Interaction in Small Groups by Peter J. Burke University of California, Riverside 1 | 2 |
From all of these results it is clear that while power and status are different, each can be derived from the other under certain conditions, but emotion plays an important role. And, since emotion is often a function of the legitimation of the powerful position/person/act, the role of legitimation needs yet to be explored in this process. More is said on this later in the chapter in the section on status and emotion. Leadership When we shift focus from the entire status structure of the group to the top person and simultaneously shift from a structure to a process orientation, we move to the study of leadership. Leadership has been a central concern in the study of groups since the very early years, with much of the early focus on the traits of good leaders (e.g., Boring, 1945). However, as Bird (1950) and others have pointed out, almost no identified traits were replicated in more than a few studies. Research then turned to identifying leadership functions by examining what leaders actually do in groups, and how leadership is accomplished (Cartwright & Zander, 1953a). Lippitt and White (1943) examined the question of what leaders do when they studied the different climates that resulted from the different actions in which authoritarian, democratic, and laissez- faire leaders engaged. As pointed out by Burke (1966), however, the impact of these leadership styles depends heavily upon the expectation of the members. Leadership that is too directive or is not directive enough (relative to the expectations of the group members) leads to problems of tension, hostility, and absenteeism. The second approach, asking how leadership is accomplished, was in some sense more fruitful as it allowed any group member to perform leadership functions. The work of Bales and his associates on leadership role differentiation may be seen in that light as it measured task leadership performance of all group members. By examining how leadership is accomplished, it became clear that the style and function of leadership were contingent on the type of group in which they occurred. The most well known theory of leadership was the contingency model of Fiedler (1978a), which sees leadership as a combination of personal and situational factors. Still somewhat of a trait theory, the model suggests that the traits necessary for effective leadership are contingent upon the circumstances of the group. Fiedler suggests there are two types of leaders: task-oriented leaders who more negatively evaluate their "least preferred co-worker" (LPC) and relationship- oriented leaders who more positively evaluate their LPC. This is viewed as a persistent trait of an individual, but its consequences depend upon the context in which leadership is exercised. Each type (high vs. low LPC) is predicted be effective under different conditions of situational control, which are a function of three factors: the leader's relations with the group (good vs. poor), the task structure (highly structured vs. less structured), and the leader's positional power (strong vs. weak). The various combinations of these three factors yield eight conditions with different degrees of situational control. By ordering the factors from most to least important, an ordering of the eight conditions of situational control is created. High LPC (task oriented) leaders are most effective in conditions of either high or low situational control, while low LPC leaders are most effective in situations of medium situational control. While the specific predictions that Fiedler's theory makes about the effectiveness of task- oriented and relationship-oriented leaders have been born out in a number of tests (Strube & Garcia, 1981), two of the eight conditions, as noted by Fiedler (1978b), are less well supported. These conditions are the good leader-member relations, structured task, and weak leader position power, and its complete opposite, the poor leader-member relations, unstructured task and strong leader positional power. At this point, it is not fully clear why these two conditions work out less well, though it may have something to do with the relative importance of the three factors which serve to order the eight conditions (Singh, Bohra, & Dalal, 1979). Fiedler's theory suggested that the most important factor was leader-member relations, and the least important was the leader's positional power (Fiedler, 1978a). Singh and his associates (1979) suggest there is evidence that this suggested order may be incorrect. By changing importance of the factors, the two cases are less anomalous, though it is not clear theoretically why either order is to be preferred. Hollander (1958; Hollander & Julian, 1969) developed a more process oriented view of leadership that aimed to understand how the leader can be both a person who pushes the group in new and innovative directions and be a person who upholds the group norms. His idea is that leadership is a relationship between leader and followers. He developed the idea of "idiosyncrasy credit," viewed as an index of status. Leaders develop credit while interacting with other group members over time by adhering to the group norms and by identifying strongly with the group. This credit can then be used later when the leader engages in idiosyncratic behavior to push the group in new directions. In a sense, idiosyncrasy credit is the legitimacy (endorsement) which the leader can gain or lose by their behavior. This idea of idiosyncrasy credit was more fully developed in later work which focused directly on the issue of the leader's