THINKING OUTSIDE THE BOX BY LOOKING INSIDE THE BOX: EXTENDING THE COGNITIVE REVOLUTION IN LEADERSHIP RESEARCH Robert G. Lord* University of Akron
Cynthia G. Emrich Purdue University
Follower and leader cognitions play critical roles in many organizational phenomena. In this review, we report and reflect on a decade of cognitive leadership research. Following a brief introduction, the review is divided into two sections—one devoted to individual and dyadic cognition, and the other to collective cognition. We identified three central themes for each section. For individual and dyadic cognition, we identified metacognitive processes and leadership, implicit leadership theories (ILTs), and network based models of ILTs as central themes. For collective cognition, we identified charisma, organizational performance and sensemaking, and transformation and change as central themes. Our goal in this review was to uncover important underlying assumptions of selected studies to illuminate more clearly the path of cognitive leadership research over the past 10 years and its potential paths over the next ten years.
INTRODUCTION Twenty-five years ago, Eden and Leviatan (1975) concluded that “leadership factors are in the mind of the respondent. It remains to be established whether or not they are more than that” (p. 741). This once radical notion catalyzed a revolutionary approach to the study of leadership, whose underlying principle is this: If leadership resides, at least in part, in the minds of followers, then it is imperative to discover what followers are thinking. In this review, we report and reflect on the status of this cognitive revolution in leadership research. We also develop 13 propositions to summarize prior literature and offer guideposts for future researchers. * Direct all correspondence to: Robert G. Lord, Department of Psychology, University of Akron, Akron, OH; e-mail: [email protected]
Leadership Quarterly, 11(4), 551–579. Copyright 2001 by Elsevier Science Inc. All rights of reproduction in any form reserved. ISSN: 1048-9843
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
An Organizing Schema The notion of schema, or an organized knowledge structure, represents the centerpiece of cognitive research because much information processing occurs in a top-down, schema-guided manner (Walsh, 1995). This review also revolves around a schema—one based on the broad distinction between individual/dyadic cognition and collective cognition. Collective and individual cognitions differ in many respects. From our perspective, the most important difference is that collective cognition is neither created nor housed in the mind of a single individual. Instead, it reflects a socially constructed understanding of the world derived from social exchanges and interactions among multiple individuals in a group or organization. Imagine group members as nodes in a social network, and their social interactions as linkages among these nodes. It is possible to capture important aspects of individual cognition by focusing solely on the nodes in this social network. It is impossible to capture collective cognition in the same way, because one must focus on the linkages among these nodes. Simply to aggregate individual cognitions to the group level ignores the socially constructed nature of collective cognition. For this reason, we distinguish between research on individual/dyadic and collective cognition. In this review, our goal was to describe studies in a manner that clearly illuminates the path of cognitive leadership research over the past 10 years and its potential paths over the next 10. After considerable reflection and discussion, we identified three key underlying assumptions that have profoundly shaped the path of cognitive leadership research in the past decade, and that have the potential to do so in this decade. Together with the earlier discussed distinction between individual/dyadic cognition and collective cognition, these three underlying assumptions comprise the schema for this review. The first assumption acknowledges a perennial question in leadership research: What is the origin of causality? Does leadership reside in a leader, or does leadership emanate from the social system in which leaders and followers interact? Scholars who assume that leadership resides in leaders tend to pose such questions as which leader traits or behaviors inspire the greatest commitment among followers. To address such a question, scholars examine or manipulate factors associated with leaders themselves. For example, Podsakoff, MacKenzie, Moorman, and Fetter (1990) conceptualized transformational leader behavior as directly affecting subordinate trust and satisfaction—a perspective that they supported in a field study using survey methodology and correlational analyses. In contrast, scholars who assume that leadership emanates from the social system tend to pose such questions as which conditions are most conducive to leader emergence. Researchers address such questions by focusing on situational and follower factors. Emrich’s (1999) study of context effects in leadership perception illustrates this perspective. She manipulated the performance context of a potential leader’s team and discovered that crisis affected perceptions of leadership qualities. These contrasting examples show how research is guided by assumptions regarding causal origin. Our aim in uncovering causal assumptions is to gauge the balance between emphasis on leaders versus social systems over the past 10 years. The second assumption involves the presumed nature and use of perceivers’
Outside the Box
leadership schemas. Do perceivers retrieve static leadership schemas to interpret, encode, and recall a target’s behavior? Lord, Foti, and De Vader’s (1984) test of leadership categorization theory is based on a static view of schema, assuming that relatively fixed prototypes, learned through extensive experience, were the basis for leadership categorization. Although this view has been helpful, more recently, Walsh (1995) stressed the need for researchers to examine changes in knowledge structures. Consistent with this recommendation, some leadership scholars have begun to develop and test more dynamic models in which perceivers recreate prototypes each time they are used (Hanges, Lord, & Dickson, 2000). The third and final assumption pertains to the nature of a study’s dependent variables. Does the study in question examine proximal intervening cognitive processes such as attention and memory or distal outcomes such as group performance and follower satisfaction? Meindl, Ehrlich, and Dukerich (1985) characterized leadership as a “large cause” in the mind of followers, implying that it should cast a large shadow over groups and organizations, with multiple and far-reaching effects in both time and space. Leader cognitions, in contrast, are focused and proximal to leaders in time. Ideally, cognitions would be studied as mediating bridges to more distal outcomes, but this is rarely done in a single study. For example, Wofford, Joplin, and Cornforth’s (1996) study of leaders’ script choices in response to information about their groups’ performances builds theories of such mediating processes without direct linkages to real organizational outcomes. Conversely, Peterson’s (1999) study of leader procedural justice and follower satisfaction works backward from outcomes to build theories of mediating processes, but only measures the outcomes, not the actual mediating cognitions. Both types of research on leader cognitions can be valuable, though they measure different aspects of leadership processes. Key Parameters for Review
Central Themes To save space and provide cohesion, we focus on central themes that emerged from the cognitive leadership literature. Within a particular theme, we describe individual studies with the goal of uncovering their underlying assumptions regarding causality, stability of schemas, and measurement focus. For individual/dyadic cognition, central themes pertaining to metacognitive processes and leadership, and network based models of ILTs emerged as central themes. For collective cognition, charisma, organizational performance, and sensemaking, and transformation and change emerged as central themes. Sources We relied on electronic retrieval of sources from several databases and surveyed over 35 journals and several books (Hogg & Tindale, 2000; Zaccaro & Klimoski, in press) to identify studies for inclusion in this review. The beginning of our timeframe coincides with the founding of The Leadership Quarterly in 1990. Nonetheless, we avoid covering the same terrain as previous reviews of the organizational (Walsh, 1995) and leadership perception (Hall & Lord, 1995) information processing
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
literatures that have appeared during this time. Yet we also have tried to build on the recommendations of such reviews by considering supra-individual knowledge structures, and the development, change, and use of knowledge structures as recommended by Walsh (1995). Following Hall and Lord (1995), we also were sensitive to levels of analysis issues and the joint role of cognitions and affect. Our concern with such issues, however, was also limited by the nature of the research we reviewed.
INDIVIDUAL AND DYADIC COGNITION The three emergent themes at this level—metacognitive processes, ILTs and attributional processes, and the contextualization of schemas—all emphasize the interaction of leaders and followers in determining cognitive processes as well as behavior. To the extent that the studies reviewed here reflect the interplay between leaders and followers, they are dynamic, even though few explicitly emphasize these dynamic aspects in their design or measures. We turn now to trends in these three themes. Metacognitive Processes and Leadership
Supervisor-Subordinates Interactions and Metacognitions Metacognitive processes are higher level cognitive systems that regulate moment to moment cognitions (e.g., scripts). They manage the internal cognitive context that, in turn, influences leader thoughts and behaviors. Understanding or changing leader behavior requires changes in metacognitive processes. Leaders may also affect subordinates by influencing their metacognitive processes. Wofford and Goodwin (1994) developed an important theoretical basis for examining leader metacognitive processes. They maintain that the immediate source of leader behavior is the activation of scripts, which are goal-oriented structures with different tracts or elements. In their model, leaders select the appropriate script or script tract, which then helps them assimilate environmental information and internal thoughts in producing behavior. To test this metacognitive theory, Wofford et al. (1996) used a behavioral simulation to obtain measures of leader thought processes and intended behavior, and to manipulate group performance feedback and causal information about group members. They found that leaders shifted to more directive scripts when feedback indicated major problems and group members were thought to have low ability and motivation. Also, measures of new script tracts were correlated with behavioral intentions, supporting a linkage between cognitions and behavior. Wofford et al.’s work is laudable in the use of protocols to measure cognitive processes and in its grounding in a clearly articulated cognitive theory. It also demonstrates that the locus of leadership behavior lies in the interplay among task performance, perception of subordinate qualities, and a leader’s own metacognitive processes. Thus, it reflects a systems perspective on leadership. In a laboratory experiment using ongoing groups, Offermann, Schroyer, and Green (1998) investigated metacognitions from an attribution theory perspective. They told leaders that their groups had performed well or poorly and gave one of three possible explanations (effort, ability or luck). Using videotapes of interactions
Outside the Box
to provide objective measures of subsequent behavior, they found that leaders changed their behavior based on this feedback, increasing talkativeness and negative comments after failure resulting from low effort or after success resulting from luck. Punishment recommendations were also highest when low performance was coupled with internal attributions (ability or effort). These findings are consistent with results from Wofford et al.’s (1996) leadership simulation that also found that performance feedback and causal attributions influenced the behavioral intentions of leaders. Using field data, Wofford, Goodwin, and Whittington (1998) also investigated metacognitive processes of leaders. Open-ended questions and content analysis were used to build measures of schemas and scripts, finding that transformational self-schema and motivational schema were indicators of transformational cognitions, which in turn, predicted transformational leadership behavior as measured by the Multifactor Leadership Questionnaire (Bass, 1985). Similarly, transactional motivational schema and self-schema were indicators of transactional managementby-exception cognitions, which, in turn, predicted transactional management-byexception behavior. Thus, behavioral reports from subordinates directly related to the leadership schema of their superiors. In an innovative study, Brown (2000) adopted Higgins’ (1998) cognitive/motivational framework to investigate the interplay between leaders and follower. Brown theorized that different motivational orientations of leaders create different metacognitive processes in subordinates, which then regulate subordinates’ task behavior. To test this idea, he developed a novel technique that combined the use of experimental primes and a motivational orientation questionnaire that measured subordinates’ perceptions of whether their supervisors emphasized ought or ideal selfguides. Two weeks after measuring their perceptions of the supervisors’ motivational orientations, subjects performed an experimental task under conditions that primed either their supervisor or a neutral situation. Consistent with Higgins’ theory, persistence on unsolvable problems was associated with an “ideal” as opposed to an “ought” orientation of subjects’ supervisors, but only under the supervisor prime condition. By demonstrating that visualization procedures can be used to bring supervisor related motivational dynamics into the laboratory, Brown has provided a technique to experimentally investigate the “large shadow” cast by real organizational supervisors on their subordinates’ metacognitive processes. In sum, several innovative studies show that leaders and followers use metacognitive processes to guide behavior. Importantly, they also show that aspects of their dyadic partner figure prominently in such metacognitions, thereby illustrating the need to adopt a systems approach to understanding leader-follower interactions whether the emphasis is on leader behavior as in Wofford’s work or on subordinate motivations as in Brown’s research. Such research also demonstrates how a socialcognitive orientation can lead to novel methodology.
