Effective leadership is influenced by the requirements of the tasks faced. But the same individuals tend to emerge as leaders as the tasks change. There seems to be a generalized competence for task leadership. Nevertheless, the competence needed for effective leadership differs from one task to another. The tasks faced by the group, the organization, and its individual members affect and are affected by the leadership. The leadership roles required by the tasks have consequences for the outcomes of the members’ satisfaction, the group’s productivity, and the organization’s performance. In the same way, the leadership involved in the tasks to be accomplished is linked to the technology used. Doing business is shifting from making and distributing products to providing services. Work is being redefined, calling for “constant learning, more higher-order thinking, and less ‘nine-to-five’ ” limits on concerns about the tasks to be accomplished (Bedone, 1993, p. 39).
Singled out for special attention has been the extent to which the task is structured or unstructured and the members’ roles to accomplish the task are clear. This, in turn, is expected to moderate the leader’s directive or participative style and its effects on outcomes. By 1986, at least 48 reasonably rigorous laboratory and survey efforts to verify this path-goal theory had been reported (Indvik, 1986b). However, before I conclude this chapter with a review of path-goal theory and its validity, I will first examine the more general issues of the extent to which various requirements of tasks and technologies have been linked with leadership and its effects. Ordinarily, technology facilitates improvements. The ability to make such improvements in products and services are important in determining whether firms remain competitive (Bedone, 1993) and government administrations and NGOs remain effective.
Life and work have been transformed in the postindus-trial world by advances in technology. For instance, the telecommunications industry has swung full circle. Early in the twentieth century, AT&T’s CEO Henry Vail set the company’s major goal as a regulated monopoly to provide customer service. Cost was secondary. Profits were regulated. With its breakup in 1984 into Baby Bells and the advent of the Internet, the telecommunication business became driven by the competitive marketplace and technology. Now quality and convenient customer service are not only seen as a good in their own right but provide important, if not most important, competitive advantages as a shrunken AT&T now competes with its former subsidiaries.
The Associated Press reported in mid-April 2008 that AT&T, the nation’s largest telecommunications provider, planned to cut about 4,600 jobs, or 1.5 percent of its work force. Most of the layoffs will be among managers, particularly in wire-line operations, including local service for large corporate customers. The company also reported its shares for the first quarter of 2008 had fallen.
A Wake-up Call. In 1956, Leavitt and Whisler predicted that by the 1980s technology would replace middle-management decision making. Middle management has indeed shrunk in numbers. In successfully coping with the change in the past several decades, the telecom firms must have adaptable cultures. Firms that were slow to adapt in the 1980s were Texaco, Coors, and General Motors. Texaco was seduced by its dominant position in the market and its prior successful growth and profits. Coors did not place enough value on its customers, shareholders, and employees. General Motors did not promote managers who showed too much leadership. On the other hand, firms with adaptable cultures such as those of Wal-Mart, PepsiCo, and Hewlett-Packard more quickly made the transition into the competitive 1990s. They conceived of leadership as an engine of change. They stressed the value of addressing constituents’ needs. They introduced adaptable new systems and favored adaptable managers.
Optimal Use of Technology. Simply regarding technology as a black box to replace personnel does not provide for optimization. Full automation is not usually the answer. What is to be sought is optimization of the human-technological interface. Sometimes there is too much human intervention with automated systems, as in the case of the automated thermostatic controls on the German cruiser Prinz Eugen being taken in 1945 to Kwajalein as an H-bomb target. Fearful of blowing up the boilers, one of two U.S. Navy cardinal sins, the American captain assigned one seaman to monitor the thermostat and a second seaman to monitor the first seaman. This resulted in continuing overcorrection.
Also suboptimal is the introduction of overengineered information systems without sufficient regard for the human user. For example, fully automated telemarketing and teleinterviewing may generate a lot of annoyed customers and poorer survey response rates than do human callers. Nonetheless, the efficiency of the human callers can be enhanced with automatic dialing, automatic recording, keyboard tallying, prompting, and so on. On occasion, Luddites may have had the right idea about tech nological advances for the wrong reasons. The costs of technological advancement may be greater than the anticipated gains. This is particularly so if we look for the hidden costs and unintended consequences of technological advances. Automation brings many other problems with it. For example, automation complacency may set in. A systems operator responsible for monitoring automated equipment as well as manual tracking will focus attention continuously on the tracking display and “look at but not see” the automated displays, like an inattentive listener at a dull lecture with blank stares on their faces.
Intellectual Stimulation and Dealing with Resistance to Change Needed. Some, but fewer, middle managers sufficiently educated in information technology (IT) are in place to implement systems of service that optimize the human and IT capabilities. The required mix of human and IT differs in each situation. Some, such as in the case of calling card dialing, may work best with full automation; others, such as handling an irate customer, may require a fully sympathetic human counselor. In two ways middle managers are challenged by the need for effective use of IT and human resources. First, they can solve the problems of optimization of IT and personnel by being intellectually stimulating as transformational leaders. Second, they can systematically review with their team and colleagues the obstacles and resistances to change and how to overcome them. They can prepare their colleagues to be ready for continuing new applications and improvements in IT.
To intellectually stimulate their direct reports, leaders and managers can reformulate the specific problem; turn to metaphors or analogies; imagine alternative states; widen, shrink, or split context; or uncover and challenge hidden, deeply rooted opinions and assumptions. They can work with direct reports to avoid premature conclusions, match competencies with subparts of the problem, and play devil’s advocate. They can do much the same with colleagues and supervisors. To deal with resistance to change, they can answer such questions as: Is the need to change perceived, and by whom? Who is the sponsor? How ready is the person, unit, or organization for change? What motives underlie resistance? How will change be implemented, supported, and maintained? Are leaders and managers engaged in continuous learning to prepare to deal with the technological imperatives they face? What needs to be done to shape the vision of a business’s future state? How much have middle managers and others bought into the new technology? How will career interests be aligned with the organization’s vision of its future state? (Bass, 1999).
The requirements of tasks affect whether a leader is needed, who emerges as a leader, how the leader behaves, and what kinds of leadership behavior result in the better performance and greater satisfaction of the followers. Early on, the evidence was mixed about how requirements of the task situation determined the traits that distinguish those who are chosen to lead. Some of the earliest research on leadership in small groups was carried out in the Soviet Union by Chevaleva-Ianovskaia and Sylla (1929), who observed that no leadership arose in spontaneous preschool groups unless special problems occurred. Caldwell and Wellman (1926) showed that the basis of choice varied according to the activities for which leaders were picked by their junior high school classmates. For example, physical abilities determined the selection of athletic leaders. Nevertheless, among the students who were studied before 1950, scholarship was high among chosen leaders in all designated situations. Dunkerley (1940) found that college women who were selected as intellectual leaders were superior in judgment, initiative, and intellectual ability and those who were picked as social leaders were superior in dress and appearance. Those who were chosen as religious leaders were least neurotic, while those who were chosen as social leaders were most neurotic. Again, some consistency was found in a leader’s performance in different situations, but there was also systematic change in what happens when a leader is transferred to new assignments and tasks.
