Robert Crandall, as chairman and president of American Airlines, once opined, “I think the ideal leader in the 21st Century will … create an environment that encourages everyone in the organization to stretch their capabilities and achieve a shared vision … give people the confidence to run farther and faster … and establish the conditions for people to be more productive, more innovative, and feel more in charge of their own lives” (McFarland, Senn, & Childress, 1993). The twenty-first century might well be called the Age of Peter Drucker. Because of him, organizational leaders scan their environments; revisit their missions; decide who are the customers and clients they serve; ban hierarchy; arrange flexible, fluid management systems; replace up-down vocabulary with team vocabulary; reject the status quo and policies that work for today and not for tomorrow; and communicate with clear, consistent mission-focused messages. Leadership is dispersed and highly involved, and responsibilities are shared. The moral compass of leaders works full-time by voice and example. Results are measured by objectives reached. The journey is celebrated (Hesselbein, November 4, 2000).
At the beginning of the second millennium, after reviewing what has happened in leadership research over the preceding 33 years, I looked forward 33 years and forecast that: in the year 2034, leaders will pay more attention to each follower than to the group they lead. Individualized consideration will have become routine. Leaders will be developed primarily online. Full second careers in management lasting 20 years will be commonplace. Women will dominate as MBA students and as educational, government, nonprofit, and service organization leaders. Empirically based ethical codes will have been established for management. Artificial intelligence will be a basis of management decision making. Biotechnology, such as genetic profiling and brain scanning, will be employed in assessing managers and leaders. There will be more concentration on leaders’ strengths than weaknesses. Mathematical operations will be used to optimize outcomes in negotiations. Leadership fads will be as strong as ever. Virtual teams and e-leadership will be the rule rather than the exception in large organizations (Bass, 2002).
The ideal twenty-first-century organizational leader will think locally and act globally, replace a bureaucratic hierarchy with a community of responsibility, communicate a strong sense of mission, and value the distribution of power, diversity, and inclusiveness (Leadership IS, 1994). Leadership research and its applications will continue to expand. Their effects will be seen in the ever-increasing attention to the “people” side of enterprise—in corporate mission statements, military doctrine, educational curricula, and nationwide community programs.
Caveat. It is impressive how wrong we can be about the future. In 1903, just before the Wright Brothers’ flight at Kitty Hawk, physicists were writing that heavier-than-air aircraft would never fly. Walter Lippmann commented in his column of April 27, 1948, in The New York Times that the Arab-Israeli conflict was “one of the simplest and most manageable.”
We can only speculate about how much the results of studies of leadership, popularized in college courses, best-selling books, and the mass media, contribute to the implicit theories of leadership espoused by college grad-uates—our future leaders. For the students believe that the productivity and morale of workers depend on leaders who express concern about the work to be done, are participatory, and offer praise and rewards for good performance (Graves, 1983). Would these same three clusters of leadership behavior have emerged among college students in 1950? In 1950, whatever was produced profitably could be sold. The largest companies were manufacturers, such as General Motors. In 2007, the profitability of the largest firms—retailers like Wal-Mart—depended on satisfied customers and the ability to negotiate lower prices with suppliers. Would students today have the same prosocial attitudes and the same cynicism were it not for continuing scandals in business and government? The predictions about leadership and management in the year 2000 made in the mid-1960s and a few years later erred mainly in that the predicted developments came sooner than expected (Bass, 1967c; Bass & Ryterband, 1974). The predictions made today are likely to be equally conservative. Evaluation of past national growth rates by Franke, Hofstede, and Bond (1991) showed that individualistic, entrepreneurial, and innovative societies such as the United States tend to be disadvantaged in growth rate, compared with more collectivistic and bureaucratic societies such as Japan. However, since the 1990s it has been the latter type of society that has been disadvantaged compared to the former (Franke & Barrett, 2004).
Paradigm shifts that took place between 1975 and 2005 are likely to settle into new concentrations of research effort. For instance, there is likely to be more research on transformational factors such as charismatic and inspirational leadership and on the cognitive processes involved in leader-subordinate relationships. In addition, there is likely to be much more sorting out of the personal and situational processes that influence leaders and followers, leading to the possibility of finding generalizations that apply to any leader and to any follower. Controlled survey and laboratory studies will increasingly extend to a wider array of methods. In the same way, issues will be broadened by the cross-fertilization of interests of the behavioral, social, and political sciences. Contributions to the field of both substance and methods from sources outside North America will increase. Concerns about the equitable distribution of power will be increasingly shared with concerns about the equitable distribution of information.
There is also likely to be increasing consolidation, as well as reanalysis, of earlier findings about leadership and the outcomes of leadership that questioned whether various contingencies were merely transient phenomena. The cultural, social, and economic changes that took place in the last half of the past century may require, as Tucker (1983) suggested, a reexamination of the instruments, structures, and relationships established earlier. Leadership researchers will need to consider whether their purpose is to make a contribution to theory or to application. If their intended audience is academia and their purpose is to make a rigorous addition to understanding, they may find there is less application of their work to human resources management, in contrast to possibly more messy broader empirical results (Weibler & Wald, 2004). Researchers will be paying more attention to the substance than the shadows (Bass, 1974). Anything that can be meta-analyzed will be meta-analyzed. Van de Vall and Bolas (1997) looked at the extent to which 120 Dutch policy research projects had an impact on organizational decisions. According to interviews, the impact of the findings had more effect on decisions if they were empirically grounded rather than based on theoretical constructs.
As new nations become complex industrialized societies, much will be learned about the leadership and management that emerge. These factors will need to fit with the cultural realities of the different nations. For instance, the new leadership styles emerging in China as a leading global market economy will be affected by the collectivism inherent in China’s rich 3,000-year-old philosophical heritage (Liancang, 1987).
Six considerations guide our peek into the future of research on leadership and its applications: (1) extrapolation from the past, (2) societal changes, (3) new technologies, (4) organizational trends, (5) changes in organizations, (6) changes in personnel practices, (7) new paradigms.
First, we look ahead by extrapolation from the past. For example, much of the course in the 2000s and beyond has been set by the trend that began in the 1950s and 1960s toward more longitudinal research rather than onetime, cross-sectional studies. Already confirmed in the 1950s and 1960s were the importance of the computer; the needs for challenging work and flexible organizations; faster communications; and more teamwork, diversity, temporary organizations, and simulation for decision making and training. Mass production and large inventories were beginning to be replaced by deliveries of just-in-time supplies and production for specific customer orders. Drucker (1994) noted that in the developed countries during the twentieth century, both qualitative and quantitative transformations occurred in the nature of work and in social, economic, and political systems. These transformations will continue through the twenty-first century. By 2010, Drucker (1998) foresaw, there will be no single dominant world economic power like the United States. To remain competitive, company leaders will need to have as good information about events and conditions outside their firms as internal reports and surveys provide about what is happening inside the firms. Retirement age will go up in developed countries, but the workforce will decrease in size as the productivity of knowledge and knowledge work continues to increase and provides the decisive competitive advantage.
The substance of research on leadership is influenced considerably by what is happening in society as a whole. Leaders will have to adapt to the changing roles and relationships in society. Will government continue to enlarge despite the efforts to reduce it? (Beckhard, 1995). The majority of the American workforce has shifted from manufacturing to service and information work. Will manufacturing, services and information work continue to move offshore? More work is being done in smaller companies (Cascio, 1995). Organizational life in the public sector and health care have become of increasing interest. Opportunities to increase productivity, particularly its quality, have become more dependent on effective human relationships and the development of personnel. In the same way, when humans are replaced with robots and computers, more high-level personnel specialists, supervisors, and team development are likely to be required. Mandated in much of western Europe, industrial democracy remains voluntary in the United States, where it is increasingly becoming a fact of life through enlightened managements and employee ownership plans. At the same time, the percentage of the workforce in U.S. labor unions continues to decline, particularly in the private sector. Labor unions remain stronger in Europe, Australia, and the U.S. public sector. Nevertheless, despite unionization, affective organizational commitment to their firms by 635 Australian employees of Generation X, born between 1960 and 1975, when controlled for age, appears no different from that of 382 Australian baby boomers born between 1941 and 1959 (Hart, Schembri, Bell, et al., 2003). Low wages (paid in the United States to 25% of workers in 2004) remain acceptable to families because both spouses tend to be working. But the many scandalous threats to the United States’ civic infrastructure have resulted in a sharp reduction of public trust in government and business institutions (Gill, 1996).
Demographics are inexorably working their will. The current generation of U.S. employees is older and more diverse in sex, color, race, and foreign birth than the preceding one. Unlike in Europe, Japan, China, and parts of the Middle East, the United States’ birth rate is not declining, and its population is continuing to increase very quickly due to continuing immigration. The trend continues toward more automation and self-service, and many immigrants are needed to fill lower-paying jobs in construction, manufacturing, agriculture, and service. Russian and other Eastern European professionals and skilled employees have migrated west in significant numbers. Illegal immigration of cheap labor from Mexico and elsewhere has reached crisis proportions. Flows of illegal immigrants from undeveloped Africa to developed Europe are occurring in the same way. Lower-paying jobs and the income gap between the rich and the poor continue to increase in the United States. The percentage of more relations-oriented personnel in the U.S. labor force has increased with the sharp rise in the employment of women. A plural society of varied ethnicity and race is replacing the ideal of a single melting pot.
The multinational firm, with its worldwide outlook, continues to expand. Multinational firms are just as likely to be headquartered in Japan, Britain, Germany, or elsewhere as in the United States. Dual careers and wage earners in the family are now the rule. Five of every six men over age 65 are moving into a new long-lived class of retirees. By 2010, Drucker (1998) expects, healthy people will stop working at 75. Foreign investment in the United States has risen sharply, particularly since 1985. A considerable percentage of U.S. personnel are now working for foreign-owned firms. There is an increasing need to remain competitive in the world marketplace with the Pacific Rim nations and with the rapidly growing European Union. The developed countries can remain competitive for a few more decades in the productivity of knowledge work (Drucker, 1998), although India and China are catching up fast.
Forms of government and economic systems from Chile to the Russian Federation changed in the 25 years from 1980 to 2005. Japan’s industry is maturing; Britain’s class structure is shifting. Ireland has gone from being one of the poorest to one of the wealthiest countries in Europe. The U.S. hegemony in 1945 in economic power is being challenged in 2008 by the European Union and East Asia. The expansion in international trade has been accompanied by rapid changes in exports and imports. In 2003, developing China exported $30 billion in manufactured and high-tech goods to the developed United States; the developed United States exported $3 billion in agricultural products to developing China. These economic and political changes are bringing new, continuing challenges for leaders at many levels in both countries in government, industry, and education. Continued social change is foreseen in the coming years.
The world climate continues to warm. The failure to control agricultural and industrial pollution continues, along with threats of pandemics of AIDS and avian influenza, a new strain of influenza resistant to medication available in 2008. Other threats include worldwide terrorism and the exploding populations in developing countries. Concern about corruption, business ethics, and the social responsibility of large corporations continues.
Leadership research and development reflect these societal developments. Mentoring will play an increasingly important role (Zey, 1988). Furthermore, a premium will be placed not only on junior managers’ abilities to deal with the human factor but on attention to quality and costs. More statesmanship and an international outlook will increasingly be required of senior executives, who also will need to remain vigilant in the face of mergers and hostile corporate takeovers. Political leaders will increasingly need to learn how to negotiate and cooperate internationally in a competitive and sometimes hostile world.
Work/Life Balance. Households with both husband and wife working full-time as well as households with single parents have focused attention on balancing the needs of work with those of home and family. Johnson & Johnson introduced its Balancing Work and Family program in 1988. It was seen as beneficial for recruiting and retention (Cole, 1998). In 1996, of 1,050 employers surveyed, 68% provided flexible scheduling; 86%, child care; 30%, elder care; and 23%, adoption benefits. Other firms were arranging job sharing and compressed work-weeks, help for spouses of relocated employees finding employment, and other benefits for help with home and family. Family friendliness was seen as a strategic objective (Vincola, 1998). Family-friendly corporations have been found to increase shareholder returns, according to an analysis by Cascio and Young (2005) of the 100 Best Companies for Working Mothers. Fathers are taking on more household and child care duties. A combined work and family life is an increasingly common experience for children that is shaping their attitudes (Riggio & Desrochers, 2005).
National Security. The increased importance of the highly equipped and trained individual soldier and small teams in the all-volunteer armed services will continue the need to promote stability and cohesiveness through effective leadership. In 2025, U.S. military thinking will need to make strong use of the political, economic, and social aspects of national security. Training and resources will need to be invested in peacekeeping as well as fighting war. More focus will be needed on small-scale interventions with greater mobility, agility, flexibility, and speed of strategy and decision making. Unless the major confrontation in the Middle East continues, overall, the active U.S. defense forces could be limited to three army divisions, three aircraft carrier battle groups, and four fighter wings. Much more may be invested in small, mobile special operations teams. A robust reserve force may be necessary, as might strong intelligence and understanding of foreign cultures and governments. What is predictable is less dependence on big force-on-force threats and conflicts and more dependence on special operations, airlifting, and close air support, together with diplomacy that integrates the use of U.S. and allied forces. These changes will make us well prepared to deal with the threats of terrorism and guerrilla warfare (Bass, 1998; D. Smith, Corbin, & Hellman, 2002).
Special operations teams in large numbers will replace the big military forces. Their missions are intensive, often secret, highly focused, and usually swift, stressful, and hazardous. Their members are cohesive “bands of brothers”–style heroes who are daring, imaginative, highly trained, and highly disciplined. They exploit short-term opportunities, surprise, weaknesses, and unpublicized vulnerabilities in the well-defended, much larger, and stronger-opposing forces. Whether army Rangers, navy Seals, British SAS (or guerrillas as in Spain or Russia in the past), the teams are led by extraordinary leaders who think “out of the box.” Their leaders need to learn more without postponing decisions, to adapt shrewdly without losing focus, to understand their opponents as well as themselves, and to command while maintaining authority and mutual loyalty. Both teams and leaders need to combine intellectual with physical courage, daring with patience, sacrifice with achievement, endurance with change, and attention to detail with openness to new approaches (Leebaert, 2006, pp. 7–11, 585–587).
International cooperation, multinational alliances, and astute diplomacy will be needed more than ever. Competitive market forces, population pressures, accelerating technological advances, and rapid societal changes further reduce the tolerance for laissez-faire leadership. But care has been taken to avoid a drift back to the promotion of autocratic behavior in the guise of active leadership. Presidents Ronald Reagan and George W. Bush both breached the constitutional limits on the powers of the president in the name of defending the nation against threats to national security. Paradoxically, they both sought to promote democracy abroad by upsetting the constitutional balance of the courts, Congress, and the executive branch at home.
A new managerial mindset is needed. The diffusion of new technologies makes innovation a competitive requirement. A premium has been placed on organizational learning. E-commerce, e-retailing, and online transactions with suppliers and customers increase revenues and reduce costs (Hitt, 2000). Networks of personal computers and interactive programming greatly increase the speed and opportunities for survey studies and the experimental manipulation of leader-follower variables. Advances in miniaturization and telemetry will no doubt be put to good use in the direct observation of neural reactions and nonverbal leader-follower interactions (Schyns & Mohr, 2004). As electronic mail crosses oceans and cultural boundaries instantly, leaders face new opportunities and challenges. Behavioral research on information systems is stimulated. Complexity theory suggests that the order, innovation, fitness, and decline of an organization depend on its networks inside and outside the organization, whether they are closer to or farther from the organization’s technological core. Leaders should help to build and maintain these networks (Marion, 1999). Kahi, Sosik, and Avolio (2003) predicted that with the Internet, some organizational boundaries will disappear. Increasingly, consumers will communicate directly with peer suppliers through networks such as eBay. Hierarchical systems will be revolutionized by networking.
E-technology can bring both positive and negative consequences. It can foster e-leadership and virtual teamwork and promote close monitoring of employees. Spreitzer (2003) sees virtual teamwork continuing to replace face-to-face supervision with more remote management. Virtuality reduces the need to travel, increases round-the-clock operations in different time zones without the need for shift work, and minimizes diversity problems. But it may foster feelings of isolation, cross-cultural misunderstanding, and reduced identification.
Computerization. Fifty years ago, Leavitt and Whisler (1958) predicted that by the 1980s, computerization would change the role of management and leadership in organizations. Much of the routine decision making and reporting by middle management would be computerized. Information technology would replace much of the need for the traditional middle management role. Much of what was predicted came to pass. The world of computerized management information systems and the change in organizations came by the 1980s, as forecast. Middle management was squeezed out, as Leavitt and Whisler predicted, because senior management could be linked by computer directly to lower levels of supervision and to operations. Employee empowerment, collaboration, and teamwork also reduced the need for frequent managerial decision making and intervention (Kielson, 1996). Byrne and Zellner (1988) estimated that more than a million U.S. managers and staff professionals lost their jobs during the 1980s and that more than a third of middle-management positions were eliminated. The tenure of the average manager in a single firm, which was 12 years in 1970, had been reduced to seven years by 1990. By that year, much of the information processing formerly done by management and staff had been taken over by computers. Information technology departments were organized to maintain and modify what was needed.
Leadership and the Information Revolution. According to Cleveland (1985), the character of the leadership role is systematically changing under the shock of the information revolution. This revolution is evidenced by the rapidly changing distribution of the U.S. workforce. In 1920, 9% of the workforce was engaged in knowledge and educational services; in 1955, 29%; and in 1975, 50%. In 2000, it was estimated that 66% of the workforce was involved in such services. The generation, communication, storage, and retrieval of information have been increasing exponentially in speed and amount. Older approaches to getting work done may provide little guidance for the future.
Six interlinked characteristics of information have radically altered the leader-follower relationship. The agenda for research on leadership entails dealing with these properties of information: (1) Information is expandable. The more we have, the more we use and the more useful it can become. Information is not a scarce resource; only time and the capacity of people limit its growth. (Of course, it is possible to be overloaded with information so that decisions based on it become delayed or less than optimum.) (2) Information is compressible. It can be concentrated, integrated, summarized, and miniaturized in its manipulation and storage. It requires little energy and the depletion of few physical resources. (3) Information can replace land, labor, and capital. Whole libraries can be packed into a computerized data bank. Automation replaces people. Organized data reduces localized inventory requirements. (4) Information is highly transportable. Telecommunications makes physical meeting sites unnecessary. People geographically distant from one another can work together in virtual teams. According to Kahai’s (2000) surveys of senior executives, virtual teams are among the top five challenges for organizations to address in the future, along with the search for workers best suited to staff virtual projects. (5) Information is diffusive. It leaks. Despite efforts to maintain secrecy, the leakage may be wholesale and pervasive. (6) Information is shareable. An information-rich environment is “a sharing environment … the standards, rules, conventions, and codes are … different from those created to manage the zero-sum bargains of market economics” (Cleveland, 1985, p. 32).
The information technology revolution enables organizations to solve problems and share innovations with agility and speed by connecting the collective knowledge of their members. Ready access to marketing, production, employee, and customer information will be important for businesses to remain competitive. By 2025, it is estimated that 60% of work will use equipment and operations that did not exist in 2000. Future managers will attend more to operations, products, and outputs. They will coordinate their subordinates rather than telling them what to do. Again, they will coordinate rather than direct consultants, contract workers, and temporary employees outside the organization (Kahai, 2000). Networks of inside and outside organizational members will become more important to managers and leaders (Ronfeldt 1993). Before the information revolution, managers faced the possibility of too little information; now they are faced with information overload. In managing operations, information about objectives without clear boundaries as they affect decision-making processes can be treated as “fuzzy sets” of objectives instead of the less realistic single objective function. When involved with different sources of information, the multiscaling of categorizations can be applied. Computer mediation enhances the speed of communication and interaction between leaders and followers (Fischer & Manstead, 2004).
Age-Status Imbalance. Younger managers and junior officers are likely to have grown up with the rise of information technology. Older senior executives and senior officers are less likely to have developed “hands-on” experience with the personal computer, the Internet, and the Web. Older executives grew up with relative information scarcity. Now they have to deal with an overload of information. Computer-challenged managers and leaders who are not young enough to have grown up with computers are being helped with short courses, self-training, and coaching (Fleener, 2004). Computer literacy is becoming universal and is being instituted faster in the developed world, more slowly in the developing nations. Computer training is furnished in elementary, secondary, and higher education as well as virtually and at work-places.
The Future of the Hierarchical Organization. Like many, Helgeson (2005) argued that top-down power is obsolete and will be replaced by networking. Nonetheless, in agreement with Leavitt (2005), some form of top-down hierarchy will remain in most organizations into the foreseeable future. As Leavitt noted, such hierarchies will persist even though we don’t like them. They will continue to be further modified substantially by the humanist movement and strengthened substantially by the systems movement. We need to learn how to live with them as they are softened by the humanizing changes occurring; at the same time, the controls are being hardened in organizations by modern information technology. We need to reconcile personal growth and fulfillment inside the hierarchy with the “roiling, multi-directional storms” of internal and external “storms of change” (Leavitt, 2005, p. 11). Hierarchy is unpopular in democratic societies because it generates dependence and mistreatment, blocks interpersonal relationships, and provides those at the top with power to become greedy and corrupt. Hierarchies may be slow, unresponsive, and inflexible. Those at the bottom may suffer from violation of democratic values, while those at the top may suffer from intrigues and conspiracies, cronyism, and personal and physical insecurities. Nevertheless, organizational hierarchies are here to stay. In contrast to the alternatives, they are efficient. Yet research continues to be needed on how they can be improved by further humanization and systemization.