Metacognitions and Self or Social Knowledge As the prior research has shown, to effectively select appropriate behavioral scripts or motivational schemas, leaders must be able accurately to assess themselves and subordinates in dynamic social systems. Thus, metacognitive processes related
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
to social perceptions are also important in understanding individual and dyadic level processes. In one of the few studies looking at metacognitions and self-perceptions of leaders, Prussia, Anderson, and Manz (1998) found that metacognitive factors related to self-perceptions were critical in translating actual leadership skills into performance outcomes. Similar reasoning underlies an extensive program of research by Eden (1992) on self-fulfilling prophecies. He finds that leaders’ beliefs regarding subordinates’ ability serve metacognitive functions that cause leaders to behave toward subordinates in a manner that confirms their expectations (Pygmalion Effects). In additions, leaders’ expectations also can be subtly transmitted to subordinates, affecting their self-efficacy, motivation, and performance as well. Interestingly, seven recent studies of Pygmalion leadership training (Eden, Geller, Gewirtz, Gordon-Terner, Inbar, Liberman, et al., 2000) found only weak effects of leadership training. One possible explanation is that trained leaders, who would be aware of Pygmalion effects, processed information about subordinates differently than leaders in previous studies who presumably were unaware of the operation of Pygmalion effects. Awareness that social information can affect judgments often differentiates between assimilation and contrast effects in social judgments. Most metacognitive studies, however, have examined the extent to which leaders’ metacognitions afford insight into how they are perceived by others or how they adjust to social demands. For example, Zaccaro, Gilbert, Thor, and Mumford (1991) discuss metacognitions related to social demands, defining social intelligence as both the ability to perceive accurately social requirement in situations and the ability to adjust behavior accordingly. Viewing self-monitoring ability as one aspect of social intelligence, Zaccaro, Foti, and Kenney (1991) and Hall, Workman, and Marchioro (1998) found self-monitoring ability to be related to leadership emergence. Other research links a leader’s self-awareness (Sosik & Megerian, 1999) and self-monitoring ability (Sosik & Dworakivsky, 1998) to transformational or charismatic leadership. In a study that may be most important for its gender-related implications, Malloy and Janowski (1992) found that aggregate metaperceptions (self-other agreement) were highly accurate, but that dyadic metaperceptions were inaccurate. They explained this finding by arguing that aggregate metaperceptions were based on a target’s awareness of the extent to which his or her behavior was leader-like (i.e., fit generally held implicit theories of leadership and stereotypes). Consistent with this logic, females perceived that they would be judged significantly lower on leadership than men did, and they received lower leadership ratings than men. At least for females, awareness of how one is likely to be judged by others may limit attempts at leadership. In sum, metacognitive studies of information processing and leadership, though limited in number, reveal a complex process by which actors use knowledge about themselves and their social situations to guide leadership activities. Feedback on task performance and from group members may be equally important. As shown in the following two propositions, which summarize major ideas, a social systems view of causality is implicit in such work.
Outside the Box
Proposition 1: Causality for leadership behavior and perceptions lies in a complex social system involving the leader, follower, task, and context. Proposition 2: The effects of social systems on: (a) leader behavior are mediated by leader metacognitive processes, whereas (b) leadership perceptions are mediated by follower metacognitive processes. Implicit Leadership Theories
Role in Leadership Perceptions As noted earlier, cognitive research has emphasized the role of followers as well as leaders in leadership perceptions (Hall & Lord, 1995). Extensive research using an information processing approach shows that individuals can be recognized as leaders based on their fit with leadership prototypes or that leadership can be inferred from favorable outcomes. Once someone is labeled as a leader, perceivers use their ILTs as a basis for retrieving previously observed leadership behaviors and often experience difficulty distinguishing between observed and unobserved behaviors that are prototypical of leadership (see Lord & Maher, 1991, for a review of this literature). More recent research (Murphy & Jones, 1993) has shown that inferred leadership based on performance outcomes occurs only with person-based encoding. In contrast, script-based encoding elicits situational explanations, which previous work (Phillips & Lord, 1981) has shown to minimize performance cue effects. Consistent with such research, Yorges, Weiss, and Strickland (1999) found that internal attributions were associated with attributions of charisma and the potential influence of a leader. Such research shows that perceptions of leadership involve comprehensive interpretations of situational factors and that leadership perceptions depend on seeing a leader’s behavior as reflecting personal qualities rather than situationally induced responses. Prototype matching processes have also been examined by Smith and Foti (1998), who stressed that prototypes involve patterns of traits and that these patterns contain important information that goes beyond their specific elements. In an experimental study of newly formed groups, they found that a pattern of high dominance, high general self-efficacy, and high intelligence was strongly associated with leadership emergence. Consistent with their predictions concerning the importance of patterns, Smith and Foti also found the three-way interaction of these traits to be significant, adding variance in leadership emergence not explained by the main effects of these three traits or their two-way interactions. Their results show that even some of the oldest ideas regarding leadership—that individuals who possessed certain traits tended to emerge as leaders—can be advanced by integration with information processing theory. Contextual Constraints on Leadership Cognitive research in leadership perceptions bring followers to center stage but, with the exception of cross-cultural and attribution theory research, largely ignores the broader context in which leadership perceptions occur. One contextual factor is the supervisory role itself, which was examined by Konst, Vonk, and Van Der Vlist (1999). Using a sentence completion technique Konst et al. found that for
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
many subjects (who were both supervisors and subordinates), the label “leadership” triggers spontaneous causal attributions and implies potency. This interpretation, however, must be tempered because Konst et al. did not actually manipulate leadership labels but instead manipulated supervisor/subordinate roles. Several studies have examined the performance context associated with leadership, paying particular attention to crisis situations. Meindl (1995) theorized that crisis situations increase reliance by perceivers on leadership as an explanatory construct; however, empirical work in applied settings shows that this relationship is more complex. Contrary to predictions, Pillai and Meindl (1998) found that charisma was negatively related to perceived crisis. This may result from the fact that in ongoing organizations, the existence of a crisis situation implies ineffective leadership. As Emrich (1999) observed, when leaders are seen as “part of the problem” rather than “part of the solution,” crisis may likely be negatively associated with charisma. Other work has effectively used both experimental designs and information processing methodology to investigate the relationship between crisis and leadership perceptions. Emrich (1999) found that a context of crisis unconsciously activated perceivers’ leadership schemas. Using a simulated selection task, she found greater false recognition of leadership behaviors for an applicant for a managerial job when the applicant was expected to manage a team in crisis versus a team that was performing well. This difference in recognition memory occurred even though perceivers in both context conditions received identical information about the applicant. Importantly, Emrich’s target was a potential job applicant and, therefore, could not be held responsible for the crisis situation as in the Pillai and Meindl (1998) study. Hunt, Boal, and Dodge (1999) also used an experimental design to show that crisis expands the type of behaviors that produce charismatic leadership perceptions. They identified two types of charisma in leadership behavior: visionary charisma, in which leaders consistently stress new interpretive schemas, and crisis-responsive charisma, in which leaders only emphasize new interpretive schemas after a crisis has developed. Manipulating both types of leadership behavior as well as the existence of crisis in a laboratory experiment, they found that vision-based leadership produced stable charismatic perceptions in both crisis and noncrisis situations; the positive effects of crisis-responsive charisma, however, decayed quickly once the crisis abated. Using an experimental design in a real classroom situation, Pillai (1996) actually created a crisis/non-crisis factor by giving bogus quiz feedback to students. She found that in a subsequent group exercise, the crisis/non-crisis factor significantly affected ratings of charisma (eta2 ⫽ .07) but not exchange behaviors, with higher charisma ratings in the crisis situation. Taken together, the work of Emrich, Pillai, Meindl, and Hunt et al. shows that perceivers integrate contextual information and ILTs regarding leadership prototypes to form leadership perceptions. The major patterns in findings are summarized in the following proposition. Proposition 3: Crisis situations tend to elicit perceptions of charisma from perceivers as long as the leader being rated is not responsible for the crisis. When the leader is responsible, crisis is negatively related to perceptions of charisma.
