Theories of leadership attempt to explain its emergence or its nature and its consequences. Theories of management focus on its governance and consequences. Models show the interplay among the variables that are perceived to be involved; they are replicas or reconstructions of reality. Both theories and models can be useful in defining research problems for the social and political scientist and in improving prediction and control in the development and application of leadership. In this chapter we will introduce briefly the most prominent theories of leadership, Many will be discussed in more detail in later chapters as appropriate.
Until the late 1940s, most theories of leadership focused on the personal traits of leaders. According to these theories, leadership depended on leaders’ abilities and personality characteristics. Then, up to the late 1960s, personal styles of leadership rose in prominence. From the late 1960s to the early 1980s, leadership studies became contingent on a mix of leaders’ and followers’ traits and situations. Leadership theories of inspiration and transformation emerged in the early 1980s and became prominent in the 1990s and at the turn of the twenty-first century. Theories of management paralleled the rise of civilization, and management practice was exemplified in the rules of planning, organizing, and controlling. Theories of management became more sophisticated in the twentieth century and were strongly influenced by the advent of computerization, information science, and globalization.
Before 1970, there were many complaints about a lack of theory to provide hypotheses for empirical research to test. Since then, diverse theories and models have been abundant as a source of hypotheses. However, relatively few of these models and theories have dominated the research community, and many have been restatements of the obvious. On the other hand, progress has been made when the models and theories have been built on astute observations and on assumptions that are the result of insightful observations. Good theories are disciplined imagination (Weick, 1995). They are internally consistent and consistent with a more general body of propositions from the social sciences. For instance, theories about reinforcement leadership have been built from what is generally known about reinforcement theories. Transformational leadership theory has similarly benefited from motivational theory.
Diagrams, speculations, and hypotheses are not theories; nor are cited references, data, and variables that are used in place of theory (Sutton & Staw, 1995). Rather, good theories usually start from one idea or a small set of ideas. They make possible a logically detailed case characterized by simplicity and interconnectedness. Predictions are presented with underlying causal logic: “The process includes abstracting, generalizing, relating, selecting, explaining, synthesizing, and idealizing.” Good theories explain how and why (Sutton & Staw, 1995, p. 389). DiMaggio (1995) adds that good theories provide “categories and … assumptions that … (clear) away conventional notions to make room for artful and exciting insights” (p. 391). Good theories emerge from various experiences. Observations and inferences from frequent contact with leaders and followers help. Connections are identified between observations and concepts that heretofore were not seen to be connected. There is a convergence of several interests and activities at the same time. Intuition and feelings supplement logical analysis. There is a desire to explain, understand, and find meaning in the real world. Ideas are confronted and confirmed or disconfirmed. Conventional wisdom may be revised or reversed. Research is restated in alternative ways. Established value judgments are challenged. Above all, to be good, theories need to be grounded in assumptions that fit the facts. Theorizing and modeling are now commonplace in providing the rationale to justify and test empirical hypotheses. These will be discussed more fully as they become relevant to the topics of later chapters.
Nothing is supposed to be as practical as a good theory (Lewin, 1947), but nothing seems more impractical than a bad one (Bass, 1974). A theory is supposed to be a way of trying to explain the facts. Unfortunately, theories about leadership sometimes obscure the facts. Much effort then has to be expended in coping with the obscurity. Poor research derives from poor theory. Poor research is often expedient, quick, and convenient. Klimoski (2005) notes that bad theory is dangerous. And according to Ghoshal (2005), bad management theories are destroying good management practices. Examples are agency theory (Jensen & Meckling, 1976) and transaction costs theory (Williamson, 1975). Agency theory assumes that managers are bent on maximizing their own interests at the expense of all the other constituencies of the organization. Transaction costs theory assumes that managers are in deadly competition with each other. Monitoring and tight control of their opportunistic behavior are needed. They cannot be trusted. To gain a comparative advantage requires a company to compete not only with other companies but also with its own suppliers, customers, employees, and regulators (Porter, 1980). These “ideologically inspired” amoral theories are being preached in many business schools and management development programs. They free future organizational leaders from any sense of moral responsibility for their decisions. and encourage rather than inhibit their opportunistic behavior (Ghoshal & Moran, 1996). Scientism rather than science makes these bad theories appealing, for their derived propositions can be reproduced with mathematical models. They would be good theories if their assumptions about human behavior were correct. In fact, governance based on the assumption that managers are self-aggrandizing and cannot be trusted results in less trustworthy managers (Osterloh & Fry, 2003). Deductions from these economic theories that control of managers requires more independent boards of directors, and that corporate performance is improved if the positions of board chairman and CEO are separated, are not supported empirically (Dalton, Daily, Ellstrand, et al., 1998). But the same could be said about cognitive, behavioral, and social theories based on faulty assumptions about human and social behavior. There was a long-held theory that efficiency increased if work was divided into parts so each worker could specialize; but the effects on the workers’ motivation were ignored.
Ferraro, Pfeffer, and Sutton (2005) point to a further complication and effect of bad theory and practice. They cite evidence that corporate managers are less interested in first evaluating whether a new fad increases effectiveness and profitability, and more interested in showing that they have adopted the latest popular theory or practice without really finding out its effects. They treat the effects as self-fulfilling prophesies. And political leaders have done the same thing, according to Keynes (1936), with damaging long-term effects on the economy and society.
With the growth of cognitive psychology we have seen cognitive theories and models of leadership added to the earlier behavioral and social theories. Zaccaro (1996) has combined cognitive, behavioral, and social approaches in a “Leaderplex Model” of executive leadership (see also Hooijberg, Hunt, & Dodge, 1997). Theories of leadership with extensive empirical support of effects on followers’ effectiveness and satisfaction can be broadly classified as instrumental, inspirational, and informal. Instrumental theories focus on the leader’s orientation to the task or to the person, on leaders’ direction or followers’ participation, on leaders’ initiative or consideration of their followers’, and on leaders’ promises and rewards or threats and disciplinary action. Inspirational theories of leadership include charismatic, transformational, and visionary theories. These focus on emotional and ideological appeals, displaying exemplary behavior, confidence, symbolism, and concentrating on intrinsic motivation. Informal leadership theories deal with the emergence and service of effective leaders who lack formal positions and authority (House, 1995).
Implicit Theories of Leadership (ILTs). Implicit theories of leadership (ILTs) are the concept of “leader” that different people have in mind. These implicit leadership theories about leaders and followers affect the relations between them. An ILT may be about instrumental, inspirational, or informal leadership. ILTs are beliefs about how leaders behave, in general, and what is expected of them (Eden & Levitan, 1975). Such theories are naive concepts of leadership, revealed (for instance) when people are asked to list a number of traits that come to mind when they think of the term “leader.” Although respondents may be affected by individual, gender, social, organizational, and cultural differences, ordinarily their similarities outweigh their differences. Therefore, it is possible to determine a generalized image or “prototype” of the concept of a leader. We can then compare prototypes that may emerge from markedly different samples, such as Americans and Chinese. In the same way, different prototypes can be discriminated among the concepts of 11 types of leaders: business leaders, sports leaders, etc. (Lord, Foti, DeVader, 1984). There is little difference between the prototype of a “leader” and that of an “effective leader.” However, the term “supervisor” receives somewhat less favorable trait ratings from college students than the terms “leader” and “effective” leader (Offerman, Kennedy, Wirtz, 1999).
For college students, T. Keller (1999) found ILTs for “leader” comparable to those of Offerman et al. (1999). In addition, she obtained modest correlations of selected ascribed traits of “leader,” self-rated personality using the NEO five-factor personality inventory, and perceived parental traits. For instance, the extent to which respondents perceived sensitivity as a characteristic of a “leader” correlated significantly (.20) with the respondents’ self-rated “openness.” Dedication as a trait of a leader correlated .23 with their self-rated “conscientiousness” and .31 with their perceived “paternal dedication.” A number of other such correlations were found, suggesting to Keller that individuals perceive a leader as mirroring traits of themselves and their fathers.
Another application of implicit leadership theory was a study to determine 378 undergraduates’ idea of characteristics required for a new leader to be accepted by a group. The list of 16 expectations formed four abstract categories: (1) taking charge, (2) learning the group’s goals, (3) being nervous, and (4) being a nice person (Kenny, Blascovich, & Shaver, 1994).
Increased interest in implicit leadership has paralleled a cognitive revolution in psychology and developments in information processing. Lord and Maher’s (1991) book, Leadership and Information Processing: Linking Perceptions and Performance, is illustrative.
In organizations, these are—in contrast to implicit theories—concepts espoused in formal documents, memorandums, and executives’ speeches. These state intentions, goals, objectives, and desired relationships of leaders and managers with their various constituencies. Or they may be theories in practice, theories actually guiding the behavior of members of the organization that can be observed, expressed, or inferred in interviews with members (Argyris, 1982).
Grounded Theory. Grounded theory is closely allied to implicit leadership theory. If a theory of leadership is to be used for diagnosis, training, and development, it must be grounded in the users’ concepts, assumptions, language, and expressions. The users include emergent leaders and followers in informal organizations as well as managers, administrators, and officials in formal organizations (Glaser & Strauss, 1967; Strauss & Corbin, 1990). Qualitative research is often grounded. But such grounding requires conceptual rigor. There may be a loss of generality and less opportunity to apply standardized measurements. Grounded research generates rather than tests theory (Parry, 1998).
“He was born to lead.” “She instinctively knew how to take charge.” “The coach was a natural leader.” Such comments assume that nature has been more important than nurture in the emergence of a particular leader. But until recently a majority of social scientists considered nurture more important than nature. Nonetheless, leaders are both born and made. The importance of health, physique, and energy in the emergence of leadership has long been recognized. Additionally, with advances in genetics and neuropsychology, as well as speculations about evolution and recognition of the biological differences between the sexes, has come an appreciation of inborn traits and the expression of genes.
As Spinoza said, “Man is a social animal.” Lawrence (1997) has enumerated four built-in evolutionary survival factors in humans that affect their organizational life, and therefore leader-follower relations: (1) a need to acquire; (2) a need to bond; (3) a need to learn; and (4) an inborn reflexive mechanism to avoid pain. They involve strong emotions and a strong motivation toward goal-oriented behavior. The acquisitive needs of followers are helped by instrumental and transactional leadership. Social-bonding needs are fulfilled by considerate and charismatic leaders. Learning needs of followers are fulfilled when leaders clarify purposes. Followers’ pain is avoided when coercive, ruthless leadership is avoided.
Great-Man Theories. Jennings (1960) reviewed the “great man” theory. For many commentators, history is shaped by the leadership of great men. Without Moses, according to these theorists, the Jews would have remained in Egypt; without Winston Churchill, the British would have given up in 1940; without Bill Gates, there would have been no firm like Microsoft.
The eighteenth-century rationalists felt that to determine the course of history, luck had to be added to the personal attributes of great men. The Russian Revolution would have taken a different course if Lenin had been hanged by the old regime instead of being exiled. For romantic philosophers such as Friedrich Nietzsche, a sudden decision by a great man (Thomas Jefferson’s decision to purchase Louisiana, for example) could alter the course of history. William James (1880) believed that certain mutations of society were due to great men, who initiated movement and prevented others from leading society in another direction. According to James, the history of the world is the history of great men, who determined what the masses could accomplish. Carlyle’s (1841) essay on heroes reinforced the concept of the leader as a person endowed with unique qualities that captured the imagination of the masses. The hero would contribute somehow, no matter where he was found. History was created by the acts of great leaders. Leaders molded the masses. (Despite the examples of Joan of Arc, Elizabeth I, and Catherine the Great, great women were ignored.) Dowd (1936) maintained that “there is no such thing as leadership by the masses. The individuals in every society possess different degrees of intelligence, energy, and moral force, and in whatever direction the masses may be influenced to go, they are always led by the superior few.” Although one of many civil rights leaders, Martin Luther King, Jr., was considered the great man whose leadership inspired the civil rights movement. The great-man theory of leadership was espoused to show how faltering or threatened organizations could be turned around by business executives like Lee Iacocca, military leaders like Douglas MacArthur, and political figures like Margaret Thatcher.
Influenced by Galton’s (1869) study of the hereditary background of great men, several early theorists attempted to explain leadership on the basis of inheritance. Woods (1913) studied 14 nations over periods of five to ten centuries and found that the conditions of reign approximated the ruler’s capability. The brothers of kings (as a result of natural endowment) also tended to become men of power and influence. Woods concluded that the man makes the nation and shapes it in accordance with his abilities. In line with the eugenics movement, Wiggam (1931) proposed that the survival of the fittest people and intermarriage among them produces an aristocratic class, which differs biologically from the lower classes. Thus an adequate supply of superior leaders depends on a proportionally high birthrate among the abler classes.
The Warrior Model of Leadership. This variant of the great-man theory appeared in several classics: Suntzu’s Art of War (c. 400 b.c.), Aristotle’s Politics (324 B.C.), Machiavelli’s The Prince (1513), Gratian’s The Art of Worldly Wisdom (1643), and Clausewitz’s On War (1833). General George Patton exemplified the warrior model. Wars are won or lost, according to this theory, depending on the leadership of the opposing forces. Thus, Napoleon’s and Julius Caesar’s tactics often spelled the difference between victory and defeat in battle. President Lincoln had to replace the commanding general of his Union army numerous times before he found, in Ulysses Grant a commander who was able and willing to use his forces, superior in size and logistics, to take the initiative and accept the casualties needed to wear down Robert E. Lee’s Confederate forces. Victorious warrior leaders win fame and power. They control flows of information. The means justify the ends for them, even if they must resort to deception, betrayal, violence, and other morally questionable acts. These acts may be delegated to subordinates so that the leaders are held blameless. Planning and preparation assume that the world is a dangerous place (Nice, 1998).
The great-man theories drew attention to the specific qualities of leaders and their identification (Kohs and Irle, 1920). L. L. Bernard (1926), Bingham (1927), Tead (1929), Page (1935), and Kilbourne (1935) all explained leadership in terms of traits of personality and character. Bird (1940) compiled a list of 79 relevant traits from 20 psychologically oriented studies. Similar reviews were done by Smith and Krueger (1933) for educators, and by W. O. Jenkins (1947) for military leaders.
