The social scientific study of leadership: quo vadis?

Citation metadata

Date: May-June 1997
From: Journal of Management(Vol. 23, Issue 3)
Publisher: Sage Publications, Inc.
Document Type: Article
Length: 32,644 words

Document controls

Main content

Abstract: 

The social scientific study of leadership started in the early 1930s. Over the years, the cultivation of knowledge about this phenomenon has been cumulative. Trait, behavioral, contingency and neo-charismatic paradigms have contributed to the development of leadership studies. However, although many things are already known about leadership, deficiencies about the field of leadership research still abound because certain issues that are critical to leadership remain untouched by researchers. These include managerial supervision, generic leader functions and specific leader functions, organizational context, diversity management, strategic leadership, leadership styles, cross cultural leadership, political leadership, management training and development, and universal or near universal effective leader behaviors .

Full Text: 

Introduction

Although the phenomenon of leadership has been around since antiquity (Bass, 1990), the systematic social scientific study of leadership did not begin until the early 1930s. As we shall show, the resulting contributions have been cumulative, and a great deal is known about leadership phenomena. However, there remain many unanswered questions. In this article, we attempt to specify some of the more important of these questions and some of the deficiencies in the present store of knowledge concerning leadership. For example, to this day, the dominant proportion of the more than 3,000 studies listed by Bass (1990) is primarily concerned with the relationship between leaders and their immediate followers, and largely ignores the kind of organization and culture in which leaders function, the relationships between leaders and superiors, external constituencies, peers, and the kind of product or service provided by the leader's organization.

The leadership literature is based on a limiting set of assumptions, mostly reflecting Western industrialized culture. Almost all of the prevailing theories of leadership, and about 98% of the empirical evidence at hand, are rather distinctly American in character: individualistic rather than collectivistic, stressing follower responsibilities rather than rights, assuming hedonism rather than commitment to duty or altruistic motivation, assuming centrality of work and democratic value orientation, and emphasizing assumptions of rationality rather than asceticism, religion, or superstition. Further, a number of important topics are largely ignored or only very recently addressed in the leadership literature.

In this article, we present brief overviews of the research paradigms that have been most prominent in the leadership literature historically, and the more prominent extant theories. We discuss the major foci of these paradigms and theories, their assumptions, their limitations, and some of the problems remaining to be resolved. We then discuss a number of issues and topics relevant to the exercise of leadership which we believe warrant serious attention, but have been largely unexplored or ignored. The outcome of this article is a specification of research issues intended to provide some new directions for the development of future leadership theory and empirical research.

The Leadership Trait Paradigm

Systematic research concerned with leadership first focused on the search for individual characteristics that universally differentiate leaders from nonleaders. This research was largely atheoretical. A large number of personal characteristics were investigated such as gender, height, physical energy and appearance as well as psychological traits and motives such as authoritarianism, intelligence, need for achievement, and need for power. The dominant part of this literature was published between 1930 and 1950.

In influential reviews of the trait literature, Gibb (1947), Jenkins (1947), and Stogdill (1948) identified several studies in which traits were associated with measures of leader effectiveness, with correlations as high as .50. Unfortunately, such findings were seldom replicated in multiple studies, and it appeared to scholars of the time that there were few, if any, universal traits associated with effective leadership. Consequently, there developed among the community of leadership scholars a near consensus that the search for universal traits was futile. It should be noted, however, that the most influential author to address this issue (Stogdill, 1948) did not call for an abandonment of the study of traits, but rather for an interactional approach in which traits would be considered as interacting with situational demands facing leaders. Substantial progress in the development of personality theory and in the operationalization of traits has been made since the early 1980s. As we shall show, trait theory has re-emerged and is alive and well. Not only have several defensible theoretical trait-related propositions been introduced in the last decade and a half, there is a modest amount of empirical evidence in support of these propositions.

Problems with the Early Trait Paradigm

In hindsight, it is easy to criticize earlier research. However, one needs to appreciate the limitations associated with early investigation of the phenomena. One problem with early trait research was that there was little empirically substantiated personality theory to guide the search for leadership traits. Consequently, there were few replicative investigations of the same traits. Also, test-measurement theory was not well developed during the time when trait studies dominated leadership research. As a result, even when common traits were studied in two or more investigations, they were usually operationalized differently. Very little information about the psychometric properties of the trait measures were reported; thus, it is possible that many of the measures had limited validity. As a consequence of the lack of theory and valid measurement instruments, both the traits studied and the way they were operationalized varied widely among investigators. Further, neither specific situational demands of leaders nor the degree to which the situation permitted the behavioral expression of personality inclinations were taken into account. Finally, trait studies were almost entirely based on samples of adolescents, supervisors and lower-level managers, rather than individuals in significant positions of leadership, such as high-level managers and chief executives with overall responsibility for organizational performance.

Revival of Trait Theory

In the early 1970s, interest in leadership traits re-emerged. Substantial advancement occurred in theory due to clarification of several theoretical issues. In addition, several new empirically supported traits have been suggested. We discuss the reemergence of the leadership trait paradigm in this section.

Theoretical Clarifications. Beginning in the mid 1970s, the study of individual dispositions as predictors and explananda for individual behavior has become more theoretical. Bem and Allen (1974), Mischel (1973), Schneider (1983), and House, Shane, and Herold (1996) have clarified when and how traits are likely to explain individual behavior. Bem and Allen (1974) argued theoretically and demonstrated empirically that traits are more predictive of behavior for some people than for others. Thus, trait-relevant predictability is a trait in itself. We speculate that this predictability can be explained by self-monitoring tendencies of individuals (cf. Snyder, 1974). High self-monitors respond more to situational cues and are, thus, less likely to express dispositional inclinations in many situations. In contrast, low self-monitors are more likely to enact their dispositions behaviorally regardless of strength of situations or situational cues.

Mischel (1973) made the important observation that the behavioral expression of dispositions is likely to be suppressed by highly constraining "strong" situations, but that dispositions will likely be enacted in "weak" situations. Strong situations are those in which there are strong behavioral norms, strong incentives for specific types of behaviors, and clear expectations concerning what behaviors are rewarded and punished. Thus, in organizations that are highly formalized and governed by well-established role expectations, norms, rules, policies and procedures, there is less opportunity for organizational members to behaviorally express their dispositional tendencies. Strength of situation was not taken into account in the early leadership trait studies. Mischel's argument has since received support in a laboratory experiment (Monson, Hesley & Chernick, 1982), and two field studies (Barrick & Mount, 1993; Lee, Asford & Bobko, 1990).

Schneider (1983) addressed one of the major criticisms of trait theory. Critics of trait theories argue that traits must be stable and predict behavior over substantial periods of time and across widely varying situations (Davis-Blake & Pfeffer, 1988). Schneider (1983) observed, however, that traits are predictive of an individual's characteristic behavior in select situations, rather than across all situations. Thus, an individual who is disposed toward aggressiveness, as indicated by some psychometric measure, is more likely to behave in an aggressive manner only in aggression-arousing situations; for example, situations in which others disagree with or threaten the individual. This tendency to aggress is a characteristic that differentiates between individuals only under aggression-arousing conditions. In other situations, individuals with an aggressive disposition are not likely to behave more aggressively than others.

House, Shane, and Herold (1996) observed that individual dispositions may be stable over extended periods of time, but not necessarily for life. Thus, traits can predict behavior in the short-term and such behavior often has long-term consequences, even for somewhat unstable traits. For example, House, Spangler, and Woycke (1991) demonstrated that U.S. presidential motives inferred from inaugural addresses were predictive of presidential charismatic behavior and presidential effectiveness throughout presidential first terms. Even if presidential motives inferred from later speeches or writings of presidents disclose patterns different from those indicated in the inaugural addresses, the motives disclosed in the inaugurals clearly had strong predictive power with respect to both presidential action throughout their first terms (House et al., 1991), and with respect to the social and economic effects of presidential actions (Spangler & House, 1991).

In addition to the above theoretical clarifications, there was also a substantial yield from earlier trait research that has gone largely unnoticed by subsequent students of leadership.

Unrecognized Yield from Early Trait Research. House and Baetz (1979) pointed out that when studies of adolescents and children are omitted from Stogdill's (1948) review, the results show a rather consistent set of relationships between some traits, followers' perceptions and indicators of leadership, with many correlations ranging from .40 to .50. The traits House and Baetz found to be rather consistently supported by reconsideration of Stogdill's review were intelligence, prosocial assertiveness (dominance as measured by the California Personality Inventory), self-confidence, energy-activity, and task-relevant knowledge.

Stogdill (1974) updated his earlier review of the trait literature based on studies conducted between 1949 and 1970. He concluded that his earlier paper had under-emphasized the possibility that certain traits exhibited by leaders might well be quite universal. Lord, DeVader, and Alliger (1986) conducted a meta-analysis of 35 of the early studies dealing with six leader traits. Lord et al. found that three traits - intelligence, dominance, and masculinity - were all significantly associated with follower perceptions of leadership, with a fourth trait, adjustment, not far behind.

These findings, while showing that the study of leader traits has considerable promise, are atheoretical and provide no explanation for the associations between the traits and leader effectiveness. In contrast, there has recently emerged a modest body of trait theory and evidence relevant to leadership and the practice of management.

Recent Trait Perspectives

There are four trait theoretical perspectives that enjoy nontrivial empirical support. These are McClelland's Achievement Motivation Theory, his Leader Motive Profile (LMP) Theory, House's Theory of Charismatic Leadership, and Kenny and Zaccaro's leader sensitivity and flexibility constructs.

Achievement Motivation. Achievement Motivation Theory was originally developed in the 1940s (McClelland, 1961), and has a great deal of relevance to the practice of leadership. However, this relevance has only recently been empirically demonstrated (House et al., 1991; House, Delbecq & Taris, 1997). Despite the fact that over 1,000 studies relevant to achievement motivation have been conducted, and the fact that these studies strongly support the theory (Spangler, 1992), this construct has been given scant attention in the organizational behavior or leadership literature. Achievement motivation is defined as a non-conscious concern for achieving excellence in accomplishments through one's individual efforts (McClelland, Atkinson, Clark & Lowell, 1958). Achievement motivated individuals set challenging goals for themselves, assume personal responsibility for goal accomplishment, are highly persistent in the pursuit of goals, take calculated risks to achieve goals and actively collect and use information for feedback purposes. Achievement motivation is theoretically predicted to contribute to effective entrepreneurship (McClelland, 1985) and effective leadership of small task-oriented groups (House et al., 1991). High achievement motivated individuals engage spontaneously in a high degree of self-regulatory behavior such as that described by social cognitive psychologists (Bandura, 1986), without training or directions from others.

It is interesting to note that the similarity between characteristic nonconscious achievement motivated schema and self-regulatory cognitions has not been addressed by current social-cognitive psychologists, despite the fact that the dominant themes of achievement motivated individuals in responses to ambiguous stimuli (pictures or statements) clearly reflect such self-regulatory thoughts, aspirations, and feelings. The responses to projective tests by high achievement motivated individuals suggest that they spontaneously and rather consistently engage in goal setting, envisioning successful performance, anticipation of obstacles and sources of social support, strategic planning to overcome obstacles, use of feedback to measure performance, and self-evaluation contingent on performance achievements.

In management positions at middle or higher levels, and particularly in organizational settings where technical requirements are few and impact on others is of fundamental importance, managerial effectiveness depends on the extent to which managers delegate effectively and motivate and coordinate others. Theoretically, high achievement motivated managers are strongly inclined to be personally involved in performing the work of their organization and are reluctant to delegate authority and responsibility. Therefore, high achievement motivation is expected to predict poor performance of high-level executives in large organizations.

Taken together, the above considerations suggest that achievement motivation will be positively related to the effectiveness of leaders of small task-oriented groups and leaders of relatively small entrepreneurial firms, and negatively related to effectiveness of middle and high-level managers in large complex organizations or in political situations. There is some evidence to support these predictions. Litwin and Stringer (1968) demonstrated experimentally that small groups managed by individuals who enacted achievement-oriented and achievement-arousing behaviors were more effective than groups with managers who did not. House et al. (1991) found achievement motivation of U.S. presidents was inversely related to archival measures of presidential effectiveness. More recently, House, Delbecq and Taris (1997) found that achievement motivation reflected in interviews with chief executives is strongly associated with indicators of organizational effectiveness in entrepreneurial firms. In contrast, achievement motivation reflected in interviews with heads of divisions of large and more bureaucratic firms, was not associated with indicators of organizational effectiveness.

Social Influence Motivation and Leader Motive Profile (LMP) Theory. Several authors have also theorized that, since the practice of management relies heavily on social influence processes, social influence motivation as measured by power motivation (McClelland, 1975; Winter, 1973), measures of desire for influence, or measures of prosocial influence motivation (often inappropriately labeled dominance), will be predictive of managerial success and leader effectiveness. This hypothesis has been supported in a large number of laboratory experiments (see House & Baetz, 1979 for a review), as well as several field studies of managerial populations (Miner & Dachler, 1973; Spangler & House, 1991; Winter, 1991). Interest in leader social influence motivation subsequently led to a more complex theory that is promising and enjoys more than modest empirical support - the Leader Motive Profile Theory (LMP theory).

LMP theory was first advanced by David McClelland in 1975. McClelland argued that the following combination of nonconscious motives are generic to, and predictive of, leader effectiveness: high power motivation, high concern for the moral exercise of power, and power motivation greater than affiliative motivation. This combination of motives is referred to by McClelland (1975) as the Leader Motive Profile (LMP). Following is a brief description of the underlying rationale of LMP theory.

Power motivation is defined as a nonconscious concern for acquiring status and having an impact on others. According to LMP theory, the power motive is necessary for leaders to be effective because it induces them to engage in social influence behavior, and such behavior is required for effective leadership. Further, highly power-motivated individuals obtain more satisfaction from the exercise of influence, and this satisfaction sustains their interest in the exercise of leadership. Theoretically, if enacted in a socially constructive manner, high power motivation should result in effective managerial performance in middle and high-level positions (McClelland, 1975; 1985). However, unless constrained by a disposition to use power in a constructive manner, power-motivated managers will exercise power in an impetuously aggressive manner for self-aggrandizing purposes, to the detriment of their subordinates and organizations.

Accordingly, it is hypothesized that individuals who have a high concern for the moral exercise of power will use power in an altruistic and collectively-oriented manner, behave ethically, and be concerned about the consequences of their own actions on others (Winter & Barenbaum, 1985). The combination of high power motivation and high disposition toward the moral exercise of power should, thus, result in leadership which induces follower trust and respect for the leader and commitment to the leader's vision.

The power motive, according to LMP theory, needs to be higher than the affiliative motive. Affiliative motivation is defined as a nonconscious concern for establishing, maintaining, and restoring close personal relationships with others. Individuals with high affiliative motivation tend to be non-assertive, submissive, and dependent on others (McClelland, 1985). Theoretically, highly affiliative motivated managers are reluctant to monitor the behavior of subordinates, to convey negative feedback to subordinates even when required, or to discipline subordinates for ethical transgressions or violations of organizational policies. Highly affiliative motivated managers are also theoretically expected to manage on the basis of personal relationships with subordinates and, therefore, show favoritism toward some. House et al. (1991) found that affiliative motivation as inferred from presidential inaugural addresses was significantly inversely related to presidential charismatic leadership and archival measures of U.S. presidential effectiveness during presidential first terms. Further, House et al. (1996) found affiliative motivation, inferred from executive interviews, to be negatively related to subordinates' perceptions of executive charisma and fairness, as well as to indicators of organizational effectiveness.

When the power motive is higher than the affiliative motive, individuals do not engage in the dysfunctional behaviors usually associated with high affiliation motivation - submissiveness, reluctance to monitor and discipline subordinates, and favoritism. Finally, when the self-aggrandizing tendency usually associated with high power motivation is inhibited by a high concern for morally responsible exercise of power (or social influence), individuals are predicted to engage in the exercise of power in an effective and socially desirable manner. These theoretical expectations were confirmed in the study by House et al. (1996).

Theoretically, the leader motive profile is predictive of managerial effectiveness under conditions where leaders need to exercise social influence in the processes of making decisions and motivating others to accept and implement decisions. In formal organizations, these conditions are found at middle and higher organizational levels and in nontechnical functions. By contrast, in smaller technologically based organizations, where group leaders can rely on technological knowledge and direct contact with subordinates rather than delegation through multiple organizational levels to make decisions, the achievement motive is theoretically predictive of leaders' effectiveness. Thus, LMP theory is limited to the boundary conditions of moderate to large nontechnologically oriented organizations (McClelland, 1975; Winter, 1978, 1991), and to managers who are separated from the work of the organization by at least one organizational level.

Several studies have demonstrated support for the LMP theory. Winter (1978) found that LMP was predictive of the career success of entry-level managers in nontechnical positions in the U.S. Navy over an eight-year interval. Both McClelland and Boyatzis (1982) and Winter (1991), in separate analyses of the same data but with different operationalizations of LMP, found similar results at AT&T over a sixteen-year interval. McClelland and Burnham (1976) found that work units of high LMP managers had employees with a greater sense of responsibility and team spirit, and perceived more clarity of organizational demands than work units of low LMP managers. House et al. (1991) found that the motive components of the LMP individually predicted U.S. presidential charisma and presidential performance effectiveness.

More recently, House et al. (1996) found that LMP theory was most predictive of chief executive charismatic leader behavior and effectiveness in small entrepreneurial organizations. This finding is inconsistent with the boundary conditions specified for LMP theory - large, complex, nontechnical organizations. However, this finding is consistent with Mischel's argument that dispositions are most likely enacted under weak psychological conditions. In entrepreneurial organizations, chief executives are less constrained by formal organizational rules and policies than are managers of large organizations.

A major deficiency in this literature concerns the specific behaviors associated with LMP. Only the study by House et al. (1996) addresses this issue. These authors found that LMP is rather strongly associated with follower reports of their superiors as being charismatic, displaying integrity, and being supportive. In conclusion, the major research issues confronting LMP theory concerns the need to specify the boundary conditions under which the theory holds, and the need for a better understanding of the behavioral manifestations of the LMP motive syndrome.

Charismatic Leadership Theory. House (1977) speculated that charismatic leaders are exceptionally self-confident, are strongly motivated to attain and assert influence, and have strong conviction in the moral correctness of their beliefs. These personality traits are asserted to be antecedents to charismatic leadership and effectiveness. Since charismatic leaders advocate change and, thus, challenge the status quo, they are likely to be strongly resisted by defenders of the status quo who are in positions of substantial power. Also, since social change is difficult to accomplish, a great deal of determination and persistence is required on the part of change agents. The need to overcome these obstacles requires substantial risk-taking and perseverance. Leaders who are motivated to assert and exercise influence are theoretically expected to advocate change and challenge the status quo. Leaders who are exceptionally self-confident, and who have strong conviction,in the moral correctness of their beliefs, are theoretically expected to be more persistent in the face of obstacles and, therefore, to be more effective.

Simonton (1987) and House et al. (1991) found social influence motivation to be associated with U.S. presidential charismatic behavior and effectiveness. Three recent studies of charismatic leaders have revealed that such leaders are exceptionally high on self-confidence (House, Spangler & Woycke, 1991; Howell & Higgins, 1990; Simonton, 1987). Self-confidence is likely to be correlated with risk taking, as well as with assertiveness and prosocial dominance mentioned above. Simonton (1994) and Spangler and House (1991), in different operationalizations of the same archival data, found U.S. presidents' risk-taking behavior to be predictive of presidential greatness and effectiveness, respectively. Strength of conviction remains to be tested as a predictor of charismatic or otherwise outstanding leadership.

More recently, the 1976 version of charismatic leadership theory (House, 1977) has been expanded to include LMP as an antecedent to charismatic leadership emergence and effectiveness (House et al., 1991). It is argued that charismatic leaders, to be effective, must mobilize a critical mass of followers in the interest of the leader's vision; thus they need to have high power motivation. Since such leaders will almost inevitably be resisted and criticized, they need to be relatively insensitive to such criticism and, thus, have low affiliative motivation. Finally, to be effective and to maintain their position, charismatic leaders must advocate a vision of a better future for the collective (social system or organization) and for followers. They must not exercise leadership in the interest of self-aggrandizement. According to charismatic theory, the emergence and effectiveness of charismatic leaders will be associated with leaders' sense of social responsibility and collective interests rather than with self-interest.

Leader Flexibility. Kenny and his associates have introduced two additional constructs into the leadership trait literature. Studies by Kenny and his associates (Kenny & Hallmark, 1992; Kenny & Zaccaro, 1983; Zaccaro, Foti & Kenny, 1991) have demonstrated the importance of leadership flexibility and social sensitivity for the leadership emergence process. Kenny and Zaccaro (1983) re-analyzed data reported by Barnlund (1962) concerning the emergence of leaders who had supervised multiple groups over an extended period of time. The tasks of these groups varied so that each one required a different constellation of skills. Kenny and Zaccaro found that between 49% and 82% of the variance in leadership emergence could be attributed to a stable, but unidentified, characteristic of emergent leaders. They speculated that the underlying basis for this characteristic may involve behavioral flexibility and social sensitivity. This speculation received support in a subsequent study conducted by Zaccaro et al. (1991). Their investigations revealed that 59% of the variance in leadership emergence was due to leader behavioral flexibility and social perceptiveness, as indicated by subjects' scores on a measure of self-monitoring. Earlier, Simonton (1987) had found that, in the U.S., presidential flexibility moderated the relationship between presidential propensity to exploit veto power and electoral mandate. Flexible presidents exploited their mandates by use of their veto power, whereas inflexible presidents did not. Further, for flexible incumbents their level of congressional support did not affect their use of veto. The use of veto by inflexible presidents was significantly influenced by congressional support, thus making them vulnerable to the number of allies they had in Congress.

Summary of Findings from Trait Research

Three salient points emerge from trait theory and research to date.

First, there appear to be a number of traits that consistently differentiate leaders from others. These are physical energy, intelligence greater than the average intelligence of followers led, (Simonton, 1994; Stogdill, 1974), prosocial influence motivation as measured by the Dominance scale of the California Personality Inventory (House & Baetz, 1979), adjustment, self-confidence, achievement motivation, and the motives of the leader motive profile. Flexibility and adjustment may also be considered for addition to this list, pending further empirical verification.