Stogdill, Shartle, Scott, et al. (1956) studied 20 naval officers who were to be transferred to new positions and the 20 officers whom they were to replace. After several months in their new positions, the transferred officers were found to have shifted their patterns of work performance, but not their interpersonal behavior, to resemble the patterns of the officers they had replaced. In other words, the officers tended to transfer their patterns of interpersonal behavior from one assignment to another but changed their patterns of work performance in response to the task requirements of the new assignments. In a similar way, the individual who emerged as the leader in one group tended to become a leader when placed in other groups, particularly if the different groups were performing similar tasks. Boyatis (1982) conceived such individuals as having the competencies to lead others—the necessary abilities, motives, drives, and behaviors. A change in task may permit new leaders to emerge.
Rotation Experiments. Barnlund (1962) rotated group members through a set of different tasks and through groups with changing memberships. The highest degree of leadership transferability occurred between literary and construction tasks and the lowest degree between coordination and mathematical tasks. Barnlund attached too much importance to changes in both task and group membership as to who emerged as leader; as Kenny and Zaccaro (1983) reported when they reanalyzed Barnlund’s data, 49% to 82% of the variance was attributable to individual consistencies and not as much to changes in tasks and group members. The correlation was .64 in the ranking in leadership across the different situations. Borg and Tupes (1958) and Blake, Mouton, and Fruchter (1954) also reported consistency of behavior in the same leader performing in different groups with varying tasks. Evidently, leaders tend to change certain aspects of their behavior in response to changes in the demands of the group’s task.1
Relevance of Ability to Task. One can think of some competencies such as effective listening linked to leadership in decision making (Johnson & Bechler, 1998), which would correlate with leadership in many different situations. And so, while we find the same individuals emerging as undergraduate leaders in different task situations, we also find that the emergence of leadership is correlated with how much the emergent leader’s abilities are relevant to the tasks the group faces. Different tasks call for different abilities, and the leaders who emerge have different competencies that are relevant to the requirements of the different tasks. Thus, Carter and Nixon (1949a) performed a complicated experiment in which the leadership performance of 100 high school boys was measured by teachers’ ratings, nominations by students, school activity records, and observers’ ratings in three group tasks. A seven-hour battery of tests was also administered. It was found that the boys’ scores on the mechanical ability test coincided with leadership in mechanical tasks on all criteria. Scores on the numerical test and for persuasiveness were correlated with leadership in intellectual tasks, while the scores on the work fluency and clerical tests predicted leadership in clerical tasks on all criteria. Scores on the reasoning test were positively related to leadership in all tasks on all criteria, while the scores on the musical interest test were negatively related.
Clifford and Cohn (1964) described how different attributes of group members, according to colleagues, correlated with nine different leadership positions. Nominations for the role of planner were significantly related to having ideas and being smart, friendly, liked, empathic, and a good influence. Nominations for swimming captain were significantly correlated with being good at swimming and being a good influence. A different pattern of characteristics was associated with each role, but none of the attributes was significantly related to being chosen for the role of banquet chairman.
Effects of Different Tasks. Different tasks had different effects. Hemphill, Pepinsky, Shevitz, et al. (1956) observed the effects of task ability on leadership processes by providing different amounts of advance information to different members of a group. Outcomes were different with different types of tasks. Individuals who were given task-relevant information before an experiment scored higher in attempted leadership in assembly and construction tasks but not in strategic and discussion tasks. Kellett, Humphrey, and Sleeth (2002) noted that some individuals are recognized as leaders because of their mental abilities and competence in dealing with complex tasks. Others are recognized as leaders because of their empathic ability, their emotional intelligence as observed in their socioemotional competence to deal successfully with problems of relationships among people. Still others are recognized as leaders for their competence in both tasks and relationships (see Chapter 19).
Competence Relative to Subordinates’ Abilities. The competence of the leader for the task and how the leader behaves as a consequence are moderated by the subordnates’ ability to handle the task. If supervisors believe their subordinates have the requisite skills, the supervisors are more likely to be consultative, participative,2 or delegative (Heller, 1969a), and so they should be (Hersey & Blanchard, 1977). It seems equally true that if subordinates do have the skills and interest for the particular tasks of consequence, their productivity and satisfaction will be greater if supervisors permit them to participate partially or fully in the decision-making process. Full participation will be less than optimum when subordinates do not have the requisite skills and interest.
Not only do group tasks affect which abilities are important for leadership, they are also linked to personality factors that seem to promote the choice of leaders. Wardlow and Greene (1952) reported that adolescent girls with high scores on tests measuring their adjustment to school and home and health problems were preferred by peers who were working with them on an intellectual task. Megargee, Bogart, and Anderson (1966) asked pairs of participants, one with a high score and one with a low score on a dominance test, to perform two different tasks. When the instructions emphasized the task, the dominant participants did not emerge as leaders significantly more often than did their partners. However, when leadership was emphasized in the instructions, the dominant members emerged as leaders in 90% of the pairs.
Effects of Task Structure. B. B. Roberts (1969) administered a battery of personality tests to leaders and followers who were studied under different conditions. Concrete and practical-thinking persons were chosen as leaders in structured tasks by all group members, but they were chosen only by the practical and concrete choosers for unstructured tasks. The theoretically oriented members were chosen as leaders in unstructured tasks by abstract, theoretical followers. Although more frequent leader initiation and direction may actually be seen in structured task situations (Stech, 1981; Wolcott, 1984), they also may be superfluous when task structure is already sufficient. More leader task orienting and objective setting may be necessary for effectiveness when tasks are ill defined (Taggar, 2001; Bain, Mann, & PirolaMerlo, 2001). But when innovation and creativity are sought and tasks are ill defined, overcontrolling leaders with too much initiation may stifle the autonomy conducive to creativity (Trevelyan, 2001). There is some optimum direction and participation for maximizing performance with ill-defined problems and problems calling for creative thinking (Mumford, Scott, Gaddis, et al. 2002).
Effects of Different Task Requirements. W. W. Burke (1965) found that followers who worked under a socially distant leader rated their groups as more satisfying and productive in a decision-making task than on a code-solving task. Dubno (1963) obtained results indicating that the effectiveness of groups was higher when quality rather than the speed of performing a task was emphasized by their leaders and when the leaders reached decisions more quickly than did the other members. But Hoyt and Stoner (1968) failed to confirm that the risky decisions made by groups were due to the leadership of risk-prone members. With leadership effects held constant, group discussions still produced group decisions that were riskier than the mean of the individual decisions of members. Nevertheless, as will be detailed later, the risk behavior of the leader is likely to depend on his or her speed in reaching decisions, on the organization’s stage of development, and on whether the group is at an early or late phase in problem solving. The task requirements of newly formed groups and organizations are different from those of mature groups and organizations, and what is required of their leaders at different stages is also likely to be different. Furthermore, just as the patterns of interpersonal behavior do not change substantially but the patterns of work performance do change when leaders are transferred, the leader’s initiation of structure and directive leadership are more likely to shift with the changing demands of tasks than are the leader’s consideration and relations orientation (Ford, 1981).
Indvik (1986b) surveyed 467 nonacademic staff at a university and found that followers’ perceptions of their ability and preference for structure systematically affected what kinds of leadership behavior, under various task circumstances, generated elevated expectations, satisfaction, and meritorious performance. For followers who preferred structure, participative leadership messages were most effective when formalization of the organization was low and the importance of the work group was high. For followers who did not prefer structure, participative leadership messages enhanced their satisfaction when the task was highly structured, the importance of the work group was low, and the formalization of the organization was high. (More will be said later about leading creative followers.) Again, Keller (1992) found among 462 personnel in three R & D departments led by charismatic and intellectually stimulating project leaders that project quality was higher for research projects than for developmental projects.