Until the 1940s, much research about leaders and leadership focused on individual traits. Leaders were seen as different from nonleaders in various attributes and tested personality traits. Two questions were usually posed: (1) What traits distinguish leaders from other people?; (2) What is the extent of the differences? The pure trait theory eventually fell into disfavor. Stogdill’s (1948) critique concluded that both person and situation had to be included to explain the emergence of leadership. But as will be seen in Chapter 5, traits are still considered of great importance in the study of leadership.
Charismatic-Transformational Leadership Theory.1 Max Weber (1924/1947) introduced a religious concept—charisma—into the social sciences to describe leaders who are perceived as endowed with extraordinary abilities. Charismatic leaders are highly expressive, articulate, and emotionally appealing. They are self-confident, determined, active, and energetic. Their followers want to identify with them, have complete faith and confidence in them, and hold them in awe. Generally, charismatic leaders have strong positive effects on their followers. House (1977) presented a theory of charismatic leadership that specified the expected behavior of charismatic leaders and their followers, stimulating renewed interest in empirical studies of this subject.
According to Hunt (1999), for the empirical study of leadership, transformational leadership was a new paradigm (Kuhn, 1964)—a change of views, preferred methods, acceptable findings, interpretations of findings, and important areas to study. Transformational leadership was first mentioned by Downton (1973) and first formalized as a theory by Burns (1978). In contrast to transactional leaders, transformational leaders were said to motivate followers to go beyond their own self-interests for the good of the group, organization, or society. Followers’ interests are raised by transformational leaders from concerns for security to concerns for achievement. Followers are encouraged to meet the challenges they face, to excel, and to self-actualize. Bass (1985a) presented models of the factors in transformational and transactional leadership. House and Aditya (1997) viewed transformational leadership as close in meaning to charismatic leadership. Bass (1985a) found that charismatic leadership was the largest factor in transformational leadership but only one of several other empirical factors with which it correlated, including inspirational leadership, intellectual stimulation, and individualized consideration. Transactional leadership encompassed contingent reward, management by exception, and passive or laissez-faire leadership. This was confirmed empirically by Avolio, Bass, and Jung (1999), among others.
By 1960, the dominant paradigm for the study of leadership had evolved from research on the traits and situations that affect leadership to something more dynamic. Leadership was now seen as contingent on traits and situations involving a transaction or exchange between the leader and the led (Hollander, 1986). In this view, leaders promise rewards and benefits to subordinates in exchange for the subordinates’ fulfillment of agreements with the leader. Even the psychoanalysts conceived of followers as complying with the leader to obtain the leader’s love. But Freud (1922) suggested that there was more to the concept of leadership than a mere exchange: the leader embodied ideals with which the follower identified. Barnard (1938) noted that personal loyalty was more powerful than “tangible inducements.” But along with Downton (1973), Burns (1978) presented the new paradigm of the transformational as opposed to the transactional leader. The transformational leader asks followers to transcend their own self-interests for the good of the group, organization, or society; to consider their long-term need for self-development rather than their need of the moment; and to become more aware of what is really important. Hence, followers are converted into leaders. Among 90 transformational leaders who were interviewed, Bennis (1984) found evidence of competence to manage attention and meaning, to articulate visions of what was possible, and to empower the collective effect of their leadership.
Burns’s conceptualization of leadership as either transforming or transactional was modified by Bass (1985a, 1985b), who proposed that transformational leadership augmented the effects of transactional leadership on the efforts, satisfaction, and effectiveness of followers. Many great transformational leaders, including Abraham Lincoln, Franklin Delano Roosevelt, and John F. Kennedy, did not shy away from being transactional as well as transformational. Lincoln, Roosevelt, and Kennedy were able to move the nation as well as play petty politics. Waldman and Bass’s (1985) analysis of surveys of senior military officers and business managers confirmed the fidelity of the model. Tichy and Devanna (1986) described the hybrid nature of transformational leadership. According to them, transformational leadership is not due just to charisma. It is “a behavioral process capable of being learned and managed. It’s a leadership process that is systematic, consisting of purposeful and organized search for changes, systematic analysis, and the capacity to move resources from areas of lesser to greater productivity … [to bring about] a strategic transformation” (p. viii).
On the basis of Kegan’s (1982) theory of the evolving self, Kuhnert and Russell (1989) designed a four-stage model of how the transformational leader develops. Factor-analytic studies done by Bass (1985b) and confirmed by Avolio, Bass, and Jung (1999) have suggested that transformational leadership can be conceptually organized along four correlated dimensions: (1) charismatic leadership, (2) inspirational leadership, (3) intellectual stimulation, and (4) individualized consideration. For House (1977) and Conger (1999) all four components are contained within their concept of charismatic leadership (House & Shamir, 1993). Later chapters detail the antecedents and consequences of transformational and charismatic leadership. The components of transactional leadership are contingent reinforcement, expressed usually as contingent reinforcement and management by exception, which will be examined at length in Chapters 15 and 16.
According to a leader’s subordinates, colleagues, and superiors, transformational-charismatic leadership correlates more highly with the leader’s effectiveness than contingent-reward leadership does. Contingent reward, in turn, correlates more highly with the leader’s effectiveness than reactive management by exception or contingent punishment do. Satisfaction with the leader follows a similar pattern (Bass & Avolio, 1989). Similar results can be obtained when different sources are used to describe the leader and to more objectively evaluate the outcomes of leadership in terms of effectiveness and satisfaction (Lowe, Kroek, & Sivasubramaniam, 1996). In the same way, House and his associates (e.g., House, Spangler, & Woyke, 1991) have reported numerous studies showing positive correlations between charismatic and effective leadership. Conger and Kanungo (1998) along with Sashkin (1988) focused on the process effects of charisma on the followers and on the leader’s need to articulate a vision to be accepted and followed.
Transformational leadership is closer to the “proto-type” of leadership that people have in mind when they describe an ideal leader and is more likely to provide a role model with which subordinates want to identify (Bass & Avolio, 1988). In practice, this means that leaders develop in their subordinates an expectation of high performance rather than merely spend time praising or reprimanding them (Gilbert, 1985). For Bradford and Cohen (1984), the manager must be more than a hero of technical competence and organizing skills. He or she must become a developer of people and a builder of teams.
Servant Leadership. Less well researched but still prominent, servant leadership was formulated by Green-leaf (1977) and was based on his experiences as an executive. According to Greenleaf, ego spurs achievement, but leaders need to curb their own egos, convert their followers into leaders, and become the first among equals. The needs of others must be the leaders’ highest priority. Power has to be shared by empowering followers. Leaders should think of themselves as servants building relationships with their followers that help their followers to grow (Buchen, 1998). Servant leaders must be oriented to the future as stewards of the human and physical resources for which they are responsible. Leaders who are stewards are similar to but not the same as servant leaders. Stewards try to balance the interests of all the different constituents of their organization: shareholders, owners, managers, peers, subordinates, customers, clients, and community (Bass, 1965; Donaldson, 1990). Servant leaders are especially concerned about constituencies with less power or more need for help. Servant leadership also shares much in common with transformational leadership: vision, influence, credibility, and trust (Farling, Gregory & Stone, 1999).
Closely aligned with servant leadership is a model of organizational leadership by Choi and Mai-Dalton (1999): self-sacrificial leaders abandon or postpone their own interests, privileges, or welfare in the way they work with their followers. They give up or postpone rewards to which they are fairly or legitimately entitled. They voluntarily give up or refrain from using their positional or personal power. In one study that used scenarios about a self-sacrificing company president and a non-self-sacrificing president, university students and white-collar employees judged the self-sacrificing president as more charismatic and legitimate. These respondents stated that they would be more likely to reciprocate in self-sacrifice if the president was self-sacrificing.
In direct opposition to trait theorists, situational theorists have argued that leadership is a matter of situational demands; that is, situational factors determine who will emerge as a leader. Particularly in the United States, situationalism was favored over the theory that leaders are born, not made. According to situationalism, the leader is a product of the situation and circumstances. not self-made and not a product of personality, drive, or ability (Stogdill, 1975).
The controversy over which is more important to lead-ership—situation or personality—is an ancient one. Plutarch’s Parallel Lives (c. a.d. 100) described how for each type of leader who emerged in Greece, one emerged under parallel conditions in Rome. Alexander the Great, for instance, had his counterpart in Julius Caesar. The great-man theorists believed that it was all a matter of personality and personality development—that Alexanders and Caesars would surface no matter what conditions surrounded them. The situationalists thought otherwise: for example, they sought to identify conditions that gave rise to the emergence of the “man on the white horse,” the dictator who appears following revolutionary upheaval, chaotic politics, social and economic distress, and a weakening of traditional institutions.
The situationalists advanced the view that the emergence of a great leader is a result of time, place, and circumstance. For Hegel, the great man was an expression of the needs of his times. What the great man did was automatically right, because he provided what was needed. The great man actually could not help what he did; he was directed and controlled by his historical environment. For example, the need for civil peace made it mandatory for Octavian to make himself sole ruler of Rome, form the Roman Principate, and destroy republicanism. If Octavian had not appeared to carry out these changes, someone else would have done the same. Herbert Spencer believed that societies evolved in a uniform, gradual, progressive manner, and no great man could change the course of this development. Engels, Marx, and their successors believed that economic necessity made history. The American Civil War was, then, an inevitable clash caused by the conflicting economic interests of North and South. Economic determinists held that obstacles to expanding production had to be overcome. The greater the obstacles, the greater this need was and the more capable the required leader had to be. But who he turned out to be was irrelevant (Hook, 1943). Mumford (1909) agreed that who emerged as a leader depended on the abilities and skills required at a given time to solve the prevailing social problems. Although these abilities and skills were innate as well as acquired, leadership, as such, stemmed from the organized phases of the social process or the habitual ways in which people adapt to each other. Thus, according to the situationalists, the national condition determined the development and emergence of great military figures. For A. J. Murphy (1941), leadership did not reside in a person but was a function of the occasion. The situation called for certain types of action; the leader did not inject leadership but was the instrumental factor through which a problem was solved. For Person (1928), any particular situation played a large part in determining the leadership qualities and the leader for that situation. Moreover, the leadership required in that situation was a product of a succession of previous leadership situations that molded the leader.
J. Schneider (1937) noted that the number of great military leaders in En gland was proportional to the number of conflicts in which the nation engaged. Spiller (1929) concluded that a broad survey of the field of human progress would show that 95% of the advance was unconnected with great men. Rather, a great man like Martin Luther King, Jr., would appear at a critically important point of a socially valued cause, would devote himself to it, and would profit greatly from the work of many others. Thus, time itself was an important variable for the situationalists. The passage of time changed the situation and the people involved. Thierry, Den Hartog, Koopman, et al. (1997) viewed Dutch people’s preferences for leadership as a matter of Dutch history and trends. Leadership was conceived as a relational process that unfolded over time. Time might be needed for trust to develop before an individual could emerge as a leader. Time made an important addition to models and theories of leadership. Repeated measurements over time are important to analyses of leaders’ performance (Hollen-beck, Ilgen, & Sego, 1994). Bogardus (1918) presented the view that the type of leadership that developed in a group was determined by the nature of the group and the problems it had to solve. Hocking (1924) went even further, suggesting that leadership resided in the group and was granted to leaders only when they put forth a program that the group was willing to follow.
A Rational-Deductive Model. Vroom and Yetton (1974) rationally linked some of the accepted facts about directive and participative decision making as assumptions. From these, they created prescriptions for the leadership style that was most likely to succeed in a given situation. They posed ten questions that leaders should ask themselves in deciding whether to be directive or participative in making decisions with their subordinates and whether to do so primarily with individual subordinates or with the whole group at once. Essentially, the prescriptions were that supervisors ought to be directive when they were confident that they knew what needed to be done and when their subordinates did not have this knowledge. Furthermore, in this situation, the subordinates would accept the decision made by the supervisor. However, if the subordinates had more information than the supervisor, if the subordinates’ acceptance and commitment were of paramount importance, and if the subordinates could be trusted to concern themselves with the organization’s interests, the supervisor should be participative.
Vroom and Jago (1988) created an improved model. Instead of requiring yes-or-no answers to the questions, they provided a five-point scale of possible answers: (1) no, (2) probably no, (3) maybe, (4) probably yes, (5) yes. The older model used only two criteria for the decision rules: acceptance of the decision by followers, and the quality of the decision. The five-point scale made it possible for the leader to prioritize possible decisions, reducing the several feasible alternatives into a single decision.2
Although wars and other crises present opportunities for the acquisition of leadership by persons who would otherwise remain submerged in the daily round of routine activities, various theorists have maintained that the situation is not in itself sufficient to account for leadership. How many crises arise that do not produce a person who is equal to the occasion? A combination of personal and situational elements needs to be considered. James (1880) pointed out that the “great man” needs help. His talents need to fit with the situation. Ulysses Grant, for instance, was a failure in private life before his emergence as the Union’s great military commander, and he failed again afterward as president. His rise to commanding general of the Army of the Potomac was delayed by the many political appointees who came before him and took turns displaying their ineptitude before an exasperated President Lincoln turned to him. Grant’s leadership in the Vicksburg campaign brought victory, despite the orders of his superior, General Halleck, to fall back toward New Orleans. But it was Grant’s persistence, and some help by congressmen, that overcame the inertia of the political appointment system; and the traits of persistence and confidence marked the style with which he hammered out his military victories (Williams, 1952).
The great man theorists and the situational theorists both attempt to explain leadership as an effect of a single set of forces, and both overlook the combined effects of individual and situational factors. In reaction, Westburgh (1931) suggested that the study of leadership must include the affective, intellectual, and action traits of the individual, as well as the specific conditions under which the individual operates. Case (1933) maintained that leadership is produced by a conjunction of three factors: (1) the personality traits of the leader, (2) the nature of the group and its members, and (3) the event confronting the group. J. F. Brown (1936) proposed five field-dynamic laws of leadership: leaders must (1) be identified as members of the group they are attempting to lead; (2) be of high interpersonal potential; (3) adapt themselves to the existing structure of relationships; (4) realize the long-term trends in the structure; (5) recognize that leadership increases in potency at the cost of reduced freedom of leadership.
