LEADERSHIP

The concept of leadership has been the focus of research and discussion of scholars in a variety of disciplines. Literary authors and philosophers provided the initial descriptions and guidance for leaders of their time. With the evolution of social sciences, scholars of political science, anthropology, sociology, psychology, and business have all explored the nature of leaders and the process of leadership. Although each of these various approaches did add a different perspective through the decades, there seem to be some who perceive a merging of disciplines in understanding leadership as we approach the close of the twentieth century. However, there are others who still maintain that leaders in different settings are fundamentally different.

As leadership is a vast field of research, only the most prevalent and unique approaches will be acknowledged in the following sections. Readers can find a more thorough review of leadership research and theories in texts such as Bass (1990), Chemers (1997), Chemers and Ayman (1993), and Dansereau and Yammarino (1998).

HISTORICAL REVIEW

The definition of leadership has varied across time and cultures. In ancient times, the focus was on kings and rulers, who received guidance from philosophers. In many civilizations around the world, authors have written essays on good and bad leadership. For example, Confucius and Mencius wrote essays on leaders' proper behavior during the end of the fourth century and the beginning of the fifth century b.c. in China. Aristotle's book, The Politics, describes the characteristics of the kings and kingship in ancient Greece (fourth century b.c.). In eleventh-century Iran, Unsuru'l-Ma'ali wrote Qabus-Nameh and Nezam Mulk Tussi wrote Siyassat Nameh, advising the kings of the time in effective governance. Machiavelli wrote The Prince, in Florence, Italy, during the sixteenth century, guiding European rulers in politics. Ibn-e-khaldun from Tunisia provided his observations and guidance to the ruling groups of North Africa in his famous book Muqiddimah in the fourteenth century.

From the end of the eighteenth century to the beginning of the twentieth century, during the evolution of industrialization, the concept of "leader" also came to include leaders of industry. Only a hundred and fifty years ago the first chief executive officer (CEO) appeared (Smoler 1999). Smoler stated that the increased volume of economic activity required the increased efficiency and administrative coordination that gave rise to this kind of leadership. The twentieth century saw steep growth in the numbers of CEOs and of top executive management positions.

At the close of the twentieth century there is a debate on differentiating between leaders and managers, and between leadership and management (Kotter 1988). The saying that some leaders and managers both lead and manage, but most leaders do not manage and most managers do not lead, may make this distinction. A popular phrase that clarifies the relation between leader and managers is, "Good managers do things right and good leaders do the right thing."

At the beginning of the twentieth century, two schools of thought on leadership evolved (Ayman 1993; Bass 1990). In Britain, Thomas Carlyle (1907) presented the "great-man theory," and in Germany, Karl Marx (1906) explained leadership from a zeitgeist approach (i.e., as being in the spirit of the time). The great-man approach focuses on the unique characteristics of the individual as the basis for being a leader. In this approach there is a strong belief that some people have leadership qualities and some do not; thus the born-leader concept. In the zeitgeist approach, the assumption is that the situation provides the opportunity for the individual to become a leader. Thus, it is not the person but the circumstances or the waves of time that puts people in leadership positions. These two approaches have been the foundation of leadership research throughout the twentieth century.

As the end of the century is approaching, the fascination with leaders seems to be at an all-time high, based on the number of publications about and for leaders (Ayman 1997). Many of these publications are inspirational and guiding. The scholars of leadership maintain a steady effort to refine the conceptualization of leadership and enhance the methodologies used in the study of leadership.

The twentieth century was the beginning of scientific and systematic study of leadership. Many disciplines have contributed to this endeavor, such as political science, anthropology, sociology, social psychology, and, most recently, industrial and organizational psychology. Each discipline has explored this phenomenon uniquely, guided by its own perspective and orientation. The psychologists' focus was initially influenced by the "greatman theory," whereas the sociologists and social psychologists were more interested in the situational perspective. As the investigations proceeded and the disciplines evolved, the interplay of these approaches and the complexity of leadership became apparent. The following sections will briefly review various approaches: the trait and behavior approaches, which evolved from the greatman theory; the situational approach, which evolved from the zeitgeist approach; and, finally, the contingency approach, which combined the two approaches. At the end of this article, some challenges facing leadership practice and research are delineated.

TRAIT APPROACH

The initial work on leadership was influenced by the great-man theory, which focuses on the characteristics of individuals in leadership situations. The primary focus of this approach is to compare the characteristics of those who are leaders and those who are not. The majority of trait research began in the early part of the twentieth century and declined somewhat in the late 1940s. In a majority of these studies, there seemed to be a search for one trait that best differentiated leaders from nonleaders. Stogdill (1948) reviewed studies in various settings and identified a list of traits that were most commonly studied to identify leaders. The traits that differentiated leaders from nonleaders included "sociability, initiative, persistence, knowing how to get things done, self-confidence, alertness to and insight into situations, cooperativeness, popularity, adaptability, verbal facility" (Bass 1990, p. 75). However, Stogdill (1948) concluded that research on leader's traits is inconclusive and that, depending on the situation, one trait may be more important than another.