Second, the effects of traits on leader behavior and leader effectiveness are enhanced to a great extent by the relevance of the traits to the situation in which the leader functions. Achievement motivation is most predictive of effectiveness and success when tasks are challenging, require a high degree of initiative, and require the assumption of personal responsibility for success. Leader flexibility is likely to be most predictive of leader effectiveness when leaders function in unstable environments, or when leaders are required to lead different people performing different tasks over time. This speculation remains to be tested empirically.

Measures indicating high social influence motivation such as the power motive, or prosocial assertiveness motivation such as the Dominance scale of the California Personality Inventory, are predictive of leader effectiveness in complex organizations in which decision implementation requires a high degree of persuasion and social influence. An important challenge to leadership scholars is now to determine the role requirements of the leaders studied and assess the effect of specifically relevant motives to the successful fulfillment of these requirements.

Third, traits have a stronger influence on leader behaviors when the situational characteristics permit the expression of individual dispositions. Such situations are termed "weak" situations, and those which suppress the expression of individual dispositions are termed "strong" situations. Thus, the behavioral manifestation of traits is stronger in weak situations and weaker in strong situations.

These conclusions need to be qualified with the caveat that they apply to leaders of task-oriented work units or task-oriented organizations, and that they are based predominantly on the study of American males. Future research may show that the traits that differentiate female leaders, or leaders in other cultures, from nonleaders are at variance with these conclusions.

Contemporary research on intelligence offers renewed potential for leadership trait research. The notion of multiple intelligences (Gardner, 1983; 1995) and Sternberg's (1988) theory of triarchic intelligence have implications for managerial roles. Leadership is embedded in a social context, and the idea of social intelligence as a required leadership trait is a powerful one. The emerging theories of intelligence imply that differences in cognitive abilities between leaders and nonleaders go beyond conventional IQ measures. Standardized measures of these newly postulated domains of intelligence will be instrumental in generating new areas for trait research in general, as well as for specific theories such as Social Information Processing theory.

Leaders: Born or Made?

A natural implication of the trait approach to leadership is the notion that leaders are to a substantial extent born, not made. This raises the question as to whether personal abilities and traits relevant to the effective exercise of leadership are genetically influenced.

Among the most influential studies in heritability of personality traits is the continuing Minnesota Study of Twins Reared Apart, begun in 1979. In one recent report, Bouchard, Lykken, McGue, Segal, and Tellegen (1990) report heritability coefficients of as much as 78% for the g-factor in mental ability, and a mean of 50% for a wide variety of personality traits measured by the Multidimensional Personality Questionnaire and the California Personality Inventory, among other variables. These coefficients express the amount of variance attributable to heritability. They imply that, at least to the extent that these traits are correlated with effective leader behavior, the potential for leadership may be genetically influenced. Included in the traits measured are several that are similar to the ones described in the above section such as adjustment, achievement and affiliative orientation, desire for social influence, tendency to behave in a socialized or responsible manner, assertiveness, interpersonal sensitivity, and interpersonal flexibility.

There is a potential problem with twin studies, however. The environments provided to monozygotic (MZA) twins reared apart are presumed to be different. The heritability coefficients are obtained as a direct estimation from the MZA Twins Reared Apart correlation, under the assumption that for twins randomly placed for adoption in early infancy, variance in personalities is due to genetic heritage, rather than shared environmental experience. In fact, however, some of the subjects recruited for the Minnesota study are twins who approached the Minnesota Twins Family Study Center, having learned of the ongoing research. At the time of their participation in the study, such twins may have been in contact with each other from anywhere between a week to 20 years (Bouchard et al., 1990). This raises the possibility that some of the twins could have developed similar interests as a result of coming together, even if they had been raised in different environments. Bouchard et al. state that "Such marked behavioral similarities between reared apart MZA twins raise the question of correlated placement: were the twins' adoptive homes selected to be similar in trait-relevant features which, in turn, induced psychological similarity?" (italics ours). Thus, it is possible that at least some of the paired twins share common experiences which, in turn, account for a nontrivial proportion of their shared variance in trait scores. Lykken (personal communication) reports that the amount of co-variance in twins' environments, measured on trait-relevant features, has been found by the Minnesota twin study researchers to range from .04 to. 10, indicating that the extent of environmental similarity has a very modest effect on the overall findings of their research. With more sophisticated control over extraneous and environmental variables, twin studies offer the potential to substantially increase our understanding of the origins of leadership.

There are several other theories that enjoy, or have enjoyed, currency in the leadership literature. These are theories of the leader behavior paradigm, theories of the contingency paradigm, theories of the neocharismatic leadership paradigm, Leader-Member Exchange theory, and the Social Information Processing Theory. These are discussed in turn in the next several sections.

The Leader Behavior Paradigm

Following the disenchantment with traits, there ensued a period of almost thirty years during which leaders where studied either by observing their behavior in laboratory settings or by asking individuals in field settings to describe the behavior of individuals in positions of authority, and relating these descriptions to various criteria of leader effectiveness. Three influential groups of investigators pursued the quest for explanations of leader effectiveness in this manner. These were Robert Bales and his associates at Harvard (Bales, 1954), members of the Ohio State Leadership Center (Stogdill & Coons, 1957), and members of the Institute for Social Research at the University of Michigan (Kahn & Katz, 1953; Likert, 1961; Mann, 1965).

Research conducted within this paradigm became known as the behavioral school of leadership. One of the major empirical contributions from the behavioral school was the identification of two broad classes of leader behaviors - task-oriented and person-oriented behaviors - which were identified by repeated factor analyses conducted by the Ohio State group, interviews by the Michigan group, and observation of emergent leaders in laboratories by the Harvard group. It should be noted that the Harvard group also identified a third dimension, individual prominence, which was somehow ignored in subsequent leadership literature. We speculate that this dimension was neglected because of the social disapproval of individual prominence-seeking. It may also be that such prominence-seeking behavior was a reflection of power motivation of the kind described earlier.

A second major contribution of the behavioral paradigm was a more refined and detailed specification of task- and person-oriented behaviors. For a good summarization of the behaviors conceptualized by various investigators, the reader is referred to Bowers and Seashore (1966). Unfortunately, there was no pattern of leader behavior which was found to be consistently associated with subordinates' satisfaction or any criteria of supervisor or manager effectiveness (House, 1971; Larson, Hunt & Osborn, 1974).

Assumptions and Limitations of the Leader Behavior Paradigm

Research conducted within the leader behavior paradigm shares several similarities with early research on leader traits. This research was based almost exclusively on observations of individuals who functioned at lower organizational levels and whose roles primarily concerned supervision, or observations of university students in laboratories, rather than observations of higher-level leaders responsible for the functioning of entire organizations. Behavioral studies were frequently based on questionnaires that sought to elicit subordinates' recall of the behavior of their superiors, presumably reflecting global historical patterns of behavior and relationships between leaders and followers, as well as specific recently enacted behaviors.

Like trait research, the research of the behavioral school was largely inductive and lacked theoretical orientation, since basic theoretical concepts had not been well developed at the time. This school was also plagued by limitations of measurement. Many of the leader behavior questionnaires were of questionable validity. For example, the scales most frequently used to measure the leader task and person orientation constructs were the leader Initiating Structure and Consideration scales developed by the Ohio State investigators. The Leader Initiating Structure and Consideration constructs were measured with several different scales that were subsequently shown to measure substantially different specific leader behaviors and to correlate differentially with various criterion variables (Schriesheim, House & Kerr, 1976).

The initial guiding assumption of the behavioral paradigm was that there are some universally effective leader behaviors, and these could be discovered by either observing leaders in action, usually in a laboratory setting, or by asking subordinates about the behavior of their immediate superiors. As with trait research, little thought was given to the specific role demands of leaders, the context in which they functioned, or differences in dispositions of leaders or followers. Failure to consider these factors was subsequently thought to be the reason for the researchers' inability to identify leader behaviors that had universal or near universal effectiveness.

Contingency Theories

Five theories were advanced to reconcile differences among the findings concerning leader behavior. These were Fiedler's Contingency Theory of Leadership (Fiedler, 1967; 1971), the Path-Goal Theory of Leader Effectiveness (House, 1971; House & Mitchell, 1974), Hersey and Blanchard's (1982) life cycle theory, the Cognitive Resource Theory (Fiedler & Garcia, 1987), and the Decision Process Theory (Vroom & Yetton, 1973).

Fiedler's Contingency Theory

This theory was the first to specify how situational variables interact with leader personality and behavior. Fiedler's Contingency Theory posited a two-way interaction between a measure of leader task-motivation versus relationship motivation, and a measure of what was initially referred to as "situational favorableness" and later relabeled "situational control" (Fiedler, personal communication, September, 1996). Situational control is the degree to which the leader can control and influence the group process. Eight conditions (octants) of situational control were defined, and hypotheses concerning their moderating effects on relationships between leader motivation and effectiveness were proposed. Approximately 200 tests of the hypotheses generated by the theory were conducted. As predicted, task- or relationship-motivated people were shown to manifest the same behavior (e.g., consideration, structuring) under selected different conditions of situational control, and different behaviors under the same conditions of situational control.

Despite its ground-breaking nature, the Contingency Theory was criticized both for conceptual reasons (Schriesheim & Kerr, 1977; also see Fiedler, 1977 for a rejoinder) and because of inconsistent empirical findings and inability to account for substantial variance in group performance (Ashour, 1973; also see Fiedler, 1973 for a rejoinder to Ashour). It should be noted for the record, however, that partial but substantial support for the Contingency Theory was indicated by two extensive meta-analyses (Peters, Hartke & Pohlman, 1985; Strube & Garcia, 1981).

Strube and Garcia found overall support for the Contingency Theory based on 33 tests that were used by Fiedler (1967) to develop the model and 145 subsequent tests of the validity of the model. However, these authors failed to find support for several octant-specific predictions. Further, the accuracy of Strube and Garcia's conclusions has been challenged by Vecchio (1983) on the basis of faulty selection of studies and inappropriate statistical analysis. He concluded that when corrections are made for the methods used by Strube and Garcia, the empirical evidence does not support the Contingency Theory. In response, Stube and Garcia (1983) have contested Vecchio's conclusions.

Peters et al. (1985) meta-analyzed the empirical data relevant to the Contingency Theory using data from studies based on within-octant interacting groups. Peters et al. found that laboratory studies showed overall support for the model with the exception of one octant, and that field studies fell short of supporting the overall model and predictions for five of the eight octants. However, Fiedler (personal communication, May 1997) states that his contingency hypothesis has always been based on three zones of situational control, rather than the eight octants. He hypothesized that task-motivated leaders perform best in situations of high and low control while relationship-motivated leaders perform best in moderate control situations. This hypothesis has been amply supported by the meta-analyses.

Path-Goal Theory

Path-Goal Theory (House, 1971; House & Mitchell, 1974) was also intended to reconcile prior conflicting findings concerning task- and person-oriented leader behavior. The theory specified a number of situational moderators of relationships between task- and person-oriented leadership and their effects. While initially promising, this theory, when tested empirically, met with mixed results. Wofford and Liska's (1993) meta-analysis of the results of 120 tests of Path-Goal Theory hypotheses showed that support for the theory was significantly greater than chance. However, the overall results were quite mixed and disappointing.

Recent reviewers of the history of Path-Goal theory have all concluded that it has not been adequately tested (Evans, 1996; Schriesheim & Nieder, 1996; Yukl, 1993). This is perhaps because it is a complex theory that specifies four leader behaviors, a number of situational and follower trait moderators, five intervening variables (follower expectancies and valences), and two dependent variables (follower satisfaction and performance). Structural modeling might be used to examine the overall fit of the theory to empirical data, controlling for extraneous variables that might affect follower satisfaction and performance. Controlled laboratory investigation would be useful in rigorously testing specific propositions of the theory. Since adequate leader behavior measures were not available and latent structure modeling had not been introduced into the organizational behavior literature in the early 1970s, tests of the overall empirical fit of the theory were inadequate.

House (1996) subsequently recognized that one of the boundary conditions for Path-Goal Theory is that followers must be able to make relatively confident and accurate(1) estimates of probabilities of goal accomplishment (performance) and receipt of extrinsic outcomes contingent on such accomplishment. Thus, the theory has a strong rationality bias. Under conditions of nonreducible uncertainty with respect to effort requirements, goals, or extrinsic rewards, or when followers or leaders are under a substantial amount of stress and, therefore, not likely to make rational and accurate expectancy estimates, the theory is not likely to hold. Thus, the absence of stress and nonreducible uncertainty constitute boundary conditions of Path-Goal Theory.

Life Cycle Theory

Hersey and Blanchard (1982) have also developed a situational leadership theory of leadership. They postulated four leadership styles: telling, selling, participating, and delegating, each appropriate for certain kinds of situations defined by subordinates' "maturity" level. That is, the prescribed leadership style is contingent on follower maturity, defined as "the degree to which followers are ready and willing to tackle the task facing the group." Hersey and Blanchard see their theory as representing a life-cycle model, analogous to a parent-child relationship where the parent gradually relinquishes control as the child matures. This model has a high degree of face validity and is the foundation for a commercial management training program sold by Hersey and Blanchard. To our knowledge, there have been few empirical tests of the theory. Vecchio (1987) found in one empirical investigation involving 303 teachers in high schools that the theory may hold only for certain types of employees. More recently employed subordinates needed and appreciated more task structuring from their leader. Given its practical implications, especially in light of Vecchio's (1987) findings, and the fact that it is the foundation of a commercially distributed training program, it seems to us that it is professionally incumbent on the authors to provide evidence relevant to the validity of the model.

Cognitive Resource Theory

Despite the criticisms of the Contingency Theory of Leadership, Fiedler persisted in his search for a better explanation of the leadership phenomena. As a consequence of Fiedler's interest in situational control, he investigated the effect of situationally induced stress on leaders and followers, one form of situational unfavorableness. As a result, his investigations led to the development of the Cognitive Resource Theory (CRT) of Leadership (Fiedler & Garcia, 1987). This is a person-by-situation interaction theory in which the person variables are leader intelligence and experience, and the situational variable is stress experienced by leaders and followers. This theory enjoys considerable empirical support (Fiedler, 1995).

One of the most important findings resulting from the research conducted by Fiedler and his associates is that, under low stress, intelligence is positively correlated, and experience negatively correlated, with performance. In contrast, under high stress, intelligence is negatively correlated with performance, and experience positively correlated. That is, when subordinates report high job- or boss-related stress, bright people perform worse than dull people. When job- or boss-related stress is low, more experienced individuals perform worse than do less experienced individuals. This implies that under conditions of high stress a highly intelligent person should rely on experience, rather than intelligence, to be effective. Yet, such a person will be inclined to apply both experience and intelligence and will, therefore, be less effective. Intelligence and experience, thus, interfere with each other. These findings are not only counter-intuitive, but also empirically supported in a substantial number of both field and laboratory studies (Fiedler, 1995).

Fiedler (1996, p. 247) states that "An oversimplified explanation of this interference effect is that we cannot think logically and analytically while at the same time reacting to emergencies and stress on the basis of over-learned previous knowledge and behavior, i.e., experience. Nor can a team carefully consider all the options and alternatives to solving a problem when their highly experienced leader tells them that 'we have gone over all of these arguments before and don't need still another study.'" Stress is the enemy of rationality (Simon, 1987). It narrows one's focus of attention and search for alternatives, causes rigidity (Staw, Sandelands & Dutton, 1981), and leads to authoritarian decision making, as well as increasing desire for and dependence on authoritarian leaders (Sales, 1972).

Cognitive Resource Theory helps to answer one of the most important questions in the leadership literature, namely, when is it more effective to be participative with followers, and when is it more effective to be directive? This question had been raised by several leadership scholars (Tannenbaum & Schmidt, 1973; Vroom & Yetton, 1973), and an enormous amount of research has been devoted to answering it (see Filley, House & Kerr, 1976, and Vroom & Jago, 1988, for reviews). The Cognitive Resource Theory contributes to a resolution of this issue.

According to the theory, leader intelligence cannot contribute to group performance unless the leader tells the group what to do, and the group members listen to the leader and do what they are told to do. Thus, under poor leader-follower relationships, directive leadership will be effective only when the leader has total control of the behavior of followers, and either a) the leaders are not under stress and can use relevant aspects of their intelligence, or b) leaders are under stress, but can use relevant experience. When the leader has little control over the behavior of followers and leader-follower relationships are troublesome, neither directive nor participative leadership will be effective - followers will neither listen to the leader, nor do what they are told to do. When leader intelligence is lacking in low stress conditions or leader experience is lacking in high stress conditions, directive leadership will be ineffective.

Further, when leader-follower relationships are good, and stress is low, participative leadership will work best when group members are more intelligent than their leader. This occurs because, relationships being good, the leader will listen to the followers. When leader-follower relationships are good, and stress is high, participative leadership will work best when group members are more experienced than their leader and the leader listens to them. Finally, when leader-follower relationships are poor, participative leadership will be ineffective because neither leader nor followers will listen to each other.

The Cognitive Resource Theory has important implications for leader selection and for situational management. Fiedler (1996, p. 248) recommends a two-step process for effective utilization of leaders: (1) recruiting and selecting individuals with required intellectual abilities, experience, and job-relevant knowledge, and (2) enabling leaders to work under conditions that allow them to make effective use of the cognitive resources for which they were hired. Fiedler (1996) reports that an experimentally administered stress reduction program increased the performance of officer candidates on an in-basket management simulation task which required intelligence rather than experience for effective performance. Stress reduction presumably increased the use of intelligence.

For inherently stressful tasks such as firefighting, or combat performance, overlearning is recommended. For stressful jobs which require both experience and intelligence, such as directing air traffic from control towers, overlearning and stress reduction procedures as well as training in coping with stress are recommended.

The above recommendations offer new directions for research. More specifically, there is a need for development and validation of selected overlearning training programs, assessment and selection procedures, stress reduction programs, superior-subordinate conflict resolution programs, situational engineering procedures, and placement programs directed at stress reduction and person-situation fit.

Decision Process Theory

Another situational theory of leadership was advanced by Vroom and Yetton (1973) and reformulated by Vroom and Jago (1988). The theory is intended to help managers make decisions that ensure high technical and economic quality solutions to problems and obtain solutions that are acceptable to subordinates, if acceptability of solutions is important for effective implementation.

The original theory (Vroom & Yetton, 1973) described seven decision-making methods which they believed to have different outcomes under different situations. These methods range from autocratic through democratic decision processes. Five methods are relevant to group decision making, and include two authoritarian processes (denoted AI and AII); two consultative processes: consultation with subordinates individually (CI) and consultation with subordinates as a group (CII); and a group process of joint decision making (GII). Of these, three (AI, AII and CI) are also relevant to individual decision making, as are two additional methods: GI, joint decision making by superior and subordinate, and DI, delegation of the decision.

Based on a review of prior literature, in addition to their own theorizing, Vroom and Yetton listed seven properties of problems they believed were relevant to decision making. They also developed seven corresponding decision rules intended to guide a leader in selecting the most appropriate decision method. The combination of five decision processes, seven problem attributes, and seven rules constituted the variables in the original normative theory. A series of relationships among these variables, cast in Boolean algebraic equations, constitute the theory.

The theory is operationalized as a prescriptive model in the form of a decision tree. The decision rules, recast in the form of seven questions answered sequentially, are designed to help managers determine appropriate decision processes. The answers to the questions permit the application of the decision tree. When the seven questions are answered, a theoretically feasible set of decision processes is identified. Fourteen possible feasible sets are specified, contingent on the presence of the seven situational variables. A manager may choose among the decisions in the feasible set and be theoretically assured that the decision will result in high decision economic and technical quality, and high acceptance by subordinates when such acceptance is theoretically required for effective implementations. Vroom and Yetton refer to each feasible set as a problem type. Thus, the theory includes fourteen problem types. The number of decisions in the feasible sets range from one to five.

Validity of the Prescriptive Model. To our knowledge, nine studies have empirically tested the Vroom-Yetton model. Margerison and Glube (1979) studied 47 owner-operators of cleaning franchises in the U.S. and Canada. The owner-operators responded to vignettes by indicating the decision processes they would use to solve the problems described in the vignettes. Respondents who conformed most closely to the Vroom-Yetton model were found to have more profitable operations and more satisfied employees. In four of these studies (Bhohnisch, Jago & Reber, 1987; Tjosvold, Wedley & Field, 1986; Vroom & Jago, 1988; Zimmer, 1978), managers were asked to recall past decisions, describing various aspects of the problem and decision outcomes. The median proportion of successful decisions reported in the four studies was 67% of all decisions by managers whose decision processes were consistent with the model's prescriptions, and 33% for managers who used decision processes inconsistent with that suggested by the theory. These figures imply that managers can double the success rate of their decisions by following the prescriptions of the model.

Two other studies (Field, 1982) examined the theory in a laboratory setting. Field (1982) examined the decision effectiveness of 276 business school students formed into groups to solve experimentally manipulated problems. He reported that the model was supported overall, although only four of the seven rules had the predicted effect. Specifically, decisions made through processes consistent with the model (i.e., within the feasible set) had higher effectiveness ratings than decisions made outside the feasible set, but the effect size was small (Cohen's d = .30). Stated differently, 49% of decisions made by processes in the feasible set and 34% of those by processes outside the feasible set were effective. However, after controlling for group membership and problem difficulty, decision processes were found to account for only 2% of the variance in decision effectiveness.

In the study by Liddell, Elsea, Parkinson and Hackett (1986, unpublished; reported in Vroom & Jago, 1988, p. 80), 54% of effective decisions were in the feasible set, and 29% were outside the feasible set. Liddell et al. did not control for differences in decision-making groups and problem difficulty. The finding by Field, referred to above, suggests that the effect of decision processes on decision effectiveness is likely to be vastly overestimated by Liddell, et al.

Overall, the theory has been supported more strongly in field studies than in laboratory studies. Both types of study, however, had their limitations. Laboratory experiments suffer from limited mundane and experimental realism (Aronson & Carlsmith, 1968). In the studies described, the laboratory setting offered no real consequence to the participants, possibly making the experimental manipulation of goal congruence or the need for subordinate acceptance less effective. On the other hand, the field studies were susceptible to bias from correlated observations, as all the variables involved had been measured using reports from the same source, namely, managers. Further, the field studies reviewed above did not control for potentially confounding variables, such as the amount of stress experienced by leaders and followers, problem difficulty, or leader or follower intelligence and experience. Since these variables have been found in prior studies to account for substantial variance in decision effectiveness (Fiedler & Garcia, 1987; Field, 1982), the effect of decision process reported in the field studies reviewed thus far is likely to be much higher than is actually the case.

