Bulletin Personality and Social Psychology

Personality and Social Psychology Bulletin http://psp.sagepub.com/

The Link between Leadership and Followership: How Affirming Social Identity Translates Vision into Action S. Alexander Haslam and Michael J. Platow Pers Soc Psychol Bull 2001 27: 1469 DOI: 10.1177/01461672012711008 The online version of this article can be found at: http://psp.sagepub.com/content/27/11/1469

Published by: http://www.sagepublications.com

On behalf of:

Society for Personality and Social Psychology

Additional services and information for Personality and Social Psychology Bulletin can be found at: Email Alerts: http://psp.sagepub.com/cgi/alerts Subscriptions: http://psp.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://psp.sagepub.com/content/27/11/1469.refs.html

Downloaded from psp.sagepub.com at Thuringer Universitats - und on April 6, 2011

PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN Haslam, Platow / SOCIAL IDENTITY AND FOLLOWERSHIP

The Link Between Leadership and Followership: How Affirming Social Identity Translates Vision Into Action S. Alexander Haslam University of Exeter Michael J. Platow La Trobe University, Australia Two experiments test the hypothesis that support for leaders is enhanced when their decisions affirm a distinct social identity that is shared with followers. In Experiment 1, participants showed less support for a leader who favored ingroup members who were relatively sympathetic to an outgroup position than for one who favored ingroup members who opposed an outgroup position. This finding was replicated in Experiment 2, which also showed that this pattern extended to support for the leader’s novel plans. Although participants indicated that they supported a leader who behaved evenhandedly toward all ingroup members as much as one whose behavior was identity-affirming, they were unwilling to back up the evenhanded leader with written comments and arguments. These data suggest that leaders’ capacity to engender active followership is contingent on their ability to promote collective interests associated with a shared ingroup identity.

Exactly how the wishes of leaders get translated into

the efforts of followers can be considered a master problem in social and organizational psychology. How is it that the words and vision of an individual (or select individuals) become the wishes and actions of a multitude? What makes workers willing to “go the extra mile” to enact the commands of their bosses? And why do people sometimes set aside their own personal ambitions to ensure the success of someone else’s? At a practical level, the implications of solving these riddles would appear to be enormous, because they provide the key to a range of behaviors on which organizational success depends. To name just three, these include initiation of structure (a leader’s ability to move toward key organizational outcomes by clarifying subordinates’ goals, roles, and tasks; Fleishman & Peters, 1962),

change responsiveness (employees’ willingness to embrace organizational change; King & Anderson, 1995), and organizational citizenship (employees’ willingness to do more than is formally asked of them; Organ, 1988; Ouwerkerk, Ellemers, & de Gilder, 1999). These questions are important at a theoretical level, too, because they represent the nexus between issues of leadership, power, and motivation, which—despite being central organizational and social psychological topics—are rarely broached in the same theoretical or empirical treatment (Pfeffer, 1997). Popular answers to the above questions often appeal to some special quality of a leader, which allows a group to exceed expectations. These are most apparent in personality approaches, which point to the ability of a leader’s inherent charisma to energize and enthuse followers (e.g., Burns, 1978). According to this view, the inspirational capacity of people such as Nelson Mandela, Norman Schwarzkopf, or Bill Gates can be traced back to their personality-based referent power. Allied with notions of transformational leadership (e.g., House, Spangler, & Woycke, 1991), these individuals are seen to achieve their impact through an ability to redefine followers’ goals, values, and aspirations (De Cremer, in press). Although this analysis captures important features of the leadership process, a core problem with such approaches is their lack of predictive power. Great leaders do indeed appear to transform the psychology of Authors’ Note: Address correspondence to S. Alexander Haslam, School of Psychology, University of Exeter, Exeter EX4 4QG, United Kingdom; e-mail: [email protected]. PSPB, Vol. 27 No. 11, November 2001 1469-1479 © 2001 by the Society for Personality and Social Psychology, Inc.

1469 Downloaded from psp.sagepub.com at Thuringer Universitats - und on April 6, 2011

1470

PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

their followers and to be perceived by them as highly charismatic, but these phenomena appear to be correlates rather than causes of the leadership process (Hogg, 1996). They are also highly contingent on social perspective and social context, so that charisma is never conferred by all followers all of the time. Accordingly, Nadler and Tushman (1990) complain that “in real time, it is unclear who will be known as visionaries and who will be known as failures” (p. 80). The fact that a person’s charisma also can increase after his or her death is also highly problematic for arguments that its source lies in the individual alone (Haslam, 2001; Turner & Haslam, 2001). Mindful of such problems, the most common way of dealing with the contextual basis of emergent charisma is to adopt a contingency approach that sees leadership as the outcome of a “perfect match” between the leader’s character and the situation that he or she confronts. The prototypical model of this form was developed by Fiedler (e.g., 1978) and proposes that effective leadership results from a happy marriage between the leader’s style (relationship oriented or task oriented) and the constellation of (a) good or bad leader-member relations, (b) high or low task structure, and (c) the leader’s strong or weak position power. Significantly, too, such models are usually favored by influential leaders themselves, so that effective management is seen as a question of having “the right person for the job at hand.” Yet, whereas contingency models address matters of context, they still neglect those of perspective. If one were to argue, for example, that Martin Luther King, Jr., was the right person to lead the civil rights movement in the 1950s and 1960s, why did only a subset of the population respond to the dream of racial tolerance that he articulated? And why do some people continue to deny his claim to greatness? To deal with these questions, a number of theorists have advocated a transactional approach to leadership that attempts to explain the transformational aspects of leadership in terms of social exchange principles. In this vein, work by Hollander (1985, 1995) focused on processes of interdependence and argued that leadership emerges from the system of interpersonal relations between leaders and followers and reflects the capacity for that system to generate rewards that are gratifying for each party. Hollander’s analysis suggested that followers follow leaders on whom they are dependent for reward (e.g., financial, emotional, intellectual) and that leaders lead followers who provide them with similar rewards in return. In Hollander’s (1995) words, “Whether leadership is called transactional or transformational, the common persisting element is the significant relational nature of the intangible rewards provided to followers by