Hook (1943) noted that there is some restriction in the range of traits that a given situation permits the emergent leader to have. Thus heroic action is decisive only when alternative courses of action are possible. Exiled to Elba, close to France, Napoleon had alternatives; exiled to Saint Helena, in the South Atlantic and more closely guarded, he had none.
Bass (1960) argued that controversy over the great man versus the situation was a pseudo problem. For any given case, some of the variance in what happens is due to the situation, some is due to the individual, and some is due to the combined effects of the individual and the situation. Mao Zedong played a critical role in the Chinese revolution, but without the chaotic state of Chinese affairs under the Kuomintang, his rise to power would not have been possible. Dansereau, Alutto, and Yammarino (1984) developed the multiple-levels analysis, to be described later. It provided a complete statistical formulation of models to examine the interplay of leader, individual follower, group, and organizational situation. Personal-situational theorists argue that theories of leadership cannot be constructed in a vacuum: they must contain elements of the person as well as elements of the situation. Any theory of leadership must take account of the interplay between the situation and the individual. Barnard (1938) and many others (C. A. Gibb, 1947; Jenkins, 1947; Lapiere, 1938; Murphy, 1941) attempted to resolve the controversy over situation versus personality by suggesting that leadership behavior is a less consistent attribute of individuals than such traits as nonsuggestibility, energy, and maturity, which are empirically associated and theoretically linked with overt leadership behavior. Leaders with a strong personal tendency to be consistent will display leadership across many situations.
Stogdill (1948) concluded that leaders’ traits must bear some relevant relationship to the characteristics of the followers. An adequate analysis of leadership needs a study not only of leaders, but of the situation. Stogdill’s position strongly influenced the theories that followed. According to Gerth and Mills (1952, pp. 405–406), “to understand leadership, attention must be paid to (1) the traits and motives of the leader as a man, (2) images that selected publics hold of him and their motives for following him, (3) the features of the role that he plays as a leader, and (4) the institutional context in which he and his followers may be involved.” C. A. Gibb (1954, p. 914) suggested that “leadership is an interactional phenomenon arising when group formation takes place.” A group structure emerges. Each member of the group is assigned a relative position within the group depending on the nature of his or her relations with the other members. It is a general phenomenon and depends on the interrela-tion of individuals pursuing a common goal. Similarly, Stogdill and Shartle (1955) proposed that leadership needs to be studied in terms of the status, interactions, perceptions, and behavior of individuals in relation to other members of an organized group. Leadership should be regarded as a relationship between persons rather than as a characteristic of an isolated individual. Data for all the members of a group should be combined and interrelated to study leadership in terms of the structural and functional dimensions of the organized interrelationships.
Wofford (1981) presented an elaborate integration of concepts and research results from behavioral studies of ability, motivation, role perception, environmental constraints, determinants of the behavior of leaders, and environmental influences. This leader-environment-follower interaction theory conceived of the leader as a person who analyzes current deficiencies in the conditions that determine the performance of followers and takes correc-tive action. The theory appeared to concentrate on the practice of management by exception. Bennis (1961) concluded that theories explaining who emerged and succeeded as a leader in an organization had to take into account: (1) the impersonal bureaucracy, (2) the informal organization and interpersonal relations, (3) the benevolent autocracy that structures the relationship between superiors and subordinates, (4) the job design that permits individual self-actualization, and (5) the integration of individual and organizational goals.
The personal-situational approach has come to dominate the forecasting of leadership potential in prospective supervisors and managers. The effort builds on attempts to match individuals’ personal history, competencies, and traits with the requirements of a job. Since the late 1970s, analyses of the inspirational, charismatic, transformational leader, and the servant leader have all looked at both the person and the situation. Crises are seen to lie behind the rise of charismatic leaders whose personal development and personality move them to succeed in taking charge. Transformational leaders are usually personally assertive. They react to their perceptions of what their followers need, but they also proactively influence what their followers want. Servant leaders are strongly service-oriented and are influenced by what their organization and followers need. A description that emerged from a conference about what was expected of senior managers illustrates the importance of both person and situation: “Leaders should be uncommon (yet congenial, vulnerable, and accessible); capable as almost to promote awe.” A leader should be “a faultless reader of signals from the environment, a diagnostician capable of taking corrective action; … an architect who builds an enabling organization” (Wilson, 1994).
Increasingly, behavioral theories postulate leadership as an interaction among the leader, the situation, and the led (e.g., Popper, 2001). Yammarino (1991) reasoned that a full explanation of leadership and situation requires a multiple-level approach: leaders, leader-follower pairs, groups, and organizations. At the level of leaders, whether and how much the leaders differ are examined. At the level of the leader-follower pair (the dyad), analyzed is how much the leader relates differently to each follower. At the level of groups, the question is to what extent each leader’s followers differ. At the level of organizations, the question is whether and how much differences in leadership depend on the organization—above and beyond any differences among the different groups and leaders. Many other analyses are possible across levels. These within-and-between analyses (WABAs) are appearing with increasing frequency.
Freud (1922), as well as many other psychoanalytically oriented writers such as Erikson (1964), Frank (1939), Fromm (1941), and H. Levinson (1970), addressed leadership at length, in terms of clinical studies. Favorite interpretations conceived of the leader as a source of love or fear, as the embodiment of the superego, and as an emotional outlet for followers’ frustration and destructive aggression (Wolman, 1971). Freud (1913) proposed that the beginnings of civilization required a struggle with the leader (father) of the primitive clan. A primal horde of sons slew their father and formed a society of equals, but the need for leadership resulted in the rise of totemism and religion, in which gods substituted for the murdered father. According to Freud (1922), group behavior is emotional and irrational. Groups of followers are both obedient to and intolerant of authority and require strength and forcefulness in their leaders. They are oppressed by, fearful of, and ruled by leaders. The group mind determines its cohesiveness. The followers’ identification with the leader shapes their identification with each other. However, within their personalities, they remain in a state of tension. Their ego (rationality) is too weak to resolve the conflict with their id (instincts) and superego (social and moral imperatives). But leaders are without emotional attachment to anyone. They are narcissistic, independent, and self-confident.
Wolfenstein (1977) and Lasswell (1960) made extensive use of psychoanalytic theories to account for political and revolutionary leaders. Lasswell held that the personality of a political person compensates for feelings of inadaquacy and low esteem by displacement, by a continuing pursuit of power to maintain personal integrity. This personal need is rationalized as “in the public interest” (Lasswell, 1962). According to Wolfenstein, the personality of a revolutionary leader is an externalized revolt against a parent. It is the Oedipus complex projected against society. Much of this psychoanalytic theorizing about leadership attempted to explain leaders’ political behavior by looking at their early childhood and families. Thus Freud and Bullitt (1932) said that Woodrow Wilson was obsessed with his articulate and impressive father. Wilson buried his resentment under an intense idealization of his father and publicly played out his private fantasies of Christlike greatness by attempting to become a new savior of the world.
For Freud, the father of the family defined the leader’s psychological world: “He is everyone’s own private leader, who mediates the transition … from inner to outer, from psychology to politics” (Strozier Offer, 1985, p. 43). Fenichel (1945) held that obedience to the father provided protection; the father could become a savior in a crisis (Bychowski, 1948). Mother figures could be as important as father figures. Strong mothers or absent fathers figured strongly in the careers of Franklin Delano Roosevelt, Douglas MacArthur, and many other world leaders (as will be noted in Chapter 34).
Psychohistory. Psychoanalysis was the theory of choice for psychohistorians attempting to understand political leaders in terms of childhood deprivation, cultural milieu, relationships with parental authority, and the psychodynamic needs of their followers. Illustrative of this approach are the psychoanalytic treatises written about Adolf Hitler (e.g., Waite, 1977). Other biographical subjects of psychoanalysts have been Abraham Lincoln, Martin Luther, and Mahatma Gandhi (Erikson, 1969). Psychohistory went beyond psychoanalysis by looking at the personality development and dynamics of leaders in their interactions with society and history. The personal conflicts of world leaders were linked to historical developments. For instance, Erikson (1968, 1969) concentrated on the importance of Martin Luther’s adolescence and Mahatma Gandhi’s adult life. Historical crises were explained in terms of personal traumas of a leader.
Kernberg (1979) focused on the schizoid, obsessive, paranoid, and narcissistic character of leaders. Kohut (1976, 1977) saw charismatic leaders as narcissists, who use their followers to maintain their self-esteem. The followers’ shame, jealousy, and hate are buried by their idealization of the leader. The pathology of leadership was also explored by Kets de Vries (1980, 1984), among others. But although the psychoanalytic study of leadership has accentuated psychopathological issues, “it is … patently absurd to label all leaders as pathological” (Strozier & Offer, 1985, p. 6). There is an imbalance in the psychoanalytical attention given to the neurotic and psychotic aspects of leaders’ behavior and the ignoring of what is healthy and creative in leaders. Therefore, the psychoanalytic view needs to be refocused to explain “that elusive fit between the leader and the led in the full richness of the unique moment of the past” (Strozier Offer, 1985, p. 7). For example, Alexander (1942) and Erikson (1964) considered mature leaders, in contrast to immature leaders, to have innate abilities to command attention, to be free from irrational conflicts, to be sensitive to the needs of others, and to be able and willing to relate emotionally to others.
Using the methods of psychohistory to delve deeper into questions concerning social insight, G. Davis (1975) explained how the psychodynamics of Theodore Roosevelt found expression in his affective insights as an adult leader. Personal recollections, published accounts, journalism, and biographies about Roosevelt as a child were meshed with an analysis of relevant cultural developments in the United States during his time. Davis concluded that Roosevelt’s psyche resolved the childhood experiences of his generation.
Crises and Charisma. Psychoanalytical theory was also used by Kets de Vries (1980) and by Hummel (1975) to show how the interaction of the personalities of leaders and their situations is dramatized in times of crisis. Kets de Vries (1980) maintains that charismatic leaders arise during crises out of a sense of their own grandiosity and the group’s sense of helpless dependency. Whether they serve well as leaders depends on whether they can test their “paranoid potential” and their sense of omnipotence against reality. In fragmented societies, such charisma may give rise to an integration of institutions and loyalties or it may spawn opposition movements (G. T. Stewart, 1974). For Hummel, “projection” by followers explains their intense love for a charismatic leader. Followers see the leader as a superhuman hero because they cannot become consciously aware of their unconscious projections. Zaleznik (1977) proposed that a true leader, in contrast to a manager, has resolved the conflicts of his id and superego and has developed strong ego ideals, embodied in his confidence and self-determination.
Cognitive-Experiential Self Theory. Freud’s initial concept has been improved in the light of new clinical insights and experiments in cognition and the psycho-dynamic unconscious. Freud’s single fundamental need to seek pleasure and avoid pain was replaced as a source of motivation by the need of a leader for a stable, co-herent conceptual framework (Rogers, 1959) and the need to overcome feelings of inferiority (Adler, 1954), among many other alternatives of consequence to the leader-follower relationship. Particularly relevant was Epstein’s (1994) integration of the cognitive and psycho-dynamic unconscious. Epstein assumed two interacting information processing systems—rational and emotional-experiential. Behavior is influenced by both. Four fundamental interacting needs are those (1) for pleasure, (2) to maintain coherence, (3) for relatedness, and (4) to enhance self-esteem. Both leaders and followers would be more likely to value emotional appeals based on hunches if they more often processed information on the basis of experience and emotions. Conversely, they would be moved by principles and deductions if they processed information more often by reasoning.
Group Dynamics. Psychoanalysis had much to say about the leader-follower development in the small group.3 According to Freud (1922), group members act like siblings in developing their ego-identification. They form a common libidinal connection with their leader (father) by incorporating his image into their superego. According to Redl (1942), the central person in the group (not necessarily the group leader) becomes an object of the members’ identification on the basis of love or fear, an object of aggression, and a support for their own egos. The central person can become a model to be admired—the members’ ego-ideal. The followers may internalize the leader’s standards of conduct or come to fear her or him as an aggressor. What had the most influence on the course of subsequent research on education and the practice of group dynamics was Bion’s (1948, 1961) sorting of leader-member relations into four “cultures”: (1) task-oriented, (2) dependent, (3) fight-flight, and (4) pairing.
Psychoanalysis also has much to say about leader-ship in therapeutic groups, although opinions differ on whether the group therapist is the group leader. For instance, Scheidlinger (1980) argued that the therapist’s leadership is important to how the group will function, as well as how much the group can contribute to successful treatment by providing a climate of safety and support for the reenactment of family-child and parent-child encounters.
Political theorists, from Plato on, had explanations, either explicit or implicit, and prescriptions for leadership. Marxism-Leninism, with its focus on economic determination in history coupled with the dictatorship of the proletariat, laid out strong messages about who shall lead and what is expected of leaders. Mao Zedong’s mass-line leadership was much more explicit. It incorporated operant conditioning, consciousness-raising in small groups, confession, self-criticism, and critical feedback. For Mao, the scattered and unsystematic ideas of the masses were to be studied to turn them into concentrated and systematic ideas that the leadership would take back and explain to the masses. This was to continue until the masses became committed to the ideas and then implemented and tested them (Barlow, 1981).
Nazi ideology was centered on the Führerprinzip, which had figured strongly in German authoritarian ideology in the nineteenth century. As propounded by the Nazi movement, unquestioning obedience and loyalty to superiors produced order and prosperity, to be shared by those who were worthy by race (the Aryans) to participate in the “new order.” The other races were to be relocated, enslaved, or exterminated (Evans, 2005). Worship in Japan and fascism in Spain, Italy, and elsewhere had a similar blend of feudalistic, authoritarian, and ethnocentric ideologies. Like kings with a divine right, like the emperor of China who pursued the “will of heaven,” the national dictator could do no wrong, so each successive level of leadership below him was equally infallible. Superiors’ decisions were to be obeyed, not questioned (Evans, 2005). In contrast, in the leadership espoused in the democratic world constitutionally elected representatives are responsible to their constituencies and follow laws based on the legislative vote of the majority, but the rights of the minority are constitutionally respected and protected.