Further, the field studies used managers' self-reports. Self-report data have been shown in prior research to be biased, that is, they tend to differ from data based on direct observations (Bass, 1957; Besco & Lawshe, 1959; Campbell, 1956; Graham & Gleno, 1970). Given that: (1) the Vroom-Yetton theory is a highly rational theory, and (2) the managers in the field studies can be expected to present themselves as highly rational, a strong rationality bias based on social desirability may be expected in the results.

An appropriate test of the model would be a comparison between the model's prescribed behavior and independently observed (as opposed to self-reported) behavior of decision makers and attributes of problems. Indeed, Vroom and Yetton state that they had to abandon such a design due to practical difficulties. Subsequently, two studies (Field & House, 1990; Heilman, Hornstein, Cage & Herschlack, 1984), achieved this by obtaining information collected from both decision makers and subordinates.

In the Field and House (1990) study, the Vroom-Yetton model was tested using both managers' and subordinates' recall of decision processes, attributes, and decision effectiveness. While the data obtained from the managers supported the model, those obtained from the subordinates did not. The opposing results were primarily due to differences in ratings of decision effectiveness and ratings of subordinates' acceptance required between managers and their subordinates.

Heilman et al. (1984) demonstrated the same phenomenon by assigning subjects to the roles of subordinate and superior in an experimental manipulation in a laboratory setting. Subordinates' ratings of decision effectiveness disconfirmed the model, while managers' ratings tended to confirm it. Similar results were obtained with subordinates' and managers' ratings of subordinate commitment.

The Reformulated Theory. Vroom and Jago (1988) noted the several limitations to the Vroom-Yetton model. They reformulated the theory adding a new criterion variable entitled "overall effectiveness." This variable is composed of decision quality, decision commitment (which replaces decision acceptance) and two additional criteria of decision effectiveness: the effect of the decision process on subordinate development and time available to make the decision, assumed to be an indicator of decision costs. The reformulated theory is expressed in mathematical functions that express the decision criteria as functions of the decision process and problem attributes.

Five additional decision rules are incorporated into the new theory. There are twelve problem attributes, twelve decision rules and eight problem types, each representing one of the eight combinations of individual and group decision outcome criteria: decision quality, decision commitment, minimum time required, and maximum subordinate development. Information on the twelve problem attributes, along with specification of the decision criteria, will enable one to identify the appropriate decision process according to eight normative models (sets of decision rules operationalized in the form of questions that are applied to a decision tree). There is one normative model, i.e., one set of questions for each decision criterion. Since there are numerous possible combinations of problem attributes, desired criteria and decision rules, the task of applying the model to practical situations can prove quite unwieldy. To facilitate application, the models have been computerized so that managers can operationalize them as expert systems to help them adopt rationally appropriate decision processes for specific problems.

The validity of two of the models (the decision quality and decision acceptance models) has been assessed in a study by Jago, Ettling, and Vroom (1985; reported in Vroom & Jago, 1988, p. 179). In this study, the investigators obtained correlation coefficients between the predictions of the models and the actual outcomes resulting from 80 decisions as follows: .38 for decision quality, .84 for subordinate's commitment, and .68 for decision effectiveness (quality multiplied by acceptance). (A correlation of 1.00 would indicate that the model could perfectly predict, without exception or error, the numerical outcomes of all 80 decisions). These results strongly support the model. However, unlike the study by Field (1982), this study employed no controls for group differences or problem difficulty. Since Field demonstrated that such controls dramatically reduce the amount of variance in decision effectiveness attributable to differences in decision processes, the findings by Jago et al. (1985) likely vastly overestimate support for the models.

As Vroom and Jago (1988, p. 179) note, "The experiment tested the predictive ability of the revised model in only a single situation, one of a million and a half combinations of possible relevant attributes. Complete confirmation of the model (theory) will certainly require years of testing."

It seems appropriate at this time to record some fundamental criticisms of the original Vroom-Yetton model by Field (1979), which are equally applicable to the reformulated model. These are: (1) that the theory assumes that the decision makers' goals are always congruent with the goal of the organization; (2) that since training is required in order to use the theory reliably, the population that may use it prescriptively is limited; (3) that the theory ignores the discussion and conference skills required of the manager to actually solve problems on a group decision; and (4) that the theory is excessively complex.

We add two additional observations. First, neither the underlying theory nor the models include consideration of manager or subordinate stress, intelligence, or experience. Since the importance of these variables has been established by investigations relevant to Cognitive Resource Theory, their omission from the Vroom and Jago theory is a serious one. Second, since the theory allows a million and a half combinations of possible relevant attributes, the theory appears to be untestable. As with the Hersey and Blanchard Life Cycle Theory, given the practical implications of the theory, and the fact that it is the foundation of a commercially distributed training program, it seems to us that it is professionally incumbent on the authors, or the distributors of the training program, to provide substantially more evidence relevant to the validity of the models. This will almost certainly entail making the theory less complex in order to be testable, or at minimum, identifying and testing a limited set of crucial aspects of the theory. Clearly, more theoretical work is needed to incorporate stress as a situational attribution and to make the theory more parsimonious; further studies are needed to establish the practical validity of the theory.

Field (1982) suggests that a theory with four situation attributes instead of seven, and two decision processes rather than five, will equally predict decision quality and acceptance. His theory is summarized briefly in the following set of decision rules, using terms found in Vroom and Jago's model:

If acceptance of the decision by subordinates is critical to effective implementation and it is not reasonably certain that subordinates would accept an autocratic decision, but they share organizational goals (or decision quality is not important) use GII; otherwise, use CII.

Cumulative Contribution of Contingency Theories

While some of the major predictions of the Contingency and Path-Goal theories were supported by the meta-analyses (Peters, Hartke & Pohlman, 1985; Strube & Garcia, 1981; Wofford & Liska, 1993), the theories did not fare well overall, and the interest of leadership scholars in these theories waned. However, both of these theories eventually led to the development of better theories - better in that the later theories explained, described, and predicted important leadership related phenomena and now enjoy considerably more empirical support.

Specifically, the Contingency Theory led to the development of the Cognitive Resource Theory of Leadership (Fiedler & Garcia, 1987), which has enjoyed considerably more support than any of the other behavioral theories.

The Path-Goal Theory led to the development of the 1976 Theory of Charismatic Leadership (House, 1977) and the 1996 version of the Path-Goal Theory. The 1976 theory has been supported in several studies conducted in both laboratory and field settings under a wide variety of circumstances (Yukl, 1993). This theory has been considerably broadened in scope and is now referred to as Value Based Leadership Theory (House et al., 1996). (See House, 1996, for a discussion of how the original Path-Goal Theory led to the development of the 1976 Charismatic Theory and a description of the 1996 version of Path-Goal Theory). Value Based Leadership theory is discussed in more detail further on.

Whereas the original Path-Goal Theory was a theory of dyadic relationships between superiors and subordinates, the 1996 version is a theory of relationships between superiors and work unit effectiveness. This theory more clearly specifies boundary conditions for specific hypotheses. It specifies eight classes of leader behavior and situational contingencies that moderate the effect of these behaviors on work unit performance. It is also intended to explain and account for follower empowerment. Because of its recency, the 1996 version remains to be tested. It does, however, reconcile conflicting findings in the leadership literature and is consistent with all extant theoretical positions concerning leadership behavior and relationships between leaders and subordinates.

In similar fashion, the Contingency Theory led to the development of Cognitive Resource Theory. Fiedler (Personal communication, December, 1996) describes the relationship between the Contingency Theory and Cognitive Resource theory:

We were looking for an explanation why high LPC (meaning Least Preferred Co-worker score) leaders perform better in moderate-control and low LPC's in high or low-control situations. One likely hypothesis was that leaders are able to utilize their cognitive resources better when LPC and situational control are "in-match", since Chemers, Hayes, Rhodewalt, and Wysocki (1985) showed that leaders who are in-match have lower stress and anxiety scores and higher job satisfaction than leaders who were out of match. (Leaders are "in match" when their behavior is, according to the Contingency Theory, theoretically appropriate for the situations in which they function.)

In other words, when these managers were under stress their intellectual abilities were not used effectively, indicating that the prediction of leader performance was contingent on the absence of stress. In addition, CRT supported the hypothesis that stress results in less mature or previously dominant behavior by showing that under stress leaders fell back on knowledge and behavior learned from experience ....

Finally, in similar fashion, the Vroom-Jago Decision Process Theory is an elaboration and extension of the earlier Vroom-Yetton Theory.

Recently Introduced Theories

With the waning of scholarly interest in the early contingency theories because of ambiguous findings, several theories were introduced during the last decade and a half. Many of these are intended to explain different aspects of the leadership phenomena. We discuss these more recent theories in this section.

Leader Member Exchange Theory

Leader Member Exchange (LMX) theory is a theory about the development and effects of separate dyadic relationships between superiors and subordinates. This theory has the potential to be a theory of systems of dyadic relationships (Graen & Uhl-Bien, 1995), but further theoretical development is required for the fulfillment of this potential. The theory prescribes a high degree of mutual influence and obligation between superiors and subordinates, and asserts that such a relationship will result in several important positive outcomes such as lower turnover, and higher subordinate performance, citizenship behavior, satisfaction, and commitment. The distinguishing feature of LMX theory is the examination of relationships, as opposed to behavior or traits of either followers or leaders. Proponents of the theory argue that the quality of "mature" superior-subordinate dyadic relationships would be more predictive of positive organizational outcomes than traits or behavior of superiors.

A precursor to LMX theory was the Vertical Dyadic Linkage Theory (VDL) (Dansereau, Graen & Haga, 1975; Graen & Cashman, 1975). Central to the VDL theory is the notion of differentiated dyadic relationships of followers with the same leader, giving rise to in-groups and out-groups. A shift of theoretical emphasis to LMX is seen by Graen and Uhl-Bien (1995) as the second stage in the evolution of VDL model, focusing on the quality of the dyadic relationships and its effects on organizational outcomes. The authors also discuss stages in the development of the LMX theory. Stage three involves investigating and describing the development of high-quality LMX relationships. Stage four represents an integration of findings from the previous stages to explain group outcomes and network phenomena.

While it is almost tautological to say that good or effective leadership consists in part of good relationships between leaders and followers, there are several questions about such relationships to which answers are not intuitively obvious. The first question concerns the attributes of high-quality relationships. Graen and Uhl-Bien (1995, p. 223) state that "The critical question addressed by the theory is: What is the proper mix of relational characteristics to promote desired outcomes?" The second question concerns the antecedent conditions and leader behaviors that enhance high-quality relationships. The third concerns the effects of variance in the quality of relationships. Finally, there is the question of the process by which variance in the quality of relationships influences leaders and followers with respect to their affect, behavior, and organization-related performance.

LMX theory conceives of leader-follower relationships in terms of social exchanges in which both leaders and followers perform effectively in response to high-quality relationships with each other. In its present stage, the questions raised above, despite a substantial volume of literature on the topic, are in the process of being answered. The theory is, therefore, still in the making. However, the major contribution of LMX theory is that it has focused attention on the superior-subordinate relationship, and opens up substantial opportunity for further research. In discussing extant literature on LMX, we take the questions raised above as our point of departure.

The first question concerns the characteristics of the relationship itself. According to LMX theory, high-quality relationships are characterized by trust, respect, and mutual obligation, generating mutual loyalty and influence between superiors and subordinates, and wide latitude of discretion for subordinates.

It is not clear, however, that these are universal attributes of high-quality relationships. It may well be that what is considered a high-quality relationship varies among individuals. For example, the effects of a training program intended to train managers to increase the quality of LMX with their subordinates is revealing (Graen, Scandura & Graen, 1986). Graen et al. (1986) report the training had a positive effect on subordinates' performance and satisfaction only (emphasis ours) when subordinates' need strength was high (Graen et al., 1986, p. 484). These findings suggest that theoretically high-quality LMX relationships were not desired, or considered to be high-quality relationships, by subordinates with low growth need strength. Thus, it is likely that individual differences among subordinates constitute a boundary condition for the prescriptive implications of LMX. Such boundary conditions are yet to be specified. We conjecture that individual differences such as level of growth need strength, desire for autonomy, and low authoritarianism of subordinates (Vroom, 1960) will serve as boundary conditions for LMX theory. We believe this speculation deserves empirical testing.

The second question concerns the antecedents to high LMX superior-subordinate relationships. High-quality relationships may be influenced by a host of situational factors, follower attributes and behaviors, and leader behaviors. LMX theory does not specify these, but implies that the behavior of subordinates influences superiors to show support, delegate to subordinates a substantial amount of discretion in conducting their work, engage in open communication, and encourage mutual influence between themselves and their subordinates. Further, Vecchio (personal communication, December, 1996) states the following.

Where the theory is somewhat less clear is in explaining how the processes of increasingly better or poorer quality working relationships would be "initiated." How these relations are "sustained" has to do with social exchange dynamics that are grounded in either a basic "hired hand - formal employment contract" set of relational assumptions or an "inner circle" set of relational assumptions.

Early on, it was argued in this literature that subordinate performance was the key determinant of in/out status, and that the gradual classification of subordinates was the result of "role episodes." In essence, these are "tryouts" or tests of a new hire. For example, a supervisor asks a new hire to do something more than the "formal contract" calls for (such as staying late on Friday). The new hire's reaction (e.g., "Sure, glad to help," versus "Grumble, grumble") gives an indication of potential trustworthiness, loyalty, etc., and leads to more or less opportunities for responsibility, personal growth experiences, etc ....

One can not speak of prescribed leader behaviors in this model without also discussing subordinate reactions to "tryouts" and the interplay of leader and subordinate influence bases over time. In that sense, the model is very different from other views of leadership.

A specification of the attributes of high-quality LMX - trust, respect, openness, latitude of discretion - is as close as the theory comes to describing or prescribing specific leader behaviors. The theory implies that any leader behavior that has a positive effect on LMX quality will be effective. However, precisely what these behaviors are is not explicitly stated, as the appropriate leader behavior is dependent on anticipated subordinate response. As will be discussed in the concluding section of this article, substantially different behaviors may be required to develop high-quality relationships in cultures that do not emphasize individualistic and egalitarian values.

Several variables have been linked with LMX quality, and these studies can help to further develop LMX theory. For instance, Kinicki and Vecchio (1994) hypothesized, and found, significant correlations between two variables - locus of control and time-based pressure - and LMX (r = .32 and r = .48, respectively). They also found a significant inverse relationship between time pressure and the variability in LMX within the unit, suggesting that supervisor perception of high time-based pressure was associated with a tendency on the part of the supervisors to form uniformly high LMX relationships with the subordinates in their units. The correlational design of the study does not permit causal inferences, although their hypotheses were based on arguments for Locus of Control and time-based pressure as antecedents to LMX.

The third question raised above concerns the effects of variance in the quality of relationships Graen and Uhl-Bien (1995, p. 229) state that findings document "significant, positive relationships between quality of exchange (LMX) and many outcome variables of interest." They cite a number of studies that indicate that there is a consistent positive correlation between measures of relationship quality reported by subordinates and satisfaction, commitment, performance, and citizenship behavior of subordinates. However, closer scrutiny indicates that empirical findings relating LMX to dependent variables are mixed and less supportive of the theory than Graen and Uhl-Bien imply. For example, empirical findings relevant to subordinates' citizenship behavior have been overstated. One of the studies they cite (Scandura, Graen & Novak, 1986) addresses the relationship between LMX and decision influence, not LMX and citizenship behavior. Another citation is a review of empirical literature on citizenship behavior and not an empirical study (Podsakoff, MacKenzie & Hui, 1993). The remaining two studies conducted by Manogran and Conlon (1993) and Yammarino and Dubinsky (1992) did, indeed, find a positive relationship between quality of superior-subordinate relationships and citizenship behavior of subordinates.

In a study not cited by Graen and Uhl-Bein, Deluga (1994) used LMX as an incremental regressor in a hierarchical regression analysis and found evidence of an association between the quality of LMX and self-reported citizenship behaviors (organizationally desirable courtesy, conscientiousness, altruism and sportsmanship, but not civic virtue). Deluga is careful not to imply causal direction in the association, and also notes the limitation introduced by the self-report measures of the variables. Since LMX and citizenship behavior were both measured with subordinate self-report scales that were included in the same questionnaire, there is the likelihood of common method-common source bias being reflected in the findings. We can only speculate about the magnitude of the correlation if objective measures or independent sources of data had been used. However, in another study also not cited by Graen and Uhl-Bien (Wayne & Green, 1993), the effect of LMX on two citizenship behaviors was assessed and found to be quite modest. This study did not suffer from a variety of problems associated with common self-reports of LMX and citizenship behavior. LMX had a significant, but low, effect on altruism (r = .25), but no significant effect on compliance (r = .09).

Studies that have assessed the association between LMX and employee turnover are also mixed, with correlations ranging from .02 to -.44. (Ferris, 1985; Graen, Liden & Hoel, 1982; Vecchio, 1985; Vecchio, Griffeth & Hom, 1986; Vecchio & Norris, 1996). The largest correlations have been obtained by Graen, the major proponent of LMX. In contrast, Ferris (1985), Vecchio (1985), Vecchio, Griffeth, and Hom (1986), and Vecchio and Norris (1996), found insignificant correlations of -.19, .02, -.08, and -.17, respectively.

Several studies have investigated the association between LMX and performance (Duarte, Goodson & Klich, 1994; Dunegan, Duchon & Uhl-Bien, 1992; Graen, Novak & Sommerkamp, 1982; Vecchio & Gobdel, 1984; Vecchio, 1987). Correlations ranging from .02 to .33 have been reported in these studies. Thus, the association between LMX and performance is not reliable. Many of these studies used superior's ratings to measure performance. Such measures are likely to be correlated with measures of leader-member exchange quality since the superiors are themselves parties to the LMX relationship. Therefore, if quality of LMX is high, as perceived by subordinates, and if the superior-subordinate perceptions are mutual, there is a strong likelihood that superiors will both like subordinates and rate their performance as high due to this liking, rather than due to the subordinates' actual performance. For example, Vecchio (1977) found positive relationships between superiors' ratings of subordinates and familiarity of superiors with subordinates, and superiors' liking of subordinates.

Turban, Jones and Rozelle (1990) showed experimentally that supervisor liking positively influenced the supervisors' expectations of subordinates' perception of LMX. Liking also enhanced the amount of psychological support and encouragement superiors provided to subordinates. Finally, liking positively influenced superiors' evaluations of subordinates' performance. Liking did not significantly influence the objectively measured performance of subordinates. Thus, it seems reasonable to conclude that positive relationships between LMX and subjectively rated performance by superiors are at least partially due to common variance associated with supervisors liking of subordinates.

Vecchio and Gobdel (1984) found modest correlations between LMX and objective measures of performance (r = -.23 and -.25 respectively, for objective measures of frequency and magnitude of errors). Duarte, Goodson and Klich (1993) examined the influence of LMX on the relationship between employee performance measured objectively and the supervisor's rating of that performance. They found that poorly performing high LMX subordinates were given high performance ratings, regardless of their actual performance. The ratings of low LMX employees were consistent with their actual performance. The correlation between LMX and objective performance for their entire sample was .07.

Further, superiors' ratings of subordinates' performance have also been shown to be significantly associated with superior-subordinate demographic similarity (Antoinette, 1992; Judge & Ferris, 1993; Lagace, 1990) and perceived similarity of superiors and subordinates (Liden, Wayne & Stilwell, 1993). Interestingly Liden et al. (1993) did not find a significant relationship between demographic similarity and LMX quality, but did find perceived similarity to be associated with LMX quality. Also, ratings of LMX reflect subordinates' ability level (Wakabayashi & Graen, 1984), work unit time pressure (Kinicki & Vecchio, 1994), and liking (Liden, Wayne & Stilwell, 1993).

As Vecchio points out, when superiors' ratings are used to relate LMX to subordinates' performance, the measurement process may introduce correlated response errors. When objective measures are used, associations between LMX and performance are less reliable. A key issue is the potential problem of social reciprocity serving as an influence on ratings obtained from supervisor and subordinate dyads (Vecchio, 1982).

One conclusion that may be drawn from the above discussion is that both LMX and performance ratings may jointly reflect the influence of multiple biases. A second conclusion is that LMX relationships are as much a function of the characteristics and behavior of subordinates as the behavior of superiors. A third conclusion is that LMX is a better explanation of the development of superior-subordinate relationships (Graen & Scandura, 1987) than of the effects of leaders on followers. As such, LMX theory may be more accurately viewed as a theory of dyadic relationships and their subjective consequences, rather than a theory that focuses primarily on leadership.

Finally, there is the question of the process by which variance in the quality of relationships influences the leaders and followers with respect to their affect, their behavior, and followers' career progress. The LMX process is very likely to be as follows: For whatever reason - demographic or perceived similarity, familiarity, liking, reputation of subordinates, social reciprocity, subordinates' ability level, and/or prior performance - selected pairs of subordinates and superiors develop high-quality LMX. Superiors express positive attitudes such as trust and respect toward these subordinates. Superiors also express a desire for reciprocal influence with subordinates and imply that they expect a high-level of mutual support and loyalty.

According to our explanatory scenario, these explicit and implicit communications convey expectations of follower loyalty, commitment, mutual obligation and possibly mutual liking. This information induces a Pygmalion effect - a self-fulfilling prophesy (Eden, 1990) and social reciprocation (Vecchio, 1982). Selected subordinates demonstrate loyalty, commitment, mutual obligation, and possibly higher performance. Superiors, in turn, grow to like these subordinates and subsequently give them high performance ratings. These ratings then influence the subordinate's reputation, and often become a matter of record. The ratings may be used formally or informally in future selection, development, and promotion decisions. Consequently, subordinates with a history of high performance ratings become promoted to higher-level positions. Thus, a positive Pygmalion effect could well result not only in high subordinate satisfaction, commitment, citizenship behavior, and performance ratings by superiors, but also in enhanced career advancement.

We would judge this to be a good, or effective, process of subordinate development and promotion were it not for the possible adverse implications it might have for the development and career advancement of subordinates who are not demographically similar, familiar and well liked. The resulting quality of LMX will be high within the in-group and low in the out-group. This will likely result in disproportional allocation of organizational rewards to "in-group" members to the exclusion of "out-group" members.