Organizational Change. The very purpose of an organization is to provide reliability, predictability, and coordination to its constituent parts. Nevertheless, its environment continues to increase in turbulence, and continuing responsiveness, flexibility, and change are required. Attention to both stability and change in organizations will continue to be required, with efforts held together by good communications. Hierarchy has been changed with the participative management and the team approach of the humanizers, and with the rise of operations research, information technology, and analytical managing by the systemizers.
The Case for Further Humanizing. Although humanizing and systemizing are two opposing approaches to modifying organizations, both are needed for improvement. “Organizations used to be [discussed] as pieces of engineering, flawed pieces maybe but capable of perfection, of precision, of full efficiency designed, planned and managed … [with] control backed by authority. … The new language of organization is about networks and alliances, adhocracy and federalism, shared values … and consensus … options not plans, the possible rather than the perfect, involvement instead of compliance, political systems (rather than) engineering … leadership rather than management” (Handy, 1994, p. 7).
Humanizers push for less hierarchy. As predicted by Bennis (1992, p. 3), traditional hierarchies are being modified by “networks, clusters, cross-functional teams, temporary systems, ad hoc task forces, lattices, modules, matrices.” Tall organizational structures are flattened. Combined with cost-cutting efforts, mergers, and acquisitions, the flatter structures have contributed to a decrease in the number of managers in the resulting new type of organization (Tomasko, 1987). The number of layers in management hierarchies continues to decrease, with larger spans of control for those who remain.
In an expanding economy, newly formed firms can absorb most of the managers made redundant in their former companies. But the decline of the percentage of managers in the U.S. workforce from its high of above 10% in 1980 is expected to continue well into the twenty-first century, exacerbated by the stock market bubble at the beginning of the millennium and the downturn in the economy in the three subsequent years. A career in a single firm from entry to retirement can no longer be counted on. Managers are redirecting their attention more toward personal goals and capabilities. The psychological contract of managers’ loyalty to the firm and the firm to the managers will continue to erode (R. Hirsch, 1987; Marks, 1988). The new contract provides managers with career-enhancing but impermanent opportunities. Managers need to take responsibility for their own careers. There is less status than before in working for a Fortune 500 company. Although 2,000 managers from many different corporations said that there was more than a 100% difference between the top performers and those below average in performance, the actual difference in pay between those at the same organizational level was 5% to 10%. Opportunities for promotion have decreased as layers of management have been reduced. To replace these motives for remaining with an organization, an organization needs to identify whom it wants to retain. Opportunities for involvement, recognition, and development need to be provided. For example, younger managers can be assigned to action learning projects. Organizations can be made more attractive through less bureaucracy, more flexible hours, and a greater sense of ownership (Mason, 1996). Among others, Clegg (1990) saw the need for organizations to move away from ratio-nalistic, deterministic, and bureaucratic forms of organizing, toward more flexible, fluid, and values-driven forms. Decisions have to take into account individuals’ values, emotions, and preferences as well as logic and empirical considerations. The individuals include suppliers and consumers outside the organization (Maas & Graf, 2004).
The Case for Further Systemizing. Systemizers support hierarchy, consistency, discipline, and order. Humans are just another resource amid these factors for the organization. Systemizers want to control, regulate, and reduce variability. They prefer to substitute machine performance for human performance and artificial intelligence for human thinking. They favor managing over leading. They are more supportive of hierarchies as the natural way to get things done than are humanizers. Both will contribute to effective organizational change into the foreseeable future. Boundaryless organizations illustrate a compromise between humanizing and systemizing. Such organizations contain fewer fixed structures and more temporary systems. Technology and systems are assembled as needed for specific projects (Davis, 1995). Boundaryless organizations are clusters of activities whose members and their goals are continually changing. Projects are more significant than positions (Kanter, Stein, & Jick, 1991). Dejobbing, or unbundling of work, illustrates another such compromise. Instead of a permanent position—a bundle of tasks—to be performed by one person, various necessary tasks are done by people inside and outside the organization. The individual member is assigned various tasks at different times (Bridges, 1995). Generally, changes in organization—following sociotechnical theory—encompass both systemizing and humanizing efforts to improve a system and employees’ work satisfaction.
Issue Management. During changes in an organization, an underresearched area is leaders’ role in building the momentum of key issues critical to the change process and reducing the attention to other issues by manipulating the dimensions of the issues. As a big change occurs, members are cognizant of a stream of issues flowing continually and competing with other issues for members’ attention. These issues of organizational change often involve internal changes rather than changes in the external environment. For organizational change to occur, leaders need to move members to attend to the important issues inside the organization in order to solve the problems they present (Dutton, 1988). Leaders can do this by increasing the salience of an issue in an ongoing stream of issues by emphasizing and publicizing it; the ambiguity of an issue by clarifying it; the immediacy of an issue by setting deadlines; the interdependence of an issue by incorporating others’ objectives; and the scope of an issue by broadening its range of activities (Hietapelto, 1994). In nonprofit organizations, Hesselbein (2004), as leader of the Girl Scouts for 13 years and leader of the Peter Drucker Foundation for the same time span, noted three particular challenging issues for nonprofit organizations. First, leaders of change must be developed with a moral compass that works full-time; they must be healers and unifiers who embody the mission, live the values, keep the faith, and remain fully responsible. Second, such leaders need to reflect and embrace diversity in the board, management, workforce, clients, and customers. Third, such leaders need to be open to collaboration, alliance, and partnerships. Nonprofits need to be in equal partnerships with government and private corporations. A more general question for nonprofit organizations is why there has been so little controlled leadership research about them. Many nonprofit foundations are dedicated to leadership development yet provide relatively little for formal leadership research in the nonprofit sector (Riggio & Orr, 2004).
Another consideration that may help predict likely future developments in substantive research on leadership and its applications is the discernment of dissatisfaction with the adequacy of current personnel practices. According to a survey in early 1988 (Anonymous, 1988b), the needs of human resource management that were not being handled well at that time and were seen as requiring emphasis were succession planning, human resource productivity, and organizational design. Less important practices that also needed improvement were team building, the measurement of morale, performance aids, job design, the stress of career development, and outplacement.
Matching Practice with Theory and Doctrine. The importance of a trusting relationship between leaders and their subordinates has been stressed as a principle for effective leadership in many previous chapters. It has been U.S. Army doctrine since 1775, but practice has often failed to match theory and doctrine. For instance, mentoring of junior officers by superiors is a well-established doctrine, but 85% of junior officers reported that they had received their support from such counseling less than a week before the Officer Evaluation Report was due (Bass, 1998).
Practices Endorsed by Popular Books. Academics are often asked what they think about a popular best-selling book on leadership and are likely to regard it with disdain, as anecdotal and not as rigorous as academic research published in peer review journals. But practicing leaders and managers are more likely to have read or heard about the popular book. Fortunately, Dickson, BeShears, Borys, et al. (2003) found that the gap between popular and academic literature on leadership was closing. They summarized the models of leadership themes they found in 30 best-selling books on leadership. At least 10 of the books were written by prominent academics. As seen in the preceding chapters, the themes in the popular press are readily found in the academic literature. They include the importance to leadership of goal orientation, self-awareness, visionary leadership, building of teamwork, development of others, assessment of the organization and environment, and management of organizational culture. Followers also read the popular literature and may expect their leaders to behave accordingly.
A last consideration is the possibility of new paradigms that can affect both future methods and the content of leadership research. In looking to the future, we must not forget the past. Leadership is a mature field of study (Hunt & Dodge, 2001, p. 436). Concepts such as transformational leadership have evolved through stages (Reichers & Schneider, 1990): introduction and elaboration, evaluation and augmentation, and consolidation and accommodation. We have already seen the rapid impact on leadership theory and research of the cognitive revolution in psychology. Of the most prominent academic theories of leadership from 1900 into the twenty-first century, information processing became very active only in the 1990s; charismatic-transformational leadership (the “new leadership”) became very active only in the 1980s; and trait theories, which had been very active before 1920, reemerged as very active in the 1990s (Antonakis, Cianciolo, & Sternberg, 2004). Developments in the mathematics of dealing with irregularities, reversals in trends, and seemingly chaotic conditions may be applied to modeling the natural discontinuities in leader-follower relationships (Lord & Maher, 1991, p. 195). The physical sciences may suggest new ways of looking at short-lived phenomena—for example, the emergence of instant leadership in a crisis followed by its equally instant disappearance. The willingness to accept two distinct ways of dealing with the same phenomenon, as is common in wave and particle physics, may lead leadership theorists to simultaneously treat the leader’s and subordinates’ different rationales for what is happening around them. Cause-and-effect analyses may be seen as the exception to mutual effects between the leader and the led. The 1980s saw an upsurge in interest in upward influence (see, for example, Kipnis & Schmidt, 1983), which has continued. It is likely that a greater number of future studies will be conducted of the reciprocity involved in the effects of upward and downward influence, as was done by Deluga (1988–1988a), who showed that subordinates’ upward influence is depressed more by transactional than by transformational leaders. Generally, we are likely to see more work on how leaders influence and are influenced by the individuals and teams they have been elected or appointed to lead. Hollander (1958) focused importance on the follower in the emergence of the leader. But the follower needed to be active, not passive (Offerman, 1997). “We are content to follow when we lead the way” (Homer, c. 750 B.C.E.).
Societal, technological, and organizational changes call for new ways of thinking, the importance of creating new knowledge, and its diffusion and utilization in the organization. Many organizations consider their personnel as human capital assets accruing from their investment in human resources (Hitt, 2000). In contrast to regarding employees as assets to be developed and retained, another strategy is hiring temporary workers. A firm’s workforce and costs can be expanded or contracted depending on its needs. Or whole functions such as payrolls can be contracted out. The workplace can be replaced by work at home, with subordinates linked to supervisors and colleagues by e-mail and telephone (Challenger, 1998). Faced with global competition, firms will put an emphasis on innovation, flexibility, and responsiveness. Leaders’ attention will shift from tangible to intellectual resources (Dess & Picken, 2000). Attention will also continue to be paid to the company image, reputation, political connections, and strategic flexibility (Hitt, 2000). Methodological and substantive issues will broaden. Korukanda and Hunt (1991) see conflicting paradigms about leadership that may be combined. For instance, leadership research can be objective or subjective. The objectivist view assumes that leadership is a natural object—a concrete reality amenable to quantification. Understanding accrues from the discovery of facts about leadership and its processes and relationships. Inferences are drawn about the statistics of the average individual. The subjectivist view is that leader and leadership are convenient labels or fictions that facilitate sense making for the observer or for introspection. Results are usually qualitative. Self-awareness and self-improvement are of central importance to the information gathered. Holistic leadership demonstrates the subjectivist approach. The leaders of the Body Shop, for instance, focus on organizational processes with a systems view of interactions between people in their various roles and relationships. They keep the big picture in mind and practice empowerment. The senior managers refer to themselves as support staff. What the people in their franchised stores do is more important to the business than they are. Self-awareness and self-improvement are stressed. A new paradigm may combine the objectivist and subjectivist views to help understand the phenomenon better.
In the public sector, government agencies will ideally become more empowering and supportive rather than serving; competitive but not monopolizing of the delivery of services; mission-driven rather than rule-driven; results-oriented; meeting the needs of the client rather than the needs of the bureaucracy; likely to run surpluses rather than deficits; market-oriented, anticipatory, participative, and decentralized (Gore, 1993; Osborne & Gaebler, 1993). Change is also appropriate if it is creative and humane and there is a dedication to public service (Denhardt, 1993). Research, graduate education, and empirical studies in the United States have had strong influences on public administration in Britain, Belgium, the Netherlands, Germany, and Scandinavia but not in France, Spain, or Italy (Nelissen, 1998). It is proposed that by 2015 diplomacy between nations will not be limited to diplomats, but, thanks to the Internet, “diplomacy” will include governments and NGOs as well as individual citizens from the general public of the nations involved. Efforts to promote such virtual diplomacy have already been initiated (United States Institute of Peace, 2002).
Changes in the Prominence of Leadership Theories. From 1991 to 2000, of the 188 articles published in The Leadership Quarterly, the most prominent were those dealing with charismatic and transformational leadership (34%), contingency theories (25%), and other, newer directions such as political leadership, strategic leadership, top-management teams, leader-Member exchange, and other multiple-level approaches. This was in contrast to fewer articles on trait theories and behavior theories of leadership. No doubt some of these will continue in prominence in the twenty-first century, while some will be replaced by newer entries stemming from new developments in genetics, technology, and European and Asian influences. In looking ahead from 1975 and 1995, Phillips (1995) suggested that, increasingly, theories and research about leadership schemas and cognition will modify trait, behavioral, and situational contingency theories.
In examining how leadership relates to outcomes, we continue to be faced with the question of what is cause and what effect. Thus, Kernberg (1984) stated that breakdowns in work effectiveness are erroneously attributed to failures of leadership. Yet often they are due to failures in systems and environmental, organizational, and group factors. Ineffective leadership is an effect rather than a cause. Supervisors may be supportive because they have productive subordinates, or subordinates may be productive because they have supportive supervisors (Bass, 1965c). Some investigations (see, for example, Greene, 1974) have shown that subordinates cause their leader’s behavior. Other studies (see, for instance, Dawson, Mes-sick, & Phillips, 1972) have found that the leader’s behavior is a cause of the subordinates’ performance. Still others (such as Jacobs, 1970) have pointed to mutual causality in an exchange in which subordinates comply because of the leader’s promise and the leader rewards subordinates for the compliance.
Although they still form a distinct minority, a sizable number of causal studies of leadership have appeared since 1970. Of 89 studies of leadership between 1970 and 1975 reviewed by Hunt, Osborn, and Schriesheim (1977), almost all the 17 laboratory studies but only 24% of the 72 field studies concentrated on causal relations. Despite the shortcomings of laboratory studies, they still make collecting causal data convenient—something that is much more difficult to accomplish in the field. However, it is edifying to see the increasing efforts to combine laboratory experiments with field studies and to search the field for corroboration of laboratory findings. But care is required, especially about the need for highly reliable measures. Regression strategies to determine cause and effect, such as path analysis, require highly reliable measures and strong enough relationships to permit the testing of alternative models.
The five-year review by Hunt, Osborn, and Schriesheim (1977) of six major journals found that 72 field studies and 17 laboratory investigations focused mainly on supervisors’ behavior and reactions to it. In the field, managers and their associates in a wide variety of organizations were the subjects of inquiry. In the laboratory, superior-subordinate relations were simulated by students. A fundamental question was whether the temporary nature of laboratory situations could faithfully represent the real-world relationship. Osborn and Vicars (1976) noted a particular source of error in trying to generalize from laboratory studies of leader-follower relations: short-term laboratory situations tend to evoke the behavior of participants on the basis of available stereotypes. Extensive interpersonal contact in real life provides a more realistic basis for behavior.
Field studies, by their very nature, are fraught with internal and external threats to the validity of their data and measurements. Nonetheless, although laboratory experiments provide rigor and control, relying exclusively on laboratory studies should be avoided. As Meehl’s (1967) paradox indicates, the more aseptic and controlled the laboratory study, the greater the precision of its outcome. In physics, greater precision increases confidence in the generality of the finding; in the social sciences, it does just the opposite. Ideally, laboratory studies should be planned in conjunction with fieldwork.
In controlled laboratory experiments, students are usually the subjects. When there is disagreement between the findings from such experiments and field studies of operating supervisors and employees, the greater rigor and control of the experiments often leads to greater confidence in their conclusions. They are preferred to field studies, although at times, field studies may obtain results that are closer to the truth. Thus, Fodor (1976) showed that the results of laboratory experiments, with practicing supervisors instead of students as participants, may be quite different from the results obtained in the field. In a rare laboratory study that utilized industrial supervisors as participants, Fodor found that, in comparison to control subjects, supervisors who were exposed to a group stress situation responded by giving significantly fewer pay raises and lower performance evaluations to compliant workers. However, field studies (such as Goodstadt & Kipnis, 1970) earlier found just the opposite; in those studies, compliant workers received significantly more pay increases under group stress than under control conditions. Nor were the laboratory results with supervisors consistent with the results that Fodor (1974) obtained with students. In all, we can expect to see an increase in efforts to examine and report laboratory experiences jointly with tests of the same hypotheses in the field.
The law of large numbers states that large random samples will be highly representative of the population from which they are drawn. The law of small numbers assumes erroneously that small samples will be similarly representative. If you have a sufficiently large sample, any difference is likely to be statistically significant. Testing the null hypothesis with a small sample of differences is a meaningless ritual (Cohen, 1994). Critics continue to voice their objections, but statistical significance testing of differences in results remains strong. Tversky and Kahneman (1971) demonstrated that the belief in the law of small numbers leads to highly inflated estimates of the amount of information contained in studies that use small samples. Schmidt and Hunter (1980) thought that much of the variation in the observed relationships in small samples could be considered random departures from a relatively simple overall generalization. The overall generalization does not need to be qualified by the particular situation involved. For example, in examining the relationship between individual competence and job performance, Schmidt, Hunter, and Urry (1976) showed that the samples were usually too small to produce acceptable levels of statistical power. Thus, if the true correlation between, say, scores on intelligence tests and the success of leaders is really about .35, any sample of 30 to 50 would yield a result that was statistically significant, from zero only 25% to 50% of the time. For a statistical power of .90, in order to reject statistically the zero relationship 90% of the time when the true correlation is .35, sample sizes of 200 or more are needed.
Erroneous Interpretations of Small Differences. In organizational research, we ordinarily measure the responses of individuals nested in their work groups. They are not random and are likely to be influenced by their groups. In the same way, if the groups are nested in different organizations and influenced by their organizational membership, they are not random (Hanges & Shteynberg, 2004). They need to be treated by the variant approach (Dansereau, Alutto, & Yammarino, 1984), by hierarchical linear modeling (Hofmann, 1997), or by factor analysis (Muethen, 1994).
We may look at ten small-sample studies of scores on intelligence tests as they relate to leadership performance and find that half are statistically significant and half are not. Then we try to infer a reason for the different findings, when in fact the differences in the various samples are null and can be accounted for by the law of large numbers. The obvious implication is our need to be cautious in interpreting the meaning of situational variance when the data from the different situations are based on small samples. With contingent analyses using small samples, we will err considerably in rejecting the null hypothesis at the 5% level of confidence that no differences exist between different contingencies. Thus, Hunter and Schmidt (1978) noted that 28% of the time (at the 5% level of confidence) we could erroneously infer differential relationships between, say, intelligence and performance for 30 black leaders compared with 30 white leaders when no true difference exists.
The problem is far from academic. Hunt, Osborn, and Schriesheim’s (1977) review of 89 reports found that 20% used samples as small as 30 or even fewer in analyses of data. The problem is compounded when we deal with leadership because ordinarily only small samples of leaders are available unless an organization is very large. Thus, we may need to reexamine carefully how much contingent results occur because of the low power of the sampling on which many are based.
Meta-analysis offers the opportunity to test how random are the various results obtained from the pool of small-sample studies of the same relationship and to arrive at a mean effect size for a given relationship. Here, however, we have to be cognizant of the inflation in the estimated effect size resulting from the adjustments for criterion unreliability and restriction in range. These adjustments may be highly unrealistic because they are based on assumptions about infinitely repeated measurements and samples with the full range of possibilities. Nevertheless, as has been noted, meta-analysis helps to tease out reliable contingent effects from those that are ephemeral, transient, or a consequence of random variation from a true mean relationship.
Erroneous Conclusions from Small Effects. Exacerbating the problem of small samples is the small size of many of the mean effects that emerge. It requires little in the way of systematic errors to distort or confound them. Systematic errors creep into the scene for a number of reasons (Webster & Starbuck, 1987). There are broad characteristics of people, such as intelligence, that appear to correlate with leadership and with outcomes when it is possible that there may be no direct relationship between the leadership and the outcomes. It may be that leaders tend to be more intelligent and that better outcomes require more intelligence, but there may be no direct link between the leadership and the outcomes. The failure to consider a contaminating third variable may result in the unwarranted blowing up of the importance of many small mean effects. Thus Woodward (1965) found that decision-making processes within an organization depended on the organization’s technology and structure. Many subsequent studies elaborated on this finding. However, Gerwin (1981) showed that the modest results obtained tended to disappear when the size of the organization was held constant. Multiple-regression control variables are likely to become the rule, rather than the exception, in future research.
Convenience and feasibility have often dictated that a sample studied was actually a selected complete subpopulation, rather than a random representation of a complete population. For example, all the supervisors in one Canadian department store are surveyed or interviewed, rather than a representative sample of all Canadian retail sales supervisors. Researchers still have to rely too much on convenience samples—samples that are most often obtained from larger rather than smaller organizations. The organizations must be cooperative and supportive or the data can be collected only on the outside or by an unobtrusive participant-observer. The combination of large organizations with accessible personnel systems, data banks, mailed surveys, and telephone interviewing will increase the possibilities of large-scale representative sampling.