According to J. M. Burns (1977), political leaders is “those processes and effects of political power in which a number of actors … spurred by aspirations, appeal to and respond to the needs … of would-be followers … for reciprocal betterment … or real change in the direction of ‘higher’ values. Political leadership is tested by the extent to which real and intended changes are achieved by leaders’ interactions with followers through the use of their power bases. Political leadership is broadly intended “real change.” It is “collectively purposeful causation” (p. 434). In established governments, “political power” refers to processes for the “authoritative allocation of values that are considered legitimate uses of power under existing … conventions, traditions, understandings, or constitutional processes. This legitimacy is usually linked to formal authority” (J. M. Burns, 1977, p. 434). Political theories of leadership explain the rise of conservative, reform, and rebel movements; the significance of historical events and forces; the mobilization of constituencies; and the importance of the leader’s personality and power. They examine the leadership of presidents, ministers, cabinets, and legislatures in democratic, authoritarian, and totalitarian states. Their focus ranges from the small community to the large urban center. They examine the results of public opinion polls and elections.
Chester Barnard had 40 years of experience with AT&T and eventually became president of New Jersey Bell. On the basis of his experience and observations, in his classic Functions of the Executive (1938), he concluded that cooperation was essential for an organization’s survival. (At the time he wrote, command and control were regarded as more essential by management theorists.) Equally important were acceptance of purpose and the ability to communicate. The executive required the capacity for affirming decisions that provided quality and morality to the coordination of organized activities and to the “formulation of purpose.” The purposes of cooperation were impossible without specialization, and the integration of the specializations demonstrated the cooperation needed for the organization to survive.
Barnard’s ideas were modified by a focus on executive decision making (Simon, 1947; March & Simon, 1958), and more recently by other organizational theorists to fit with changes in society. Thus Guskin (1999) suggested that organizational instability may call for executive command and control until stable conditions are reestablished. Perloff (1999) argued that Barnard had left out the overriding importance of trust in the executive by his (or her) various constituencies (board members, managers, employees, customers, suppliers, government agents, and community). Fowler (1999) saw another factor as having been ignored: the importance to the executive of being allied with reliable advisers who tell what he needs to hear, not what he wants to hear. Among many others, Clifton (1999) added to Barnard’s principles of purpose and the need for communication the psychological and humanistic need of the organization for a sense of shared mission, promoting employees’ participation in decisions, matching employees’ strengths and assignments, and offering incentives for measured superior performance.
Economic Theories of Organization. Economic theories are based on assumptions about the motivation of “economic man.” For instance, according to Agency Theory (Meckling & Jensen, 1976), owners of firms aim to maximize their profits. Nonowner managers are agents of the owners and cannot be trusted to have the same interests as the owners. They are likely to be interested in maximizing their own compensation and advancement. As economic persons, they are selfish, opportunistic, and individualistic, and they may ignore what is best for the organization in favor of what is best for themselves. They will need to be monitored closely by the owners. Williamson (1975) saw the firm as a marketplace in which all members are economically motivated in competition with one another. Bargaining replaces cooperation in getting things done. The interactions between leaders and followers are negotiations influenced by their differences in resources and power. Economic theories are in marked contrast to humanistic theories, which see people as basically cooperative, as having a strong sense of responsibility, and as willing to work toward common goals (Donaldson, 1990).
Social Psychological Theories. Social psychological leadership theories were influenced by the principles of American democracy and individual freedom. The human being is by nature a motivated organism. The organization is by nature structured and controlled. It is the function of leadership to modify the organization to provide freedom for individuals to realize their motivational potential for the fulfillment of their needs and to contribute to the accomplishment of organizational goals.
McGregor. One prominent analysis was McGregor’s (1960, 1966) postulation of two types of organizational leadership—Theory X and Theory Y. Theory X assumed that people are passive: they resist organizational needs and attempts to direct and motivate them to fit these needs. Theory Y assumes that people already have motivation and a desire for responsibility: organizational conditions should be arranged to make it possible for people to fulfill their needs with efforts toward achieving organizational objectives.
Argyris. In his maturity-immaturity theory, Argyris (1957, 1962, 1964a) perceived a fundamental conflict between the organization and the individual. It is the nature of organizations to structure members’ roles and to control their performance in the interest of achieving specified objectives. It is the individual’s nature to be self-directive and to seek fulfillment through exercising initiative and responsibility. An organization will be most effective when its leadership provides the means whereby followers may make a creative contribution to it as a natural outgrowth of their own needs for growth, self-expression, and maturity. Most organizations pursue a one-way model of how people are supposed to relate to others. The model has a single loop or one-way link from the more powerful to the less powerful. In this model, there is preference for: (1) unilateral control; (2) a win-or-lose orientation toward others; (3) concealment of feelings; and (4) a rational censoring of information, freedom, and risk. In contrast, the model espoused by Argyris (1983) is double-looped in that it comprises: (1) a learning orientation; (2) a low-defensive, high-information environment; and (3) joint control by the more powerful and the less powerful. Choice is free and informed. This double-loop model ought to be more effective in the long run for both the individual and the organization.
Heavily influenced by Kurt Lewin (1947), Likert (1961a, 1961b, 1967) argued that leadership is a relative process because leaders must take into account the expectations, values, and interpersonal skills of those with whom they are interacting. Leaders must present behaviors and organizational processes that the followers perceive to be supportive of their efforts and of their sense of personal worth. Leaders will involve followers in making decisions that affect their welfare and work. They will use their influence to further the task performance and personal welfare of followers. Leaders will enhance the cohesiveness of the group and the members’ motivation to be productive by providing followers with freedom for responsible decision making and the exercise of initiative.
Blake and Mouton. Blake and Mouton (1964, 1965) conceptualized leadership in terms of a managerial grid. Concern for people was represented by one axis of a two-dimensional grid; concern for production was represented by the other axis. Leaders may be high or low on both axes, or they may be high on one axis and low on the other. The leader who rates high on both axes develops followers who are committed to the accomplishment of work and have a sense of interdependence through a common stake in the organization’s purposes. Relationships of trust and respect for the leader emerge as well. Misumi and Peterson (1985) reviewed a line of theory and testing by Misumi and colleagues in Japan. Called Performance-Maintenance (PM) theory, it was similar to Blake and Mouton’s concern for performance and concern for production. Optimum supervision occurred when both P and M were high rather than low.
Maslow. Maslow’s theory of “eupsychian” management (1965) was derived from his observations of people at work in industry. Maslow stressed that it is important for managers to develop their subordinates’ self-esteem and psychological health and emphasized the need for self-actualization so that everyone would have an opportunity to realize his or her own capacity. Eupsychian management distinguished between the person who was trying to be a democratic superior and one who was spontaneously democratic. According to this theory, the unconscious and the depths of personality had to be probed in the search for enlightened management. On the basis of such probes, different leaders would be chosen for different situations. The Blackfoot Indians were an example. The Blackfoots gave power to a leader only on an ad hoc basis for the situation in which it was warranted. Such leadership ought not to be left to self-seekers with a neurotic need for power, but should be given to those who are best suited to be leaders for the designated situation—those who can set things straight, who can do what needs to be done.
Hersey and Blanchard. Hersey and Blanchard’s (1969a, 1972) life cycle theory of leadership synthesizes Blake and Mouton’s (1964) managerial grid, Argyris’s (1964a), maturity-immaturity theory and the Ohio State leadership study with regard to concepts of consideration and initiation of structure (Stogdill & Shartle, 1955). According to Hersey and Blanchard’s situational leadership model, the leader’s style of behavior should be related to the maturity of the subordinates As the subordinates mature, the leader should decrease emphasis on structuring tasks and increase emphasis on consideration. As the subordinates continue to mature, there could be an eventual decrease in consideration. Maturity is defined in terms of subordinates’ experience, motivation to achieve, and willingness and ability to accept responsibility.
Rost. Rost’s (1993) theory is at an extreme. Most theories of leadership focus on the leader. Rost argued for eliminating the distinction between leader and follower. Leadership is a process. Influenced by Maslow, Rost held that whoever had the information and motivation in a given set of informal relationships should be able to temporarily be the source of influence in the ideal group, a group without a formal distribution of power, authority, and responsibility.
Social psychology formed the basis of many other theories of leadership. Many other social interaction and social learning theories explain the leader-follower relationship as a consequence of the leader’s interaction with the followers as well as with the circumstances involved. Interaction theories of leadership such as Gibb’s (1958) study are characterized by a complex combination of the leader’s personality; the followers’ needs, values, attitudes, and personality; and the group’s structure of interpersonal relations, character, task, and environmental setting. What happens also may be explained in terms of the leader’s role and its attainment, reinforcement of change, paths to goals, and the effects of contingencies.
Leader-Rote Theories. In informal groups, structure develops and roles are taken so as to permit one person or perhaps a few persons to emerge as leaders. According to leader-role theory, the characteristics of the individual member and the demands of the situation interact so that during the course of the members’ interactions, groups become structured in terms of positions and roles. Leaders are expected to play a role that differs from the roles of other group members. Homans (1950) developed a theory of the leader’s role based on action, interaction, and sentiments. He assumed that an increase in the frequency of interaction by group members and their participation in common activities was associated with an increase in their mutual liking and in the clarity of the group’s norms. The higher the status of persons within the group, the more nearly their activities would conform to the group’s norms, the wider their range of interactions would be, and the larger the number of group members with whom they would originate interactions would become.
In formal organizations, leaders behave according to what their colleagues expect of them and how—in their own perception—their roles are formally defined. The leaders’ and colleagues’ perceptions and expectations of their roles are further influenced by the organization’s formal policies and procedures, by informal communications with colleagues, by past experience, and by their own needs and values (Kahn & Quinn, 1970). There is also an effect on the ratings of their performance as leaders, which depend on whose expectations are most salient for them in defining their roles (Tsui, 1995). The expectations will change as an organization changes. New and conflicting requirements will call for changes in roles (Eggleston & Bhagat, 1993). Managers ordinarily must cope with conflicts among different sources of information about their roles. Osborn & Hunt (1975a) argued that what the organization prescribes for the managers’ routine activities is not leadership; rather, leadership involves only the discretionary activities that leaders perform when the prescriptions fail to tell them what to do. Hunt, Osborn, and Martin (1981, p. 3) presented a well-supported theory to explain why some leaders act efficiently in response “to specific opportunities and problems which the unit is not designed to handle.”
Rules and procedures created by an organization can make the leadership role redundant. Kerr and Jermier (1978) pioneered the analysis of the substitutes for leadership. Nevertheless, a wrong inference can be drawn here—that more available regulations necessarily reduce the discretionary behavior of leaders. On the contrary, Hunt, Osborn, and Martin (1981) predicted and found that leaders in units with more rules, policies, and procedures were expected to respond with more discretionary use of those rules and procedures and actually did so. Jones (1983) analyzed the leader’s role in terms of controls imposed by the organization. Jones argued that such controls of the work flow, of the way a task is structured, and of the way jobs are formalized may provide supervisors with as much influence over what goes on as does their power to discipline subordinates.
Theories on Attaining the Leadership Role. These theories attempt to explain who emerges as a leader of a group and why. Hemphill (1954) argues that leaders emerge in situations in which components of group tasks are interdependent and are related to the solution of a common problem among group members. Fundamental to his theory is the concept of structure in interaction or predictable interaction activity. The role structure of the group and the office of the leader are defined by institutionalized expectations with respect to initiation of structure in interaction. The probability that an attempted act of leadership will succeed is a function of the members’ perceptions of their freedom to accept or reject the suggested structure in interaction. When such a structure leads to the solution of common problems, it acquires value and strengthens the expectation that all group members will conform to it. Thus initiation of structure in interaction is attempted leadership. Hollander and Julian (1969) hold that an emergent leader, instead of being just another undifferentiated member of the group, is the member who has built up idiosyncrasy credit with his or her followers by successive successful attempts to lead.
Role Expectations. According to Tsui (1984), superiors, peers, subordinates, and others indicate what they expect of the manager’s role. The expectations of these various constituencies can be in conflict. Superiors may expect the manager’s role to be that of a monitor; subordinates may see it as that of a coach. If the expectations are met, the leader gains reputational effectiveness. Consistent with Hemphill, Stogdill (1959) developed an expectancy-reinforcement theory of such role attainment. This theory attempted to explain the emergence and persistence of successful leadership in initially unstructured groups. It also tried to understand what leadership is and how it comes into existence. As group members interact and engage in shared tasks, they reinforce the expectation that their actions and interactions will continue in accord with their previous performance. Thus the members’ roles are defined by confirmed expectations of the performances and interactions they will be permitted in order to contribute to the group. The potential of members to be successful leaders is the extent to which they initiate expectations and maintain structure in interaction. For Stein, Hoffman, Cooley, et al. (1979), emergent leaders are the group members who are most willing and able to perform those roles and functions that enable the group to accomplish its tasks. They guide and encourage others to contribute to the process. Such leadership will appear in phases that parallel Tuckman’s (1965) stages of group development: orientation, conflict, and emergence. Some emergent leaders take charge early; others move ahead with collaborators; and still others fail to maintain their initial success as leaders.
Path-Goal Theory.5 The reinforcement of change in the subordinate by the leader is a prominent aspect of path-goal theory. Georgopoulos, Mahoney, and Jones (1957) and M. G. Evans (1970a) suggested that the successful leaders showed followers the rewards available to them. House (1971) maintained that the leader also showed the followers what paths (behaviors) to follow to obtain the rewards. The leader clarified the goals of the followers as well as the paths to reach those goals. This clarification enhanced the psychological state of the followers and aroused them to increase their efforts to perform well. Followers achieve satisfaction from the job to be done. The leaders enhanced satisfaction with the work itself and provided valued extrinsic rewards like recommendations for pay increases, contingent on the subordinates’ performance. (The leader needed to be able to control the rewards that subordinates value.)
The situation determined which behavior by the leader could accomplish these path-goal purposes Two important situational aspects were how competent the subordinates were and how highly structured the task was (House & Dessler, 1974). To reconcile the theory with experimental results, House (1972) proposed that the effects of a leader’s behavior were contingent on three kinds of “moderator variables”: (1) task variables, such as role clarity, routine, and externally imposed controls; (2) environmental variables; and (3) individual differences in preferences, expectations, and personality.