The above scenario describes a naturally occurring process. It does not imply intentional favoritism, discrimination or bias toward selected individuals or minorities. However, it does explain one important process by which such bias can occur in organizations. The conclusion to be drawn from this discussion is that supervisors, managers, and Human Resource Management specialists need to be made aware of the potential biasing processes associated with high-quality LMX, and procedural checks and balances need to be applied to minimize such biases, if this is indeed possible. Otherwise, the development of high-quality LMX relations could result in organizationally dysfunctional consequences and discrimination against out-group subordinates.

One possible approach to minimizing selective bias that produces favored treatment toward some subordinates consists of training superiors to offer high-quality LMX relationships to all subordinates. The most convincing study concerning the effects of high-quality LMX on performance is a quasi field experiment in which supervisors were trained to improve the quality of LMX with their subordinates (Graen, Liden & Hoel, 1982). The initial step toward doing so advocated in the training program was for managers to make an offer to subordinates indicating that they, the managers, desired to have high-quality superior-subordinate relationships. The dependent variables in this study were subordinate company records of employee productivity and error rate. These were based on objective, quantified data.

Graen and Uhl-Bien (1995, p. 230) report that the experimental results showed that followers who accepted offers by the trained supervisors to develop high-quality LMX relationships improved their performance "dramatically." However, the training had an effect on subordinates' performance and satisfaction only when subordinates' need strength was high (Graen, Scandura & Graen, 1986, p. 484; emphasis ours).

Also, the training had a positive effect on the performance of subordinates whose initial report of LMX, prior to the training program of their superiors, was low (Scandura & Graen, 1984). Thus, the training had a remedial effect and improved subordinates' performance for a select portion of the individuals reporting to the trained managers - specifically high growth need strength subordinates with prior unfavorable LMX relationships with superiors.

The findings concerning the moderating effects of subordinates' growth need strength suggest that the kinds of leader behaviors required to foster high-quality LMX relationships will be different from one subordinate to another. This, of course, is consistent with VDL and LMX theory. Individuals with low growth need strength are likely to respond much less positively to superiors' attempts to enhance LMX, and likely to require different behavior of their superiors. We would also expect that individuals with low preferences for autonomy are likely to respond much less positively to delegation of discretion (Vroom, 1960) which, according to LMX Theory, is generally associated with high-quality LMX. The theory does not suggest appropriate leader behaviors for subordinates with low growth need strength, subordinates with low preferences for autonomy, or subordinates with adverse attitudes toward broad latitude of discretion.

It is not clear that the trained supervisors in the Graen et al. (1982) study did, in fact, create improved relations, of the kind characterized by LMX, with their subordinates. Since the subjects were aware that they were participating in an experiment, the performance improvement could have resulted from a Hawthorne effect - an effort by subordinates to fulfill the objectives of the program, independent of improved LMX. Further, since the subordinates of the supervisors in the experimental group had opportunities to interact with subordinates of supervisors in the control group, diffusion of treatment cannot be ruled out. If the workers from the two groups compared notes, the experimental manipulation could become evident, leading to enhanced differences in performance.

The improved performance could have also resulted from non-LMX-oriented behaviors on the part of the supervisors, such as closer supervision, increased directiveness, or increased pressure for productivity. This study warrants replication with additional controls to determine exactly what process accounts for the increased subordinate productivity among high growth need, low LMX subordinates.

To our knowledge, there has been less attention devoted to specific leader behaviors that foster high-quality relationships than to the effects of such relationships. The theory implies that the superiors need to be supportive, delegate to subordinates a substantial amount of discretion in conducting their work, engage in open communication, and encourage mutual influence between themselves and their subordinates. In this regard, the leader behaviors implied are those conceived by earlier researchers as person-oriented leader behaviors and general, rather than close, supervision (Kahn & Katz, 1953). There is little empirical evidence relevant to this implication, however.

Opportunities for Further Development of LMX Theory. VDL Theory makes an important contribution by drawing attention to the phenomenon of differentiated dyads in leader-follower relationships, and the importance of such relationships. As a theory, LMX is still developing. A theory of superior-subordinate relationships should explain the dynamics of relationship formation. Antecedents to high-quality, as well as low-quality, LMX need to be incorporated into the theory. Existing results of research probably suffer from a host of biases, including those to be expected from correlated observations and incomplete measures of critical constructs. On the plus side, a number of LMX studies have employed longitudinal designs. The potential for LMX to develop into a theory of relationships at higher-levels of aggregation is promising, and is already conceptualized by Graen and his associates (Graen & Uhl-Bien, 1995). Such development would benefit greatly from improved research designs and rigorous instrument validation procedures.

Implicit Leadership Theory

This theory was advanced by Robert Lord and his associates (Lord, Binning, Rush & Thomas, 1978; Lord, DeVader & Alliger, 1986; Lord, Foti & De Vader, 1984; Lord & Maher, 1991) and represents an important contribution to leadership research. Implicit Leadership Theory addresses the evaluations people make about leaders, and the cognitive processes underlying evaluations and perceptions of leadership. Lord and Maher (1991, p. 11) define leadership as the process of being perceived by others as a leader. This is a perspective that had not been explicitly addressed in other works on leadership. According to the theory, all specified leader behaviors would still not make an individual a leader unless that person is also perceived as a leader. From this point of view, then, leadership traits may be seen as important constructions of perceivers that help them make sense of social situations (Mischel, 1973).

According to this theory, leadership perceptions are formed through either deliberate and controlled inferential or automatic and spontaneous recognition-based processes (Lord & Maher, 1991). Once formed, such perceptions provide a cognitive framework for the evaluation of future behavior as well as for performance. Lord and Maher (1991, p. 63) argue that follower perceptions of leadership have implications at lower levels as well as at higher levels in the organization, taking care to distinguish between supervision and leadership at lower levels.

Lord et al. (1984) assert that leadership perceptions can be explained in terms of categorization theory. They argue that leadership perceptions form a number of hierarchically organized cognitive categories, each of which is represented by a prototype. These prototypes are formed through exposure to social events and interpersonal interactions. Prior knowledge about human behavior and underlying traits comprise implicit leadership theories. Subsequently, a person is categorized based on the prototype, in the observer's implicit theory, with which there is maximal fit of the observed person's behavior.

Lord and his associates have produced evidence in laboratory studies for the existence of leader prototypes. In a typical study, Lord et al. (1984, Study 3) gave college students short vignettes about a hypothetical manager, manipulating several behaviors in the description to create three experimental conditions - prototypical, neutral, and anti-prototypical behaviors. They, then, asked the students to rate the person described in the vignette on a number of dependent variables, such as the likelihood of engaging in each of the twenty-five behaviors, as well as on leadership, using a five-point scale. This study showed that the mean rating of leadership in the prototypical condition was the highest, followed by neutral and anti-prototypical conditions in that order. Level of prototypicality explained 53% of the variance in leadership perceptions (Lord et al., 1984).

This theoretical perspective also raises the possibility that there are some universally endorsed leader attributes and behaviors that comprise implicit leadership theories. Investigators in the GLOBE research program(2) are currently in the process of identifying the culturally endorsed implicit leadership theories (CILTs) in each of 60 cultures. The resulting knowledge will be helpful for understanding what is expected of leaders and how leadership is enacted cross culturally. We discuss the cross cultural implications of CILTs further on.

Besides the view of leadership as being dependent on follower perceptions, the categorization perspective has implications for organizational leadership. The research and theorizing by Lord and his associates suggest that different prototypes exist for leadership in different roles and contexts. Strategic leadership and supervisory leadership, discussed further on, can thus be better understood in terms of effective management of the prototype matching processes.

Future Directions for Implicit Leadership Theory. Some limitations of prior research need mention, however. One consideration concerns the ecological validity of the laboratory methods used by Lord and his associates. Cronshaw and Lord (1987) used videotaped materials, instead of vignettes, as stimuli to assess implicit theories. The results obtained were not as strong as when short, written vignettes were used (Lord & Maher, 1991, p. 38). If one were to hold videotaped stimuli to have more ecological validity, the weaker results raise further questions about the generalizability and applicability of findings based on vignettes.

The theoretical foundation of the Implicit Leadership Theory is strongly embedded in the information processing paradigm, which has typically involved the measurement of reaction time to distinguish between controlled and automatic information processing. Lack of reaction time measurement in the vignette study described above makes it difficult to draw valid conclusions about the cognitive processes underlying subjects' responses. It is not that the role of reaction time has gone unrecognized, however. Lord et al. (1984, Study 2) demonstrated a significant correlation between prototypicality ratings and reaction time (r = -.42). In the vignette study, which did not measure reaction time, the difference in the mean ratings of leadership across conditions supports the 'existence of prototypical leader behaviors in subjects' cognitive schemas but does not allow inferences about the matching process. If the third study reported by Lord et al. (1984) had employed time measurements of leadership ratings, differences across conditions would have provided evidence relevant to prototype matching. Several of the other studies reported by Lord and Maher (1991) have also not used time measurements.

With appropriate methodology and design, research on cognitive processes involved in the functioning of implicit leadership theories can be applied to study perceptions of leadership by followers in different situations. One relatively unexplored avenue of research concerns cultural universality of cognitive processes. For instance, do controlled and automatic processes for given situations hold for individualistic as well as collectivist cultures?

Neocharismatic Theory

A major paradigm shift occurred in the mid 1970s. The new paradigm consists of a number of leadership theories that are of a common genre. The theories of this paradigm include the 1976 Theory of Charismatic Leadership (House, 1977), the Theory of Transformational Leadership suggested by Burns (1978) and further developed and operationalized by Bass (1985), the Attributional Theory of Charismatic Leadership (Conger & Kanungo, 1987), the visionary theories advanced by Kousnes and Posner (1987) and Bennis and Nanus (1985), operationalized by Sashkin (1988), and extended by Nanus (1992), and the Value Based Theory of Leadership (House et al, 1996), which is an extended version of the 1976 Theory of Charismatic Leadership (House, 1977).

These theories are all of a common genre. Bryman (1993) refers to this class of theories as "the New Leadership theories." They have several common characteristics. First, they all attempt to explain how leaders are able to lead organizations to attain outstanding accomplishments such as the founding and growing of successful entrepreneurial firms, corporate turnarounds in the face of overwhelming competition, military victories in the face of superior forces, and leadership of successful social reform for independence from colonial rule or political tyranny. Second, the theories of this paradigm also attempt to explain how certain leaders are able to achieve extraordinary levels of follower motivation, admiration, respect, trust, commitment, dedication, loyalty, and performance. Third, they stress symbolic and emotionally appealing leader behaviors, such as visionary, frame alignment, empowering, role modeling, image building, exceptional, risk-taking, and supportive behaviors, as well as cognitively oriented behavior, such as adapting, showing versatility and environmental sensitivity, and intellectual stimulation. Finally, the leader effects specified in these theories include follower self-esteem, motive arousal and emotions, and identification with the leader's vision, values, and the collective, as well as the traditional dependent variables of earlier leadership theories: follower satisfaction and performance.

House and Shamir (1993) argue that the leader behaviors specified above, with the exception of supporting, adapting, versatility exhibiting and environmentally sensitive behaviors, constitute a charismatic leadership syndrome, and provide a theoretical rationale for this syndrome. Support for the theoretical main effects of many of the leader behaviors of the charismatic syndrome has been demonstrated at several levels of analysis, including dyads (Howell & Frost, 1989), small informal groups (Howell & Higgins, 1990; Pillai & Meindl, 1991), formal work units (Curphy, 1990; Hatter & Bass, 1988), major subunits of large complex organizations (Howell & Avolio, 1993; Koene, Pennings & Schreuder, 1993), and on different variables such as overall performance of complex organizations (Koh, Terborg & Steers, 1991; Roberts, 1985; Trice & Beyer, 1986; Waldman, Ramirez & House, 1996), and U.S. presidential administrations (House et al., 1991; Simonton, 1987).

The evidence supporting this genre of theory is derived from a wide variety of samples including informal leaders of task groups (Howell & Higgins, 1990), military officers (Bass, 1985), educational administrators (Koh, Terborg & Steers, 1991), supervisors (Hatter & Bass, 1988), middle managers (Howell & Avolio, 1993), subjects in laboratory experiments (Howell & Frost, 1989), U.S. presidents (secondary data, House et al., 1991), chief executive officers of Fortune 500 firms (Waldman, Ramirez & House, 1996), and high-level executives of large Canadian firms (Javidan & Carl, 1997; Waldman, House & Ramirez, 1996), and Canadian government agencies (Javidan & Carl, 1997), and Egyptian firms (Messallam & House, 1997). The evidence shows that the effects of charismatic leader behaviors are rather widely generalizable in the United States and that they may well generalize across cultures. For instance, studies based on the charisma scale of the Multifacet Leadership Questionnaire (MLQ) (Bass & Avolio, 1989) have demonstrated similar findings in India (Pereira, 1987), Singapore (Koh, Terborg & Steers, 1991), The Netherlands (Koene et al., 1993), China, Japan (Bass, in press, 1997), and Canada (Waldman, House & Ramirez, 1996).

A recent meta-analysis by Lowe, Kroeck and Sivasubramaniam (1996) of 32 correlations between leader charisma as measured by the MLQ, and independent ratings of leader effectiveness demonstrated a mean corrected correlation of .35. A second meta-analysis by these authors, based on fifteen correlations between charisma and subordinates' ratings of their superiors' effectiveness, demonstrated a corrected correlation of .81. Corrected correlations between criterion variables and charisma were higher than corrected correlations between criterion variables and measures of intellectual stimulation, individualized consideration, contingent reward, and management by exception.

In summary, the studies based on various neocharismatic theories clearly show that this genre of leadership results in a high-level of follower motivation and commitment and well-above-average organizational performance, especially under conditions of crises or uncertainty (House et al., 1991; Pillai & Meindl, 1991; Waldman, House & Ramirez, 1997; Waldman, Ramirez & House, 1997).

Differences Among the Neocharismatic Theories. As shown by House and Shamir (1993), the theories of the neocharismatic paradigm differ with respect to their leader behaviors. These differences need to be reconciled, or else the contingencies under which the differing behaviors are important need to be specified. House and Shamir (1993) provide a theoretical integration of the leader behaviors of charismatic, transformational, and visionary leader theories. The leader behavior syndrome specified by House and Shamir (1993), discussed above, remains to be validated.

Lindholm (1990) argues that the term charisma refers primarily to socially undesirable and destructive leadership. Howell and House (1992) disagree, and distinguish between two kinds of charismatic leadership: personalized (self-aggrandizing, exploitative, authoritarian) and socialized (altruistic, collectively oriented, and egalitarian). Bass (1997) argues that transformational theory subsumes charismatic theory. House and Shamir (1993) see transformational, charismatic, and visionary leadership as essentially the same, in that all of these theories include among their dependent variables the affective states of followers, and all of them stress leader behavior that is symbolic, appealing to follower emotions, and highly motive arousing.

The above quibbles reflect rather minor differences among the theories and the opinions of the theorists - arguments over subsidiary definitions and fine tuning rather than arguments that are fundamental to the central concept of the neocharismatic paradigm: leader behaviors that account for outstanding leadership. More substantive differences between the Value Based Leadership Theory (House et al., 1996) and the other theories of the neocharismatic paradigm do exist, however.

The Value Based Leadership Theory is an extension of the 1976 Theory of Charismatic Leadership (House, 1977). Value Based Leadership Theory specifies the Leader Motive Profile and Leader self-confidence and conviction as predictors of charismatic leader behaviors. It also specifies a set of contextual conditions which are asserted to facilitate the enactment of leader dispositions and the emergence and effectiveness of value based leadership. More specifically, it is predicted that the emergence and effectiveness of value based leaders will be enhanced when the environment involves a high degree of stress and uncertainty, the organizational task is closely related to dominant values of the society, the situation offers at least some opportunity for "moral" involvement, goals cannot be easily specified and measured, and when the leader cannot link extrinsic rewards to individual performance.

The moderating effect of environmental uncertainty on the relationship between value based leadership and organizational effectiveness has been supported in studies by Pillai and Meindl (1991), Waldman, Ramirez and House (1996), Waldman, House and Ramirez (1996), and House et al. (1996). The effects of the other contingencies specified in the Value Based Leadership Theory require empirical testing (Shamir, House & Arthur, 1993).

A Short Scale for the Measurement of Charismatic Leadership. The MLQ (Bass & Avolio, 1989) is the most widely used measure of neocharismatic leadership. To date, at least 75 studies have been conducted using this questionnaire (Lowe et al., 1996). The MLQ has been used primarily to test transformational leadership theory. It includes three subscales relevant to neocharismatic theory: charisma, individualized consideration, and intellectual stimulation. All of these subscales have respectable psychometric properties.

The charisma subscale, however, does not include all of the theoretical behaviors of charismatic leadership. For example, Podsakoff et al. (1990) and House et al. (1996), have demonstrated that there are more dimensions of neocharismatic (transformational, charismatic, visionary) leadership than those included in the MLQ. Using an exploratory factor analysis, Curphy (1990), based on a sample of 11,000 respondents, found that the factorial structure of MLQ responses did not conform to the theoretical dimensions of Transformational Theory. Howell and Avolio (1993), however, confirmed the theoretical dimensions of Transformational Theory using confirmatory factor analysis.

Despite the inconsistency of these findings, the charisma scale of the MLQ is quite appropriate and useful for measuring variance in charismatic leadership. This seven-item scale includes three follower-reported leader behaviors: providing a vision of what lies ahead, showing determination in pursuit of goals, and communicating high performance expectations. It also includes four outcomes: making others feel good, generating respect, instilling confidence, and transmitting a sense of mission. Bass (Personal communication, September, 1996) reports that the MLQ charisma scale correlates above .90, with a longer behaviorally descriptive scale which includes items relevant to leaders' expressions of values, emphasis on commitment, setting high standards, stressing a sense of mission, talking optimistically about the future, expressing confidence, making personal sacrifices, providing encouragement to followers, and displaying conviction in ideals and values. Thus, the MLQ captures the essence of charismatic leadership parsimoniously and is useful for measuring variance in charisma, but not for specific charismatic leader behaviors.

Future Directions for the Neocharismatic Paradigm. As with all extant theories, the theories of this paradigm are not without problems. For example, the neocharismatic theories offer inadequate or untested explanations of the process by which the theoretical leader behaviors are linked to, and influence, the affective states of followers. Transformational Leadership Theory (Bass, 1985) rests on Maslow's theory of motivation, which has been largely disproved by empirical tests (Wahba & Bridwell, 1975). The theory of charisma advanced by Conger and Kanungo (1987) rests on an attributional explanation of the effects of charismatic leaders. The specific attribution processes to which the theory alludes are not clearly specified.

In addition, there is no explication of the processes whereby the leader behaviors specified by Visionary Theory (Bennis & Nanus, 1985; Kousnes & Posner, 1994; Nanus, 1992; Sashkin, 1988) have their theoretical effects on followers, groups, work units, or organizations. The self-concept and motive-arousal theory of motivation advanced by House and Shamir (1993), which underlies Value Based Leadership Theory, while more explicit and elaborate, remains to be tested. Shamir (1991) has reviewed several alternative explanations of the charisma phenomenon, none of which have been empirically explored. Thus, the processes by which charismatic leader behaviors have their effects remain to be empirically demonstrated.

Further, there is little evidence that charismatic, transformational, or visionary leadership does indeed transform individuals, groups, large divisions of organizations, or total organizations, despite claims that they do so. It may well be that such leaders induce changes in followers' psychological states, but that these states do not continue after the separation of leader and follower. There is no evidence demonstrating stable and long-term effects of leaders on follower self-esteem, motives, desires, preferences, or values.

Only the Value Based Leadership Theory addresses the relationship of organizational context and emergence and effectiveness of charismatic leader behavior. Studies by Pillai and Meindl (1991) and House et al. (1991) have shown that crises facilitate the emergence of charismatic leadership. Studies by Waldman, Ramirez and House (1996) and Waldman, House and Ramirez (1996) show that charismatic leadership is most effective under conditions of environmental uncertainty. House et al. (1996) found that charismatic leaderships more effective for CEOs in entrepreneurial firms than chief executives of divisions of large industrial or commercial organizations. Other than these studies, there is no empirical evidence concerning the relationship between contextual social conditions and charismatic leadership emergence and effectiveness.

As noted in an excellent discussion by Bryman (1993), there is an important need to better understand the routinization of charisma and the loss of charisma. More specifically, little is known about the effects of routinization, how routinization of charismatic leadership takes place, and the effects of charismatic leadership not accompanied by routinization. Finally, the processes by which loss of charisma occurs and the consequences of loss of charisma are also important topics that need theoretical development and empirical testing. These issues are discussed in greater detail by Bryman (1993).

Additional Opportunities for Future Research

There are several more general issues relevant to the leadership phenomenon that are largely ignored in current leadership literature. In this section, we discuss the distinction between leadership and managerial supervision, and between generic functions and specific behaviors. The influence of organizational context, the management of diversity, and other aspects of leadership offering scope for future research are identified.

Leadership Versus Management and Supervision

The first issue concerns the domain of leadership inquiry. In an article in the Harvard Business Review, Abraham Zaleznik (1977) argued forcefully that there is a difference between leadership and management. Subsequently, Bennis and Nanus drew the distinction between managers and leaders with the statement "Managers do things right. Leaders do the right things" (Bennis & Nanus, 1985, on dust jacket). Although Zaleznik did not cast his argument in terms of an attack on leadership research and theory, the implication was clear: leadership researchers had been almost exclusively studying management and supervision and not leadership. This implication is valid in that the research through the 1970s was almost exclusively based on lower- and middle-level managers and almost exclusively concerned the manner by which they supervised their immediate subordinates.

Interpretative interview and focus group research in 38 countries, involved in the GLOBE study referred to earlier, revealed a rather consistent view by respondents that leadership and management include clearly different activities. Leadership was generally viewed as involving the articulation of an organizational vision, introduction of major organizational change, providing inspiration, and dealing with highly stressful and troublesome aspects of the external environments of organizations. Management was generally viewed as implementation of the leader's vision and changes introduced by leaders, and the maintenance and administration of organizational infrastructures. These qualitative findings are suggestive but need to be demonstrated with more rigorous methodology.

Elsewhere, House (1996) has argued that distinctions between management, supervisory leadership, and general or strategic leadership are important because they help to understand why the academic literature entitled "leadership" has been criticized as irrelevant to the solution of practical problems, and has so infrequently been consulted by practicing managers and applied to the problems of leading organizations or societies.