Single-Source Variance. Convenience also leads to reliance on the same respondent for descriptions or evaluations of a leader’s behavior and the outcomes of the leadership. Methods for correcting for the built-in correlation that is due to such Single-source variance have been developed. They include removing from the data the first general factor that is assumed to be due to the rater’s propensities (Podsakoff & Organ, 1986), using one respondent to describe the leader and another to describe the outcome, and the varient approach of breaking the analysis of covariance into that among leaders and among between multiple respondents rating the same leader (Avolio, Bass, & Yammarino, 1988). However, so far, no one corrective action seems to be free of problems.
Lord Kelvin overreacted when he declared that in the pursuit of understanding, “if you can’t measure it, you don’t know what you are talking about.” Handy (1994) retorted that when you cannot easily measure the phenomenon you wish to understand, it is wrong to disregard it if the phenomenon is difficult to measure or to argue that it is unimportant or that it doesn’t exist.
Need for Balanced Use of Standardized Measures. A wide diversity of environmental, task, group, interpersonal, and personal variables have been employed in leadership research, each to a lesser extent, usually in multivariate fashion, as antecedent conditions, correlates, or moderators. Standardized scales for some variables, such as role clarity and role conflict, have been adopted, despite the many different but closely parallel conceptualizations of leadership style. Instruments such as Fiedler’s Least Preferred Co-worker (LPC) scale and some form of the Leader Behavior Description Questionnaire (LBDQ) have dominated research on leadership in the past 40 years. This use of common standardized instruments has made possible a great deal of comparison across studies. It has already been noted how even small changes in instruments may lead to large differences in outcomes. But this concentration of leadership measures has kept the research establishment from looking at many other, possibly more important, aspects of leadership behavior. A balance is needed. On the one hand, researchers need to avoid inventing new measures of the same attributes when old ones with satisfactory reliability and validity are available. On the other hand, concentrated efforts with measures other than the most popular ones are needed, particularly if they can be joined in a nomothetic network with the well-used instruments.
Limitations. Elaborate theories may spin out tales of curvilinear complexity and multiplicative effects. Yet analyses over 50 years based on such efforts have usually added little beyond error to the prediction equation, even though the record is improving with more sophisticated measurements. Theory building should not go too far beyond what is empirically possible. Although contingency models are intuitively appealing (Yntema & Torgerson, 1961), contingency hypotheses (how X relates to Y depends on Z), Korman (1973) noted, should be entertained in a theory only if they are empirically supported and necessary. Consistent with what has already been said about the erroneous interpretation of small differences, we may see a return to the positing of simpler relations based on larger samples.
Most observational studies use the single, mutually exclusive coding of the various categorizations of observed data. But such coding is no longer necessary. Statistical and computer programs are now available to analyze the multiple coding of multipurpose activities, multiple contacts, and simultaneous roles played (Martinko & Gardner, 1985).
Cross-lagged correlational analyses to demonstrate leader-subordinate relationships also have their limitations. For instance, Greene (1976a) noted that initial changes in leadership behavior between Time 1 and Time 2 may fail to be associated with parallel changes in subordinates during that same period, but they may show up between Time 2 and Time 3. Thus, three, not two, data collection points are needed. In addition, it is difficult, if not impossible, to rule out the possibilities of confounding uncontrolled changes in third variables on both the leader and the subordinate. Hence, the increased use of repeated measurements over many time periods, such as that reported by Howard and Bray (1988), is needed.
Yukl (1982) noted that factor-scaled questionnaires fail to include important items of leadership behavior that are correlated with two or more factors. Infrequent behaviors are also likely to be missed. Frequency is overemphasized while the leader’s sequencing, timing, and style of execution are neglected. Often, the context in which the behavior appears is also missed. Richer information is also likely to be obtained from questionnaires if in addition to the items that purport to be objective behavioral descriptions, researchers use more items that tap the observers’ gut feelings and estimates about the potential for future assignments of the leader being observed. We need to avoid, because of special interest in the leader-subordinate dyad, eliminating descriptions and evaluations of overall leader-group relations, for it is clear that both are important. Conversely, items about the interaction of the leader with the whole group fail to provide information about the leader’s dyadic relationship with each individual member of the group (Yammarino & Bass, 1988).
Simple versus Complex Hypothesis Testing. We face a dilemma. On the one hand, “the chain of relationships between leader behavior and outcome is long and complex” (Lieberman, 1976a). Much of what needs to be understood is missed if we simply try to relate leadership behavior, particularly generalized leadership behavior, to final group outcomes. The relationship must be considered in terms of the group’s norms, cohesiveness, and so on, as well as of the leader’s characteristics. Members’ expectations may be more important in determining group outcomes than anything the leader can do. Multivariate analyses, complex models, and contingent and moderator analyses need to be used. On the other hand, the limitations that have been noted in the measurements available, and the extent to which situational variations from one small sample to another, are likely to be random rather than true effects. At the same time, second-or higher-order interactions are likely to render inconsistent simple moderator effects from large sample to large sample. Thus, when faced with complex models, we need to be more open to experience and greater clinical understanding of data that demand less mathematical rigor (Bass, 1974; L. G. Cronbach, 1975). On the other hand, structural equation models (Gavin & Cheung, 2004) have become increasingly popular in dealing with both the direct and indirect paths between independent, intervening, and dependent variables. Each path or partial correlation corrects for the other paths and helps understand the network of relations.
Models may be built a priori on the basis of logic and prior information, and then tested to see if they fit the data obtained in the test. Or models can be built a posteriori to fit the obtained data. There are problems with each approach. Korman and Tanofsky (1975) pointed out that a priori models are hard to use because of the difficulties of accurately estimating the necessary parameters. Yet a posteriori empirical models may be fraught with psychometric error. They may be likely to exploit random effects. Both kinds of models may be more helpful in identifying the important elements for study rather than discerning the final true relationships.
Field studies can reflect the complexities of the real world, but laboratory studies and their controls do not provide the solution to dealing with complexity. For instance, in the study of the effects of sex differences on leadership, the short-term, artificial nature of leadership created in the laboratory results in participants’ relying on stereotypes that influence their responses to the leader. This reliance on stereotypes is less likely to occur in field studies in which leader-subordinate relations are long-term. The laboratory engenders exaggerated sex-role demands, and thus the effects of sex differences in the laboratory are not replicated in the field setting (Osborn & Vicars, 1976).
Need for Qualitative Methods. These methods may be more suitable for providing confidence in the results when complex hypotheses are involved. Bryman (2004) was able to enumerate 65 acceptable published qualitative studies of leadership between Pettigew (1979) and Vangen and Huxham (2003), although the historical and management literatures are replete with many more qualitative comparisons. Additionally, Amabile, Schatzel, Moneta, et al. (2004) completed analyses of daily diary narratives written by leaders’ subordinates.
Bryman, Bresnen, Beardsworth, and Keil (1988) examined the situational factors that construction project leaders take into account when deciding what style of leadership to adopt. Interviews revealed that site managers continually adjusted their leadership styles to suit varying circumstances, such as time pressure, their subordinates’ personalities, and the degree of control they had. The authors suggested that qualitative research can uncover a wider array of contextual variables. Such variables are grounded in people’s experiences and therefore are more accessible to leaders and researchers alike. The search for meaning and significance in the behaviors of leaders and their followers, as well as in related events, is aided by qualitative research (Van Maanen, 1979). People’s actions can be explained in terms of the total context in which they occur instead of the isolated or manipulated elements within the situation (H. Smith, 1975).
Qualitative research is likely to begin with deductions from a theory or a set of general propositions (Orpen, 1987) and then to proceed as a detective might to track down patterns, searching for consistencies in the qualitative information. Qualitative research need not begin with precoded systems but can depend on analytic inductions (Strong, 1984). Creativity and controlled imagination are required to move from specific findings to general conclusions (Mintzberg, 1979). Thus, the movement is a circular process involving the search for and collection of specific data, development of crude hypotheses, and then the examination of the data (or new data) to see how well the inferred hypotheses fit the data (Brogdan & Taylor, 1975). More such qualitative research is needed and is likely to find its way into the study of leadership as the limitations of quantitative methods in dealing with organizational complexities become increasingly apparent (Orpen, 1987).
McCall and Lombardo (1978) advocated more leadership research using the ethnographic methods of non-statistical naturalists to detect the subtleties and nuances involved in the leadership process. Greater attention needs to be paid to unconscious motives that affect leaders’ and followers’ perspectives. Often, qualitative research can deal better with the art and craft of leadership than more objective qualitative analyses. That is, there is much art in leadership that is difficult or impossible to put into a test tube. Nevertheless, there is much regularity in this art that can be made understandable by detecting and describing the patterns that appear. Orpen (1987) regretted that, for the most part, qualitative and quantitative research are likely to parallel or remain independent of each other, rather than being integrated, even though they could do much to complement each other. This was demonstrated by Gibbons (1986), for example, who made use of both quantitative surveys and in-depth interviews of the same executives.
A risk in qualitative research is that we may learn more about the investigator than about the complex scene being investigated. Chafee (1987) noted the divergent conclusions reached by four investigators who looked at the same qualitative data about college presidents. The investigators’ conclusions diverged, she argued, because the investigators came from different theoretical traditions: organizational theory, leadership theory, strategic theory, and anthropological theory. The same key words had different meanings for them. That is, the researchers used the same terminology but drew different inferences from the same qualitative database. Ideally, studies need to triangulate qualitative with quantitative methods. They need to check the results of one against the other. Bryman (2003) was able to enumerate 11 such studies between Rosener (1990) and Voelek (2003). Rosener followed up a quantitative survey of woman managers to generate research questions that became the basis of a full qualitative study that followed. Voelek used qualitative data about university library managers to support and expand quantitative results about them.
Webster and Starbuck (1987) pointed to another handicapping bias in research on leadership. Within each research discipline—industrial/organizational psychology, social psychology, sociology, political science, educational administration, and so on—there are shared sets of beliefs, values, and techniques. The empiricists in each discipline prefer to measure certain sets of variables of consequence. Thus, the dimensions of leaders’ and subordinates’ behavior examined and the methods employed to do so continue to be narrow (Greene, 1976c). The narrowness depends on the disciplinary background of the investigator. There appears to be relatively little consensus on substantive issues of consequence (Campbell, 1982), especially across interdisciplinary boundaries. Despite this lack of consensus, we are likely to see, in this century, an increase in the use of common methods in the study of leadership, depending on whether the investigator is a behavioral, social, or political scientist or an historian, but one with a continuing wide range of substantive foci.1
Leniency. The need will remain in future leadership practice and research to guard against leniency bias. Perlmutter (1954) found that the greater leaders’ abilities to influence other group members, the greater the number of favorable traits applied to the leaders—and the more socially desirable the traits attributed to them. Schriesheim, Kinicki, and Schriesheim (1979) demonstrated the strong leniency effect in the LBDQ-XII consideration scale. A specific leader will earn higher LBDQ scores from respondents who give favorable ratings to leaders in general. Similar leniency is likely to be found in related measures of leaders’ relations orientation, participation, and support. Leniency may account for much of the association between subordinates’ descriptions of their leaders’ consideration and the subordinates’ satisfaction. Measures of the initiation of structure that are free of coercive items and measures of consideration that are free of leniency (if such is possible) will provide more precise measurements against which to pit situational and personal variables for study. The Multifactor Leadership Questionnaire’s scales to measure transformational and transactional leadership appear to be relatively free of leniency effects (Bass & Avolio, 1989).
Errors in Leaders’ Self-Ratings. Except for using it for personal development and expectations and exposing the gap between self-and others’ ratings, future leadership research needs to avoid depending on self-ratings, as many studies still do. In earlier chapters, it was repeatedly pointed out that there is little or no relationship between leaders’ self-descriptions and descriptions of them by their subordinates or superiors. Leaders’ self-ratings and self-reports are suspect. Thus, in an intensive interview study and work flow analysis of 34 pairs of supervisors and subordinates, Webber (1980) found that the supervisors reported that they initiated almost twice as much verbal interaction with their subordinates as the subordinates perceived had occurred. At a considerably lower level, subordinates also overestimated the extent to which they initiated interactions with their superiors. Leaders’ self-ratings have consistently been found to relate poorly or not at all to various dependent variables (Schriesheim & Kerr, 1977b). Most are probably contaminated by social desirability (Schriesheim & Kerr, 1974). They contain self-serving, self-vindicating biases and are likely to generate descriptions of what leaders think is expected of them in their organization and society, rather than an accurate portrayal of their behavior relative to other leaders’. The manager who assures everyone that there is always full consultation on subordinates’ problems since the manager’s door is “always open” is not uncommon. Then there is the self-described “democratic” manager who announces that the organization is going to be democratic or else sanctions will be imposed.
No wonder so little correlation is found between subordinates’ perceptions of what the leader does, what they think he will do, and what he should do, on the one hand, and the leader’s self-reports on the same issues, on the other. Thus Rees and O’Karma (1980) observed that city department managers and their subordinates differed significantly in how the managers would behave in different situations described by the LEAD questionnaire. Holton (1984) reported similar discrepancies between leaders of a Cooperative Extension Service program and their staff subordinates. Nanko (1981) noted that although 1,800 elementary school teachers, using the LBDQ, judged that their supervisor was doing a poor job, each of the supervisors saw him-or herself as doing a good job. The correlations of responses from the two sources were close to zero. Burt (1984) found that 32 heads of hospital departments gave themselves better ratings on the LBDQ than their 379 employees gave them. In the same way, according to Dalessio (1983), working business students and their bosses disagreed about their bosses’ leadership. As was noted earlier, supervisors see themselves as having bigger and more important jobs than their bosses say the supervisors have (Haas, Porat, & Vaughan, 1969). Leaders generally see themselves as more transformational than subordinates see them (Bass & Avolio, 1988). According to a survey using the Multifactor Leadership Questionnaire of a representative sample of 186 surface fleet officers in the U.S. Navy, the correlations of the self-versus subordinates’ ratings of leadership were only .21 for charismatic leadership and .21 for individualized consideration (Bass & Yammarino, 1989). The self-subordinate correlations for the remaining five scales were close to zero.
In Birnbaum’s (1986) survey of 252 college presidents about the effectiveness of the institutional leadership of the average presidents, their predecessors, and themselves, the presidents gave themselves a mean rating of 77.3; the average president, 65.6; and his or her predecessor, 52.0, on a scale of 0 to 100. A serious issue raised by these discrepant results is the reliance that so many studies of chief executive officers (CEOs) place on interviews with CEOs about themselves. Given the low or even zero correlations between what leaders do and their colleagues’ descriptions of their behavior, we need to proceed with great care in drawing any inferences from leader-only data. In meeting with small informal groups of CEOs, I found that CEOs are likely to be highly and selectively biased in their self-descriptions. This is not to say that the leaders’ self-descriptions are necessarily incorrect and the colleagues’ descriptions of the leaders are necessarily correct. Rather, it is important for researchers to avoid depending solely on leaders’ self-descriptions. Moreover, leaders obviously differ. Some are much more congruent than others in the extent to which their self-reports match those provided by their colleagues. Such congruence with their subordinates can increase the subordinates’ morale (Browne & Neitzel, 1952), the quality of the leader-subordinate relationships (Graeri & Schie-mann, 1978), satisfaction with communications (White, Crino, & Hatfield, 1985), and superiors’ evaluations of the leaders (Bass & Yammarino, 1989).
Increasing the Validity of Self-Ratings. Various techniques help curb inflated self-ratings: (1) In forced-choice methods, a self-rater chooses from pairs of self-descriptions that appear equally favorable or socially desirable but in which only one of the pair is valid for prediction or assessment purposes. (2) Self-raters may be asked to complete a questionnaire as if they were applying for a job. Then they are asked to complete the same questionnaire as if it were to be used by a trusted adviser. (3) Self-raters are attached to a bogus pipeline, a physiological apparatus that purportedly is able to register their genuine responses. A meta-analysis of 31 studies demonstrated that, compared to a control group without the bogus pipeline, socially desirable self-ratings were lower (Roese & Jamieson, 1993).
Training and research efforts will, over time, make greater use of superiors’, peers’, and subordinates’ ratings and less of leaders’ self-ratings of their purported behavior. But, as will be discussed later, leaders’ own perceptions, attributions, cognitions, and opinions will continue to be of considerable research importance as a link to what leaders actually do.
Other Systematic Errors. Ratings of leaders and followers are built around memories of their behavior. What leaders or followers have actually done may register less in current evaluations of them than false memories of what they did. Past events may be remembered as far more important than they actually were. The “Rashomon effect” is illustrative: four people witness the same violent event; each later recalls a highly differing version of the same event. In future, we should expect further studies of the effects of false memories on ratings and evaluations.
Among all the biases in perceptions and ratings, a few are most salient in the study of leadership. One such tendency is for subordinates to see more good in their own relations with their leader than in the quality of leader-subordinate relations they observe elsewhere. Adams, Prince, Instone, and Rice (1984) found that when 400 freshman cadets at West Point described good leadership incidents, the incidents usually involved their interactions with their leaders, whereas the cadets perceived bad leadership as occurring mainly in other units. This perception may be part of the larger phenomenon of “we-they” bias, that is, seeing “our group” as better than “their group.” In the future, the use of multiple sources of information with demonstrable convergent validity and suitable corrections of such self-other bias will be more routine.
Consistent with earlier work,2 Larson (1982) and Larson, Lingle, and Scerbo (1984) demonstrated that raters’ responses to LBDQ descriptions of leaders who were viewed in action on a videotape were affected by whether the respondents were cued before viewing the tape about whether the leader they were looking at was considered effective or ineffective. Such cuing before they observed the videotape of the leader in action moved them to selectively encode into memory different kinds of information. Given their implicit theories about the effects of initiation and consideration, the raters saw the supposedly effective leader as higher in both. Binning and Fernandez (1986) showed that another bias in the LBDQ descriptions was due to the differential availability to the rater’s memory of the different items. The availability of items to the rater’s memory is greater the more familiar, dramatic, specific, positively salient, retrievable, and imaginable the items of leadership behavior in the LBDQ are. An average correlation of 48 was found between the extent to which an item of leadership behavior was seen as descriptive of a leader and the availability of the item to memory. The correlation was higher if information about the leader was more limited. In the future, standardized corrections in the LBDQ and similar questionnaires are likely to be introduced to adjust for these availability biases.
An aspect of the bias in self-fulfilling prophecy was detected by Schoorman (1988), who verified that the supervisors of 354 subordinates who had originally participated in the decision to hire them tended to bias their subsequent performance appraisals in a favorable direction if they had supported the decision to hire. They tended to do the reverse if they had opposed the decision.
These consistencies among persons because of their individual predilections and the tendency to be influenced by personal memories reinforce the need to have multiple sources of data about leaders and outcomes attributed to the leaders. Or else, as was mentioned earlier, corrections need to be made to allow for single-source bias when only a single source of information is possible. An example is when a dyadic analysis of the leader with each subordinate calls for each subordinate to describe both the felt quality of relations with the leader and how much the subordinate is committed to the organization’s goals.
Phillips and Lord (1986) noted that leadership questionnaires need to be designed according to the type of accuracy being sought, that is, whether it is behavioral or classificatory. Also, the theory (or lack of it) underlying the questionnaire needs to be taken into account when the questionnaire is employed, to assess interventions that attempt to change leadership behavior.
At times, the need to describe leaders’ performance in terms of their one-on-one relationships with each of their subordinates is more conceptually pleasing than empirically fruitful. However, the dyadic approach can be applied with considerable utility to a variety of research questions. For example, sanctioning, punitive leadership behavior seems to be the result, rather than the cause, of the inadequate performance of specific subordinates. It follows that dyadic analysis should reveal strong differential dealings of supervisors with their subordinates on the basis of leaders’ differential judgments about each of their subordinates.
Indeed, we are likely to find a variety of important consequences stemming from Hollander’s (1978) observation that leaders have “A” lists and “B” lists of subordinates. The As are closer, and the Bs are more distant. Work-oriented leaders are likely to relegate the more incompetent of their subordinates to the B list and to treat them more punitively. But the As will be expected to be more loyal and obedient and will be required to maintain higher standards of performance. However, person-oriented leaders may exert extra effort with their black sheep and may even think that their most competent subordinates are sources of conflict because their performance exceeds the group’s norms. The linkages to Least Preferred Co-worker need to be investigated.
Several different approaches have originated, to tease out individual, dyadic, and group effects. One approach is to hold leader-group effects constant to see what happens within each leader’s group. Vecchio (1982) obtained data on attitudes and performance from in-group and out-group members of 48 four-man military groups. Within-group leader-Member differences predicted attitudes, but only after between-group leader differences were held constant. Performance outcomes appeared to be unaffected by dyadic or group leadership. Katerberg and Hom (1981) used a stepwise hierarchical regression analysis. They first determined the contribution of each of 31 U.S. Army National Guard units to LBDQ descriptions of the units’ leaders, then the contribution of the individuals within the units. Although the effects of within-unit variations in LBDQ scores were stronger than between-unit variations, both significantly predicted the subordinates’ reactions.
Still another approach uses an adaptive process regression technique that determines the artificially assembled groups of an original pool of subordinates that will result in the best possible predictions. The nature of these groups then gives the investigator an idea of which possible groupings may moderate the dyadic leader-subordinate relationship. Berkes and Rauch (1981) found such artificial clusters among 800 police officers. One moderating cluster of those officers operated under less specified procedures. Another cluster was younger, better educated, and more interdependent in their work. Less role clarifying and more participative leadership were required for satisfying the young, educated, interdependent officers.