Contingency Theories. Along with House’s path-goal theory, Fiedler’s contingency theory (1967a) dominated much of the research on leadership during the 1970s and 1980s. For Fiedler, the effectiveness of task-oriented and relations-oriented leaders was contingent on the demands imposed by the situation. Leaders were assessed as task oriented or relations oriented according to the way they judged their least preferred coworker. A situation was favorable to the leader if the leader was esteemed by the group; if the task was structured, clear, simple, and easy to solve; and if the leader had legitimacy and power by virtue of his or her position. The task oriented leader was most likely to be effective in situations that were most favorable or most unfavorable to him or her. The relations oriented leader was most likely to be effective in situations between the two extremes of favorable and unfavorable.
Most person-situation theorists focused on how the leader ought to be developed to adapt best to the needs of the situation. But Fiedler’s research and theory emphasized that the leader ought to be placed in the situation for which he or she is best suited. Task oriented people should be selected to lead in situations that are very favorable or unfavorable to the leaders; relations-oriented people should be selected to lead in situations that are neither very high nor very low in favorability. Otherwise, leaders needed to learn how to change a situation to match their orientation. Therefore, Fiedler, Chemers, and Mahar (1976) developed a method to help a leader “match” his or her appropriate situation: the leader was helped to change the situation or to adjust better to its favorability or unfavorability. Fiedler, Chemers, and Mahar’s (1976) leadership-training program consisted of first identifying the trainee’s particular style—task or relations orientation—and then teaching the trainee how to analyze and classify leadership situations for their favorableness, or situational control. The next elements to be considered were the best fit of the situation and style and how to change one’s style to suit the occasion or how to change the situation to fit one’s style better. Fiedler’s five decades of work involved a progression from empirical discoveries to the formation of the theory to the practical application of the theory and to validation of the applications.6
Situational contingencies. Many others have proposed that personal traits resulting in the emergence or success of a leader would be influenced by the task or goals, the followers’ traits, and the organizational context of the situation (Bass, 1960).
Additional Theories and Models of Interactive Processes. Numerous additional elaborations have appeared to account for leadership and for leader-follower relations as an interactive process. For instance, Fulk and Wendler (1982) and Greene (1975) agreed that if subordinates (followers) perform well, the leader displays more consideration, which then leads to increased satisfaction for the followers. If the followers do not perform well, the leader displays more structuring behavior and the followers’ satisfaction does not increase.
Communication Theories. Communications and rhetoric provide another point of departure for theories about leader-follower interactions. For example, Sharf (1978) created a rhetorical framework based on a theory by Burke (1969) to analyze the relative success of emerging leaders in small groups in obtaining cooperation from the other members and in resolving the struggle for leadership status. When applied to recorded discussions of small, leaderless task groups, the analyses revealed the importance of going beyond symbolic divisions in the emergence of leadership.
Multiple-Linkage Model. Yukl (1971) agreed that the leader’s initiation of structure enhances subordinates’ ability to cope with a situation; the leader’s consideration for the welfare of subordinates enhances the subordinates’ satisfaction with the situation. Then Yukl (1981) greatly expanded the interaction framework with a multiple-linkage model suggesting that the subordinates’ effort and skill in performing a task, the leader’s role, the resources available, and the group’s cohesiveness all moderate the effects of the leader’s behavior on group outcomes. The model also differentiated between leadership required for short-term effectiveness and that required for long-term effectiveness. Yukl and Kanuk (1979) provided evidence that, in contrast to performance outcomes, subordinates’ satisfaction resulted from different patterns of behavior by the leader and mediating conditions.
Multiple-Screen or Cognitive Resources Model. Another interaction approach to understanding the relations of the leader and the led is the multiple-screen model, which attempted to explain the relationship between the leader’s intelligence and the group’s performance. Fiedler and Leister (1977) suggested and provided empirical support for the proposal that intelligent leaders can generate effective groups if the leaders have good relations with their bosses. If relations are poor, then experienced rather than intelligent leaders bring about more productive groups. Experience is more important to effective leadership if leader-boss relations are poor.7
Exchange theories assume that group members make contributions at a cost to themselves and receive benefits at a cost to the group, the organization, or other members. As with Burns’s (1978) transactional leadership, leaders exchange rewards or the avoidance of discipline for followers’ satisfactory performance. Interaction continues because leader and followers find the social exchange rewarding. Blau (1964) began with the fact that most people consider it rewarding to be elevated to a position of high status. Also, it is rewarding for members to associate with their high-status leaders. But leaders tend to deplete their power when members have discharged their obligations to the leaders. The leaders than replenish their power by rendering valuable services to the group. They benefit as much as anyone else from following their good suggestions rather than somebody else’s poorer ones. Followers’ compliance constitutes a surplus profit that the leader earns.8
Vertical-Dyad Linkage (VDL) and Leader-Member Exchange (LMX) Theory. Graen’s (1976) vertical-dyad linkage (subsequently called leader-member exchange) was one of many interaction theories that were based on the assumption that social interaction represents a form of exchange. The vertical dyad of leader and subordinate (VDL) is an interaction linkage of mutual influence. It emphasizes the relationship between the leader and each individual follower rather than between the leader and the group as a whole. In corresponding LMX theory, Graen (1976) assumed that the leader behaves differently toward each follower and that these differences must be analyzed separately. This theory is in opposition to most earlier theories, which assumed that the leader behaves in much the same way toward all group members and that behavioral descriptions from group members can be averaged to obtain an accurate description of the general behavior of the leader. According to Graen, leaders categorize followers as belonging to an in-group or an out-group, and the leader behaves differently toward members of these two groups. In-group members can be more independent of the leader and receive more attention from the leader, as well as more of the other rewards. As a consequence, in-group members perform better and are more satisfied than out-group members (Vecchio & Gobdel, 1984). Numerous, extensive empirical investigations by Graen and associates of vertical-dyad effects have been published and will be discussed more fully in Chapters 15 and 16.
Organizationally Defined Expectations. T. O. Jacobs (1970) formulated a social-exchange theory and buttressed it with a wide range of research findings. According to Jacobs, the group provides the leader with status and esteem in exchange for the leader’s unique contributions to the attainment of the group’s goals. Authority relationships in formal organizations define role expectations that enable group members to perform their tasks and to interact without the use of power. Leadership implies an equitable exchange relationship between the leader and the followers. When role obligations are mutually acknowledged, each party can satisfy the expectations of the other on an equitable basis.
Aaronovich and Khotin (1929) reported using differentially cued reinforcements to alter the leadership behavior of monkeys in uncovering boxes of food. Mawhinney and Ford (1977) reinterpreted path-goal theory in terms of operant conditioning. W. E. Scott (1977) saw a need to replace the concept that leadership is due to influence or persuasion with an analysis of the observable behaviors of leaders that change the behavior of subordinates. All these behavioral theories emphasized reinforcement and making the receipt of rewards or the avoidance of punishment contingent on the subordinate’s behaving as required. According to Davis and Luthans (1979, p. 239), “The leader’s behavior is a cue to evoke the subordinate’s task behavior. The subordinate’s task behavior, in turn, can act as a consequence for the leader, which, in turn, reinforces, punishes, or extinguishes the leader’s subsequent behavior. Similarly, the subordinate’s behavior has its own consequences, which serve to reinforce, punish, or extinguish this behavior. The consequences for the subordinate’s behavior may be related to the leader’s subsequent behavior [to] the work itself, and its outcomes, or [to] other organization members.”
Supervisors do not directly cause subordinates’ behavior; they merely set an occasion or provide a discriminative stimulus. The behavior of subordinates depends on its consequences. Environmental cues, discriminative stimuli, behaviors, and consequences form a behavioral contingency for analysis. Thus Sims (1977) conducted one of many investigations demonstrating that a leader’s positive rewarding behavior will improve a subordinate’s performance, particularly if the reward is contingent on the quality or quantity of the subordinate’s performance.
Concentrating on the followers’ reactions, Kerr and Abelson (1981) developed a model to represent extreme, rather than ordinary, leader-subordinate interactions. The day-to-day behavior of the leader may be relatively unimportant to the supervisor-subordinate relationship, compared with the leader’s behavior when a subordinate experiences an intense demand or when the leader experiences a highly unexpected response. This model shows that subordinates can become so accustomed to frequent leadership activity that the effects of this activity are minimal and even dampening.
The importance of reinforcement in the leader-follower relationship will be examined in various contexts in many of the later chapters that deal with power relationships and exchange relationships. Reinforcement is central to the operant model of supervision as well as to pathgoal theory.
Operant Model of Supervision. This model derives from Skinner’s (1969) theories of behavioral conditioning. It describes what leaders should do “compellingly and consistently to motivate subordinates” (Komaki & Citera, 1990, p. 91). Three categories of behaviors are specified: (1) antecedents of subordinates’ performance (instructions, rules, training, goals); (2) monitoring of performance (work sampling, subordinates’ self-reporting, secondary sources) and; (3) consequences of performance (feedback, recognition, correcting). For an empirical example, more effective insurance managers spent more time than less effective insurance managers monitoring their agents (Komaki, 1986). Racing sailboat captains who monitored their crews hoisting their sails and provided feedback correlated (.51, .47) with the racing results (Komaki, Deselles, Bòwman, 1989).
An early theoretical emphasis on the perceptual and cognitive aspects of leadership was provided by Goffman (1959), who analyzed social behavior as theater. That is, Goffman evaluated the roles, membership, and phenomena of groups in terms of actors, audience, front stage, and backstage. According to Goffman, social learning created disparities between the leader’s intentions and the followers’ understanding of what the leader was trying to do. As noted earlier, Quinn and Hall (1983) developed an integrated theory of leadership based on competing perceptual and cognitive dimensions such as leaders’ flexibility versus their control, and leaders’ internal versus external focus. Carrier (1984) constructed cognitive maps to locate traits of leadership in reference to these dimensions. For example, the trait of dominance was placed in a location that is high in both control and internal focus.
Perceptual and cognitive theories offer several advantages. They make use of advances in cognitive psychology and are immediately applicable in diagnosis and leadership education. They include theories about social and cognitive processes, attributions, integration with behavioral models, information processing, systems analysis, and rational-deductive decision trees.
Cognitive Processes. Before the late 1970s, empirical leadership research and theory paid little attention to leaders’ behavior as a consequence of perceptions and thoughts about purposes, followers, or the task and situation. An exception was the use of stimulated recall, in which leaders and members of small groups were audio-taped, then listened to the tapes, and then indicated on a second tape recording what they had been thinking at the time of the original discussion (Bass, McGehee, & Hawkins, 1953). Conference videotaping would provide on a replication of the videotape sotto voce commentary by the participants about their thoughts concerning what was originally happening and what they were trying to do.
Leaders’ and followers’ perceptions and cognition rose in importance. Lord and Maher (1985) first showed how leadership could be best understood by attention to cognitive processes, the importance of implicit theories of leadership, and the “prototypes” or models of leadership that people form mentally. Concepts that have emerged in theories about cognitive processes include attention, encoding, attribution, memory storage, memory retrieval, evaluations, expectations, and attributions.
To appreciate a leader’s behavior, we need to understand the leader’s scripts and strategies. A script is a conceptual structure. It is held in the memory of events, objects, roles, feelings, and outcomes. There is a sequential pattern to the script’s structure of familiar circumstances and tasks. According to Wofford (1998), if a leader uses a script to confront a familiar situation and finds in feedback no discrepancies between goal and outcome, he or she will continue to process the script. The process depends on recognition and is automatic. Otherwise, the process is inferential and is based on reflection of the recent past, integration of performance information, and causal attributions (Cronshaw & Lord, 1987). The leader is likely to include alternative paths to the goal in the script, so that if feedback indicates that the original path is blocked, an alternative will be tried. If no alternative path works, the leader changes strategy. A strategy is an original construction of a new script formulated in a pattern to deal with a specific situation for which available scripts have not worked. Or an alternative strategy may be retrieved from memory. Experienced and trained leaders should have a greater number of available scripts. Also, they will need to change strategies less often (Wofford, 1998).
Another cognitive concept, the schema, is an organized knowledge structure. According to Lord Emerich (2001), it is necessary to consider the difference between an individual and a collective schema. An individual schema, as the term implies, is in the mind of an individual; it reflects perceived networks of implicit theories of leadership and categorizations of types of leaders into prototypes. A matched prototype is selected. A collective schema is “a socially constructed understanding of the world derived from social exchanges and interactions among multiple individuals in a group or organization” (p. 552). Central themes include organizational performance, “sense making,” and transformations.
McCormick and Martinko (2005) have combined social cognitive theory with causal reasoning to provide an understanding of a group leader’s thoughts and be-havior and the group’s performance. McCormick and Martinko assume that people regulate their own thoughts and can control their own actions. According to Locke and Latham (1990), people actively monitor the performance environment, develop functional task strategies, implement plans, and monitor results. McCormick and Martinko hold that leaders’ and followers’ self-regulation is guided by attention and attributions of the causal reasoning process as well as schemata about task-relevant knowledge, skills, abilities, personal goals, action plans, and beliefs in their own efficacy.
Self-efficacy is a cognitive concept in Bandura’s (1985) social cognitive theory. It is the belief that one has the capability to handle prospective problem situations (Bandura, 1995). A leader’s choice of goals and strategies will be affected by a belief in self-efficacy (Kane & Baltes, 1998). McCormick (2001) has built on Bandura to formulate a Social Cognitive Leadership Model. Engaged in self-regulation, the would-be leader, high in self-efficacy, is likely to generate attempts to lead and persist in persuasive efforts despite resistance (Savard & Rogers, 1992). If the leader’s belief is valid, the leadership should be effective (Chemers, Watson, & May, 2000).
Propositions that follow include: (1) Leaders who are not self-serving and are objective in their perceptions of their environment generate more efficacious behavior. (2) Leaders have an implicit model or schema that in-fluences how they perceive their causal relationships. (3) Leaders’ task schema are influenced by their beliefs about causation in their environment. (4) Leaders’ self-efficacy influences their goals, task strategies, and use of task schema. (5) Lofty goals are a result of perceptions of “resource-rich” task schemata and a strong belief in self-efficacy (McCormick & Martinko, 2005).