leaders. This gets to the heart of motivations to follow” (p. 82). Hollander’s research dealt with much of the subjectivity that surrounds attributions of leadership and also highlighted the role that followers play in conferring leadership on any given individual. Because it is followers who end up doing the work on which any leader’s success depends, his work also has played an important role in recognizing followership as an essential component of the leadership equation. Nevertheless, as the above quotation suggests, the notions of cost and benefit on which transactional principles depend are themselves somewhat intangible. Perceptions of mutual benefit do indeed appear to be central to leadership, but similar to judgments of charisma, they appear to be outcomes of the leadership process that are much easier to specify after the event than before (Tyler, 1999). In short, leadership—and the social phenomena in which it is implicated—appears to be more than just a process of interpersonal exchange: a matter of “you scratch my back, I’ll scratch yours.” One attempt to explain what that “more” is is provided within the framework of social identity and selfcategorization theories (Tajfel & Turner, 1979; Turner, 1985; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). These theories are both founded on an assumption that people’s group membership (and the psychological consequences of that membership) has a major role to play in shaping their reactions and contributions to social life—a role that cannot be reduced to their psychological status as individuals. This is because group behavior is underpinned not by people’s personality or their sense of uniqueness (personal identity) but by a sense of shared group membership (social identity). As defined by Tajfel (1972), social identity refers to that part of individuals’ self-concept associated with their membership in social groups. It is the social self (as defined by “we”) rather than the personal self (“I”). When a given social identity becomes salient, individuals perceive their motivations and perspectives to be psychologically interchangeable with those of others who share the same social identity (Turner, 1982). This is argued to lead to behavior that is qualitatively distinct from that which is predicated on personal identity—because it is shaped by, and oriented toward, the interests of the group as a whole rather than those of the individual in isolation. Self-categorization theory makes detailed predictions both about when a particular social identity will become salient and about the psychological and behavioral impact of social identity salience (e.g., Oakes, 1987; Oakes, Haslam, & Turner, 1994; Turner, Oakes, Haslam, & McGarty, 1994). For present purposes, though, it is relevant to note that one factor that affects individuals’ will-


Haslam, Platow / SOCIAL IDENTITY AND FOLLOWERSHIP ingness to define themselves (and others) in terms of a given category membership is fit—the degree to which a particular categorization helps perceivers to make sense of the social world. Comparative fit is enhanced when the perceived differences within categories are smaller than the differences between them (e.g., where “we are different from them”; Haslam & Turner, 1992), and normative fit is enhanced when a particular category accords with perceivers’ expectations and theories about the social world (e.g., where “we are better than them”; Reynolds, Turner, & Haslam, 2000). When it comes to specifying the impact of social identity salience, self-categorization theory makes four predictions that are pertinent to the analysis of organizational behavior in general (e.g., see Haslam, 2001; Haslam, Powell, & Turner, 2000; Oakes et al., 1994; Turner, 1991; Turner & Haslam, 2001; Turner et al., 1987). First, this is expected to lead individuals to see themselves as relatively interchangeable representatives of a particular social category, sharing self-defining norms, values, and goals with other members of that category (ingroup members). Second, it provides group members with a common perspective on reality that leads them (a) to expect to agree with each other on issues related to their group membership and (b) to actively seek agreement through processes of mutual influence. Third, it provides group members with motivation (and expectations of an ability) to coordinate their behavior with reference to emergent group norms (e.g., those that define the ingroup as different to, and better than, other groups). Finally, it leads group members to work collaboratively to advance the interests of the group as a whole (their collective self-interest), even to the detriment of themselves as individuals (their personal self-interest). In this way, social identity salience can be seen to provide the psychological foundations for a range of key organizational phenomena, including information exchange, consensus-seeking, cooperation, trust, empowerment, group productivity, and collective action (see Haslam, 2001, for a detailed elaboration of these ideas). Significantly, too, these arguments also pave the way for a novel analysis of the process through which leaders and followers prove capable of mutual support and enhancement. In particular, the above analysis suggests that for true leadership to emerge—that is, for leaders to motivate followers to contribute to the achievement of group goals (Smith, 1995)—leaders and followers must define themselves in terms of a shared social identity. More specifically, we can assert that leadership centers around the process of creating, coordinating, and controlling a social self-categorical relationship that defines what leader and follower have in common and

1471

that makes them “special.” In Reicher and Hopkins’s (2001) terminology, leaders must therefore be “entrepreneurs of identity.” The success of their leadership hinges on an ability to turn “me” and “you” into “us” and to define a social project that gives that sense of “us-ness” meaning and purpose. In particular, it follows from principles of fit that the success of such an endeavor is likely to depend, among other things, on the capacity of the leader to act in a way that affirms and advances the ingroup’s position relative to salient outgroups. In this regard, a major implication of the foregoing analysis is that leaders’ capacity to marshal support for their plans and to ensure that followers act on them will be enhanced to the extent that they are able to promote the collective interests and aspirations of the group (Haslam et al., 1998). And because much of the demand for leadership emerges in intergroup contexts—where an ingroup is in conflict or competition with an outgroup—it also follows that this will often involve promotion of the ingroup at the expense of an outgroup (Sherif, 1962). Among other things, this means that group members will be most inspired by a leader (a) who demonstrates a willingness to support those ingroup members who epitomize what makes “us” different to “them” and (b) whose actions imply that “we are better than them.” It is this prediction that the present research is designed to test. The studies here follow on from previous research by Platow, Hoar, Reid, Harley, and Morrison (1997; see also Platow, Mills, & Morrison, 2000; Platow, Reid, & Andrew, 1998; Platow & van Knippenberg, in press), in which individuals were exposed to leaders who had allocated resources to ingroup and outgroup members in either an ingroupfavoring or evenhanded fashion. Consistent with the above arguments, Platow et al. (1997) found that whereas evenhandedness was seen as more fair than ingroup favoritism, this pattern was not mirrored in leadership endorsement. Indeed, in some circumstances, participants preferred ingroup-favoring leaders over fair ones. In the present studies, we push these ideas further to make the more subtle point that support for a leader will be contingent on his or her treatment of different ingroup members and their relations with outgroups. In both of the experiments in this article, university students served as the relevant ingroup. Within this ingroup, the leader treated members who had different opinions either in an evenhanded manner (i.e., treated ingroup members fairly; cf. Wit & Wilke, 1988) or differentially favored some members with specific opinions (i.e., treated ingroup members unfairly). In the unfair conditions, leaders either favored ingroup members whose opinions were different from an outgroup (behavior that affirmed the distinct identity of the