The use of dyadic analyses of leaders’ relations with individual peers, individual superiors, and relevant others will continue to increase owing to the continuing interest in leader-Member exchange theory (Graen, 1976) and its quantitative possibilities. However, allowance clearly has to be made for leaders not only to relate on some dimensions, at least, in the same way to every group member, but to the generation of effects that transcend the average member. For instance, an assembly bonus effect in productivity may be due to a leader’s structuring the group as a whole, which goes beyond the contributions of the group’s members.
The Variant Approach. As was noted in earlier chapters, the most ambitious, comprehensive, and possibly defensible effort, the varient3 approach, was developed by Dansereau, Alutto, and Yammarino (1984), who used this method to reexamine the reports that originally supported the importance of the individual leader-Member exchange dyad over the leader-group relationship in how much negotiating latitude the leader provides each member and the group as a whole. The analyses concluded that both individual member effects and group effects may be equally salient (Nachman, Dansereau, & Naughton, 1983).
The varient approach is a paradigm for formulating and testing theories that explicitly consider both the involved variables and entities—the individual, the group, and/or higher levels of analysis (Yammarino, 1988). The future will increasingly see its use to clarify: (1) the differences among leaders in their average leadership style (ALS); and (2) their vertical dyad linkages (VDL) or leader-member exchanges (LMX), as was done by Markham, Dansereau, Alutto, and Dumas (1983). More superior-subordinate communications will be profitably analyzed by the varient method (Dansereau & Markham, 1987). A higher level of effects may be examined, such as the effects of whole departments or organizations and their different policies on the leaders’ relations with their group members as a whole, as well as with various individual members.
New nomological networks of variables will be clarified by WABA procedures. For example, a leadership behavior may be related or unrelated to two outcomes. Or the two outcome variables may be related, but the leadership behavior may be related to one outcome, not to the other. Another possibility is that the two outcome variables may be unrelated and leadership behavior may be related to one outcome, but not to the other. The leader’s dyadic and group relations may affect the members’ satisfaction but not the members’ productivity, for instance. New moderator variables and contingency effects will be identified with WABA. One possibility is that variables are directly related; that leaders’ behavior, for instance, directly affects outcomes. Another possibility is that a relationship between two variables (leadership behavior and an outcome, for example) is dependent or contingent on the values of a third variable, such as company regulations. This indirect contingency is a moderated effect. Here, leaders’ behaviors could relate strongly to dyadic and group outcomes in departments with few regulations but not in departments with many regulations (Dansereau, Alutto, Markham, & Dumas, 1982).
New approaches to the study of leadership will continue to proliferate. The method of choice for study and application depends on the way the leader-follower phenomenon is conceptualized. Among the most likely approaches are cognitive and information processing, phenomenology, motivation, psychodynamics, and behavioral observations and reports. At the one extreme are cognitive investigators who focus on perceptions, causal attributions, and expectations. At the other are behavioral investigators who concern themselves with stimulating conditions, behavioral repertoires, and reinforcements that are contingent on subordinates’ performance. The vast array of possible variables to study must continually be pruned. For surveys and experiments to be manageable, researchers must single out for study different aspects of the total process of leader-follower interactions. But the models they construct that focus on one aspect or another are not necessarily inconsistent with each other. The potential exists for considerable integration of the models.
As noted in Chapter 4, by the 1940s, a variety of methods was available to study leadership, including observation and time sampling, sociometry, position and office holding, analysis of biographies and case histories, and judged requirements for leadership. Measurements were being obtained from psychological tests, questionnaires, rating scales, and interviews. Dimensionalizing by factor analysis was possible (but time-consuming). Along with refinements in these procedures since the 1940s, promising new methods have emerged.
Applications of new methods (or new applications of old methods) have accompanied the emergence of interest in cognitive and information processes in lieu of descriptions of the behavior of leaders and followers. Since 1970, empirical examination has found a decline in behavioral and an increase in cognitive studies (Robins, Gosling, & Craik, 1999). In leadership reports, the focus is on leaders’ and followers’ schemas for actively organizing in formation and the scripts they employ to give specific meanings to situations. The making of meaning or sense making “is the process of arranging our understanding of experience so we can know what has happened … what is happening, so we can predict what will happen” (Drath & Palus, 1994, p. 2). Prototypes help interpret the world of leadership. Information about specific situations is encoded, stored in memory, and retrieved in terms of category structures and inferential strategies (Pervin, 1985). Galbraith and De Noble (2004) presented a cognitive framework for historical leadership stories.
Cognitive-perceptual methods, first developed in the 1940s and 1950s to study learning and decision-making processes, the dynamics of personality, and counseling, were subsequently introduced into the study of leadership. These methods include protocol analysis, stimulated recall, and the repertory grid. In protocol analysis, individuals’ thoughts, feelings, and emotions are recorded as they engage in the activity under investigation. With stimulated recall, an audio or video record of the activity or the protocol is played back, and the individuals describe more fully the thoughts, ideas, and feelings they had when the activity was in progress (Burgoyne & Hodgson, 1984). To construct a repertory grid, a respondent sorts colleagues, noting the categories used to differentiate them (Kelly, 1955). Coghill (1981) showed the profitability of studying the perceptual categories that managers naturally use to assess one another. He applied Kelly’s repertory grid method with 90 managers to provide data to illustrate the importance of perceptual mediation, personal constructs, and implicit theories.
The methodology of another line of cognitive inquiry has been phenomenological. The inability to resolve the controversy surrounding Fiedler’s (1967a) Contingency Theory after more than three decades of theorization and empirical research led Bar-Tal (1989) to argue that what may be most important in understanding leadership is to determine leaders’ and followers’ phenomenological fields rather than to continue to pursue a positivistic, mechanistic, or statistical explanation of their interaction. To expedite this methodology, the researcher may disclose as much about himself or herself to the focal leader as the leader does to the researcher. Basic emotional processes can then be explored, along with remote and accessible aspects of their respective lives. Barriers between the researcher and the leader to knowing and sharing are expected to be minimized in this way (Massarik, 1983).
If the focus is on perception or cognition rather than behavior, methods that exploit attribution phenomena can be applied with profit to the leader-follower relationship. As noted in earlier chapters, the behavior of leaders toward their subordinates appears to be strongly determined by the reasons (ability or motivation) the leaders give for their subordinates’ performance—how much it is a matter of luck or of the situation or whether it is under the subordinates’ control. In the same way, the subordinates’ attributions of the reasons for the leaders’ behavior will strongly relate to the subordinates’ satisfaction. Leaders whom subordinates perceive to be incompetent but willing seem to be more forgivable than leaders whom subordinates perceive to be competent but unwilling.
As already noted, conscious perceptions can determine a leader’s subsequent efforts. Thus Nebeker and Mitchell (1974) found that differences in leaders’ behavior could be explained by the leaders’ expectations that a certain style of leadership would be effective in a given situation. At the same time, subordinates’ descriptions of their leaders’ behavior may be distorted by their implicit theories about leadership, particularly when they lack real information about the situation and are therefore inclined to fall back on stereotypes (Schriesheim & deNisi, 1978). The use of cognitive, as opposed to exclusively behavioral, methods is likely to parallel the same continuing developments in psychology in general.
Still underutilized are sociometric designs, such as the “Work with” sociometry4 of Stogdill and Haase (1957), which provides a measure of interaction structures that can be applied to a variety of research designs. Three-dimensional holographic sociometry (Bradley, 1987) is in its infancy. It is likely to have a promising future as it opens up the contributions of the sociometric structure of triadic relationships within an organization to the performance of leaders.
Documents, recordings, and protocols can be analyzed from particular theoretical perspectives and coding models based on motivational and psychoanalytical theories. Keyword coding of protocols provides a powerful tool. The projective inferential methodology, pioneered by Winter (1973) to analyze the needs and values of leaders from their speeches, documents, and biographical materials, was detailed in Chapter 8. The approach is likely to provide important empirical support for and understanding of studies of charismatic and inspirational leadership. Less prominent is the inductive methodology that Demause (1982) derived from psychoanalysis. Demause argued that a group’s fantasies, shared by the leader and his or her followers, which play a crucial role in charismatic leadership, can be detected from a psychoanalytic examination of the leader’s and followers’ speeches, documents, and body language. The group’s fantasies can be teased out from the metaphors and similes that the leader and followers use, and body language, feeling tones, and emotional states can be significant. Thus the “killing” of bills and “dead” halts in negotiations are relevant to the psychohistorian, as are unusual or gratuitous word usages (in the 2004 presidential campaign, saying “the ‘L word,’” for example, implied that “liberal” was a dirty word). Unusual repetition and phrasing in discussions, speeches, documents, and minutes of meetings imply potent messages for the psychoanalyst. Symbolism has obvious implications; even its absence during long periods without imagery may be symptomatic of the severe repression of the group’s fantasy (Scheidlinger, 1980).
Increasingly, metaphors from science will be used for taxonomies, as, for example, Field (1989b) did in likening leaders to stars, pulsars, quasars, and black holes. (Stars communicate positive expectations to followers and induce positive performance. Pulsars communicate positive, but unsustainable and unrealistic, expectations. Quasars are pessimistic, but followers succeed despite them. Black holes’ pessimism is a drag on followers.)
Zaleznik (1984) reviewed a number of promising schemes for looking at the “text” of leadership—the meanings, intentions, and motives that are subconscious as well as conscious in the spoken or written words employed. The deeper structure underlying the surface interactions of the leader and the followers can be discerned from thematic interpretation. Such interpretation requires grasping a multiplicity of complex conceptual structures in the text. These structures may be cultural or social forces or individual values, meanings, and intentions. They may be seen in the linguistic study of speech patterns or of the formation of symbols. Hidden meanings and unconscious ideas and fantasies are brought to light. Clues to what is hidden in the text may include: the percentage of positive, neutral, and negative words and phrases, emotional or nonemotional; the percentage of references to the past, present, conditional, or future; the percentage of action verbs; the percentage of interruptions, hidden reversals, or recycling; and delayed responses and silences.
The formation of groups and movements can be organized around such shared fantasies and provide the basis for displacing personal inadequacies. For instance, Demause (1982) applied a psychoanalytical coding scheme to the Nixon tapes to complete a fantasy analysis of them. The analysis found that a leader’s personal embarrassments can become substitutes for policy; goals can disappear, while actions become irresistible.
Psychohistory has been the psychoanalysts’ exclusive turf. However, documentary analyses (House, 1988b) and survey methodology (Bass, Avolio, & Goodheim, 1987) are beginning to be applied to psychohistory. Indeed, if motivational analysis is to move off the psycho-analyst’s couch, it will need to take individual, political, and social psychology, as well as relevant aspects of sociology and anthropology, into account to elucidate the performance of historical figures (Strozier & Offer, 1985). In the hands of psychoanalysts, psychohistory has been limited to an understanding of collective actions from an individualistic orientation. It has focused too much on neurotic mechanisms and the inner person, whose ideas and thoughts are interpreted by an analyst who is far distant in place and time from the leader (Gay, 1985).
Recourse to more objective methods that are also projective techniques is likely to move researchers toward establishing more confidence in the findings about the hidden and unconscious motives and the implicit theories of leadership that affect interactions. For example, Boal, Hunt, and Sorenson (1988) constructed a “Leadership Quotes” questionnaire. Respondents indicated whether each of 133 famous quotations (such as “People are more easily led than driven,” “The world needs able men who can create and lead others,” and “Happy the kings whose thrones are founded on their peoples’ hearts” were related to leadership and whether the leadership was effective or ineffective. Multiple factors emerged that dealt with the direction of activities, influential interactions, legitimate power to lead, and initiation of structure.
In the future, we are likely to see more joining of motivational analyses of the biographies of world-class leaders, analyses of their speeches and writings, news accounts of their performances, and evaluations of them by historians and political scientists. Much of this joining is already to be seen, for example, in the work of House, Woycke, and Fodor (1988).
Behavioral accounts are still the most popular. Purely behavioral explanations cannot be dismissed. One can look at leadership as a perceptual phenomenon under certain conditions or as a behavioral phenomenon under other conditions and accurately explain what is happening in both instances. Research on leadership has been heavily dependent on subordinates’ reported perceptions of their leaders’ behavior. Yet Gilmore, Beehr, and Richter (1979) demonstrated, in a laboratory setting, that participants failed to perceive that their leaders (who had been instructed to be high or low in initiation and consideration) actually differed in their behavior. Nevertheless, higher-quality work by the subordinates resulted if the leaders’ behavior was actually high rather than low in initiation and consideration. The lowest-quality work occurred when the leader’s initiation of structure was high but consideration was low. It is clear that studies that do not depend solely on perceptions are still needed.
Among the promising behaviorally-oriented methods is one that shifts attention away from the leader’s frequency of behavior to the leader’s intensity of behavior. Influenced by opponent process theory, Sheridan, Kerr, and Abelson (1982) examined the intensity, rather than the frequency, of a leader’s actions. Seven dimensions of leadership behavior were scaled: task direction, participation, consideration, performance feedback, integrity, performance reward, and representation. Intensive task direction was illustrated by “taking time to instruct subordinates in proper work techniques in their jobs.” The lack of intensive task direction was illustrated by “bringing new employees into the group without providing any direction or indication of their job responsibilities.”
The use of multidimensional scaling in the study of leadership also has not been exploited as much as it could be. One example was provided by Misumi (1985), who described the use of the Quantification of Pattern Classification as a multidimensional scaling method for assigning quantitative values to categorizations of leadership and their effects. At the same time, not to be ignored are old methods that can be applied to new taxonomies of leadership behavior. For example, Van Fleet and Yukl (1986b) trained students to code more than 2,500 entries for each of 23 behavior categories as the students encountered them in biographies and autobiographies of military leaders. Clarifying work roles, setting goals, monitoring the environment, planning, and inspiration were the most frequent entries. Bass, Avolio, and Good-heim (1987) asked students to complete the Multifactor Leadership Survey questionnaire about the world-class leaders whose biographies they read. Inspirational leadership emerged as an important factor, along with other transformational factors.
Davis and Luthans (1984) demonstrated the usefulness of idiosyncratic observational efforts. They observed over time what happens when a single leader, in this case a production manager, takes a specific action: the introduction of a new scheduling form with complete instructions and deadlines. The investigators found that the new form resulted in improvements in the quality and quantity of production that could be explained from the cause-effect linkages in what they were able to observe in detail. The Leadership Observation System (LOS) was developed by Luthans and Lockwood (1984) to be used by trained participants and observers to record independently the behavior of managers using time sampling. Twelve leadership behavior categories with numerous subcategories emerged from a Delphi process that began with 100 categories of possible managerial behaviors. A typical category was planning/coordinating. Within this category were subcategories, such as setting goals and objectives. In a study of 120 managers in a number of organizations, Lockwood (1981) found that LOS had greater convergent and discriminant validity than parallel descriptions of the same managers with the LBDQ-Form XII or the Managerial Behavior Scales. Different conclusions about managers’ consideration and initiation of structure were reached when the different methods were applied.
At odds with Strong’s atheoretical approach (1984) to ethnographic analysis has been the construction of new observational procedures based on explicit theories and models. Although observational studies of leadership are not new, among the most promising current approaches are those that use observations of leadership based on a formal model of interrelated measurements. Particularly notable has been the emergence of increasingly sophisticated methodologies to observe behaviors and their linkage to theory in their construction. SYMLOG (Bales, Cohen, & Williamson, 1979) uses teams of up to five observers, each of whom is on his or her own time schedule. The observers pick out salient acts of the members of academic self-study groups. Twenty-six general behaviors of a member, such as “active, dominant, talks a lot,” and 26 value descriptions, such as “making others feel happy,” can be recorded and profiled in terms of three dimensions: friendly versus unfriendly, dominant versus submissive, and instrumentally controlled versus emotionally expressive. Members rate one another on these dimensions. Their images of one another reflect the polarization or unification of the group. The observed leadership behaviors and inferred values are examined for their decisive influence on the polarization or unification.
Even more closely tied to theory, Komaki, Zlotnick, and Jensen (1986) began with a theory-based taxonomy of supervisory behavior and then proceeded to develop appropriate observational measures. The taxonomy included three categories based on the theory of operant conditioning: antecedents, monitoring, and consequences of performance. Further classification broke down the observed supervisory behavior into own and solitary performance and work and nonwork-related behavior. Archival records could be consulted in addition to observed work samples and others’ reports and self-reports about designated actions. Satisfactory interrelated reliabilities were obtained, along with differential patterns of effective and ineffective supervisory behavior (Komaki, 1986). The results were particularly promising because they can be linked to the voluminous psychological research that deals with operant conditioning.
Martinko and Gardner (1984a) designed and utilized a model involving field observation of leaders’ external and work environments, competencies, needs, and causal attributions. Also included were categories of behavior (such as planning, organizing, directing, and rewarding) and outcomes. With the procedures and model, the time and event patterns of more effective and less effective school principals could be readily distinguished. For instance, in comparison to principals who were low performers, those who were high performers used more diverse media for communication and initiated more contacts with teachers and students. Metcalfe (1984) analyzed the behavior of leaders in taped appraisal interviews. A behavior that distinguished effective from ineffective interviewers was the greater frequency with which leaders invited participation from their followers.
Increasingly, along with the theory-based observations of behavior for research on leadership, simulations that feature interactive computer programs portraying problem situations likely will become common. Research on maximizing the utility of such programs and evaluating their impact on the performance of leaders will be likely in the decades ahead. Simulation offers the opportunity to examine perceptions, cognitions, motivations, and behaviors simultaneously. In a live field study, observers have no control over what they are observing. This fact places a premium on high-fidelity simulation as a method of inquiry. With high-fidelity simulation, the antecedent conditions and contingencies can be controlled, to a considerable degree, and time can be compressed to facilitate the observational processes and linkages to outcomes. The future is likely to see increasing reliance on simulations to provide opportunities for research on cognitive processes as well as on observed behavior.
Developed from earlier attempts with small-group exercises, in-basket tests, computerized business games, and larger organizational games, high-fidelity simulations have become a fine art that provides opportunities to manipulate variables experimentally in laboratory conditions that approach field studies. Subjects may be hired for a week to do, as far as they know, real clerical work in a seemingly real office. Computer simulations provide complete visual and auditory displays in which leader-subordinate relations can be examined. Complex business games that last for many months are conducted, in which many of the elements of real-world decisions must be dealt with satisfactorily and in which leaders may find themselves involved in labor negotiations and with real boards of directors. Complex military and political games played by real military and political leaders can both provide training and clarify the most probable effects on decisions for the real world. The future will bring increasing fidelity to the games.
As illustrated in Chapter 35, the Looking Glass is a finely developed simulation and parallel-in-miniature to managing Corning Glass. The Looking Glass re-creates for players the hierarchy of positions, necessary requirements, and interactions to running a sample organization. Salient variables can be manipulated in a collapsed time frame. Follow-up critiques of players in these simulations make it possible to gain insight into the processes and relationships perceived to be of consequence to the players (McCall & Lombardo, 1982). Videotaping can store the experience for later intensive analysis.
Mathematical modeling is another approach to simulation that has a great deal of unexploited potential for leadership research. Zahn and Wolf (1981) used the Markov model to chain events that related the current hypothetical task-oriented or relations-oriented performance of supervisors to future states of supervisory-subordinate relationships. The model could be manipulated to see designed hypothetical effects on outcomes. Long-term behavior was determined to be highly variable and versatile. Both the leader and the subordinates could influence the system.
What if? is a thought experiment or a mental simulation of possible events at turning points of history and its leaders. It is counterfactual. What if Lee Harvey Oswald’s bullet had missed John F. Kennedy in 1963? Would Lyndon Johnson’s reforms have been delayed? Would the United States still be entangled in Vietnam? What if the Germans had not arranged in 1917 for Lenin to go by train from his exile in Switzerland through Germany and German-occupied land, to emerge in St. Petersburg? Would the Bolsheviks have taken over from the Democratic Kerensky regime for the next 70 years following the October Revolution? Would the monarchy have been restored? Would fascism have taken over, so that Russia would have been allied with Germany, Italy, and Japan in World War II? How different would world history have been if the Supreme Court’s 5–4 decision had not terminated the reexamination of the Florida balloting and given the 2000 presidential election to the less experienced, conservative Republican George W. Bush instead of to the more experienced Al Gore, a liberal Democrat who had won the popular vote but not the electoral college vote?
No one approach is fully adequate, by itself, to understand the leadership process. Yukl (1982) pointed out that cognitive processes, for instance, are unlikely to be reliably detected from observation alone. Yukl and Van Fleet (1982) found, when analyzing four studies to identify effective patterns of leadership behavior among military cadets and U.S. Air Force officers, that different results materialize when nonquestionnaire methods, such as the critical incidents technique, are employed than when questionnaires are used in the same situation. They noted that leadership consideration was important to effective leadership in the critical incidents analyses but not in the questionnaire surveys, whereas coordination was important to effective leadership in the questionnaire surveys but not in the critical incidents studies.
There will continue to be proponents of one method over another. For instance, there will continue to be proponents of a cognitive-perceptual approach and advocates of a behavioral approach to understanding. Still others will emphasize social factors. Hooijberg, Hunt, and Dodge (1997) suggested that a full account of leader-outcome results would require the use of more complex approaches—cognitive, behavioral and social—for comprehensive explanations of leadership, followership, and their interaction.