Early industrial psychologists interested in selection maintained their focus on the trait approach. Therefore, primary pursuit of trait approach was by practitioners who were interested in selection and succession planning of managers in work settings. Later, Lord and colleagues (1986), with the assistance of meta-analytical techniques, reanalyzed past research by Stogdill (1948) and Mann (1959) and, across studies, identified certain characteristics more associated with leaders than with nonleaders. The result of these studies was to identify intelligence, masculinity, and dominance as strongly related to leadership, whereas adaptability, introversion/extroversion, and conservatism had a weaker relationship. The adjustment and flexibility competency has received attention in recent years. In various studies, adaptability, as operationalized by the self-monitoring scale (Snyder 1979), has also been used to predict leader emergence and effectiveness with some degree of success (Zaccaro et al. 1991; Ayman and Chemers 1991). Hogan and colleagues (1994) revived interest in the trait approach by introducing a multitrait approach to the study of leaders. The traits included in the profile are known as the "big five" and consist of extroversion, emotional stability, openness, intellect, and surgency. In addition to the big five, there are also other measures (e.g., Myers Briggs) that are used to identify leader's profiles. However, results for predicting effective leaders from these traits are inconclusive. Overall, most of the trait research has focused on differentiating the characteristics of leaders from nonleaders, and on measuring leadership potential.

BEHAVIORAL APPROACH

The first known study focusing on leader behavior examined the differing effects of democratic, autocratic, and laissez-faire leadership behaviors (Lewin et al. 1939; Lippitt and White 1943). These studies were conducted with groups of Boy Scouts led by trained graduate students. The results of these studies showed that groups with democratic leaders had the best-satisfied members. Those with autocratic leaders demonstrated the highest level of task activity, but only when the leader was present. Starting in the 1950s, the interest of the U.S. scholar gravitated toward understanding the leader's behaviors and the relationship between these behaviors and effectiveness. Three main research centers concurrently studied leader behavior in small teams at Ohio State University, led by Stogdill and his associates; at the University of Michigan, led by Likert and his colleagues; and at Harvard, led by Bales and his collaborators. Although these studies were conducted in different work settings such as the military, education, insurance companies, car manufacturing, and laboratory settings, they all found the same results. These results led to two categories of leader behavior. One category dealt with behaviors that establish and maintain relationships, commonly referred to as considerate, people-oriented, or socio-emotional behaviors. The second category focused on behaviors that get the task accomplished, commonly referred to as initiating-structure, production-oriented, or task-focused behaviors. Although many measures were designed to assess the leader's behaviors, the most prominent are three Ohio State measures: the Leader Behavior Description Questionnaire (LBDQ), the Subordinate Behavior Description, and the Leader Opinion Questionnaire (Cook et al. 1985).

The leadership behavior researchers in other countries identified two similar categories of structure and consideration, but also found some additional categories. Unlike the findings in the United States, the two behaviors in Iran were found to be intertwined, resulting in the concept of "benevolent paternalism," or the "father figure" leader (Ayman and Chemers 1983). In India, Sinha (1984) identified "nuturant-task" behavior as an addition to the two main behavioral categories. In Japan, Misumi (1985) introduced a measure that had behavioral categories similar to those in the U.S. findings, but the behaviors were assessed in context. He referred to this model as maintenance-production (MP) behaviors.

In the 1960s and 1970s, researchers investigated the relationship between these two categories of behaviors and various indexes of effectiveness (e.g., team satisfaction, performance, turnover, and grievance) in different work settings. Fisher and Edward's meta-analysis (1988) is among the studies that supported the overall effects of these two behaviors (cited in Bass 1990).

Since the middle of the 1980s, a paradigm shift in the study of leadership behavior emerged, which moved scholars from focusing on consideration and initiating structure behaviors, often referred to as "transactional leadership," to what is referred to as "transformational leadership" behavior. This movement started with the work of McGregor Burns (1978) and House (1977), and was further developed by two groups of scholars, Bass and Avolio (see Bass 1985); Avolio and Bass (1988) in the United States, and Conger and Kanungo (1987) in Canada. In this new approach, the leader provided guidance to ensure that the work was done properly and that subordinates were happy, and also was responsible for developing a relationship with employees based on mutual trust. It was proposed that transformational leadership not only changes the subordinate, but is also evolutionary for the leader and the task. A recent meta-analysis has provided a promising review of the effects of transformational leadership behavior in a variety of settings (Lowe et al. 1996). Also, it was through these approaches that the emphasis on empowerment in leadership evolved (Conger and Kanungo 1988b). Empowering, coaching, and facilitating are behaviors that were not noted extensively in leadership literature until the middle of the 1980s. During this period many inspirational and guiding documents were presented by various authors (e.g., Kouzes and Posner 1987).

Three main categories of criticism have been directed at the behavioral approaches. The first, advanced by Korman (1966) and by Kerr and colleagues (1974), argued that situational factors moderate the relationship between leader's behavior and outcomes. In addition, Fisher and Edwards (1988) more recently reviewed the studies using LBDQ and outcome variables, and acknowledged that there is a need for studying moderators.

The second criticism is based on the fact that almost all leader behavior measures are based on perception of self or the perception of others (Ayman 1993). In the 1980s, a series of studies demonstrated that perceived behavior is contaminated with various factors that are not necessarily related to the leader's behavior (Lord and Maher 1991). In these studies it was demonstrated that people's memory of the behaviors that occurred can be affected by what they think of the leader (e.g., Larson 1982; Philips and Lord 1982). The implication of this criticism is critical in exploration of diversity and cross-cultural leadership, topics that will be discussed later.

The third criticism focused on the level of conceptualization and analysis. This criticism is based on the premise that leaders may treat different people differently. That is to say, is leadership a group-level phenomenon, dyadic, or is it in the eyes of the beholder? Recently, a multilevel approach has been investigated in relation to various leader behavior paradigms (see Dansereau and Yammarino 1998).