1472


ingroup) or those whose opinions were similar to the outgroup (behavior that negated ingroup identity). Our primary hypothesis was that support for the leader would vary across these three conditions, with active followership being most pronounced where leader behavior was identity-affirming. EXPERIMENT 1

Students were presented with information about a leader of a university student council, Chris, who had to nominate members of his council board for a financial prize to be awarded for service to the university as a whole. As in Tajfel, Flament, Billig, and Bundy’s (1971) original minimal group studies, this scenario was developed with a view to ruling out personal self-interest as a possible motive for subsequent responding: Participants did not stand to benefit from any decision Chris had made. The decision occurred in the past and in an unidentified organization. Participants were told that the council had recently been engaged in work relating to key government policies. In three conditions, participants were informed that Chris had nominated for the prize either (a) four progovernment and two antigovernment members, (b) three progovernment and three antigovernment members, or (c) two progovernment and four antigovernment members. Their task was then to indicate how fair and how sensible they perceived Chris’s decision to be and how much they supported it. To control for effects that might simply reflect support for (or opposition to) the government, participants were told that the policies on which the council had been working had either been ones where their student ingroup was aligned with the position of the government (a decision to tighten gun control) or were ones where the ingroup was opposed to the government (a decision to cut university funding). This meant that the selection of four progovernment candidates was an identityaffirming decision where the ingroup-defining issue was gun control but an identity-negating decision where the defining issue was university cuts. On the other hand, the selection of four antigovernment candidates was an identity-affirming decision where the defining issue was university cuts but an identity-negating decision where the defining issue was gun control. The experiment thus had three relevant conditions: one in which Chris’s decision was identity-affirming because he favored group members who supported a normative ingroup position over those who supported an outgroup position, one in which his decision was evenhanded, and one in which his decision was identitynegating because he favored group members who supported an outgroup position over those who supported a normative ingroup position.

Following previous work by Tyler (1994) and Platow et al. (1997, 1998), we anticipated that the leader would be perceived as most fair in the condition where he behaved evenhandedly. However, in addition, we also expected that an identity-affirming strategy (i.e., one in which the leader favored members who opposed an outgroup position) would be seen as more fair than one that was identity-negating (where the leader favored members who supported an outgroup position). Significantly, though, and in contrast to predictions that support for the leader would simply reflect his perceived fairness and sense of justice (e.g., Tyler, 1994), we anticipated that participants would see the identity-affirming strategy to be just as sensible as the strategy of fairness and would also give it equally strong (if not stronger) support. We thus expected that the key predictor of support for the leader would not be his fairness but rather the extent to which he behaved in a way that promoted the distinct identity of the ingroup. Method PARTICIPANTS AND DESIGN

Participants (N = 90) were randomly assigned to one of six conditions arranged in a 2 (focal issue: university cuts/gun control) ´ 3 (leader nomination strategy: identityaffirming/evenhanded/identity-negating) factorial design. There were 15 participants in each condition. MATERIALS AND PROCEDURE

Participants were handed a one-page questionnaire that introduced the research as a study of “attitudes to student leaders.” They were told that the researchers wanted to focus on attitudes toward a particular student leader named Chris who was head of the Students’ Representative Council at an Australian university and who, in the course of his presidency, had had to nominate six members of the council board for an award reflecting service to the university—an award with a prize of $600. In one of six possible permutations, the questionnaire then informed participants that the council had recently been doing most of its work on issues relating to the government’s decision to [cut University funding/tighten gun control] and that Chris had chosen to nominate [two people who had adopted a progovernment stance and four who had taken an antigovernment stance/three people who had adopted a progovernment stance and three who had taken an antigovernment stance/four people who had adopted a progovernment stance and two who had taken an antigovernment stance]. Having read this information, participants were asked to indicate (a) how fair they perceived Chris’s nomination decision to be (on a 7-point scale with endpoints labeled not at all fair [1] and very fair [7]), (b) how sensible they perceived Chris’s nomination decision to be (on


Haslam, Platow / SOCIAL IDENTITY AND FOLLOWERSHIP a 7-point scale with endpoints labeled not at all sensible [1] and very sensible [7]), and (c) how strongly they supported Chris’s nomination decision (on a 7-point scale with endpoints labeled do not support at all [1] and strongly support [7]). When all participants had completed the questionnaire, their responses were collected and the class was debriefed. Results ANALYSIS OF VARIANCE

Scores on the three dependent measures were analyzed by means of 2 (topic) ´ 3 (leader nomination strategy) analyses of variance. On all three measures, these analyses revealed main effects for topic. Specifically, there was evidence that participants perceived the leader’s nomination strategy to be fairer and more sensible, and also supported it more, when it related to the guns issue rather than to university cuts (fairness Ms = 4.62 and 3.56, respectively), F(1, 84) = 8.71, p < .01 (sensibleness Ms = 4.60 and 3.89, respectively), F(1, 84) = 4.59, p < .05 (support Ms = 4.11 and 3.31, respectively), F(1, 84) = 3.78, p = .05. More important, though, responses on all three measures revealed predicted main effects for leader nomination strategy. Means and univariate F values are presented in Table 1. To decompose these effects, responses on this measure were submitted to one-way analysis of variance with post hoc comparison of means performed by Tukey’s method (with a set at .05). The results of this analysis are also presented in Table 1. Here it can be seen that the leader’s strategy was seen to be most fair when it was evenhanded (M = 4.87) but to be significantly less fair when it was identity-negating (M = 3.33), with the perceived fairness of an identity-affirming strategy falling between these values (M = 4.07). A strategy of evenhandedness (M = 4.83) also was seen as more sensible than one that was identity-negating (M = 3.43), but so too was an identity-affirming strategy (M = 4.47), with the perceived sensibleness of evenhandedness and identityaffirming strategies not being significantly different. However, when it came to follower support for the leader’s strategy, this was strongest when Chris’s strategy had been identity-affirming (M = 4.23) and significantly less strong when his actions had been identity-negating (M = 3.00), with support for an evenhanded strategy falling between these values (M = 3.90). REGRESSION ANALYSIS