legitimacy, which was derived from his or her prior acceptance in the process of emerging as leader (Hollander, 1993; Julian, Hollander, & Regula, 1969). Three basic sources of legitimacy were seen as important in group member's accepting a leader. The leader's competence and task success were two factors that increased the legitimacy of the leader (thus forecasting Ridgeway and Berger's (1986) theory on the sources of legitimation). However, these two factors interacted in a complex fashion with the third factor, election vs. appointment of the leader, which can be seen to reflect what Zelditch and Walker (1984) called endorsement and authorization (Julian et al., 1969). For elected leaders, there was low satisfaction with an incompetent leader, irrespective of the leader's success or failure. For successful leaders, however, there was satisfaction only for the competent. Among appointed leaders, the pattern shifted. There was satisfaction with successful leaders, whether or not they were competent, while competence moderated the satisfaction with leaders who failed, with the more competent still enjoying some satisfaction among the members. A very different approach to the study of leadership was initiated by Moreno (Moreno & Jennings, 1960) in the context of what he called sociometry, or the measurement of social configurations (see also the chapter on social networks in this volume). Based on the idea that there are positive and negative connections between persons in groups (each based on particular criteria, e.g., live with, work with, play with, etc.), sociometry maps these connections by asking group members to select (or reject) others based on the criteria. Additional information is gathered to help understand the pattern of choices and help draw conclusions from those patterns. Some people are chosen by a lot of others, some are chosen by no others, some are rejected. Those who are relatively over chosen may be considered to be leaders in this approach (Jennings, 1950). It is stressed in this approach that it is not just the pattern of choices that is important, but understanding the basis of that pattern (for example, the characteristics or the chooser and chosen). This approach to the identification of individuals in different positions within a group (for example stars or isolates) found acceptance in therapeutic (e.g., Passariello & Newnes, 1988), organizational (e.g., Patzer, 1976), and educational settings (e.g., Hallinan & Smith, 1985). In more mainstream sociology, this approach moved away from notions of leadership and developed into the study of formal networks (H. C. White, Boorman, & Breiger, 1976) as well as the study of larger social networks and the ways people are tied into them (examples include Burt & Janicik, 1996; Butts, 2001; Granovetter, 1983). Gender and Leadership The relationship between gender and leadership has been extensively explored in hundreds of studies. Using meta-analyses, Eagly and associates (see below) have broken down the gender and leadership issue into four areas: emergence, effectiveness, style, and evaluation. Eagly and Johnson (1990) first looked at the difference of leadership styles by gender. Here, context made a difference. In organizational studies, males and females did not differ in style. In laboratory experiments, however, stereotypical results were obtained; males were more likely to be task oriented and females more likely to be interpersonally oriented in their orientations. One difference consistent with sex-role stereotypes that was found across all settings was that males tended to be more autocratic and females tended to be more democratic. The question of emergence of leadership was examined in initially leaderless groups (Eagly & Karau, 1991). The results of a survey of studies showed that in general men emerged as leaders more often than women, and this was especially true in short-term task oriented groups. On the other hand, women were slightly more likely to emerge as social leaders. These results are consistent with the tendency in our culture for men to be more task oriented and for women to be more relationally oriented and socially facilitative (Eagly, 1987). The third meta-analysis by Eagly and her colleagues concerned reactions to and evaluations of male and female leaders (Eagly, Makhijani, & Klonsky, 1992). In the 147 reports investigated, there was a slight tendency for female leaders to be derogated more than male leaders, but again, context and style made a difference. Female leaders who used a masculine style (autocratic, task-oriented) were more likely to be devalued. Female leaders were also more likely to be devalued if they occupied male-dominated roles, or when the evaluators were males. Interestingly, however, ratings by subordinates reversed these evaluations. Male subordinates rated female leaders more positively and female subordinates rated male leaders more positively. Finally, with respect to the effectiveness of leaders, Eagly and her associates found no overall differences in the effectiveness of male and female leaders (Eagly, Karau, & Makhijani, 1995). However, in particular environments, there were differences with leaders being more effective in gender congruent environments. Additionally, it was found that males were more effective in roles that were numerically dominated by male leaders and subordinates. Not inconsistent with these results, Brown (1979), in a review of 32 female leadership studies, found that in laboratory studies of students compared with managerial studies, female leaders were less effective, suggesting that stereotypes working to the detriment of female leaders