Cynicism. Reichers, Wanous, and Austin (1997) conducted interviews, surveys, and follow-ups with approximately 2,000 unionized workers and their managers. Forty-three percent of the hourly workers and 23% of the managers were classified as cynical about change. (Many previous change efforts had been unsuccessful.) The results pointed to the detrimental effects of followers’ cynicism about leaders’ efforts to make changes in the organization. The cynics were likely to be pessimistic about the success of changes and accused those responsible for the change efforts to be lazy and incompetent. The cynics felt that they lacked information about the changes, respect from supervisors and union representatives, and opportunities to participate in decisions making. They had experienced failures in previous change efforts and were personally disposed to be cynical. They lacked commitment, satisfaction, and motivation to make the changes.
Followers’ Competence. There is an obvious correlation between followers’ competence and their effectiveness unless they are in conflict with one another about objectives, have contrasting values, and are poorly led and poorly motivated. However, an extremely competent follower may usurp the leadership, “take control and then begin to undermine the influence of others. The work climate is poisoned, and morale declines. Cooperation turns to competition, then ill-will, and then into subtle forms of sabotage” (Quinn, 1996, p. 116).
Task interdependence has been found to be an important determinant of whether leaders should be appointed or elected. Basik, Gershenoff, and Foti (1999) reported that when members’ tasks were highly interdependent, team performance was high if team leaders were elected, but when members’ tasks were low in interdependence, the teams did better with assigned leaders. Valenzi, Miller, Eldridge, et al. (1972) reviewed previous research on the impact of leaders’ behavior on task requirements. They concluded that task interdependence, structure, routineness, complexity, and intellectual (but not manipulative) requirements systematically alter the amount and kind of leadership that is most effective. On the basis of this review, a survey was conducted. Using stepwise regression analyses, Bass, Valenzi, Farrow, and Solomon (1975) found, for 78 managers described by their 407 subordinates, that tasks with clear objectives were seen to result in more direction and consultation by the managers. Routine tasks were associated with less participative leadership. More complex tasks correlated with negotiative leadership and more frequent delegation. Delegation was also reported to be more frequent among executives if their subordinate managers had to engage in planning, coordination, and other managerial activities. I confirmed more frequent leader direction in an unpublished follow-up of the effect of structure on 340 managers and 1,300 subordinates. As noted just below, more frequent direction may occur with strong structure, when actually direction is less needed.
Leaders of groups and individuals with structured, as opposed to unstructured, tasks have been considered in a great many studies.3 For instance, Hill and Hughes (1974) reported that leaders displayed more socioemotional behaviors, both positive and negative, when subordinates performed unstructured as opposed to structured tasks. Widely used to examine the effects of structure is House and Dessler’s (1974) measure of task structure, a ten-item questionnaire on which respondents describe the extent to which their tasks are simple, repetitive, and unambiguous.
Task structure varies in how much is specified, is certain, and has to be done. For example, Lawrence and Lorsch (1967b) assumed that task structure went from the lowest to the highest as one moved from fundamental research to applied research, to sales, and finally to production. In six organizations, they found that production personnel (whose work was most specified and certain) and fundamental research personnel (whose tasks were least specified and certain) both preferred task-oriented leaders. Members in the sales subsystem (which had moderately specified and certain work) preferred more interpersonal, socially oriented leaders. Structure may be strong or weak, tightly or loosely arranged, clear or ambiguous (Mischel, 1977). When it is strong, tight, and clear, everyone sees the situation in the same way. It induces uniform expectations and provides clear incentives for appropriate responses. Less guidance and directive leadership are needed. When structure is weak, loose, and ambiguous, there is more acceptance of guidance, di rection, and charismatic leadership (Shamir & Howell, 1999).
Fiedler’s Task Structure. A tenet of Fiedler’s (1967a) contingency model was based on the effect of task structure. For Fiedler, task structure creates a more favorable situation for the leader. A task-oriented leader will be more effective when there is either a great deal of task structure or very little task structure. A relations-oriented leader will be more effective if task structure is moderate. Although the conclusions remain controversial (see, e.g., Graen, Alvares, Orris, & Martella, 1970; Graen, Orris, & Alvares, 1971), considerable empirical support for this proposition was presented in Chapter 19. However, the effects of task structure may be much weaker than other contingent situational elements that favor the leader (W. Hill, 1969a).
Some Effects on Leadership Behavior and Outcomes. Wofford (1971) observed that unstructured tasks elicited more achievement-oriented and organizing managerial behavior than did structured tasks. Lord (1975) found that the degree of task structure in a situation was negatively related to the occurrence of facilitative leadership behavior. An unexpected curvilinear effect emerged. Instrumental (task-directed) leadership was most effective with tasks that were moderate in structure. Shaw and Blum (1966) found that structured problems were better served by directive supervision. Thus, when the problem on which five-person groups were working was highly structured so that clear procedures could be followed, directive supervision led to quick results. Yet when leaders initiated structure a great deal, as was seen in a study by Badin (1974) of 42 work groups in a manufacturing firm, their effectiveness was reduced if the groups were already highly structured. The correlation of effectiveness and Initiating Structure was −.56. But in the less structured of the 42 groups, the effectiveness of the groups correlated .20 with the extent to which the first-line supervisors initiated structure.
Filley, House, and Kerr (1979) found that task structure moderated the relationship between participation and performance and between participation and the attitudes of subordinates. Participative leadership had no effect or was contraindicated when tasks were machine-paced, mechanized, and highly structured in other ways. But when tasks were unstructured, the effects of participative leadership on both the attitudes and productivity of subordinates were consistently positive. Hanaway (1985) examined the effects of the uncertainty of tasks on the initiating and search behavior of 18 upper-level and 32 lower-level administrators of school districts. The lower-level administrators were observed to be more reticent to take actions when conditions were uncertain, but the results for the upper-level administrators were less clear.
Role Clarity. Generally, people, particularly those with a great need for structure, prefer to work in clear task settings. Valenzi and Dessler (1978) showed that among 284 employees in two electronics firms, satisfaction was uniformly high when role clarity was high. The leaders’ consideration promoted even more satisfaction. In addition, Benson, Kemery, Sauser, and Tankesley (1985) found that a high need for clarity was related to low job satisfaction among 370 university employees above and beyond the effects of their role ambiguity on the employees’ dissatisfaction.
The determination of perceived role ambiguity or perceived role clarity is a way to discern how much structure exists in a group. Kinicki and Schriesheim (1978) studied the role clarity of 173 freshmen in 16 classes. Role clarity was measured by Rizzo, House, and Lirtzman’s (1970) scale dealing with clear objectives, responsibilities, expectations, and explanations about what has to be done. They found that students were more satisfied with relations-oriented teachers, particularly when role clarity in the situation was low. But the students were more productive with directive teachers when role clarity was low, not when it was high. As Siegall and Cummings (1986) demonstrated, satisfaction with supervision is enhanced if a subordinate’s role is initially ambiguous and if the supervisor contributes to clarifying the role by issuing instructions and directions. Fulk and Wendler (1982) found, among 308 clerical and managerial employees, that role clarity was associated with nonpunitive, task-oriented leadership and contingent reward. Such leadership was satisfying. But role clarity could also emerge as a consequence of arbitrary and punitive leadership behavior that was dissatisfying. However, Schriesheim and Murphy (1976) failed to find that role clarity moderated the relationship between any kind of leadership behavior and the satisfaction and performance of subordinates. To help further in the understanding of the above evidence, later in this chapter, when the path-goal theory and the specific tests to validate it are reviewed, more will be said about the extent to which structure and role clarity affect the impact of leadership on performance and satisfaction.
