Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Ivan Steiner in 1974 lamented in an infamous article “whatever happened to the group in social psychology?” Certainly an examination of published research in social psychology shows a sharp decline in research on small groups, and in particular face-to-face interacting groups (Wittenbaum & Moreland, 2008). However, in an answer to Steiner’s question, Levine and Moreland (1990) replied “groups are alive and well and living in other disciplines.” Small groups can be generally defined as three or more people interacting together with a common purpose (Moreland, 2010; but see Williams, 2010). Although there are no longer concentrated centers for research on these small groups, researchers interested in their operation now have homes in a multitude of host departments, from psychology, sociology, and communications, to organizational behavior, education, and industrial engineering. A comprehensive review of literature on interaction in small groups requires a much broader search than in earlier years.

Despite this recent spread of group research across disciplines, research on small groups originated at the intersection of sociological and psychological perspectives. Sociologists were interested in how religious, political, economic, and educational systems function to sustain society, that is, how groups served to maintain the social order. For example, Durkheim (1897) described the loss of group ties as a loss of identity so important to the individual that it could lead one to commit suicide. Cooley (1902) also described the central importance of the group to the individual, suggesting for example that primary groups, such as the family, provide a focus for socialization of the individual. Psychologists were interested in issues of how individuals react in group settings. For example, LeBon (1895) described how groups or mobs could take over an individual’s will to create a collective mind. Triplett (1898), who is often credited as conducting the first experiment in social psychology, was interested in how the mere presence of others can affect our behavior. Somewhat different research traditions have grown out of those two originating fields with varying emphasis on the groups versus the individuals as the important foci, with sociology treating each person as an element of a larger system and psychology focusing on the individual as the key to understanding the group.

As small group research grew as an empirical science, theoretical perspectives emerged that reflected these two historical influences. For example, expectation states theory (Berger, Conner, & Fisek, 1974) describes the emergence and development of status hierarchies in small groups and has been particularly useful in more recent years in understanding processes that underlie role differentiation, status, and leadership in groups. Social exchange theory (Homans, 1961) described the exchange of activity, tangible or intangible, costly or rewarding, of individuals in interaction with one another and has been particularly useful in understanding processes that underlie inclusion and exclusion in groups, such as cohesion and coalition formation. The functional perspective (Hollingshead et al., 2005; McGrath, 1984) describes processes that lead groups to behave efficiently and has been particularly useful in understanding processes that underlie group performance.

These theoretical perspectives are reflected in the content sections of this chapter. Research on power and status structures in groups is somewhat more identified with the sociological tradition, whereas research on performance processes in groups is somewhat more identified with the psychological tradition. However, all of the content sections of this chapter stress interaction processes in small groups as they gain status, coordinate efforts, and exchange information.

We begin with a brief review of the primary methods used to study small groups. That is followed by a discussion of research on power, status, and leadership, reviewing research on role differentiation, the development of status structures, and theories of leadership. A discussion of inclusion and exclusion processes in groups follows, including research on cohesiveness and social identity versus ostracism, social rejection, and schisms in groups, as well as a discussion of majority and minority influence. The final content section covers motivation and coordination processes in groups, including social facilitation and social loafing, transactive memory systems, and information exchange.

Methodological Approaches

Researchers have applied a variety of methodological approaches to the study of small groups (Forsyth, 2006). Forsyth describes the three primary approaches as case studies, experimental studies, and non-experimental correlational studies.

Case studies are in-depth examinations of a single or just a few select groups. Classic examples of the case study approach include Janis’s examination of decision making in high profile groups that made historically poor decisions (Janis, 1972) and Whyte’s (1943) detailed observations of a street gang. More recent examples are Hackman’s (1990) descriptions of groups that work and groups that don’t, Pescosolido’s (2002) detailed examination of leadership and emotional processes in jazz groups and rowing crews, and Pascoe’s (2003) description of negotiated identities among adolescent boys. The advantage of the case study approach is that the researcher can create a highly in depth description of the structure and functioning of the chosen groups. However, the drawback is the potential lack of generalizability of those details to other groups (McGrath, Martin, & Kulka, 1982).

Experimental studies have been a constant presence in the field since Lewin’s influence in the late 1930s and early 1940s (Lewin, 1947; Lewin, Lippitt, & White, 1939). Lewin demonstrated that important group processes can be measured and manipulated, and his Center for Group Dynamics legitimized the study of groups as an experimental field (Lewin, 1951). Examples of contemporary research using the experimental approach are bountiful and include Williams (2009) experimental work on ostracism using a virtual ball toss paradigm (Cyberball) and Stasser and Titus’s (1985, 1987) work on information sharing in decision making groups. The advantage of experimental studies is the researcher’s ability to manipulate processes of interest in order to establish causal relationships among variables, as well as to carefully control for possible confounding variables. Drawbacks, however, include the limited number of factors that can be examined, as well as the potential lack of realism of context and lack of representativeness of many research participants (McGrath, Martin, & Kulka, 1982).

Non-experimental approaches explore the correlational relationships among variables, such as in Newcomb’s (1943) classic study of opinion change among college students over time. Many recent examples of non-experimental approaches can be found in organizational psychology, where the ecological validity of using intact groups often takes precedent over experimental rigor (Austin, 2003; Cole, Walter, & Bruch, 2008). Although interesting relationships can be documented in this manner, causal relationships cannot be established given that there is no manipulation of potential causal agents.

Perhaps the most interesting methodological contributions of group research have been in the creation of observation systems or other methods for investigating unique group processes. Much of what we know about role differentiation in groups is based on research by Bales’ (1950) using the Interaction Process Analysis (IPA) system. The IPA system classifies group behaviors into one of 12 categories, half of which quantify task activities (e.g., questioning or attempting answers) and the other half of which quantify socioemotional activities (e.g., positive actions such as agreement and negative actions such as disagreement). At one time, the IPA was the most widely used method for studying group interaction. In the more recently developed SYMLOG system (Systematic Multiple Level Observation of Groups; Bales, 1980, 1988), observers code group members’ behaviors in terms of the dimensions of dominance/submissiveness, friendliness/unfriendliness, and acceptance/non-acceptance of authority. The coding of interaction content has risen again recently with the development of the information sharing paradigm mentioned above, given its fairly simple coding of information that is contributed to the group discussion as either shared across members or unique to specific group members.

Another assessment technique that was developed for studying patterned relationships among group members by mathematically indexing group relations was sociometry (Moreno, 1934, 1960). In this approach, individuals are asked an attraction question regarding their group members (e.g., whom do you like the most?). Based on these responses, researchers construct a sociogram, or a graphical representation of a pattern of relationships among group members. Various group members’ positions in the sociogram can be described using a mathematical summary. Research using sociometry also seems to be increasing. More advanced statistical techniques have aided in this mathematical representation (Wasserman & Faust, 1994), as has the development of computer programs such as Sociometrics (Walsh, 2003) that can generate sociograms.

These different methodological approaches tend to characterize different topical areas of research. Research on power and status has tended to use non-experimental approaches, although research on leadership in particular has combined different approaches. Research on inclusion and exclusion processes has tended to use experimental approaches with many exceptions. Research on motivation and coordination has been somewhat exclusively dominated by experimental approaches. These areas are reviewed more thoroughly in the following sections.

