness increases conformity, some companies find it desirable to move workers frequently and attempt in other ways to inhibit the formation of interpersonal ties and identification with the group. Other organizations attempt to mobilize small group processes to support their goals.
Successful performance requires that a group have the necessary resources (material resources and members' skills, knowledge, and competencies) and time needed to accomplish its tasks. In addition, issues of coordination and motivation arise. When it confronts a disjunctive unitary task, the group simply must assure that the "best" member has the opportunity, recognition, and authorization to function and is motivated to do so. The only coordination needed may be to prevent interference from other members. For other types of tasks the quality, sequence, and articulation of many or all members' contributions are important (Miller and Hamblin 1963).
Allocation of opportunity to participate and evaluation of members' actions constitute elements of the status structures of groups. The explanation of how interaction inequalities in task groups are developed and maintained is the concern of expectation states theory (Foschi 1997; Foschi and Lawler 1994).
Group members hold expectations about the nature, quality, and value of each other's performances. Their expectations influence the quality of those performances and affect the evaluation of performances after they occur; they are in this sense self-fulfilling prophecies.
Though expectations may derive from first-hand task experience within the group, they also are based on "external" status characteristics of the members. Diffuse status characteristics such as age, race, gender, or perceived social rank may influence expectations whether or not they are objectively relevant to the group's task and goals. Inequalities in participation, evaluation of performance outputs, and influence over the group's decisions reflect inequalities in status characteristics that members bring to the group. These inequalities tend to be maintained within the group regardless of their pertinence. Evaluations of performance output depend on previous evaluations, and expectations that arise out of group interaction influence subsequent interaction to produce their own confirmation (Berger et al. 1972, 1980). Thus the degree of influence exerted by group members and the impact of their contributions to the group's effort may not be highly correlated with their task-related competence and abilities.
Processes of influence and conformity may degrade performance quality. Majorities generate social pressure whether or not they are competent. Techniques have been devised to control these effects by regulating the kind of interaction that can take place. Some procedures, such as Multi-Attribute Utility Analysis, require the clear identification of task elements and their accomplishment in specified sequences. Group members can interact freely but must adhere to the task stages. Other approaches impose rules for communication in decision-making processes.
Many studies have been concerned with evaluating the effects of different patterns of interaction, primarily communication, on performance. "Brainstorming," a group interaction technique in which members generate as many ideas as possible within a given time period without evaluation or criticism, is intended to overcome inhibiting social influence processes while taking advantage of those which stimulate creativity. However, research indicates that a brainstorming group is generally less effective in producing ideas than are the same number of individuals working alone. The Delphi Method and the Nominal Group Technique are two approaches to regulating the combination of individual effort with group feedback or interaction to reduce the deleterious effects of social relations within the group, conformity, and personalized conflict. In the Delphi Method individuals, without communicating with each other, make judgments that are combined into a group "product." The results are made known to the participants, who then make another round of judgments. This procedure is repeated until a final group judgment is attained. The Nominal Group Technique, which is used for developing plans or ideas or for choosing a correct or best solution, begins by having individuals work separately to generate plans, ideas, or judgments. The group then collectively lists and evaluates the material that was produced individually. These methods have advocates, but the desirability of some of their results is questionable and the time and cost involved in their utilization may be significant (McGrath 1984).
A substantial body of research has compared the relative effectiveness of structured networks of communication available to members of problem-solving groups. Typically, groups of three to five persons were required to combine information distributed across the individual members, communicating only through channels provided by the experimenters. Various networks of communication channels have been investigated to see how they affect a group's efficiency and the members' satisfaction. The networks differ in terms of how centralized or open they are. The most centralized network compels all messages to flow through one position, while the most open permits direct communication among all the members.
The conclusions from this research are that centralized networks are most efficient in dealing with simple tasks but that group members tend to be dissatisfied, except for the person occupying the central position. In more complex tasks the advantages of centralization are lost. Burgess (1969) suggested that the network experiments were basically flawed in failing to provide meaningful consequences for group performance and in studying groups only for brief periods, while they were learning to use the networks. His research demonstrated that when subjects had enough time to learn to use the channels provided, and received rewards based on performance, the type of network made no difference. Given time and motivation, groups adapted efficiently to overcome the structural constraints.