To explore our hypotheses further, a series of regression analyses were performed to examine the predictive role of key independent variables in the effects obtained on the three dependent measures. To represent the impact of the leader’s fairness, the leader reward variable was recoded with a value of 1 (representing high

TABLE 1:

1473

Experiment 1: Followers’ Reactions to Leader’s Nomination Strategy Leader’s Nomination Strategy

Measure Fairness of leader’s nomination strategy Sensibleness of nomination strategy Support for nomination strategy

Identity- Even IdentityAffirming handed Negating F(2, 84)

4.07ab

4.87a

3.33b

5.52**

4.47a

4.83a

3.43b

6.16**

4.23a

3.90ab

3.00b

3.14*

NOTE: Means in the same row with no common subscript are significantly different (p < .05). *p < .05. **p < .01.

fairness) being given to the evenhanded condition and values of 0 (representing low fairness) being given to both ingroup-favoring and outgroup-favoring conditions. To represent the impact of the leader’s identity affirmation, the leader reward variable was recoded with a value of 1 (representing high affirmation) being given to the identity-affirming condition, a value of 0 (representing intermediate affirmation) being given to the evenhandedness condition, and a value of –1 (representing low affirmation) being given to the identity-negating condition. These loadings were based on the difference in the number of pro- and antigovernment counselors nominated by the leader. Each of the three dependent variables were then simultaneously regressed onto fairness, identity affirmation, and topic. The first of these analyses indicated that the only significant predictor of the perceived fairness of the leader’s strategy was his actual fairness ( = .29), t(88) = 2.91, p < .01. However, the perceived sensibleness of the leader’s strategy was predicted independently by both identity affirmation ( = .25), t(88) = 2.50, p < .02, and fairness ( = .25), t(88) = 2.47, p < .02. Finally, though, when it came to predicting support for the leader’s strategy, the only predictor was identity affirmation ( = .25), t(88) = 2.42, p < .02. Discussion The above results provide straightforward support for our experimental hypotheses. Specifically, it is clear that although the leader was perceived as most fair when he behaved toward ingroup members in an evenhanded manner, this demonstration of fairness was not the key determinant of his capacity to marshal follower support. Instead, support for the leader was strongest when he chose to reward ingroup members who opposed a salient outgroup, and thus, it was his capacity to affirm a distinct social identity that emerged as the key predictor of followership.


1474


Beyond these basic patterns, it is also interesting to note that there was a trend for a strategy of identity affirmation to be seen as more fair than one of identitynegation and also that the perceived sensibleness of the leader’s behavior reflected his capacity both to affirm identity and to behave equitably. Again though, although this concern for equity was seen as sensible, its impact was less evident when followers had to express their support for the leader. This pattern is thus consistent with our general hypothesis that whereas a leader’s concern for fairness may be seen as desirable in the abstract (e.g., see Tyler, 1994), it may prove to be a dispensable luxury when it comes to the cut-and-thrust of actual political behavior (Pfeffer, 1992). Having said that, it is also true that this study is only a first step in examining the determinants of followership. Certainly, in most situations (e.g., in organizational contexts) leadership is contingent on a leader being able to do more than simply secure the support of her or his followers. Instead, it will depend on the leader’s capacity to motivate followers to contribute to the activities required to translate individual vision into collective action and social reality (Haslam & Platow, in press). In particular, this will often involve (a) support on one issue being carried over into support on another and (b) that support being translated into more tangible follower behavior. It was with the goal of examining these processes that a second study was conducted. EXPERIMENT 2

The design of this second study was modeled closely on that of Experiment 1. However, it also included measures designed to examine the impact of a leader’s behavior toward ingroup members on (a) followers’ support for novel plans and (b) followers’ efforts that would help to realize those plans. As in the previous experiment, participants were told about the decision of a student leader (Chris) concerning nominations of student counselors for a university prize. However, as well as this, they also were shown a videotaped interview with Chris in which he answered questions about his presidency. The interview focused on his feelings about his decision to nominate student counselors for a community-service award. The details of this decision were not discussed but the interviewee indicated in a fairly relaxed manner that he had been hurt by some students’ reaction but also encouraged by others’ and that, if given his time again, he would not do anything any differently. At the end of the interview he also briefly touched on his plans for the future and indicated (without giving clear reasons) that he thought it would be a good idea to lobby the university to provide permanent billboard sites around campus.

Having seen the video, students were asked the same questions as in Experiment 1 (i.e., relating to Chris’s decision to nominate particular counselors for the prize) but they also were asked how fair they thought his proposal to have permanent billboards around campus to be, how sensible it was, and how much they supported it. Moreover, they also were given the opportunity to write down arguments relating to the proposal’s worth and utility. The key issue here, then, was the extent to which participants took advantage of this opportunity to engage in intellectual work (e.g., rationalization and justification) that would help to promote Chris’s vision. Would they be willing to add some argumentative flesh to the skeleton plans foreshadowed in the interview? Or would their support prove to be hollow and insubstantial? We expected that whereas results on preliminary measures would replicate those obtained in Experiment 1, these effects also would be carried over into these additional measures. In particular, we expected that support for the billboard policy would be strongest when the leader had a history of affirming the distinct identity of the ingroup and that this same history would play a key role in participants’ willingness to exert intellectual effort to back up his plans for the future. Method PARTICIPANTS AND DESIGN

Participants included 96 students who were enrolled in a 1st-year psychology course; they participated in this experiment in six sessions containing between 12 and 21 people. Within each session, participants were randomly assigned to one of six conditions arranged in a 2 (focal issue: university cuts/gun control) ´ 3 (leader nomination strategy: identity-affirming/evenhanded/identitynegating) factorial design. There were 16 participants in each condition. The study was conducted 9 months after Experiment 1 and involved a different cohort of students. MATERIALS AND PROCEDURE