Power, Status, and Leadership

The roles of power, status, and leadership have been at the forefront of small groups research since early work by Lewin and Bales. This section begins with a discussion of role differentiation, followed by work on status structures with a theoretical focus on expectation states theory. We conclude with a discussion of leadership, a topic that continues to generate abundant research.

Role Differentiation

As mentioned above, work on role differentiation comprises some of the earliest and most influential small group research on status and power. The current section focuses on Bales and Moreno’s early contributions to this area.

Bales (1950, 1970) documented the development of roles in unstructured groups of Harvard students, finding that stable roles developed within the first hour of interaction. For example, the most talkative member in the group was addressed most by other group members, his/her ideas were rated the most favorably, and he/she was indicated as the leader. Similar effects have also been found in more heterogeneous groups such as juries (e.g., Strodtbeck, James, & Hawkins, 1957).

Bales and colleagues also found evidence for the emergence of two types of leaders: task leaders and socioemotional leaders (Bales & Slater, 1955). Task leaders are those group members who engage in behaviors that are designed to facilitate task completion. Task activities involve behaviors such as delegation and the coordination of group communication. Socioemotional leaders are those group members who engage in behaviors that are designed to facilitate positive relations within the group. Socioemotional activities involve behaviors such as encouragement and compromising. Bales and colleagues argue that task and relationship roles emerge because these activities are often at odds with one another (Bales, 1955, 1958). When task leaders work towards the completion of goals, they necessarily engage in behaviors that do not promote positive group relations, such as criticizing and giving orders. Thus, most groups have two leaders: one specializing in task activities and the other specializing in socioemotional activities.

Bales investigated the emergence of task and relationship roles using Interaction Process Analysis (IPA) (Bales, 1950), which as discussed above, classifies communicative behaviors into either task or socioemotional activities. Bales found that the majority of group members performed either task or socioemotional activities, with few group members performing both. Although in the first session a little over half of leaders engaged in a good deal of task and socioemotional activities, by the fourth session less than 10 % of leaders specialized in both types of activities (Bales, 1953, 1958; Bales & Slater, 1955; Slater, 1955). It is important to note, however, that these role differentiation effects are not obtained when the task leader is assigned by the experimenter (Burke, 1968), nor when group members are given strong motivation to complete the task (Burke, 1967). Thus, role differentiation between task and socioemotional activities may be more likely in low legitimation circumstances (Burke, 2003).

Moreno (1934, 1960) also conducted early work on role differentiation using the sociometric technique described above. By assessing patterns of attraction and influence among group members, the sociogram depicts the group structure and can be used to calculate roles. For example, those group members who are liked by many other group members have high status and are called “populars.” In sum, work by Bales and Moreno set the groundwork for studying status and power in small groups.

Status Structures

A wealth of research has come out of Bales’ early work on role differentiation. In particular, Berger and colleagues developed expectation states theory in an effort to understand the development of group status structures (Berger et al., 1974; Berger, Fisek, Norman, & Zelditch, 1977; Berger & Zelditch, 1998). This theory was developed to explain the status structure specifically in task groups that have a collective orientation (Correll & Ridgeway, 2003). A collective orientation is the perception of group members that working in a group is necessary to complete their task goals.

The main idea of expectation states theory concerns the influence of performance expectations on the development of status hierarchies. Expectation states theory argues that people use available social status cues to form performance expectations, anticipations of each group members’ likelihood of making meaningful contributions to the group task. These expectations are often unconscious, but substantially influence subsequent behavior. Those group members who are expected to make more valuable contributions to the group are given more opportunities to contribute to the group. For example, they are given more opportunities to speak and offer ideas. Additionally, these ideas are more likely to be perceived positively by group members. Thus, performance expectations influence the development of status hierarchies in self-fulfilling ways.

Expectation states theory posits three processes that impact the development of performance expectations: socially significant status characteristics, social rewards, and behavioral interchange patterns (Correll & Ridgeway, 2003). Socially significant status characteristics are characteristics that differ across people and for which there are normative beliefs regarding the advantages of certain characteristics over others. Expectations states theory posits two types of status characteristics, specific and diffuse. Specific characteristics are those characteristics associated with performance expectations for only a certain set of relevant tasks. For example, group members’ mathematic ability may be used as a status characteristic in groups performing a primarily mathematical task, but not in groups performing a primarily verbal task. Diffuse characteristics are those characteristics associated with general performance expectations. For example, gender is a diffuse characteristic, as men are generally expected to be more competent than women (Ridgeway & Bourg, 2004). Other examples of diffuse characteristics include race, physical attractiveness, and seating position (Berger, Webster, Ridgeway, & Rosenholtz, 1986; Strodtbeck & Hook, 1961).

Although the majority of work on performance expectations focuses on the influence of socially significant status characteristics, social rewards and behavioral interchange patterns can also play important roles in the development of performance expectations. In particular, individuals form performance expectations that are congruent with the relative distribution of rewards (Berger, Fisek, Norman, & Wagner, 1985; Stewart & Moore, 1992). Thus, performance expectations can serve to legitimize the differential allocation of rewards among group members. Individuals also form performance expectations that are congruent with behavioral interchange patterns, such that group members who exhibit more assertive and dominant behavior (e.g., high frequencies of participation) are associated with more favorable performance expectancies than group members who exhibit more deferential behavior. Although space constraints prevent a more detailed discussion of expectations states theory, a great deal of research has grown out of this theory (for a more in-depth review see Correll & Ridgeway, 2003).

Leadership

Perhaps the majority of current small group work on power and status concerns leadership. Indeed, there are a variety of current psychological and sociological journals devoted to the study of leadership. Work on leadership often explores either leader emergence or leader effectiveness (Forsyth, 2006). As the previous section briefly mentioned some leadership emergence work, the current section focuses on leadership effectiveness.

Psychological and sociological research on leadership effectiveness has a long-standing tradition, stemming in part from seminal work by Lewin and colleagues (1939). These researchers proposed three types of leadership: authoritarian leadership, in which the leader does not take followers’ opinions into account, democratic leadership, in which decisions are made by the group as a whole, and laissez-faire leadership, in which the leader seldom engages with the group. These types of leadership were shown to differentially affect group processes, including the time followers spent working when the leader was present, the time followers spent working when the leader was not present, and aggressive behaviors. In general, the followers favored democratic leaders over authoritarian and laissez-faire leaders.

More recent work also posits leadership typologies that impact group performance. For example, the theory of transformational leadership (Bass, 1997) distinguishes between transactional and transformational leadership. Transactional leadership involves an emphasis on clear expectations, rewards, and punishments, including active strategies such as managing by exception and passive strategies such as addressing severe problems. Transformational leadership is associated with charisma, inspiration, intellectual stimulation, and individualized consideration. Empirical support has been obtained for Bass’s assertion of the superiority of transformational leadership on some outcome measures, such as productivity and follower satisfaction (Conger, 1999; Lowe, Kroeck, & Sivasubramaniam, 1996; Wang, Oh, Courtright, & Colbert, 2011).

Other theoretical approaches to leadership, such as the contingency model, the situational leadership model, and leader-member exchange theory, propose that situational factors operate in concert with leadership style to impact leader effectiveness. Fiedler’s (1978, 1981, 1996) contingency model suggests that leadership effectiveness is a function of the leader’s motivational style and situational control. Motivational style is measured using the least preferred co-worker scale, a uni-dimensional measure indexing degree of task- versus relationship-orientation. A leader’s situational control is posited to be a function of leader-member relations, task structure, and leader power. Favorable situations involve good leader-member relations, structured tasks, and strong leader power. Task oriented leadership is expected to be associated with greater effectiveness in either highly favorable or unfavorable situations, while relationship oriented leadership is expected to be preferable in moderately favorable or unfavorable situations.