may hold more in the laboratory context. INTEGRATION AND COHESION. I now consider the second issue area of group integration and cohesion. Understanding the sources of the degree to which members of a group are attracted to the group, attracted to others in the group, like the other individuals in the group, or want to stay in the group has been a long-standing goal of group researchers. Each of these (attraction, liking, and staying) has been defined as evidence of cohesion by various researchers (Forsyth, 1999). Some researchers have pooled them all together. Schachter (1953), for example, defines cohesion as the "total field of forces acting on members to remain in the group." Hogg and his colleagues, on the other hand, have taken a different tact to distinguish between attraction to others in the group (personal attraction) and attraction to the group (social attraction). They have defined group cohesion uniquely in terms of social cohesion in order to distinguish the group from interpersonal relations (M. Hogg, 1987) Several approaches to understanding the sources of group cohesion have been taken over time, including ingroup-outgroup distinctions, interaction, exchange, and identity processes. Simmel (1955) observed quite early that outgroup conflict serves to create ingroup cohesion, and the early experiments by Sherif on boys groups at camp verified this quite dramatically (cf. Sherif, 1966). In more recent work based on social identity theory, the mere distinction between an ingroup and an outgroup, even in the absence of conflict is sufficient to bring some cohesion (M. Hogg, 1987; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). Interaction that is facilitated by a social and physical environment that is conducive to people frequently meeting and interacting with each other brings about a sense of community, cohesiveness, and sharing (Festinger et al., 1950). This, in turn, brings pressures on individuals to share in the group norms and be considered part of the group (Schachter, 1953). The effects of interaction, however, may be seen to vary with the type of interaction and the emotional reactions of members to the interaction. Negative emotions are divisive and positive emotions are integrative (Kemper, 1991). This has been seen very strongly in marital interactions (Gottman, 1993; Gottman, Coan, Carrere, & Swanson, 1998; Tallman, Rotolo, & Gray, 2001). The notion of positive interaction as a source of group cohesion has been taken up by Lawler and others from an exchange perspective in the theory of relational cohesion (Lawler, 1999; 2000; Lawler & Yoon, 1993; 1996). In this theory, a series of successful exchanges, engaged in over time, leads to positive emotions, which in turn lead to relational cohesion or group commitments. The greater the frequency of exchanges, the greater will be the "emotional buzz" that arises from the exchange process, and the greater will be the degree of cohesion. This theory has been elaborated and extended to build a stronger framework for the role of emotions in not only group cohesion, but also other manifestations of "groupiness" including interpersonal trust, strong norms, and reciprocal typifications (Lawler, 2002). An identity theory approach to this issue was taken by Burke and Stets (1999) who suggest that it is not the exchange process as such that brings about cohesion and commitment, but the process of self-verification in the group context, or what they term mutual self- verification. They suggest that if, in the process of verifying their own group identities, each person in the group helps to verify the identities of other group members, a mutual dependence comes into existence. The process of mutual verification over time builds trust among the group members who come to rely on each other, and the trust, in turn, builds commitment and positive emotional feelings for the other group members. It is recognized that self-verification may involve exchange behavior as in the theory of relational cohesion, but it goes beyond to involve all social behavior. This theory was supported in a study of marital interaction (Burke & Stets, 1999). All of the above theories have involved interaction as an important process that builds cohesion. A more cognitive approach involving dissonance was suggested in an early paper by Aronson and Mills (1959). They tested the common observation that people who go through a great deal of trouble or pain to attain something tend to value it more highly. An experimental situation was set up in which some people had to undergo an embarrassing test (two forms that were more (severe condition) and less (mild condition) embarrassing) to obtain membership in an ongoing discussion group, while others did not undergo any test. Afterward, all respondents at this "first meeting" were asked to simply listen to the discussion of the group since they had not yet had a chance to read the material that was to be discussed. The results showed that those who underwent the severe form of the test rated the participants and the discussion much more highly than those who had the mild test or had no test. These results were explained by dissonance theory (Festinger, 1957), suggesting that those in the severe condition paid a high price to belong and adjusted their attitude and feelings about the group to be consistent with the knowledge that they paid a high price. A cognitive approach is also taken in social identity/self-categorization theory. Knowledge of membership in a group (ingroup) automatically creates an outgroup, feelings of being like others in the ingroup, and behavior that favors the ingroup. This