SITUATION AND LEADERSHIP

In the 1950s, parallel to the development of leader behavioral studies, many scholars examined situational factors that gave rise to the identification of a leader without examining their characteristics. Some examined seating arrangement, or distance of the individuals (e.g., Bass et al. 1953; Howells and Becker 1962). The results of these studies demonstrated that individuals at the head of the table and who commanded more space were more often identified as the leader. Other studies examined communication patterns (Leavitt 1951) in the work team. The results demonstrated that those who have access to most of the information and are the center of the information flow are most often identified as leaders. The frequency of contribution to the team's goal achievement and verbal visibility was another contributor to the emergence of the leader (Bavelas et al. 1965). The position in the organization and assigned roles were also demonstrated to predict how a new leader behaved (Shartle 1951). That is, the best predictor of leadership behavior was the behavior of the leader's boss or predecessor. By the 1960s, there was some evidence that different tasks elicited different behaviors from the leaders (e.g., Morris and Hackman 1969).

In the 1980s, substitutes for leadership theory (Kerr and Jermier 1978; Howell et al. 1986) diverted the attention of leadership researchers again to the situation and provided a taxonomy for the role of the situation in leadership. This new approach categorized situations as neutralizers, enhancers, supplements, and substitutes (Schriesheim 1997). This approach has received mixed results, but the contribution of its taxonomy and methodological clarification has clarified the role of the situation in leadership research (Howell 1997).

In conclusion, the situational studies demonstrated that regardless of the leader's values, traits, or other characteristics, situational factors can influence the emergence of the leader and the behavior of the leader. Visibility and control of information are some of the necessary situational factors that assisted an individual being recognized as a leader. The more recent work on the situation has further extended the role of the situation in leadership effectiveness.

LEADERSHIP EMERGENCE

Early situational research and trait approaches focused on leadership emergence. Most studies on leader behavior attended to the effectiveness of leaders and teams. Hollander's concept of idiosyncrasy credit (1958) examined the exchange that occurs in the work team and leader emergence process. This concept gave rise to a series of studies (e.g., Hollander 1978; Hollander and Julian 1970) that demonstrated that, in a team, the members who contribute to the team's goal achievement and are loyal to the values held by the team establish credibility or idiosyncrasy credits. These studies demonstrated that idiosyncrasy credits then could be spent as the individual took risks to lead the group. Hollander's work may be seen as the first work that brought leadership and the management of innovation and change close together.

Hollander's work also demonstrated that the exchange process in teams resulted in the emergence of the member with the highest credits as the team leader. Later, the work of Graen (1976) further expanded the nature of this exchange in the vertical-dyad linkage (VDL) model. In the VDL model, the leader and the subordinates establish their relationship by negotiating their roles. Therefore, in work groups, there are some members who form the core group and are closer to the leader. Also, there may be other members who stay on the periphery of the work group.

The results of the initial studies were criticized, and unfortunately this line of research was abandoned due to methodological limitations (Dansereau et al. 1975). However, the notion of exchange between the leader and his or her subordinates is still a fertile area of research, known as leader-member exchange theory (e.g., Graen and Uhl-bien 1995).

CONTINGENCY APPROACHES

Korman (1966) and Stogdill (1948) criticized past research on leader behavior and traits, and called for the study of leadership to be situation-based. In the early 1970s, two main categories of contingency approaches emerged, one based on the leader's trait and the other based on the leader's behavior. The first scholar who operationalized the contingency approach to leadership and presented a trait-based model in this paradigm was Fred Fiedler. Several narrative reviews of the model have been written (see Ayman et al. 1998; Fiedler 1978). This model has been an impetus for many studies and some have criticized it over the years. However, several meta-analyses have provided some validating support on the thirty years of research (e.g., Strube and Garcia 1981; Peters et al. 1985).

The model is composed of three elements: one is the leader's motivational orientation, as measured by the "Least Preferred Coworker" scale; the second is the degree of leader's control and influence in a situation; and the third is the match between the situation and the leader's trait. In this model, three aspects of the situation were examined: leader-members relationship, task structure, and position power. The model's assumption is that leaders who are more focused on getting their job done than their relationship with people, do better than their counterparts in situations of high and low control and they are known to be in-match. Similarly, in situations of moderate control, those leaders with more focus on relationships will be more effective and are known as being in-match. When task oriented leaders are in moderate control condition and relationship oriented leaders are in high or low situational control conditions, they are referred to as being out-of-match (Ayman et al. 1995 and Ayman et al. 1998).

The strength of this model has been primarily in predicting team performance. More recently, subordinates' satisfaction and leader's experience of stress have also been included as outcomes examined by the model. Another strength of the model is that the information about the dependent variables and the independent variables are gathered from different sources. Leaders usually provide their motivational orientation by completing a quasi-projective test of Least Preferred Coworker. Sometimes the leader and other times the experimenter assessed the situation. A third party primarily has provided the assessments of the performance.

One area of potential improvement for the model is to include more than one trait. Recently, additional traits have been incorporated into the contingency model research. In 1993, Fiedler proposed the cognitive resource theory. This theory included the leader's intelligence and experience and proposed that their contribution to the team's success may have contradictory effects depending on the stressfulness of the situation. Also, Ayman and Chemers (1991) included the leader's self-monitoring capability in their study, which was conducted with Mexican managers. Results showed that subordinates' work satisfaction was higher for out-of-match leaders who are high self-monitors than for those who were lower self-monitors. This may indicate that self-monitoring trait compensate for leaders who are out of match based on the contingency model of leadership effectiveness. However, the subordinates of in-match, high, self-monitoring leaders were not as satisfied with their work as the low self-monitoring in-match leaders. Therefore, self-monitoring seemed to only work for those who find themselves in situations that are not conducive to their internal state. As this is only one study, future work on these or other traits is warranted.