The situational leadership model also proposes that leadership effectiveness is a function of both leader behavior and group situation (Hersey & Blanchard, 1976, 1982; Hersey, Blanchard, & Johnson, 2001). This model suggests that leader behavior varies along two dimensions, supportive and directive, and that leadership effectiveness depends on a match between this behavior and the maturity of the followers. While initially followers require high degrees of directive behavior and low degrees of supportive behavior, high degrees of supportive behavior are also required as time progresses. When the group is near to their goal, followers may not require much leadership, benefiting from low degrees of both directive and supportive behavior.

Finally, leader-member exchange theory emphasizes the dyadic relationships between the leader and each single follower and its impact on leadership effectiveness (Deluga, 1998; Linden, Sparrowe, & Wayne, 1997). This theory suggests that leaders have different relationships with each follower, and particularly good working relationships are associated with a variety of positive outcomes. For example, in these relationships leaders may provide followers with greater responsibilities or rewards. Additionally, these followers express greater satisfaction with the leader and are less likely to leave the group (Gerstner & Day, 1997). These relationships often also affect group structure, as groups divide into an inner group, those followers with particularly positive relationships with the leader, and an outer group, those individuals with less positive relationships with the leader.

Gender and Leadership

Within work on leadership there is a substantial sub-literature investigating gender and leadership. Although recent progress towards gender equality in leadership has been made, women continue to be underrepresented in leadership positions such as electoral offices and Fortune 500 committee board chairs (Catalyst, 2011; Center for American Women and Politics, 2010). Research in this subarea explores gender differences in reactions to and perceptions of leaders, in addition to gender differences in leadership emergence and style.

Traditional gender stereotypes describe and prescribe women as having communal characteristics, such as kindness and sensitivity, and men as having agentic characteristics, such as self-confidence and aggression (Burgess & Borgida, 1999). However, traditional leadership stereotypes are primarily agentic. This incongruity between females and leadership stereotypes puts female leaders at a disadvantage, requiring them to achieve a delicate balance of agentic and communal behavior (Eagly & Sczesny, 2009). Notably, this incongruity can be thought of in the context of expectation states theory, as gender is a diffuse characteristic that signals competence (Ridgeway & Bourg, 2004).

The incongruity between females and leadership stereotypes can result in women receiving backlash for success in stereotypically male domains and in less favorable attitudes towards female leaders as compared with male leaders (Eagly & Karau, 2002; Eagly, Makhijani, & Klonsky, 1992; Heilman, Wallen, Fuchs, & Tamkins, 2004). Research also suggests that stereotypes play a role in ratings of leader effectiveness. For example, leader effectiveness ratings are more favorable when the description of the leader’s role is congruent with the leader’s gender (Eagly, Karau, & Makhijani, 1995).

Although backlash against agentic women continues to be demonstrated, there is evidence that leadership stereotypes are becoming less masculine (Koenig, Eagly, Mitchell, & Ristikari, 2011). Thus, while early work on gender and leadership investigated the “glass ceiling,” a concealed barrier to women’s rise in leadership, more recent work takes a different approach, suggesting that the obstacles to female leadership are better described by the term “labyrinth,” which emphasizes their intricate nature (Eagly & Carli, 2007). For example, work by Ryan and Haslam (2007) suggests that female leaders may be disadvantaged as they are perceived as better suited than men for risky leadership positions, a phenomenon they refer to as the glass cliff. By appointing women to more risky leadership positions than men, female leaders are more likely to fail than male leaders. Thus, this differential appointment can serve to perpetuate the stereotype that men are better leaders than women.

Although there is little evidence of subject gender effects on reactions towards female leaders, there is evidence that women operate differently than men in groups with respect to status behaviors. In mixed sex groups men tend to engage in greater task-oriented behavior than women, including talking more and making more suggestions regarding the task (Ridgeway, 2001; Ridgeway & Smith-Lovin, 1999). In mixed-sex groups men also tend to engage in more nonverbal displays of power than women, including looking more at group members when speaking but looking less at group members when listening (Dovidio, Brown, Heltman, Ellyson, & Keating, 1998). However, these verbal and nonverbal findings reverse when groups engage in stereotypically feminine tasks (Dovidio et al., 1998). Additionally, there is little evidence for similar gender differences in same-sex group behavior (for exceptions see Hutson-Comeaux & Kelly, 1996; Ridgeway, Diekema, & Johnson, 1995).

Gender differences have also been documented in leadership emergence and style. Eagly and Karau’s (1991) meta-analysis suggests that men are more likely than women to emerge as leaders of initially leaderless groups. However, there were moderators of this effect. Men were especially likely to emerge as leaders in short-term groups, and consistent with gender stereotypes, women were more likely than men to emerge as social leaders.

In addition to work demonstrating differences in the circumstances in which women and men emerge as leaders, gender differences have also been documented in leadership style. Eagly and Johnson’s (1990) meta-analysis suggests that male leaders tend to be task-oriented and more autocratic while female leaders tend to be interpersonally-oriented and democratic. Notably, female leaders were evaluated more negatively to the extent that their leadership style reflected more masculine, autocratic elements (Eagly et al., 1992). More recent work also suggests that female leaders are more likely than male leaders to engage in transformational leadership and contingent reward behaviors, one component of transactional leadership (Eagly, Johannesen-Schmidt, & van Engen, 2003). There is also evidence suggesting that all female groups develop less centralized leadership structures than all male groups, a structure that can be associated with positive outcomes including better performance (Berdahl & Anderson, 2005).

Current Directions in Leadership

As mentioned earlier, leadership research is currently a very popular topic of study within small group work on power and status. Similar to small group work in general, research on leadership spans a variety of disciplines, such as social psychology, industrial/organizational psychology, and sociology. To conclude our discussion of power and status, we briefly discuss some of the current directions in leadership research.

Recently, researchers have not only been interested in exploring ideal forms of leadership, but in exploring destructive leadership as well. Einarsen, Aasland, and Skogstad (2007) define destructive leadership as leadership that results in outcomes that are detrimental to the legitimate goals of the organization. There are many different types of destructive leadership, although laissez-faire leadership may be the most common (Aasland, Skogstad, Notelaers, Nielsen, & Einarsen, 2010). Destructive leadership has been linked to individual difference variables such as low self-efficacy, as well as situational variables such as the support of authority (Mumford, Gessner, Connelly, O’Connor, & Clifton, 1993). Indeed, theoretical work by Padilla, Hogan, and Kaiser (2007) proposes that the damaging outcomes of destructive leadership are a function of destructive leaders, susceptible followers, and environments that are conducive to destructive leader behaviors.

Recently, small groups researchers have also been interested in the effects of mood on group dynamics. In particular, researchers have been exploring the process of emotional contagion, whereby moods are transferred among group members (Barsade, 2002; Kelly & Barsade, 2001). This interest has also expanded to work on leadership. For example, Sy, Cote, and Saavedra (2005) demonstrated that the mood of group leaders transfers to their followers and can subsequently affect group dynamics. Groups in positive moods demonstrated greater coordination but less effort than those groups in negative moods.