At the start of the experiment, participants were handed a one-page questionnaire that introduced the research as a study of “attitudes to student leaders.” As in Experiment 1, they were told that the researchers were interested in attitudes toward a particular student leader named Chris. Participants were then shown a short video in which an off-camera interviewer (the first author) had an informal discussion with a 24-year-old male who was supposedly the head of the Students’ Representative Council at an Australian university. The interview focused on how Chris had felt about the reaction that followed his decision to nominate particular counselors for a prize reflecting service to the university community. No mention was made of the actual decision itself and


Haslam, Platow / SOCIAL IDENTITY AND FOLLOWERSHIP his comments were reasonably level-headed and nondescript. Toward the end of the interview, Chris also was asked about his plans for the future. Here he talked about his desire to finish his degree and then think about the possibility of obtaining further qualifications, but he also suggested that in his time remaining as president he would like to lobby the university to provide permanent billboards around campus, because he perceived this to be “a useful and worthwhile cause.” After they had seen the video, participants were given a one-sided questionnaire that informed them, in identical fashion to Experiment 1, that the council had recently been doing most of its work on issues relating to the government’s decision to [cut University funding/tighten gun control] and that Chris had chosen to nominate [two people who had adopted a progovernment stance and four who had taken an antigovernment stance/three people who had adopted a progovernment stance and three who had taken an antigovernment stance/four people who had adopted a progovernment stance and two who had taken an antigovernment stance]. Participants then responded on 7-point scales to the same three questions as in Experiment 1. Beneath these questions, participants were then provided with a short statement reminding them of Chris’s proposal to lobby the university to provide permanent billboards. They also were told that these would cost about $3,000 to erect. There then followed three questions, asking participants to indicate on 7-point scales how fair they thought the decision to lobby for permanent billboards (where 1 = not at all fair and 7 = very fair), how sensible it was (where 1 = not at all sensible and 7 = very sensible), and how much they supported it (where 1 = do not support at all and 7 = strongly support). As well as this, participants were asked to list on the back of the questionnaire “any arguments that you think are relevant to the decision to erect permanent billboard sites (e.g., ones for or against the policy).” After all participants had been given an opportunity to respond to this item, questionnaires were collected and participants were debriefed. Results ANALYSIS OF VARIANCE

Responses on the three rating-scale items for the two issues (reactions to the counselor nomination, reactions to the billboard plan) were analyzed by means of a 2 (topic) ´ 3 (leader nomination strategy) analysis of variance. The only effects to emerge from this analysis were main effects for the leader’s nomination strategy. On measures reflecting reactions to the leader’s nomination of counselors, there was significant variation in perceived fairness and sensibleness as well as follower

TABLE 2:

1475

Experiment 2: Followers’ Reactions to Leader’s Nomination Strategy and Novel Plans Leader’s Nomination Strategy

Measure Fairness of leader’s nomination strategy Sensibleness of nomination strategy Support for nomination strategy Fairness of leader’s billboard plan Sensibleness of billboard plan Support for billboard plan No. of arguments opposing billboard plan No. of arguments supporting billboard plan

Identity- Even IdentityAffirming handed Negating F(2, 90)

3.88ab

4.69a

3.47b

5.96**

4.19a

4.56a

3.13b

9.50**

4.13a

4.03a

3.00b

5.36**

4.28 4.09 3.72a

4.34 3.91 3.66ab

3.84 3.44 2.72b

1.00 1.29 3.70*

0.38

0.59

0.59

0.80

1.03a

0.44b

0.25b

8.14**

NOTE: Means in the same row with no common subscript are significantly different (p < .05). *p < .05. **p < .01.

support for the strategy. On measures relating to Chris’s billboard plan, there was only significant variation in follower support. Means and univariate F values are presented in Table 2. As in the previous study, these effects were decomposed further by submitting responses on these measures to one-way analyses of variance with post hoc comparison of means performed by Tukey’s method (with a set at .05). The results of this analysis also are presented in Table 2. As in Experiment 1, it can be seen that the leader’s strategy was seen to be most fair when it was evenhanded (M = 4.69) but to be significantly less fair when it was identity-negating (M = 3.47), with the perceived fairness of an identity-affirming strategy again falling between these values (M = 3.88). Evenhandedness (M = 4.56) also was seen as more sensible than an identitynegating strategy (M = 3.13), but so too was an identityaffirming strategy (M = 4.19), with the perceived sensibleness of evenhandedness and identity-affirming strategies not being significantly different. Support for the leader’s nomination strategy was stronger when his behavior had been identity-affirming (M = 4.13) and evenhanded (M = 4.03) than when it had been identity-negating (M = 3.00). Mirroring this pattern, support for the leader’s novel billboard plan also was stronger when his previous nomination strategy had been identity-affirming (M = 3.72) or evenhanded (M = 3.66) than when it had been identity-negating (M = 2.72). To analyze the pattern of arguments generated by participants in response to the leader’s billboard plan, these were first coded by two independent raters. Each rater


1476


was given the task of identifying the number of distinct arguments and defining them as supporting or challenging the proposal. Perfect agreement was made in classifying all but five of the participants’ responses, and in each of these cases the minor difference of opinion was resolved through brief discussion. The resulting scores were then subjected to a 2 (topic) ´ 3 (leader nomination strategy) ´ 2 (argument content: pro/anti) analysis of variance with repeated measures on the last factor. Two effects emerged from this analysis. The first was a two-way interaction between topic and nomination strategy, F(2, 90) = 4.60, p < .02. This resulted from the fact that participants were particularly likely to generate arguments in the condition where the leader had behaved in an identity-affirming manner and the issue concerned university cuts (the mean number of arguments generated in this condition was 1.06; the highest mean in any other condition was 0.56). This analysis also revealed the predicted and more substantial two-way interaction between leader nomination strategy and argument content, F(2, 90) = 11.45, p < .001. As the means presented in Table 2 indicate, this arose from the fact that participants were particularly likely to generate arguments supporting the leader’s billboard scheme when his nomination strategy had served to affirm ingroup identity. This point was confirmed by separate one-way analyses of variance performed on the number of arguments supporting and opposing the leader’s plan in each of the three nomination strategy conditions. As can be seen from Table 2, the only effect to emerge here resulted from variation in the number of supporting arguments generated by followers. Post hoc tests (Tukey’s with a = .05) indicated that more of these were produced when the leader had rewarded ingroup members in a manner that was identity-affirming (M = 1.03) than when his strategy had been evenhanded (M = 0.44) or identitynegating (M = 0.25). REGRESSION ANALYSIS