Clarity of Objectives. Nagata (1965) observed that groups with goal-relevant tasks enabled leaders to exercise more influence on the members than on members in groups in which the tasks were not relevant to the goal. To some extent, clear objectives may substitute for structured relationships or clear role relationships in getting a job done, particularly if little coordination is required. Multiple regression analyses by Bass, Farrow, and Valenzi (1977) for 250 managers and their 924 subordinates suggested that the effectiveness of work groups was significantly greater if the mangers had clear objectives. Furthermore, when these organizational members were competent and motivated to attain their objectives and could operate independently or in cooperation with one another, the effectiveness of their performance was strongly associated with the clarity of their objectives and the adequacy of their resources.
Error criticality is the extent to which mistakes at work produce negative outcomes. In health care, error criticality is high for those in nursing and pharmacy departments, for their mistakes can lead to patient injury or death. Their motivation to perform well is not increased or decreased by the quality of their supervision or by job design (autonomy and variety). On the other hand, error critically is generally low for those in hospital administrative units. Compared to nursing or the pharmacy, mistakes in administrative work ordinarily are not life-threatening. Administrative employees’ motivation to work is more strongly affected by the quality of their supervision, by their morale, by their job autonomy, and the job variety in their jobs, according to a study by Morgeson (2005) of 189 hospital employees.
One’s felt autonomy and discretion to do one’s job appear to be complex perceptions that are affected by various factors, such as the organization’s technology and the frames of reference shared with coworkers, as well as the style of leadership employed. Presumably, leaders who initiate a great deal of structure by definition, reduce their subordinates’ autonomy. But leaders’ consideration should enhance subordinates’ feelings of autonomy. Ferris and Rowland (1981) suggested that the leaders’ initiating structure and consideration systematically affect the subordinates’ perceptions of autonomy in this way, which, in turn, may or may not contribute to the subordinates’ performance.
The same leadership will have different effects depending on whether subordinates’ autonomy is high or low. Johns (1978) found a correlation of .29 between leaders’ initiation of structure and subordinates’ job satisfaction when 232 union employees reported that their autonomy was high, and a correlation of .01 when they indicated that their autonomy was low. The leaders’ consideration and the subordinates’ job satisfaction correlated only .20 when the subordinates’ autonomy was high but .52 when it was low. More than 1,300 subordinates’ descriptions of 340 managers indicated that if managers were viewed as delegative and negotiative, their subordinates felt in possession of more discretionary opportunities (Bass & Valenzi, 1974). Managers were also less likely to be seen as directive under such conditions (Bass, Valenzi, Farrow, & Solomon, 1975).
Innovation and Creativity. Discretionary opportunities enhance innovation and creativity. Such opportunities are likely when leaders and subordinates have good relations with each other. Scott and Bruce (1994), Scott and Bruce (1998), and Tierney, Farmer, and Graen (1999) reported positive correlations between the quality of leader-member exchange (LMX), task innovation, and creativity of managers, scientists, and engineers. Amabile and Gryskiewicz (1987) interviewed 120 R & D scientists to elicit two critical incidents from each of them, one to illustrate the occurrence of high creativity, the other to provide an instance of low creativity. Consistent with what has been said about participative leadership and creativity, in describing the incidents of high creativity, more than 74% of the interviewees mentioned freedom to decide what to do and how to do one’s work, a sense of control over one’s work and ideas, freedom from contraints, and an open atmosphere. This finding agrees with conclusions reached by Andrews and Farris (1967) and others, although Amabile and Gryskiewicz hastened to add that there is a limit to the amount of such freedom, for it must be bounded by the team supervisor’s coordination of the team’s overall efforts (Pelz & Andrews, 1966b). In presenting these incidents of high creativity, one-third to one-half the 120 scientists revealed that their leaders’ enthusiasm, interest, and commitment to new ideas and challenges encouraged their creativity.4 Supportive leadership was seen in the willingness to take risks, provide recognition for success, and clarify what was needed. The leaders accepted failure without destructive criticism and avoided excessive evaluations. Such leaders did not dwell on the status quo. Keller (1989) reported that the productivity of 30 project research teams correlated .57 with the intellectual stimulation of their leaders but close to zero with the same leadership in 36 project development teams. According to Allen, Katz, Grady, et al. (1988), the performance of 181 project teams was greater when their leaders kept them current in the science and knowledge they needed to know. Frischer (1993) found that when 38 managers of new product development empowered and gave a sense of responsibility to their subordinates, the subordinates perceived a greater innovative climate. In agreement, Judge, Gryxell, and Dooley (1997) noted that R & D managers enhanced innovation by giving their subordinates operational autonomy, personal recognition, cohesiveness, and a continuity of slack resources.
Jobs can require that employees carry out the same few tasks in a repetitive cycle or may involve a greater variety of tasks that are more variable in sequence. Optimal performance by leaders depends on whether the work involves uniform, recurring, repetitive tasks or considerable variability (Valenzi, Miller, Eldridge, et al., 1972). In their classic study of workers on the assembly line, Walker and Guest (1952) emphasized the extent to which supervision was likely to be more directive when the tasks to be performed were extremely routine. Likewise, Bass, Valenzi, Farrow, and Solomon (1975) reported more directive leadership as well as less supervisory delegation in work groups that were carrying out routine tasks. Similarly, Ford (1981) showed that in 35 departments in a book-publishing firm, a bank, and a university, the routineness of work uniquely correlated .36 with the extent to which the departmental manager initiated structure but only .05 with his or her consideration as a leader.
In a study involving 16 departments in 10 organizations, R. H. Hall (1962) distinguished between uniform, easily routinized, standardized activities and nonuniform, difficult-to-routinize, creative activities. He found that departments and hierarchical levels whose activities were more nonroutine were also less bureaucratic than were those departments and levels that were oriented toward routine activities. In nonroutine situations, the atmosphere was more personal, had less hierarchical emphasis, and required fewer procedures and regulations. Consistent with all these findings, Heller and Yukl (1969) found that production and finance managers (who supervised more routinized work) tended to use centralized decision making, while general and personnel managers (who supervised less routinized work) were more participative.
The linkage of routineness to the greater initiation of structure by leaders is consistent with the hypothesis, verified by Kipnis (1984) for several levels of managers in Australia and the United States, that employees who work on routine tasks are likely to be undervalued by their supervisors. Despite this finding, Jiambalvo and Pratt (1982) reported that in four large accounting firms, considerate leadership behavior by senior accountants increased the involvement in tasks among staff assistants who performed relatively simple tasks more than it did among those who performed relatively complex tasks.
Variety. Hackman and Oldham’s (1975) Job Diagnostic Survey included job variety as an important variable that was likely to relate to the motivation to work. Using ratings based on the survey, Johns (1978) showed that with job variety, the leaders’ initiating structure generated greater job satisfaction and fewer intentions to quit among subordinates. Without such variety, the leaders’ initiation was unrelated to the subordinates’ satisfaction and increased their intentions to quit. However, the leaders’ consideration was strongly associated with satisfaction, regardless of whether the job was varied or routine. Yet only when variety was absent was the leaders’ lack of consideration associated with the subordinates’ intentions to quit. In contrast, Brief, Aldag, Russell, and Rude (1981) found, in an investigation of police officers, relatively little of the expected effects of the variety of job skills on the favorability of police officers’ attitudes toward the citizenry.