Research has also explored the effects of leadership prototypicality, often from a social identity perspective (discussed in greater detail in the upcoming section). Generally, leaders who are more prototypical of the groups they lead are linked to a variety of positive outcomes. For example, prototypical leaders are evaluated more positively than non-prototypical leaders following a failure to reach an ideal goal (Giessner & van Knippenberg, 2008). These positive perceptions of prototypical leaders are most likely to occur for group members that are highly identified with the group (Platow & van Knippenberg, 2001). This relationship is due in part to the fact that highly identified group members perceive leaders as acting in line with the group’s interests (van Dijke & De Cremer, 2010).

Processes of Inclusion and Exclusion in Groups

This section will discuss processes of inclusion and exclusion in groups. We will discuss why people exclude others and detail the consequences of exclusion. Schisms in groups, particularly coalition formation, will also be discussed in this section. Finally, minority and majority influence will be discussed including details of how these factions form and the benefits and consequences of their formation.

Inclusion Processes in Groups

Attraction (how attracted an individual is to the group), liking (how much an individual likes the group), and staying (how much the individuals want to stay in the group), are three things that determine a group’s cohesion (Forsyth, 2006). Schachter (1953) defined cohesion as the “total field of forces acting on members to remain in the group.” This definition combined many different aspects of cohesion into one. However, Hogg (1987) separated attraction to the group itself (social attraction) from attraction to members in the group (personal attraction). Thus, Hogg considers the attraction to the group itself as the main determinant of cohesion.

Many different theories provide explanations as to how groups become cohesive. Lawler’s exchange theory perspective discusses how emotions may influence group cohesion. This theory suggests that multiple positive interactions lead to positive emotions that will lead to cohesion (Lawler, 1999, 2000). Mutual self-verification may also play a role in how groups become cohesive. Mutual self-verification occurs when group members develop a mutual dependence by each verifying others’ identities while in the process of verifying their own group identity (Burke & Stets, 1999). This leads to increased trust and commitment resulting in positive emotions toward the group that increase cohesion. Cohesion can also develop through the reduction of cognitive dissonance, whereby individuals are motivated to relieve the discomfort they feel when they experience inconsistency (e.g., undergoing a negative initiation into a group that one doesn’t feel very positively about) (Festinger, 1957). Aronson and Mills (1959) demonstrated this when they had participants engage in a severely embarrassing test, a mildly embarrassing test, or no test at all in order to participate in a group discussion. Those who engaged in the severely embarrassing test rated the group discussion and the group members more positively than those in the other conditions. This presumably occurred because the high cost of taking the embarrassing test caused people to change their beliefs about the group in order to justify the embarrassment.

Impact of Cohesion

Cohesive groups sound like they should work well together and be effective entities. Therefore, a lot of effort is put into making teams cohesive, for example with businesses employing team-building activities. However, while cohesion may result in positive effects for the group, it can also lead to negative effects.

Though seldom empirically studied, cohesion is a necessary antecedent condition for the phenomenon of groupthink (Janis, 1972, 1982). Groupthink occurs when highly cohesive groups tasked with making a decision attempt to minimize conflict to the detriment of actually discussing and evaluating alternative courses of action. Groupthink has been implicated in several catastrophic events (Janis, 1971; Moorhead, Ference, & Neck, 1991). For example, in January of 1986, the Challenger shuttle exploded after take-off. After investigating why this tragedy occurred, it was determined that a faulty piece of equipment and abnormally cold temperatures had contributed to the explosion. Engineers had expressed concerns about this equipment, but because the launch had already been delayed for many days, the officials were eager to launch the shuttle immediately and encouraged the engineers to change their diagnosis. This led to the engineers proclaiming that the shuttle was good to launch, leading to a national tragedy. These poor group decisions were believed to have occurred due to groupthink (Moorhead et al., 1991).

Highly cohesive groups may encourage conformity to group norms and ideas and discourage dissent amongst group members. If individuals do not provide alternative options or produce unique output, it could lead to poorer decision-making. On the other hand, meta-analytic studies have shown that cohesion is generally related to better performance (Mullen & Copper, 1994). However, the reverse relationship, with better performance leading to improved cohesion, is stronger. In addition, cohesion that is based on commitment to the task is more strongly related to performance than is cohesion that is based on interpersonal attraction to group members.

A largely accepted view of cohesion is that members of highly cohesive groups will adopt and adhere to the norms of that group, and therefore cohesive groups will perform well or poorly depending on the level of these performance norms. When the group has a norm of high productivity and high performance, then cohesive groups will perform well. However, when the group has a norm of low productivity and low performance, then cohesive groups will perform poorly. Seashore (1954) found that cohesive groups with high performance goals were more productive, and cohesive groups with low performance goals were less productive, demonstrating that the norms of highly cohesive groups determine how the group performs. Gully, Devine, and Whitney (1995) similarly suggest that if the group develops norms that encourage playfulness, openness, and disagreement, performance may improve. If dissent is encouraged, people are free to speak up about new and different ideas and suggest alternatives to decisions that may improve performance. However, if the norms that a group develops lead to conformity or take attention away from the task at hand, performance will suffer (Gully et al., 1995).

Social Identity Theory

Other potential sources of cohesion are identified by social identity theory, which suggests that some sources of cohesion can lead to positive consequences for one’s ingroup while leading to negative consequences for one’s outgroup. Social identity theory explains how an individual’s identity may be derived from their group membership. It suggests that an individual’s self-concept is tied to their group membership, primarily through the process of social categorization. Individuals distinguish between their ingroup and their outgroup in ways that enhance the positive qualities of the ingroup, and thus enhance their self identity. Notably, although the current work focuses on social identity theory as it applies to small groups, this theory also applies to large groups and social categories (see Chap. 18 by Hogg, 2013).

Social identity theory posits several important principles (Brewer, 2010). First, the ingroup accentuation principle states that perceived within-group differences are minimized and perceived between-group differences are maximized. Second, the ingroup favoritism principle suggests that people selectively associate positive characteristics with members of the ingroup but not with members of the outgroup. Finally, the social competition principle suggests that there is comparison and perceived competition between groups (Turner, 1975).

These three principles help to accomplish the motives associated with social identity theory. The self-esteem enhancement motive suggests that attributing positive characteristics to the ingroup, and simultaneously perceiving few within-group differences, allows people to be able to vicariously enhance their self-esteem through their group identity (Abrams & Hogg, 1988). An example of how self-esteem enhancement relates to social identity theory is through the process of BIRGing (Basking In Reflected Glory) (Cialdini et al., 1976). When a person associates the self strongly with a group and a member of that group succeeds at something, the other members can then associate that positivity with themselves, thus enhancing their own self-esteem.

Another possible motive considered by social identity theory is a need for security and belonging. Baumeister and Leary (1995) suggest that there is a universal need for belonging that motivates us to attach ourselves to groups. However, Brewer’s (1991) theory of optimal distinctiveness suggests that this motive may be selectively applied. This theory posits two identity needs: inclusion, which is satisfied through the ingroup, and differentiation, which is satisfied through differences between the ingroup and outgroup. Thus, there is an optimal level of distinctiveness that balances differentiation and inclusion. Research has shown that threats to inclusion enhance self-stereotyping on traits characteristic of the group (Brewer & Pickett, 1999). Thus, it seems a need for security and belonging may be a motive that acts through finding the perfect balance between inclusion with the ingroup and differentiation from the outgroup.