Following the strategy adopted in Experiment 1, a series of regression analyses was performed to examine the predictive role of key independent variables in the above effects. Here, the different conditions were recoded to reflect strategies of fairness and identity affirmation in the manner described previously. The same three measures as in Experiment 1 (pertaining to the fairness, sensibleness, and support for the leader’s nomination strategy), as well as support for his billboard plan and the number of arguments supporting it, were then simultaneously regressed onto fairness and identity affirmation (as well as the focal topic). The first three of these analyses revealed an identical pattern to that obtained in Experiment 1. Specifically,

the only significant predictor of the perceived fairness of the leader’s strategy was his actual fairness ( = .32), t(94) = 3.25, p < .01. The perceived sensibleness of the leader’s strategy was predicted independently by both identity affirmation ( = .29), t(94) = 3.09, p < .01, and fairness ( = .29), t(94) = 3.05, p < .01. Again, though, the only predictor of support for the leader’s strategy was identity affirmation ( = .29), t(94) = 2.96, p < .01. Moreover, the pattern obtained on this last measure also was reproduced in support for the billboard plan and the number of arguments articulated in its defense. Accordingly, when it came to this novel issue, identity affirmation was the only variable that predicted willingness to support the leader ( = .24), t(94) = 2.43, p < .02, or to generate arguments in support of his ideas ( = .37), t(94) = 4.00, p < .001. Discussion The results of this study replicate those of Experiment 1 but also extend them in important ways. In particular, they demonstrate that followers’ support for leadership strategies that affirm ingroup identity is extended both (a) to support for novel plans and (b) to willingness to exert effort that helps realize those plans. Moreover, one particularly interesting feature of this study’s findings is that they reveal differences in responses to evenhandedness across these two additional measures. As in the previous study, evenhandedness was recognized as more fair and more sensible than a strategy of identity-negation. Similar responses to an evenhanded leader and one who affirmed identity were observed on these measures and similar levels of support emerged in relation to both the initial nomination strategy and the envisioned billboard scheme. Significantly though, it is clear that when participants were given the opportunity to do some intellectual work that would help promote the leader’s new plan, the support that participants in the evenhandedness condition had displayed in their rating-scale response evaporated. On this measure of follower effort, the only factor that seemed to encourage the participants to do some work on the leader’s behalf was knowledge that he had previously supported ingroup members who were opposed to the outgroup. If the leader had simply been fair, his would-be followers generated very few suggestions and exerted no more effort than participants who believed that his nomination strategy had been identity-negating. In short, only when the leader had a history of standing up for the group and affirming its distinct identity was the group prepared to stand up for him and do the work (in this case, the intellectual justification and rationalization) necessary for his vision to be realized. In relation to the three initial measures that were reproduced from Experiment 1, the only substantial dif-


Haslam, Platow / SOCIAL IDENTITY AND FOLLOWERSHIP ference in the pattern of findings obtained in this second study pertained to effects for focal topic. In Experiment 1, respondents had generally been more enthusiastic about leaders whose behavior related to gun control rather than to university cuts, but in this study, there was no such difference. The precise reasons for this are unclear. However, one possible explanation may lie in the fact that the first study was conducted shortly after the issue of gun control became a major topic of debate in Australia (following the massacre of tourists in Port Arthur). Accordingly, the fact that this was a focal issue for student leaders may have had a positive impact on students as a whole. Yet by the time the second study was conducted, this issue was less topical; therefore, attention to it may not have elicited such a strong positive reaction. If this explanation is correct, it underlines the importance of social context in determining reaction to leader behavior (a point widely accepted in the leadership literature; e.g., Cooper & McGaugh, 1963; Fiedler, 1978). Beyond this, though, the pattern appears to have no direct bearing on the core theoretical issues that our research was designed to address. GENERAL DISCUSSION

Considered together, the findings of these experiments provide strong support for theoretical arguments derived from social identity and self-categorization theories. Specifically, they suggest that a leader’s capacity to display true leadership by motivating followers to participate in activities that promote a vision that he or she has identified as important is itself contingent on the leader’s ability to behave in a way that exemplifies the values and ideals that are shared by the group they lead and that make it distinct from comparison outgroups (Tajfel & Turner, 1979). In both experiments, the predictions derived from this approach were tested against the view that participants would be willing to follow a leader who simply embodied values of equity and decency by behaving evenhandedly toward various ingroup members. Consistent with previous research by Platow et al. (1997, 1998), our findings indicated that this evenhandedness was recognized as fair and that it was seen as having some rational basis (i.e., to be quite sensible), but nevertheless, it was not a significant determinant of participants’ support for the leader. On the contrary, in both studies, support for a leader’s decision to nominate particular ingroup members for a prize was strongest when that nomination favored ingroup members (students) who opposed rather than supported a relevant (anti–gun control or pro–university cuts) outgroup. Other researchers have made the point, however, that there is a potential world of difference between supporting a leader’s behavior in the present and willingness to embrace his vision for the future (e.g., Hollander, 1985,