A type of process loss observed in both physical and cognitive tasks is the reduction in effort people put into group performance compared with the effort they make when working individually. This effect, called social loafing, has been observed in numerous cultures and is related to group size: As groups get larger, individual effort tends to diminish. Social impact theory explains social loafing in terms of a conflict between a person's sense of responsibility and his or her feeling that inaction is the safest or least costly course of behavior. Diffusion of personal responsibility occurs in group situations and reduces the blame for inaction. Also, a lack of individual identification and the absence of evaluation by others become more likely as group size increases. Thus both rewards for effort and punishment for lack of effort become less certain and less consistent. While laziness and "goldbricking" often occur in individual situations, those behaviors can be concealed more easily in a crowd. Nonetheless, members who identify with and value a group and hold strongly to its norms will exert effort on behalf of the group (Hogg 1992). As was noted above, there is a positive association between group cohesion and productivity in groups with norms that support good performance. This relationship is reciprocal: Group success tends to increase cohesion.
The social loafing effect is confounded with problems of coordination, since both increase as the number of persons involved in a task gets larger. In addition, members' impatience and/or frustration with coordination problems can undermine their motivation and sense of responsibility, exacerbating social loafing.
Although researchers have paid much attention to process loss, there also are many significant process gains in group interaction. Social facilitation, stimulation, learning, socioemotional support, development and reinforcement of identity, and even conformity processes can enhance creativity, productivity, and effectiveness. While questions of individual versus group superiority may be provocative, most human endeavor occurs in group contexts and requires group effort.
The recognition that productivity can be affected substantially by the functions and quality of leadership in a group has stimulated much research and theorizing about leadership styles and effectiveness. An early experimental program conducted by Lewin et al. (1939) systematically varied the behavior of adult leaders of clubs of 11-year-old boys engaged in craft work and recreational activity. The leaders were trained to enact democratic, authoritarian (autocratic), and laissez-faire styles of leadership, and those styles were experienced by each club for several weeks. The resulting changes in the boys' task performance, social relationships and interaction, and motivation as well as some aspects of group "climate" were intensively documented and analyzed. The results of the study, while complex, generally favored the democratic leadership style both for producing increased motivation and originality and for fostering more mutual friendliness and group-mindedness among the boys. Group members preferred the democratic leader to either the autocratic or the laissez-faire leader.
A contingency model of leadership effectiveness has been developed by Fiedler (1981), whose concepts of task-motivated versus relationship-motivated style recall Bales's identification of instrumental and socioemotional leadership functions. A leader's effectiveness results from the combination of style and situation: Task leaders are most effective in situations that are either highly favorable or unfavorable, while relationship leaders are most effective in middle-range situations. The contingency model, though supported by a large body of research, is questioned by some who feel that effective leaders are those who deal with both task and relationship elements of group situations.
COOPERATION AND COMPETITION IN GROUPS
Two different orientations are evident in research on competition and cooperation within groups. In one approach cooperation and competition are treated as imposed external conditions that influence the quality of group interaction and task performance. Alternatively, cooperation and competition have been studied as dependent behaviors that are affected by reward and risk contingencies, the availability of communication, and other situational factors.
Numerous studies have compared the productivity and efficiency of groups working under cooperative conditions (defined as working for group goals) and competitive conditions (defined as working for individual goals). The concept of cooperation in early research usually specified only mutual dependency of outcomes, with little attention paid to the interdependency of the members' task activities. The findings indicated that the efficiency of work under competition was greater than that under cooperation for tasks that did not require coordination of effort. Some research indicated that cooperative groups worked together more frequently and were more highly coordinated.
Analysis of research focusing on the nature of tasks used as criteria in comparing cooperative and competitive reward structures points to the importance of "means interdependence," the degree to which group members are reliant on one another (Schmitt 1981, 1998). When tasks are simple, requiring no division of labor or sharing of information or resources, the advantage of cooperative contingencies seems to hold. However, cooperative contingencies are typically superior for tasks high in means interdependence involving distribution of effort, coordination of responses, or information sharing. Also, the long-term consequences of different reward structures may be substantial in natural groups, involving issues of morale and sustained member motivation that seldom arise in relatively brief laboratory studies. Such consequences can depend on the way in which competitive payoffs are determined; the possibility that some group members may become perpetual "losers" while others are constant winners will affect the efforts of all the members.
In a number of cases an additional element of competition between groups has been found to increase the productivity of internally cooperative groups. Turner (1987) observed that competition (for mutual distinctiveness) can develop between groups even in the absence of conflicts of interest. This striving to enhance positive social identity increases group cohesiveness and solidarity, making cooperation more likely.
Laboratory research treating cooperation as a dependent effect has focused on the participants' choice of cooperative rather than competitive behaviors and the distribution and coordination of responses. The effects of threat and communication were investigated in a well-known "trucking game" study (Deutsch and Krauss 1962). Two subjects could cooperate by taking turns using a "short route" to reach a destination and thus make money. Cooperation was reduced when one subject could block the route with a gate ("unilateral threat") and was extremely rare when both subjects had gates ("bilateral threat"). Communication between subjects did not increase cooperation under the threat conditions.