Social identity theorists have used the minimal group paradigm to investigate many of these propositions (Tajfel, 1970). This paradigm allows researchers to examine the minimal conditions required for intergroup dynamics to occur. In this paradigm, groups are formed based on arbitrary traits that individuals share. For example, groups may be formed based on preferring a certain artist (Tajfel) or being overestimators or underestimators on a dot-estimation task (Gerard & Hoyt, 1974). When using this paradigm researchers have found that individuals will allocate more rewards to ingroup members than to outgroup members (Tajfel, Billig, Bundy, & Flarnent, 1971) and individuals will evaluate ingroup members more positively than outgroup members (Brewer, 1979). The minimal group paradigm has shown that intergroup bias can develop in the absence of interaction, history, personal identification of group members, or any direct benefits to the self (Brewer, 2010).

Outgroup derogation, or negative behavior directed towards other groups, has also been explained by social identity theory. However, it is not thought to be an inevitable component. Research using the minimal group paradigm has shown that while individuals consistently provide the ingroup with more rewards, they are not very likely to provide the outgroup with more negative outcomes.

The famous Robber’s Cave experiment provides an example of social identity theory at work (Sherif & Sherif, 1953). Young boys were admitted into an experiment under the guise of going to summer camp. The boys first interacted with all of the others at the camp. They ate meals together and engaged in activities together. However, after a bit of time, the camp counselors (experimenters) separated the boys into two different groups (the Eagles and the Rattlers). Once separated, the boys ate meals and engaged in activities only with the boys in their own groups and not with the other boys. Competition between the two groups became fierce and liking and cooperation within groups increased. Putting labels on the two separate groups led to increased positivity within each group and negativity toward the other group.

There has also been qualitative research done examining social identity in adolescents in everyday life. Adler and Adler (1995) discuss how cliques function as a socialization experience, particularly through inclusion and exclusion. Cliques are friendship groups with internal hierarchies that are exclusive in that they do not grant membership to anyone who wants to join, thus forming a strong social identity for those in the clique and perhaps even those not in the clique (through their identification as an outsider). There are efforts to become included in cliques by those not in them. Additionally, once in a clique, there are inclusionary efforts directed toward reaching a higher position within the clique or simply maintaining membership in the clique. These techniques include recruitment (being invited to the group), application (seeking membership), realigning friendships within the clique, and ingratiation. Because cliques are exclusive in nature, they exclude others by requiring permission to join. There are many exclusionary techniques used by cliques, such as subjugating the out-group (including excluding and rejecting children not in the clique), subjugating those within the group (usually done by higher status members to members with lower status), compliance (getting lower status members to do things they don’t necessarily want to do), stigmatization (particularly of in-group members), and expulsion from the clique. If expulsion occurs, it is often difficult to form new friendships because the people outside the clique remember how former members treated them and see them as elitists. This shows how social identity can be both a benefit to those in the in-group, and a detriment to those in an out-group, yet it also displays how the benefits of being in the in-group are potentially harmful.

More qualitative research on identity in adolescents has shown the importance of the masculine identity to adolescent boys (Pascoe, 2003). In high school, athleticism and sports are key to this masculine identity, and so the “jocks” are the epitome of masculinity. However, not every teenage boy is a “jock.” Those who are not create their own way of displaying a masculine identity through their associated group identity. For example, one boy who was involved in drama and performing in plays discussed how the drama parties are very sexually charged and emphasizes his pride in flirting with girls during his first play. Thus, he is able to emphasize an overall masculine identity while also emphasizing his identity with his social group.

Processes of Exclusion in Groups

Being included in groups is important to the self and an individual’s need for belonging. More importantly, having an ingroup that satisfies the need for inclusion and an outgroup that satisfies the need for differentiation seems to help individuals reach an optimal level of distinctiveness. However, people also exclude members from their groups. Although excluding members of one’s own group may be seen as more unusual than including them, we discuss why people may exclude fellow group members and the consequences of such exclusion.

Belonging is considered a fundamental human need necessary for survival (Baumeister & Leary, 1995). Evolutionary theory suggests that belonging developed as a need because in prehistoric times people needed the protection of a group to survive. Belonging to a group helped people retrieve food and water necessary for survival and provided safety. When people’s belonging needs are thwarted, they experience exclusion, rejection, loneliness, or even nothing at all (a numbness). In addition to threatening the need to belong, being excluded also threatens individuals’ needs for self-esteem, control, and a meaningful existence. Experiencing exclusion then leads to the activation of an internal system designed to help fortify the threatened need for belonging (Williams, 2009).

This system includes three stages: reflexive, reflective, and resignation. The first reflexive stage occurs quickly and crudely. Because belonging to a group was necessary for survival, Williams (2009) proposes that there is likely to be over-detection of ostracism (being excluded and ignored) in order to protect people from being alone and vulnerable. There are very few moderators of this reflexive stage of ostracism, and people’s needs are depleted regardless of the circumstances. Much of the experimental research on ostracism has been conducted with Cyberball, a virtual ball-toss game where pre-programmed figures throw a ball back and forth to each other and to a naïve participant. In the ostracism condition the participant is thrown the ball a couple of times and then never again. People feel ostracized when they are not thrown the ball and report negative psychological and emotional consequences. Research has shown that people detect and react to ostracism even when they are told the other opponents are a computer (Zadro, Williams, & Richardson, 2004), when they believe they are being ostracized by hated groups (e.g., the KKK) (Gonsalkorale & Williams, 2007), when the ball is a “bomb,” and when being thrown the ball decreases one’s payoff (van Beest & Williams, 2006). Additionally, the phenomenon of being “out-of-the-loop” occurs when people are excluded on particular information, yet they are not ignored, thus representing a form of partial ostracism (Jones, Carter-Sowell, Kelly, & Williams, 2009). While this may seem like a minor form of ostracism, people who are out-of-the-loop also experience the depleted needs and negative emotions that occur with full ostracism experiences.

The reflection stage occurs when people attempt to repair the negative psychological consequences of being ostracized. People may respond in one of two ways – through aggression in an effort to fortify their need for control, or through attempting to be included which fortifies the need for belonging. For example, female participants were less likely to socially loaf after being ostracized (Williams & Sommer, 1997), perhaps due to a motivation to gain favor with those who ostracized them. On the other hand, when participants were ostracized and stripped of control, they were more likely to show aggressive behavior towards a stranger by allocating more hot sauce to a person who did not like hot sauce (Warburton, Williams, & Cairns, 2006). Additionally, Wesselmann, Butler, Williams, and Pickett (2010) had naïve participants “get-acquainted” by interacting with other participants (confederates) and then told them that either all or none of the group members wanted to work with them further. When participants expected rejection they were less aggressive, but when participants did not expect rejection their need for control was depleted and thus they were more aggressive. Thus, the reflection stage can lead to fortifying the need for belonging through prosocial acts, whereas fortifying the need for control can lead to aggressive acts.

The final resignation phase deals with the long-term effects of chronic ostracism. Two key components of this phase are affective numbness and a lack of self-regulation (Baumeister, DeWall, Ciarocco, & Twenge, 2006). Williams’ (2009) model suggests that over time a person’s resources become depleted and they are no longer able to fortify their threatened needs. However, there is little experimental research on this phase.