1477

1995). In this regard, the most significant feature of our second study was that it established the validity of this assertion by showing that although support for a leader’s act of evenhandedness was quite strong and also was reflected in an apparent willingness to endorse his future plans, this support was not translated into any activity that would help make those plans a reality. Consistent with the social identity approach, what we observed instead was that followers were only willing to “go the extra mile” on behalf of a leader when he himself had previously gone to some trouble to ensure that he represented and affirmed the values of the group (and did not merely “sit on the fence”). These findings are consistent with the argument that the key to leadership is leaders’ and followers’ belief that they are partners in a shared social categorical relationship that positively differentiates their ingroup from other comparison outgroups (Duck & Fielding, 1999; Haslam, 2001; Haslam & Platow, in press; Hogg, Hains, & Mason, 1998; Reicher, Drury, Hopkins, & Stott, in press; Reicher & Hopkins, 2001). In considering these various claims and observations, it seems appropriate to reflect on how they vary from those that might be made by proponents of other theoretical approaches. In particular, this is because our arguments have much in common with transactional approaches to leadership, which suggest that the key to a leader’s success is his or her ability to behave in a way that advances the interests of as many group members as possible. As we outlined in the introduction, these arguments are particularly associated with the work of Hollander (1985, 1995). Hollander correctly points out that leadership is contingent on the enthusiasm of followers, but he argues that this followership flows from principles of social exchange such that group members’ willingness to exert effort on behalf of the leader is directly related to their perception that the leader is willing to exert effort on behalf of them. The key point of difference between these arguments and those that we have attempted to develop in this article relates to the level of self-categorization at which participants are thought to interrelate. In particular, as Tyler (1999) noted, whereas principles of social exchange generally assume that people are motivated by personal reward, we have argued that followers are instead motivated by group-level concerns. In contrast to the arguments of exchange theorists, we would therefore argue that participants in this study were not motivated by personal self-interest (as if followership was determined by a satisfactory answer to the question “What’s in this for me?”) but rather by collective self-interest (an awareness of “What’s in this for us”). Indeed, as in Tajfel et al.’s (1971) original minimal group studies, the scenarios in our experiments were explicitly designed in such a way


1478


as to rule out the possibility of personal gain because the leader’s behavior was oriented toward specific people (student counselors) who were connected to the participants only by their membership in a common social category (as students who were pro–gun control or anti– university cuts). It seems reasonable and parsimonious to conclude, then, that sensitivity to the values and interests of this social self-category was the primary motivator of the participants in this study. It is also worth noting that although we construe the key independent variable in both studies as having involved a manipulation of identity affirmation, it could be argued that it was simply a manipulation of attitude congruence—with the observed effects merely showing that people follow leaders who hold similar beliefs to their own. We think there is some validity to this point, in so far as attitudes are clearly a component of social identity (with attitudinal similarity helping to define an ingroup; Turner, 1985). However, were attitude congruence the only determinant of followership, one would have expected the evenhanded leader to inspire as much followership as the leader we have characterized as identity affirming given that people generally endorse attitudes commensurate with fairness (Platow et al., 1998; Tyler, 1994). We would argue, then, that social identity principles must be invoked to explain which leader attitudes it is that people value in any given social context. Turning again to the questions with which this article opened, it therefore seems to be the case that the primary factor that motivates group members to engage in collective projects of the form typically identified by leaders is a belief that the leader is representative of their collective aspirations as group members. Because the meaning of the group changes (e.g., as a function of normative and comparative context; Turner et al., 1994), the exact nature of those aspirations will necessarily vary (Platow et al., 1997; Reicher et al., in press). This is one reason why leaders who adopt rigid positions on grouprelated issues are destined to be swept away with the tides of historical change. The same analysis also suggests that the key to successful leadership does not lie in the enduring and fixed personalities of leaders (which are a major component of most contingency theories) but in the higher order relationships between leaders and followers. It is only where these are established and affirmed that a leader can successfully translate his or her consideration for group members into initiation of structure. Paradoxically, however, we would expect that it is in those situations where this affirmation of identity is associated with group success that attributions to the leader’s personal charisma are most likely to be made (Haslam et al., in press; Meindl & Ehrlich, 1987). The charisma here is real. However, contrary to popular managerial wisdom, it is not the hallmark of maverick or independent

personality but the manifestation of an emergent social identity in which leaders and followers are creatively united. REFERENCES Burns, J. M. (1978). Leadership. New York: Harper & Row. Cooper, J. B., & McGaugh, J. L. (1963). Leadership: Integrating principles of social psychology. In C. A. Gibb (Ed.), Leadership: Selected readings. Baltimore: Penguin. De Cremer, D. (in press). Charismatic leadership and co-operation in social dilemmas: A matter of transforming motives? Journal of Applied Psychology. Duck, J. M., & Fielding, K. S. (1999). Leaders and subgroups: One of us or one of them? Group Processes and Intergroup Relations, 2, 203230. Fiedler, F. E. (1978). The contingency model and the dynamics of the leadership process. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 11). New York: Academic Press. Fleishman, E. A., & Peters, D. A. (1962). Interpersonal values, leadership attitudes, and managerial success. Personnel Psychology, 15, 43-56. Haslam, S. A. (2001). Psychology in organizations: The social identity approach. Thousand Oaks, CA: Sage. Haslam, S. A., McGarty, C., Brown, P. M., Eggins, R. A, Morrison, B. E., & Reynolds, K. J. (1998). Inspecting the emperor’s clothes: Evidence that randomly selected leaders can enhance group performance. Group Dynamics: Theory, Research and Practice, 2, 168-184. Haslam, S. A., & Platow, M. (in press). Your wish is our command: The role of shared social identity in translating a leader’s vision into a follower’s task. In M. A. Hogg & D. Terry (Eds.), Social identity processes in organizations. New York: Taylor & Francis. Haslam, S. A., Platow, M. J., Turner, J. C., Reynolds, K. J., McGarty, C., Oakes, P. J., Johnson, S., Ryan, M. K., & Veenstra, K. (in press). Social identity and the romance of leadership: The importance of being seen to be “doing it for us.” Group Processes and Intergroup Relations. Haslam, S. A., Powell, C., & Turner, J. C. (2000). Social identity, selfcategorization and work motivation: Rethinking the contribution of the group to positive and sustainable organizational outcomes. Applied Psychology: An International Review, 49, 319-339. Haslam, S. A., & Turner, J. C. (1992). Context-dependent variation in social stereotyping 2: The relationship between frame of reference, self-categorization and accentuation. European Journal of Social Psychology, 22, 251-278. Hogg, M. A. (1996). Intragroup processes, group structure and social identity. In W. P. Robinson (Ed.), Social groups and identities: The developing legacy of Henri Tajfel (pp. 65-93). Oxford, UK: Butterworth-Heineman. Hogg, M. A., Hains, S. C., & Mason, I. (1998). Identification and leadership in small groups: Salience, frame of reference, and leader stereotypicality effects on leader evaluations. Journal of Personality and Social Psychology, 75, 1248-1263. Hollander, E. P. (1985). Leadership and power. In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (3rd ed., pp. 485-537). New York: Random House. Hollander, E. P. (1995). Organizational leadership and followership. In P. Collett & A. Furnham (Eds.), Social psychology at work: Essays in honour of Michael Argyle (pp. 69-87). London: Routledge. House, R. J., Spangler, W. D., & Woycke, J. (1991). Personality and charisma in the U.S. presidency: A psychological theory of leader effectiveness. Administrative Science Quarterly, 36, 364-396. King, N., & Anderson, N. (1995). Innovation and change in organizations. London: Routledge. Meindl, J. R., & Ehrlich, S. B. (1987). The romance of leadership and the evaluation of organizational performance. Academy of Management Journal, 30, 91-109. Nadler, D. A., & Tushman, M. L. (1990). Beyond the charismatic leader: Leadership and organizational change. California Management Review, 32, 77-97.