Consider the following: In 1988 and again in 1993, Time magazine published cover articles addressing the need for leadership in the U.S. political system. Not a single reference was made to any academic studies conducted by leadership scholars. Thus, despite approximately 3,000 empirical studies conducted by academic researchers generally referred to as leadership studies (Bass, 1990), this literature seems to have been completely ignored by policy makers, the press, and practicing managers.

What, then, is the essential difference between leadership and management? Yukl (1994) clarifies the issue by noting that leadership and management involve separate processes, but need not involve separate people. Yukl (1994, p. 4) notes that "... the essence of the argument seems to be that managers are oriented toward stability and leaders are oriented toward innovation; managers get people to do things more efficiently, whereas leaders get people to agree about what things should be done."

We believe it is also useful to further distinguish between strategic and supervisory leadership. Strategic leadership is directed toward giving purpose, meaning, and guidance to organizations. This is accomplished by the provision of a vision of the organization which has inspirational appeal to members of the organization and to external constituencies on which it is dependent. Strategic leadership includes: making strategic decisions concerning the products and services of organizations and markets; selection of key executives; and allocation of resources to major organizational components; formulation of organizational goals and strategy; providing direction for the organization with respect to the organization's domain; conceptualizing and installing organizational designs and major infrastructures, such as compensation, information, and control systems; representing the organization to critical constituencies such as representatives of financial institutions, government agencies, customer interest groups, and labor; and negotiating with such constituencies for legitimacy and resources.

Supervisory leadership is defined as behavior intended to provide guidance, support, and corrective feedback for the day-to-day activities of work unit members. Supervisory leadership consists essentially of the task- and person-oriented leader behaviors specified in the leader behavior paradigm.

In contrast to leadership, we define management as the behavior of a person in a position of formal authority, intended to obtain compliance of organizational members with their normal role or position requirements. Management consists of implementing the vision and strategy provided by leaders, coordinating and staffing the components of organizations, administering the infra-structures of organizations, and handling the day-to-day problems that inevitably emerge in the process of strategy and policy implementation and ongoing organizational functioning.

It is possible for managers to be leaders and leaders to be managers. Managers become leaders by providing vision, direction, strategy, and inspiration to their organizational units, and behaving in a manner that reinforces the vision and its inherent values. Leaders often must perform many of the management functions described above.

Unfortunately, leadership research to date has concentrated predominantly on only two aspects of leadership. The first is supervisory leadership as operationalized in the leader trait, behavior, and contingency paradigms, and Leader Member Exchange and Cognitive Resource theories. The second aspect concerns giving purpose, meaning, and guidance to organizational members as operationalized by the neocharismatic paradigm. As a consequence, little attention has been given to strategic leadership and to several organizationally related functions of leaders.

The Need for More Organizational Focus

The dominant portion of leadership theories and research is primarily concerned with relationships between leaders and their immediate followers or with supervisory behaviors. It is almost as though leadership scholars, including the first author of this article, have believed that leader-follower relationships exist in a vacuum. While it is unlikely that scholars believe this, the fact is that the organizational and environmental context in which leadership is enacted has been almost completely ignored. (An exception to this statement is the work of Finkelstein & Hambrick, 1996).

We believe this state of affairs is likely due to several reasons. First, it reflects the psychological tradition of early investigators. The study of leadership began with a very limited focus - the search for universal leader traits. The tradition of psychological inquiry has provided the guiding theories and methods for leadership research since the early 1930s. It was natural for psychologists to focus on the traits and subsequently on behavioral aspects of leadership, rather than on organizational aspects, because traits and behaviors are clearly psychological in nature.

Second, it has always been, and is today, easier to study supervisors and managers at lower organizational levels than higher-level executives because relatively large samples of lower-level subjects are more accessible to investigators. The consequence is that leadership studied at lower levels is almost exclusively supervisory, rather than strategic leadership.

Third, early traditional management theorists focused on the rational analytic functions of managing such as planning, organizing, coordinating and controlling, to the exclusion of human issues such as motivation and enhancement of follower abilities. Somehow, management theory was dealt with as distinct from leadership theory, and seldom have the authors of the two literatures attended to the work of each other (See House & Miner, 1969, for an early analysis of the division of these two literatures).

The result of this state of affairs is that there is little theory or evidence concerning the kinds of leader behaviors required in various organizational settings. It is likely that either different behaviors or differential importance of behaviors will be associated with differences in organizations. Organizational variables such as size, organizational environment, and type of strategy, technology, and organizational form are all likely to impose different demands on leaders and, thus, require specific leader behaviors (see discussion section entitled Generic Leadership Functions and Specific Leader Behaviors further on).

The conclusion to be drawn from this discussion is that there are several fruitful topics for future investigation including relationships between leaders and superiors and peers, the influence of external constituencies on leader behavior, the role of leadership in strategic management, the influence of organizational variables on leader behavior, and the management of organizational infrastructures.

Strategic Leadership

The study of strategic leadership focuses on executives who have overall responsibility for an organization (Finkelstein & Hambrick, 1996, p. 2). Until only recently, this topic has been largely unresearched. Prior to about the mid 1980s, there were very few empirical studies of the strategic leadership process or strategic leader behavior (See Finkelstein & Hambrick 1996, for an excellent review of these studies).

The relative neglect of strategic leadership as a subject of empirical investigation is ironic since the study of effective organizational policies and strategies has been one of the most prominent foci of business school education ever since the founding of the earliest business schools: Wharton in 1891 and Harvard in 1908. The focus of courses on these topics has been dominantly the content, or types of policies and strategies that lead to effective or ineffective overall organizational performance. Michael Porter's books (Porter, 1985; 1986) are perhaps the most frequently used and best examples of this focus of scholarship. Most of the strategic management literature emphasizes the alignment between industrial, market, and economic characteristics and organizational strategies, goals and processes. Relatively little attention has been paid to the processes by which strategic leaders affect organizations.

The various forms of strategies and their relevant considerations, and the history and content of this literature, are summarized well by Mintzberg (1994). Until approximately the mid 1980s educational courses and writings on this subject have been largely atheoretical and based almost exclusively on case studies. The dominant part of this literature is published in the Academy of Management Journal, the Journal of Management, and the Strategic Management Journal.

A small body of theoretical literature concerning the role of leaders and top management teams, and the processes of strategy formulation and implementation, has only recently emerged. This literature concerns the processes by which trip level managers make strategic decisions, and to a lesser extent the behavior of such managers as leaders of their organizations, and the composition of top management teams (Jackson & Ruderman, 1995).

A comprehensive summary of this literature, and a theoretical conceptual framework for understanding and studying strategic leadership, is provided by Finkelstein and Hambrick (1996). Their second chapter is entitled "Do Top Managers Matter?" Finkelstein and Hambrick identify most of the few empirical studies that adequately assess the effects of top managers on overall organizational performance. Based on studies of the characteristics of top executives that have been found to be associated with organizational effectiveness, they conclude that top managers do indeed matter but that they are often constrained by factors in their environments, organizational inertia resulting from fixed costs and prior commitments, and limitations of the executives themselves. These constraints limit executive discretion, the effects of their decisions, and the effects of their behavior on overall organizational performance.

There have been six empirical studies that adequately address the issue as to whether top managers have an influence on organizational performance (Fiedler, 1996; House, Spangler & Woycke, 1991; Smith, Carson & Alexander, 1984; Thomas, 1988; Waldman, House & Ramirez, 1996; Waldman, Ramirez & House, 1996). The studies by House et al. (1991), and Smith et al. (1984) were based on U.S. presidents and ministers of churches, respectively. The studies by Thomas and by Waldman and his associates are based on profit-making organizations. In all of these studies, it was found that the leaders of organizations have significant influence on the overall organizational performance of their organizations.

There have also been a substantial number of studies of the effects of top management succession on organizational performance in both profit and nonprofit sectors. A review of these studies is provided by House and Singh (1987). Collectively, these studies show that, more often than not, executive succession makes a substantial difference in the performance of organizations. However, the limitations on the effects of executives are substantial, resulting from a number of sources as well as the conditions under which chief executives assume their role. For example, executive succession has been found to result in increased organizational performance when the predecessor is a non-dominant individual and has left the organization, and when the succeeding CEO has the ability to cope with organizational uncertainties, a history of competence, relevant knowledge and external influence, and personal attributes that match the demands of the organization. Further, executive succession has also been found to result in increased organizational performance when succession is orderly and planned, the organizations' members do not have close personal ties, and organizations are relatively less bureaucratic (House & Singh, 1987). The latter two conditions are likely to allow newly appointed CEOs to more easily make strategic changes.

There remain many unresolved questions and issues to be studied concerning strategic management. Following are some of the more important of these:

* Are there any generic and universal leader behaviors or personal attributes that differentiate effective from ineffective top managers?

* How do contextual factors such as prior sunk costs, high fixed costs, historical events, industrial characteristics, organizational size, international competition, economic globalization, market structure, and organizational demography influence the work of top managers and moderate the effects of their behavior on organizational performance?

* What are the most common limitations on the effects of top managers?

* How do these limitations operate and can they be modified by top managers in the short or the long run?

* Can the roles of top management be shared? If so, how can the top management tasks and leadership functions be divided and what aspects can be delegated?

* What are the processes by which top managers have their effects on, for instance, decision choice, policy formulation and direction, development and management of infrastructures, motivation and inspiration, and representation to critical constituencies? (Interestingly, Waldman, House & Ramirez [1996], found that charismatic leaders make more strategic changes following their appointment than do noncharismatic leaders.)

Generic Leadership Functions and Specific Leader Behaviors

In the current leadership literature, dimensions of leader behavior, specific leadership behaviors, and leadership style are often used interchangeably by some authors, while others use these terms to refer to different aspects of leadership. It is useful to draw distinctions between these terms and present some working definitions.

Organizations are both task performing and social institutions. As such, it is logical and reasonable to expect that there is a generic set of task-oriented leadership functions that must be performed to ensure organizational performance, and a generic set of organizational maintenance-oriented leadership functions that must be performed for the effective social integration of organizational members, units, and activities (Misumi, 1985).

The generic, task-oriented functions include the task-oriented leader behaviors specified in the leader behavior paradigm and the strategic leader behaviors specified above. The generic maintenance-oriented functions include the person-oriented behaviors specified in the leader behavior paradigm, in addition to ensuring collaborative interaction among organizational members, establishing a supportive social climate, and providing infra-structures and management practices to ensure equitable compensation to organizational members, member training and development, protection of their interests, and work-related satisfaction.

Generic Functions and Specific Behavioral Manifestations of Leadership. It is useful to distinguish between generic leadership functions and specific behavioral manifestations of the generic functions which we refer to as specific leader behaviors. The generic functions are enacted by several specific behaviors. Thus, the genetic leadership functions represent broad classes of specific leader behaviors. We refer to these classes of behaviors as generic leadership functions because they are likely to be required of one or more leaders in the normal functioning of groups or organizations, although not necessarily required at all times, or of the same leader at the same time. The enactment of these generic functions is accomplished by the exercise of specific behaviors which vary as a function of the nature of the work performed by organizational members, their competence, the personality of the leader, and the cultural context in which leaders function.

The distinction between generic leadership functions and specific leader behaviors involved in enactment of these functions is nicely illustrated in a study by Smith, Misumi, Tayeb, Paterson, and Bond (1989). Their study is based on questionnaire responses of supervisors from six electronics-producing companies in Britain, Hong Kong, Japan, and Korea. These investigators demonstrated that a select set of questionnaire items describe generic task- and maintenance-oriented leadership functions. The existence of generic functions was demonstrated by similar factor structures representing these two leadership functions in each of the four countries. Several additional items also loaded on these factors in some countries, but not in others. These country-specific items reflected specific behavioral manifestations of the generic functions (factors).

For example, factors labeled maintenance-oriented leadership in all four countries, consisting of several common items were found. Several specific behavioral items associated with this factor in Japan described supervisors as speaking about a subordinate's personal difficulties with others in the subordinate's absence rather than speaking personally with the subordinate, and sending written memos rather than giving directions on a face-to-face basis. In contrast, specific items associated with the same factor in the U.S. described supervisors as being consultative and participative, dealing with subordinates on a face-to-face basis, and not sending written memos.

From these examples, it can be seen that the generic person-oriented function is enacted by specific behaviors that differ markedly across settings, in ways which are understandable within each setting. Communication with subordinates through others or through written memos avoids direct confrontation, and permits face-saving in the event of disagreement. This form of communication is consistent with the Japanese cultural norm of maintaining in-group harmony and preserving face.

Consultation and direct face-to-face communication favored by U.S. respondents reflect the cultural norms of egalitarianism and directness of Anglo-Saxon countries. Thus, while the generic function of maintenance-oriented leadership is reflected in common dimensions (statistical factors) in all four cultures, some of the specific behavioral manifestations of this generic leadership function vary among countries.

Using both generic and behavior specific measures has the advantage of more precise measurement. This approach allows investigators to relate specific behavioral descriptors to generic leadership functions. This approach also offers the advantage of being able to describe leadership as it is uniquely manifested in each culture, organization, or organizational unit studied. An understanding of generic leadership functions is necessary for theoretical understanding of the leadership phenomenon. An understanding and description of specific leader behaviors is more appropriate for applied purposes such as training or organizational development efforts and for understanding cultural and organization specific leadership phenomena. Misumi (1985) has demonstrated the usefulness of using specific leader behaviors in an extensive leadership research program conducted in Japan.

In addition to task- and person-oriented leadership functions, there are a number of other possible generic leadership functions required for effective organizational leadership. The evidence relative to the Neocharismatic Leadership Paradigm suggests additional leadership functions: communicating a desirable vision of the future of the organization to provide direction, communicating core values to guide the functioning of the organization, and to infuse organizations with values and meaning, providing inspirational leadership by setting personal examples of the core values and serving as a role model of the core organizational values. These leadership functions are likely to be generic strategic leadership functions. The specific behaviors by which these generic functions may be enacted by leaders include: making public speeches, and communicating their messages through written or electronic media, personal example, or by personal visits during which the leaders communicate their message directly to organizational members.

Research by Mintzberg (1973), based on intensive observation of five high-level executives from 28 to 53 hours each, suggests ten generic leader behaviors. Mintzberg identified ten "roles" that he inferred from his observations. Mintzberg believes that these roles represent the dominant classes of activities that high-level managers perform. These roles are divided into three groups: interpersonal roles including figurehead, liaison, and leader roles; informational roles - including monitor, disseminator, and spokesman roles; and decisional roles - entrepreneur, disturbance handler, resource allocator, and negotiator roles. Empirical research supports Mintzberg's classification of managerial activities (Pathet & Lau, 1982; also see Bass 1990, pages 385-401, for a review of studies pertinent to managerial roles). Obviously, these roles overlap with task- and person-oriented generic leadership functions, and the strategic and neocharismatic leader functions reviewed above. Thus, it is possible that Mintzberg's roles represent generic leadership functions which are, and must be, performed for effective leadership of almost any organization. Of course the relative emphasis placed on these roles will vary as a function of organizational technology, the day-to-day problems faced by leaders, and the task environment of their organizations.

We believe it is likely that there are indeed several leadership functions generic to the exercise of leadership and universally expected, accepted, and effective across organizations, industries, and cultures. Among these are the task- and maintenance-oriented leader functions, the functions of the neocharismatic leadership paradigm, the strategic leadership functions described above, and the functions suggested by Mintzberg's research. At this stage, this belief represents theoretical speculation and remains to be developed theoretically and demonstrated empirically. The rather obvious conclusion to be drawn from the above discussion is that research needs to be directed toward establishing the generic functions, the conditions requiring their performance, and the specific leader behaviors required to enact these functions.

Leadership Styles

The manner in which specific behaviors are expressed may vary substantially. For example, the same leader behavior can be expressed with a high degree of emotion or with equanimity, assertively or nonassertively, in a friendly versus an impersonal or even unfriendly manner, harshly or gently, or autocratically, consultatively, or democratically. Variance in leader mannerisms of this kind is likely due to leader personality and to cultural norms. We refer to the manner by which leaders express specific behaviors as leader style. With the exception of autocratic, consultative, and democratic styles, there has been little research concerning the way leader styles vary, how leader style affects subordinates, and the conditions under which the various aspects of leader styles are more or less appropriate. We speculate that the judged appropriateness, acceptance, and effectiveness of leader styles will be primarily a function of congruence with the norms of the culture in which the leader functions. However, to our knowledge, there has been no empirical research in this area.

The Management of Diversity

Consider the following quotations, taken from Chen and Van Velsor (1996):

"... there is only a very limited knowledge base regarding leadership behaviors of nontraditional and non-Western leaders" (p. 292).

"... it is important to explore and discover the various mechanisms through which diversity impacts organizational life" (p. 295).

"Global leaders, like leaders of any diverse group, should be cultural integrators and facilitators" (p. 297).

The above quotations are taken from the concluding article of a two- volume series on leadership and diversity published in the Leadership Quarterly 7(1 & 2, 1996). These statements clearly illustrate the problems associated with, and the importance of, increasing available knowledge relevant to leadership of diverse subordinates. It is evident from these articles, as well as a growing number of other sources (Chemers, Oskamp & Costanzo, 1995; Loden & Loeser, 1991; Morrison, 1992; Triandis, Dunnette & Hough, 1994), that the issue of leading diverse individuals and groups requires substantial theoretical development and empirical research. At the present time, the literature on this issue is largely speculative and anecdotal.

Organizational diversity can be objectively assessed by identification of the proportions of organizational members with different gender, ethnic, racial, religious, historical, and national backgrounds. The essential aspect of diversity is that it is associated with differences in dispositions and modal behavior patterns of members from diverse groups. The differing dispositions take the form of social identities, meanings imputed to events, social expectations, values, beliefs, assumptions, and motives. When there is commonality within groups with respect to these dispositions, group norms usually emerge. Group norms are assumed to have powerful directing, motivating, and controlling effects on group members and on their interaction with members of other groups.

Perhaps the beginning point for a research agenda on diversity leadership is to search for an understanding of how differing individual dispositions and group norms influence interactions and behavior of individuals and groups in organizations. At the group level of analysis, one important consideration concerns the conditions under which highly diverse groups are able to work together, and possibly even achieve synergy. A second consideration concerns how dispositions and norms influence intergroup cooperation. A third consideration concerns how diverse individuals or groups that are unable to work together collaboratively might be managed to avoid conflict and enhance performance effectiveness. A fourth consideration concerns the management of geographically separated organizational units that operate in countries other than that of their parent company. For example, Yetton (Personal communication, 1996) has found that the appointment of Australian-educated Asians to manage divisions of Australian firms operating in Asia is highly effective. This approach is less costly, and the Australian-educated Asians are more likely to be culturally attuned than Australian executives, even if the Australian executives are provided cultural training in preparation for Asian assignments.

A fifth consideration concerns whether diversity does, indeed, have the potent effects commonly assumed. It may be that there are a set of organizational imperatives which take precedence over the management of diversity, and that when these imperatives are met, the effects of diversity are dampened, or even completely muted. For example, Anderson (1983) found ethnic diversity to make little difference in the kinds of leader behaviors practiced and accepted in an organization with members from Caucasian and various Polynesian ethnic backgrounds. This finding is especially surprising, since several of the Polynesian groups were historically antagonistic toward each other.

Anderson speculates that when work is well managed and coordinated, and when employees understand their role demands and task requirements, the effect of diversity is muted. If Anderson's speculation proves to be generalizable, then the current concern with diversity leadership may reflect an exaggerated view of the problems associated with diversity. While diverse groups may well have different preferences that need to be considered, and equal opportunity needs to be assured for minorities, the expectation that diversity will inevitably result in intragroup conflict, the imposition of extremely difficult role demands on managers, or problems that are unusually difficult to solve, may be exaggerated. It is possible that the dysfunctional effects of diversity will be substantially less when diverse members of organizations have a positive relationship with their organization, when rewards are equitably distributed to members of all diverse groups, when the organizational culture is generally supportive and adequately employee-oriented, and when the work is well organized and roles are clear. We know little about the potential dampening effect resulting from management practices suggested by Anderson. This possibility deserves empirical investigation.

Chen and Velsor (1996, p. 290) note that "... the more relations are diverse between the leader and followers, and among members, the more the researcher needs to go into the dyadic dynamics in order to capture what is really going on." At the dyadic level of analysis, LMX Theory and the 1971 version of Path-Goal Theory appear to be applicable. The problem with applying these theories is that, as they are currently stated, they reflect Western values and assumptions, such as individualism, and emphasis on rationality rather than on tradition, spirituality, or superstition.

In order to effectively apply LMX theory, it is necessary to understand what constitutes desired, acceptable, and effective relationships between superiors and subordinates. In collectivistic societies, individuals define their self-concepts in terms of, and take substantial satisfaction from, group identification (Markus & Kitayama 1991; Triandis, 1995). The interpretation of specific behaviors may also differ across cultures, leading to a breakdown in the predicted course of LMX development in multicultural environments (Aditya, 1994). Failure to understand the nature of expected relationships among members of diverse cultures is likely to limit the application of LMX Theory in its present stage of development.

Differing expectations of diverse group members will also limit the applicability of the 1971 version of Path-Goal Theory. The leader behaviors specified by this theory and indeed by most prevailing theories of leadership reflect a highly individualistic orientation and may be unacceptable and inappropriate in collective societies. Paternalistic leader behaviors, and behaviors directed toward within group harmony, conflict avoidance, face saving, and group maintenance, are likely to be more effective in collectivistic societies, and to be less effective or even dysfunctional in individualistic societies (House, Wright & Aditya, 1997, in press). The 1996 version of Path-Goal theory is a theory of work unit effectiveness, and may be more applicable in collectivist societies than the 1971 version, which is a dyadic theory of the effects of leader behavior on individuals. Empirical research is required to determine the validity of the 1996 version of Path-Goal Theory.

Cross Cultural Leadership

Similar problems with the application of most extant leadership theories are likely to arise from differences in dispositions and norms across cultures. More specifically, diverse cultures have varying norms and varying reactions with respect to a wide range of management and leadership practices, such as degree of assertiveness and aggressiveness of leaders, use of extrinsic or intrinsic incentives, task or achievement orientation, short- versus long-term orientation, and leader consideration. Substantial evidence shows that cultures vary on the value they assign to the above leader behaviors, and that individuals from different cultures have different reactions to these leadership and management practices (House et al., 1997). Yet, to date, we have no empirically supported theories that adequately consider cultural differences of organizational members.