Outside the Box
Content, Source, and Use of ILTs In perhaps the most comprehensive assessment of ILT content, Offermann, Kennedy, and Wirtz (1994) had multiple samples generate and rate items that were associated with business leaders. Their results suggest that leadership knowledge structures contain eight broad dimensions: sensitivity, dedication, tyranny, charisma, attractiveness, masculinity, intelligence, and strength. In addition, these factors did not differ across sex groupings or across stimulus conditions that described a “leader,” an “effective leader,” or a “supervisor.” Other research (Kenney, Schwartz-Kenney, & Blascovich, 1996) found support for implicit theories of leaders “worthy of influence.” These ILTs specify subordinate expectations that, in turn, influence leader acceptance. Thus, studies of ILT content have practical implications for understanding the image that leaders must project to garner social influence and, thereby, increase their impact on subordinates. A novel question addressed by Keller (1999) concerns the source of ILTs. It is generally theorized that ILTs develop from experience with leaders, but an individual’s values or personality may influence the type of situations one selects, the type of behavior one elicits from leaders, or the type of experiences that are noticed and stored in memory. Using Offermann et al.’s (1994) dimensions of ILTs, Keller found several instances where subjects’ personalities predicted their ILTs. For example, individuals who were higher on the Big Five dimensions of agreeableness, dedication, and extroversion had ILTs that were higher on sensitivity, dedication, and charisma, respectively. Keller also reasoned that different types of leaders influence our leadership experiences and hence our ILTs. Theorizing that parents are the most consistent leaders to whom college students have been exposed, she also asked students to assess whether Offerman et al.’s eight ILT dimensions described the personality of their parents. Dedication and tyranny of parents predicted these same dimensions in offsprings’ ILTs, with the effects of fathers being substantially larger than those of mothers. An interesting follow-up question would be whether ILTs for female leaders fit better with their mothers’ traits. Few studies have examined the relation of implicit theories to leadership processes. One exception is the work of Engle and Lord (1997), which examined implicit performance theories (IPTs) and ILTs of both superiors and their subordinates. Results showed that the normativeness of supervisors’ ILTs and IPTs was strongly related to liking and LMX when measured from both subordinates’ and supervisors’ perspectives. Subordinates’ implicit theories, however, showed minimal relation to dependent variables. Engle and Lord also found that congruence on IPTs predicted both supervisor-rated LMX and supervisor liking for subordinates. Thus, high quality exchanges seemed to form around dyads that defined good performance similarly. Cultural Specificity of ILTs Cognitive theory has shown that schemas such as ILTs are learned through experience. Consequently one would expect different cultures to produce different leadership experience and different ILTs. This simple logic has motivated several cross-cultural comparisons of ILTs (Chong & Thomas, 1997; Den Hartog, House, Hanges, Ruiz-Quintanilla, and Dorfmann, 1999; Gerstner & Day, 1994; Thomas, 1997). The study of cross-cultural effects of ILTs has also been spurred by the
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
globalization of business in the past decade. For example, Gerstner and Day compared the leadership prototypes of eight groups of students studying in the United States who had different national origins. They found no universal leadership traits, but several traits were common to subgroups of countries. Gerstner and Day also went beyond mere description by constructing a multidimensional scaling of countries based on prototype similarity. They found a three-dimensional Multidimensional Scaling (MDS) solution to be strongly related to each country’s values on the Hofstede’s dimensions of Power Distance, Uncertainty Avoidance, and Individualism. This result implies that leadership prototypes reflect the higher-level values that characterize a culture, a process examined in detail by Lord and Brown (2000). The most comprehensive assessment of cross-cultural differences in leadership prototypes comes from the work of Robert House and his GLOBE partners. In a study involving 15,022 middle managers from 60 different societies/cultures, Den Hartog et al. (1999) examined the generality of both specific items and higher order factors associated with effective leadership. Three of their six second-order factors (Charismatic/Value Based, Team Oriented, and Participative) were prototypical of outstanding leadership in all cultures. Similarly, several sets of items pertaining to integrity, charisma and team-orientation were universally endorsed. The items ruthless, asocial, irritable, loner, egocentric, non-explicit, non-cooperative, and dictatorial were universally rejected. Such work provides a substantial descriptive basis for cross-cultural cognitive theories of leadership, but it does not appear to test any specific cognitive theory. However, this work is firmly grounded in categorization theories of leadership, and it illustrates the effective use of cognitive methodology. The following two general propositions summarize key ideas in the literature on ILTs. Proposition 4: The most immediate determinant of leadership perceptions (leader behavior) is the activation of perceivers’ (leaders’) ILTs. Proposition 5: The contextual sensitivity of perceivers’ (or leaders’) metacognitive processes operates through their differential activation of appropriate ILTs. Network Based Models of ILTs
Dynamic Quality of ILTs While most research on ILTs has been content focused, a few researchers have emphasized underlying processes, questioning the more static views of category learning. One particularly promising area, connectionist modeling, explains how contextual information such as culture, crisis, or leader qualities can automatically influence ILTs. Hanges et al. (2000) adopt such a dynamic perspective on ILTs, proposing that leadership prototypes and schema can be conceptualized in terms of connectionist networks. Connectionist networks organize simple processing units so that when they are mutually activated, meaningful constructs are created. Such networks regenerate, rather than retrieve, schemas or prototypes each time they are used, subject to
Outside the Box
current constraints. Thus, connectionist networks explain how context effects can occur. Another important aspect of such networks is that they reflect patterns of relations and, as Smith and Foti (1998) have shown, patterns may be as important as specific attributes in explaining leadership emergence. Connectionist networks can also complete patterns from partial information, thereby explaining how perceivers generalize from known information to form expectations about unobserved behaviors or traits. Thus, connectionist networks can explain why schema consistent false-positives often occur in descriptions of leadership behavior. Lord, Brown, and Harvey (in press) have also argued that connectionist networks underlie both the prototypes used in leadership perceptions and the scripts used to generate behavior. They maintain that networks can adjust prototypes used in leadership perceptions and the scripts used to generate behavior. They maintain that networks can adjust prototypes (or behavior) to context because they use parallel processes to rapidly integrate situational constraints with past learning. Because connectionist networks can involve preconscious processes, they also are a means by which peripheral factors can influence definitions of leadership or expectations for behavior. Physical features associated with race, gender, or ethnicity, for example, may prime specific components in prototypes, leading perceivers to expect different behaviors from potential male as compared to female leaders, without any awareness that such processes are operating. Smith (1997) used a priming methodology to demonstrate such effects. Equally important, factors such as the nonverbal behavior of a leader or follower emotions, that may not always be consciously available to individuals, can also operate through connectionist networks. Along these lines, Awamleh and Gardner (1999) report that the way in which leaders delivered information had much greater effects than did vision content on perceptions of leader charisma and effectiveness.
Representation of Contextual Constraint and Metacognitions Though offering new insights into how cognitive structures can guide leadership perceptions and processes, connectionist models have not yet been empirically tested in the leadership domain. They do, however, offer a specific processing model of how general ILTs could automatically be adjusted to reflect a variety of contextual constraints related to leadership. This view is consistent with research that shows that ILTs change with contextual factors associated with leader ethnicity (Chong & Thomas, 1997), leader roles (Konst et al., 1999), close or distant positions in organizational hierarchies (Yagil, 1998), crisis situations (e.g. Meindl and Pillai’s work), cultural views (Gerstner & Day, 1994), task type (Hall et al., 1998; Sutton, 1998), and the extent of one’s identification with a group (Hains, Hogg, & Duck, 1997; Hogg, Haines, & Mason, 1998; Van Vugt & De Cremer, 1999). More traditional views of ILTs cannot accommodate such flexibility, because they would require enough experience in each new context to learn a new leadership prototype. One can also view dynamic adjustment through connectionist networks as a process that underlies the metacognitive aspects of leadership discussed earlier. This research showed that leaders adjusted their behavior based on attributional information and knowledge of subordinates’ performance (Offerman et al., 1998; Wofford et al., 1996, 1998). Although leaders may select from pre-existing script
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
tracts, as suggested by Wofford and Goodwin (1994), they may also construct script tracts on the spot. Both script tract retrieval (reconstruction) and construction of new, context-appropriate script tracts can be produced by using connectionist networks and contextual constraints. The following proposition summarizes these ideas. Proposition 6: The differential activation of ILTs in various contexts is produced by dynamic, connectionist network-based processes that create new, contextually tuned ILTs each time they are used.