Communication sometimes has been found to increase cooperation in some of the many studies using the "Prisoner's Dilemma." In this situation, two participants benefit moderately if both choose to cooperate and lose substantially if both "defect." If either one chooses to cooperate while the other defects, the cooperator suffers a very large loss and the defector's outcome is highly favorable. Thus, cooperation involves risk while defection implies motives of self-protection, exploitation, or both. The structure of outcomes is paradoxical: The rational choices of each individual lead to poor collective consequences.
The rates of cooperation observed in these studies are low. The Prisoner's Dilemma epitomizes the class of situations called social traps, in which individual (usually short-term) "rational" self-interest conflicts with the (usually longer-term) well-being of the group, leading to collective irrationality (Kollock 1998).
Inequity of outcomes and the presence of risk have been found to reduce cooperation across a wide range of experimental research (Marwell and Schmitt 1975). Beneficial effects of communication were dependent on the timing of its availability and the pattern of behavior that had occurred before communication took place.
Studies of cooperation and competition have addressed problems of motivation and coordination, issues of equity, the effects of short-term and long-term consequences, and the relationship of individual outcomes to collective outcomes. The analysis of these topics is a notable feature of recent small group research, particularly as concern with social traps and dilemmas resonates with the environmental and social issues facing contemporary society.
Aronson, Elliot, and Judson Mills 1959 "The Effect of Severity of Initiation on Liking for a Group." Journalof Abnormal and Social Psychology 59:177–181.
Asch, Solomon E. 1951 "Effects of Group Pressure upon the Modification and Distortion of Judgments." In H. Guetzkow, ed., Groups, Leadership, and Men. Pittsburgh: Carnegie.
Bales, Robert F. 1950 Interaction Process Analysis: A Methodfor the Study of Small Groups. Cambridge, Mass.: Addison-Wesley.
—— 1970 Personality and Interpersonal Behavior. New York: Holt, Rinehart and Winston.
——, and Stephen P. Cohen 1979 SYMLOG: A Systemfor the Multiple Level Observation of Groups. New York: Free Press.
Berger, Joseph, Bernard P. Cohen, and Morris Zelditch, Jr. 1972 "Status Characteristics and Social Interaction." American Sociological Review 37:241–255.
——, Susan J. Rosenholtz, and Morris Zelditch, Jr. 1980 "Status Organizing Processes." Annual Reviewof Sociology 6:479–508.
Bollen, Kenneth A., and Rick H. Hoyle 1990 "Perceived Cohesiveness: A Conceptual and Empirical Examination." Social Forces 69:479–504.
Borgatta, Edgar F. 1962 "A Systematic Study of Interaction Process Scores, Peer and Self-Assessments, Personality and Other Variables." Genetic Psychology Monographs 65:269–290.
—— 1981 "The Small Groups Movement: Historical Notes." American Behavioral Scientist 24:607–618.
Burgess, Robert L. 1969 "Communication Networks: An Experimental Evaluation." In R. L. Burgess and D. Bushell, Jr., eds., Behavioral Sociology: The Experimental Analysis of Social Process. New York: Columbia University Press.
Cartwright, Dorwin 1973 "Determinants of Scientific Progress: The Case of Research on the Risky Shift." American Psychologist 28:222–231.
——, and Alvin Zander, eds., 1968 Group Dynamics:Research and Theory, 3rd ed. New York: Harper & Row.
Cooley, Charles H. 1909 Social Organization. New York: Charles Scribner's Sons.
Deutsch, Morton, and Robert M. Krauss 1962 "Studies of Interpersonal Bargaining." Journal of Conflict Resolution 6:52–76.
Evans, Charles R., and Kenneth L. Dion 1991 "Group Cohesion and Performance: A Meta-Analysis." SmallGroup Research 22(2):175–186.
Evans, Nancy J., and Paul A. Jarvis 1986 "The Group Attitude Scale: A Measure of Attraction to Group." Small Group Behavior 17:203–216.
Festinger, Leon, Henry W. Riecken, and Stanley Schachter 1964 When Prophecy Fails. New York: Harper & Row.
Fiedler, Fred E. 1981 "Leadership Effectiveness." American Behavioral Scientist 24:619–632.
Forsyth, Donelson R. 1999 Group Dynamics, 3rd ed. Belmont, Calif.: Brooks/Cole-Wadsworth.