In a study that distinguished working on uniform and nonuniform tasks, Pelz (cited in Litwak, 1961) found a higher correlation between the motivation to work and productivity when those engaged in nonuniform tasks were permitted by their supervisors to make their own job decisions. But for those involved with uniform tasks, there was a higher correlation between motivation and productivity when freedom to make decisions was restricted. Nonetheless, Katz, Maccoby, and Morse (1950) found that supervisors of high-producing sections were significantly more likely to provide general rather than close supervision, even though they were supervising routine clerical work in a life insurance company. However, in a subsequent study of less routine railroad work (Katz, Maccoby, Gurin, & Floor, 1951), little difference was found between the closeness of supervision by supervisors of high-and low-producing sections. In these and the many other related studies that followed,5 R. Likert could find no diminution in the utility of participative (System 4) leadership in routine jobs as compared to those with more variety. But Griffin (1980) suggested otherwise. He contrasted the leadership-outcome relations among employees who had “high-scope” tasks (varied, involving, and autonomous) and those who had “low-scope” tasks (simple, routine). In his first survey of 129 employees, the scope of the employees’ tasks did not correlate with the leadership style of their supervisors. However, for a subset of employees with high-scope jobs but a low need for personal growth, directive supervision subsequently resulted in greater satisfaction. On the other hand, employees with low-scope jobs were more satisfied with their subsequent supervision if their supervisors were more supportive and practiced more management by exception. Such supervisors were described by such statements as “My supervisor doesn’t bother me as long as I do a good job” and “My supervisor leaves me alone and lets me work.”
Low-Scope Tasks and Passive Leadership. Although passive management by exception has generally been downgraded by subordinates as a satisfying style of supervision,6 the results suggest that employees in low-scope jobs prefer and feel better when their supervisors practice management by exception. This suggestion was confirmed in a survey of 195 full-time employees by Algattan (1985). When the scope of the tasks that subordinates performed was low and the subordinates had little need for growth, leaders’ maintenance of the status quo was more positively related to outcomes than was more active directive, participative, or task-oriented leadership.
Technologies are simple if they can readily be understood and communicated between experts. Complex technologies cannot. As they move from simple to complex technologies, leaders need to shift from gatekeepers of information to facilitating and insulating information networks. These networks have to create knowledge, not just transfer information and knowledge (Kash & Ry-croft, 1996, 1997). Bell (1967) viewed the complexity of tasks in terms of the degree of predictability of the demands of the work, the amount of discretion exercised, the extent of responsibility, and the number of different tasks performed. When informal groups face complex tasks, leaders emerge who have the cognitive abilities to handle the tasks and to help others to do so. They may also have the socioemotional skills to deal with the others facing the complex tasks (Kellet, Humphey & Sleeth, 2002). Among supervisors in a hospital, Bell (1967) found that the more complex the subordinates’ tasks or the supervisor’s job, the narrower the supervisor’s span of control. But the complexity of tasks did not influence how closely the subordinates were supervised. Barrow’s (1976) study indicated that leaders exhibited more task orientation when faced with more complex tasks, but the complexity of the tasks did not affect their tendency to be punitive. However, Cuthbertson (1982), in a survey of 175 subordinates and their 25 supervisors in a central office of a school district, found more directive leadership than she expected in relation to the complexity of the tasks involved. Among 61 to 68 telecommunications units, Osborn and Hunt (1979) obtained contrasting patterns of correlations with leadership, the structural complexity of the units led, and the difficulty of the units’ tasks. They found that the leadership that was required was unrelated to the complexity of the units led or the difficulty of the tasks. However, the actual amount of support and role clarification by the leaders, as expected, correlated between .24 and .31 with structural complexity of the units but unexpectedly between −.30 and −.42 with the difficulty of the task.
Barrow (1976) observed, in a simulation using 120 male college students as leaders, that more initiation of structure was caused by increasing the complexity of the task but autocratic behavior was generated more by workers; poor performance than by increasing the complexity of the task. Considerate leadership was evoked by improvements in the workers’ performance. Relevant to these results, Wofford (1971) found that a personal interaction (relations-oriented) manager was more effective for complex operations. A self-oriented, autocratic manager was more suited to situations with simple work schedules. Hammer and Turk (1985) showed that supervisors were less likely to perform group maintenance activities if they supervised employees who worked with intensive technology, as in a repair shop or an R & D laboratory where tasks were complex and nonroutinized. On the other hand, supervisors of workers who were engaged in long-linked technologies, such as mass production assembly lines, and whose tasks were interdependent were more likely to engage in network activities. In addition, supervisors in the intensive-technology situation felt, to a greater extent, that they had the authority to reward subordinates, whereas supervisors in the long-linked technology settings felt they had more authority to discharge employees and, in turn, were less likely to be seen as experts.
The effects of the increasing complexity of tasks on requirements for leadership are illustrated by the changing military scene. Those who are now being selected to serve as military leaders in the next several decades will have to operate under conditions for which there is less public consensus than existed during World War II. They will be expected to know how to use minimal force in unconventional conflicts in which they will be trying to keep the peace. Fighting, as it continues in Iraq, is intense, lethal, and destructive. For this type of fighting, an understanding of the local values of what is right, good, and important, along with intellectual sensitivity, is particularly important. Officers are required to respond thoughtfully to increasingly ambiguous circumstances. They need to inspire subordinates with a vision of the future that strengthens the subordinates’ loyalty and commitment rather than merely fosters the subordinates’ grudging obedience. Gal’s (1986) forecasts fairly accurately described the task and technology requirements for coalition forces in the Persian Gulf War in 1991 and the wars in Afghanistan and Iraq and their continuing insurgent aftermaths.
Difficulty of Tasks. A factor analysis of 104 different experimental tasks by M. E. Shaw (1963b) disclosed three factors that contribute to the effects of the complexity of tasks: (1) the difficulty of a task (the number of operations, skills, and knowledge required to complete the task); (2) the multiplicity of correct solutions; and (3) the requirements for cooperation (integrated efforts). C. G. Morris (1966a, 1966b) found that as the difficulty of the tasks increased for groups, there was a concomitant increase in leaders’ and members’ attempts to structure answers, propose solutions, and seek evaluations. However, tasks of intermediate difficulty generated the highest frequency of attempts to structure the problem, followed by tasks that were the most difficult. Easy tasks produced the most irrelevant interactions. At the same time, Nagata (1966) found that groups with easy tasks exhibited more role differentiation and permitted leaders to exercise more influence than did groups with difficult tasks. Nevertheless, Bass, Pryer, Gaier, and Flint (1958) observed fewer attempts to lead in groups with easy problems. But it is important to recognize that the difficulty of a task is in the eye of the beholder. Manz, Adsit, Campbell, and Mathison-Hance (1988) surveyed 3,580 managers in a large firm about the hindrances to their performance. Better-performing managers paid more attention to external hindrances, such as inadequate appraisal systems and the absence of opportunities for promotion. Poorer performers focused more on deficiencies in skills, such as the lack of interpersonal or technical abilities.