COLLECTIVE COGNITION Three important themes emerged from our search for collective level cognitive leadership research: charisma, performance and sensemaking, and organizational transformation and change. The term collective cognition may appear to be a misnomer when applied to the studies in this section because, with few exceptions, the studies do not examine collective cognition as defined when we described our organizing schema—that is, as “a socially constructed understanding of the world derived from social exchanges and interaction”. Some of these studies examined key organizational outcomes such as customer ratings of team performance (Edmondson, 1999), while others aggregated individual level responses such as perceptions of work unit culture to the team level (Pillai & Meindl, 1998). Nonetheless, these studies can be characterized as collective because they examined the broader effects that leaders exert on groups, organizations, and nations. Charisma Conger and Kanungo (1988) declared charisma a “poorly-understood phenomenon” (p. 6). Since then, scholars have increasingly invoked cognitive explanations for charismatic behavior and its effects (Emrich, Brower, Feldman, & Garland, 2000; Fiol, Harris, & House, 2000; Gardner & Avolio, 1998; Pillai & Meindl, 1998; Shamir, House, & Arthur, 1993; Simonton, 1994). The cognitive aspects of these studies are still formative in nature; in all of them, however, cognitive principles are either invoked post hoc to explain findings (e.g., Emrich et al., 2000) or offered as underpinnings for as yet untested theoretical propositions (e.g., Shamir et al., 1993). None of these studies examined cognitive processes in real time, but they acknowledge the key role of follower cognitions in the charismatic relationship and, in doing so, provide excellent blueprints for future cognitive research on charisma.
Communication Charismatic leaders are believed to inspire “performance beyond expectations” (Bass, 1985) by appealing to followers’ fundamental and enduring motives rather than by promising followers material rewards (House & Shamir, 1993). In their self-concept based theory of charisma, Shamir et al. (1993) posit that charismatic leaders inspire followers by linking their visions to important aspects of followers’ self-concepts and self-esteem, by increasing followers’ identification with the leaders,
Outside the Box
by heightening followers’ collective identities, and by increasing followers’ self and collective efficacy. They analyzed Jesse Jackson’s 1988 speech to the Democratic National Convention as a test case for their theory. Jackson, for example, emphasized that he had grown up in poverty and thus was similar to many of his followers who struggled to improve their own lives. By stressing his similarity he presumably increased followers’ identification with him. He also invoked the metaphor of a quilt to illustrate that, as a collective, followers could create something that none could create individually. Though this research used an innovative methodology, Shamir et al.’s hypotheses could be tested more directly by manipulating speech content and then examining the effects on individuals’ self-concepts, self-esteem, attitudes, and behavioral intentions. Fiol, Harris, and House (1999) combined Lewin’s field theory with a semiotics approach to study the means through which the charismatic leader/follower interaction generates extraordinary social change. Their central thesis was that charismatic leaders create compatibility between their own and followers’ values through a process of frame breaking, frame moving, and frame re-aligning. Fiol, Harris, and House proposed several communication strategies for achieving this transformation. One strategy involved the charismatic leader’s use of more inclusive language such as “we” and “us” rather than exclusive language such as “I” and “you” (see also Brewer & Gardner, 1996). Another strategy involved the charismatic leader’s use of “strategic ambiguity” or language that is sufficiently abstract to accommodate multiple interpretations. In a study of various speeches of twentieth century U.S. presidents, Fiol and her colleagues found strong support for the hypotheses that charismatic presidents used these two (and other) communication strategies more frequently than did their non-charismatic counterparts. Given the archival and correlational nature of their study, Fiol et al. (1999) were unable to examine followers’ reactions to these various communication strategies; they did, however, suggest that future research could “incorporate empirical data concerning if, how, and when follower frames actually shift during the change process” (p. 474). Most recently, Emrich et al. (2000) examined the relationships among presidential charisma, greatness, and two types of rhetoric: abstract and pallid concept-based rhetoric (e.g., “help,” “dependable,” “commitment,” and “idea”) versus concrete and vivid image-based rhetoric (e.g., “hand,” “rock,” “heart,” and “dream”). They did so by examining two sets of speeches of elected U.S. presidents, from Washington through Reagan. The distinction between image-based and concept-based rhetoric originated in the study of artistic and literary change and greatness (e.g., Martindale, 1975, 1981, 1986). Although seemingly far afield from presidential charisma and greatness, Emrich et al. reasoned that artists, writers, and leaders face a similar challenge: to cultivate a loyal following by offering something new, different, and better. This theoretically grounded “hunch” found support in their two studies. U.S. presidents who engaged in more image-based rhetoric in their inaugural addresses were judged to be more charismatic (Study 1). U.S. presidents who engaged in more image-based rhetoric in speeches that they delivered at pivotal points in their administrations were judged to be both more charismatic and higher in greatness (Study 2). By drawing from research in multiple disciplines, Emrich et al. developed a
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
tentative theoretical account for these observed correlations. This account centered on “four initial inputs” into charismatic perception: attention, comprehension, emotions, and elaboration and memory. The overarching message of these two studies of presidential rhetoric was that a leader’s ability to articulate his or her vision in a way that enables followers to “see” in their minds’ eyes what the leader hopes to accomplish is one key to charisma. As with other recent studies of charisma, Emrich et al. stop short of examining underlying cognitive processes. Nonetheless, it would be straightforward to develop speeches that vary in their proportions of image-based rhetoric and then to examine the effects of this manipulation on the proposed four initial inputs into charismatic perception.
Systems and Structures Gardner and Avolio’s (1998) model of the charismatic relationship offers one of the few truly systems-based perspectives on charisma. In their model, they acknowledge not only the roles of leaders and followers but also the social environment. To construct this model, they adopt a dramaturgical perspective wherein “‘meaning’ is both the product of human interaction, and the defining quality of the ‘social act’” (p. 33). Cognitive processes in the form of information processing play a key role in this model. For example, to examine followers’ roles in the charismatic relationship, Gardner and Avolio introduce Lord and Maher’s (1993) work on leadership and social information processing. From this perspective, charismatic leader behavior constitutes novel or unusual behavior that requires followers to engage in more effortful, controlled processing to interpret and encode. Eventually, Gardner and Avolio argue, followers create a subtype—the charismatic leader subtype—as an addendum to their general leader prototype. This subtype would be reserved for those leaders who engage in the type of visionary and image building behavior that typifies the charismatic leader. This information-processing approach helps to explain how it is that charismatic leaders retain a “special place” in the minds of followers. Pillai and Miendl (1998) recently conducted one of the few empirical studies of the roles of organizational structure and culture on charismatic perception. They examined 101 work units of a government health agency—an agency that operates in a turbulent environment frequently punctuated by crisis. Pillai and Meindl discovered that respondents perceived their leaders to be more charismatic when their work units were organic in structure and collectivist in orientation and, as discussed previously, their leaders were not associated with crisis situations. To conclude this section, charisma is alive and well as a topic of investigation in our field (see also Keller, 1992; Pillai, 1996; and Yagil, 1998 for additional noteworthy studies). Though the perspectives of researchers exhibited considerable variety, the following two propositions encompass much of the thinking regarding charisma. Proposition 7: Charismatic leaders use a variety of communication based techniques (e.g., metaphors, collective pronouns, strategic ambiguity, and image-based rhetoric) to link their own visions or values to followers’ values and self-structures.
Outside the Box
Proposition 8: Charismatic leadership is dependent on facilitating social systems and structures. The studies just reviewed demonstrate the fruitfulness of both multi disciplinary and cognitive approaches to the perennial enigma that is charisma. They also demonstrate the range of assumptions that scholars bring to their research. The studies that focused on charismatic communication clearly emphasized the causal role of the leader in charismatic perception. According to some scholars, charismatic leaders articulate their visions in a way that is superior in many respects to the non-charismatic leader (Awamleh & Gardner, 1999; Emrich et al., 2000; Fiol et al., 2000; Shamir et al., 1993; Yagil, 1998). Yet none of these scholars acknowledged the larger social system within which charismatic leaders and their followers interact. In contrast, social systems and structures constituted a central focus of both Gardner and Avolio’s (1998) and Pillai and Meindl’s (1998) work (see also Pillai, 1996). Only Gardner and Avolio (1998) were explicit about the nature of perceivers’ leadership categories, however, which pertains to our second assumption. They adopted Lord and Maher’s (1991) information processing approach, which assumes that perceivers possess fairly stable representations of leaders that they retrieve to interpret and encode a target’s behavior. By default, we classify the other studies in the same static/retrieval mode. The final assumption concerns the measurement focus or range of effects that studies examined. All of the above-reviewed studies of charisma focused on outcomes—including perceptions of charisma, leader effectiveness, and work unit performance. Therefore, one fruitful avenue for future research would be a focus on intervening cognitive processes in the charismatic relationship. Performance and Sensemaking A potential criticism of leadership research is its focus on perceptions of leader behavior, traits, and styles to the exclusion of a focus on key outcomes such as leader, team, and organizational performance and sensemaking. Most of the studies in this section examine directly the link between various qualities of leaders and measures of either their own performance or the performance of their teams or organizations. All of them emphasize leader sensemaking as a key to effective leadership.
Team Performance Teams vary in their tasks and, therefore, in the type of leadership needed to facilitate team processes. Keller (1992) tested this idea in a study of 66 project groups in three industrial research and development organizations. He found considerable support for the prediction that transformational leadership would be more valuable in research teams because of the ill-structured and novel nature of their tasks, but that transactional leadership would be more valuable in development and service teams because their “work generally focuses on incremental technological improvements” (p. 492). Interestingly, team performance at Time 1 affected team members’ leadership ratings at Time 2 such that the more favorable team members’ percep-
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
tions of project quality, the more transformational they perceived their team leaders. Keller’s study is particularly interesting because it adopted a systems approach to the study of performance in project teams. It incorporated both person (i.e., leader behavior) and contextual (i.e., team task) variables. In addition, it recognized that the perceptual process contains a feedback loop from team performance to perceptions of leader behavior. In a recent study, Edmondson (1999) introduced and examined the concept of team psychological safety for interpersonal risk taking in a multi-method, multiphase field study of 51 manufacturing work teams. As hypothesized, team psychological safety positively affected both teams’ learning behavior and performance. The safer team members felt in taking risks in the team, the more they indicated that their team was always working to improve their process, and the more favorably customers rated their teams’ work performance. The relationship between team psychological safety and team performance was completely mediated by team learning. Interestingly, team leader coaching was a more consistent predictor of team psychological safety than was a context of support for the team, thereby demonstrating that leaders can have positive effects that go beyond securing tangible resources for their teams. Both of the studies in this section can be summarized by the following proposition. Proposition 9: Effective leadership involves sensemaking that links important outcomes such as performance to social (e.g., leadership) and psychological (e.g., safety) aspects of contexts, thereby facilitating organizational learning.