Foschi, Martha 1997 "On Scope Conditions." SmallGroup Research 28(4):535–555.
——, and Edward J. Lawler, eds. 1994 Group Processes:Sociological Analyses. Chicago: Nelson-Hall.
Goffman, Erving 1959 The Presentation of Self in EverydayLife. Garden City, N.Y.: Doubleday.
Hare, A. Paul 1982 Creativity in Small Groups. Beverly Hills, Calif.: Sage.
——, Herbert H. Blumberg, Martin F. Davies, and M. Valerie Kent 1992 Small Group Research: A Handbook. Norwood, N.J.: Ablex.
——, Edgar F. Borgatta, and Robert F. Bales, eds. 1965 Small Groups: Studies in Social Interaction, rev. ed. New York: Knopf.
Hogg, Michael A. 1987 "Social Identity and Group Cohesiveness." In J. C. Turner, ed., Rediscovering theSocial Group. Oxford, UK: Basil Blackwell.
——, 1992 The Social Psychology of Group Cohesiveness:From Attraction to Social Identity. New York: New York University Press.
Homans, George C. 1950 The Human Group. New York: Harcourt, Brace.
—— 1974 Social Behavior: Its Elementary Forms, rev. ed. New York: Harcourt Brace Jovanovich.
Janis, Irving L. 1982 Groupthink: Psychological Studies ofPolicy Decisions and Fiascos, 2nd ed. Boston: Houghton-Mifflin.
Kellerman, Henry 1981 "The Deep Structures of Group Cohesion." In H. Kellerman, ed., Group Cohesion:Theoretical and Clinical Perspectives. New York: Grune & Stratton.
Kelley, Harold H., and John W. Thibaut 1978 Interpersonal Relations: A Theory of Interdependence. New York: Wiley.
Kollock, Peter 1998 "Social Dilemmas: The Anatomy of Cooperation." Annual Review of Sociology 24:183–214.
Lewin, Kurt, Ronald Lippit, and Ralph K. White 1939 "Patterns of Aggressive Behavior in Experimentally Created 'Social Environments."' Journal of Social Psychology 10:271–299.
Marwell, Gerald, and David R. Schmitt 1975 Cooperation:An Experimental Analysis. New York: Academic Press.
McGrath, Joseph E. 1984 Groups: Interaction and Performance. Englewood Cliffs, N.J.: Prentice-Hall.
Milgram, Stanley 1974 Obedience to Authority. New York: Harper & Row.
Miller, L. Keith, and Robert L. Hamblin 1963 "Interdependence, Differential Rewarding, and Productivity." American Sociological Review 28:768–778.
Moreno, Jacob L. 1953 Who Shall Survive? rev. ed. Beacon, N.Y.: Beacon House.
Moscovici, Serge 1980 "Toward a Theory of Conversion Behavior." In L. Berkowitz, ed., Advances In Experimental Social Psychology, vol. 13. New York: Academic Press.
Ofshe, Richard A., ed. 1973 Interpersonal Behavior inSmall Groups. Englewood Cliffs, N.J.: Prentice-Hall.
Olmstead, Michael, and A. Paul Hare 1978 The SmallGroup, 2nd ed. New York: Random House.
Schmitt, David R. 1981 "Performance under Cooperation or Competition." American Behavioral Scientist 24:649–679.
——, 1998 "Effects of Reward Distribution and Performance Feedback on Competitive Responding." Journal of the Experimental Analysis of Behavior 69:263–273.
Sherif, Muzafer 1935 "A Study of Some Social Factors in Perception." Archives of Psychology 27, no. 187.
Sidowski, Joseph B. 1957 "Reward and Punishment in a Minimal Social Situation." Journal of ExperimentalPsychology 54:318–326.
Steiner, Ivan D. 1972 Group Process and Productivity. New York: Academic Press.
Tajfel, Henri 1981 Human Groups and Social Categories. Cambridge; UK: Cambridge University Press.
Thibaut, John W., and Harold H. Kelley 1959 The SocialPsychology of Groups. New York: Wiley.
Triplett, N. 1897 "The Dynamogenic Factors in Pacemaking and Competition." American Journal of Psychology 9:507–533.
Turner, John C. 1987 Rediscovering the Social Group: ASelf-Categorization Theory. Oxford, UK: Basil Blackwell.
Whyte, William F. 1955 Street Corner Society: The SocialStructure of an Italian Slum, 2nd ed. Chicago: University of Chicago Press.
Zajonc, Robert B. 1966 Social Psychology: An ExperimentalApproach. Belmont, Calif.: Brooks/Cole.
Robert W. Shotola