Attribution Theories. DeVries (1997) constructed a model of attribution that linked impression formation, categories, prototypes, perceptions of leadership ideals, and leadership success. Each leader and follower is perceived to have his or her own implicit theories of leadership. If we want to understand the behavior of individual leaders, we must begin by attempting to find out what they are thinking about the specific situation. Whether they are seen as acting like leaders depends on their own and their followers’ implicit theories about leadership (Eden & Leviatan, 1975). We observe the behavior of leaders and infer that the causes of these behaviors are various personal traits or external constraints. If these causes match our naive assumptions about what leaders should do, then we use the term “leadership” to describe the persons who we observed. Thus Calder (1977) says that leadership changes from a scientific concept to a study of the social reality of group members and observers—a study in how the term is used and when it is used, and assumptions about the development and nature of leadership. For Calder, leadership is a perception of followers that caters to their perceptual needs. Attributions of leadership by observers and group members are biased by their individual social realities (Mitchell, Larson, & Green, 1977), which accounts for the low correlations that are often found between supervisors’, peers’, and subordinates’ ratings of the same leaders (Bernardin & Alvares, 1975; Ilgen & Fugii, 1976; T. R. Mitchell, 1970a), as well as for the confounding of evaluations of the performance of subordinates and the behavior of leaders (Rush, Thomas, & Lord, 1977).
Green and Mitchell (1979) formulated a model to study such attributional processes in leaders. They explained that a leader’s behavior is a consequence of his or her interpretation of the subordinates’ performance. Thus a leader presented with an incident of a subordinate’s poor performance (such as low productivity, lateness, a missed deadline, or disruptive behavior) will form an implicit theory about the subordinate and the situation, judging that the cause of the incident was the subordinate’s own personality, ability, or effort, or an externality, such as lack of support, a difficult task, or insufficient information. Causality is attributed more to the subordinate than to the situation if the subordinate has had a history of poor performance and if the poor performance has severe outcomes (Mitchell & Wood, 1979). In such circumstances, the leader will focus remedial action on the subordinate, rather than on the situation, even if the situation was the cause of the problem. Meindl, Ehrlich, and Dukerich (1985), along with Pfeffer (1977) and Calder (1977), agreed that there is a tendency to attribute more of the cause than is actually warranted to the subordinate rather than to the situational circumstances.9
The Romance of Leadership. Meindl, Ehrlich, and Dukerich (1985) argued that leadership is in the eye of the follower. It is a social construction in the followers’ minds and is about their thoughts of how leaders are cognitively structured and represented. Followers place more emphasis on the image of the leader than on the actual behavior or effects of the leader. They are much less under the control and influence of the leader per se than under the control and influence of “the social forces that govern the social construction process itself” (Meindl, 1995, p. 330). The object of study is not the actual personality or behavior of the leader, but personality and behavior as constructed or imagined by the followers. Ordinarily, when correlations are found between the supposed leadership and the effectiveness of an organization, the traditional interpretation is that the effectiveness was the result of the leaders’ performance. But interpreted as a romantic notion, leadership is imagined by the followers. Its purpose, as noted above, is to cater to the perceptual needs of the followers (Calder, 1977). Symbols and rituals reinforce the importance of the leader (Pfeffer, 1977). Meindl (1990, 1993) found empirical support for the romantic perspective. Nonetheless, a leader does have real effects on followers, although some effects may be imagined by them. Much additional evidence in this book will attest to the objective effectiveness of leadership when measured independently of followers’ opinions. For instance, Smith, Carson, and Alexander (1984) found that leaders do moderately influence the performance of their organizations. Day and Lord (1988) demonstrated that the executive leadership of organizations can account for 45% of these organizations’ effective performance. But the question remains in many studies: how much is real and how much is romance, and in what circumstances?
Integrated Cognitive and Behavioral Models and Theories. Although attributions by followers can strengthen explanations, any explanation needs to be based on behavior. Still, Lord and Maher (1993) assume that to be a leader one must be perceived as leader. Traits are schemata, knowledge structures, and “sense makers.” Perceptions of distant executive leaders depend mainly on attributions and inferential processing from events and outcomes. Prototypes, scripts, implicit theories, and categories grow out of experiences. They are stored in long-term memory and are activated automatically. Controlled processing is intentional information processing based on short-term memory. The distinction explains a loss of charisma as followers shift from inferential to controlled processing. To solve problems, experts use automatic processing; novices use controlled processing. Experts use meaningful categories. Diagnosis is implicit, intuitive, or both. Solutions and evaluations are scripts and heuristics available to the expert. Novices categorize the same problems by using surface or environmental features, and solutions evaluations.
Luthans (1977) left room for cognitive processes to enter the scenario “to assign concepts to behavior and to infer relationships between events.” Luthan’s functional analysis of the leader-subordinate dynamic used his S-O-B-C model, in which S is the antecedent stimulus, O is the organism’s covert processes, B is the behavior, and C is the consequence. Chemers (1993) provided a behavioral model of leadership, which also took account of cognitive processes. The first of three dimensions of the model was relative to persons: it was the development of interpersonal relationships. The second dimension was resource utilization relevant to the task to be accomplished. The third dimension was the leader’s image management, the effort of the leader to establish and maintain a specific image in the group, such as competence or conscientiousness. As already noted, Hoojberg, Hunt, and Dodge (1997) integrated cognitive, social, and behavioral complexity into a “Leaderplex” model to generate empirical leadership research on global organizations, team-based organizations, diversity, hierarchy, and charisma.
In a mixed model proposed by Bass (1960), leadership deals with the observed effort of one member in a group to change the motivation, understanding, or behavior of other members. Change will be observed in the followers if the leading member is successful in influencing the others. Motivation is increased by changing the followers’ expectations of being rewarded or punished. Leaders acquire their position by virtue of their perceived ability to reinforce the behavior of group members by granting or denying rewards or punishments. Since the group’s effectiveness is evaluated in terms of its ability to reward its members, leaders are valued when they enable a group to provide expected rewards. The congruence of a leader’s perceived status (the value of the position held, and esteem) with the leader’s ability and value as a person regardless of the position can account for the leader’s success. Incongruence generates conflict and failure. This emphasis on congruence is also found in Halal’s (1974) general theory. A particular style of leadership is congruent with specific technologies of tasks and specific motivations of subordinates. Adaptation occurs to achieve greater congruence.
Information Processing. Newell and Simon (1972) focused on the problem solver’s “subjective problem space.” This “space” contains encodings of goals, initial situations, intermediate states, rules, constraints, and other relevant aspects of the task environment. Lord (1976) saw the utility of studying the shared problem spaces of leaders and followers when they tackle a common task. For example, a leader was expected to devote more effort to developing an orientation and definition of the problem or the group when the actual task lacked structure.
Social cues and symbols take on more importance for an understanding of leadership if this information processing approach is employed. In addition to encoding, information processing involves selective attention, comprehension, storage, retention, retrieval, and judgment. According to both theory and evidence, perceptions of leaders are based largely on spontaneous recognition. Moreover, the cognitive category of leadership is hierarchically organized. Perceptions and expectations of the attributes and behavior of leaders are widely shared (Lord, 1976, 1985; Lord, Binning, Rush, et al., 1978). Recognition-based processes are dominated by automatic processing, categorizations, and implicit theories, ordinarily involved in face-to-face interactions at lower levels in the organization. Inferential processes are attributional, controlled, and inferred from events and outcomes ordinarily associated with distant executive leaders (Lord & Maher, 1991).
An open-systems point of view implies sensitivity to the larger environment and organization in which leaders and their subordinates are embedded. To convert inputs into outputs, flows of energy and of information must occur in the system. In open systems, the effects of the outputs on the environment are feedback and new inputs. The relations within the system grow and become more intricate with repeated input-output cycles. The cyclical conversion process can be increased in rate and intensity. Leaders or followers can import and introduce more information. Directive leaders do this alone; if followers are included, the process is participative. Energy can be increased by selecting as leaders and followers more highly motivated individuals or by increasing the reinforcements that accrue from outputs (Katz & Kahn, 1966). Agency theory (Meckling & Jensen, 1976) suggested that managers put their own interests ahead of those of owners and shareholders, as when they provide themselves with “golden parachutes.” But open-systems theory sees the manager as a good steward (Donaldson, 1990), concerned with aligning as much as possible the interests of all the constituents of the organization: owners, shareholders, management, employees, and community. The worth of the organization depends on this alignment (Bass, 1952).
Bryson and Kelley (1978) created a systems model for understanding emergence, stability, and change in organizations in which formal leaders are elected, such as cooperatives, professional associations, and legislatures. They made a list of clusters of individual, procedural, structural, and environment variables that were likely to be of consequence to each other on the basis of earlier formulations by Peabody (1976) and Van de Ven (1976).
Change-Induction and Therapeutic Groups. Lieberman (1976a) explained change-induction groups, such as psychotherapy groups, encounter groups, self-help groups, and consciousness-raising groups, in terms of systems analysis. Five structural characteristics of the system were seen to affect the change-induction process: (1) the psychological distance between the participant and the leader; (2) felt causes, sources, and cures of psychological misery; (3) the extent to which the group is seen as a social microcosm; (4) the degree to which members stress differentiation rather than similarity; and (5) the relationship between the cognitive and expressive behavior of the leader.
Macro-and Microlevels. Many models and theories of leadership have been embedded in larger organizational models and theories. For example, Bowers and Seashore’s (1966) four-factor theory of leadership is part of a larger systems theory of organizations. Osborn and Hunt (1975a, 1975b) formulated an adaptive-reactive model of leadership to incorporate environmental constraints and organizational demands as antecedents of the behavior of leaders. Likewise, Bass and Valenzi (1974) used systems theory to construct an open-systems model of leader-follower relationships. According to their model, the systems are open to the outside environment and are sensitive to the constraints imposed on them from outside. The system imports energy (power) and information from outside, converts it, and exports goods and services. The Bass-Valenzi model (Bass, 1976) proposes that whether leaders are directive, negotiative, consultative, participative, or delegative depends on their perceptions of the system’s inputs and the relations within the system. The leader and his or her immediate work group form an open system of inputs (organizational, task, and work-group variables), relations within the system (power and information differentials), and outputs (productivity and satisfaction). For instance, the Bass-Valenzi model posits that leaders will be more directive if they perceive that they have more power and information than their subordinates. They will consult if they perceive that they have the power but that their subordinates have the necessary information to solve the group’s problems. They will delegate when they perceive that their subordinates have both power and information. They will negotiate when they perceive that they have the information but not the power. A small-space analysis of empirical data by Shapira (1976) supported these propositions.10
Bass (1960) argued that the emergence of leadership success in influencing the group and its effectiveness (the group’s actual achievement of its goals such as reward or the avoidance of punishment) depended on the interaction potential in the situation—the physical, psychological, and social proximity among the would-be leader and the followers. The likelihood that individuals would interact depends on the size of the group; the geographic and social proximity of the individuals; their opportunity for contact, intimacy, and familiarity; mutuality of esteem and attraction; and homogeneity of abilities and attitudes. Monge and Kirste (1975) extended the examination of proximity as a time-and-space opportunity, showing the positive association of proximity with the potential to interact as well as its contribution to satisfaction with the interaction.
Starting with open-systems theory and Jaques’s (1978) general theory of bureaucratic organizations, Jacobs and Jaques (1987) formulated a theory to explain the requirements of leadership at successively higher echelons of large bureaucratic organizations such as the U.S. Department of Defense or General Motors. To operate successfully, these organizations must have an appropriate structure, which Jacobs and Jaques specified as no more than five operating echelons and two additional higher headquarters echelons. At each echelon, the complexity of the environment must be understood and clearly transmitted to the next echelon below to reduce uncertainty there. A reduction of uncertainty will add value to productivity at that echelon and define how it must adapt to remain competitive. At each echelon, the role of the leadership is to ensure the accuracy of the uncertainty reduction process and the availability of resources for the required adaptive changes. To accomplish this goal, leaders at successively higher echelons increasingly must have “the capacity to deal with more uncertain and more abstract concepts” and with longer time spans for accomplishment and evaluation.
At the lowest three echelons, leaders must focus on how they can contribute to the organization’s productivity above and beyond the rules and policies that have been laid down for them by higher authority. At the next two echelons, leaders must concern themselves with how to maintain and improve their organizational arrangements. At the highest echelons, leadership involves strategic decision making in a “nearly unbounded” environment.
Multiple Levels Approach. It is possible to look at the same behavior at three levels of a system—individual, group, and organizational—although the operational character of a construct will change if we move from one level to another. For instance, individualized consideration, a component of transformational leadership, would be addressed at the individual level by a question such as, “Does the leader spend time with newcomers to help orient them to their jobs?” At the team or group level, the same issue would be addressed by, “Do the members of the team provide useful advice to newcomers?” At the organizational level, the question might be, “Are there special policies and programs for orienting newcomers?” (Avolio Bass, 1995). We need to specify at which level we are operating. For example, a leader may be described somewhat differently by each member of a team but uniformly by the team as a whole. Conger (1995) conceived a need to add a fourth level—the intrapsychic—within the leader’s mentality. Data about a leader can form an additional level when they are collected during several different periods or phases. Markham, Yammarino, and Palanski (in press) found that the “leader-member exchange” had an effect on performance, which was greater in predicting performance for 25 manager’s groups composed of their 110 subordinates when each leader-member dyad was taken into account, rather than each manager’s group average of the exchange results.
Cognitive, behavioral, and interactional explanations are likely to be needed to account fully for leader-follower relations and the outcomes of these relations. Gilmore, Beehr, and Richter (1979) instructed leaders in an experiment to display either a lot of or a little initiative and a lot of or a little consideration. Although the participants who were subjected to the leadership failed to perceive that their leaders’ behavior actually differed, a lot of actual (but not perceived) initiation, coupled with a lot of actual (but not perceived) consideration by the leaders, resulted in better-quality work by the participants. The quality of the participants’ work was lower when the leaders displayed a great deal of initiative but little consideration. Evidently, under certain conditions it is more profitable to make use of behavioral theories to understand the behavior of leaders. Under other circumstances, such as when leaders and subordinates must act on the basis of their interpretations of a situation, perceptual and cognitive theories are more useful. Some theories or aspects of theories may account better for the leadership that handles short-term disturbances; other theories may deal better with the leadership that corrects chronic deficiencies over the long term.