The contingency models of leader behavior examined the moderators that influence the relationship between leaders' behaviors and the out-come variables. As there are multiple leader behaviors and multiple ways that the situation can be conceptualized, many models have emerged. Two models have received substantial empirical attention (i.e., path goal theory [House 1971]) and the normative model of decision making [Vroom and Yetton 1973; Vroom and Jago 1974]). However, other models have made conceptual contributions, such as the multiple-influence model of leadership (Hunt 1991) and the multiple linkage model (Yukle 1989). These models' contributions have been in the development of taxonomy for situational factors and managerial behavior (see Chemers 1998). The behavior domains studied in this category of contingency models are managerial, supervisory, and decision-making behavior. However, most managers and leaders are also involved with managing interpersonal conflict, and this behavior domain has not received much attention by leadership researchers. In this section we will primarily review path goal theory and the decision-making model, and will acknowledge the relevance of conflict management to leadership.

Path goal theory (House 1971) was based on Vroom's expectancy model of motivation. Its proposition was that the leader as a supervisor needed to help the subordinates by clearing their path to goal attainment. Therefore, the model predicted that the leader's behavior would be dependent on the situation, which was determined by the characteristics of the subordinates and the nature of their work. Therefore it predicted that subordinates with high ability were more satisfied and less stressed when their leaders were perceived as behaving considerately. Subordinates with complex tasks were more satisfied with leaders who were perceived as more directive. The model acknowledged four behavioral choices for the leaders: participative, supportive, directive, and achievement-focused. However, the majority of the research used Ohio State's measure of LBDQ-XII to define the leader's behavior as either considerate or structuring. In most of the studies, the descriptions of the situation, the subordinates' behavior, and the leader's behavior were assessed according to the subordinates' perceptions. Research on path goal theory received mixed results. In the majority of the studies either the subordinates' characteristics or their tasks were investigated. Also, the studies were primarily supportive, predicting subordinates' satisfaction but not their performance.

The decision-making model (Vroom and Yetton 1973; Vroom and Jago 1974) was mostly assessed from the leader's perspective. It provided the leader with choices of team decision-making strategies (e.g., very participative, partially participative, and very autocratic). The success of these choices depended on the situations. The model provided strategic questions that would help a leader determine the extent to which he or she should be participative or autocratic based on eight different criteria. These questions assessed the leader's knowledge and the subordinates' supportiveness in a given situation. The key finding was that subordinates who were part of the decision-making process were more satisfied, compared to those who were not part of this process.

Studies that have examined conflict management strategies of leaders are limited. The most commonly used measures are those by Rahim (1985) and by K. W. Thomas (1992). The two approaches identify five conflict management styles (competitive, withdrawing, accommodating, compromising, and collaborating) that represent the degree to which leaders are focused on themselves or the task, or the extent to which they are concerned about the subordinates' welfare.

In the same vein, the subordinates' choice of action with the leader was also examined. The situational factors studied so far have been the status of the actor in relationship to the target and the presence of an interpersonal conflict. The evidence seems to show that leaders have been perceived as using competitive and withdrawal strategies and subordinates tend to use more accommodating and compromising strategies. This line of research is new and requires more development. There is little evidence connecting choices of conflict management strategies to outcomes such as subordinate satisfaction or leader's influence or team performance.

Overall, the strength of behavioral approaches lies in their focus on the leader's behavior, which is assumed to present an aspect of the leadership process. These approaches are also sensitive to the moderating factor of situational elements; that is, they assume that not all successful or effective leaders behave the same way across situations. The behavioral approach provided some clarity in leadership process that allowed trainers to design training and development programs guiding leaders in how to behave.

These models have also faced challenges. First, the assessment of leader behavior and the situation in these models were dependent on the perception of the leader or the subordinates, thus susceptible to social cognitive influences, which will be briefly discussed in the next section. Second, most of these studies were based on a single source of information, that is, the same person who provided the predictors (e.g., the leader's behavior descriptions and the situation) also provided the assessment of the outcome and the moderator (i.e., situational factors). Finally, these models initially examined leadership only at a team level; thus assumed leaders treated everyone similarly. However, the criticism is that leader behavior may vary with each individual.

NEW TOPICS AND FUTURE DIRECTIONS

After a century of empirical research, there are theoretical and methodological issues that have drawn the attention of the leadership researchers, theorists, and practitioners alike. Theoretically, three issues that need to be explored are: (1) the role of social cognition in leadership, (2) the role of diversity and multiculturalism in leadership, and (3) moving toward an integrative theory. From a methodological perspective, four issues need to be considered: single sources variance, level of conceptualization and analyses, a systematic conceptualization and measurement of the situation, and measurement of leadership effectiveness.

Many scholars have recognized the role of social cognition in leadership studies (e.g., Calder 1977; Martinko and Gardner 1987). Some have argued that leadership is a product of our imaginations (Miendl 1990), whereas others have respected the role of perception in the study of leadership (Ayman 1993; Lord and Maher 1991). From these discussions and the studies that have transpired, it has been substantiated that people have an image of a leader in "their mind's eye." This schema or the mental map does affect their perception and evaluation of leaders they meet (Larson 1982; Lord et al. 1978).