Power structures can affect whether certain people are excluded. Markovsky and colleagues (Markovsky, Skvoretz, Willer, Lovaglia, & Erger, 1993) suggest that when there is a stronger power structure, such as in organizations, it is easier to exclude those lower in the hierarchy. However, when there is a weaker power structure, such as in friendship groups, where everyone is equal and there is no hierarchy, it is more difficult to exclude others. The power structure can also affect how a group functions. When there is a stronger power structure, people may be less committed as they feel their position is not secure. Thus, people may take precautions against investing too much in those who may later abandon them, leading them to feel the negative effects of exclusion.

Schisms in Groups

When groups experience difficulty operating or clash on their opinions or ideas, schisms may occur leading to the development of factions or coalitions. Schism refers to when members of a group decide to leave the group to form a separate group or to join an already existing but different group (Sani & Todman, 2002). Caplow (1956) developed a typology of three-person groups useful for predicting the factions that may form in them. The coalitions that form include everyone being of equal strength, everyone being unequal in strength, or somewhere in between these extremes. For example, one member may have slightly more strength than the other two, one member could have more strength than the other two combined, or the combined strength of any two members could be more than the third alone. One consistent finding was that the weakest member was likely to initiate the formation of a coalition (Vinacke & Arkoff, 1957). However, this research was criticized for only focusing on the outcome of coalition formation and not on the process that led to that outcome.

Sometimes coalitions can no longer function within the group. When this occurs, group schisms can result where a former subgroup leaves the group entirely to form a new group or join with a different existing group. Sani and Todman (2002) developed a model detailing what makes schisms occur in groups. Schisms often occur because a new norm changes a central aspect of the group’s identity. Sani (2005) later found that first there is a feeling that the group identity has been changed, which leads to negative emotions and a decrease in group identification and perceived group entitativity. The combination of decreased group identification and negative emotions then increase schismatic intentions.

Minority-Majority Influence

While schisms can be very negative, sometimes having a faction with a different opinion may not be detrimental to the group and may even improve its functioning. Having minority opinions in groups often leads to positive consequences for the group and its effectiveness, despite the fact that majorities often prevail over minorities. This section will detail how majorities and minorities develop and the ensuing consequences of having minorities and majorities.

When a group has a majority it may be difficult for those who don’t agree with them to express their opinions. Asch (1956) first demonstrated this by having participants make a judgment about which of three lines matched a target line. On critical trials, confederates chose a noticeably incorrect line to see if the participant would conform to the incorrect majority. Asch found that participants conformed about one third of the time. He also found that when even one confederate disagreed with the incorrect majority, the participant was more likely to disagree as well.

One important key to majority influence is that they stick together and do not falter. If one person dissents, it leads the way for others to dissent and minorities may form within the group. Latané and Wolf (1981) suggest in their social impact theory that when the size of the majority is kept constant, increasing the size of an opposing minority leads to less conformity to the majority. They also suggest that the first few minority members have the most impact, and each additional minority member adds less impact than the previous member. Tanford and Penrod’s (1984) social influence model suggests something similar in that after the first few minority members, each additional minority member does not increase the amount of minority influence. Their model also showed that minority influence was most likely to occur in smaller groups with at least two minority members and when the minority was consistent.

Moscovici (1985) developed a dual-process theory of minority versus majority influence that described when each would be likely to emerge in a group and when they would have influence. Conflicts arise in groups due to differing viewpoints which leads one of two salient norms to be activated – a norm of reduction of conflict or a norm of originality. When a norm of reducing conflict arises, the conflict will be resolved through normative influence and agreement with the majority because majorities hold status. A comparison process occurs whereby minorities will compare their views with the majority view. Agreement with the majority results because of a need for consensus, not because of an actual change in viewpoints, leading to compliance with the majority’s viewpoint (i.e., public agreement without private acceptance).

On the other hand, the norm of originality leads to the goal of creating conflict rather than reducing it in order to challenge the majority view. If a minority is consistent and confident, the others will undertake a validation process whereby they critically evaluate the minority viewpoint and attempt to verify it with reality (Moscovici, 1985). This can lead to conversion to the minority viewpoint, which is shown through private judgments and indirect measures but not necessarily publicly (for fear of being seen as a deviant). A meta-analysis conducted on minority influence showed that consistent behavioral styles mediated minority influence (Wood, Lundgren, Ouellette, Busceme, & Blackstone, 1994). Wood and colleagues (1994) also showed that minority influence had the largest impact on private and indirect measures rather than on the public opinion. Thus, both majority and minority factions may exert influence on the group, though through different means.

As seen in Moscovici’s theory, minorities may stimulate different thought processes in groups. Minorities are more likely to appear and be convincing when a norm of originality is salient. Nemeth’s (1986) theory of convergent-divergent thinking also suggests that minorities can be beneficial to a group. In her view, it is not important that majorities conform to the minority, but rather that divergent thought processes accompany a minority viewpoint. Divergent thinking leads to the generation of more original ideas and the detection of more unique and novel solutions. Majority viewpoints, on the other hand, lead to convergent thinking, which involves focusing on the majority position and no other possibilities. Convergent thinking can therefore lead to poorer decisions and less original ideas. It often does not even matter if the minority is correct or incorrect, but rather simply that different types of thinking emerge when a minority opinion is present (Nemeth, 1976; Nemeth & Wachtler, 1983). In addition to coming up with more original ideas, groups with a minority have also been shown to make better decisions (Schulz-Hardt, Brodbeck, Mojzisch, Kerschreiter, & Frey, 2006).

While the evidence supporting minority viewpoints leading to more creativity is very strong, majority viewpoints can also lead to more creativity. When judges rated arguments generated for a position by the minority or the majority, they found that the minority arguments were more original and convincing than the majority arguments (Kenworthy, Hewstone, Levine, Martin, & Willis, 2008). However, when comparing minority arguments, majority arguments, and arguments of people in an “equal-faction” condition, both the minority and majority arguments were seen as stronger and more persuasive than those in the equal-faction condition, though the minority’s arguments were rated as more creative than both the majority’s and the equal-faction’s (Kenworthy et al., 2008).

If minorities are successful in converting others to their view then they become the majority, which can have both positive and negative effects on both the minority and the majority. Those losing the majority position experienced decreases in perceptions of group-self similarity, attraction to the group, and expectations of positive interactions with the group (Prislin, Limbert, & Bauer, 2000). However, those gaining the majority position did not experience increases in these factors. Thus, Prislin and colleagues (2000) suggest that after changes in power, a group may be particularly fragile due to the disintegrative forces that result from the loss of power of the previous majority. When these power changes occur, the new minority group experiences a decrease in desire to belong to the group. Over time, there was no positive evaluation of the group from the new majority. There was, however, an increase in preference for belonging to the group from the new majority (Prislin & Christensen, 2005). So the new minority experiences more negative consequences, and at least initially, does not necessarily experience more positive consequences. In addition, new majorities also show stronger identification with the group when they think people moved to their new positions for genuine reasons, as opposed to superficial or unexplained reasons (Prislin, Levine, & Christensen, 2006). A more negative finding was that new majorities who were low in perceived control abused their newfound power by displaying in-group favoritism and outgroup hostility (Prislin, Sawicki, & Williams, 2011). Out-group hostility was more likely to occur when the new majorities were low in perceived control and received social support for their negative behaviors. Thus, it seems there may be more negative consequences to a drastic power shift in a group than positive.