Organizational Performance Day and Lord (1992) examined the relationship between CEO expertise in problem solving and organizational sensemaking. In this study, they compared CEOs’ (experts) problem categorizations with those of MBA students’ (novices). Based on work on expert processing, Day and Lord distinguished between two approaches to problem solving—meaning-based and surface-based. Meaning-based problem solving is organized around “implicit principles and abstractions,” while surface-based problem solving is organized around explicit information in stimulus materials. As predicted, they found that CEOs, who tended to focus on the meaning dimension of organizational problems, were able to diagnose problems more quickly than MBA students, who tended to focus on the surface features. Clearly, speed in diagnosing problems would be a critical skill in today’s increasingly fast-paced business organizations. Interestingly, the number of categories that CEOs used to solve the problems was positively related to the number of processes and services offered by their firms. Given the correlational nature of these results, it is impossible to determine the direction of causality. Day and Lord speculated that CEOs developed knowledge structures whose complexity matched that of their firms’ offerings. It is possible, however, that firms’ offerings reflected, instead, the complexity of CEOs’ knowledge structures. This question would be interesting for future research on leader expertise and organizational sense-making.
Outside the Box
Denison, Hooijberg, and Quinn’s (1995) and Hooijberg, Hunt, and Dodge’s (1997) research on leader complexity is consistent with this latter interpretation— that leader complexity affects a wide range of key outcomes in organizations. Denison and his colleagues (1995), however, stopped short of linking a leader’s behavioral complexity with organizational performance. In a subsequent theory development article, Hooijberg, Hunt, and Dodge (1997) broadened their concept of leader complexity to include social complexity, or the “leader’s capacity to differentiate the personal and relational aspects of a social situation and integrate them in a manner that results in increased understanding or changed action-intention valences” (p. 382). Although theoretical in nature, Hooijberg et al.’s (1997) discussion provides a brief overview of social network analysis as a measure of leader social complexity. Using this analysis, one could measure the number of ties that a leader has with others, the range and closeness of these ties, a leader’s access to others, and the asymmetry between a leader’s in-coming and out-going interactions. The higher a given leader’s scores on these indicators, the greater his or her social complexity. Adopting a systems approach, Hooijberg and his colleagues hypothesized that a leader’s cognitive, behavioral, and social complexity would indirectly affect organizational effectiveness via its effect on followers. They offered a number of specific propositions for each link in their leaderplex model. One interesting avenue for future research would be to examine the potential transmission of complexity from leader to followers. A final study in this section on organizational performance and sensemaking is noteworthy. In one of the rare empirical studies of leaders’ roles in information technology (IT) management, Armstrong and Sambamurthy examined the role of senior leadership in IT assimilation to enhance their business performance. They surveyed Chief Information Officers (CIOs) and other members of the top management teams of 153 firms. As predicted, they found that senior leaders’ IT and business knowledge positively predicted a firm’s IT assimilation. They further discovered that senior leaders’ IT and business knowledge were greater in firms than espoused a transformational IT vision. They also discovered only partial support for the moderating role of a transformational IT vision in these relationships. Even so, this study is important in that it is the first to demonstrate the links among senior leadership knowledge, transformational vision, and IT assimilation. Future research might profitably track the development or transmission of senior leaders’ knowledge to IT assimilation. Two general propositions are suggested by this research. Proposition 10: Effective top-level leaders (or teams) develop vision relevant expert knowledge structures through social processes involving experience, contact with key constituencies, and use of information technologies. Proposition 11: Individual-level traits and abilities (intelligence, extroversion, social complexity, business knowledge) moderate the effectiveness of such collective schema construction processes.
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
The studies that we have reviewed in this section demonstrate that leaders have the potential to make a difference in teams and organizations. These studies demonstrated moderate variability in the three underlying assumptions outlined earlier. With regard to the origin of causality, several adopted a middle-ground perspective by acknowledging the dual roles of leaders and their social systems in organizational performance (Day & Lord, 1992; Edmondson, 1999). In contrast, several placed the greatest emphasis on the leader’s role (Armstrong & Sambamurthy, 1999; Denison et al., 1995; Hooijberg et al., 1997). None of the studies in this section addressed either implicitly or explicitly the nature of the leadership category or the perceptual process, because all of them focused on leaders’ sensemaking (and the effects thereof) rather than on perceivers’ sensemaking. The greatest variability among the studies in this section occurred in their dependent variable foci. The work of Edmondson (1999), Day and Lord (1992), and Hooijberg et al. (1997) adopted a middle-ground approach: Each examined some combination of intervening processes and outcomes. Denison et al. (1995) and Armstrong and Sambamurthy (1999) examined leadership outcomes including theorized organizational effectiveness and IT assimilation, respectively. Together, these studies illustrate a wide range of approaches to the study of leaders’ performance and sensemaking. Organizational Change and Transformation In today’s fast-paced and rapidly evolving organizations, both leaders and followers must cope with radical change and transformation. All of the studies that we review in this section deal with this organizational change and transformation, but they do so from a variety of perspectives and use a variety of techniques. All are cognitive in the sense of placing an emphasis on how organizational actors—whether leaders or followers—anticipate, make sense of and react to change. These studies do not fall neatly into separate sub-themes, however, and so we simply review and reflect on them in chronological order. Friedman and Saul (1991) argued that scholars should consider more than just the economic consequences of leadership changes in organizations; they should consider organization members’ reactions as well. They surveyed the human resources (HR) directors of 235 Fortune 500 companies, because they believed that HR directors would be most knowledgeable about affective “disruption” (e.g., “confusion” and “fear”), executive turnover, and organization members’ morale— the three key outcome variables in their study. Overall, HR directors reported fairly low disruption and executive turnover but higher company morale following leadership changes. One surprising finding was that the reason for the leadership change (e.g., retirement vs. board-initiated) had no effect on members’ reactions. An important contribution of Friedman and Saul’s study is their theoretical model of members’ reactions to leadership changes. In this model, they proposed that various contextual factors affect disruption, turnover, and morale via two mediators: successor-induced changes and members’ expectations. They did not examine either proposed mediator in their study; they did, however, observe that members typically expect significant changes to unfold in the aftermath of executive
Outside the Box
succession and that contextual factors could affect members’ expectations. A promising avenue for future research based on this model would be to examine directly the interplay between leaders and followers in the aftermath of executive succession. One particularly intriguing possibility is that members’ expectations, in part, may drive any changes that follow a change in their organization’s top leadership. Hambrick, Geletkanycz, and Fredrickson (1993) observed that many organizations do not adapt effectively to environmental change, in part because of topexecutive’s commitment to the status quo (i.e., current organizational strategies and profiles). This study is unique in its attempt to get inside the heads of leaders rather than to merely impute intervening processes from their demographic characteristics. Hambrick et al. discovered that the more favorable firms’ performances, the stronger CEOs’ beliefs that their successors should possess qualities similar to their own and pursue similar organizational strategies. Interestingly, the empirical relationship between organizational performance and commitment to maintaining similar organizational leadership was not strong (r ⬍ .10) in the Hambrick et al. Study; those CEOs who were most committed to “cloning” themselves could not justify this commitment on empirical grounds. In future research, it would be valuable to explore in greater depth the basis of CEOs’ commitments to the status quo, and to examine the link between these commitments and objective, rather than self-report, measures of firm performance. Nutt and Backoff (1993) developed a set of propositions to address the problem of how to effect radical change or transformation in public organizations. Public organizations, they argued, face particularly tough obstacles to transformation because of their limited latitude, vague goals, and abundance of rules. Three approaches had been proposed for overcoming these obstacles: strategic management, strategic leadership, and empowerment. Nutt and Backoff explored the potential for strategic leadership to effect radical change by reviewing and contrasting various leadership theories. They observed that, although all leadership theories stressed the importance of leaders articulating a vision, they “ignore the steps needed to uncover a vision, leaving the origins of the vision on which strategy is built unclear” (p. 334). Based on this observation, they concluded that strategic management and strategic leadership were complementary approaches, with the former developing the transformational vision and the latter implementing it. We would argue, however, that leaders can and do develop—and do not merely implement—visions of radical change and transformational in public organizations. The “fault,” in this instance, may lie not in strategic leadership per se, but in theories of strategic leadership. Future theory and research could address more directly the vision development process. Is a transformational vision something that leaders themselves create, co-create with followers, or have “thrust upon them” by other organizational stakeholders? In a recent study of 36 computer companies, Eisenhardt and Tabrizi (1995) studied product innovation as a means through which organizations adapt. They pitted two competing models against each other. The compression model assumes that product innovation is a “predictable or certain process, one that can be planned out as a series of discrete steps” (p. 88). In contrast, the experiential model assumes that “the key to fast product development is, then, rapidly building intuition and
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