When it was arranged for students to instruct others on easy and hard tasks in a laboratory setting, the “instructors” used less punishment when trainees performed difficult tasks than when they performed easy ones. A meta-analysis by Tubbs (1986) of 87 studies tested Locke’s (1968) hypotheses on how the difficulty and specificity of goals enhance the speed and quantity of work that subordinates perform. Motivating leaders (in this case the experimenters) were more structuring with specific, difficult goals than were leaders who told subordinates just to do their best or did not tell them anything about the goals. Difficult goals resulted in higher motivation and performance of participants in short-term laboratory studies, but according to the meta-analysis, such results were somewhat less likely to materialize in survey studies of workers in longer-term assignments, perhaps because sometimes assignments might be too difficult, as noted next.
Fast pace, time pressure, and the need to meet deadlines contribute to the difficulty of tasks. For research scientists and engineers, Hall and Lawler (1971) found that the pressure to do high-quality work and to help their company attain its financial goals contributed to the successful performance of their research laboratories. Andrews and Farris (1972) also reported that time pressures experienced by scientists and engineers correlated positively with their subsequent performance. However, pressure that was perceived as unreasonable or excessive resulted in poor or decreased performance. Such excessive pressure could result in the setting of unrealistically high goals (Forward & Zander, 1971). Excessive pressure by a supervisor is likely to be contraindicated when tasks become more difficult with multitasking by operators who are novices and when the pace is conducive to errors because of its speed. Nonetheless, Shamir and Howell (1999) argued that a charismatic leader could emerge and be effective if the followers faced challenging and complex tasks, required intense efforts, and if performance goals were ambiguous and extrinsic rewards could not be linked to performance.
Temporal Complexity. Halbesleben, Novicevic, Harvey, et al. (2003) called attention to the importance of time and timing to giving organizations competitive advantage. Awareness of the dimensions of temporal complexity is an important competency for leading in creativity and innovation. Leaders need to be aware of the time frame of a task, such as the deadline for completing it. They need to pay attention to the timing of required activities. They need to be aware of the limited time spans of processes and events within the time scope of the task. They need to be aware of individual differences in pauses and gaps in the work and to support the simultaneity, sequence, and synchronization of operators, events, and processes. Leaders need to appreciate timelessness, losing one’s sense of time while engaged in creative thinking, which contributes to optimal flow (Csikszentmihalyi, 1997) for creativity and innovation (Sosik, Kahai, & Avolio, 1999).
Uncertainty of Tasks. Wilson and Rhodes (1997) found in a set of experiments that coordination among followers facing uncertainty depended on particular kinds of leadership. According to De Mayer, Loch, and Pich (2002), there are four kinds of uncertainty and the leadership required as a consequence: (1) Variation. A task may vary, causing uncertainty, for example, about the time needed for completion. Here the leader needs to monitor deviations, set limits, or take corrective action, and arrange for buffers at critical points. (2) Foreseen uncertainty. A few known factors may influence carrying out the task in unpredictable ways. The leadership needs to plan for contingencies and alternate ways of completing the task. (3) Unforeseen uncertainty. Some factors influencing the task cannot be predicted. The leadership needs to plan for foreseen uncertainty but also scan for and recognize unexpected influences and arrange new contingent actions accordingly. (4) Chaos. Unforeseen events completely invalidate previous plans and efforts. The leadership needs to reexamine its assumptions, redefine the task, and do contingent planning based on incremental learning, constant feedback, and continual iterations, to gradually select the final approach.
Multiplicity of Solutions. Shaw and Blum (1966) noted that directive supervision was more effective if the problem called for agreement on a single solution, whereas participative leadership paid off when multiple, divergent solutions were needed. Experimental groups of five members each performed three tasks that required different types of solutions. Directive supervision was more effective when the problem called for a single final decision or involved the convergence of judgments into some final product. On the other hand, when the problem required multiple, divergent final solutions, participative approaches were more effective.
Participative leadership is suggested by the assembly bonus effect—groups achieve better solutions to problems with multiple alternative possibilities than does their average member working alone (Bass, 1960; Steiner, 1972). (The group decision may, however, not be as good as the decision achieved by the best member working alone.) This assembly bonus effect occurs unless individual members already have the requisite information to solve the problem alone, for instance, when every member is a professional expert who is highly trained to deal with the same standard types of problems in the same way. There will be less of an assembly bonus if the addition of members produces interference rather than nonredundant information. According to Heller (1969a), the primary reason managers in 15 firms reported using participative leadership was to improve the technical quality of complex decisions. In fact, some form of consultation is mandatory in highly technically oriented organizations, for the available technical expertise does not fully reside with supervisors but is distributed, to some degree, among subordinates.
Required Cooperation and Interdependence of Tasks. Adam Smith, in The Theory of Moral Sentiments, called attention to the paradox he had noted in The Wealth of Nations that the economy was driven by self-interest, yet people also recognized that the pursuit of self-interest required cooperation from others. Conflicting evidence has emerged here. To emerge as a leader in an initially leaderless group, one needs to both compete with and cooperate with the other members of the group (Bass, 1954). Cooperation and the interdependence of task performance among subordinates have been found to be promoted by participative leadership in some research studies. Other investigations have failed to find any effects, and evidence from still other laboratory and survey studies suggests that when subordinates engage in interdependent tasks, directive leadership and initiation of structure by the leader are more efficacious.
O’Brien (1969b) theorized that the equalization of power and participative leadership would be appropriate for tasks that require a great deal of cooperation. A power differential between the superior and the subordinates would be more effective in situations in which subordinates carry out tasks independently of one another. Vroom and Mann’s (1960) results were illustrative. Vroom and Mann studied drivers and positioners in a package delivery company. The positioner’s job required a high degree of interdependence and considerable interaction with coworkers and with the supervisor. The driver’s job involved little interpersonal interaction and considerable independence in work activity. In line with expectations, the positioners favored democratic leaders and the drivers preferred authoritarian leaders.
Bass, Valenzi, Farrow, and Solomon (1975) failed to find statistically significant correlations between leadership styles and the interdependence of tasks of work group members. But larger-scale follow-up analysis by Bass and Valenzi (1974) indicated that more directive leadership and consultative leadership were associated with the interdependence of tasks. Nevertheless, negotiative leadership was greater when members worked independently of one another.
In a survey of 25 departmental managers and their 445 departmental associates, Ford (1981) did not find that the amount of interdependence in the workflow within the departments did not account for initiation of structure or consideration by the managers beyond that already explained by the routineness or uncertainty of departmental tasks. On the other hand, Lord and Rowzee’s (1979) laboratory experiment with four-person teams showed that when tasks required a high degree of interdependence among the subjects, more frequent directive leadership behaviors emerged to develop orientation, plans, and coordination. Consistent with these results, Fry, Kerr, and Lee (1986) found that among 22 high school and college teams in eight sports, winning coaches of highly interdependent sports teams (like basketball) were described by their players as displaying more initiation of structure and less consideration than were winning coaches of sports teams, such as golf, that required little or no interdependence. Coaches of winning teams that were highly interdependent also displayed more initiation and less consideration than did coaches of losing teams that were highly interdependent.
To adequately explain these results requires attention to another variable: the competence of the team members. Directive leadership to clarify roles may be needed more when team members must work in collaboration but the team members are novices. Participative leadership that is focused on commitment becomes more important for high-quality collaborative efforts when the members are already trained and experienced.