Research on diversity and multiculturalism in leadership has gained momentum in the later part of the century. Initially many studies in the United States focused on gender. This may have been partially due to sample availability. Studies have demonstrated that overall, in the United States and other cultures, across age and gender, the image of a leader is more similar to the image of a man than a woman (e.g., Ayman-Nolley et al. 1993; Schein et al. 1989, 1996). Although many meta-analyses have reflected that, across studies, men and women leaders' behavior differences are not very noticeable (Eagly and Johnson 1990), and that the differences that were found were mostly on task behaviors. However, the authors identified various moderators in determining a leader's gender and behavior relationship, such as self/subordinate rating, male-dominated setting, and role congenial position (Eagly and Johnson 1990). In measuring leader effectiveness across indexes such as overall satisfaction and performance, no gender difference was found. However, women leaders were favored more than men were when subordinates' satisfaction was measured, and men did better than women when performance was measured. In addition, the moderator of the gender role congeniality of the job was identified as a major predictor of this relationship. Therefore, it is possible that the dominant masculine image of a leader (i.e., more prevalent in traditional male-dominated roles) has been the major challenge of women in breaking the glass ceiling (Morrison and Van Glinow 1990).

As leadership process is dependent on perception and social judgement, the saliency of the stimuli's characteristics is important in this process. Although most of the diversity studies on leadership have been on gender differences, other diversity indexes also need to be examined. In studying leaders who are diverse, it is critical that the superficial diversity elements (i.e., those that can be seen, such as sex and color) and the deep diversity elements (i.e., those that cannot be seen, such as traits and education) be considered separately (Harrison et al. 1998). For example, Korabik (1997) has presented evidence that women may perceive themselves differently based on their sex-role orientation. But observers perceive women mostly based on the feminine stereotypes.

In addition, the process of social interaction between leaders and subordinates involves the leader as the actor and others as observers who judge this leader. In this case, both the leader's characteristics and the characteristics of subordinates need to be considered. In gender studies already reviewed, the gender composition of the group has been shown to affect the selection of a man or a woman as the leader (Eagly and Karau 1991) and to impact on men and women leaders' evaluation (Eagly et al. 1995). Research has demonstrated that in mixed-gender teams, men are more often favored than women. So, in studying leaders of other cultures or ethnic groups, it is not sufficient to study their behavior or expectations with their own culture; rather, it is critical to examine their interaction with people from other cultures (Ayman 1997). For example, Thomas and Ravlin (1995) found that American autoworkers rated more negatively on the dimension of trust when Japanese managers behaved like American leaders, but perceived these Japanese managers as more similar and effective.

Recently, several researchers and scholars have conducted cross-paradigm studies and provided integrated models of leadership. In the last five years there have been studies that examined the relationship between traits and leader behavior, such as Myers-Briggs and transformational behavior (Roush and Atwater 1992) and Bem's sex-role inventory and transformational leadership (Hackman et al. 1992). There are also studies connecting leader member exchange with transformational leadership (Basu and Green 1997). More efforts to connecting various leadership research approaches are warranted. Chemers (1997) has presented an integrative model of leadership enabling researchers to approach leadership research with a more comprehensive understanding in the future. That is, although each study may be limited in scope, using a model like that of Chemers can help place it in the universe of leadership research. Doing cross-theory studies of leadership will enhance the understanding of this complex phenomenon.

The methodological challenge facing leadership research may be categorized in three topics (i.e., single-source variance, level issues, and measurement of the situation and the effectiveness of the leader). Many models reviewed in this article were validated on a single source of information. In other words, the same person who provided the predictor (e.g., the leader's behavior description) also provided the criterion (e.g., satisfaction or performance). This can artificially inflate the relationships under investigation (Yukle and Van Fleet 1992). One solution that some have used is that one subordinate describes the leader's behavior and another provides the effectiveness measure (Avolio et al. 1991). Although this is a step in the right direction, still both the leader and the subordinates have a vested interest in the relationship and thus may provide a biased perception.

Level of conceptualization and analysis is another methodological challenge that recently became the focus of leadership researchers' attention (Dansereau et al. 1995). These authors propose that the level of variables used in leadership studies needs to consider whether it is assessing the phenomenon at an individual, a dyadic, a group, or an organizational level. Also, it is important to consider the level issues in design and analysis of a study. In some studies the measures and the analyses may stay in the same level or go across levels to examine a meso-level relationship. The gravity of this issue and the methodologies involved in this procedure are reflected in a two-volume book edited by Dansereau and Yammarino (1998). In these volumes, many leadership approaches and theories are examined with the multilevel question in mind. It is not necessary for all models to be validated at all levels. However, it is valuable to understand the parameters of the various models in relation to the level issue.

The third challenge is the assessment of leadership situation and leadership effectiveness. Although in situational studies and in contingency models some element of the situation is operationalized, a more systematic, cohesive model for evaluating these variables across theories would assist the field's evolution. Fiedler and Chemers (Ayman and Romano 1998) have identified three elements in the situation that affect the leaders' perceived control (leader-member relation, task structure, and position power). These aspects of situational control parallel French and Raven's sources of power (Ayman et al. 1998). The relative degree of importance of these aspects of situation has been substantiated both in contingency model research (Fiedler 1978) and in sources-of-power research (Podsakoff and Schriesheim 1985). Also, Hunt (1991) identified different levels within the organization as another situational factor affecting leadership. What is important to remember is that the prominent models of leadership reviewed here were validated in various work settings and industries. These models have been able to explain leadership regardless of whether the person is a political leader, a hospital administrator, a factory manager, a retail manager, a school principal, or a military leader. Consequently, future work is needed on a framework for identifying elements and dimensions that are important to consider in leadership situations. For example, one dimension that has been identified is the leader's power base. What are others?