In a discipline, consensus about a new paradigm usually develops slowly. Nevertheless, a number of new substantive issues quickly aroused attention in the last quarter of the twentieth century and the twenty-first century. For example, charismatic and transformational leadership can be studied in small groups and in the histories of world-class leaders with the same theoretical framework. Historical studies of leaders in the past were stories about prominent figures who had influenced events (Luecke, 1994). Relatively recent and continuing cultural, behavioral, social, and/or political science explanations of institutions, followers, leaders, and events are in force. More than just narratives of the past, they are stories about elites and ordinary people, individual choices and collective experiences, unusual and normal events, and social, political, educational, and economic problems. Historians provide descriptive statistics and maps to buttress their propositions. In the future, some may analytically test the validity of hypotheses about qualitative and quantitative changes linked to the leaders (Fischer, 1989). Illustrative is the investigation of the constructiveness and destructiveness of charismatic leaders by psychologists (O’Connor, Mumford, Clifton, Gessner & Connelly, 1995). The centuries-old person-versus-situation controversy is likely to remain with us, with more attention in the future to genetics and hormonal effects. New co-twin studies are to be expected of the heritability of leadership behavior. Hormonal levels will also be considered. Thus Dabbs (2000) suggested that testosterone level is a biological aspect of temperament. Coupled with a proclivity for action, this factor has predicted firefighting and emergency medical performance. It may predict a leader’s power, motivation, dominance, boldness, focused attention, frame of mind, and leadership style, along with the emergence of violent or conciliatory behavior. Hormonal levels may become increasingly important as women take more positions of leadership.
Jago (1982) believed that much more needed to be done to compare the validity of universal and contingent theories of leadership. Since then, mounting evidence has supported the point of view that both personal factors and contingencies are involved. How much of each is important is an empirical question for continuing research (Bass, 1997).
There has been a renewal of interest in the personal factors of leadership. Although research designed solely to isolate the characteristics of leaders was thought to have reached the point of diminishing returns, it has been revived in a new form with the emergence of interest in the life-span development of leaders, particularly transformational leaders. The need to learn more about what contributes to the self-confidence, self-determination, and freedom from inner conflict of the charismatic leader is apparent. The quest for a greater understanding of personality and the personality dynamics that affect leaders’ performance transcends situational considerations. Programmatic efforts to look at personality, as such, are also suggested by the developments of assessment centers, as well as long-term predictions of success as a manager. Increasingly, character is seen as of prime importance to leadership.
Paige (1977) proposed a hypothesis that still awaits empirical confirmation: leaders choose tasks and engage in them on the basis of personal considerations, such as their age. Thus, what could be a tempting challenge to a 40-year-old executive might be sidestepped by his or her 60-year-old counterpart. More often than not, even when there is a lack of supporting evidence, personality is seen as interacting with situational variables to account for leadership and group performance. It may be more important in some situations than others and with some people more than others. It may be dominant in cultural settings where, to be a leader, one must epitomize the authoritarianism rooted in the culture. Autocrats may behave the way they do because of their personalities; democrats may behave the way they do as a consequence of both their personalities and other immediate situational and personal considerations (Farrow & Bass, 1977).
Research that is designed to test the effects on the group of the interaction of the characteristics of the leader and the followers has generally been effective in producing valuable insights into leader-follower relations. Research in the 1950s and 1960s on different combinations of authoritarian leaders and followers continues to suggest designs that could be used to test the effects of other leader-follower characteristics. Although much has been done with reference to the need for achievement, the need for affiliation, internal and external loss of control, and so on, many other personality dimensions can be examined in the same way to test the interaction of the leader’s personality, values, and behaviors with the followers’ personalities, values, and behaviors and the effect of such interaction on the group. Interest is strong in examining how the leader’s style of thinking, as measured by the Myers-Briggs Type Inventory, interacts with the followers’ styles. To be kept in mind is the likelihood that extreme homogeneity in leader-follower characteristics may be dysfunctional to satisfactory problem solving by the group that requires flexibility and creativity. With increased interest in charismatic-transformational leadership, time orientation of leaders to the past, present, or future is an underresearched issue likely to be examined more frequently in the future (Daltrey & Langer, 1984).
Suggestions continue to be made that research on leadership—in addition to its narrowness—has been concentrating on the wrong thing. The aim of science is to understand. Understanding is checked by prediction. Adequate prediction can produce control. Practitioners are anxious to provide such control when their understanding is far from perfect. Research on leadership is faulted when it fails to improve such control immediately. Basic research should be judged on whether it contributes to understanding.
The failure of adequate prediction to make the suitable control of behavior possible may also be partly a consequence of the unpredictable and uncontrollable elements in the real-world performance of leaders. Mintzberg (1973) and many others since have concluded that instead of a systematic, steady, orderly attack on one problem at a time, the practicing manager is more likely to be observed devoting short bursts of time to different problems. Frequently interrupted, the manager unsystematically responds to a diversity of demands from superiors, peers, clients, and subordinates. (This behavior helps explain the popularity of management training in time management.) Given the large array of diverse situations that are the daily regimen of the general manager, M. W. McCall (1977) suggested that much of the research to determine which type of leadership style is most effective in particular conditions may remain an impractical academic exercise because the demand characteristics of the manager’s role result in the manager optimizing his or her performance by “proficient superficiality” (Mintzberg, 1973). Insufficient time is given to different problems to deal with each of them fully and adequately.
A promising approach looks at how the leader’s identity, prototypicality, and representativeness of the team, the followers’ social identification with the team, and team members’ desires for closure to reduce uncertainty contribute to various criteria of satisfaction and effectiveness. Pierro, Cicero, Bonaiuto, et al. (2005) surveyed 242 manager professionals and white-collar and blue-collar employee members of three Italian firms in petrochemicals, manufacturing, and electronics. Leader prototypicality correlated .41 with perceived leader effectiveness and .20 with job satisfaction. Intentions to quit were high when leader prototypicality was low, but only when members were also high in preference for closure. The members’ self-rated performance was high when the leader’s prototypicality was high, but only when members were high in preference for closure.
Innovation and Creativity. Mumford, Connelly, and Gaddis (2003) concluded from a set of experiments that for leaders to promote creativity and innovation, they must possess creative thinking skills. These may provide a critical perspective in evaluative operations in collaboration with followers. More needs to be learned about the leader’s contribution to the followers’ creativity. The leader’s competence in this regard may depend on: (1) his or her awareness of temporal complexity in a “knowledge-rich,” hypercompetitive technological environment (Halbesleben, Novicevic, Harvey, et al., 2003); (2) whether the leader accepts, challenges, or synthesizes different ways of doing things (Sternberg, Kaufman, & Pretz, 2003); (3) how transformational leadership contributes to innovation (Jung, Chow, & Wu, 2003); (4) how the transformational leader’s unconventional behavior can add to the group’s creativity (Jaussi & Dionne, 2003); (5) how transformational leadership and members’ anonymity can make a difference (Kahai, Sosik, & Avolio, 2003); (6) the leader’s emotional intelligence (Zhou & George, 2003).
We need to learn more about why, when jobs are routine, extrinsic rewards contribute to the innovativeness of cognitively adaptive followers but, when jobs are complex, the results of extrinsic rewards may be negative (Baer, Oldham, and Cummings, 2003). Mumford and Licuan (2004) concluded that further research is needed on: (1) how leaders choose the performance strategies that influence the innovations they are willing to support; (2) how leaders establish a creative group identity; (3) how leaders need to understand the importance of emotions in the leadership of creativity in the groups they lead. Research is also needed that distinguishes: (1) leading idea generation and idea implementation; (2) leadership of routine groups and leadership of creative groups; (3) leadership of groups working on significant innovations being implemented. We need to learn the cognitive requirements that contribute to a follower’s creativity (Reiter & Illies, 2004). Further evidence needs to be found about how a leader’s consultation increases followers’ creativity and how clarification of roles and objectives affects it (Amabile, Schatzel, Moneta, et al., 2004). Such clarity provides clear objectives, commitment to excellence, and support for innovation (West, Borrill, Dawson, et al., 2003). Research is needed on: (1) how to be participative and delegative leaders and yet be able to impart expert knowledge to promote subordinates’ creativity (Krause, 2004); (2) about what contextual features in the leader-subordinate relationship promote followers’ creativity (Shalley & Gilson, 2004); (3) how team members with different preferences in the creative process can be brought to think together innovatively (Basadur, 2004). Further studies are also required of the champions of innovation, how they informally influence idea generation in the organization and pursue a flexible role orientation. Cognitive mapping of their mental models is needed (Howell & Boies, 2004).
Leader-Member Exchange. In addition to seeing the leader’s role as less orderly, Graen (1976) shifted attention away from the unilateral and consistent group-oriented behavior of the leader toward a focus on the mutual role relationships of the leader with each different member of the group. The quality of this reciprocal relationship contains considerable explanatory value, and researchers should continue to explore its antecedents and consequences with methods described earlier, to ascertain its importance relative to average leader-group effects. Nonetheless, Dienesch and Liden (1986) argued that to be more fully utilized, leader-Member exchange needs to be conceived of as a multidimensional variable involving perceived contributions, loyalty, and affect. Research is needed on the different ways in which the leader-Member relationship varies, and the different outcomes that will occur as a consequence.
Various personal characteristics—rates of talking and interaction, capacity to interact, ability to perform a task, dominance, exclusive possession of information, initiation of spontaneity, provision of freedom to the group, and acceptance of group members—have been found to be associated with emergence as a leader. It has not been demonstrated, however, that these are the same variables that enable the leader to retain his or her role. There is a need for research that isolates the factors that facilitate retention of a position of leadership once it has been attained. There is an abundance of evidence about factors that contribute to the emergence of a leader; there is a dearth of research on factors that enable a leader to consolidate his or her position once it has been obtained. In the same way, the characteristics required to win election to political office may be quite different from those required for success and effectiveness as an officeholder. An important problem that has been ignored concerns the effects of training on the retention of the leadership role. What kind of training strengthens or weakens an individual’s chances of retaining the leadership position?
Issues of Succession. Increasing empirical attention to the replacement and succession problem would seem to have great utility. There are many issues to be explored in this regard. Nathan (1989) concluded that whether a succession had a positive or negative effect or no effect on an organization depended on the study’s method, the criterion of effectiveness, and the level of analysis. Continuing good organizational performance will slow successions; continuing poor organizational performance will speed up successions. Several questions needing future research include: What are the optimum rates of succession for different types of groups, organizations, or institutions? How is the succession used as a tool of strategic change? How do successors cope best with initial resistance? Do the dynamics of the succession process change over time? Is the cost of accelerating the recruitment and compensation of successor CEOs justified? (In 1965, the compensation of CEOs of major U.S. corporations averaged 24 times that of their average worker. The ratio was 38 times in 1978, 71 times in 1989, and 262 times in 2005–$10.9 million for the average CEO compared to $41,861 for the average worker.)
Many other questions about the succession process can also be posed: Does a successor executive ignore the former occupant’s performance in an executive position, see it competitively, or attempt to build on it? Should the former occupant be involved in the choice of the successor? Does the former occupant help or hinder the succession process? When occupants are “lame ducks” and must give up office in a designated amount of time, how does the time limit shape their objectives, planning, and power? How does the loss of institutional knowledge affect legislatures whose members cannot run for reelection after two or three terms?
Types of Successors. Sonnenfeld (1988) was able to interview 50 CEOs and their senior management about the CEOs’ retirement experience and succession plans—interviews that were difficult to obtain. Four types of CEOs were identified: (1) Monarchs did not leave office willingly but were overthrown. (2) Generals left office reluctantly but instituted campaigns to return to office. (3) Ambassadors remained active in the firm after leaving office and supported their successors; (4) Governors left the firm cleanly and cut all formal contact with it. Each type had advantages and disadvantages. The monarch had brilliant visions, assumed personal responsibility for problems, created order in the face of environmental disturbances, and was completely dedicated. The firm grew in assets, sales, employees, and shareholder support. At the same time, the monarch stubbornly defended old strategies and was reluctant to develop the next generation of leadership. The general was capable of building strong leaders but was cautious about leadership transitions and was ready to return to office if needed. But the general fueled internal rivalries, encouraged resistance to successors, and even undermined them. The ambassador provided long service, continuity of command, and wisdom as an elder statesman. As a member of many outside boards, with wide contacts and community interests, the ambassador remained available for assignment as an external representative of the firm and was familiar with other approaches to management. But his continued presence became awkward, he supported inappropriate fads, and he needed to defend his record. He wanted to remain available to give advice that was unwanted. The governor was in office for only a short time. His organization was a large, stable, and formal bureaucracy. He had a wide range of outside business interests, which may have distracted him from his responsibilities of office. The firm’s performance lagged. He gave his successor little help except for the freedom to revamp his strategy. When he retired, he completely cut connections with his organization. Many of Sonnenfeld’s qualitative conclusions and typology can form the basis of future quantitative studies of executive retirement and succession.
Those who tend to conceive leadership as nothing but a form of social power obscure important relationships between leadership and power and restrict the range of research on the problem. More research is needed on the interaction of power with leadership behavior and personality. Questions remain: (1) Will followers respond positively to a leader with power who exhibits task-oriented rather than relations-oriented behavior? (2) How will the personality of the followers and the urgency to complete the task affect their response? (3) Are there patterns of a leader’s personality that may mitigate the adverse effects of coercive power? (4) Can a sense of humor exert a strong moderating effect? (5) Will coercive power and strong control be more readily accepted when the leader’s and followers’ values are highly similar than when they are not? (6) What factors tend to legitimate different forms of power among members of a group? (7) What factors tend to legitimate different forms of power in the eyes of observers who are not members of the group? (8) What contributes to a leader’s continuing to be held in high regard despite his making one mistake after another? (9) How do presidents and television evangelists maintain their power and charisma regardless of the quality of their performance? (10) What are the limits to the possibilities of “damage control”? (11) When does reward power become adverse in its effects?
Power Equalization. Research on the equalization of power has been attended by serious difficulties. Appropriate control organizations are seldom available for comparative purposes when some form of power equalization is introduced into an organization. External forces are almost impossible to control. The effects of social, political, and economic factors may be stronger than any variations that can be introduced. Simulated organizations may provide a better setting for the study of the equalization of power in the large organization.
Conflict of Roles and Ideas. Research on role conflict, conducted primarily in formal organizations, is deficient in experimental controls. Here again, it would be advantageous to study role conflict in simulated organizations in which conflict-inducing demands can be varied and controlled. Research is needed to determine the extent to which various styles of the personality and behavior of leaders are subject to role conflict. We need to know more about how controversy—conflicts of ideas—can result in creative outcomes. Tjosvold (1985d) pointed to the importance of goal interdependence, confirmation, and collaborative influence in this regard. These factors would be relevant to strategic decision making, policy setting, and participative leadership. A small number of competently executed studies are available on the legitimization of the leadership role. More studies are needed. Little is yet known about what makes a leader legitimate and why and what the effects of such legitimacy are. The importance of the subject merits a much higher rate of activity in this area. In future research, it would be desirable to determine the effect of variations in the followers’ characteristics, as well as variations in a leader’s characteristics. One would like to know, for example, for what types of follower a given characteristic or pattern of behavior tends to legitimate the leadership role. In addition, it would be useful to know under what conditions a given pattern of leadership behavior is regarded as legitimate or illegitimate and by what type of follower. From a practical point of view, the problem of the legitimization of leaders is one that should be given high priority in future research. With the expected continued increase in knowledge work, the formal leader of a knowledge work group needs to share leadership. Leadership in knowledge work done in teams should rotate to the worker with the knowledge, skills, and abilities for the work at hand. Leadership will rotate when knowledge tasks are highly complex and interdependent and when much creativity is required (Pearce, 2004). Much is still to be learned about how superleaders (see Chapter 20) share their power by converting and legitimatizing others to lead themselves (Goel, Manz, Neck, et al., 1995). They convert leaders into self-leaders. The suppressive effects of power on followers suggest the need for more subtle approaches to data collection than direct questioning. We should also be seeing more multiple methods examining the same power and leadership issues. Authority and responsibility, linked to power but conceptually different, could be examined more fully in their distribution as sources of organizational pathology.
Conflict Management. What should be done for managers who want to be participative but have poor skills in conflict management? Are there high costs for such managers who bring their subordinates together to resolve conflicts among them? (Crouch & Yetton, 1987). Burgess, Salas, Cannon-Bowers, et al. (1992) prepared training guidelines for teams under stress. To correct errors leaders needed to learn transactional practices such as contingent rewarding and management by exception (MBE). MBE included troubleshooting and double-checking of team members’ performance focused on the immediate task. Transformational leadership was suggested for handling future crises, understanding the mission, aligning individual and team goals, providing support, and establishing trust and cooperation.
Empirical studies are needed of the mathematical approach that Brams (1993) developed for resolving conflicts. Opportunities should be expanded for learning more about crisis management based on the increased willingness of decision makers to air their mistakes. The disclosure by U.S. and Soviet decision makers in 1989, 27 years after the Cuban missile crisis in 1962, of how each side misread the other’s intentions and almost jointly precipitated World War III is illustrative (Keller, 1989). Likewise, Whyte’s (1989) addition to the factors that underlie “groupthink” may further elucidate events when leadership fails during a crisis. Whyte noted the tendency for a group to polarize around the point of view initially dominant in the group. This tendency usually frames the leader’s and group’s decision as a choice between two or more unattractive options. Threats are not seen as possible opportunities. Risk seeking occurs as a way of gambling for a reasonable chance to avoid the certain losses. Much still has to be learned about the U.S. leadership’s initiating a preemptive war against Iraq, a secular state, when it was fundamentalist religious Muslims in Saudi Arabia who were mainly responsible for the 2001 terrorist attacks on the United States. How much was due to the groupthink of the neoconservative advisers in the Bush administration? How much was due to the false intelligence that the Iraqis had weapons of mass destruction? How much did Saddam Hussain’s game-playing behavior about inspections contribute to the sense of threat from Iraq? How much was due to failures, secrecy, and poor communications among the 15 U.S. security agencies? How much did the excessive layers of bureaucracy affect decision making? How much did the destruction of the World Trade Center on September 11, 2001, and the desire for retaliation contribute (The 9/11 Commission Report, 2004)? The mishandling of the Hurricane Katrina disaster in 2005 and its flooding of New Orleans and the Gulf Coast will be the subject of studies for years to come.
In the business world we need to learn more about the human costs and benefits of acquisitions and mergers. A great many fail to achieve the expected benefits from the combining of firms. Failures are estimated to be as high as 80% due to choice of partners, lack of clarity of strategy, incompatibility of organizational cultures, and inadequate preparation of their organizations’ members for the event (Marks & Mirvis, 2000).
To some degree, all research on leadership styles prior to Burns’s (1978) introduction of the concepts of transformational versus transactional leadership could be conceived of as being about democratic, autocratic, or laissez-faire leadership, which takes us back to where it all began, in 1938 with Lewin and Lippitt’s seminal experiment. Each of these three styles is described by either the amount of overall activity of the leaders or the extent to which the leaders are oriented toward being completely work-oriented or completely person-oriented.
Ordinarily, leaders’ performance will be better if they are more active than inactive. But activity does not guarantee effective, satisfied, and cohesive groups. A number of hypotheses need to be tested further. For instance, it has been suggested that leaders tend to be more autocratic and directive when it is easier and more comfortable for them to be so because of their own personalities and because goals are clear and structure is given. But it has also been proposed that leaders really need to be directive and structuring when goals are not clear and structure is not given. Another suggestion is that leaders tend to be more democratic, participative, relations-oriented, and considerate when they are concerned about the need for input and reactions from subordinates. Again, how different personalities and situational demands play their parts can be tested in designated organizational settings to determine the importance of selection in contrast to training of leaders.
Dimensions. Many dimensions of autocratic and democratic leadership have been proposed to describe leaders’ different styles. The relationships among some of the dimensions have been clarified and altered. Borrowing from work motivation theory, researchers have purged the initiation-of-structure scales of their coercive components and have relabeled them instrumental (to the success of the task). Similarly, consideration is now seen as supportive leadership, and its focus is now on leaders’ attention to the different needs of individual subordinates rather than to the use of group methods.
The situational leadership approach, as in the Hersey-Blanchard model, and an approach to the best way of integrating task and relations orientation, as advocated by Blake and Mouton, need to be compared with a variety of methods and measurement instruments. Studies ought to be conducted that make the comparisons in simulated organizations and in controlled laboratory settings, as well as in content analyses of documents and time-sampling observations in the field.
It should be clear that factor analyses can establish only how leaders are distributed empirically in the eyes of observers of their performance on given dimensions, such as directive, negotiative, consultative, participative, and delegative.5 However, leaders in real-life positions tend to reveal performance on all the dimensions but in different amounts, although the dimensions are conceptually distinct but not empirically uncorrelated.
One can stress the utility of consultation (the manager’s most frequent style) as a useful style in general. But, as Vroom (1976b) noted, concrete situations may demand otherwise. And leaders do change their styles in response to situational conditions (Hill & Hughes, 1974).
The differential effects of task-related and person-related clusters of leadership styles suggest that two central needs are: a response allocation analysis, to sort out the different styles from one another conceptually; and more factorial analyses to determine the empirical communalities among the styles. Researchers need to take into account the well-established negative impact on the group of the inactivity of a laissez-faire leader with the possibly but not necessarily positive impact of various types of active intervention by the leader.