Required Interdependencies among Members. Thompson (1967) distinguished among three kinds of interdependencies: pooled, sequential, and reciprocal. In the pooled circumstance, each individual, such as a baseball player at bat, performs alone, but the results have collective effects. In the sequential effect, one person, such as a running quarterback in a football play for whom others will block the opposition, depends on others earlier in a sequence. Reciprocal interdependence occurs when each person must interact with others, such as in basketball. Leadership and management in conditions of pooled interdependence require continuing attention to tactical judgments and the development of individual performers, but cohesiveness may not be as important. In sequential performance, leadership and management must attend more to planning and the preparation of the team. The requirements for performance are tighter and more highly specified and scheduled. Cohesiveness is more important. In reciprocal interdependence, satisfactory mutual adjustments are of the greatest importance and necessitate that the most attention be paid to relationships and continuing cooperation (Keidel, 1984).
Kabanoff and O’Brien (1979) studied leadership when members of a team either had to coordinate their efforts (work on subtasks arranged in an order of precedence) or collaborate (work simultaneously with one another on every subtask). Teams that had to coordinate their efforts were more productive, especially when the leaders were more task-competent. But the leaders’ task competence was irrelevant to productivity in the collaborative task situation. This finding may be explained by a suggestion by Hill and Hughes (1974) that there is a greater emphasis on the leaders’ socioemotional function than on their task function in the collaborative situation. As a consequence, their task competence is relatively less important when collaboration is required.
As was just observed, a distinction that is important for understanding what type of competence will be demanded of a leader is whether socioemotional or task requirements will be emphasized for leadership. This is the most frequent role differentiation that occurs in discussion groups. In groups, when the demands of the task are high, being liked does not contribute much to leadership and socioemotional skills are not highly valued (Slater, 1955). On the other hand, in social and personal development groups, such as therapy groups, sensitivity training groups, social clubs, and gangs, the socioemotional function is emphasized. As was concluded in Chapter 19, in most kinds of groups, both types of leadership usually need to be present. A. S. Miles (1970) reported that student leaders who rated high on both task ability and socioemotional ability were considered most influential. Empirically, the differentiation between the two types of abilities often is not found. For instance, Gustafson and Harrell (1970) observed relatively little differentiation between task and socioemotional roles in experimental groups. Similarly, V. Williams (1965) noted that some types of group structures were able to operate effectively without differentiating task specialists from socioemotional specialists. Kellet, Humphrey, and Sleeth (2002) noted that either or both could be identified as a leader.
To sort out the effects, Olmsted (1954) gave one set of groups instructions that were designed to induce socio-emotional concerns for group processes and the satisfaction of members. The directions given to the second set of groups emphasized the accomplishment of tasks and maintenance of impersonal relationships among members. The most talkative members in the task-directed groups talked longer than did their counterparts in the socioemotional groups, perhaps as the result of a group norm related to the intensity of participation. Task-directed groups tended to develop stable leadership status structures. Members of socioemotional groups continued to jockey for position for a longer time. Olmsted (1955) later found that the amount of agreement among members in discussion groups moderated the effects. In groups that achieved a high degree of consensus on solutions to problems. The highest participator (who was more likely to be a leader than a follower) usually received the highest rating for helping the group meet the requirements of its task. In low-consensus groups, the highest participator was not rated high.
Adam Smith (1775) noted that civilization arose with the division of labor. Instead of the same primitive humans both farming and fishing, one farmed and another fished. Such specialization of energies made for more efficiency. A merchant trader specialty emerged when the farmer bartered or sold his corn to the fisherman through the merchant trader. With the advent of the Industrial Revolution, instead of one craftsman making an entire pin, efficiency was increased by dividing the work in the pin factory into various simpler subtasks, each carried out by a different worker (Herman, 2001). Unfortunately, job simplification, as opposed to crafting a whole product, also brought with it boredom and dissatisfaction, which leaders had to deal with until it was widely recognized in the mid–twentieth century that making the same worker responsible for a whole, more meaningful task, product or process, a sociotechnical design of work, was likely to be more satisfying and efficient. There was less need for coordinating leadership among subtasks carried out by different workers. There has been an increasing interest in designing tasks and work to take account of both the task and the socioemotional requirements.
When technical and social systems are integrated, joint optimization is needed. Bamforth observed that long-wall coal miners became much more productive and satisfied when they were able to return to the earlier, traditional premechanization methods of working as a team instead of on individual assignments as directed by a supervisor (Trist & Bamforth, 1951). Along with Fred Emery (1959, 1967), Trist (1971, 1981) set out the principles of socio-technical systems design. They were consistent with Lewin’s (1939) theories and experiments in group dynamics (Fox, 1995).
The critical dimensions, methods, and principles of socioeconomic systems concern control and coordination and who was mainly responsible for them—managers or workers. The principle was that decisions should be made at the lowest organizational level possible. Instead of specialists, engineers, or supervisors making the decisions for corrections and changes in the production process or production line, the operators themselves, suitably trained, make the corrections and recalibrations and perform preventive maintenance. Operators are empowered to load, unload, and monitor machinery as well as to program them. In Britain, this change from specialist to operator control in an electronics company resulted in reducing downtime in computerized equipment that saved 10 hours of production time. In time, it became a 70% reduction as workers became more skilled and increased their intrinsic job satisfaction. Machine reliability was increased (Wall, Corbett, & Clegg, 1990; Jackson & Wall, 1991). In Japan, this was seen as “giving wisdom to the machine” (Ralls, 1994, p. 36). In 1987, a survey of 584 machine-manufacturing plants found that plants where machine operators routinely programmed the machines were 30% more efficient than plants where they did not. Such worker control in the more efficient plants depended on providing workers with technical training (Kelly, 1995).
According to Salzman (1992), the introduction of operator control met with a lot of resistence. Of 200 engineering design and engineering textbooks published between 1938 and 1989, only 42 mentioned worker roles in production and technology. The roles were mainly subservient to the machinery. Illustrative of the prevailing orientation in engineering design was the admonition that “the operator should not read an (ohm) meter. She should not be told what an ohm is. She should not be asked to make borderline judgments—Any adjusting may upset (her) … any slight gain in efficiency will be lost in retraining time” (Gibson, 1968). Nevertheless, as computers and technology have become more user-friendly, operators have become able to maintain, modify, and program their own equipment and machinery. This allows computer specialists and engineers to serve as coaches and to do more specialized work (Ralls, 1994).
Considerable consultation with the workers is seen as fundamental to establishing the bases for meeting both the requirements of the task and the workers’ socioemotional needs. Participative leadership becomes mandatory in the actual operations and is built into the design. For instance, minimal critical specification is a principle of sociotechnical design. According to this principle, the design process should identify what is essential to be accomplished in a task, and no more should be specified than what is essential. For example, although the design process may be precise about what needs to be done, it should leave maximum latitude about the method the employee may use to accomplish the task. Again, the multifunctional principle of sociotechnical design proposes that for an organization to be sufficiently adaptive to meet environmental demands, it is necessary for its members to be willing and able to perform more than one function or to perform the same function in a variety of ways to meet changing circumstances. Clearly, both task-oriented and relations-oriented leadership are needed to obtain the requisite employee performance and the commitment of employees to such fluid arrangements (Bass & Barrett, 1981).