Leadership effectiveness (e.g., group performance, leader's performance, group cohesion, sub-ordinate satisfaction, organizational commitment, employee turnover rate) has been measured by various indexes and by various sources such as self report, superiors, and subordinates. Some models were more effective in predicting leaders emergence than effectiveness. Others have been more successful in predicting the well-being and cohesion of the group, such as subordinate satisfaction, turnover in the group, and subordinates stress. Several have predicted group performance measures by subordinates' evaluation or a superiors' assessment, or have used performance criteria such as financial indexes, sales, or goals met (e.g., sport teams' scores). It is important to consider the parameter of models and the measures used to assess leadership. Some models may be better in predicting one criterion than another.

In conclusion, some have said, "Leadership is paradoxical; popular thinking . . . emphasizes the importance of . . . leadership establishing excellent organizations, but many academic publications assert that leadership is consequential" (Day and Lord, in Hunt 1991, p. 1). However, as Fiedler said in a personal comment to the author: "This is a pretzel shape universe that requires a pretzel shape model." The complexity of the world and the challenges it presents may call for a different form of leadership from the one-person-focused leadership. Since the team approach is being used in many organizations, it may be that the future of leadership will also be a team leadership approach. Compared to many fields of science, study of leadership is very young. Although challenges are present and the phenomenon is nebulous, leaders would attest that there are some things going on that make things work, but they are hard to explain. Future researchers are charged to systematically overcome the challenges and to collectively gain control of understanding of leadership in a global dynamic environment.

references

Avolio, B., and B. M. Bass 1988 "Transformational Leadership, Charisma and Beyond." In J. G. Hunt, B. R. Balaga, H. P. Dachler, and C. Schriesheim, eds., Emerging Leadership Vistas. Elmsford, N.Y.: Pergamon.

Avolio, B. J., F. J. Yammarion, and B. M. Bass 1991 "Identifying Common Methods Variance with Data Collected from a Single Source: An Unresolved Sticky Issue." Journal of Management 17:571–587.

Ayman, R. 1993 "Leadership Perception: The Role of Gender and Culture." In M. M. Chemers and R. Ayman, eds., Leadership Theory and Research: Perspectives and Directions. New York: Academic.

——1997 "Organizational Leadership." In S. W. Sadava and D. R. McCreary, eds., Applied Social Psychology. Upper Saddle River, N.J.: Prentice Hall.

——, and M. M. Chemers 1983 "Relationship of Supervisory Behavior Ratings to Work Group Effectiveness and Subordinate Satisfaction." Journal of Applied Psychology 68:338–341.

——1991 "The Effect of Leadership Match on Subordinate Satisfaction in Mexican Organizations: Some Moderating Influence of Self-Monitoring." Applied Psychology: An International Review 40:299–314.

——, and F. Fiedler 1998 "The Contingency Model of Leadership Effectiveness: Its Levels of Analysis." In F. Dansereau and F. J. Yammarino, eds., Leadership: The Multiple-Level Approaches—Classical and New Wave. Stamford, Conn.: JAI.

——, and R. Romano 1998 "Appendix: Measures and Assessments for the Contingency Model of Leadership Effectiveness." In F. Dansereau and F. J. Yammarino, eds., Leadership: The Multiple-Level Approaches—Classical and New Wave. Stamford, Conn.: JAI.

Ayman-Nolley, S., R. Ayman, and J. Becker 1993 "Gender Affects Children's Drawings of a Leader." Paper Presented At The Annual Convention Of American Psychological Society, Chicago (June).

Bass, B. M. 1985 Leadership and Performance beyond Expectations. New York: Free Press.

——1990 Bass and Stogdill's Handbook of Leadership: Theory, Research, and Managerial Applications, 3rd ed. New York: Free Press.

——, S. Kulbeck, and C. R. Wurster 1953 "Factors Influencing the Reliability and Validity of Leaderless Group Discussion Assessment." Journal of Applied Psychology 37:26–30.

Basu, R., and S. G. Green 1997 "Leader-Member Exchange and Transformational Leadership: An Empirical Examination of Innovative Behaviors in Leader-Member Dyads." Journal of Applied Social Psychology 27:477–499.

Bavelas, A., A. H. Hastorf, A. E. Gross, and W. R. Kite 1965 "Experiments on the Alteration of Group Structure." Journal of Experimental Social Psychology 1:310–320.

Burns, J. M. 1978 Leadership. New York: Harper and Row.

Calder, B. J. 1977 "Attribution Theory of Leadership." In B. M. Staw and G. R. Salanick, eds., New Directions in Organizational Behavior. Chicago: St. Clair.

Carlyle, T. (1841) 1907 Heroes and Hero Worship. Boston: Adams.

Chemers, M. M. 1997 An Integrative Theory of Leadership. Mahwah, N.J.: Lawrence Erlbaum.

——, and R. Ayman 1993 Leadership Theory and Research: Perspectives and Directions. New York: Academic.

Conger, J. A., and R. A. Kanungo 1987 "Towards a Behavioral Theory of Charismatic Leadership in Organizational Settings." Academy of Management Review 12:637–647.

——1988 "The Empowerment Process: Integrating Theory and Practice." Academy of Management Review 13:471–482.

Cook, J. D., S. J. Hepworth, T. D. Wall, and P. B. Warr 1985 The Experience of Work: A Compendium and Review of 249 Measures and Their Use. New York: Academic.

Dansereau, F., G. Graen, and W. J. Haga 1975 "A Vertical Dyad Linkage Approach to Leadership in Formal Organizations." Organizational Behavior and Human Performance 13:46–78.

——, and F. J. Yammarino (eds.) 1998 "Leadership: The Multiple-Level Approaches," vols. I and II. Stamford, Conn.: JAI.