Theory and research are needed about the various kinds of considerate behavior. Consideration, as measured by the LBDQ, tends to be noncontingent, unlike the contingent rewarding of the MLQ. For instance, LBDQ items about how much the leader helps and does favors for the group do not refer to the group’s performance. Nor does LBDQ consideration correlate highly with the individualized consideration factor score of the MLQ (Seltzer & Bass, 1990), which tends to concentrate on how much the leader is concerned about followers’ differential development needs.
Before the 1980s, behavioral research on leadership concentrated on the transactional exchange between the leader and the led. The leader clarified what needed to be done and the benefits of compliance to the self-interests of the followers. In the new paradigm, the transformational leader moves followers to transcend their own interests for the good of the group, organization, or society (Burns, 1978; Bass, 1985a). Paraphrasing Zaleznik (1977), transformational leaders, like charismatics, attract strong feelings of identification from their subordinates.
The 1980s were a decade in which empirical research was initiated with this new paradigm in mind. As noted in Chapter 25, a number of survey questionnaires, such as the Multifactor Leadership Questionnaire (MLQ), were used extensively in a variety of organizations and countries (Bass, 1997). In the twenty-first century, the work has expanded into new areas, such as the empirical connection of transformational leadership with corporate social responsibility (Waldman, Siegel, & Javidan, 2004).6 Much more still needs to be done, especially in teasing out the dynamics involved. Some empirical support has been obtained of followers’ strong beliefs in the leader as a person. Do these beliefs go beyond the leader’s actual competence? Followers have faith that the leader will make it possible for the group to succeed. Are they justified in their faith? Do the followers willingly give the leader too much power to act in crises? Do such leaders transform followers into leaders? Podsakoff, MacKenzie, Moorman, and Fetter (1990) have shown the importance of trust as a variable intervening between transformational leadership and outcomes in performance and satisfaction.
Generally, transformational leadership augments transactional leadership (Waldman, Bass, & Yammarino, 1988). But under what conditions do they conflict? How can one add the fostering of the pursuit of group and organizational goals (in transformational leadership) to the promotion of self-interest (in transactional leadership)? Much more research lies ahead about the components of transformational and transactional leadership: charisma and inspiration, intellectual stimulation, individualized consideration, contingent rewarding, and active management by exception (which includes contingent but not noncontingent punishment).
Transformational leadership theory has now gone from being novel and interesting to initial acceptance; to critiques of what is missing … a much deeper investigation [is needed] of what constitutes the constructs of transformational leadership and how it can be measured, developed, and projected through technology in virtual teams and organizations. (Avolio & Yammarino, 2002, p. 388)
Charisma and Inspirational Leadership. Leaders of political, social, and labor and student movements as well as managers, ministers, battalion commanders, teachers, coaches, and directors can be found who fit the description of persons to whom followers form deep emotional attachments and who, in turn, inspire their followers to go beyond their own interests for superordinate goals, higher goals than the followers previously recognized. Even in hardened bureaucracies, there are leaders with knowledge of the system, good connections, and the ability to mobilize and husband resources. Such leaders keep their eyes on the bigger issues and take the risks required for “creative administration,” which gives them the large amount of idiosyncrasy credit necessary to arouse in subordinates complete faith and trust in them and a willingness to strive for the higher goals they set forth as challenges for the group.
To understand the high expectations of success generated by the charismatic leader, we can look at Field and Van Seters’ (1989) proposal to combine the self-altering prophecies of the Pygmalion and Galatea effects to generate an energized group, organization, or society of high expectations in which members act with positive expectations about one another. This group is labeled the “Metharme effect” by Field (1989), after the daughter of Pygmalion and Galatea (Field & Van Seters, 1989). House and Shamir (1993) and Shamir, House, and Arthur (1993) began a theoretical explanation of the effects of transactional and transformational leadership. Transactional leaders focus on pragmatic paths and goals that can be reached by followers. Transformational leaders produce in their followers a greater sense of: (1) collective identity in their self-concepts; (2) consistency between their self-concepts and their actions for the leader; (3) self-worth and self-esteem; (4) similarity between their self-concept and perception of the leader; (5) collective efficacy; (6) meaning in their work and lives. Charismatic/transformational leaders arouse in followers unconscious motives of achievement, power, and affiliation.
Idealized Influence. Handy (1994) noted that Maslow’s (1954) hierarchy of needs from safety and security to self-realization failed to include a need for idealization, a higher need than self-realization, not as self-centered as the rest of the hierarchy. Bass and Avolio (1989–2003) began to substitute idealized influence for charismatic leadership because of charisma’s popular association with flamboyance and malevolent dictators. Idealized influence was thought, like inspirational leadership, to correlate highly with charismatic leadership. For Handy (1994, p. 275), Maslow’s hierarchy was all self-centered. At a higher level was idealization, “the pursuit of an ideal or a cause that is more than oneself.” Handy’s concept may relate better to Howell and House’s (1992) concept of socialized charisma than to personalized charisma. Is this also true for idealized influence?
Collective Motivation. An interesting research issue is raised by the question of how collective motivation is promoted in organizations. Collective motivation on the part of employees was seen by Lawler (1982) to be a matter of outcomes that they valued occurring from the performance of the organization and their own contributions to that performance. But Staw (1984) argued that collective motivation depends on employees’ identification with the organization and their contributions to the organization’s performance. Transformational leadership theory would go further and imply that collective motivation is generated by the employees’ identification with their leader and the employees’ and their leader’s identification with the organization and its goals and values. This requires fuller testing empirically.
The Importance of Identification. House, Howell, & Shamir (2002) use the identification of followers with a transformational/charismatic leader to explain why the followers are willing to forgo their own self-interests for the collective mission. The theory they present raises many questions for the future, including the evolutionary processes that have made social identification such a strong force. In what ways is identification with a charismatic leader and his or her organization important to members’ decision making, sense making, satisfaction, and commitment? Why do strong emotions accompany identification with the charismatic leader and mission? How and when does identification cause conformity, groupthink, blind obedience, and abandonment of rationality?
The themes of neo-Weberian charismatic leadership research that are likely to continue are: (1) charisma in everyday life of work, care, friendship, and leisure; (2) focus on the relationships between the leaders and the led, with special emphasis on emotional interactions; (3) a determination of the extent to which the charismatic leader or the charismatic mission drives followers; (4) an explanation of the charismatic mission is institutionalized, routinized, and maintained in the succession (Jermier, 1993). Additionally, research is needed on: (5) how the charismatic leader influences followers’ self-concept, self-efficacy, and social identification (Yukl, 1994); (6) how much motivational impetus and routinization in the mission is lost in a succession when the charismatic is replaced, departs, or dies; (7) how much the routinization of the charismatic mission impedes needed organizational changed to facilitate adaptation to the changed environment; (8) how much is lost in the succession without the driving force of the departed charismatic leader (Bryman, 1993).
Intellectual Stimulation. As noted earlier, innovation and creativity often depend on leadership (Mumford & Licuan, 2004). Mueller (1980) described the “leading-edge” leader who deals with “fuzzy futures.” This type of leader is able to simplify problems and to get to the crux of complex matters while the rest of the crowd is still trying to identify the problem.7 Research is needed on this “rapid reification” and how to integrate and relate intellectual stimulation with the logical and intuitive attributes of leadership. With the continuing importance of technological innovation in the twenty-first century, we will see more research on the subject, such as that by Berson, Dionne, and Jaussi (2002) connecting task, technology, people, and innovative processes both quantitatively and qualitatively.
Individualized Consideration. Future questions to be answered are: How do leaders maintain their recognition for fairness when they treat each of their followers individually? How much do the effects of individualized consideration using the MLQ correlate with the effects of consideration using the LBDQ? When is an individually considerate leader seen by followers as weak? How different is individualized consideration from relations-oriented leadership? Can participation proceed without individualized consideration?
Management by Exception: Contingent Reward and Punishment. Podsakoff and Schriesheim (1985) enumerated a number of unanswered questions about a leader’s use of contingent reinforcement for which research is needed. What is the best sequencing of contingent and noncontingent rewards and punishments (when necessary)? How do they interact? Can contingent punishment for poor performance work if it is offset by promises of rewards for appropriate performance? What rewards make promises more or less effective? How important are specificity and focus on the subordinate’s performance or products to the efficacy of contingent praise for good performance (as Kanouse, Gumpert, and Canavan-Gumpert [1984] suggested)? How does contingent reinforcement of the individual affect the group? How does contingent reinforcement of the group affect the individual? Further exploration is needed about the conditions when active management by exception is successful and/or effective and when it is not. Is there evidence of its changing in practice when followers are novices versus when they are experts?
The Hierarchy of Effectiveness. A hierarchy of leader-outcome relationships based on survey questionnaire methods and routines has been established in a number of industrial, educational, and military settings. Transformational leadership, particularly charismatic leadership, correlates above .70 with subordinates’ satisfaction with the leader and the leader’s perceived effectiveness. The transactional contingent reward correlates closer to .40 with these same outcomes in satisfaction, as can the transformational aspect of contingent rewarding of psychological benefits (Antonakis, 2000). Transactional passive management by exception correlates closer to zero with these outcomes (Yammarino & Bass, 1989). Studies of the more conceptually differentiated personal communication style of the transformational leader, along with his or her mastery of impression management and authenticity, will be important. Will Handy’s (1994) idealization, the pursuit of ideals, help or hinder effectiveness? Researchers need to examine the conditions that promote the emergence of a transformational leader and how to facilitate this emergence. They need to determine the consequences of moving from an emphasis on social exchange, which characterizes the transactional leader, to leadership that mobilizes and directs members toward higher objectives. They need to study the costs and obvious dangers, as well as the benefits, of transformational leadership. A step in this direction will be to examine further the communication differences between self-oriented and prosocial leaders.
Handling Stress and Conflict. Investigators should be searching for how a transformational or transactional leader moves followers toward acceptance of super-ordinate goals. How can followers be aroused to self-transcendence? How can a leader move a group from complacency, hasty responses, inertia, or defensiveness in the face of threats to complete and adequate vigilance? If a group is focused primarily on its lower-level needs for safety and security, how can a leader move it toward concern for recognition and achievement? If a group is under stress that is too high for coping with the complexity of the situation, how can the leader steady and calm the group?
Micro-Macro Linkage. Hollander (1985) remained dissatisfied with the overconcentration of research on the microleadership of small groups. More macroleadership studies are needed of how leaders affect the strategic functions of institutions and organizations. For instance, Khan (2002) reported on the transformational and transactional leadership of the executives in 157 sponsoring high-tech Japanese-U.S. alliances, finding that, as predicted, transformational leadership by both the Japanese and American executives resulted in lower innovative effectiveness in the alliance.
The transformational-transactional distinction offers an avenue for tying together research on microleadership and macroleadership. Gordon and Yukl (2004) see the need to link microleadership with strategic leadership. As noted in Chapter 12, transformational leadership—charismatic, intellectually stimulating, inspirational leadership—can be manipulated in the laboratory, reliably observed in small groups playing business games, rated reliably in large organizations at all levels, and used to differentiate among world-class leaders. At the same time, the paradigm lends itself to investigations across the disciplines of political science, organizational science, anthropology, sociology, and psychology.
Selection and Training. We need to be able to train the average supervisor in the sensitivities and interpersonal competencies that are required of a transformational leader. We need to be able to select potential transformational leaders who may not show up well on many currently available predictors of transactional leadership. It may be less difficult to enable leaders to learn how and when to be constructive with contingent rewards or corrective with management by exception. Self-reports, incidents, and 360-degree ratings from 200 executives and 500 leaders in one community were collected during and after their participation in Full Range Leadership Development programs, indicating modest improvements by the trainees in transformational leadership and reduced management by exception. Similar results were found for U.S. Army platoon sergeants and lieutenants from 72 platoons rated by their more than 3,000 subordinates, peers, and superiors (Avolio & Bass, 1996).
Contraindications. Theory and research are needed on the costs and dangers of transformational leaders. Thus Bennis (1989) pointed to the destructive consequences of considering only such successful transformational leaders as Steve Jobs and Lee Iacocca as heroes. Lippman-Blumen (2005, p. 220) warned against heroes with noble visions who are actually toxic for us. We need to “escape the magnetic pull of subtle toxic leaders … heroes [who] … attract us with challenges and visions that entrap us before we perceive their negative potential, … their dark call to battle, genocide, and unjust war.” The toxic leader is pseudotransformational. Toxic leaders consciously feed followers with illusions that enhance the leaders’ power and impair followers’ capacity to act independently. They play to followers’ basic fears and needs. They mislead followers through deliberate untruths. They treat their own followers well but persuade them to hate others. They are amoral and lack integrity. Their egos and arrogance blind them to their own faults that limit the possibilities of their self-improvement (Lipman-Blumen, 2005, pp. 20–21). Much needs to be learned about the moral development of the toxic pseudotransformational leader when the charismatic leader is a false messiah.
Bass (1985a) speculated on organizational and societal circumstances that might militate against the success and effectiveness of transformational leaders. Spangler and Braiotta (1990) located such an example. Chairmen of audit committees of boards of directors need to be more concerned about good monitoring, auditing, and controlling. For them, transactional leadership may be as effective as transformational leadership. Contingent reward and active management by exception correlated .43 and .39, respectively, with effectiveness judged by colleagues in accounting and auditing inside and outside the firm. Parallel correlations with effectiveness of individualized consideration, intellectual stimulation, and inspirational leadership were .40, .30, and .25 respectively.
Questions for the future: What makes laissez-faire leadership ineffective and empowerment effective? When is laissez-faire leadership appropriate and effective?
Political leadership needs to be seen in a positive light as well as a Machiavellian one. Although the evidence (possibly as a consequence of social desirability) points to a general rejection of Machiavellian approaches (with-holding information, bluffing, agreeing but delaying, and maintaining social distance), organizational and social decisions can often be understood as a consequence of coalition formation, negotiations, and other beneficial political processes. Decisions are based on the relative power of those involved rather than the merits of the issue.8 Ammeter, Douglass, Gardner, et al. (2004) agreed that since organizations are an arena for politics (Mintzberg, 1983), they can be beneficial as well as harmful. They can be altruistic as well as self-serving. Ammeter et al. (2002) theorized that the effects of political leadership in organizations were to be studied, not deprecated Like Cyert & March (1963), they took a neutral stance on organizational politics. Their model proposed that the antecedents and outcomes of political influence depend on many intervening variables, such as the combinations of tactics used, the need to justify decisions, the facilitators of political effectiveness, and the leader’s motivation, reputation, and interpersonal style. Also important were the status of the leaders and followers and the question of whether the political behavior was directed toward individuals on a one-to-one basis, in coalitions, or in networks.
Culbert and McDonough (1980) considered effective brokering inside and outside their own groups as a challenge for leaders. Inside the group, the leader should try to negotiate arrangements with the members that make for the best possible mix of members’ serving their own interests and meeting the needs of the group. Externally, the leader as a representative can often do much to increase the group’s resources and opportunities through effective negotiations with higher authority, with constituencies, and with opposition. Whereas, within organizations such negotiating behavior with individual subordinates may be a counterproductive leadership style,9 the leader’s successful playing of the role of broker is worthy of further exploration.
The effective political use of authority and power is still a highly underresearched area. Particularly important and underresearched is how the would-be organizational politician can be converted into an organizational statesman. It is possible to be an “honest broker,” for whom the marshaling of evidence becomes as significant as the relative interests of the parties in a negotiation. In this regard, the similarities and differences between legislative leadership and executive leadership need to be examined more fully. We are likely to see a continuing interest in the specifics of the influence tactics employed by leaders and managers to obtain desired outcomes from their superiors, their subordinates, their organization, and outsiders (Kipnis, Schmidt, Swaffin-Smith, & Wilkinson, 1984), as well as the use of increasingly refined instruments to do so (Schriesheim & Hinken, 1986). Hermann, Snyder, and Cunningham (1980) posed a number of questions about the way the organizational politician uses staffs, advisers, and confidants: How are they chosen? To what extent do they insulate the leader from the rest of the organization? To what extent do they amplify, distance, or reduce incoming and outgoing communications by the leader?
Drake and Moberg (1986) called for research on how leaders modify their language to create consistent forms of calibration and impression. Contingencies in the phrasing of requests need to be examined. Tactics need to be related to desired outcomes. The effective use of disclaimers needs to be better understood, along with other ways in which the forms of language can foster interaction.
It is not feasible for even the most active leader of a complex business to have a hand on every lever … not practical to juggle with a fast-moving present … and the imponderable forces of the future. … The direct order [needs to be] replaced by a framework—a set of values to guide the independent actions [of others]. Devising that framework needs sympathetic thought, specific to the objective. … The well-tempered leader … will combine tactical skill, judgment, and the ability to coalesce the intentions [of others]…. [He will be] an impresario skilled in generating enthusiasm and … the aspirations of [people with talent]. (Benton, 1990, p. 349)
Ciulla (2004) suggested that the writings of classical philosophers such as Plato and Confucius contained many principles of ethics worthy of modern empirical research. These classics contain themes and values that offer well-grounded ideas about who we are, what we should be like, and how we should live. In reviewing the ethics of leadership, Bennis (2004) concluded that particularly significant for further study will be the continuing examination of leadership from a cross-cultural perspective. We need to understand each other’s symbols, values, and mind-sets. For instance, Westerners need to develop a better understanding of leadership in the Islamic world, especially its tribalism, a handicap to globalization. Bennis also suggested that future efforts need to focus on leadership that continues to involve rhetoric, artifice, and perceived authenticity. The public figure is mainly a virtual leader, depending on the media to carry his or her message to the public. Impressions from the media create the perceived character of the political leader. We need to do more research about how the media affect the reputation and behavior of public leaders. Spin begins where reality ends. Trust in leaders falls as belief in the media’s manipulation of reality increases. Instant polling allows leaders to change their positions in midspeech. More must be learned about how modern leadership is affected by the Internet as it grows in size and importance as an informal channel of information, rumor, and gossip (Bennis, 2004, pp. 340–341). Much of this has equal relevance to political leadership in the private sector, about which little is known (House, 1999)
Both legislative and the organizational politicians may adopt influence tactics that border on the unethical. Flagrant lying, cheating, and stealing are clearly unethical managerial practices. But the extent to which telling white lies, cutting corners on standards, and making unauthorized use of company or government property are clearly unethical varies with the circumstances and individual differences in the concepts of right and wrong. Despite the importance that it should have, this is a severely underresearched area of empirical research on managerial and leadership behavior. A 1988 Louis Harris poll of 1,031 office workers and 150 senior executives reported that while 89% of the employees thought that it was very important for managers to be honest, upright, and ethical in their dealings with employees and the community, only 41% said this was actually true of their current employers (Anonymous, 1988a).
Weber (1989) found in interviews with 37 managers that they reasoned morally in order to conform to majority opinion and their organization’s rules rather than due to any universal principles of morality. For Burns (1978), the fostering of moral virtue was fundamental to being a transformational leader. The self-seeking charismatic, by definition, could not truly be a transformational leader. But Howell (1988) and Bass (1989) suggested that many charismatic leaders have two faces. They can show their prosocial one, which is transforming, and keep their self-aggrandizing face from interfering with the process.
For Kuhnert and Lewis (1987), the transformational leader develops from having a concern for personal goals and agendas to having more mature, higher levels of values and obligations. At the highest level of development are to be found the endorsement of universal ethical principles of justice, the equality of human rights, and respect for human dignity. There is a belief in and a commitment to these principles as valid and as an end in themselves. Thus, the values of life and liberty led Paige (1977) to argue that the pursuit of nonviolent means for dealing with problem solving has to be a central focus in the education of political leaders. Somewhat less idealistic is the extent to which values are relative to the group and circumstance; sometimes they may conflict with one another. But the organization and society are maintained by the decision to do what is right. At the other extreme, if the ethical sense is absent, one obeys the rules to avoid being punished (Kohlberg, 1969).10
Chapter 11 reviewed the few studies that could be found that dealt with managers’ attitudes toward corrupt practices. Revealing to a competitor one’s own company’s bid for a payoff was almost universally condemned, but taking petty payoffs from outside contractors for awarding them contracts was much less frowned upon. Insider trading is seen as more a matter of illegality11 than immorality (Pitt, 1985). The Enron scandal of 2001 showed how widespread and devastating corruption could be, as it involved fraudulent behavior by the CEO, vice presidents, leading financial officers, accountants, and outside auditing agencies. It resulted in the bankruptcy of the firm, the destruction of the auditing agency, the loss of employment by thousands of employees, and the employees’ loss of pensions and savings. Thousands of shareholders were left with worthless shares (St. Petersburg Times, 2006).
An experimental program was launched in 1986 to apply Misumi’s (1985) PM leadership in seven Chinese factories, hospitals, and agencies. An idiographic character function of leadership was added to the survey feedback process with emphasis on the moral character of the leader. Along with elements of effective human relations familiar to the West was included in moral character “the commitment to remain within the law and resist temptations for personal gain” (Peterson, 1989, p. 33). Also included was willingness to follow the Communist Party line even when one’s personal views conflicted with it.
In the West, professional organizations, such as those in the health professions or in public accounting, tend to monitor violations of their rules on a case-by-case basis. But beyond this monitoring, critical-incidents survey methods used by the American Psychological Association (1953) are available for gathering both clear-cut and borderline incidents of immoral behavior by managers and the various alternative ways in which managers and leaders could have acted that would have been morally acceptable in terms of informed consensual opinion or universal standards. We are still waiting to see such an effort in the fields of organizational management, led by the accounting profession. But we are more often likely to see questions of commitment to ethical behavior added to survey feedback.