is true even when the ingroup is a minimal group, i.e., one to which the respondent is randomly assigned, in which there is no interaction and no meeting of other persons in the ingroup (or outgroup), when in fact there is no group as such. Being named as part of a "group" is sufficient to bring about deindividuation and feelings of belongingness. INTERACTION The study of interaction in groups is the study of the process of individuals acting and reacting to each other over time. As earlier mentioned, this is what the Bales' IPA coding system is designed to capture. One of the early uses of this coding system was to understand the evolution of relationships in a triad (Mills, 1953). Mills was interested in Simmel's hypothesis that triads tend to break into a pair and an "other." He examined the interaction between the most active two members of triads and classified their relationship as solidary if each supported the other, otherwise as conflicting, dominating, or contending. The solidary relationship could be viewed as a coalition of two against one, and when examined over time was the most stable of the relationships. The dominant and contending relationships were the least stable and tended to become conflicting over time. The conflict relationship was of medium stability but tended over time to change to one of the other forms, with more changing to the solidary (coalition) form than either contending or dominant. Thus, the coalition is stable and other forms tend over time to become coalitions of two against one in the triads. The study of coalitions in the three person group and the conditions under which they would form became an issue that was central for a number of years in social psychology following the methodology initiated by Vinacke and Arkoff (1957) to test some ideas suggested earlier by Caplow (1956). Caplow had analyzed triads and distinguished six basic types, depending upon the relative power of the three members. For example, all members having equal power was type I, or one person having more power than the other two (who are equal, but whose combined power is greater than the first person) was type II, and so on. The relative power of the different members was then used to predict what coalitions would be formed. Vinacke and Arkoff (1957) confirmed most of these predictions and suggested that initial power was the determining factor in the formation of coalitions, with the weaker member more often initiating the formation of coalitions in the manner Simmel predicted with respect to tertius gaudens. However, further tests of this question under more strict conditions failed to confirm this finding (Stryker & Psathas, 1960). Kelley and Arrowood (1960) pointed to another problem with the Vinacke-Arkoff procedure for setting up power differences in the triad. They suggested that several of the triad types were in fact structurally equivalent, even though the assigned power/points were different. By altering the experimental procedure slightly, Kelley and Arrowood showed that, in these structurally equivalent triads, participants learned over time that the point variations in structurally equivalent games were irrelevant and did not need to be considered in dividing up the coalition's profits. The fact that people initially attend to the points indicates the degree to which people look for signs of status and power in our culture. By the early 1970s a number of theories concerning the formation of coalitions had emerged (Caplow, 1968; Chertkoff, 1971; Gamson, 1961; Laing & Morrison, 1973). One issue in most of these initial studies of coalition formation was that the formation (or not) of coalitions was the only outcome. The process of interaction and negotiation to achieve these outcomes was ignored. As this issue was addressed, there was a shift in the studies to the process of bargaining and exchange (Chertkoff, Skov, & Catt, 1980; Friend, Laing, & Morrison, 1974). Exchange Research in Groups Beginning with the work of Emerson (1972a; 1972b) exchange theory began to study the concept of power. Power was defined as the inverse of the degree to which one person depends upon another in a network of interaction. In this way, power was viewed as emerging from the network of relations and the distribution of resources (Willer, 1999). The earlier work on power and coalitions in the triad can be seen in this network approach. Especially important was that network exchange theory took exchange theory from transactions between persons to the study of transactions between persons embedded in networks. The insight of Emerson was that the power of A over B was in part a function of the alternatives that A has to exchange with persons other than B. When A negotiates with B, she has an advantage if she has an alternative source in C. If A has no alternatives A's power is thereby reduced. This means that it is the structure of the network of relations that is an important determinant of power. This idea was not totally new given the work of Bavelas and Leavitt (Bavelas, 1953; Leavitt, 1951), who showed that the structure of contacts and information flow in a group had a strong impact on leadership and power. Although they looked more at information flow than exchanges, they showed that centrally located persons had more power and were more satisfied with their job than more peripherally located persons, and they had higher evaluations of the job the group completed. To understand the nature of the relationship between exchange structures and power (in the context of negotiated exchanges in which persons negotiate the distribution of some good), a number of different theories