Dansereau, F., F. J. Yammarino, and S. E. Markham 1995 "Leadership: The Multiple-Level Approaches." Leadership Quarterly 6:97–110.

Eagly, A. H., and B. T. Johnson 1990 "Gender and Leadership Style: A Meta-Analysis." Psychological Bulletin 108:233–256.

——, and S. J. Karau 1991 "Gender and the Emergence of Leaders: A Meta-Analysis." Journal of Personality and Social Psychology 60:685–710.

——, and M. G. Makhijani 1995 "Gender and the Effectiveness of Leaders: A Meta-Analysis." Psychological Bulletin 117(1):125–145.

Fiedler, F. E. 1978 "The Contingency Model and the Dynamics of the Leadership Process." In L. Berkowitz, ed., Advances in Experimental Social Psychology, vol. 11. New York: Academic.

——1993 "The Leadership Situation and the Black Box in Contingency Theories." In M. M. Chemers and R. Ayman, eds., Leadership Theory and Research: Perspectives and Directions. New York: Academic.

——1995 "Cognitive Resource and Leadership." Applied Psychology: An International Review 44:5–28.

——, and M. M. Chemers 1984 Improving Leadership Effectiveness: The Leader Match Concept, 2nd ed. New York: John Wiley.

Fisher, B. M., and J. E. Edwards 1988 "Consideration and Initiating Structure and Their Relationships with Leader Effectiveness: A Meta Analysis." In Best Papers Proceedings, Academy of Management. Anaheim, Calif.: Academy of Management.

Graen, G. 1976 "Role Making Processes within Complex Organizations." In M. D. Dunnette, ed., Hand-book of Industrial and Organizational Psychology. Chicago: Rand McNally.

Graen, G. B., and Uhl-Bien 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." The Leadership Quarterly 6:219–247.

Hackman, M. Z., A. H. Furniss, M. J. Hills, and T. J. Paterson 1992 "Perceptions of Gender-Role Characteristics and Transformational and Transactional Leadership Behaviors." Perceptual and Motor Skills 75:311–319.

Harrison, D. A., K. H. Price, and M. P. Bell 1998 "Beyond Relational Demography: Time and the Effects of Surface-and Deep-Level Diversity on Work Group Cohesion." Academy of Management Journal 41:96–107.

Hogan, R., G. J. Curphy, and J. Hogan 1994 "What We Know about Leadership: Effectiveness and Personality." American Psychologist 49(6):493–504.

Hollander, E. P. 1958 "Conformity, Status, and Idiosyn-crasy Credit." Psychological Review 65:117–127.

——1978 Leadership Dynamics: A Practical Guide to Effective Relationships. New York: Free Press.

——, and J. W. Julian 1970 "Studies in Leader Legitimacy, Influence, and Innovation." In L. Berkowitx, ed., Advances in Experimental Social Psychology, vol. 5. New York: Academic.

House, R. J. 1971 "A Path Goal Theory of Leader Effectiveness." Administrative Science Quarterly 16:321–338.

——1977 "A 1976 Theory of Charismatic Leadership." In J. G. Hunt and L. L. Larson, eds., Leadership: The Cutting Edge. Carbondale, Ill.: Southern Illinois University Press.

Howell, J. P. 1997 "Substitutes for Leadership: "Their Meaning And Measurement"—An Historical Assessment." Leadership Quarterly 8:113–116.

——, P. W. Dorfman, and S. Kerr 1986 "Moderator Variables in Leadership Research." Academy of Management Review 11:88–102.

Howells, L. T., and S. W. Becker 1962 "Seating Arrangement and Leadership Emergence." Journal of Abnormal and Social Psychology 64:148–150.

Hunt, J. G. 1991 Leadership: A New Synthesis. Newbury Park, Calif.: Sage.

Kerr, S., and J. M. Jermier 1978 "Substitutes for Leadership: Their Meaning and Measurement." Organizational Behavior and Human Performance 22:374–403.

Kerr, S. C., C. J. Schriesheim, C. J. Murphy, and R. M. Stogdill 1974 "Towards a Contingency Theory of Leadership Based upon the Consideration and Initiating Structure Literature." Organizational Behavior and Human Performance 12:62–82.

Korabik, K. 1997 "Applied Gender Issues." In S. W. Sadava and D. R. McCreary, eds., Applied Social Psychology. Upper Saddle River, N.J.: Prentice Hall.

Korman, A. K. 1966 "Consideration, Initiating Structure, and Organizational Criteria—A Review." Personnel Psychology 19:349–361.

Kotter, J. P. 1988 The Leadership Factor. New York: Free Press.

Kouzes, J. M., and B. Z. Posner 1987 The Leadership Challenge: How To Get Extraordinary Things Done in Organizations. San Francisco: Jossey-Bass.

Larson, J. R., Jr. 1982 "Cognitive Mechanisms Mediating the Impact of Implicit Theories of Leader Behavior on Leader Behavior Rating." Organizational Behavior and Human Performance 29:129–140.

Leavitt, H. J. 1951 "Some Effects of Certain Communication Patterns on Group Performance." Journal of Abnormal and Social Psychology 46:38–50.

Lewin, K., R. Lippitt, and R. K. White 1939 "Patterns of Aggressive Behavior in Experimentally Created Social Climates." Journal of Social Psychology 10:271–301.

Lippitt, R., and R. K. White 1943 "The Social Climate of Children's Groups." In R. G. Baker, J. S. Kounin, and H. F. Wright, eds., Child Behavior and Development. San Francisco: Freeman.