Motivation and Coordination

Successful group interaction involves the coordination of many processes. Group members must coordinate both physical and mental efforts, and must be motivated to put forth effort within the group. Group members must also coordinate the exchange of information among group members both within ongoing group interactions and in the retrieval and storage of information important to the group. Much of the research that has examined these coordination processes looks at group performance as the primary dependent variable of interest, and thus takes a more functional perspective on group interaction (Hollingshead et al., 2005). That is, these studies focus on describing the relationships of inputs and processes to effective group performance.

Process Losses and Gains

Some of the earliest studies of coordination problems in groups were conducted by Max Ringelmann (1913), an agricultural engineer. Ringelmann was interested in measuring the efficiency with which various agricultural tasks could be accomplished given a variety of configurations of resources (Kravitz & Martin, 1986). For example, is a field plowed more efficiently with a team of three horses or with a team of two horses. In comparing workgroups (both human and animal) of varying sizes, Ringelmann found a consistent tendency for groups to work less efficiently as group size increased, a tendency referred to now as the Ringelmann effect (Steiner, 1974). That is, when predicting the efforts of a group based on the performance of individuals, groups always performed less efficiently than individuals. Ringelmann attributed this loss of efficiency primarily to two processes. First, groups suffer from physical coordination problems as group members each apply their efforts at fractionally different times (Kravitz & Martin). Second, when interacting in a group, members simply are less motivated to work hard, a phenomenon now known as social loafing (Williams, Harkins, & Latane, 1981).

Although physical coordination losses have not received a great deal of empirical attention, social loafing has been widely studied (see reviews by Karau & Williams, 2001, and Williams, Harkins, & Karau, 2003). More specifically defined, social loafing is the reduction of individual effort exerted by people when working in groups compared to when working alone. Social loafing is widespread across many different types of effortful tasks – from shouting and cheering to generating unusual uses for common objects. However, the presence of certain conditions seems to decrease the amount of social loafing: (1) As identifiability of individual efforts increases, loafing decreases. The presence of other people often makes individual contributions difficult to distinguish, therefore increasing the likelihood that loafing will occur (Williams et al., 1981). (2) As the potential for evaluation increases, loafing deceases (Harkins & Symanski, 1988). (3) When the task is described as meaningful, loafing also decreases. That is, people loaf more on tasks that are easy or perceived as unimportant (Harkins & Petty, 1982).

Free-riding and the sucker effect are both potential causes of social loafing. Free-riding also describes a situation where effort is reduced in a collective setting. Free-riding refers to situations where individuals reduce their efforts when there are others who will do the work (Kerr & Bruun, 1983), that is when one feels that one’s efforts are dispensable, or not essential to completion of the group’s goal. The sucker effect is similar, and refers to a reduction in effort that is at least partially motivated by not wanting to be the victim of loafing who ends up doing most of the work (Schnake, 1991).

Although motivation and coordination processes can lead to less effective group performance, groups can also be the setting for motivation and coordination gains (Kerr, 2001; Williams & Karau, 1991). For example, Williams and Karau found evidence of social compensation, or the tendency to increase effort in group settings, under conditions where the task was meaningful, but one’s partner was unskilled at the task. Two factors must be present for social compensation to occur. First, the individual must think that effort from their partner will not be forthcoming. Second, the outcome on the task must be valued by the individual (Karau & Williams, 2001). More generally, Karau and Williams (1993) argue in the collective effort model (CEM) that motivation is determined by the joint factors of expectations about one’s ability to reach a goal, and the value of that goal. Therefore in situations where expectations are high concerning reaching a goal and that goal is highly valued, motivation is also high, and collective performance can actually exceed predictions.

Motivation gains when working in a group have also been identified by Kerr and colleagues (Hertel, Kerr & Messé, 2000; Lount, Kerr, Messe, Seok, & Park, 2008). When working on a conjunctive task, where performance is determined at the level of the weakest group member (Steiner, 1972), performance of that weaker group member is sometimes improved relative to their individual performance. This phenomenon is referred to as the Kohler effect in honor of Otto Kohler who identified it in the mid-1920s (Kohler, 1926). In contrast to free-riders, the Kohler Effect stems at least in part from the weaker members feeling indispensable to the group’s efforts (Hertel et al., 2000).

Transactive Memory

Social loafing, social compensation, and social facilitation often occur in the context of coordination of cognitive efforts, although they have been applied to physical efforts as well. However, another type of coordination of cognitive efforts that occurs in groups involves the management of memory. Transactive memory refers to a shared division of cognitive labor within a group or dyad with respect to the encoding, storage, retrieval, and communication of information (Wegner, 1987). Transactive memory is defined as having two components (Wegner, Giuliano, & Hertel, 1985). (1) An organized store of knowledge contained entirely in the individual memory systems of group members, and (2) a set of knowledge-relevant transactive encoding, storage, and retrieval processes that occur among group members.

The successful operation of these processes is dependent on the formation of transactive memory structures – the organizing schemes that connect knowledge held by each individual to knowledge held by others in the system (Wegner et al., 1985). These transactive memory structures are therefore meta-memory processes that contain knowledge about what information others in the group know. Meta-memory may prompt a group member to remind another group member that he or she possesses knowledge that they may have failed to retrieve on their own. In such a way, a group’s memory may actually exceed the sum of individual memory systems (Wegner, 1987).

Transactive memory systems can have both differentiated and integrated structures (Hollingshead, 2001; Wegner et al., 1985). A differentiated structure represents the specialized knowledge held by each individual in the transactive memory system (Gupta & Hollingshead, 2010). An integrated structure represents the shared knowledge systems; it specifies what individuals know in common and that they all know it. Differentiated structures may be particularly important when a group is attempting to collectively store and retrieve large amounts of information. On the other hand, integrated systems may be particularly important in situations where mistakes are particularly dangerous, and therefore redundancy of information is necessary. Hollingshead examined the role of cognitive interdependence in the development of differentiated or integrated memory structures in collective information processing. She demonstrated that differentiated structure developed when interdependent individuals have an incentive to remember different domains of information, whereas an integrated structure developed when interdependent individuals have an incentive to remember common information.

Transactive memory processes have been examined within two primary research paradigms. In the partner paradigm, individuals work on a task that requires that they learn and retrieve information in different knowledge domains with an assigned partner (e.g., Hollingshead, 1998a, 1998b; Wegner, Erber & Raymond, 1991). In some conditions, the partner is known by the participants (e.g., an intimate partner, a coworker), and in other conditions, the partner is a stranger (Hollingshead, 1998a). Prior to working on the task, the participants are asked about their own and their partner’s expertise in each knowledge domain. Participants who knew their partners had more shared agreement about relative expertise, were more likely to specialize in their areas of relative expertise, and performed better than did pairs of strangers (Wegner et al., 1991).

The second research approach involves training group members either together or individually, with transactive memory processes more likely to develop when members are trained together and therefore are exposed to the areas of expertise of other group members (Liang, Moreland, & Argote, 1995; Moreland, 1999; Moreland, Argote, & Krishnan, 1996, 1998). Across a series of studies, Moreland and colleagues were able to show that the benefits to performance from having been trained together were indeed due to the development of transactive memory systems, and not simply greater cohesion (Moreland et al., 1996), more practice (Moreland et al., 1998), or better communication (Moreland & Myaskovsky, 2000).