A Theory of Cross Cultural Leadership. House et al. (1997) have advanced a theory of cross-cultural leadership. This theory asserts that expected, accepted, and effective leader behavior varies by cultures. According to the theory, the importance placed on, and effectiveness of, person- and task-oriented leader behaviors are contingent on the culturally endorsed implicit theories of leadership (CILTs) of the broader social system. Person- and task-oriented leader behaviors that are congruent with CILTs will be more effective than behaviors that are incongruent. However, it is argued that some leader behaviors that are incongruent with CILTs will also be effective. These behaviors are those concerned with the introduction and implementation of major changes in the culture or in the organizations within the culture. More specifically, the behaviors described by the neocharismatic leadership paradigm are expected to be effective in bringing about constructive change, regardless whether they are congruent or incongruent with the norms of the broader culture.

The theory specifies the process by which cultures influence the kinds of leader behaviors that will be accepted, effective, and enacted in specific cultures. The theory also specifies a number of interactions between cultural norms, organizational practices, leader behaviors and leader acceptance, effectiveness, and enactment. The theory remains to be tested. For a more detailed discussion of the theory, the reader is referred to the chapter by House et al. (1997).

The Importance of CILTs. Implicit Leadership Theory (Lord & Maher, 1991), reviewed above, can be extended to the cross-cultural rather than the dyadic context. As suggested by Lord and Maher (1991), and Ayman (1993), implicit leadership theory is especially useful for the cross-cultural study of leadership. If cultural values and norms influence commonly held implicit theories of leadership, then one would expect these theories to vary by culture. A knowledge of CILTs would provide information about what is expected of leaders, and about the influence granted to leaders. More specifically, CILTs indicate the extent to which individuals in positions of leadership are expected to be change-oriented, risk-oriented, visionary, directive, and proactive, as opposed to being reactive, nondirective, risk-averse enactors of prevailing social consensus, and maintainers of the status quo. Knowledge of such differential expectations of leaders will contribute to prevailing leadership theory and will provide the basis for future behavioral survey and experimental research on the effects of adhering to or violating cultural expectations by leaders.

Knowledge of culture-specific and universal aspects of CILTs will help to better understand the scope of cultural influences on leadership and leader-related variables: the extent to which there are universal cultural forces, the extent to which such forces permeate management practices, and the extent to which the cultural syndromes resulting from cultural forces vary across cultures. Thus, knowledge of culture-specific and universal culture syndromes will be useful in understanding the very nature of culture, which is, as yet, theoretically not well defined.

Knowledge of CILTs will facilitate cross-cultural communication and be useful for improving relationships between members of different cultures. This knowledge could also be used for management selection and for the content of management training programs for managers assigned to work in cultures other than their own. As argued by House et al. (1997), it is likely that there are both some leader behaviors that are universally accepted and effective, and some for which acceptance and effectiveness is culture-specific; but little is known about such behaviors and their cultural-specificity or universality at present. For a more elaborate discussion of this issue, see House et al. (1997).

Clearly, research on cross-cultural leadership is needed and likely to become more and more important as the world becomes a "global village" with a near-common market for the major countries of the world.

Toward a Theory of Political Leadership

It is rather amazing that there is no theory of political leadership in complex organizations, given the fact that social psychologists have long studied power and influence processes in organizations (e.g., Kipnis, 1984; Kipnis, Schmidt & Wilkinson, 1980; Schmidt & Yeh, 1992; Yukl, Falbe & Youn, 1993). Most observers of organizations will admit that politics and political behaviors take place in organizations, that such influencing behaviors often are necessary to achieve organizational objectives, and that such behaviors can have profound effects on not only the targets of such influence, but on those who exercise them (Kipnis, 1976, 1987) and on the functioning and effectiveness of organizations.

At minimum, an adequate theory of political leadership would include the following components:

* A definition of political behavior,

* A specification of the conditions that enhance or impede the exercise of political behavior,

* A specification of the sources of influence on which leaders and organizational members can draw,

* A specification of the countervailing political forces that exist in organizations,

* A specification of the motives and personality traits relevant to the exercise of power and political behavior in organizations,

* A description of the kinds of behavioral tactics enacted in the pursuit of political objectives,

* A description of when such tactics will be used,

* A description of how politically motivated behavior becomes legitimized,

* The moderating effects of organizational context on relationships between politically motivated behaviors and their effects, and

* The ultimate effects of politically motivated behavior on organizational performance and survival.

Many of these building blocks for a theory of political behavior are currently available, but await theoretical integration. Following is a brief description of the literature dealing with these building blocks.

Several authors have offered definitions of political behavior. The essence of these definitions is that such behavior is driven by self-interest and is not explicitly condoned or condemned by organizational policy or norms (Porter, Allen & Angle, 1981). Thus defined, political behavior is neither inherently good nor evil, neither detrimental nor functional for organizational performance, and can be enacted in pursuit of organizational and individual interests mutually or divisively.

The conditions that give rise to and enhance the exercise of political behavior are described by Pfeffer (1981). Many of the sources of influence available to organizational members have been specified in most organizational behavior textbooks and in many articles. (See Bacharach & Lawler, 1980; Mintzberg, 1973; Yukl, 1994). Strategic Contingency Theory of power in organizations describes sources of organizational work unit power (Hickson, Hinnings, Lee, Schneck & Pennings, 1971). Many of the countervailing political forces that exist in organizations have been described by Mintzberg (1973). There is a substantial amount of literature on the motives and dispositions related to politically oriented behavior, including the power motive (McClelland, 1985; Winter, 1973), the moral responsibility disposition (Winter, 1973), and Machiavellianism (Christie & Geis, 1970).

House (1988) provided a review of research concerning personality and power in organizations. Organizational influence tactics are described by a number of authors (Kipnis, Schmidt, Swaffin-Smith & Wilkinson, 1984; Mintzberg, 1993; Yukl, 1994). Xin and Tsui, (1996) provide a comprehensive review of the empirical evidence on influence tactics. The effects of the exercise of such tactics on powerholders are well documented by Kipnis (1976). Relationships between personality variables and the acquisition and exercise of power are discussed in detail by House (1988). Finally, House (1991) has presented a theory which explicates some of the organizational variables which moderate or influence the acquisition, distribution, and exercise of power in organizations.

While many of the psychological aspects of political behavior in organizations have been studied, there is much less information about how organizational context influences the exercise of political. behavior or its outcomes. There is little theory or available evidence concerning the legitimization of politically motivated behavior, the moderating effects of organizational context on relationships between politically motivated behaviors and their effects, and the ultimate effects of politically motivated behavior on organizational performance and survival. Clearly, political leader behavior in organizations is a topic worthy of theoretical development and empirical investigation.

Distributed Leadership Revisited

Classical Management Theory argues that there should be a single chain of command within organizations and that every individual should have one, and only one, boss (Fayol, 1930). Adherence to a single chain of command ensures that there is only one person to whom each organizational member reports and from whom the individual takes orders. Consequently, adherence to a single chain of command increases managerial control and reduces role conflict experienced by subordinates (House, 1970). However, there is some speculation, and some preliminary evidence, to suggest that concentration of leadership in a single chain of command may be less optimal than shared leadership responsibility among two or more individuals in certain task environments.

In the very recent past, there has emerged a consensus, or near consensus, among Program Directors of the Dwight D. Eisenhower Leadership Development program that indeed distributed leadership, for purposes of this program, is more effective. This is a program funded by the U.S. Department of Education to encourage leadership education for undergraduate students and citizen community organizations. This consensus reflects a strong egalitarian orientation (Eisenhower Leadership Group, 1996) and is best articulated by Astin and Astin (1996, p. 16), who state "A leader is not necessarily a person who holds some formal position of leadership or who is perceived as a leader by others." These authors define a leader as "... one who is able to effect positive change for the betterment of others, the community, and society." They go on to state that "All people, in other words, are potential leaders. Moreover, the process of leadership cannot be described simply in terms of the behavior of an individual: rather, leadership involves collaborative relationships that lead to collective action grounded in the shared values of people who work together to effect positive change."

Distributed leadership can take three forms: delegated, leadership, co-leadership, and peer leadership. We discuss each of these in turn.

Delegated Leadership. It is likely that when the job of managing involves rather large complex organizations, several generic functions of management can be divided among two or more leaders, and performed contemporaneously. For example, one member of a top management team might perform spokesperson and external coordination functions, while another performs internal administrative functions, and yet another conducts the analyses required for strategy formulation. We refer to this approach as delegated leadership. The top management generic leadership functions are divided among executives, in this manner, at both General Electric (Elderkin & Bartlett, 1993) and Microsoft (Schlender, 1995). In both firms, delegated leadership seems to be working quite effectively as these firms have attained exceptionally high levels of growth and profitability over the last decade by managing in this manner. It should be noted, however, that while the generic leadership functions are delegated, there is one leader in each firm who has final decision authority over all strategic decisions. Little is known about how such delegated leadership functions, the constraints on such delegation, the demographic and personality variables that are relevant to effective integration of delegated leadership, and the coordinative mechanisms required to make delegated leadership effective.

Co-Leadership. Co-leadership concerns the division of roles. Co-leadership was first suggested, in a report of the Harvard Laboratory Studies, by Robert Bales (1954, p. 320):

The laboratory findings, while still tentative, indicate that the man who is judged by the group members to have the "best ideas" contributing to the decision is not generally the "best-liked." There are two separate roles - that of task leader and that of social leader. If a man comes into a task-leadership position because he is popular or best liked, he is ordinarily confronted with a choice: (1) If he chooses to try to keep the task leadership of the group, he tends to lose some of his popularity and to collect some dislikes. (2) If he chooses to try to keep his popularity, he tends to lose the task leadership. People differ in the way they solve this dilemma, although most tend to prefer to keep the popularity rather than the task leadership.

Bales states that the dilemma becomes more acute with time. At the end of the group's first meeting, there is a 50% chance that the task leader will be the most liked. This reduces to 25% at the end of the second meeting, about 16% by the third, and even less by the fourth meeting.

It is rare, according to Bales, for one person to be able to hold both roles; instead, the tendency is for these positions to be held by two different persons, both of whom contribute to the effective performance of the group: "The task leader helps to keep the group engaged in the work, but the pressure of decision and work tends to provoke irritation and injure the unity of the group. The best-liked man helps to restore this unity of the group and to keep the members of the group aware of their importance as particular individuals, whose special needs and values are respected." In some small organizations, the chief executive and the workers' representative may occupy these roles.

The stability of the group is maximized when the two leaders recognize each other's roles and work together. Examples of such coalition may be found at home, in the family, or in the workplace.

Some findings relevant to co-leadership have been reported by Waldersee and Eagleson (1996) with respect to the distribution of the task and person-oriented leadership functions among top managers of 44 hotels owned by a Sydney-based hotel-management corporation. These authors found that the implementation of a major change program was substantially more effective when the task- and person-oriented leadership functions were divided among at least two members of the hotel change program steering team. Many situations may require division of the task- and person-oriented behaviors. For example, some members may display aggressiveness toward competitors, while others display empathy toward in-group members. Or, "good cop-bad cop" roles might be distributed to different individuals for the purpose of bill collecting, interrogation, or difficult persuasion of competitors or others. These, as well as Bales' observations and suggestions, remain to be further investigated and subjected to empirical testing.

Peer-Leadership. It is also possible that some of the specific leader behaviors required to enact generic functions can be distributed throughout the entire group or work unit being managed. Thus, several individuals, would enact the same specific leader behaviors contemporaneously. We refer to this approach as "peer leadership." Peer leadership was found by Bowers and Seashore (1966) and Eagleson (1996) to be associated with a high level of organizational performance.

Bowers and Seashore studied the degree to which individuals in agencies of a leading life insurance company engaged in the following behaviors: supportive leadership, goal emphasis, work facilitation, and interaction facilitation. They measured the extent to which each of these behaviors was enacted by both superiors and peers. They found correlations ranging from .49 to .82 between the degree to which superiors and peers enacted these behaviors. They found that peer leadership often had a higher correlation with agency performance than leadership exercised by the formal manager of the agency. They also found that the highest correlations between manager and peer leadership measures were those in which the same behaviors were measured by scales administered to both managers and group members: manager and peer-work facilitation, manager and peer-goal emphasis, manager and peer interaction facilitation, and manager and peer support. This suggests that the manager sets the example of appropriate peer leader behavior. The research by Bowers and Seashore (1966) clearly demonstrates that the exercise of leader behaviors can be shared by members of work units, as well as conducted by formal work unit managers.

Future Opportunities for Research on Distributed Leadership. The above discussion suggests several topics for future research:

* What generic functions can be distributed?

* What specific behaviors can be enacted as peer leadership?

* What styles are compatible and complementary and what kinds are incompatible?

* To what extent (if at all) is distributed leadership more effective when it is consciously planned and formally implemented than when it emerges naturally and informally?

* What are the conditions under which each of the three forms of co-leadership are likely to be effective or ineffective?

Management Training and Development

That management training and development efforts will result in improved management appears to be taken as an article of faith by many corporations, professional management associations, and consultants. Yet, despite the immense amount of investment in management training on the part of corporations and government, there is little evidence that such training results in more effective management behavior. Burke and Day (1986) conducted a meta-analysis of seventy different management training studies. They found a moderate increase in knowledge with respect to prescribed leadership practices; but, interestingly, some studies also showed negative effects. They concluded that they could recommend only two training methods that had been empirically validated

by rigorous procedures. These were the Sorcher and Goldstein (1972) Behavioral Modeling Approach and the Fiedler and Chemers (1984) Leader-Match Training Program. An interpersonal relations training program designed for supervisors has also been validated by Latham and Saari (1979). This training program was designed on the basis of social learning theory principles. The performance of the trained supervisors' work units were shown to increase by about 20% as a result of training, using suitable multiple methods of assessment. Finally, the program evaluated by Graen and his associates, described earlier, appears to be effective for the management of subordinates with high growth needs.

All of these training programs concern supervisory behaviors, rather than the enactment of neocharismatic leader behaviors, strategic leadership, or the ten roles specified by Mintzberg. Further, none of the commercially available programs dealing with these latter topics have been subjected to rigorous evaluation. Fiedler (1996, p. 244) states that "All of the reviews of leadership training ... stress that we know very little about the processes of leadership and managerial training that contribute to organizational performance. At least one reason for this lack of knowledge is the scarcity of meaningful and rigorous research ...."

Fiedler (1996, p. 244) concludes that "...we must certainly continue to ask whether we are teaching what business students and managers really need to know, and we need to examine carefully how much training and development programs contribute to organizational performance." To this, we would add that there is also a need to develop methods to ensure transfer of training to on-the-job application, and a need to teach managers the conditions under which they can make effective use of what they have learned in training programs.

Social influences in the work environment of leadership trainees can produce either functional or dysfunctional consequences of training. Three studies of the effects of supervisory leadership training indicate that when existing management practices do not support the practices taught in management training programs, there will be increased role conflict for the trainees, increased stress between the trainees and the members of the organization with whom they interact, and decreased job performance (Fleishman, Harris & Burtt, 1955; House, 1960; Sykes, 1962). In one study, Sykes (1962) found that, as a result of attending a management training program, some managers became frustrated with existing management practices in their own organizations, and sought jobs in other organizations.

Not only may management training stimulate turnover among managers, but the managers who leave their organizations will also be those most mobile in the labor market, reflecting their superior experience and prior performance (Sykes, 1962). Thus, management training programs have the potential to stimulate higher performing managers to leave their present organizations. Surprisingly, little attention has been paid to the congruence between the organizational environment of individuals who participate in leadership training programs, and the content of the training programs. Clearly, this issue requires research and the development of methods to ensure that leadership training efforts not only stress appropriate leader behavior and make valid prescriptions, but also that they do not result in such dysfunctional consequences as those specified above.

Finally, there is a need to understand the limits of behavioral, cognitive, and motivational flexibility of leaders. As mentioned above, there is reason to believe that many individuals are not able to substantially vary their cognitive style or orientation, their dominant motives, or their global behavioral patterns (Fiedler, 1967; Sherif, Sherif & Nabergale, 1965). There is no available evidence that shows that individuals can substantially alter autocratic, participative, charismatic, task-oriented, or person-oriented behavior patterns. Yet, this is what is often taught in management training programs. There are undoubtedly limits to the effects that can be expected from management training and development efforts.

There are at least three plausible competitive hypotheses relevant to the effects of management training and development that need to be investigated. First, it can be hypothesized that only those managers who are disposed toward the exercise of effective leadership will profit from management training. For example, Graen et al. found that LMX training had a positive effect for only those managers with high growth need strength. Second, and more optimistically, it can be hypothesized that management training, like athletic coaching, will improve all who are trained, but that those managers who are disposed toward effective management practices, much like individuals with athletic potential, will improve most. Third, it can be hypothesized that management training will only result in improved management performance when the trained managers return to jobs in which the practices taught in the training programs are supported by their immediate superiors and existing management practices. Clearly, research directed toward tests of these competitive hypotheses is warranted.

Universal or Near Universal Effective Leader Behaviors

There is reason to believe that some generic leadership functions may be universally acceptable and effective, regardless of the dispositions and norms of diverse groups, although enacted with different behaviors depending on the situation or culture. First, there are undoubtedly several problems that are universally associated with the management of large complex organizations and with the management of groups. For example, the need to ensure task orientation and the need to develop and maintain cohesiveness and collaboration among organizational members are likely to be present in all complex organizations and in all organizational subunits. Second, as House and Baetz (1979) noted, effective leadership requires a disposition to be influential. This disposition may well result in some universal influence-oriented behaviors. Third, many of the strategic contingencies facing organizations may well be universal, or near universal. For example, all organizations that function in competitive environments must of necessity conduct negotiations and transactions with external constituencies for the attainment of resources and legitimacy.

While the logic suggesting universality of leader behaviors is compelling, there is only sparse empirical evidence relevant to this issue. Following is a brief review of available evidence. A twelve-country study by Bass, Burger, Doktor, and Barrett (1979) revealed that managers from all countries indicated a desire to get work done, while using less authority. Smith and Peterson (1994) found that managers in 25 countries representing a wide variety of cultures report satisfaction with events for which they were delegated substantial discretion.

Interpretive interviews and focus group research in 38 countries involved in the GLOBE study, mentioned earlier, suggests that the behaviors specified in the neocharismatic leadership paradigm might well be universally accepted and preferred. In all 38 countries, managers were asked in focus groups to describe leader attributes and behaviors that enhance outstanding leader performance. In all countries, managers described behaviors similar to those of the neocharismatic leader behavior syndrome. Bass (1997) reports studies using the MLQ transformational leadership scales in China, the U.S., Netherlands, Singapore, England, and Japan. In all of these countries, transformational leadership was found to be positively related to leader effectiveness and subordinate satisfaction. Research by Messallam and House (1997) in Egypt, by Javidan and Carl (1997) in Canada, and by Geyer and Steyrer in Germany (1994) also yielded similar findings. However, the specific behaviors and the mannerisms (styles) with which these apparently generic leadership functions are enacted may vary substantially among leaders, and may be differentially required for diverse groups. For example, charismatic leader behaviors may be enacted aggressively, as exemplified by General George Patton, Fidel Castro, or Theodore Roosevelt, or in a quiet, unemotional, and nonaggressive manner as exemplified by Mahatma Gandhi, Nelson Mandela, or Mother Teresa.

Conclusion

In this article, we have reviewed the more prominent extant theories of leadership. We have also reviewed a select set of topics that we believe offer opportunities for important contributions to knowledge about the leadership phenomenon. Our choice of topics was guided by our own knowledge base, and by our judgment about what is most important for the formulation of an agenda for future leadership research.

There are many other important topics that could also have been included in this review: leader selection, the effect of leadership at multiple levels of analysis, and formal education as an aid to leadership effectiveness, to name only a few. The effects of leadership at multiple levels of analysis has been discussed at length with respect to the theories reviewed in this literature, in two volumes of Leadership Quarterly (Special issues on multiple-level approaches, 6, 2 & 3, 1995). Leader selection is discussed at length in many industrial/organizational psychology texts, and the research issues on this topic are rather well established. The role of formal education as an aid to leadership is discussed at length in the Dwight D. Eisenhower Leadership publication, cited above, as well (Astin & Astin, 1996). Undoubtedly there are other topics worthy of discussion.

Cultural Limitations of Extant Theory

All of the currently prominent theories are authored by American scholars. One limitation of these theories is that they do not address the issue as to whether they can be generalized to other cultures. It is very likely that most of these theories are culture-bound, reflecting U.S. assumptions, values, and beliefs. For example, both LMX theory and the original Path-Goal theory are based on the presumption that leadership consists basically of dyadic relationships between leaders and followers. This assumption is clearly a reflection of the individualistic orientation of dominant main-stream U.S. culture. Transformational Theory asserts that effective leadership involves the exercise of individualized consideration toward subordinates. This leader behavior may well violate the cultural norms of highly collectivistic societies. It is likely that group-oriented consideration will be more readily accepted and effective in collectivistic societies. Transformational Theory also asserts that effective leadership includes the exercise of intellectual stimulation, which involves encouraging subordinates to be independent and to approach problems in new ways. This leader behavior reflects the achievement and entrepreneurial orientation of the U.S. culture and may violate norms of dependency and conformity which characterize many other cultures.

There is also a need for identification of emic manifestations of generic leader behaviors in all cultures to which theories of leadership might be applied. It is very doubtful that the generic leadership functions adequately describe the exercise of leadership in all cultures. For example, in the Asian, Scandinavian, and, Dutch cultures, the expression of individuality is considered socially undesirable. In these cultures, singling out individuals with public praise is likely to result in embarrassment, rather than gratification. Similarly, the qualitative GLOBE research described earlier suggests that in these cultures, highly assertive behavior is also considered socially undesirable, whereas in the U.S. and in Eastern European countries, such behavior appears to be not only accepted but, expected of leaders.

In the U.S., superior-subordinate relationships that foster independence and allow subordinates to experience autonomy and openness are most generally accepted and preferred as a result of egalitarian norms (Hofstede, 1980a, 1980b; House, Wright & Aditya, 1997, in press). LMX Theory reflects this American cultural preference, and prescribes reciprocal influence between superiors and subordinates and a high degree of job autonomy for subordinates. In South East Asia, superior-subordinate relationships that are less open and that foster face saving are most generally accepted and preferred. This often involves indirect conflict resolution tactics, rather than open discussion of differences which is implied in LMX theory.