have developed, each trying to increase its scope and predictive accuracy over others. These include power-dependence theory (Cook, Emerson, Gilmore, & Yamagishi, 1983), elementary-relations theory (cf. Willer & Markovsky, 1993), network exchange theory (Markovsky, Willer, & Patton, 1988), expected-value theory (Friedkin, 1992), core-theory (Bienenstock & Bonacich, 1992), and, as extensions to network exchange theory, resistance theory (Heckathorn, 1983; Willer, 1981), and resistance and degree (Lovaglia, 1995a). Such a proliferation of theories can only take place when there is a great deal of research activity and interest in the issues. Some of these are discussed elsewhere in the chapter on Social Exchange Theory in this volume. Negotiated exchange is not the only kind of exchange, and Molm has engaged in a program of research on non-negotiated exchange that examines not only the distribution of "goods" (positive outcomes such as rewards), but also "bads" (negative outcomes such as punishments) (Molm, 1997). In non-negotiated exchange, people unilaterally give rewards or punishments to others. These may or may not be returned at some point in the future. This is the pattern, for example, in giving birthday gifts. No immediate return is expected, and no negotiation takes place beforehand. By giving out punishments, people exercise what Molm calls coercive power. Coercive power is quite unlike reward power. It is not induced by a coercive power advantage. Rather, it is used purposefully, though sparingly, primarily by people who are disadvantaged in reward power. There is also more individual variability in its use (Molm, 1997). Development of Status Structures. A process orientation was used in more current work on the evolution or development of status structures in groups. One study, following the Bales tradition of studying freely interacting groups, examined the emergence of a status and influence ordering, focusing on total interaction rates (Fisek & Ofshe, 1970). These researchers found that about half the groups differentiated quickly in member participation rates, while the other half went for a long time with nearly equal participation rates among the three members, developing a dominance structure over time. For those that evolved a status structure early, Fiske and Ofshe argued that characteristics of individuals influenced the early development of structure. In a series of more recent studies, Shelly and Troyer further examine this question of the origins of a status ordering (Robert K. Shelly & Troyer, 2001a; 2001b). Drawing from expectation states theory, they hypothesize that persons who are initially advantaged with some status characteristic (task skill, status legitimation, or being liked) are likely to emerge quickly as a high contributor in initially leaderless task-oriented groups, and that advantage on more than one dimension should accelerate this process. Compared with a control group, however, only the person with initially advantaged task skill emerged more quickly. Those who had the advantages other than task skill did not emerge as highest contributor more quickly than the highest participator in the control group (with no initial status advantage). Additionally, being initially advantaged on more than one attribute did not lead to quicker emergence as a high contributor. On the other hand, those advantaged by status or skill tended to take longer turns and were rated as more influential by others in the group (Robert K. Shelly, Troyer, Munroe, & Burger, 1999). However, when the initial advantage combined status and skill, the combination apparently led to greater legitimacy and the top person did not need to dominate the discussion, so that while influence was high, speeches were shorter (Robert K. Shelly et al., 1999). The results of these studies suggest that the process of status emergence is not a simple, straight-forward process. Contrary to Fiske and Ofshe's (1970) expectations that initial status advantage helps the status evolution process, Shelly and Troyer's (2001a) results show that does not always happen. What is happening in the status emergence process in groups of freely interacting individuals remains a topic for future investigation. Another approach to the evolution of leadership and status structures in freely interacting groups was taken by Riley and Burke (1995) who approached the issue from the point of view of identity processes. They argued that individual behavior is a function of the set of self-meanings that individuals have in roles. People try to portray in their behavior the meanings they hold for themselves in the role, and work to counter disturbances to these self-meanings that arise in the situation the process of self-verification. They investigated the set of meanings of a task leader identity and found that when group members are able to portray the level of leadership they hold in their identities, and it is confirmed by the reactions of others in the group, they are most satisfied with their performance in the group. When discrepancies occur, they are less satisfied and work to reduce the discrepancy over time. From this we see that the leadership performances are a function not only of the status characteristics individuals hold, but also of individual's identities and the normal identity process of self-verification. Since there is initial variability across group members with respect to the (amount of) task leader identity, there resulted a distribution in task leadership performance and a corresponding status structure. Status Structures and Emotion While most of the work examining the relationship