Lord, R. G., J. F. Binning, M. C. Rush, and J. C. Thomas 1978 "The Effect of Performance Cues and Leader Behavior on Questionnaire Rating of Leadership Behavior." Organizational Behavior and Human Performance 21:27–39.

Lord, R. G., C. L. Devader, and G. M. Alliger 1986 "A Meta-Analysis of the Relation between Personality Traits and Leadership Perceptions: An Application of Validity Generalization Procedures." Journal of Applied Psychology 71:402–410.

Lord, R. G., and K. J. Maher 1991 Leadership and Information Processing: Linking Perceptions and Performance. Boston: Harper Collins.

Lowe, K. B., K. G. Kroeck, and C. A Campo 1996 "Effectiveness Correlates of Transformational and Transactional Leadership: A Meta-Analytic Review of the MLQ Literature." The Leadership Quarterly 7:385–426.

Mann, R. D. 1959 "A Review of the Relationship between Personality and Performance in Small Groups." Psychological Bulletin 56:241–270.

Martinko, M. J., and W. L. Gardner 1987 "The Leader/Member Attribution Process." Academy of Management Review 12:235–249.

Marx, K. 1906 Capital. Chicago: Charles H. Kerr.

Miendl, J. R. 1990 "On Leadership: An Alternative to the Conventional Wisdom." In B. A. Staw, ed., Research in Organizational Behavior, vol. 12. New York: JAI.

Misumi, J. 1985 The Behavior Science of Leadership: An Interdisciplinary Japanese Research Program. Ann Arbor: University Of Michigan Press.

Morris, C. G., and J. R. Hackman 1969 "Behavorial Correlates of Perceived Leadership." Journal of Personality and Social Psychology 13:350–361.

Morrison, A. M., and M. A. Von Glinow 1990 "Women and Minority in Management." American Psychologist 45:200–208.

Peters, L. H., D. D. Hartke, and J. F. Pohlmann 1985 "Fielder's Contingency Theory of Leadership: An Application of the Meta-Analysis Procedures of Schmitt and Hunter." Psychological Bulletin 97:274–285.

Philips, J. S., and R. G. Lord 1982 "Schematic Information Processing and Perception of Leadership in Problem-Solving Groups." Journal of Applied Psychology 67:486–492.

Podsakoff, P. M., and C. A. Schriesheim 1985 "Field Studies of French and Raven's Bases of Power. Critique, Reanalysis, and Suggestions for Future Research." Psychological Bulletin 97:387–411.

Rahim, M. A. 1985 "Referent Role and Styles of Handling Interpersonal Conflict." The Journal of Social Psychology 126:79–86.

Roush, P. E., and L. Atwater 1992 "Using the MBTI to Understand Transformational Leadership and Self-Perception Accuracy." Military Psychology 4:17–34.

Schein, V. E., R. Mueller, and C. Jacobson 1989 "The Relationship Between Sex-Role Stereotypes and Requisite Management Characteristics among College Students." Sex Role 20:103–110.

Schein, V. E., R. Mueller, T. Lituchy, and J. Liu 1996 "Think Manager—Think Male: A Global Phenomenon?" Journal of Organizational Behavior 17:33–41.

Schriesheim, C. A. 1997 "Substitute for Leadership Theory: Development and Basic Concepts." Leadership Quarterly 8:103–108.

Shartle, C. L. 1951 "Studies in Naval Leadership." In H. Guetzkow, ed., Groups, Leadership, and Men. Pittsburgh: Carnegie.

Sinha, J. B. P. 1984 "A Model of Effective Leadership Styles in India." International Studies of Management and Organization 14:86–98.

Smoler, F. 1999 "The History of the CEO." Worth 5:150–154.

Snyder, M. 1979 "Self-Monitoring Process." In L. Berkowitz, ed., Advances in Experimental Social Psychology, vol. 12. New York: Academic.

Stogdill, R. M. 1948 "Personal Factors Associated with Leadership: A Survey of the Literature." Journal of Psychology 25:35–71. Strube, M. J., and J. E. Garcia 1981 "A Meta-Analysis Investigation of Fiedler's Contingency Model of Leadership Effectiveness." Psychological Bulletin 90:307–321.

Thomas, D. C., and E. C. Ravlin 1995 "Responses of Employees to Cultural Adaptation by a Foreign Manager." Journal of Applied Psychology 80:133–147.

Thomas, K. W. 1992 "Conflict and Negotiation Processes in Organizations." In M. D. Dunnette and L. M. Hough, eds., Handbook of Industrial and Organizational Psychology, 2nd ed., vol. 3. Palo Alto. Calif.: Consulting Psychologists.

Vroom V. H., and A. G. Jago 1974 "Decision-Making as a Social Process: Normative and Descriptive Models of Leader Behavior." Decision Sciences 5:743–769.

Vroom V.H., and P. W. Yetton 1973 Leadership and Decision-Making. Pittsburgh: University Of Pittsburgh Press.

Yukle, G. 1989 Leadership in Organizations, 2nd ed. Englewood Cliffs, N.J.: Prentice Hall.

——, and D. D. Van Fleet 1992 "Theory and Research on Leadership in Organizations." In M. D. Dunnette and L. M. Hough, eds., Handbook of Industrial and Organizational Psychology, 2nd ed., vol. 3. Palo Alto. Calif.: Consulting Psychologists.

Zaccaro, S. J., R. J. Foti, and D. A. Kenny 1991 "Self-Monitoring and Trait-Based Variance in Leadership: An Investigation of Leader Flexibility across Multiple Group Situations." Journal of Applied Psychology 76:179–185.

Roya Ayman, Ph.D.