flexible options in order to learn quickly about and shift with uncertain environments” (p. 91). Eisenhardt and Tabrizi found the greatest support for the experiential model that emphasizes the uncertainty associated with product innovation. Most importantly for our purposes, they found that “strong” (i.e., hierarchically central) leaders did in fact more effectively motivate and focus their project teams during the innovation process. This finding illustrates the role of leaders in organizational change and performance. Nonetheless, future research is needed to examine more closely the effect of strong leaders on project team members’ focus on the “big picture,” because the nature of this link remains unclear. Scholars have long assumed that executives are subject to “selective perception” as a result of their functional backgrounds such that they are more likely notice aspects of their organization or environment that are related to their experiences and expertise. Waller, Huber, and Glick (1995) examined the effects of executives’ functional background on their perceptions of change in their organizations’ effectiveness or environments. Interestingly, they discovered that, among 63 top executives (i.e., CEOs or COOs), functional background affected perceptions of changes in their organizations’ effectiveness but not their organizations’ environments. To explain this finding, Waller and her colleagues draw from past research and theory in both behavioral and cognitive psychology. They reasoned that as executives progressed upward in their organizations, they experienced more consistent reinforcement of accurate perceptions of organizational effectiveness than environments. For this reason, executives would have developed more complex and firmly entrenched effectiveness schemas. This study provided important insights into the effects of leaders’ sensemaking on perceptions of organizational change. Future research in this area could examine different techniques for increasing the accuracy of executives’ perceptions of changes in their organizations’ environments, because such changes likely have implications for subsequent organizational effectiveness. Though the studies in this section on organizational transformation and change are eclectic, an underlying theme is reflected in the following proposition. Proposition 12: Organizational transformation and change processes involve a system in which both schema-guided leader actions and follower reactions (or expectations) determine outcomes. The studies in this section examined the roles of leaders and, in two cases, followers, in effecting organizational transformation and change. Hambrick et al. (1993) viewed leaders as causal origins. Other scholars adopted either a middle ground or systems perspective with regard to causality. It is more difficult to uncover scholars’ assumptions regarding the nature of the leadership category and the perceptual process. In Friedman and Saul’s (1991) study of members’ reactions to CEO succession, some evidence suggests that members’ expectations of significant change following a leadership change are based in fairly static representations of new leaders as change agents. Hambrick et al.’s CEOs showed considerable stability in their ideal leadership schemas when expressing beliefs that their successors should possess qualities similar to their own. Waller et al. (1995) explicitly addressed the schema concept in their study by demonstrating that leaders’ schemas regarding
Outside the Box
cues to their organizations’ effectiveness (but not environments) remain stable over the course of their careers. The final assumption involves the dependent variable foci of these studies. Hambrick, Waller, and their colleagues examined leaders’ schemas and discussed the implications of these schemas for key organizational outcomes. In contrast, Friedman and Saul (1991) and Eisenhardt and Tabrizi (1995) focused solely on organizational outcomes. Though variable in their assumptions, these studies all demonstrate the key role of leaders and followers in effecting organizational transformation and change. Integration of Individual/Dyadic and Collective Cognitions Collective cognitions must have an individual level analog, and we suggest here that a useful analog is a connectionist level network. Thus, we believe that connectionist networks can help bridge the individual/dyadic and collective aspects of leadership cognitions. Consider the work of Hooijberg on the Leaderplex Model (Hooijberg & Schneider, in press; Hooijberg et al., 1997) in comparison to work on leadership perceptions. As discussed previously, Hooijberg’s model argues that cognitive, behavioral, and social complexity allows leaders to interact effectively with a variety of role sets or stakeholders. Through this process, top-level leaders develop a comprehensive vision of an organization and future organizational environments, which then can be applied in specific strategic decision making situations. Although this model is quite different from those in most studies of leadership perception, the underlying processes may be quite similar across these models. In all of them, abstract interpretive structures are created through experience with diverse examples—whether they represent varying role set in Hooijberg’s theory or different exemplars in leadership prototype research. Once learned, these abstract and general structures are then applied in a contextually sensitive manner to create either a vision of the future organization or a model of desired leadership qualities. These structures then guide actions in the relevant substantive domain, strategy and social perceptions, respectively. Further, these structures have the potential to be altered as feedback develops on the effectiveness of strategic decisions or the behavior of a particular leader. This reasoning is reflected in the following proposition. Proposition 13: Collective and individual level processes are linked by two processes: (a) collective experiences that help individuals develop more comprehensive knowledge structures and ILTs, and (b) social stimuli that serve as critical inputs which activate and contextually constrain the retrieval (reconstruction) or knowledge structures and ILTs in use. With this perspective in mind, it is useful to revisit the underlying assumptions in cognitive research discussed at the outset of this review. First, while causality is often thought to originate in a leader, this is likely to be an oversimplification. Leadership processes always involve an interaction of leader, subordinate, and contextual qualities that are mediated by leader cognitions. Even in the shortrun, leadership schema are shaped by processes external to a leader as they are
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
reconstructed in a contextually appropriate manner; in the long-run, a leader’s underlying knowledge is created and cued by social processes. Certainly, leader qualities are important in this process as they influence how effectively leaders can interact with diverse role sets, the type of social networks that leaders find most comfortable, and how well they learn from experience (Lord, in press). Yet leaders are only part of larger social/organizational systems. Second, we have emphasized that schema generation processes are dynamic. Neither cognitive structures, such as scripts or schema, nor specific memories are retrieved intact from long-term memory. Instead, they are recreated in a contextually sensitive manner and are refined on a moment-to-moment basis as new information is assimilated. Finally we suggest that the measurement focus should be on the operation of such intervening processes as well as on final outcomes of leadership. For example, research on ILTs might profit from more interest in how they are formed and used, rather than merely specifying ILT content. Areas for Future Research The above discussion suggests several fruitful areas for future research, such as dynamic leadership processes, the interplay between individual cognitions and social/contextual factors, and the relation of intervening and outcome processes. One area that has only been briefly mentioned is the interplay between affective and cognitive processes. Prior reviews (Hall & Lord, 1995; Walsh, 1995) indicated a need to investigate the effects of leadership on both affect and cognitions, yet we found little literature on this topic. We stress again the need for such research. The process of extracting, regenerating, and refining schema all occur within an affective context comprised of a leader’s own mood and emotions as well as the affect of other individuals. As noted by Srull and Wyer (1989), we always encode affect in social processes, so affective constraints on connectionist networks should be ubiquitous. Affective process continually interact with cognitive processes, and affective reactions are likely to be essential for understanding how leadership schema develop, change, and are used. Further, affective reactions may reflect individual interpretations of social events, or they may be collectively constructed as Bartel and Saavedra (2000) demonstrated recently. Other promising areas involve the effects of self-concepts and self-evaluative processes. Several researchers (Day, in press; Lord, Brown, & Freiberg, 1999; Gardner & Avolio, 1998; Lord, Brown, & Freiberg; Shamir et al., 1993) stress that leaders affect the identities of followers, thereby influencing their self-regulatory processes. Brown’s (2000) research provides an innovative demonstration of such processes, while Lord and Brown (in press) show how such processes can be constrained by organizational and cultural values. Several studies also suggest that self-efficacy can limit leadership processes, whether arising from internal processes such as leader’s dispositional optimism (Chemers, Watson, & May, 2000) or external processes such as the demographic heterogeneity of a leader’s group (Mayo, Pastor, & Meindl, 1996). We suspect that such self-regulatory effects operate though the type of automatic, network based process we have described and involve both affective and cognitive components.
Outside the Box
In the area of collective cognition two paths appear particularly promising for future research. First, research in leadership and collective cognition seems especially poised at this time to focus on the cognitive processes that mediate the kinds of interesting and important relationships that scholars have demonstrated in the past decade. Several have offered tentative explanations for intervening processes (e.g., Emrich et al., 2000; Friedman & Saul, 1991; Gardner & Avolio, 1998), but few have examined these explanations empirically. A second path leads toward an examination of the “linkages among the nodes” in the social network that comprises collective cognitions. Almost without exception, leadership scholars have examined collective level issues of team learning, organizational adaptation, and so forth, by aggregating individuals’ responses to the team or organizational level of analysis, respectively. As we pointed out earlier, it is possible to capture individual cognition by gathering individual-level data. It is impossible to capture collective cognition in this way, because the social interactions and shared sensemaking that is the “stuff” of collective cognition does not reside in the minds of solitary individuals. Moreover, social stimuli may be critical constraints on the collective regeneration of collective schema as Walsh (1995) suggested. Scholars who study groups have developed a number of process-focused techniques that could prove valuable to the study of leadership and collective cognition. One such technique was developed recently by Bartel and Saavedra (2000) in their study of work group moods. Briefly, they created an inventory of facial, vocal, and postural indicators of group mood and then trained observers to note the frequency of these indicators in groups’ interactions. These observers used this inventory first to independently and then to collectively assess work groups’ moods. In their discussion, Bartel and Saavedra noted the potential to use their approach to trace “how both influential . . . and behaviorally expressive members build on and shape each other’s moods” (p. 223). This same issue could be explored in future cognition-oriented studies of leadership. As noted earlier, although leadership clearly contains a strong affective component, much remains unknown about leaders’ effects on the moods and emotions of followers (Hall & Lord, 1995; Walsh, 1995). For example, affect is believed to play a critical role in followers’ responses to charismatic leaders. Future research could use Bartel and Saavedra’s inventory and method to examine “emotional comparison” and “emotional contagion” in followers’ responses to the kinds of communication techniques that Shamir, Fiol, Emrich, and their colleagues have theorized to play pivotal roles in the construction of charisma.