between status and emotion looks at the effect of status processes on emotion following the Bales tradition (for example that high status persons have more positive emotional responses than low status persons (e.g., Lucas & Lovaglia, 1998)), the reverse connection has begun to be examined. Shelly (1993; 2001; Robert K. Shelly & Webster, 1997) in a series of experiments has shown that sentiment structures in a group (patterns of liking and disliking) have a small impact on performance expectations the status ordering in a group. Lovaglia and Houser (1996) have further shown that emotion can strengthen or weaken existing status structures. Through experimental work they showed negative emotions reduce compliance with a high status person, and when emotional reactions are consistent with the existing status structure, they reinforce it, but when the emotional reactions are inconsistent (higher status person showing positive emotion and lower status persons showing negative emotions), status differences are lowered. Another theory that grows out of expectation states theory goes back to the issue of role- differentiation in that it examines emotional responses to task behavior. Ridgeway and Johnson (1990) develop expectation states theory ideas to incorporate emotional behavior which, as noted, had been excluded from the theory (Berger & Conner, 1974). Ridgeway and Johnson point out that agreements and disagreements, which Bales had considered to be socioemotional behavior, were redefined within the expectation states tradition as instrumental behaviors involving influence. They suggest that there is in fact a dual nature to agreement and disagreement. They have both instrumental and expressive properties, and this plays a key role in group processes. While agreement and disagreement have a task focus as responses to how the task is to be accomplished, reactions to these behaviors are likely to be more emotional. Drawing on the sociology of emotion literature, Ridgeway and Johnson (1990) suggest that the way in which these emotional reactions play out, however, depends upon the expectation states of the group members. When a person who has lower task expectations for the self than the other experiences a disagreement, he or she is likely to take it as a sign of their lesser ability and react depressively and not express overt emotional behavior. However, when the person has higher task expectations for the self than the other, a disagreement is likely to lead to anger and the expression of negative socioemotional behavior. The result of these processes is to reinforce the status structure by feeding more negative reactions to lower status members. Agreements, on the other hand, lead to positive reactions and encourage the person to continue with their contributions to the task, thus reinforcing the emerging structure of task contributions. Because of the correlation between negative reactions and the status structure, positive reactions will tend to go to higher status members because lower status persons will be less inclined to disagree with higher status persons. In this way, socioemotional activity will be also patterned around the status structure and tend to reinforce it. However, when the status structure is ill-determined, status struggles can lead to more negative socioemotional activity in the group that is contained in part by what Ridgeway and Johnson (1990) term legitimacy dynamics. Here, third persons become important. These other parties may act to reinforce and accept the suggestions of one person rather than another, thus lending legitimacy (and status) to the person whose ideas are accepted. In this way, status structures can emerge through consensual evaluation of the quality of the contributions of different group members. All of these dynamics, thus, tend to produce and/or reinforce a single status structure based on and supporting the task contributions of members. For this to work, however, there must either be consensus on the initial status rankings or on the relative contributions of different members, with no member pushing the group too hard on the task so as to loose legitimacy. Consistent with the findings of Burke (1974b), this suggests that the legitimacy conditions for leadership when met produce and maintain a single status ordering. What Ridgeway and Johnson (1990) do not discuss is what happens when the high task status person uses too much negative behavior to put down contributions and/or challenges by the lower status persons or when the legitimacy conditions for leadership are not met. In this case, a second status ordering may well emerge based on socioemotional contributions (Burke, 1974a). Group polarization. Outside the area of status processes, one of the most researched processes in groups has to do with the finding that in making social judgments that involve some level of risk, the decisions of individuals in a group prior to discussion and the decision of the group (or of individuals) after a discussion are quite different (Stoner, 1961). In the initial problems that were given, a shift toward a more risky decision than the average of the individual pre-discussion decisions was noted. Group discussion seemed to intensify people's opinions. Later studies in exploring this phenomenon showed that for some problems there was a shift toward a more conservative decision following discussion. The general phenomenon came to be known as group polarization and was documented in many contexts and cultures (Fraser, Gouge, & Billig, 1971; Gologor, 1977). Researchers were so captivated by this finding, and the ensuing experiments and theories attempting