Information Exchange

Effective group interaction also involves the coordinated exchange of information among group members. In order to solve problems and make effective decisions, group members must gather information, identify and evaluate alternatives, and ultimately choose an alternative to be implemented. This focus on information exchange has been referred to as group level information processing – the degree to which information, ideas, or cognitive processes are shared among the group members and how this sharing of information affects both individual- and group-level outcomes (Hinsz, Tindale, & Vollrath, 1997; Propp, 1997).

Group Decision Making

Much of the research in the area of group decision making, both past and present, has focused on barriers to the effective pooling of information. For example, the faulty decision making processes involved in groupthink (Janis, 1972, 1982), such as self-censorship and stereotyping, lead to symptoms of groupthink, such as poor information search and a deficient generation of alternative courses of action, ultimately resulting in disastrous decisions. Janis and Mann (1977) listed a number of potential techniques that group members use to limit discussion and avoid making a decision, such as ignoring alternatives and/or satisficing – adopting a minimally acceptable solution rather than searching for an optimal one.

For the past two and a half decades, a dominant paradigm has arisen for examining factors that lead to more or less optimal patterns of information exchange in groups. In the mid-1980s, Stasser and Titus (1985, 1987) developed the information sharing paradigm, whereby information to be used in making a decision is distributed unequally among group members, such that some information is distributed in common to all group members, and some is distributed uniquely to only a few group members. Although optimal decision making usually involves the appropriate pooling of all information, both that known in common and that known uniquely, Stasser and Titus (1985) found that group members tended to focus on information that they knew in common, an outcome that became known as the shared information bias. Note that focusing the discussion on shared information will generally only lead to defective decision making under conditions where information that is uniquely known conflicts with information that is known in common, a situation known as a hidden profile (Stasser & Titus). However, dozens of studies have demonstrated that the shared information bias is particularly difficult to overcome and that hidden profiles are difficult to uncover (Greitemeyer & Schulz-Hardt, 2003; Larson et al., 1998; Wittenbaum & Park, 2001).

A number of factors have been identified that contribute to the shared information bias. First, shared information simply has a greater probability of entering a discussion since more people are aware of that piece of information. Winquist and Larson (1998) have found that this sampling advantage is particularly true early in the group discussion. However, the quality of information changes as well once it becomes shared, through a process known as mutual enhancement (Wittenbaum & Bowman, 2004). Shared information is seen as more credible and valid because it is known by more than one person (Wittenbaum & Bowman, 2004). In addition, providers of shared, credible information are seen as better contributors to the group discussion. And finally, being agreed with (e.g., someone else can agree that the information exists) is positive and validating for the self (Wittenbaum & Bowman, 2004).

Although the shared information bias has proven to be a fairly robust phenomenon, there are some factors that have been found to improve information exchange. For example, making group members responsible for domains of expertise increases the likelihood that they will contribute the information, both shared and unshared (Stewart & Stasser, 1995), to the group discussion. Framing the decision as an intellective task, one for which there is a correct answer, rather than as a judgmental task, where consensus simply consists of resolving preferences, also increases information exchange (Kelly & Spoor, 2004; Stasser & Stewart, 1992). In the original experiment, creating a situation where there is pre-discussion disagreement concerning the best option, rather than creating a hidden profile, also increases information exchange (Stasser & Titus, 1985). Aspects of how the information is presented, such as making unshared information salient, providing a low total amount of information, or providing a low proportion of shared to unshared information, can also increase the exchange of unshared information (Wittenbaum & Stasser, 1996). Finally, eliminating or reducing temporal constraints on discussion time can also increase the discussion of unshared information (Kelly & Karau, 1999; Larson, Christensen, Franz, & Abbott, 1998; Winquist & Larson, 1998).

Group Polarization

Group members not only exchange facts in coming to a decision, but they also exchange persuasive arguments and preference information. A pervasive phenomenon that results from the exchange of initial thoughts and opinions about a decision is that group members shift their own positions in a direction that is consistent with the most popular trend in the group. That is, the group’s opinion shifts toward either a more risky or a more conservative position depending on the initial average inclinations of group members (Moscovici & Zavalloni, 1969; Myers & Lamm, 1976). This phenomenon, known as group polarization, has been demonstrated in dozens of studies across many different types of groups. For example, work by Davis and colleagues (Davis, Bray, & Holt, 1977; Davis, Kameda, & Stasson, 1992) on jury deliberations has shown that if a two-thirds majority of jurors are initially in favor of a verdict (either guilty or not guilty), that verdict, in most cases, will ultimately prevail as a unanimous verdict.

Two primary processes are thought to underlie group polarization. First, social comparison theory suggests that we prefer to hold positions that are slightly more correct than the positions held by others (Festinger, 1954). Once group discussion shows us that, although we hold a popular position, we are not much different than others in our correctness, we shift to a more extreme position (Myers, 1978; Weigold & Schlenker, 1991). That is, polarization occurs because of a group’s ability to exert normative social influence, or influence based on a member’s desire to be well thought of and liked (Kaplan & Miller, 1987).

Second, group polarization is thought to occur because of exposure to more persuasive arguments that are generated for the majority position (Brauer, Judd, & Gliner, 1995; Burnstein & Vinokur, 1977; Vinokur & Burnstein, 1978). Thus, persuasive arguments theory emphasizes the initial distribution of positions. It assumes that arguments for each position are distributed in proportion to the distribution of pre-decision positions of the members. According to this theory, group polarization occurs due to informational influence, or influence that is based on an exchange of facts and information (Kaplan & Miller, 1987).

Both of these processes have been shown to be sufficient to induce group polarization. For example, if group members are provided with arguments concerning a group decision, but no actual positional or preference information, polarization will still occur (Sanders & Baron, 1977). Conversely, if group members are provided with positional or preference information only, but no actual arguments in favor of those positions, polarization again occurs (Goethals & Zanna, 1979). In most situations, however, these processes work in tandem to produce the positional shifts.

Social identity theory researchers suggest that when group members identify strongly with the group, polarization effects are likely to be even greater (Abrams, Wetherell, Cochrane, Hogg, & Turner, 1990). To the extent that initial preferences are seen as an indication of a shared norm within a valued group, group members will be more likely to move in the direction of that norm (Spears, Lea, & Lee, 1990).

Conclusions

Despite the shift in emphasis within sociology from interaction in groups to issues of status and power structure, social exchange, and emotions, a more disciplinarily inclusive look at the field demonstrates that research on groups is plentiful. Recent developments in technology may also open up new avenues for group research. For example, highly portable and reliable recording devices are readily available for use in studying groups outside a laboratory setting. Social media platforms, such as Facebook, have expanded the definition of what constitutes a group by making interaction with large numbers of people easily attainable. Indeed, computer-mediated group interaction has allowed for group members to be distributed over distances rather than being co-located.

One of the implications of these technological advances is a possible change in the type and structure of groups under investigation. These shifts in what constitutes a group may change the relative importance of various group processes across different group types. These changes may also once again change the focus of group research to new topics.

Finally, the multi-disciplinary nature of today’s group research is apparent in this review. In fact, Wittenbaum and Moreland (2008), in their review of trends in small group research, conclude that although group research is abundant, it is not particularly active in the area of social psychology. Whether the technological changes mentioned above will reinvigorate group research in social psychology will need to be examined in future reviews.