In cultures characterized by norms and widespread acceptance of high power concentration, subordinates may find job autonomy and reciprocal influence between themselves and their superiors to be incongruent with cultural norms and therefore stressful and unacceptable. In many cultures, such as Korea, supervisors are expected to be paternalistic toward subordinates (Kim, 1994). Among other attributes, paternalism involves dependence of subordinates on their supervisors for satisfaction of many of their personal as well as job-related needs. Thus, job autonomy, as recommended by LMX theory, may in fact violate Korean cultural norms. When applying or conducting research on LMX theory in cultures other than Anglo-Saxon cultures, it would be advisable to first determine the kinds of superior-subordinate relationships that are considered acceptable and supportive, and that have the theoretical effects asserted by LMX Theory. One could then trace the origins of culturally specific high-quality LMX to discover the attributes and behaviors of both supervisors and subordinates that lead to such relationships.

It is possible, even likely, that the kinds of leader behaviors required to foster high-quality superior-subordinate relationships will vary from one culture to another as well (House et al., 1997). For example, as discussed above in the section entitled Generic Leadership Functions and Specific Leader Behaviors, rather radically different behaviors are considered to be employee-oriented and supportive in the United States, Japan, and Korea. While the LMX studies conducted in Japan (Wakabayashi, Graen, Graen & Graen, 1988) found positive effects of LMX on outcomes, these studies did not investigate specific leader behaviors that led to high-quality LMX.

Cognitive Resource Theory assumes that reactions to stress are rather universal. However, in societies that emphasize suppression of emotions and educate members to engage in emotional control, this assumption may not hold. Fiedler (1996), it may be recalled, reported that an experimentally administered stress reduction program increased the performance of officer candidates on an in-basket management simulation task. It may well be that, in cultures in which individuals are socialized to engage in meditation, the effects of stress asserted in Cognitive Resource Theory may not occur.

Available evidence, discussed above, suggests that the generic leadership functions specified in the neocharismatic leadership paradigm may be universal, or near universal, under high-levels of uncertainty or stress. Further, the specific emic leader behaviors by which the neocharismatic leadership functions are enacted have not yet been identified. This possibility remains to be investigated more rigorously.

The Cumulative Gain

The social scientific study of leadership began in the early 1930s. Sixty years is a relatively short time for the history of a scientific endeavor. The development of a body of informed and empirically supported leadership literature has been truly cumulative. Early trait research was shown, in retrospect, to have identified traits that have some claim to universality, at least in the U.S. (Lord et al., 1986). Early research by Barnlund (1962) led to the identification of leader flexibility and sensitivity to followers as a predictor of leader emergence (Zaccaro et al., 1991). The behavioral school identified two generic dimensions of leader behavior, task- and person-oriented leadership, that continue to have importance in accounting for leader effectiveness (Bass, 1990). The Contingency Theory (Fiedler, 1967) led to the development of Cognitive Resource Theory. Weber's conceptualization of charisma and Path-Goal Theory led to conceptualization of the 1976 Theory of Charismatic Leadership, which in turn contributed to the development of the Value-Based Leadership Theory. McClelland's theory of individual motivation led to the development of the Leader Motive Profile. Conceptual contributions by Burns (1978) and Bennis and Nanus (1985) led to operationalizations of transformational leadership by Bass (1985) and visionary leadership by Sashkin (1988).

In the early years, leadership research was, of necessity, almost completely atheoretical and purely inductive. It is decidedly more theoretical at the present time, and current trait theory explains earlier trait findings. While task- and person-oriented behaviors have continued to be important to leadership, these have been supplemented with the behaviors specified in the neocharismatic leadership paradigm. With the advent of the neocharismatic paradigm came a broadening of the domain of scientific leadership research. Attention became focused on symbolic and emotionally appealing, as well as supportive and instrumental, leader behavior. We now have several theories of leadership, each of which explain, describe, and predict an important aspect of leadership. All of the prevailing theories of leadership reviewed in this article offer useful insights about the leadership phenomenon, and most enjoy at least some empirical support. All of the prevailing theories suggest new directions for further investigation.

Early trait and behavioral approaches to the study of leadership contributed in important ways that have to some extent gone unnoticed or unappreciated by current scholars. We have called attention to these contributions and have attempted to show how they can, with additional research and theoretical development, be incorporated into current theories. The early contingency theories stimulated a substantial amount of empirical research which further advanced our understanding and led to the formulation of better theories.

Some Prevailing Problems

A problem with the current study of leadership is that it continues to focus excessively on superior-subordinate relationships to the exclusion of several functions that leaders perform and to the exclusion of organizational and environmental variables that are crucial to effective leadership performance. A second major problem is that scholars of the traditional management and leadership literatures seldom take advantage of each others' contributions and, consequently, these two literatures are not adequately integrated. A third problem is that, historically, leadership research has been primarily concerned with generic leadership functions, to the exclusion of specific behavioral manifestations of these functions. Further, the diverse styles (mannerisms) by which leader behaviors are enacted have been largely ignored. The result is that much of our understanding about leadership is not easily operationalized in practical settings. Finally, most of the investigations concerning the prevailing theories are conducted by the authors of the theories or their associates. When conducting meta-analyses, authorship of this kind should be examined as a moderating variable.

Despite the problems, we expect the accumulation of knowledge about leadership to continue. The several issues and topics for further research identified in this article will be, we hope, of use to scholars in extending the study of leadership to new horizons.

Notes

1. The literature on overconfidence suggests a nonlinear relationship between confidence and accuracy of estimates, in the form of an inverted U, where the peak of the curve would be the beginning of overconfidence. Our discussion here pertains to the first half of the curve.

2. The Global Leadership and Organizational Behavior Effectiveness Program (GLOBE) is directed toward understanding how societal cultural variables influence organizational values, organizational practices, culturally endorsed norms, implicit theories relevant to leadership, and the exercise of leadership in each of the cultures studied. One hundred sixty country co-investigators from 60 countries are participating in the GLOBE program. GLOBE is funded by the Dwight D. Eisenhower Development Program of the U.S. Department of Education.

References

Aditya, R.N. (1994). Managing the multinational sales force - Cross cultural exchange in the dyad. Pp. 178-183 in Proceedings of the American Marketing Association Summer Educators' Conference. San Francisco, CA: American Marketing Association.

Anderson, L.R. (1983). Management of the mixed-culture work group. Organizational Behavior and Human Performance, 31: 303-330.

Antoinette, P. (1992). Leader-member exchange quality: The relationship of similarity, competence, and selected personality variables. Doctoral Dissertation, The Louisiana State University & Agricultural & Mechanical Col. Dissertation Abstracts International, 53(9), 3286A.

Aronson, E. & Carlsmith, J.M. (1968). Experimentation in social psychology. Pp. 1-79 in G. Lindzey & E. Aronson (Eds.), The handbook of social psychology, 2nd ed., Vol. 2. Reading, MA: Addison-Wesley.

Ashour, A.S. (1973). Further discussion of Fiedler's contingency model of leadership effectiveness: An evaluation. Organizational Behavior and Human Performance, 9: 339-355.

Astin, H.S. & Astin, A.W. (1996). A social change model of leadership development guidebook, version 3. Los Angeles, CA: Higher Education Research Institute, University of California, Los Angeles.

Ayman, R. (1993). Leadership perception: The role of gender and culture. In M. M. Chemers & R. Ayman (Eds.), Leadership theory and research. New York: Academic Press.

Bacharach, S.B. & Lawler, E.J. (1980). Power and politics in organizations - The social psychology of conflict, coalitions, and bargaining. San Francisco: Jossey-Bass.

Bales, R.F. (1954). In conference. Harvard Business Review, 32(2): 44-50.

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.

Barnlund, D.C. (1962). Consistency of emergent leadership in groups with changing tasks and members. Speech Monographs, 29: 45-52.

Barrick, M.R. & Mount, M.K. (1993). Autonomy as a moderator of the relationships between the Big Five personality dimensions and job performance. Journal of Applied Psychology, 78(1): 111-118.

Bass, B.M. (1957). Leadership opinions and related characteristics of salesmen and sales supervisors. Pp. 134-139 in R.M. Stogdill & A.E. Coons (Eds.), Leader behavior: Its description and measurement. Research Monograph No. 88, Bureau of Business Research, Ohio State University.

Bass, B.M. (1985). Leadership and performance beyond expectations. New York: Free Press.

Bass, B.M. (1990). Bass and Stogdill's Handbook of leadership: Theory, research and managerial applications. (3rd ed.). New York: Free Press.

Bass, B.M. (1997). Does the transactional-transformational paradigm transcend organizational and national boundaries? American Psychologist, 22(2): 130-142.

Bass, B.M. & Avolio, B.J. (1989). The multifactor leadership questionnaire. Palo Alto, CA: Consulting Psychologists Press.

Bass, B.M., Burger, P.C., Doktor, R. & Barrett, G.V. (1979). Assessment of managers: An international comparison. New York: Free Press.

Bem, D. & Allen, A. (1974). On predicting some of the people some of the time: The search for cross situational consistencies in behavior. Psychological Review, 81: 506-520.

Bennis, W. & Nanus, B. (1985). Leaders: The strategies for taking charge. New York: Harper & Row.

Besco, R.O. & Lawshe, C.H. (1959). Foreman leadership as perceived by superiors and subordinates. Personnel Psychology, (12): 573-582.

Bhohnisch, W., Jago, A.G. & Reber, G. (1987). Predicting Austrian leader behavior from a U.S. measure of behavioral intent: A cross-cultural replication. Manuscript submitted for publication.

Bouchard, T.J., Jr., Lykken, D.T., McGue, M., Segal, N.L. & Tellegen, A. (1990). Sources of human psychological differences: The Minnesota study of twins reared apart. Science, 250: 223-228.

Bowers D.G. & Seashore, S.E. (1966). Predicting organizational effectiveness with a four-factor theory of leadership. Administrative Science Quarterly, 11: 238-263.

Bryman, A. (1993). Charismatic leadership in business organizations: Some neglected issues. Leadership Quarterly, 4(3/4): 289-304.

Burke, M.J. & Day, R.R. (1986). A cumulative study of training. Journal of Applied Psychology, 71: 232-265.

Burns, J.M. (1978) Leadership. New York: Harper & Row.

Campbell, D.C. (1956). Leadership and its effects on the group. Monograph No. 83, Bureau of Business Research, Columbus, OH: Ohio State University.

Chemers, M.M., Hayes, R., Rhodewalt, F. & Wysocki, J. (1985). A person-environment analysis of job stress: A contingency model explanation. Journal of Personality & Social Psychology, 49(3): 628-635.

Chemers, M.M., Oskamp, S. & Costanzo, M.A. (1995). Diversity in organizations. Thousand Oaks, CA: Sage.

Chen, C.C. & Van Velsor, E. (1996). New directions for research and practice in diversity leadership. Leadership Quarterly, 7(2): 285-302.

Christie, R. & Geis, F.L., (1970). Studies in Machiavellianism. New York: Academic Press.

Conger, J.A. & Kanungo, R.A. (1987). Toward a behavioral theory of charismatic leadership in organizational settings. Academy of Management Review, 12: 637-647.

Cronshaw, S.F. & Lord, G.R. (1987). Effects of categorization, attribution, and encoding processes on leadership perceptions. Journal of Applied Psychology, 72(1): 97-106.

Curphy, G.J. (1990). An empirical evaluation of Bass' (1985) theory of transformational transactional leadership, Unpublished Doctoral Dissertation, University of Minnesota.

Dansereau, F., Graen, G.B. & Haga, W. (1975). A vertical dyad linkage approach to leadership in formal organizations. Organizational Behavior and Human Performance, 13: 46-78.

Davis-Blake, A. & Pfeffer, J. (1988). Just a mirage: The search for dispositional effects in organizational research. Academy of Management Review, 14(3): 385-400.

Deluga, R.J. (1994). Supervisor trust building, leader-member exchange and organizational citizenship behavior. Journal of Occupational & Organizational Psychology, 67: 315-326.

Duarte, N.T., Goodson, J.R. & Klich, N.R. (1993). How do I like thee? Let me appraise the ways. Journal of Organizational Behavior, 14: 239-249.

Duarte, N.T., Goodson, J.R. & Klich, N.R. (1994). Effects of dyadic quality and duration on performance appraisal. Academy of Management Journal, 37: 499-521.

Dunegan, K.J., Duchon, D. & Uhl-Bien, M. (1992). Examining the link between leader-member exchange and subordinate performance - The role of task analyzability and variety as moderators. Journal of Management, 18: 59-76.

Eagleson, G. (1996).Notes for the workshop on leadership development. Unpublished manuscript, Australian Graduate School of Management, Sidney, Australia.

Eden, D. (1990). Pygmalion in management: Productivity as a self-fulfilling prophecy. Lexington, MA: Lexington Books.

Eisenhower Leadership Group. (1996). Democracy at risk: How schools can lead. Washington, D.C.: United States Department of Education.

Elderkin, K.W. & Bartlett, C.A. (1993). General electric: Jack Welch's second wave (A). Harvard Business School, case number: 9-391-248.

Evans, M.G. (1996). R.J. House's Path-goal theory of leader effectiveness. The Leadership Quarterly, 7(3): 305-309.

Fayol, H. (1949). General and industrial management. London: Isaac Pitman.

Ferris, G.R. (1985). Role of leadership in the employee withdrawal process: A constructive replication. Journal of Applied Psychology, 70(4): 777-781.

Fiedler, F.E. (1995) Cognitive resources and leadership performance. Applied Psychology - An International Review, 44: 5-28.

Fiedler, F.E. & Chemers, M.M. (1984). Improving leadership effectiveness: The leader-match concept. New York: Wiley.

Fiedler, F.E. (1967). A theory of leadership effectiveness. New York: McGraw-Hill.

Fiedler, F.E. (1971). Validation and extention of the contingency model of leadership effectiveness: A review of empirical findings. Psychological Bulletin, 76: 128-148.

Fiedler, F.E. (1973). The contingency model: A reply to Ashour. Organizational Performance and Human Behavior, 9: 356-368.

Fiedler, F.E. (1977). A rejoinder to Schriesheim and Kerr's premature obituary of the Contingency Model. Pp. 45-50 in J.G. Hunt & L.L. Larson (Eds.), Leadership: The cutting edge. Carbondale, IL: Southern Illinois University Press.

Fiedler, F.E. (1995). Cognitive resources and leadership performance. Applied Psychology - An International Review, 44: 5-28.

Fiedler, F.E. (1996). Research on leadership selection and training: One view of the future. Administrative Science Quarterly, 41: 241-250.

Fiedler, F.E. Personal communication. May, 1997.

Fiedler, F.E. & Garcia, J.E. (1987). New approaches to effective leadership: Cognitive resources and organizational performance. New York: Wiley.

Field, R.H.G. (1979). A critique of the Vroom-Yetton contingency model of leadership behavior. Academy of Management Review, 4: 249-257.

Field, R.H.G. & House, R.J. (1990). A test of the Vroom-Yetton model using manager and subordinate reports. Journal of Applied Psychology, 75:(3) 362-366.

Field, R.H.G. (1982). A test of the Vroom-Yetton normative model of leadership. Journal of Applied Psychology, 67: 523-532.

Filley, A.C., House, R.J. & Kerr, S. (1976). Managerial process and organizational behavior. Glenview, IL: Scott, Foresman, and Co.

Finkelstein, S. & Hambrick, D. (1996). Strategic leadership: Top executives and their effects on organizations. 457. St. Paul, MN: West Publishing.

Fleishman, E.A., Harris, E.F. & Bunt, H.E. (1955). Leadership and supervision in industry. Columbus, OH: Bureau of Educational Research, Ohio State University.

Gardner, H. (1983). Frames of mind: The theory of multiple intelligences. New York: Basic Books.

Gardner, H. (1995). Leading minds: An anatomy of leadership. New York: Basic Books.

Geyer, A.L.J. & Steyrer, J. (1994). Transformationale Fuhrung, klassische Fuhrungstheorien und Erfolgsindikatoren von Bankbetrieben. Zeitschrift fur Betriebswirtschaft 64(8): 961-979.

Gibb, C.A. (1947). The principles and traits of leadership. Journal of Abnormal and Social Psychology, 42: 267-284.

Graen, G. & Cashman, J.F. (1975). A role-making model of leadership in formal organizations: A developmental approach. Pp. 143-165 in J.G. Hunt & L.L. Larson (Eds.), Leadership frontiers. Kent, OH: Kent State University Press.

Graen, G., Novak, M.A. & Sommerkamp, P. (1982). The effects of Leader-Member Exchange and job design on productivity and satisfaction: Testing a dual attachment model. Organizational Behavior and Human Performance, 30: 109-131.

Graen, G.B. & Scandura, T. (1987). Toward a psychology of dyadic organizing. Pp. 175-208 in B. Staw & L.L. Cummings (Eds.), Research in organizational behavior, Vol. 9. Greenwich, CT: JAI Press.

Graen, G.B. & Uhl-Bien, M. (1995). Relationship-based approach to leadership: Development of leader-member exchange (LMX) theory of leadership over 25 years: Applying a multi-level multi-domain perspective. Leadership Quarterly, 6(2): 219-247.

Graen, G.B., Liden, R.C. & Hoel, W. (1982). Role of leadership in employee withdrawal process. Journal of Applied Psychology, 67(6): 868-872.

Graen, G.B., Scandura, T. & Graen, M.R. (1986). A field experimental test of the moderating effects of growth need strength on productivity. Journal of Applied Psychology, 71: 484-491.

Graham, W.K. & Gleno, T. (1970). Perception of leader behavior and evaluation of leaders across organizational levels. Experimental Publication System, Ms. 144A: Issue 4.

Hatter, J.J. & Bass, B.M. (1988). Supervisor's evaluations and subordinates' perceptions of transformational leadership. Journal of Applied Psychology, 73(3): 695-702.

Heilman, M.E., Hornstein, H.A., Cage, J.H., & Herschlag, J.K. (1984). Reactions to prescribed leader behavior as a function of role-perspective: The case of the Vroom-Yetton model. Journal of Applied Science, 69: 50-60.

Hersey, P. & Blanchard, K. (1982). Management of organizational behavior: Utilizing human resources. Englewood Cliffs, NJ: Prentice-Hall.

Hickson, D.J., Hinnings, C.R., Lee, C.A., Schneck, R.H. & Pennings, J.M. (1971). A strategic contingencies' theory of intraorganizational power. Administrative Science Quarterly, 16: 216-229.c

Hofstede, G. (1980a). Culture's consequences: International differences in work related values. Beverly Hills, CA: Sage.

Hofstede, G. (1980b). Motivation, leadership and organization: Do American theories apply abroad? Organizational Dynamics, 9(1): 42-63.

House, R.J., Delbecq, A.L. & Taris, T. (1997). Value based leadership: A theory and an empirical test. Working paper, Reginald H. Jones Center for Strategic Management, Wharton School of Management.

House, R.J., Shane, S. & Herold, D. (1996). Rumors of the death of dispositional theory and research in Organizational Behavior are greatly exaggerated. Academy of Management Review, 21(1): 203-224.

House, R.J. (1960). An experiment in the use of selected methods for improving the effectiveness of communication training for management. Unpublished Doctoral Dissertation. Ohio State University.

House, R.J. (1970). Role conflict and multiple authority in complex organizations. California Management Review, 12(4): 53-60.

House, R.J. (1971). A path goal theory of leader effectiveness. Administrative Science Quarterly, 16: 321-338.

House, R.J. (1977). A 1976 theory of charismatic leadership. Pp 189-207 in J.G. Hunt and L.L. Larson (Eds.), Leadership: The cutting edge. Carbondale, IL: Southern Illinois University Press.

House, R.J. (1988). Power and personality in complex organizations. Pp. 305-357 in B.M. Staw & L.L. Cummings (Eds.), Research in organizational behavior, Vol. 10. Greenwich, CT: JAI Press.

House, R.J. (1991). The distribution and exercise of power in complex organizations: A MESO-theory. Leadership Quarterly, 2(1): 23-58.

House, R.J. (1996). Path-goal theory of leadership: Lessons, legacy and a reformulated theory. The Leadership Quarterly, 7(3): 323-352.

House, R.J. & Baetz, M.L. (1979). Leadership: Some empirical generalizations and new research directions. Pp. 341-423 in B. Staw (Ed.), Research In organizational behavior, Vol. 1. Greenwich, CT: JAI Press.

House, R.J. & Miner, J.M. (1969). Span of control-group size interaction. Administrative Science Quarterly, 14(3): 451-464.

House, R.J. & Mitchell, T.R. (1974). Path-goal theory of leadership. Journal of Contemporary Business, 3: 81-97.

House, R.J., Spangler, D. & Woycke, J. (1991). Personality and charisma in the U.S. Presidency: A psychological theory of leadership effectiveness. Administrative Science Quarterly, 36: 364-396.

House, R.J. & Shamir, B. (1993). Towards the integration of transformational, charismatic and visionary theories. Pp. 81-107 in M.M. Chemers & R.Ayman (Eds.). Leadership theory and research: Perspectives and directions. San Diego, CA: Academic Press.

House, R.J. & Singh, J.V. (1987). Organizational behavior: Some new directions for I/O psychology. Annual Review of Psychology, 38: 669-718.

House, R.J., Wright, N. & Aditya, R.N. (1997). Cross cultural research on organizational leadership: A critical analysis and a proposed theory. In P.C. Earley & M. Erez (Eds.). New perspectives on international industrial and organizational psychology. San Francisco, CA: Jossey-Bass.

Howell, J. & Avolio, B.J. (1993). Transformational leadership, transactional leadership, locus of control, and support for innovation: Key predictors of consolidated-business-unit performance. Journal of Applied Psychology, 78(4): 891-902.

Howell, J.M. & House, R.J. (1992). Socialized and personalized charisma: An essay on the bright and dark sides of leadership. Unpublished manuscript. School of Business Administration, The University of Western Ontario.

Howell, J.M. & Higgins, C. (1990). Champions of technological innovation. Administrative Science Quarterly, 35: 317-341.

Howell, J.M. & Frost, P.J. (1989). A laboratory study of charismatic leadership. Organizational Behavior and Human Decision Processes, 43(2): 243-269.