to explain it, that the number of publications, many of which had only minor variations on the theme, skyrocketed to the point that some journals were threatening a moratorium on publishing any more research on the topic. In spite of this plethora of research, there is still no accepted single explanation. Four different explanations exist, and it may be that all (or none) are an accurate account of the phenomenon. These four explanations, in the order in which they were proposed, are: an extension of Festinger's (1954) social comparison theory, a persuasive arguments theory, social decision theory, and intergroup differentiation theory. The social comparison theory suggests that on such issues (for example, in the shift to risk side), people have opinions based on a general culture which supports risk. Each feels that s/he is risky. Only when discussion occurs, however, do people see that they are not as risky as they thought compared to others. This argument has received considerable support (Goethals & Zanna, 1979; Sanders & Baron, 1977). The persuasive arguments theory focuses on the content of the discussion, and suggests that the more arguments that are presented in one direction (risky) or the other (caution), the more people will move their own opinions in that direction. This is coupled with the ideas that there are more arguments in one direction or the other for a given issue depending upon the culture, and that an individual will not be familiar with all the arguments. As the arguments come out in the group discussion, people are moved in the culturally supported direction. This argument also has received considerable support, some in direct contrast to the social comparison theory (Burnstein & Vinokur, 1977; Vinokur & Burnstein, 1978). A third argument, based on social decision theory, suggests that groups have implicit rules about how they will make decisions. Groups that generally favor risk taking on an issue seem to adopt a rule that says if only one person favors the risky decision, ignore it. However, if two or more favor the risky decision, then take it. For groups with a more cautious approach, a similar rule applies in the other direction. This approach also has be confirmed (e.g., Davis, Kameda, & Stasson, 1992). An intergroup relations approach was proposed by Wetherell (1987). She raised two questions that the other theories did not handle: "what makes a persuasive argument persuasive?" and "why are some kinds of extremity desirable?" The approach she suggests draws upon self- categorization theory (Turner, 1985) to argue that group members have an idea of the characteristics of the prototypical group member, and, wanting to be good group members, they emulate the prototypical member. But, prototypicality is, in part, defined by the presence of an outgroup (if only implicitly) (M. Hogg, 1987). Thus, for a group that sees itself as somewhat risky, the prototypical group member would be even more risky to distinguish the in-group from the conservative outgroup. Thus, in being good, prototypical group members, they conform to the more extreme prototypical standard. This theory also has received support (Mackie, 1986). While the principles underlying each of the four current explanations have been confirmed, each also finds some fault with the other explanations. It is possible that each contributes part of the overall explanation (Isenberg, 1986), or that each holds under certain conditions that are not clear (R. Brown, 2000). It is also possible that each would be subsumed in a more general theory should that be developed. All of these theories consider each person's initial position as a point on some continuum. However, we also know that there is likely variance around that point in a sense people's opinions form a probability distribution rather than a point. The shape of this probability distribution may play some role in the dynamics with people finding it easier to change in one direction or another, with some people caring more than others, with some people being influenced more than others, and so on. Clearly, there is more room for work on this issue. Conclusion Contrary to the worry expressed by Steiner (1974), the study of small groups and small group processes is alive and well. It has become pervasive and diverse, however, across many disciplines and research issues so that it is difficult to see the whole. Indeed, in this brief review I have covered only a small part of the research on groups and group processes. 8 What can we make of the current trends? As indicated, much of the research on groups in sociological social psychology is conducted on individual reactions, choices, perceptions, feelings, and so forth in constrained (e.g., limited channels of communication) or artificial (e.g., interacting with a computer simulated other) social situations. And, while this is entirely appropriate for answering certain theoretical issues about particular processes, it does miss phenomena that only occur in the process of interaction. It is not enough to know only what a person sees, feels, or thinks to know how the interaction will pattern itself. As suggested in the emerging field of complexity theory (Gottman, 1991), people adjust and readjust to each other in a dynamic fashion that cannot be replicated in a study of individual reactions and perceptions. The emergence of norms, of roles, of culture, of group development, in short, of what Parsons and Bales called the emerging social system has mostly been ignored. Hopefully, as indicated in some current work, interest in these issues is gaining ascendancy and work will continue to develop. |
|