CONCLUSION In this review we have developed several key propositions that can serve as guideposts to future cognitively oriented leadership research. These propositions are summarized in Table 1. Collectively, they are the beginnings of an information processing-based theory of organizational leadership that emerged from the collective efforts of leadership researchers. We have also noted several trends that indicate a maturation of the field, in that
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
earlier calls by Walsh (1995) and Hall and Lord (1995) for information processing research have been reflected in the extant literature. For example, Walsh lamented scholars’ overemphasis on documenting the contents of knowledge structures in the broadly based organizational information processing literature. The leadership research reviewed in this article continues to reflect such tendencies, particularly in the cross-cultural area; much of the leadership research, however, does go beyond such surface level issues. Considerable research emphasizing metacognitions attempt to explain how schema are used to generate leadership behavior or perceptions, and theoretical models based on connectionist networks explain how such schema develop and change—two issues that Walsh (1995) identified as needing attention only five years earlier. Research and theory on the development and use of collective cognitions, particularly in constructing a leader’s vision, is another area where researchers have responded to prior calls for needed research. In addition, the interplay between individual-level social abilities and the development of collective cognitions, as suggested by Hooijberg’s work, may be a key approach to developing cross-level information processing research that can build on the theoretical perspectives developed by Hall and Lord (1995) and Walsh (1995). Such developments indicate to us that information-processing studies of leadership are contributing to a deeper understanding of key processes, and they make us optimistic regarding future leadership research. More than two decades have passed since Eden and Leviatan (1975) concluded that “leadership factors are in the mind of the respondent. It remains to be established whether or not they are more than that” (p. 741). The cognitive revolution that was sparked, in part, by these words shows no signs of abating. Indeed, cognition is accorded a central place in a surprising (to us, at least) proportion of current models of leadership. These models and their associated empirical findings have established that leadership factors reside not only in the minds of followers, but in the minds of leaders, in leaders’ behavior and attitudes, and in the social context in which leaders and their followers interact.
REFERENCES Armstrong, C. P., & Sambamurthy, V. (1999). Information technology assimilation in firms: The influence of senior leadership and IT infrastructures. Information Systems Research, 10, 304–327. Awamleh, R., & Gardner, W. L. (1999). Perceptions of leader charisma and effectiveness: The effects of vision content, delivery and organizational performance. Leadership Quarterly, 10, 345–373. Bartel, C. A., & Saavedra, R. (2000). The collective construction of work group moods. Administrative Science Quarterly, 45, 197–231. Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press. Brewer, M. B., & Gardner, W. L. (1996). Who is this “we” question? Levels of collective identity and self-representation. Journal of Personality and Social Psychology, 71, 83–93. Brown, D. J. (2000). The implied presence of a supervisor and its effect on a subordinate’s self-concept. Unpublished doctoral dissertation, University of Akron.
Outside the Box
Chemers, M. M., Watson, C. B., & May, S. T. (2000). Dispositional affect and leadership effectiveness: A comparison of self-esteem, optimism, and efficacy. Personality and Social Psychology Bulletin, 26, 267–277. Chong, L. M. A., & Thomas, D. C. (1997). Leadership perceptions in cross-cultural context: Pakeha and Pacific Islanders in New Zealand. Leadership Quarterly, 8, 275–293. Conger, J. A., & Kanungo, R. N. (1988). Charismatic leadership: The elusive factor in organizational effectiveness. San Francisco: Jossey-Bass. Day, D. V. (in press). Assessment of leadership outcomes: The interrelated concerns of organization, image, and reputation. In S. J. Zaccaro & R. J. Klimoski (Eds.), The nature of organizational leadership: Understanding the performance imperatives confronting today’s leaders. San Francisco: Jossey-Bass. Day, D. V., & Lord, R. G. (1992.) Expertise and problem categorization: The role of expert processing in organizational sense-making. Journal of Management Studies, 29, 35–47. Den Hartog, D. N., House, R. J., Hanges, P. J., Ruiz-Quintanilla, A., Dorfman, P. W., & Associates (1999). Culture specific and cross-culturally generalizable implicit leadership theories: Are the attributes of charismatic/transformational leadership universally endorsed? Leadership Quarterly, 10, 219–258. Denison, D. R., Hooijberg, R., & Quinn, R. E. (1995). Paradox and performance: Toward a theory of behavioral complexity in managerial leadership. Organization Science, 6, 524–540. Eden, D. (1992). Leadership and expectations: Pygmalion effects and other self-fulfilling prophecies in organizations. Leadership Quarterly, 3, 271–305. Eden, D., Geller, D., Gewirtz, A., Gordon-Terner, R., Inbar, I., Liberman, M., et al. (2000). Implanting pygmalion leadership style through workshop training: Seven field experiments. Leadership Quarterly, 11, 171–210. Eden, D., & Levitan, U. (1975). Implicit leadership theory as a determinant of the factor structure underlying supervisory behavior scales. Journal of Applied Psychology, 60, 736–741. Edmondson, A. (1999). Psychological team safety and learning behavior in work teams. Administrative Science Quarterly, 44, 350–383. Eisenhardt, K. M., & Tabrizi, B. N. (1995). Accelerating adaptive processes: Product innovation in the global computer industry. Administrative Science Quarterly, 40, 84–110. Engle, E. M., & Lord, R. G. (1997). Implicit theories, self-schemas, and leader-member exchange. Academy of Management Journal, 40, 988–1010. Emrich, C. G. (1999). Context effects in leadership perception. Personality and Social Psychology Bulletin, 25, 991–1006. Emrich, C. G., Brower, H. H., Feldman, J. M., & Garland, H. (2000). Images in words: Presidential rhetoric, charisma, and greatness. Manuscript submitted for publication. Paper presented at the meeting of the Academy of Management, Toronto, Ontario, Canada. Fiol, C. M., Harris, D., & House, R. J. (1999). Charismatic leadership: Strategies for effecting social change. Leadership Quarterly, 10, 449–482. Friedman, S. D., & Saul, K. (1991). A leader’s wake: Organization member reactions to CEO succession. Journal of Management, 17, 619–642. Gardner, W. L., & Avolio, B. (1998). The charismatic relationship: A dramaturgical perspective. Academy of Management Review, 23, 32–58. Gerstner, C. R., & Day, D. V. (1994). Cross-cultural comparison of leadership prototypes. Leadership Quarterly, 2, 121–134. Hains, S. C., Hogg, M. A., & Duck, J. M. (1997). Self-categorization and leadership: Effects
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
of group prototypicality and leader stereotypicality. Personality and Social Psychology Bulletin, 23, 1087–1099. Hall, R. J., & Lord, R. G. (1995). Multi-level information-processing explanations of followers’ leadership perceptions. Leadership Quarterly, 6, 265–287. Hall, R. J., Workman, J. W., & Marchioro, C. A. (1998). Sex, task, and behavioral flexibility on leadership perceptions. Organizational Behavior and Human Decision Processes, 74, 1–32. Hambrick, D. C., Geletkanycz, M. S., & Fredrickson, J. W. (1993). Top executive commitment to the status quo: Some tests of its determinants. Strategic Management Journal, 14, 401–418. Hanges, P. J., Lord, R. G., & Dickson, M. W. (2000). An information-processing perspective on leadership and culture: A case for connectionist architecture. Applied Psychology: An International Review, 49, 133–161. Higgins, E. T. (1998). Promotion and prevention: Regulatory focus as a motivational principle. In M. P. Zanna (Ed.), Advances in experimental social psychology (pp. 1–46). San Diego, CA: Academic Press. 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. Hogg, M. A., & Tindale, R. S. (2000). Blackwell handbook of social psychology, Vol. 3: Group processes. Oxford: Blackwell. Hooijberg, R., Hunt, J. G., & Dodge, G. E. (1997). Leadership complexity and development of the Leaderplex Model. Journal of Management, 23, 375–408. Hooijberg, R., & Schneider, M. (in press). Behavioral complexity and social intelligence: How executive leaders use their stakeholders to form a systems perspective. In S. J. Zaccaro & R. J. Klimoski (Eds.), The nature of organizational leadership: Understanding the performance imperatives confronting today’s leaders. San Francisco: Jossey-Bass. House, R. J., & Shamir, B. (1993). Toward the integration of transformational, charismatic, and visionary theories of leadership. In M. Chemers & R. Ayman (Eds.), Leadership: Perspectives and research directions (pp. 81–107). New York: Academic Press. Hunt, J. G., Boal, K. B., & Dodge, G. E. (1999). The effects of visionary and crisis-responsive charisma on followers: An experimental examination of two kinds of charismatic leadership. Leadership Quarterly, 10, 423–448. Keller, R. T. (1992). Transformational leadership and the performance of research and development project groups. Journal of Management, 18, 489–501. Keller, T. (1999). Images of the familiar: Individual differences and implicit leadership theories. Leadership Quarterly, 10, 589–607. Kenney, R. A., Schwartz-Kenney, B. M., & Blascovich, J. (1996). Implicit leadership theories: Defining leaders described as worthy of influence. Personality and Social Psychology Bulletin, 22, 1128–1143. Konst, D., Vonk, R., & Van Der Vlist, R. (1999). Inferences about causes and consequences of behavior of leaders and subordinates. Journal of Organizational Behavior, 20, 261–271. Lord, R. G. (in press). Developing a systems theory of executive leadership. In S. J. Zaccaro & R. J. Klimoski (Eds.), The nature of organizational leadership: Understanding the performance imperatives confronting today’s leaders. San Francisco: Jossey-Bass. Lord, R. G., & Brown, D. J. (2000). Leadership, values, and subordinate self-concepts. Manuscript submitted for publication. Lord, R. G., Brown, D. J., & Freiberg, S. J. (1999). Understanding the dynamics of leadership: The role of follower self-concepts in the leader/follower relationship. Organizational Behavior and Human Decision Processes, 78, 167–203.