Winter (1978, 1979a) developed a complex model that combined aspects of the trait, reinforcement, behavioral, and cognitive approaches and the feedback loops of systems analysis. Winter’s model was based on a battery of tests of skills and behavioral competence measures for over 1,000 naval personnel and their leaders. Figure 3.1 shows the emergent model that links various skills with particular performances. The model is based on empirical cluster analyses and subsequent regression analyses.
Greater optimization (assigning tasks to those subordinates who are most likely to do them well and making trade-offs between the requirements of the tasks and individual needs) and setting goals both contributed to more delegation by the leader. Increased monitoring by the leader resulted in more positive expectations, disciplining, advice, and counsel. It also contributed to more feedback, which in turn led to more disciplining and giving more advice and counsel.
Zand (1997) found that effective executives make use of knowledge, trust, and power. These three factors are needed to create an impelling vision and to direct a broad course of action. Working with their staff, the executives select a path, sharpen concepts, learn from mistakes, make adjustments, and refine their strategy and implementation as they go along, Johnston (1981) used many of the preceding theories to construct a model of a holistic leader-follower” grid. To represent the leader-follower interchange adequately, Johnston borrowed from Jung’s (1968) psychoanalytic theory of life cycles, Berne’s (1964) transactional analysis, McGregor’s (1960) theory X and theory Y, Rogers’s (1951) nondirective counseling, and Tannenbaum and Schmidt’s (1958) model of decision making. Tomassini, Solomon, Romney, et al. (1982) also constructed a cognitive-behavioral model in which the leader’s influence interacted with the subordinate’s work behavior, and they identified situations that circumscribe what a leader can do.
Figure 3.1 Flow Chart of Navy Leadership and Management Processes in Terms of the Cross-Validated Competencies
SOURCE: Adapted from Winter (1978).
As noted in Chapter 4, Stogdill (1948) reviewed leadership studies between 1904 and 1947 and found that they employed one of six methods: (1) observation and time sampling of behavior in group situations; (2) choice of associates by voting, naming, ranking, or sociometrics; (3) nomination by qualified observers; (4) identification of persons occupying leadership positions; (5) analysis of biographical and case history data; and (6) listing of traits essential to leadership. Research on the traits of leaders tended to rely heavily on tests and questionnaires for the collection of data. With the development of interaction models and theories, it became important to know what was happening in the group. Expert or trained observers were used to keep a running record of the behavior of group members, and the leaders and followers also might be asked to report their feelings and observations. In some cases, observers merely reported what they saw; in other cases, they were provided with checklists of behaviors or processes to be reported. Bales (1950) developed a checklist for observations of a set of behavioral categories. Carter and associates (1951); Mann (1979); and Komaki, Zlotnick, and Jensen (1986), among others, also developed observational categories and procedures. Bass, Gaier, Farese, et al. (1957) relied on the changes in correlations among members’ judgments from before to after group discussion to measure how influential each member had been.
Observational studies identified behaviors that the trait theorists did not anticipate. Whereas the trait theorists were interested in the leader’s personality, the interaction experimenters were more concerned with observable interactional behaviors of leaders and their followers. The two approaches could produce moderately correlated results. Jacoby (1974) was able to demonstrate substantial construct validity based on agreement among three methods of assessing opinion leadership—(1) self-designating, (2) sociometric, and (3) key informant.
Laboratory experiments and field surveys have often been the method of choice in cognitive and behavioral investigations. Experiment and survey together provided more convincing evidence. By now, it is fair to say that almost every procedure known to social science has been applied specifically to the study of leadership. These procedures have included content analysis and thematic analysis of autobiographies, biographies, and case studies; verbal protocols; individual structured interviews face-to-face and by computer; panel interviews; interview boards; news and journal reports of historical events, stories, and anecdotes; historical accounts; memorandums; minutes of meetings; speeches; biodata; sociometry, using face-to-face and electronic communication patterns; autologs; cognitive maps; observers’ logs of leaders’ activities; and ratings by observers, superiors, peers, subordinates, and outsiders, such as clients, customers, and suppliers. Many analyses were based on leaders’ self-ratings; but to increase confidence, these increasingly were supported by investigations using two or more independent approaches. For example, Heller (1969a) first collected survey-questionnaire data from managers. After analyzing the results, he gathered the managers together in panels to interpret and confirm or refute the findings. Also, focused interviews of panels of voters provided the basic ideas for political media campaigns that were then followed up by public opinion polls of representative samples of voters (Kerman & Hadley, 1986).
For experiments, leaders were described in fictional sketches as being reacted to by readers as if they were followers. Leaders were self-selected, appointed, nominated, elected, emergent, or simulated. Methods were quantitative or purely qualitative. but there was increasing use of each method to buttress the other. The historiometic approach was the oldest method, beginning with Quetelet’s (1835/1968) research on leading dramatists. Historiometric research on leadership was exemplified early on by Cox’s (1913) study of the influence of monarchs, and more recently by O’Connor, Mumford, Clifton, et al. (1995) on famous and infamous world leaders. Simonton (1999) enumerated psychological studies of eminent persons such as famous sports stars, Nobel laureates, and chess grand masters. Quantitative methods are used to analyze historical sources of information. Psychometric research applied surveys, interviews, and psychological tests to study individual leaders. Modern psychobiographical research began with Freud’s (1910/1964) psychoanalytic study of Leonardo da Vinci. An intensive study of a single individual, Roosevelt, The Lion and the Fox is illustrative (J. M. Burns, 1956). Psychobiographical research can also deal with multiple cases. It has usually been qualitative. Politicians and presidents are most often the subject of investigation. Comparative research looks at a small sample of leaders and contrasts them according to biographies, interviews, media accounts, and other sources of information and events. Among the many quantitative and qualitative methods, some of those used more in leadership are presented briefly below.
Of the 188 articles published by the Leadership Quarterly between 1990 and 1999, descriptive statistics were provided by 87%, and simple inferential statistics by 69%. From 25% to 32% of the articles made use of multivariate analyses. Less frequently employed were confirmatory factor analysis (14%), structural equation modeling (13%), and multiple-levels analyses (13%; Lowe & Gardner, 2001). The use of quantitative methods in social science such as Q sorting, confirmatory factor analysis, and structural equation modeling increased with the availability of computer analysis. Other quantitative methods in social science most applicable to studies of leadership were multiple-levels analysis, rotational designs, and designs for the avoidance of common error variance. Meta-analysis made possible valid statistical summaries of the results of replicated and similar studies using the same or comparable variables.
Q Sorting. Q sorts are measurements, usually rankings on a set of ratings or tests within each of the same individuals instead of the more common measurements between individuals on the same set. Individuals rather than measurements are correlated. When factor analysis is applied to the matrix of Q sorts, factor types with similar profiles usually emerge, instead of factors of measurements (Cattell, 1946). Q sorting was also applied to matching ratings in order of importance of job requirements with ratings of each candidate’s profile of knowledge, skills, and abilities (O’Reilly, 1977). In a large consumer products company, the consistency of ratings of 60 statements of requirements for the position of production manager by 10 job experts was .96. The agreement among the rated manager, a peer, and the boss about the requirements was .81. The correlation of the person-job fit ranged from 2.44. to .86. A high correlation meant a very good fit; a high negative correlation signified a very bad fit. The rank order of the fits correlated .98 with the rank order of performance of the managers rated by the human resources staff and superiors. Similar findings were reported in three further investigations (Caldwell & O’Reilly, undated).
Confirmatory Factor Analysis. Unlike the more traditional exploratory factor analysis, which extracts underlying factors from a matrix of correlations of individuals and variables, confirmatory factor analysis begins with a model of the assumed factors and provides several tests of the goodness of fit between the model and the factors. For instance, Avolio, Bass, and Jung (1999) determined that six factors of the “Full Range of Leadership” provided the best fit of 14 samples of data from scores on the Multifactor Leadership Questionnaire.
Structural Equation Modeling. This is a systematic analysis making use of partial and multiple regressions to explore or confirm causal relations between a set of predictors and a set of outcomes. The fit of the obtained model and the hypothesized model can be tested. The strengths of the relations between variables can be obtained from path coefficients similar to partial correlations. Computer programs like LISREL facilitate the analysis for exploration and confirmation.
WABA: Within-and-Between Analysis. Dansereau, Alutto, and Yammarino (1984) began a continuing line of investigation of variant analyses which quantitatively shows how much of the ratings of the leader in a single study are due to different levels of analysis. The “person level” acknowledges the importance of consistent individual differences among leaders, followers, or both. The “dyad level” recognizes the importance of one-to-one relationships between a leader and each of the followers. The “group level” or “team level” takes cognizance of the face-to-face relations between a leader and a set of followers as a unit. For instance, the leader may treat the followers in the same set in the same way or different ways. The “collective level” or “organizational level” deals with hierarchically organized groups of groups, as when the same or different organizational policies apply to all groups (Yammarino, Dansereau, & Kennedy, 2001). The focus is between entities such as dyads or groups, not within them. Here, as in traditional analysis of variance, differences between entities are viewed as valid; differences within entities are viewed as random error. This is a between-units case in which the members of the unit are homogeneous and the whole unit is important. The “parts” view differs from the traditional analysis of variance in focusing within entities such as dyads or groups. Differences within entities are valid; differences between entities are considered random error. This is the within-units case in which members are heterogeneous; each member’s position relative to others is important. Dyads can be viewed as parts of a group. In this case of dyads-within-groups, leaders differ from each other in the same group, and the same pattern is repeated in the other groups. Thus each leader may form a favored in-group of subordinates and a less favored out-group. Dyads may be completely independent of the groups to which they belong. Then they are viewed as whole dyads, independent and homogeneous entities. Analysis of variance (WABA 1) as well as co-variance (WABA II) can be applied to quantitatively decompose a single survey of ratings of leaders and their correlated effectiveness.
Further breakdowns may test whether there is an equitable balance of leader and subordinate, whether relations between dyads are stronger in some groups than others, and whether consensus is the same in some dyads in some groups but not others. Cross-level effects from dyads to groups can also be specified (Yammarino, 1995; Dansereau, Yammarino, & Markham, 1995). In a study of insurance and retail salespersons’ attitudes about factors controllable by management and supervisory ratings of the salespersons’ performance, Yammarino and Dubinsky (1990) found differences between the groups of aggregated ratings as well as individual differences within the groups. Some groups of supervisors and subordinates were rated higher than others across the groups; some supervisors and subordinates in the same group were rated higher than others. Castro (2002) contrasted WABA with other multi-level methods of leadership research. These included intraclass correlation (rwg), hierarchical linear modeling (HLM), and random group resampling (RGR).
Intraclass Correlation. The expression rwg is the ratio of the difference in the variance of ratings between leaders’ and followers’ ratings of each of the same leaders, compared with the variance of ratings between leaders (suitably corrected for the number of leaders). This index provides the reliability of the ratings for the same and different leaders, the agreement among raters and whether aggregation of the ratings is justified (James, Demeree, & Wolf, 1984).
Hierarchical Linear Modeling (HLM) This is a two-level approach. First, the regression equations for the un-aggregated ratings (the individual level) and outcomes are calculated. Then, the intercepts, slopes, error, and regressions become the data for the group level of a parallel analysis. The two levels of analysis permit tests of significance of the individual ratings and the aggregated grouped ratings. HLM can be used to test moderator effects across levels and longitudinal changes, if not limited by the need to assume normality of the data, and if there are a sufficient number of cases (Bryk & Raudenbush, 1992).
Random Group Resampling (RGR). Based on Fischer’s permutation test, proposed in 1930, RGR is related to the bootstrap and jackknife (Efron & Tibshirani, 1993). A pseudo group of the same size as the first actual group of the first followers’ leadership ratings is created by sorting the data from lowest to highest and randomly forming a pseudo group without replacement. The remaining data are combined into a second pseudo group. The means of the two pseudo groups are calculated. Then, Monte Carlo techniques are applied to create a set of 1,000 pseudo groups and their distribution of differences in means. Finally, the mean group differences obtained from the actual data are tested as to whether they came from this distribution of differences of the pseudo groups and whether the average data at the group level or the individual differences in followers need to be considered. RGR can be used to test the significance of WABA II results for the co-variance between predictors and outcomes (Bliese Halverson, 2002).
Rotational Designs. To test for personal versus situational effects, the same individuals are rotated in a systematic way through a set of conditions. An individual is placed in different groups to examine the consistency of his or her emergent leadership in each group (Kenny & Hallmark, 1992). For instance, Zaccaro, Foti, and Kenny (1991) first tested participants on selected traits. Next, participants were systematically rotated through a set of group tasks in different groups. The investigators then were able to determine that the effects of the traits on the participants’ emergent leadership in the different groups was much stronger than the differences from group to group, indicating that personality was much more important than the situation in determining who would lead.
Increasing Response Rates to Survey Questionnaires. Responses to mail questionnaires can fall to 20% or lower. One way to test whether the respondents are a valid or biased sample is to send out successive waves of the questionnaires to see if answers do not change as the numbers of respondents increase with each wave. Rogelberg and Luong (1998) listed techniques supported by research to increase the response rate: (1) Notify potential respondents in advance. (2) Follow nonrespondents with reminders. (3) Provide incentives. (4) Use appeals. (5) Keep surveys to a reasonable length. (6) Facilitate the returns by providing first-class-stamped envelopes. (7) Ask easy and interesting questions first and demographic questions last. (8) Tell who the sponsors of the research are. (9) Ensure respondents’ anonymity. Telephone questioning can also increase response rates. Response rates can come close to 100% when questionnaires are handed out and completed in an assembly. According to a meta-analysis by Cycyota and Harrison (2002), response rates of executives can be increased by reaching them through their social networks such as associations, professional groups, and colleagues.