A Model for Ethical Analysis. For Socrates, ethics was the search for the good life in which one’s actions are in accord with the truth. According to Steidlmeier (1987), ethics is creative—the search for human fulfillment and the choosing of it as good and beautiful. Ethics is practical in purpose. It seeks the full flowering of the human person and excellence in the actualization of the human capacity. Moral virtues are life-giving patterns of behavior; moral vices are destructive ones.
Most discussions of professional ethics focus on the destructive vices. They are negative and guilt-ridden. The focus is on what should not be, rather than on what should be. Discussions are defensive and dogmatic. Steidlmeir accented the positive. He suggested that we need to clarify our worldview and our principles, which should be integrated into our educational and developmental processes. Unfortunately, principles of self-regulation and codes of conduct often conflict with the avoidance of costs and the maintenance of one’s competitive edge. There is a need to recognize the costs and to share them with public officials, stockholders, management, employees, and the public to gain societywide benefits. Weber (1989) pointed to the need to examine how the organization impinges on the manager’s maturity of moral reasoning. Beyond this, there is a need to determine the connection between moral reasoning and moral behavior and how each depends on the issue involved.
What’s needed is a model that can fit both the hortatory literature on business and political leaders’ ethics and social responsibility and the survey evidence of current leaders’ and managers’ beliefs and practices. Forrest, Cochran, Ray, and Robin (1989) made such an attempt but were unable to account for the ethical opinions of 315 surveyed managers and 577 surveyed teachers. The investigators extracted four bipolarities from the prescriptive literature: (1) economic versus social purpose, (2) company policy versus individual managers’ discretion, (3) recurring versus one-time problems, (4) utilitarian versus ideological decision-making rules.
Organizational Citizenship. Conceptualizing organizations as a collection of constituencies (March & Simon, 1958) remains a powerful source for understanding the individual organizational leader’s performance. Much of the success of leaders with their subordinates may depend on how well they negotiate arrangements with other units. By now, the need to focus more attention on the manager’s horizontal interactions with peers in other units and with clients (Dubin, 1962b) has become a platitude. However, these leader-leader interactions could be studied with considerable profitability (Hermann, Snyder, & Cunningham, 1980).
Whistle-blowers. In opposition to the political balancing of self-interest, a new concept is taking shape that transcends the formal horizontal or vertical relationship. It is that of the good organizational citizen who goes beyond his or her immediate self-interest to promote the well-being of the organization, its members, and society as a whole. “Whistle-blowers” are illustrative. There is a need in both the public and private sectors to provide more support for valid whistle-blowers and less for their corrupt superiors. More often than not, the whistle-blower is likely to suffer more punishment than the corrupt superior. Nonetheless, big rewards are being paid to whistle-blowers based on how much money was earned and how much in fines was assessed in the fraud or other crimes. What may follow the “me-too” generations of the past 40 years is more good organizational citizenship. Or there may be a reversion to the “my-group” generation of the 1950s (like the East Asians of today), which was concerned primarily with peer pressure. What happens will have considerable implications for leadership.
Ethical resolve continues to remain more important than philanthropical concerns to senior management, according to surveys of several hundred CEOs in 1988 and 1989; but ethics are less important to them than legal or economic considerations. Nor does ethical resolve appear to relate one way or another to a corporation’s financial performance (Aupperle & Simmons, 1989). Nevertheless, research on the morality of business leaders and the citizenry is likely to increase substantially in the twenty-first century if the continuing political lobbying scandals and the power of money to win elections are any indication. The increase in the consideration of the effects of spirituality and religious orientation on transformational leadership also may have ethical implications (Twigg, 2004).
In a review of 89 studies published between 1970 and 1975, Hunt, Osborn, and Schriesheim (1977) reported that 61% used only a single criterion of effective leadership and 43% of that group emphasized performance. However, a greater use of multiple criteria was noted in field studies. Most field studies (81%) used criteria obtained from a different informational source than the predictors.
Studies have been conducted on followers’ beliefs, satisfaction, and behavior, as well as productivity, drive, and cohesiveness. However, certain variables, such as group productivity and followers’ satisfaction, have been overemphasized at the expense of other variables. For instance, the mean true correlation was .30 between job performance and job satisfaction for 54,417 individuals in 312 samples, according to a meta-analysis by Judge, Thoresen, Bono, et al. (2001). Followers’ efforts, which presumably should be affected by leadership, have been neglected as an outcome due to leadership.
Use of a Supermeasure. Morgan and Rao (2002) advocated using one single criterion of effectiveness in service organizations that align their strategy with operating systems serving well-defined markets. One supermeasure focuses the attention of these service firms on their customers’ needs better than a complex set of criterion measures does. For instance, Continental Airlines found that on-time performance of its flights encompassed an effective incentive system that was simple to implement, communicate, and build a strategy around—one that provided a feeling of success when Continental’s on-time arrivals ranked highest among airlines. Nordstrom, a department store, identified customer satisfaction as its supermea-sure, Marriott supported careful, selective hiring of staff for its economy service. Fairfield Inns selected its super-measure of cleanliness, friendliness, and efficiency.
Consecutive Criteria. Anderson, Lievens, van Damm, et al. (2004) noted that flexible forms of work teams and rapidly changing organizations mean that the criteria for a single leadership job are less relevant to predicting performance than are the criteria for a role involving a multiplicity of assignments (Howard, 1995). Instead of predicting person-job fit, the effort has to be in two stages. First, there is a need to determine the composition of a changeable job role; then there is a need to predict the person-role fit. Person-role fit may entail considering, along with aptitude predictors of adaptability, flexibility, innovativeness, openness to change, and skills needed for the future (Herriot & Anderson, 2004).
Evaluations Are Relative and Subject to Change. Evaluations of leaders are subject to revision. Shartle (1956) observed that ten years too late, a firm might discover it had fired its most effective manager. Often vilified while he was in office, President Harry S. Truman’s performance was upgraded considerably in the years that followed. Seen as being more interested in playing golf than in taking decisive actions, President D. Eisenhower was later found to have played an active role in substituting clandestine forces in Iran, Central America, and elsewhere for open displays of U.S. military power actions. These actions had delayed and devastating effects on those countries and on subsequent U.S. relations with them. Donald Burr, at first evaluated as a hero who devised a radical new approach to managing airlines, became the villain who failed to consult with others and overexpanded PEOPLExpress to the point where it was thrust into bankruptcy.
Evaluations are relative. Well known are the stories about executives who introduce belttightening measures into a department, ride roughshod over subordinates to reduce costs, maximize immediate productivity, and are then promoted to repeat the process higher up. They leave behind a shambles of dissatisfaction and conflict to be blamed on their successors. Peters and Waterman’s (1982) popular assessment of exemplary firms that represented excellence in management was based on a highly subjective evaluation of excellent management. When a fuller analysis of the firms’ financial performance was compared with those of 1,000 firms evaluated by Forbes, the performance of the first group of firms was seen as not particularly better (Aupperle, Acar, & Booth, 1986).
What scholars view as effective leadership and how followers view the same performance may be quite different. In contrast to presidents from Woodrow Wilson through Richard M. Nixon, who fulfilled nearly 75% of their campaign pledges, Jimmy Carter did almost as well but was seen as a failure by the public because he did not fulfill the major campaign promises that were important to voters: reducing unemployment and inflation (Krukones, 1985). Despite many mistakes made in his first administration, George W. Bush won reelection in 2004. But his popularity fell from 90% to 30% as his administration’s mistakes continued; and by 2006 he could not figure out how to withdraw from the war in Iraq he had started in 2003. The discrepancies between scholarly appraisals and the public’s evaluation of a leader’s performance are particularly apparent in the case of demagogues.
There is a need to distinguish between the successful influence of a leader and his or her long-term effectiveness, as determined by the contribution that the “influence” makes to attaining the long-term goals and well-being of the organization or nation. For instance, researchers need to broaden the criteria of consequence and to attach the appropriate longer-term value of leadership to the development of followers, the organization, and the nation (McCall, 1977). Human resources accounting is an effort in this direction. As managerial accountants increasingly become interested in the behavioral side of a firm’s assets and liabilities, there should be increasing opportunities to measure such effects (Caplan & Landekich, 1974). Utility analysis is a promising tool (Cascio & Ramos, 1986). But considerable care will be required to weed out the cause-effect relationships. First Lady Nancy Reagan was effective in the short run in the nation with her campaign “Just say no to drugs,” which was largely forgotten once her husband left office, but Eleanor Roosevelt had lasting effects on the world with her influence as an advocate for human rights and the United Nations (Wills, 1994).
Additional questions about the evaluations of the effectiveness of leaders were posed by Hermann, Snyder, and Cunningham (1980). Who determines the criteria for effective leadership, and what difference does it make? (Shareholders will emphasize stock prices; employees, job satisfaction.) Are output measures, such as the size of profits, votes obtained, or church membership, sufficient measures of the effectiveness of leaders, or are leadership-process variables more important? How do researchers avoid having leaders directly influence the raters to render favorable judgments about their effectiveness, a phenomenon examined by Dubin (1979)?
Why does a leader’s behavior affect the group’s outcomes? Why do the group’s outcomes affect a leader’s behavior? The leadership roles serve the necessary functions of maintaining role structure, role freedom, goal direction, and cohesive group action in the performance of a task. These roles do not necessarily specify the patterns of behavior required but, rather, the purposes they serve. These purposes include dealing with goal direction, individual members’ goals, and relations within the group; maintaining the group’s role structure, constraints and freedom; and promoting individual and group motivation, individual satisfaction and group cohesion, and the group’s norms and requirements.
In the emergent situation, leadership is created by the group in response to its own necessities even though an impatient member may impose himself or herself upon the group before the group expresses a demand for such services. In formal organizations, the necessity is acknowledged by the appointment or election of leaders. The effective performance of a task is usually dependent on the appointed or emergent leader’s enactment of the leadership roles to carry out the necessary functions. How are the role enactments affected by changes in membership, task, and organization? Are there satisfactory substitutes for leadership, such as overtraining of followers, computerization and automation, or policies and regulations? How far can self-management be taken with such substitutes?
Computers can be substituted for a supervisor for monitoring of and feedback (to both employees and a higher authority) about an individual employee’s productivity, wasting of time, frequency of errors, and presence at or absence from the workstation. Computers can trouble-shoot and inform employees about corrections they should make and how to make them. Are such monitoring and feedback likely to be favorably accepted by the employee, in contrast to less close human monitoring? Are some aspects more coercive than others? What if this type of monitoring and feedback represents a diminution of the job for the employee, with the associated effects of increased boredom, dissatisfaction, and loss of self-esteem?
Katz and Kahn (1966) pointed the way. A systems approach looks at the leader as someone imbedded in a system with multiple inputs from the environment, the organization, the immediate work group supervised, the task and the leader’s behavior. Other factors include his or her relationships with subordinates and outputs in terms of effective performance and satisfactions (Bass, 1976). Hunt, Osborn, and Schriesheim (1977) thought this kind of systems view of contingent variables was particularly important for future research. For instance, studies of military leadership confirmed that combat and noncombat conditions require different patterns of effective leadership behavior. Such differences are likely to vary by the officers’ level. Upward and lateral influence in the system can be as important as downward influence to the accomplishment of missions (Van Fleet & Yukl, 1986). Gal (1989) argued that military leadership in the future will be derived from the commitment of subordinates to rise to the challenge of complex, risky, uncertain, and dangerous situations.
According to Demming (1986), managers need to know whether the system in which they are involved is stable and predictable. Are the system’s goals and norms stable? Inefficiencies that show up repeatedly call for managers’ efforts to improve the situation. Any one positive outcome by itself, such as the satisfaction of customers or meeting the competition, may mean little to the overall health of the system. Zero-defects programs or new equipment alone may also be of little consequence to the system. Management needs to focus proactively on the constant improvement of the design and processes of the system that lie within its capabilities.
Much needs to be learned about how leaders manage the boundary between what is inside and outside their mission in the larger organization and how to protect it. How do they create the necessary environment for productivity within their boundaries? To what must they attend, and what can they ignore? How do they help others to make sense of arrangements and situations (Gilmore, 1982)?
Adams and Yoder (1985) advocated that particular attention be paid to the need to consider leader-subordinate relations in the broader setting in which actions take place—the wider organization and external environ-ment—as was done in Chapter 25. Hermann, Snyder, and Cunningham (1980) noted the absence of research about how a leader’s behavior is affected by a variety of specific contextual issues. Will a number of groups need to be coordinated, or can one group handle the problems? Is the organizational problem a crisis, or is it routine? Is the problem familiar or unfamiliar to the organization? Is the leadership operating with immediate subordinates or at a distance? Similarly, Behling and Schriesheim (1976) and Schriesheirn and Neider (1988a) saw the need for more future research about the externally oriented activities of managers with higher authority, other department heads, clients, and the community. In all, a more complex set of roles will be seen as part of managerial leadership.
The application of a systems approach to leadership was reflected in the principles set forth for senior executives by the Center for Creative Leadership. Executive leaders need to get in touch with “what is really going on” at different levels in the organization. They need to articulate a vision with clear goals, standards, and priorities, along with a plan for developing a climate that will support the effort. Key leaders at various levels of the organization will need to set examples of such support. Periodic measurements of the effort and its success are required. Risk taking and trust building need to be encouraged. Competitiveness that reduces information sharing needs to be discouraged. Suitable appraisal and reward systems for outstanding individual and team accomplishments should follow, with continuing attention to effective personnel staffing, development, and succession to fit the programmatic effort.
In contrast to microlevel research on the leader in relation to his or her subordinates and immediate superiors and to macrolevel research on the aggregated data that relate to the total organization and its environment, mesolevel research involves the individual leader, on the one hand, and the total organization and environment surrounding the leader, on the other. More such mesolevel research is needed (House, 1988a).
The Learning Organization. Senge (1990) introduced the concept of the learning organization. It adapts to the changes in its environment. Its leaders learn how to move its members to adjust to needed changes. The leaders and members learn to be more flexible. New perspectives are adopted when the organization moves from national to global markets. Leaders and members adapt to the needs to diversify the workforce and customers. In the learning organization, the permanent bundle of specific tasks that form a job is replaced with a temporary combination of tasks. Some tasks may be shifted to others in the organization or outsourced. Needed also in the learning organization is consultative leadership that articulates what needs to change, careful listening, the giving and accepting of feedback, coaching, mentoring, calculated risk taking, and careful experimentation (Bass, 2000).
A few efforts have been made to identify the discretionary opportunities that shape a leader’s behavior. Much more study is needed to determine what is and is not under a leader’s control. Such knowledge should reduce subordinates’ unrealistic expectations about the leader, as well as the leader’s own experiences of violated expectations. In selection and placement testing, researchers should be able to generate higher validity coefficients between personality attributes and the portion of a leader’s behavior that is discretionary.
Delegation remains a relatively unexplored management option despite the evidence of its important contribution to the leader’s and the organization’s effectiveness. Miller and Toulouse (1986) showed that the extent to which delegation of authority was practiced in 97 small businesses in Quebec correlated .31 with profitability and .34 with sales growth of the business relative to others in the same industry. The effects of delegation were stronger than any others examined, such as the business strategies employed, how they were devised, and the CEO’s personal characteristics.
Studies of the delegative process are likely to be important in future research on leadership, with more finetuning of delegation as a style of leadership. Delegation is clearly not participation. Although Vroom and Yelton (1973); Bass, Valenzi, Farrow, and Solomon (1975); and Leana (1986, 1987), among others, contributed to some understanding of delegation, it still remains the least researched style.
As a step forward in the research effort, Schriesheim and Neider (1988) conceived of three forms of delegation instead of one. Research is needed to learn whether each of the forms is likely to interact differently with the motivation, commitment, satisfaction, and performance of the leaders and subordinates: (1) advisory delegation, in which the subordinate makes a decision after first getting a recommendation from the leader; (2) informed delegation, in which the subordinate makes the decision after first getting needed information from the leader; (3) nonsuggestive delegation, in which the subordinate makes the decision without any input from the leader. Of the three forms, advisory delegation is likely to coincide with a strong commitment from the subordinates.
Delegation has been regarded as the style at the other end of the continuum of styles from autocratic directive leadership. Yet an autocrat may delegate because of lack of interest or time to handle the problem directly. A transactional leader may delegate in exchange for subordinate support. A transformational leader will use delegation to develop his or her subordinates. A consultative leader may delegate after being convinced of the subordinate’s competence and motivations to handle the problem. Delegation may be the choice of a participant consensus. A laissez-faire leader may delegate to avoid taking the blame for a possible failure.
There is a need to conceptualize delegation as a process and to research what makes it effective or ineffective.
Personal traits that are associated with the emergence of a leader appear to transcend situational demands. It is still axiomatic that some of the variance in the emergence and effectiveness of leadership is due to personal traits, and some is due to situational and interactive factors. Nevertheless, a meta-analysis by Lord, DeVader, and Alliger (1986) of data from Mann (1959) and additional data suggested that both Stogdill (1948) and Mann seriously underestimated the importance of personal traits relative to situational effects. The meta-analysis showed that the theoretical mean correlations predicting emergence as a leader between such traits as intelligence, extroversion, and assertiveness, and actual emergence as a leader were much higher than had been realized. These high correlations occurred even when suitable adjustments were made for restrictions in the range of the subjects of the studies and for the unreliability of the measures employed. Also taken into account in the meta-analysis was the likelihood of random variation from one study to the next, which explained much of the situational variance and argued against the need for contingent explanations. Assertiveness, for example, is necessary for leaders, regardless of the situation.
Contingent models are still required to explain how the styles required for effective leadership vary with the demands of a situation. For example, Barbieri (1983) used path analyses to show that much of the 39% of the variance in revenues from the sales of mature products generated by marketing managers and their sales representatives could be accounted for by the managers’ extroversion, relations orientation, and intuitive thinking. But revenues from the sales of innovative products were accounted for by the managers’ task orientation and the sales representatives’ experience, practical orientation, and intelligence. At the same time, the branch managers’ unwillingness to delegate was not a contingent effect; it contributed positively to the sales of both kinds of products.
In sum, despite meta-analytic support for generalizing across situations about personal traits associated with the emergence of leadership, such as intelligence, assertiveness, and extroversion, applications of contingent models of leadership are likely to remain prolific sources of further examination and validation. This occurs particularly with reference to the styles of leadership that are most effective in designated circumstances. More causal, longer-term, larger-scale contingent studies of styles of leadership and their effectiveness are likely to appear during the coming decades.
Research on the various assessment methods employed and their integration into decisions about assessees need to be studied further. The phrasing of more than 100 such research questions emerged from a brainstorming session on the subject (Jeswald, 1971), and most of them still remain unanswered or partly unanswered almost four decades later. For example, much more remains to be learned about the kinds of biodata and biometric data that can be profitably used for assessment.
The development and evaluation of more sophisticated theory-based, in contrast to empirically-originated, biodata forms of greater validity are needed. These forms can be developed and evaluated by asking questions that generate responses more likely to relate to the personal styles that are of consequence to subsequent performance as a leader. Thus researchers can ask questions about how the applicant coped with “the most important challenge you faced” or “your most disappointing experience.” Appropriate scoring, consistent with theoretical expectations, can be developed (Kuhnert & Russell, 1989). A list of such dynamic items for application blanks was suggested by Bass and Barrett (1972)—for example, “May we consult your present employer?”
In this endeavor, as well as in the development of new psychological tests and measurements, experimental findings about personality and social psychology can be put to better use. More complex simulations that duplicate critical aspects of on-the-job performance should be employed. The subtleties of the deep structures involved can be introduced into such simulations. More should be done to assess creative, flexible intelligence, as well as less socially desirable aspects of personality, such as dogmatism and authoritarianism, that may influence leadership behavior. More research is needed on newer personality dimensions that are of consequence, such as the tolerance for ambiguity. A lot more needs to be known about the abilities needed to cope with higher-level managerial responsibilities, including the ability to translate ideas into actions. More sophisticated assessments of leaders’ performance in decision-making processes are desirable. Justification of the greater expense of assessment centers over ordinary selection methods calls for more utility analyses such as the one presented by Cascio and Ramos (1986). The contextual conditions surrounding the assessment process also need to be taken into account more fully (Bentz, 1987). For example, Kleinmann, Kuptsch, and Köller (1996) showed that for 119 students in a German assessment center, construct validity was higher if the dimensions on which they were being assessed were transparent, that is, if the students were informed in advance about the assessment dimensions and the behavior that was required of them. In the same way, McClelland (1998) noted that for good competency assessment in interviews, the competencies assessed and how they may be improved should be made public instead of kept secret as in the case of traditional testing of traits.
Expected Developments in Assessment. FernándesBallester (1995) suggested that in the coming decades, new evaluations of mental processes will appear that examine human cognitive functioning, the dynamic assessment of intelligence, and applications of Item Response Theory to computerized and adaptive intelligence tests and personality inventories. Also, new ways of reducing biases in self-reports will emerge. New biophysical measurements of emotion and new ways of assessing person-situation interactions will appear. Many uses for assessment will accrue from the availability of virtual reality. Further studies will be needed in the utility of cognitive, social, and emotional assessment by individual tests in comparison to group situational tests. Assessment technology will have spread to many more countries in many more languages. Further developments can be expected in controlling response set, faking, and desirability of responses.