Outside the Box
Lord, R. G., Brown, D. J., & Harvey, J. L. (in press). System constraints on leadership perceptions, behavior, and influence: An example of connectionist level processes. In M. A. Hogg & R. S. Tindale (Eds.), Blackwell handbook of social psychology, Vol. 3: Group processes. Oxford: Blackwell. Lord, R. G., Foti, R. J., & De Vader, C. L. (1984). A test of leadership categorization theory: Internal structure, information processing, and leadership perceptions. Organizational Behavior and Human Performance, 34, 343–378. Lord, R. G., & Maher, K. J. (1991). Leadership and information processing: Linking perceptions and performance. Boston: Routledge. Malloy, T. E., & Janowski, C. L. (1992). Perceptions and metaperceptions of leadership: Components, accuracy, and dispositional correlates. Personality and Social Psychology Bulletin, 18, 700–708. Martindale, C. (1975). Romantic progression: The psychology of literary history. Washington, D C: Hemisphere. Martindale, C. (1981). Cognition and consciousness. Homewood, IL: Dorsey. Martindale, C. (1986). Aesthetic evolution. Poetics, 15, 439–473. Mayo, M., Pastor, J. C., & Meindl, J. R. (1996). The effects of group heterogeneity on the self-perceived efficacy of group leaders. Leadership Quarterly, 7, 265–284. Meindl, J. R. (1995). The romance of leadership as a follower-centric theory: A social constructionist approach. Leadership Quarterly, 6, 329–341. Meindl, J. R., Ehrlich, S. B., & Dukerich, J. M. (1985). The romance of leadership. Administrative Science Quarterly, 30, 78–102. Murphy, M. R., & Jones, A. P. (1993). The influence of performance cues and observational focus on performance rating accuracy. Journal of Applied Social Psychology, 23, 1523– 1545. Nutt, P. C., & Backoff, R. W. (1993). Transforming public organizations with strategic management and strategic leadership. Journal of Management, 19, 299–347. Offermann, L. R., Kennedy, John K., Jr., & Wirtz, P. W. (1994). Implicit leadership theories: Content, structure, and generalizability. Leadership Quarterly, 5, 43–58. Offermann, L. R., Schroyer, C. J., & Green, S. K. (1998). Leader attributions for subordinate performance: Consequences for subsequent leader interactive behaviors and ratings. Journal of Applied Social Psychology, 28, 1125–1139. Peterson, R. S. (1999). Can you have too much of a good thing? The limits of voice for improving satisfaction with leaders. Personality and Social Psychology Bulletin, 25, 313–324. Phillips, J. S., & Lord, R. G. (1981). Causal attributions and perceptions of leadership. Organizational Behavior and Human Performance, 28, 143–163. Pillai, R. (1996). Crisis and the emergence of charismatic leadership in groups: An experimental investigation. Journal of Applied Social Psychology, 26, 543–562. Pillai, R., & Meindl, J. R. (1998). Context and charisma: A “meso” level examination of the relationship or organic structure, collectivism, and crisis to charismatic leadership. Journal of Management, 24, 643–671. Podsakoff, P. M., MacKenie, S. B., Moorman, R. H., & Fetter, R. (1990). Transformational leader behaviors and their effects on followers’ trust in leader, satisfaction, and organizational citizenship behaviors. Leadership Quarterly, 1, 107–142. Prussia, G. E., Anderson, J. S., & Manz, C. C. (1998). Self-leadership and performance outcomes: The mediating influence of self-efficacy. Journal of Organizational Behavior, 19, 523–538. Shamir, B., House, R. J., & Arthur, M. B. (1993). The motivational effects of charismatic leadership: A self-concept based theory. Organizational Science, 4, 577–594.
LEADERSHIP QUARTERLY Vol. 11 No. 4 2000
Simonton, D. K. (1994). Greatness: Who makes history and why. New York: Guilford Press. Smith, J. A., & Foti, R. J. (1998). A pattern approach to the study of leader emergence. Leadership Quarterly, 9, 147–160. Smith, W. G. (1997). Implicit gender and leader cognitive categorization: A test of the catastrophe model. Unpublished doctoral dissertation, University of Akron. Sosik, J. J., & Dworakivsky, A. C. (1998). Self-concept based aspects of the charismatic leader: More than meets the eye. Leadership Quarterly, 9, 503–526. Sosik, J. J., & Megerian, L. E. (1999). Understanding leader emotional intelligence and performance: The role of self-other agreement on transformational leadership perceptions. Group & Organizational Management, 24, 367–390. Srull, T. K., & Wyer, R. S., Jr. (1989). Person memory and judgment. Psychological Review, 96, 58–83. Sutton, J. L. (1998). Effects of task content, task type, and leader gender on perceptions of gender differences in leadership effectiveness. Unpublished doctoral dissertation, Texas Tech University, Lubbock. Van Vugt, M., & De Cremer, D. (1999). Leadership in social dilemmas: The effects of group identification of collective actions to provide public goods. Journal of Personality and Social Psychology, 76, 587–599. Waller, M. J., Huber, G. P., & Glick, W. H. (1995). Functional background as a determinant of executives’ selective perception. Academy of Management Journal, 38, 943–974. Walsh, J. P. (1995). Managerial and organizational cognition: Notes from a trip down memory lane. Organization Science, 6, 280–321. Wofford, J. C., & Goodwin, V. L. (1994). A cognitive interpretation of transactional and transformational leadership theories. Leadership Quarterly, 5, 161–186. Wofford, J. C., Goodwin, V. L., & Whittington, J. L. (1998). A field study of a cognitive approach to understanding transformational and transactional leadership. Leadership Quarterly, 9, 55–84. Wofford, J. C., Joplin, J. R. W., & Comforth, B. (1996). Use of simultaneous verbal protocols in analysis of group leaders’ cognitions. Psychological Reports, 79, 847–858. Yagil, D. (1998). Charismatic leadership and organizational hierarchy: Attributions of charisma to close and distant leaders. Leadership Quarterly, 9, 161–176. Yorges, S. L., Weiss, H. M., & Strickland, O. J. (1999). The effect of leader outcomes on influence, attributions, and perceptions of charisma. Journal of Applied Psychology, 84, 428–436. Zaccaro, S. J., Foti, R. J., & Kenney, D. A. (1991). Self-monitoring and trait-based variance in leadership: An investigation of leader flexibility across multiple group situations. Journal of Applied Psychology, 76, 308–315. Zaccaro, S. J., Gilbert, J. A., Thor, K. K., & Mumford, M. D. (1991). Leadership and social intelligence: Linking social perspectiveness and behavioral flexibility to leadership effectiveness. Leadership Quarterly, 2, 317–342. Zaccaro, S. J., & Klimoski, R. J. (in press). The nature of organizational leadership: Understanding the performance imperatives confronting today’s leaders. San Francisco: JosseyBass.
Outside the Box
APPENDIX A Propositions summarizing a social information processing approach to leadership
Individual Level Proposition 1. Causality for leadership behavior and perceptions lies in a complex social system involving the leader, follower, task, and context. Proposition 2. The effects of social systems on: (a) leader behavior are mediated by leader metacognitive processes, whereas (b) leadership perceptions are mediated by follower metacognitive processes. Proposition 3. Crisis situations tend to elicit perceptions of charisma from perceivers as long as the leader being rated is not responsible for the crisis. When the leader is responsible, crisis is negatively related to perceived charisma. Proposition 4. The most immediate determinant of leadership perceptions (leader behavior) is the activation of perceivers’ (leaders’) ILTs. Proposition 5. The contextual sensitivity of perceivers’ (or leaders’) metacognitive processes operates through their differential activation of appropriate ILTs. Proposition 6. The differential activation of ILTs in various contexts is produced by dynamic, connectionist network-based processes that create new, contextually tuned ILTs each time they are used. Collective Level Proposition 7. Charismatic leaders use a variety of communication based techniques (e.g., metaphors, collective pronouns, strategic ambiguity, and image-based rhetoric) to link their own visions or values to followers’ values and self-structures. Proposition 8. Charismatic leadership is dependent on facilitating social systems and structures. Proposition 9. Effective leadership involves sensemaking that links important outcomes such as performance to social (e.g., leadership) and psychological (e.g., safety) aspects of contexts, thereby facilitating organizational learning. Proposition 10. Effective top-level leaders (or teams) develop vision relevant expert knowledge structures through social processes involving experience, contact with key constituencies, and use of information technologies. Proposition 11. Individual level traits and abilities (intelligence, extroversion, social complexity, business knowledge) moderate the effectiveness of such collective schema construction processes. Proposition 12. Organizational change processes involve a system in which both schema-guided leader actions and follower reactions or expectations determine outcomes. Proposition 13. Collective and individual level processes are linked by two processes: (a) collective experiences that help to develop more comprehensive knowledge structures and ILTs in individuals, and (b) social stimuli that serve as critical inputs that activate and contextually constrain the retrieval (reconstruction) of knowledge structures and ILTs in use.