Rating Inaccuracies. Starbuck and Mezias (1996) reviewed 210 publications about misperceptions that are sources of inaccurate ratings by managers. These articles dealt with managerial perceptions of organizations and their internal and external environments. Ten other publications provided objective data for analysis. The authors concluded that perceptual data play an extremely important role in studies and theories about managerial behavior. Objective situations defined by perceptual data may not correlate with each other. Employees fail to agree about the properties of their organization. Perceptions may be very inaccurate. Needed in organizational and leadership research are studies to measure the errors and biases in perceptions and their determinants.
Anonymity. Anonymity eliminates the fear of reprisal, particularly when subordinates want to rate superiors unfavorably (Antonini, 1994). The extent to which subordinates’ favorable or unfavorable ratings were reciprocated by superiors was examined in anonymous and nonanonymous conditions for 241 teams of female undergraduates in groups of six with an assigned leader. They conducted a discussion to reach decisions about the value of 15 items if they were stranded in a desert. Significant reciprocity of evaluations between the leader and the led occurred in the nonanonymous but not the anonymous condition (Haeggberg & Chen, 2000).
Avoiding Common or Same-Source Variance. Common methods variance is the overlap between two variables due to a common bias rather than to a relationship between the underlying constructs. For instance, a correlation between two measurements may be partially due to the fact that the measurements were obtained with the same method, at the same time, or by the same rater. Common or same-source error is greatest when two sets of data are collected at the same time, from the same respondent, on the same instrument, by the same method, and about the same trait. Campbell and Fiske (1959) suggested that each respondent be assessed on each of at least two different traits by at least two different methods to take common variance into account This multitrait, multimethod procedure is one of six ways proposed to eliminate or control error. None is foolproof, according to Podsakóff and Organ (1986), who suggested applying WABA analysis in which each rater is matched with a counterpart. The first rater rates one variable—say, leadership—and the second rates an outcome. (See Avolio, Yammarino, & Bass, 1991, for a more detailed exposition.)
Meta-Analyses. Meta-analysis is an effort to estimate the population or true effect of an analysis from results obtained from comparable samples. Three major types of meta-analysis have been developed by Rosenthal and Rubin (1988), Hedges and Olkin (1985), and Hunter and Schmidt (1990). The basic ideas go back to the early 1930s (Johnson, Muller, & Salas, 1995). Rosenthal and Rubin convert each sample of results into standardized scores (Z’s) with one-tail probabilities for significance. Effect sizes derive from Fisher’s conversion of correlations to Z’s. Weighted means are generated for comparing the relevant dependent variables. Hedges and Olkin convert samples of results into standard deviation units (g’s). These need correction because they overestimate the population effect size, especially with small samples. The results are combined, their consistency is tested, and their variability is explained by use of models with moderators. Hunter and Schmidt’s approach is probably the best-known of the three approaches. It does not correct the biases in effect size or with moderators. Like the other approaches, it weights the means of each sample according to size. It corrects effect sizes for sampling error, attenuation due to restriction of range in each study, and the reliability of the variables used. By 2006, we were able to report numerous meta-analyses to support conclusions about many aspects of leadership and management and their effects.
Indexes of Change. To measure whether a change has occurred in the leader, the led, or the organization as a consequence of learning, three measures were proposed by Golembiewski, Billingsley, and Yeager (1976). Alpha change is a simple rise or fall in a measurement of the level of a state of affairs calibrated to reflect a one-to-one change in the concept assessed. Beta change is a rise or fall in the measurement where there is a recalibration to reflect systematic modification in relation to the concept assessed (see also Terborg, Howard, & Maxwell, 1960). Gamma change occurs if there is a change in any of these: perspective, frame of reference, or concept assessed. Factor analysis was applied to register changes in factor structures from before to after an intervention (Golembiewski et al., 1976; Schmitt, 1982).
Critique. MacCallum (1998) pointed to a need to correct interpretations drawn from quantitative methods in leadership and organizational research including structural equation modeling, factor analysis, analysis of variance, multiple levels of analysis, and basing measurement reliability on coefficient alpha. He also noted methodological problems in event history analysis. He favored making more use of archival data, using moderated regression, and comparing the goodness of fit of several models rather than depending for significance on the fit of a single model.
Increasingly, we are seeing efforts to use triangulation of quantitative and qualitative methods to reinforce conclusions from each. A quantitative survey is accompanied by a qualitative in-depth interview of selected respondents. The themes in a qualitative description of a case are categorized, counted, and content-analyzed.
The 188 articles in the Leadership Quarterly in the 1990s used about half as many qualitative methods as quantitative methods. The most favored qualitative methods were content analyses (53%), case studies (45%), and grounded theory (24%; Lowe & Gardner, 2001). Results would have favored more qualitative methodology if the Journal of Leadership Studies or political science and sociological journals had been sampled instead of the psychology-oriented Leadership Quarterly.
Strauss and Corbin (1990) noted what was required to collect reliable and valuable observational data. Days and situations for observation need to be routine. Interviews need to be conducted with a cross-sectional sample representative of the population of subjects. Coding of data needs to be done through successive iterations from a smaller initial list to a more complex final list. These data need to be checked against recordings of the initial observations and interviews. From a background in cultural anthropology which depends heavily on qualitative research, Conger (1998) argued that qualitative methods, although time-intensive and complex, are the method of choice for capturing insights into contextual effects and longitudinal developments in the study of leadership. Running summaries are kept of the large amounts of the observations and data collected. General categories are refined as new data are gathered. Ideas, concepts, and theory emerge and evolve. We have already presented the qualitative endeavor of psychohistory. Psychoanalytic concepts are employed to infer the causes underlying the development and performance of individual leaders and their followers. Freud and Bullitt’s (1932) interpretation of President Wilson’s career is illustrative of an ideo-graphic approach to psychohistory. Qualitative research may be idiographic or nomothetic. The ideographic approach is an intensive study of a single case; for example, Kofodimos, Kaplan, and Drath (1986) studied the character and development of a single manager. They included his work life and private life to try to understand why he behaved as he did. The nomothetic approach tries to draw inferences from a more limited exposure to a large number of cases. Most of the studies presented in this book are nomothetic.
Interviews. Some kind of structured or open-ended interviews begin many qualitative analyses. With a background like Conger’s, Sayles (1964) applied the anthropologist’s approach to interviewing. He sought to build a coherent, chronological account of problems, cases, and issues. His interviews focused on learning about events, transitions, and conflicts. Using triangulation, he would verify or reject new information about critical interactions. He learned how events were perceived by different managers in diverse roles and was able to “reconstruct the recurring social process underpinning the tough human challenges of the organization” (Sayles, 1999, p. 9). Sayles saw the challenges of leadership as the “heart” of management. Leaders’ interventions had to be skillfully timed, sequenced, and executed.
Waldman, Lituchy, Gopalakrishnan, et al. (1998) conducted a qualitative analysis of managers. They drew themes and categorical schemes from interviews in a manufacturing plant, a hospital, and a police force engaged in quality-improvement programs. They uncovered two alternative paths of managerial commitment to the quality-improvement process. One path, following top management’s vision and commitment, was characterised by continued commitment, planned adjustments, and a cultural shift favoring quality improvement. The other path, from top management leading down, was one of wavering commitment, knee-jerk reactions, and cynicism about organizational change. From the CNN broadcast Pinnacle, verbatim interviews with CEOs, videotapes of 30 of 80 interviewees were randomly selected by Piotrowski and Armstrong (1989) and rated on 25 personality-lifestyle dimensions by two raters. The two raters agreed that the CEOs exhibited clear values, tolerance of frustration, and egalitarian attitudes. They were also oriented toward tasks, people, results, and compulsivity. They were workaholics and needed more social recognition and family life.
Content Analysis. Content analysis is often employed in qualitative studies. After identifying the research question, the text to be examined, and the unit of analysis, an initial coding scheme is tried on a pilot text. The coding scheme is purified and the observational, interview, or written data are collected. The reliability of the trained coders is assessed. Modifications of the training and the coding may be required and tried out on a second pilot. The construct validity of the categorization is determined. If reliability and validity are satisfactory, the data are coded and analyzed. Insch, Moore, and Murphy (1997) provided examples of content analyses, methods of analysis, and their strengths and weaknesses.
Cognitive Mapping. Sims and Siew-Kim (1993) uncovered the tension involved in simultaneously being a woman, an Asian, and a manager by asking each of nine collaborating senior Singaporean woman managers to draw a cognitive map showing issues and experiences in becoming and being an Asian woman. Subsequently, each woman provided incidents about managing or being managed. The research was grounded in the women’s own language, expressions, thought systems, beliefs, and concepts: the women explained what they thought was important and why. Nine themes emerged: (1) integrity, (2) coping with male insecurity, (3) western versus Taoist and Confucian values, (4) supporting friends and superiors, (5) female leadership, (6) fighting, (7) individual versus group rights, (8) not wishing to undermine others, and (9) high standards for oneself.
Repertory Grid. Although George Kelly (1955) originated this technique in the United States in the 1930s to use in counseling students, it gained more widespread use in management development and organizational research in Britain. It focused on people’s implicit constructs, concepts, and theories; and the language they used to describe cognitions, perceptions, and behaviors. In its least structured form, interviewees are asked to identify the persons in a group who are most similar to each other. Next, they are asked to use words and phrases to tell in what ways these people are similar. Then interviewees are asked to identify the person in the group who is most different from the first two chosen persons and in what ways. The process is repeated for each member of the group. Content analysis of the responses follows. (For a more structured approach to the repertory grid, see Esterby-Smith, Thorpe, & Holman, 1996.)
Stories. Boje (1995) gained access to the Disney studio archives of audiotapes and videotapes of Walt Disney and other Disney leaders’ speeches, work interactions, documentary interviews, and conversations. Also included were television shows, films, cartoon shorts, working meetings, and stockholders’ meetings, as well as many public relations films. Boje was interested in stories about this storytelling organization and was able to infer from them that Walt Disney was a tyrant. Implications will be discussed in Chapter 17.
Spicochi and Tyran (2002) provided two examples of the role of leaders as story tellers and how followers made sense of the stories. These stories were attempts to communicate a vision of transformation in the health care industry, where there is need for flexibility and responsiveness to a continually changing environment. The authors suggested a number of ways leaders can make their stories effective. First, in order to learn how their stories will be interpreted by employees, they should first learn about employees’ reactions to past organizational change and how their attitudes have developed. Clarify the vision. Understand how employees will see the vision as affecting them. Believe in the vision. Minimize ambiguity. Appear in person to tell the story. Incorporate a range of stories from all levels of the organization to make sense for the leadership. Empathize with your audience and focus on their understanding. Finally, be ready to listen and to further clarify the meaning in the story.
Drawing from the Humanities and the Arts. According to Yammarino (2002), “Leadership is part art and part science. … An inclusive approach recognizes the complementary roles of the humanities, the arts, and the social sciences in the pursuit of leadership theory, research, and practice.” History, drama, literature, and art are underutilized by students of leadership. Yet they could provide a stronger base for theory, building to avoid developing theories of leadership that are “academic amnesia” and “leadership déjà vu” (Hunt & Dodge, 2000). A good example of an historical source that has not been used for developing a modern theory of leadership is the Anabasis by Xenophon (c. 400 B.C.), whose account of the successful march of the 10,000 Greek soldiers through a hostile Persian empire illustrated the importance of cohesion, commitment, confidence, and leadership. Gordon (2002) notes that much of the emphasis on power in such historical sources and events has not been used to formulate testable concepts and theories of leadership.
McNamara’s Fallacy. Camille Cavour and then Lord Curzon declared as early as the 1860s that if you can’t measure something, you don’t know what you are talking about. And so body counts were used invalidly to measure victory in battle in Vietnam. The need to qualify quantitative results was provided by a sophisticated view of quantitative data—“McNamara’s Fallacy.” This fallacy is named after Robert McNamara, who as U.S. Secretary of Defense tried to quantify military outcomes. The fallacy can be paraphrased as follows. First, measure what can easily be measured. This is OK. Second, disregard what can’t be easily measured or give it an arbitrary quantitative value. This is artificial and misleading. Third, assume that what can’t be easily measured is unimportant. This is blindness. Fourth, assume that what can’t be easily measured really doesn’t exist. This is suicide. The fallacy illustrated the need for qualitative methods to support quantitative analyses (Handy, 1994).
Among the valid methods combining quantitative and qualitative methods are the subjects of biographies assessed by personality tests and scales, triangulation, and data analyses about historical figures.
Biographies and Personality Tests. Biographies of political personages can become the basis for raters to complete personality tests about them. The tests are then scored to provide quantitative data about the political figures (Immelman, 1998).
Triangulation. Analogous to locating a target at the junction of two lines on a map, triangulation refers to combining results of a qualitative investigation with the results of a parallel quantitative investigation (Jick, 1979). For instance, Berson, Jung, and Tirmizi (1997) buttressed a quantitative analysis showing that light infantry platoon leaders, rated as transformational in their home station, subsequently led more effective platoons in near-combat. They also included a multi-level analysis in the process. The investigators found that the categorized responses of the observer-controllers to open-ended questions about the leaders and the platoons in action were predicted by the home station ratings, near-combat data, and U.S. Army leadership doctrine.
Modern historians with access to archival documents and data can use both to provide lessons for leaders and managers. Historians of southern history have made extensive use of data to show that slavery in the South was profitable to plantation owners. Apologists for slavery pointed to the humanitarianism of slave owners who bore the expense of caring for aged slaves who could no longer work. Mortality data show that most slaves did not live long enough to “retire.”
Many of the most prominent theories and methods related to the study of leadership were reviewed briefly and will reappear in more detail in later chapters. The long history of reliance on greatman theories naturally led to a search for traits of leadership and theories of traits. Genetics has also increased in interest. In reaction, there arose an equally strong emphasis on environmental theories. Political, psychoanalytical, group dynamics, and humanistic theories appeared along with those built on communications and leaders’ behavior, perception, cognition, and cognitive resources. Finally, syntheses were achieved in theories and methods of interacting persons and situations, built around exchanges, transformations, attributions, information processing, role attainment, reinforced change, paths to goals, contingencies of leader and situation, open-systems analysis, rational deduction, and multiple levels of analysis. Empirical research and modern theories of leadership began with studies of the personal factors that contributed to the emergence and success of a leader.