On one hand, researchers can point with pride to the large array of positive evidence about the efforts to train and educate leaders and the success that has been experienced.12 On the other, it is still possible to decry the many continuing inadequacies of leadership training and education. Kerr (2004) summarized the results of a symposium of book authors on the subject of how organizations can best prepare leaders to lead and manage. All four authors agreed, and argued convincingly, that developing leader-managers early in their careers is important. But this effort has to be limited, for their turnover is much greater than those of long-term employees and so much less is known about them. This places a premium on careful selection and appraisal. Secondly, work assignments have to be seen as a developmental tool. An assignment should provide a learning opportunity. But more often, an already experienced employee is given an assignment to ensure its successful completion.
Leadership development is a continuing process. Researchers need to learn a lot more about how experience with subordinates, peers, and superiors, as well as with family and friends, shapes one’s subsequent performance as a leader. Research on the persistence and transfer of leadership has produced a convincing body of evidence indicating that one’s past success as a leader contributes to one’s future success as a leader.
What experiences are necessary to make an effective leader? Does one need to be trained in business to be an effective industrial leader? Does one need to be trained in theology to make an effective pope? Does one need to be a lawyer to be an effective legislative leader? Does one need legislative experience to be an effective state governor or president?
Innovative and creative approaches for teaching leadership using films, classics, and fiction are provided by Pillai and Stites-Doe (2003). For instance, some of the subjects covered in college courses include: (1) What is leadership, and is it learned or innate? (2) What are the qualities and behaviors of a great leader, and are they the same for all cultures? (3) Are political leaders history’s slaves? (4) Are there unique challenges for women leaders? (5) How do great leaders handle mistakes and failures? (Nathan, 2003).
Based on a review of 20 studies, Atwater, Brett, and Waldman (2003) see both benefits and risks in multirater feedback. Improvement based on feedback was not uniformly found. Negative feedback was likely to result in reduced performance and commitment, anger, questioning of the validity of the feedback process, and discouragement. To be useful, rating systems need to be ongoing and integrated with a means of training and development. Conger and Toegel (2003) foresee a systematic set of action projects being assigned to an employee so that learning from them can build and be integrated. Qualitative data needs to be included. (In the future, we will see more analyses of qualitative data for feedback.) For Uhl-Bien (2003), relationship skills are the social capital that organizations need. Awareness of the other person’s needs, mutual obligation, trust, and respect are present in high-quality relationships. Without them, teamwork suffers. Leadership development requires more attention in the future to training in relations skills.
Kuhnert and Lewis (1987) raised a number of unanswered substantive questions about the development of transformational leadership. Are there observable changes in transformational leaders’ behavior as a function of the leaders’ personalities? Is transactional leadership a less mature developmental phase than transformational leadership? Are leaders’ developmental phases invariant? Does one have to be a transactional leader before he or she can become transformational? What happens when leaders and followers operate at different developmental levels?
The Genetic Factor. The nature-nurture controversy is by no means settled. Advances in genetics and biology need to be incorporated into leadership models. Large-scale studies of heredity suggest that genes contribute to energy levels, intelligence, long life, interests, assertiveness, a sense of well-being, the ability to take risks, and even job satisfaction. The genetic factor needs to be taken into account in any complete examination of leadership. There are remarkable correlations in these kinds of traits and outcomes among identical twins reared apart, in contrast to people who are genetically unrelated.
Purposes of Leadership Education and Training. Stogdill (1974) faulted research on training for focusing too much on the extent to which training produces attitudinal and behavioral changes in trainees. He demanded more evidence on the impact of the training of leaders on followers. Yet researchers still need to link particular training efforts with particular behavioral changes. For instance, training may understandably increase a trainee’s sensitivity; nevertheless, increased sensitivity may be counterproductive on the job. Sensitivity training may incapacitate the leader for coping with strong opposition, threats, and challenges to the legitimacy of his or her status. What is at fault here is not the training as such, but an inadequate analysis of the situational demands on the leader. We are beginning to see much greater precision in such analyses. Over time, it is expected that the application of sophisticated analyses of managers’, administrators’, and leaders’ roles to training will be much more intensive.
For Rost and Barker (2000), leadership education, such as that furnished by secondary and higher education, needs to be distinguished from organizational leadership training. Leadership education should produce citizens for a democratic society who understand the dynamics of the change process and the politics of distributing scarce resources to facilitate change. Participants need to learn how: to get into the decision-making process; collect, interpret, and disseminate information about the issues involved; develop a reputation; think critically in public discourse; and confront opposition. Cherrey and Isgar (1998) propose as goals for teaching students leadership in the twenty-first century: (1) understanding the complexity of an interrelated system such as a multinational NGO with many alliances and diverse units; (2) continually reflecting and learning with a commitment to the greater good; (3) valuing diversity and embracing inclusiveness; (4) practicing collective leadership; (5) focusing on the leadership process, not on the leader; (6) linking academic and extracurricular activities that provide opportunities to learn about leadership. For Ayman, Adams, Fisher, et al. (2003), who evaluated the leadership development programs at 63 U.S. universities, fewer than 10% appear to offer the complete program envisaged: (1) Long-term involvement, (2) the facts about what works, (3) insights gained from reading and case discussions, (4) opportunities for reflection, (5) opportunities to lead, (6) increased self-awareness as a result of personal feedback. The attention to the importance of individual merit, competitiveness, and the “bottom line” needs to be leavened with the importance of cooperation, corporate social responsibility, and ethics (Berry, 1997). Aupperle and Simmons (1989) compared the forced-choice tested attitudes of 245 CEOs in 1981 and 1988 and found there was no change in how much emphasis they placed on their corporate social responsibility. In order of importance, from most to least important, were economic, legal, ethical, and philanthropic responsibilities. The U.S. Army, Navy, and Air Force military academies come closer to meeting Ayman, Adams, and Fisher et al.’s (2003) requirements for leadership education. Also closer to the requirements are colleges such as Jepson, which offer a four-year curriculum in leadership. In 1998, there were almost 700 leadership development programs in U.S. academic institutions (Honan, 1998). With increasing recognition of the importance of leadership, we can expect more to appear in the future, as well as full four-year programs in leadership.
Increased Reliance on Intuition. While new and better techniques for analytical decision making will be developed in the coming years, we will also learn and be trained in how to make better use of intuition—apprehending something without completely understanding it. Intuition is based on experiences and reflective thinking about the experiences. It uses metaphors and imagery. There will be less reluctance to depend on intuition if reflection connects the known information, perspectives, and feelings about the issue. It can add to what analysis can provide for making a decision in the absence of complete information (Cartwright, 2004).
Improvements in Evaluation. Evidence of the utility of focusing on specific behavioral training and its specific effects is to be found in the attention now being paid to behavioral objectives, behavioral modeling, and applications of social learning theory. But in many evaluative studies of leadership training, it is impossible to determine the method or combination of methods employed. Evaluative reports need to describe in detail both the method of training and the content taught. In addition, more attention needs to be paid to the threats to the validity of training designs and their evaluations (Kane, 1976). For example, for the purposes of unbiased evaluations, if possible, superiors should not know who has received training and who has not after trainees return to their jobs. Likewise, trainees should not be told whether they are serving as experimental subjects or controls (as in holdout designs).
Multinationalism and Diversity. The globalization of industry will mean increasing attention to international managers, as such, with the probable submergence of surface differences in national styles and performance, but with needed increases in attention to underlying differences in institutions, cultures, and governments. At the same time, with the rise in power and influence and the further economic development of countries whose cultures are alien to the West, attitudes, values, interests, and beliefs that are different from those in the West and that affect leader-follower relations will emerge. They are likely to emphasize tradition and collectivism more heavily. Globalization has been accompanied by the emergence of a global civil society whose problems challenge the international community. Learning from the experiences of corporations and institutions can be helpful in satisfying the “need for leadership to promote peace, democracy, and development within (nations) and between (nations) … the national leader’s (has a) responsibility to serve his people … and global civil society’s efforts to promote peace and development worldwide … without losing sight of the larger picture linking leadership to human development” (Safty, 2000).
In the United States, studies of black, Hispanic, Asian-American, and women leaders will continue to increase, reflecting their movement into higher levels of leadership and management in industry, government, and education. A greater openness to line management by Jewish Americans, Asian Americans, and other minorities is likely to increase with the increasing technological requirements for remaining competitive. As more women move into the higher reaches of management, we may see more of an emphasis on androgyny than on masculinity as being favorable to effective management (Adams & Yoder, 1985). Extensive commentaries will be replaced by more empirical studies. More empirical studies will also be needed, considering how quickly outdated conclusions from them can become.
Because of the emotional content involved, care needs to be taken not to accept at face value leaders’ and subordinates’ cross-race and cross-sex opinions and descriptions. At the same time, more attention has to be paid to the underlying feelings of rejection, contempt, guilt, and threat in mainstreamers and outsiders that do not surface because of superficial socialization, socially desirable responses, or mistrust of the investigators. This is another reason why empirical monitoring of such dimensions as the self-confidence of blacks, women, Hispanics, and particularly Asian Americans is essential. It should be recognized that an effective, polite, mutual acceptance can be maintained at one level while underneath, a wall of misunderstanding still exists.
Little is known about Anglo-Hispanic leader-subordinate relations. Yet Latinos are now the largest U.S. minority. Research here will be made particularly difficult by the wide differences among Mexican Americans, Puerto Ricans, Cuban Americans, and the large immigrant groups from the central Caribbean and elsewhere in Latin America.
So far, the preponderance of evidence endorses the need for minority members serving as leaders in majority environments to emulate the original white male manager. However, more and timely research will be needed on the accommodations made by minority members to the duality of their roles as both managers and members of a minority group. Leadership is also a completely different matter for minority community leaders, who ordinarily need to identify more with their own subculture than do their followers.
Work/Family Balance. There will be more focus on family-friendly policies. More will need to be learned about the policies needed for the different types of child care by single mothers and single fathers, divorcées, homosexual couples, and grandparents. More research is needed on the full range of working family forms and incomes for policies across socioeconomic levels. Beyond this can lie a future where families engage in educational experiences for mutual benefit, from learning computer programs to learning foreign languages. Increasingly, we are likely to see government policies toward families becoming friendlier, with more flexible work weeks, longer part-time hours and more sharing of the costs of child care as well as greater portability of pension benefits, care for parents, and solutions to the problems of overscheduled lives. Globalization may bring the United States closer to many European nations in family friendliness (Murphy & Halpern, 2005).
One of the areas of leadership among many that I have neglected is leadership of organized criminal gangs such as the Mafia, Colombian drug lords, and the Japanese Yakuza. The gangs pursue a variety of illegal activities and also invest in legal businesses. Their leadership tends to be autocratic and coercive, supported by disciplinary rules and codes of honor. Informants and undercover agents unveil the secrets of their operations. Fictional portrayals are common in literature and TV and, like The Sopranos, can be used to propose principles of leadership, but can serve only as illustrations (Yammarino, 1996).
Many important issues remain relatively unexplored. Others have been overworked. Much original, creative leadership research has appeared since the early twentieth century and especially since the 1940s. But it has also been accompanied by a wasteful repetition of tests of shopworn hypotheses. The atheoretical research published before 1965 was too unfocused, but much of the research that followed was based on naive, uncritical theorizing and retarded the process of new discoveries. Nevertheless, much has been accomplished. Based on his 32 years as a military leader and 10 years directing the Center of Creative Leadership, Ulmer (1997) found that much had been accomplished in leadership research by then, although there had been diversions. Yukl (1998, p. 508) concluded that the field “has witnessed an increase in the richness of research questions and the variety of approaches used to study them, and … appears to be undergoing an accelerating pace of discovery.”
Critics complain that leadership is fundamentally anti-democratic and antithetical to nonleaders’ rights of expression and power sharing. Other critics complain that despite all the research, nothing is known about leadership! Still others declare that leadership is a figment of the imagination or that leadership as a research subject is moribund and has reached a “dead end.” Nonetheless, this and the preceding chapters attest to the health and well-being of the subject. A wide range of new and challenging substantive issues have been created: implicit leadership, substitutes for leadership, prototypicality effects, neocharismatic and transformational leadership, strategic leadership, virtual leadership, leader-Member exchange, impression management, self-management, and upward influence. Among the methodologies introduced have been analysis by WABA, meta-analysis, and triangulation of quantitative and qualitative findings. More replicated studies have made possible many more meta-analyses to increase confidence in results. Both personal traits and situations have been shown to be important to leadership.
Considerable advances have occurred in the development and use of theories. Linkages have been forged to general theories, for example, between contingent reinforcement leadership and the general theory of reinforcement. Testable theories have been constructed and tested. No longer does one have to wait for the demise of the theorist before a theory of leadership is abandoned or modified because of its failure to be supported empirically.
The ability to predict future leadership performance accurately has increased substantially during the past half century. Many studies have taken advantage of the miniaturization of wireless physiological recording and the personal computer. More will be doing so in the future. The same is true for electronic mail, videoconferencing, and other forms of telecommunication, which will also increase in importance in linking leaders and subordinates who are distant from each other.
New complex and effective research designs are now available. These designs would be equally useful for testing other relationships that have not been explored. The production of critically needed information could be greatly accelerated by feeding new combinations of variables into research designs that have been demonstrated to be effective. Efforts are increasing to bring together active leaders, organizational consultants, and research investigators to discern common problems and interests and to increase the focus of this research on realities and the application of research results (see, e.g., Clark & Clark, 1989).
In the social sciences, old theories used to wither away with the retirement of the originators and their disciples. However, although some theories of leadership have gone or are going this route, other theories of leadership, such as path-goal theory, the contingency theories, and charismatic/transformational leadership have by no means withered. Rather, they have been systematically tested and reshaped as a consequence of the results of tests. And new models have been fashioned on top of them.
The preceding suggestions for future research indicate that the possibilities are far from exhausted. Leadership presents a lively, challenging field for research and innovative applications.
Optimism or Pessimism? The optimist sees the wine bottle as half full; the pessimist sees it as half empty (and sour to boot). It is easier to be destructive than constructive, particularly when one lacks information. “Know-nothings” argue that the past years have essentially been a waste and it is now necessary to start afresh. Miner (1975) proposed that the concept of leadership be abandoned altogether. Since then, he has changed his mind. M. W. McCall (1977) thought that leadership would remain an enigmatic subject because of its many definitions, but in Chapter 1, I noted that the definitions were evolving in a systematic fashion and that, at any rate, the diverse definitions do not seem to detract from model and theory building. The many models and theories that have sprung forth in the past five decades are not as divergent and conflicting as one might expect from the diverse definitions of leadership on which they may be built. Ways are needed to juxtapose one model with another to test which one yields more plausible explanations. Fiedler and associates tested Fiedler’s contingency theory; Vroom and associates tested their deductive model; Likert and his associates produced massive amounts of support for System 4. Points of theoretical disagreement can be found and critical experiments developed, as was done by Locke, Latham, and Erez (1987) and Blake and Mouton (1985) to determine which model best fits the data.
To some extent, the problem may be that each theoretical point of view is supported by a different array of measurements and situational circumstances. Part of the problem may also be that situational factors determine some kinds of leadership behavior but not others. Autocratic leaders—rigid, inflexible, and self-assured—may try to lead in any situation in which they are placed. Democratic leaders may attempt to lead only if they feel competent and supported in the particular situation. The nondiscretionary behavior of leaders will be determined by organizational and environmental matters that are not within their control. But leaders’ discretionary performance will be much more a matter of their personal predilections than of situational differences alone.
It is one thing to say that researchers know nothing because they do not obtain consistent results. The lack of consistency may be attributed to a lack of knowledge. Yet it may be that to achieve consistent results requires accounting for a complexity of variables and that, as researchers do so, they increase their understanding of what is happening. What is needed are better measurements; a broader appreciation of which situational variables are more important and which are less so (Korman, 1974); and, as was noted earlier, larger samples. Calas (1986) and Calas and Smircich (1987) joined the “cry and dismay” about the narrow substantive themes and methodological means of the past in what has been published about leadership by its establishment of scientific investigators. Nevertheless, the researchers believe that the field has the potential to become: much more open to broader approaches in the search for understanding; accepting of expanded horizons; reflective of more adequate connections with salient substantive issues of the real world outside the community of researchers; and populated with researchers learning from experience, observation, and narrative interpretations of meaning about what they write, as well as from continued dependence on field surveys and controlled experiments. Historians, anthropologists, and political scientists have much to say about leadership in their narrative reports. Illustrative is Schruijer’s (1997) historical essay on the early-medieval Christian bishops’ management of meaning between the fourth and eighth centuries. They carefully substituted and imitated pagan practices and beliefs to satisfy people’s need for miracles in a dangerous world. And because they tolerated and encouraged magical practices, they succeeded in redirecting loyalty and commitment from pagan gods to Christian saints.
The choice of topics needs to be broadened to reflect more of the long-lasting issues that leaders, as well as the leadership research establishment, confront. However, one side of the establishment has continued to complain about the substantive and methodological narrowness of research in the field. The other side has been equally vociferous about the need for more rigor and discipline. Both types of investigators make contributions, particularly when they can join forces or pursue the expansion of horizons from a disciplinary or interdisciplinary basis. New lines of investigation should be expected and welcomed. For instance, Morgeson (2005) demonstrated the utility of examining the impact of novel and disruptive events on self-managing teams. This was followed shortly by Morgeson and DeRue (2006) demonstrating the impact of criticality, urgency, and duration on the disruption of teams, and influence on the intervention by the team leader. The future should see more efforts to develop understanding of how leadership is affected by events. The future should also see more research on developing, in both leaders and followers, positive psychological capacities such as hope, optimism, confidence, and resiliency (Anonymous, 2006).
Some collectively disparage the thousands of research studies of leadership.13 But T. R. Mitchell (1979) concluded that theory and research were continuing to develop and that much of what had been done was being used in practice: “There was reason for controlled optimism. Yet the challenges are still there for the years ahead.” His words are even truer a quarter of a century later. At the end of the twentieth century, House (1999) concluded that, “there is a substantial amount of available knowledge concerning the exercise and effectiveness of leadership, and there are sufficient number of remaining issues and questions to occupy the time of social scientists for a considerable, and indefinite duration.”
This book should provide an antidote for the arguments of those continuing to bemoan the supposedly unknowable, elusive, mysterious nature of leadership. I expect and hope that I have been able to show that a considerable body of theory, method, evidence, and understanding is available about how, why, and under what conditions leadership, in all its rich variations—implied, observed, described, and/or evaluated—energizes and exerts its influence—sometimes for the worse, more often for the better.
In 1967, I looked ahead to the year 2000 and forecast accurately that organizations would be more flexible, that providing challenging work would be a more common practice, and that online computers would greatly facilitate management. Sensitivity training would be more proficient; and knowledge would be more important to management. There would be increased attention to the behavioral and social sciences and a decline in middle management (Bass, 1967, Wilcox & Bush, 2003). In 1998, I speculated about the U.S. Army in 2025 (Bass, 1998). In 2002, I looked ahead to 2034 and suggested what future leadership research and applications might be like. Many of these predictions are noted next (Bass, 2002).
Methodological Trends. More studies will combine data from individual, team, and organizational levels for within and between analyses (WABA). Old fads will decline, but new ones may take their place, thanks to newspaper and magazine articles, journals and books on the subject. The fads are well known by managers. As noted earlier, academic researchers are now often the source of the information, so that we can expect an increasing coalescence of the fads and scientific literature (Gibson & Tesone, 2001; Gill, 1992). With the increase in more reliable and valid measurements, mathematical operations and model testing will increase in application. Organizational charts will describe networks of leadership, communications, responsibility, and accountability. We will have seriously underestimated developments in information technology, nanotechnology, biotechnology, and electronic brain and body scanning applicable to understanding leaders’ and followers’ cognitions, behavior, and perceptions. Genetic profiling and hormone levels will have been introduced into the study of leadership. Artificial intelligence will assist managerial decision making. More longitudinal studies contrasting and combining economic, social, and behavioral studies will have been completed.
Substantive Trends. Coaching, mentoring, and individualized consideration will have become routinized, much of it available for delivery online. Most training will be delivered from Web sites, the Internet, and intra-nets. The “not invented here” syndrome will disappear in a global economy. Most industries and businesses will be dominated by a few gigantic multinational corporations. Purely transactional bureaucratic organizations will have given way to more transformational ones. In the developed world, women will be in the majority in legislatures, government, and business management and ownership. They will foster more networking, relationships, and concerns for social justice, equity, and fairness. In the United States, more Hispanics, Blacks, Asian Americans, and American Indians, as well as those of southern and eastern European ancestry, will have increased proportionately in senior management. Second careers for leaders after retirement will be commonplace. Working to the age of 85 will not be unusual.
Ethical codes will be constructed empirically, based on experience with borderline cases and judgments about them by managers, ethicists, and workers. Changes in the environment will be mirrored in changes in the organization. At all levels, adaptability will be essential. Leaders will be prized for their innovativeness, responsiveness, and flexibility and their greater frequency of transformational leadership. Improvements in health and medical care will make diseases such as cancer, Alzheimer’s, stroke, and heart disease less likely to disrupt careers acting as Robert Crowley’s “wild cards in history.”
As Aelfric, a schoolteacher and abbot (c. 955–1020), said, “We dare not lengthen this book much more, lest it … stir up men’s antipathy because of its size” (Lacy & Danziger, 1999).