Jackson, S.E. & Ruderman, M.N. (Eds.). (1995). Diversity in work teams. Washington, DC: American Psychological Association.

Javidan, M. & Carl, D. (1997). Motivational consequences of charismatic leadership: An empirical investigation.

Jenkins, W. O. (1947). A review of leadership studies with particular reference to military problems. Psychological Bulletin, 44: 54-79.

Judge, T.A. & Ferris, G.R. (1993). Social context of performance evaluation decisions. Academy of Management Journal, 36(1): 80-105.

Kahn, R.L. & Katz, D. (1953). Leadership practices in relation to productivity and morale. In D. Cartwright & A. Zander (Eds.), Group dynamics. New York: Harper & Row.

Kenny, D.A. & Hallmark, B.W. (1992). Rotation designs in leadership research. Special issue: Individual differences and leadership: II. Leadership Quarterly, 3(1): 25-41.

Kenny, D.A. & Zaccaro, S.J. (1983). An estimate of the variance due to traits in leadership. Journal of Applied Psychology, 68(4): 678-685.

Kim, U. (1994). Significance of paternalism and communalism in the occupational welfare system of Korean firms: A national survey. In U. Kim, H.C. Triandis, C. Kagitcibasi, S-C. Choi & G. Yoon (Eds.). Individualism and collectivism: Theory, method, and applications. Newbury Park, CA: Sage.

Kinicki, A.J., & Vecchio, R.P. (1994). Influences on the quality of supervisor-subordinate relations: The role of time-pressure, organizational commitment, and locus of control. Journal of Organizational Behavior, 15: 75-82.

Kipnis, D. (1976). The powerholders. Chicago, IL: The University of Chicago Press.

Kipnis, D., Schmidt, S. & Wilkinson, I. (1980). Intraorganizational influence tactics. Journal of Applied Psychology, 65(4): 440-452.

Kipnis, D. (1984). The use of power in organizations and in interpersonal settings. Pp. 172-210 in S. Oskamp (Ed.), Applied Social Psychology Annual 5. Beverly Hills, CA: Sage.

Kipnis, D., Schmidt, S.M., Swaffin-Smith, C. & Wilkinson, I. (1984). Patterns of managerial influence: Shotgun managers, tacticians and bystanders. Organizational Dynamics, 12(3): 58-67.

Kipnis, D. (1987). Psychology and behavioral technology. American Psychologist, 42: 30-36.

Koene, H., Pennings, H. & Schreuder, M. (1993). Leadership, culture, and organizational effectiveness. In K.E. Clark & M.B. Clark, (Eds.), The impact of leadership. Greensboro, NC: The Center for Creative Leadership.

Koh, W.L., Terborg, J.R. & Steers, R.M. (1991). The impact of transformational leadership on organizational commitment: Organizational citizenship behavior, teacher satisfaction and student performance in Singapore. Paper presented at the Academy of Management, Fontenbleu, FL.

Kousnes, J.M. & Posner, B.Z. (1987). The leadership challenge: How to get extraordinary things done in organizations. San Francisco: Jossey-Bass.

Kouzes, J.M., & Posner, B.Z. (1987). The Leadership Challenge: How to Get Extraordinary. Things done in Organizations. San Francisco: Jossey Bass.

Lagace, R.R. (1990). Leader-member exchange: Antecedents and consequences of the cadre and hired hand. Journal of Personal Selling and Sales Management, 10(1): 11-19.

Larson, L.L., Hunt, J.G. & Osborn, R.N. (1974). Correlates of leadership and demographic variables in three organizational settings. Journal of Business Research, 2: 335-347.

Latham, G.P. & Saari, L.M. (1979). Importance of supportive relationships in goal setting. Journal of Applied Psychology, 64(2): 151-156.

Lee, C., Asford & Bobko, P. (1990). Interactive effects of "type A" behavior and perceived control on worker performance, job satisfaction, and somatic complaints. Academy of Management Journal, 33: 870-881.

Liden, R., Wayne, S. & Stilwell, S. (1993). A longitudinal study of the early development of leader-member exchanges. Journal of Applied Psychology, 78 (4): 662-674.

Likert, R. (1961). New patterns of management. New York: McGraw-Hill.

Lindholm, C. (1990). Charisma. Cambridge, MA: Basil Blackwell.

Litwin, G.H. & Stringer, R.A., Jr. (1968). Motivation and organizational climate. Boston: Harvard Business School Press.

Loden, M. & Loeser, R.H. (1991). Working diversity: Managing the differences. Bureaucrat, 20(1): 21-25.

Lord, R.G. & Maher, K.J. (1991). Leadership and information processing: Linking perception and performance. Boston: Unwin Hyman.

Lord, R.G., Binning, J.F., Rush, M.C. & Thomas, J.C. (1978). The effects of performance cues and leader behavior on questionnaire ratings of leadership behavior. Organizational Behavior & Human Performance, 21 (1): 27-39.

Lord, R.G., DeVader, C.L. & Alliger, G.M. (1986). A meta-analysis of the relation between personality traits and leadership perceptions: An application of validity generalizations procedures. Journal of Applied Psychology, 71(3): 402-410.

Lord, R.G., Foti, R. & DeVader, C.L. (1984). A test of leadership categorization theory: Internal structure, information processing, and leadership perceptions. Organizational Behavior and Human Performance, 34: 343-378.

Lowe, K.B., Kroeck, K.G. & Sivasubramaniam, N. (1996) Effectiveness correlates of transformational and transactional leadership: A meta-analytic review of the MLQ literature. Leadership Quarterly, 7(3): 385-425.

Manogran, P. & Conlon, E.J. (1993). A leader-member exchange approach to explaining organizational citizenship behavior. Pp. 249-253 in Proceedings of the Annual Meeting of Academy of Management. Atlanta, GA: Academy of Management.

Mann, F.C. (1965). Toward an understanding of the leadership role in formal organization. Pp. 68-103 in R. Dublin (Ed.), Leadership and productivity. San Francisco: Chandler.

Mann, R.D. (1959). A review of the relationship between personality and performance in small groups. Psychological Bulletin, 56: 241-270.

Margerison, C. & Glube, R. (1979). Leadership decision-making: An empirical test of the Vroom and Yetton Model. Journal of Management Studies, 16: 45-55.

Markus, H. & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion and motivation. Psychological Review, 98: 224-253.

McClelland, D.C. (1961). The achieving society. New York: Van Nostrand Reinhold.

McClelland, D.C. (1985). Human motivation. Glenview, IL: Scott, Foresman.

McClelland, D.C. & Boyatzis, R.E. (1982). Leadership motive pattern and long-term success in management. Journal of Applied Psychology, 67(6): 737-743.

McClelland, D.C. & Burnham, D.H. (1976). Power is the great motivator. Harvard Business Review, 54(2): 100-110.

McClelland, D.C. (1975). Power: The inner experience. N.Y.: Irvington.

McClelland, D.C., Atkinson, J.W., Clark, R.A. & Lowell, E.L. (1958). The achievement motive. New York: Appleton-Century-Crofts. Reissued with new introduction by John W. Atkinson, New York: Irvington, 1976.

Messallam, A. & House, R.J. (1997). A test of the value based theory of leadership based on Egyptian CEOs. In preparation.

Miner, J.B. & Dachler, H.P. (1973). Personnel attitudes and motivations. Annual Review of Psychology, 24: 379-402.

Mintzberg, H. (1973). The nature of managerial work. New York: Harper & Row.

Mintzberg, H. (1994). Rise and fall of strategic planning: Reconceiving roles for planning, plans, and planners. New York: Free Press.

Mischel, W. (1973). Toward a cognitive social learning reconceptualization of personality. Psychological Review, 80: 252-283.

Misumi, J. (1985). The behavioral science of leadership: An interdisciplinary Japanese research program. Ann Arbor, MI: University of Michigan Press.c

Monson, T.C., Hesley, J.W. & Chernick, L. (1982). Specifying when personality traits can and cannot predict behavior: An alternative to abandoning the attempt to predict single act criteria. Journal of Personality and Social Psychology, 3: 385-499.

Morrison, Ann M. (1992). Developing diversity and organization. Business Quarterly, 57,(1): 42-48.

Nanus, B. (1992). Visionary leadership - Creating a compelling sense of direction for your organization. San Francisco: Jossey-Bass.

Pathet, C.M. & Lau, A.W. (1982). Managerial roles, skills and effective performance. Proceedings, Academy of Management, New York, 95-99.

Pereira, D. (1987). Factors associated with transformational leadership in an Indian engineering firm. Paper presented at Administrative Science Association of Canada held in Vancouver.

Peters, L.H., Hartke, D.D. & Pohlman, J.T. (1985). Fiedler's contingency model of leadership: An application of the meta-analysis procedure of Schmidt and Hunter. Psychological Bulletin, 97: 274-285.

Pfeffer, J. (1981). Power in organizations. Massachusetts: Pitman Publishing Inc.

Pillai, R. & Meindl, J. R. (1991). The effects of a crisis on the emergence of charismatic leadership: A laboratory study. Pp. 420-425 in Best Paper Proceedings, Annual Meeting of the Academy of Management, Miami, FL.

Podsakoff, P.M., MacKenzie, S. B. & Hui, C. (1993). Organizational citizenship behaviors and managerial evaluations of employee performance: A review and suggestions for future research. Pp. 1-40 in G.R. Ferris (Ed.), Research in personnel and human resources management, Vol. 11. Greenwich, CT: JAI Press.

Podsakoff, P.M., MacKenzie, S.B., Moorman, R.H. & Fetter, R. (1990). Transformational leader behaviors and their effects on followers' trust in leader, satisfaction, and organizational citizenship behaviors. Leadership Quarterly, 1: 107-42, 94-98, 140-143.

Porter, L., Allen, R.W. & Angle, H.L. (1981). The politics of upward influence in organizations. In B. Staw & L.L. Cummings (Eds.), Research in organizational behavior, Vol. 3. Greenwich, CT: JAI Press.

Porter, M.E. (1985). Competitive advantage: Creating and sustaining superior performance. New York: Free Press.

Porter, M.E. (Ed.) (1986). Competition in global industries. Boston: Harvard Business School Press.

Roberts, N.C. (1985). Transforming leadership: A process of collective action. Human Relations, 38: 1023-1046.

Sales, S. (1972). Authoritarianism: But as for me, give me liberty, or give me maybe, a great big, strong, powerful leader I can honor, admire, respect and obey, Psychology Today: 94-98, 140-143.

Sashkin, M. (1988). The visionary leader. Pp. 122-160 in J.A. Conger & R.A. Kanungo (Eds.), Charismatic leadership: The elusive factor in organizational effectiveness. San Francisco: Jossey-Bass.

Scandura, T. & Graen, G.B. (1984). Moderating effects of initial leader-member exchange status on the effects of a leadership intervention. Journal of Applied Psychology, 69(3): 428-436.

Scandura, T.A., Graen, G.B. & Novak, M.A. (1986). When managers decide not to decide autocratically. Journal of Applied Psychology, 71(4): 579-584.

Schlender, B. (1995). What Bill Gates really wants. Fortune Magazine, 131(1): 34-36.

Schmidt, S.M. & Yeh, R. (1992). The structure of leader influence: A cross-national comparison. Journal of Cross-Cultural Psychology, 23(2): 251-264.

Schneider, B. (1983). Interactional psychology & organizational behavior. Pp 1-31 in L.L. Cummings & B.M. Staw, (Eds.). Research in organizational behavior, Vol. 5. Greenwich, CT: JAI Press.

Schriesheim, C. & Neider, L.L. (1996). Path-goal leadership theory: The long and winding road. The Leadership Quarterly, 7(3): 317-321.

Schriesheim, C., House, R.J. & Kerr, S. (1976). Leader initiating structure: A reconciliation of discrepant research results and some empirical tests. Organizational Behavior and Human Performance, 15(2): 297-321.

Schriesheim, C. & Kerr, S. (1977). Theories and measures of leadership: A critical appraisal of present and future directions. Pp. 9-44 in J.C. Hunt & L.L. Larson (Eds.), Leadership: The cutting edge. Carbondale, IL: Southern Illinois University Press.

Shamir, B., House, R.J. & Arthur, M. (1993). The motivational effects of charismatic leadership. Organizational Science, 4(4): 577-594.

Shamir, B. (1991). The charismatic relationship: Alternative explanations and predictions. Leadership Quarterly, 2: 81-104.

Sherif, C.W., Sherif, M. & Nebergale, R.E. (1965). Attitude and attitude change. Philadelphia, PA: W.B. Saunders Co.

Shirakashi, S. (1980). The interaction effects for behavior of least preferred coworker (LPC) score and group-task situations: A reanalysis. Commercial Review of Seinan Gakuin University.

Simon, H.A. (1987). Making management decisions: The role of intuition and emotion. Academy of Management Executive, 1: 57-64.

Simonton, D.K. (1987). Presidential inflexibility and veto behavior: Two individual-situational interactions. Journal of Personality, 55(1): 1-18.

Simonton, D.K. (1994). Greatness: Who makes history and why. New York: Guilford.

Smith, J.E., Carson, K.P. & Alexander, R.A. (1984). Leadership: It can make a difference. Academy of Management Journal, 27: 765-776.

Smith, P.B. & Peterson, M.F. (1994). Leadership as event management: A cross cultural survey based on managers from 25 nations. Paper presented at the International Congress of Psychology, Madrid.

Smith, P.B., Misumi, J., Tayeb, M.H., Paterson, M. & Bond, M.H. (1989). On the generality of leadership style across cultures. Journal of Occupational Psychology, 62: 97-110.

Snyder, M. (1974). Self-monitoring of expressive behavior. Journal of Personality and Social Psychology, 30: 526-537.

Sorcher, M. & Goldstein, A.P. (1972). A behavior modeling approach in training. Personnel Administration, 35: 35-41.

Spangler, W.D. (1992). The validity of questionnaire and TAT measures of need for achievement: Two meta analyses, Psychological Bulletin, 112(1): 140-154.

Spangler, W.D. & House, R.J. (1991). Presidential effectiveness and the leadership motive profile, Journal of Personality and Social Psychology, 60(3): 439-455.

Staw, B.M., Sandelands, L.E. & Dutton, J.E. (1981). Threat-rigidity effects in organizational behavior: A multilevel analysis. Administrative Science Quarterly, 26:501-524.

Sternberg, R. (1988). The triarchic mind: A new theory of intelligence. New York: Viking-Penguin.

Stilwell, C.D. (1993). Initial performance expectations, strength of those expectations, and social interaction behavior of subordinates: Their effect on the development process in leader-member exchange theory. Dissertation Abstracts International, 54: 596.

Stogdill, R.M. & Coons, A.E. (1957). Leader behavior: Its description and measurement. Columbus, OH: Ohio State University Press for Bureau of Business Research.

Stogdill, R.M. (1948). Personal factors associated with leadership: A survey of the literature. Journal of Psychology, 25: 35-71.

Stogdill, R.M. (1974). Handbook of leadership: A survey of theory and research. New York: Free Press.

Stogdill, R.M. & Coons, A.E. (1957). Leader behavior: Its description and measurement. Columbus, OH: Ohio State University Press for Bureau of Business Research.

Strube, M.J. & Garcia, J.E. (1981). A meta-analytical investigation of Fiedler's contingency model of leadership effectiveness. Psychological Bulletin, 90: 307-321.

Strube, M.J. & Garcia, J.E. (1983). On the proper interpretation of empirical findings: Stube and Garcia (1981) revisited. Psychological Bulletin, 93(3): 600-603.

Sykes, A.J.M. (1962). The effects of a supervisory training course in changing supervisors' perceptions and expectations of the role of management. Human Relations, 15(3): 227-243.

Tannenbaum, R. & Schmidt, W.H. (1973). How to choose a leadership pattern. Harvard Business Review, 51:162-180.

Thomas, A.B. (1988). Does leadership make a difference to organizational performance? Administrative Science Quarterly, 33: 588-400.

Tjosvold, D., Wedley, W.C., & Field, R.H.G. (1986). Constructive controversy, the Vroom-Yetton model, and managerial decision-making. Journal of Occupational Behavior, 7:125-138.

Triandis, H.C. (1995). Individualism & collectivism. Boulder, CO: Westview Press.

Triandis, H.C., Dunnette, M.D. & Hough, L.M. (Eds.) (1994). Handbook of industrial and organizational psychology, Vol 4, 2nd ed. Palo Alto, CA: Consulting Psychologists Press, Inc.

Trice, H.M. & Beyer, J.M. (1986). Charisma and its routinization in two social movement organizations. Pp. 113-164 in B.M. Staw & L.L. Cummings (Eds.). Research in organizational behavior, Vol. 8. Greenwich, CT: JAI Press.

Turban, D.B., Jones, A.P. & Rozelle, R.M. (1990). Influences of supervisor liking of a subordinate and the reward context on the treatment and evaluation of that subordinate. Motivation & Emotion, 14: 215-233.

Vecchio, R.P. (1977). An empirical examination of the validity of Fiedler's model of leadership effectiveness. Organizational Behavior & Human Performance, 19(1): 180-206.

Vecchio, R.P. (1982). A further test of leadership effects due to between-group variation and within-group variation. Journal of Applied Psychology, 67(2): 200-208.

Vecchio, R.P. (1983). Assessing the validity of Fiedler's contingency model of leader effectiveness: A closer look at Strube and Garcia. Psychological Bulletin, 93(2): 404-408.

Vecchio, R.P. (1985). Predicting employee turnover from Leader-Member Exchange: A failure to replicate. Academy of Management Journal, 28(2): 478-485.

Vecchio, R.P. (1987). Situational leadership theory: An examination of a prescriptive theory. Journal of Applied Psychology, 72(3): 444-451.

Vecchio, R.P. & Gobdel, B.C. (1984). The vertical dyad linkage model of leadership: Problems and prospects. Organizational Behavior and Human Performance, 34: 5-20.

Vecchio, R.P., Griffeth, R.W. & Horn, P.W. (1986). The predictive utility of the vertical dyad linkage approach. Journal of Social Psychology, 126: 617-625.

Vecchio, R.P. & Norris, W.R. (1996). Predicting employee turnover from performance, satisfaction and leader-member exchange. Journal of Business and Psychology, 49: 436-458.

Vroom, V.H. (1960). Some personality, determinants of the effects of participation. Englewood Cliffs, NJ: Prentice-Hall.

Vroom, V.H. & Jago, A.G. (1988). The new leadership: Managing participation in organizations. Englewood Cliffs, NJ: Prentice-Hall.

Vroom, V.H. & Yetton, P.W. (1973). Leadership and decision-making. Pittsburgh, PA: University of Pittsburgh Press.

Wahba, M. & Bridwell, L. (1975). Maslow reconsidered: A review of the need hierarchy theory. Organizational Behavior and Human Performance, 15(2): 212-240.

Wakabayashi, M., Graen, G.B., Graen, M.R. & Graen, M.G. (1988). Japanese management process: Mobility into middle management. Journal of Applied Psychology, 73(2): 217-227.

Wakabayashi, M. & Graen, G. (1984). The Japanese career progress: A seven year follow up. Journal of Applied Psychology, 69(4): 603-614.

Waldersee, R. & Eagleson, G. (1996). The efficacy of distributed leadership in implementing change. Unpublished paper, Australian Graduate School of Management, University of New South Wales.

Waldman, D., House, R.J. & Ramirez, G. (1996). A Replication of the effects of U.S. CEO charismatic leadership on firm profitability under conditions of certainty and uncertainty based on Canadian executives. Unpublished Manuscript, the Wharton School, University of Pennsylvania.

Waldman, D., Ramirez, G. & House, R.J. (1996). The effects of U.S. CEO leader behavior on firm profits under conditions of environmental certainty and uncertainty: A longitudinal investigation. Working paper, Reginal Jones Center for Strategic Management, Wharton School of Management.

Wayne, S.J. & Green, S.A. (1993). The effects of leader-member exchange on employee citizenship and impression management behavior. Human Relations, 46: 1431-1440.

Willner, A.R. (1984). The spellbinders.' Charismatic political leadership. New Haven, CT: Yale University Press.

Winter, D.G. & Barenbaum, N.B. (1985). Responsibility and the power motive in women and men. Journal of Personality, 53: 335-355. Reprinted in A.J. Stewart & M.B. Lykes (Eds.). (1985). Gender and personality (pp. 247-267). Durham, NC: Duke University Press.

Winter, D.G. (1973). The power motive. New York: Free Press.

Winter, D.G. (1978) Navy leadership and management competencies: Convergence among tests, interviews, and performance ratings. Boston: McBer and Company.

Winter, D.G. (1991). A motivational model of leadership: Predicting long-term management success from TAT measures of power motivation and responsibility. Leadership Quarterly, 2(2): 67-80.

Wofford, J.C. & Liska, L.Z. (1993). Path-Goal theories of leadership: A meta-analysis. Journal of Management, 19: 857-876.

Xin, K.R. & Tsui, A.S. (1996). Different strokes for different folks? Influence tactics by Asian-American and Caucasian-American managers. Leadership Quarterly, 7(1): 109-132.

Yammarino, F.J. & Dubinsky, A.J. (1992). Superior-subordinate relationships: A multiple level of analysis approach. Human Relations, 45: 575-600.

Yukl, G. (1993). A retrospective on Robert House's 1976 theory of charismatic leadership and recent revisions. Leadership Quarterly, 4(3/4): 367-373.

Yukl, G. (1994). Leadership in organizations, 3rd ed. Englewood Cliffs, NJ: Prentice Hall.

Yukl, G., Falbe, C.M. & Youn, J.Y. (1993). Patterns of influence behavior for managers. Group and Organization Management, 18: 5-24.

Zaccaro, S.J., Foti, R.J. & Kenny, D.A. (1991). Self-monitoring and trait-based variance is leadership: An investigation of leader flexibility across multiple group situations. Journal of Applied Psychology, 76(2): 308-315.

Zaleznik, A. (1977). Managers and leaders: Are they different? Harvard Business Review, 55(3): 67-78.

Zimmer, R.J. (1978). Validating the Vroom-Yetton normative model of leader behavior in field sales force management and measuring the training effects of TELOS on the leader behavior of district managers. Unpublished doctoral dissertation, Virginia Polytechnic Institute and State University.

Source Citation

Source Citation   

Gale Document Number: GALE|A20147092