Haslam, Platow / SOCIAL IDENTITY AND FOLLOWERSHIP Oakes, P. J. (1987). The salience of social categories. In J. C. Turner, M. A. Hogg, P. J. Oakes, S. D. Reicher, & M. S. Wetherell (Eds.), Rediscovering the social group: A self-categorization theory (pp. 117141). Oxford, UK: Blackwell. Oakes, P. J., Haslam, S. A., & Turner, J. C. (1994). Stereotyping and social reality. Oxford, UK: Blackwell. Organ, D. W. (1988). Organizational citizenship behavior: The good soldier syndrome. Lexington, MA: Lexington. Ouwerkerk, J. W., Ellemers, N., & de Gilder, D. (1999). Social identification, affective commitment and individual effort on behalf of the group. In N. Ellemers, R. Spears, & B. J. Doosje (Eds.), Social identity: Context, commitment, content (pp. 184-204). Oxford, UK: Blackwell. Pfeffer, J. (1992). Managing with power. Boston: Harvard Business School Press. Pfeffer, J. (1997). New directions for organizational theory: Problems and prospects. New York: Oxford University Press. Platow, M., Reid, S., & Andrew, S. (1998). Leadership endorsement: The role of distributive and procedural behavior in interpersonal and intergroup contexts. Group Processes and Intergroup Relations, 1, 35-47. Platow, M. J., Hoar, S., Reid, S., Harley, K., & Morrison, D. (1997). Endorsement of distributively fair or unfair leaders in interpersonal and intergroup situations. European Journal of Social Psychology, 27, 465-494. Platow, M. J., Mills, D., & Morrison, D. (2000). The effects of social context, source fairness, and perceived self-source similarity on social influence: A self-categorization analysis. European Journal of Social Psychology, 30, 69-81. Platow, M. J., & van Knippenberg, D. (in press). The effects of leader ingroup prototypicality and distributive fairness on leadership endorsements. Personality and Social Psychology Bulletin. Reicher, S. D., Drury, J., Hopkins, N., & Stott, C. (in press). A model of crowd prototypes and crowd leadership. In C. Barker (Ed.), Leadership and social movements. Manchester, UK: Manchester University Press. Reicher, S. D., & Hopkins, N. (2001). Self and nation. London: Sage Ltd. Reynolds, K. J., Turner, J. C., & Haslam, S. A. (2000). When are we better than them and they worse than us? A closer look at social discrimination in positive and negative domains. Journal of Personality and Social Psychology, 78, 64-80. Sherif, M. (1962). Intergroup relations and leadership: Introductory statement. In M. Sherif (Ed.), Intergroup relations and leadership:

1479

Approaches and research in industrial, ethnic, cultural, and political areas (pp. 3-21). New York: John Wiley. Smith, P. M. (1995). Leadership. In A.S.R. Manstead & M.R.C. Hewstone (Eds.), The Blackwell dictionary of social psychology (pp. 358-362). Oxford, UK: Blackwell. Tajfel, H. (1972). La catégorisation sociale [Social categorization]. In S. Moscovici (Ed.), Introduction à la psychologie sociale [Introduction to social psychology] (pp. 272-302). Paris: Larouse. Tajfel, H., Flament, C., Billig, M. G., & Bundy, R. F. (1971). Social categorization and intergroup behaviour. European Journal of Social Psychology, 1, 149-177. Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33-47). Monterey, CA: Brooks/Cole. Turner, J. C. (1982). Towards a cognitive redefinition of the social group. In H. Tajfel (Ed.), Social identity and intergroup relations (pp. 15-40). Cambridge, UK: Cambridge University Press. Turner, J. C. (1985). Social categorization and the self-concept: A social cognitive theory of group behaviour. In E. J. Lawler (Ed.), Advances in group processes (Vol. 2., pp. 77-122). Greenwich, CT: JAI. Turner, J. C. (1991). Social influence. Buckingham, UK: Open University Press. Turner, J. C., & Haslam, S. A. (2001). Social identity, organizations and leadership. In M. E. Turner (Ed.), Groups at work: Advances in theory and research (pp. 25-65). Hillsdale, NJ: Lawrence Erlbaum. Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987). Rediscovering the social group: A self-categorization theory. Oxford, UK: Blackwell. Turner, J. C., Oakes, P. J., Haslam, S. A., & McGarty, C. A. (1994). Self and collective: Cognition and social context. Personality and Social Psychology Bulletin, 20, 454-463. Tyler, T. R. (1994). Psychological models of the justice motive: Antecedents of distributive and procedural justice. Journal of Personality and Social Psychology, 67, 850-863. Tyler, T. R. (1999). Why people co-operate with organizations: An identity-based perspective. In B. M. Staw & R. Sutton (Eds.), Research in organizational behaviour (Vol. 21, pp. 201-246). Greenwich, CT: JAI. Wit, A., & Wilke, H. (1988). Subordinates’ endorsement of an allocating leader in a commons dilemma: An equity theoretical approach. Journal of Economic Psychology, 9, 151-168. Received February 21, 2000 Revision accepted November 15, 2000