Sociotechical principles are consistent with E. E. Lawler’s (1986) high-involvement management practices that promote more effective uses of technology and the introduction of new technologies. With suitable training, operators are enabled to make relevant decisions about maintaining and improving their technology-assisted work. Their involvement and alignment with organizational productivity goals are fostered by gain sharing, skill-based pay, teamwork, flexible assignments, reduced status differentiation between operators and supervisors, and engagement in problem solving. Such practices are key to successful implementing of technology in the work place (Klein & Ralls, 1994). At first, sociotechnical designs were applied only to routine, linear work. But by 1995, they had been found useful for improving services and noncontinuous processes, processing transactions, and nonroutine and professional work (Fox, 1995).
The requirements of the task change as a group progresses in its solution of a problem. Given these changing requirements during the course of group problem solving, Valenzi, Miller, Eldridge, et al. (1972) concluded that effective leadership for one phase of problem solving may be different than it is for another phase. Early on, the group usually engages in the divergent generation of alternatives. In this phase, broad participation is needed. As a consequence, Doyle (1971) considered democratic leaders to be most effective. But in the convergent, final, synthesizing phase, when coordination becomes more important, groups with leaders of high status were particularly effective. Becker and Baloff (1969) also suggested that optimal leader-subordinate relations may depend on whether the group’s immediate task involves information processing, the generation of alternatives, or decision making.
Ghiselli (1966a) observed experimental groups at various stages of problem solving. The presence of a strongly self-confident decision maker, along with highly intelligent, confident, and cohesive followers, was associated with the better performance of the group in the later stages of the group’s development. Nonetheless, initial performance was poor in its early, storming phase. Presumably, such confidence, competence, and motivation generated many conflicting alternatives that later formed the basis of high-quality decisions. Sample and Wilson (1965) also studied groups in different phases of problem solving. They found that task-oriented leaders quickly structured the group procedures during the planning phase and were then able to play a more relaxed role in the operational phase. Relations-oriented leaders, on the other hand, tended to hold group discussions during the planning stage, and the work did not get organized. As a result, these leaders had to try to organize procedures during the operational stage, with only partial success. In such circumstances, groups under task-oriented leaders performed more effectively.
Korten (1968) advised that if the final product of a task was practical, more directive supervision was in order. If the outcome was theoretical, participative leadership was likely to be more useful. Carter and Nixon (1949b) found that different participants emerged as leaders, depending on whether intellectual, clerical, or mechanical assembly tasks were involved. Carter, Haythorn, and Howell (1950) studied the effects of six types of tasks (reasoning, intellectual construction, clerical, discussion, motor cooperation, and mechanical assembly) on leadership in initially leaderless groups. Although there was some generality of leadership performance across all tasks, two clusters of tasks made a difference in who emerged as a leader. The leaders of the intellectual tasks were different from the leaders of the tasks involving doing things with one’s hands. As was mentioned earlier, C. G. Morris (1966a, 1966b) varied the type of task and the difficulty of the task for 108 groups. The variance in the leaders’ behavior was related more to the type of task than to the difficulty of the task. Discussion tasks elicited significantly more structuring of problems and more explanatory and defensive comments by leaders. Production tasks resulted in more structuring of proposed solutions, disagreement, and procedural comments. Problem-solving tasks were similar to discussion tasks but led to more irrelevant activity and less structuring of problems.
Among others, Mumford and Gustafeson (1988) saw an upsurge in organizations of an interest in stimulating creativity, generation of new ideas, and their conversion into action. Traditionally, such creativity had usually been seen as the product of a lone individual with a flash of insight about gravity as he observed an apple falling from a tree. Nonetheless, the research of Pelz (1963) and Pelz and Andrews (1966a, b) recognized early on the importance of the leadership of laboratory directors and leaders at the National Institutes of Health to the productivity of their scientists. Suggestion systems had been around since the end of the nineteenth century, but the interest in leaders who fostered creativity and innovation among their followers as a core purpose was sparked by rapid advances in technology, the information revolution, and global competition (Dess & Pickens, 2000). Tierney and Farmer (2003) developed a reliable and valid Creativity Leadership Index using five samples to assess the leadership of creativity. The data were based on responses from 1,219 employees.
The importance of autonomy and discretion was mentioned earlier in promoting innovation and creativity. Supervisory support was important to the generating of patents by 171 design and manufacturing engineers, design drafters, tool makers, and technicians in two manufacturing plants (Oldham & Cummings, 1996). On the other hand, when their group leader’s supervisory control was tight and the introduction of new ideas was discouraged among 200 scientists, the scientists’ creativity was squelched (Andrews, 1967). Mumford, Scott, Gaddis, et al. (2002) noted dilemmas facing the leader of creative people. Creative tasks are ill defined. To opimize creativity, the leader needs to avoid imposing too much or too little structure. Followers’ goals need to be aligned with the organization’s. Furthermore, the leader must evaluate the proposed work but not be overrestrictive in what is to be done.
Slusher, Van Dyke, and Rose (1972) found in a small sample study of nine engineering design groups that it was more important to group productivity for group leaders to initiate structure and be considerate than to be technically proficient. However, in most subsequent studies, it was seen as more important for creative group leaders to be technically proficient as well as expert problem solvers. Additionally, they needed to have general planning skills (Mumford, 2000). Glassman (1989) suggested that leaders of creative workers need to be able to negotiate disagreements and to listen nonevaluatively. Leading R&D activities requires the ability to communicate clearly with colleagues and top management. The R&D leader must appreciate the values and intrinsic motivation of professionals and the importance to them of challenging work and freedom to explore promising lines of inquiry.
According to Strohmeier (1998), technical leaders and senior scientists are specialized experts. They keep abreast of the literature in their field, value being recognized for patents, publications, and conference presentations, and are highly task-oriented and less concerned with relationships. Technical leaders need broad technical competence to maintain credibility with their technical followers. Yet they need to depend on their followers for additional specialized knowledge. Jaussi and Dionne (2003) showed experimentally, with 364 undergraduates in 79 groups, that leaders who were unconventional in their behavior tended to contribute indirectly to the creativity of their groups. Mumford, Scott, Gaddis, et al. (2002) provided more details about satisfying and effective leading of creative and innovative people.
Tasks at Different Organizational Levels. Systematic changes in the type of tasks that are performed occur as leaders move up the organizational ladder.7 Different types of tasks call for different kinds of leadership. At the production and operations levels, the processes are direct and concrete. Work and service are with tangible materials and methods that are accomplished by people, tools, and machines. At these levels, leaders deal with issues of routines, pacing, meeting deadlines, and balancing the need for immediate production or service with the need for the development of individual subordinates and the need to prepare subordinates for future operations. Trust is based on personal contact and knowledge. Leadership involves face-to-face interaction and interpersonal skills. At higher organizational levels, leadership deals much more with buffering the rational production and service at lower levels from the turbulence of the external environment. At these levels, the subordinates’ operations are monitored indirectly, and the coordination and integration of efforts with the market and the external environment become the tasks of consequence (Jaques, 1978). Not unexpectedly, Bass, Valenzi, Farrow, and Solomon (1975) found that the more subordinates’ work involved planning, coordination, evaluation, and other managerial activities, the more frequently their superior was likely to delegate decision making to them. However, follow-up analyses on a larger scale added that such subordinates were also given more direction.
Changes in the Organization’s Tasks. As the organization matures, its tasks change. Consequently, the type of leadership that is needed also changes, especially at higher organizational levels. Brenner (1972) and Lavoie and Culbert (1978), among others, described the commonly observed progression of organizations through development, maturity, and decline as different tasks arise. Different stages of an organization’s development call for the emergence of different styles of executive leadership in the different phases. Early on, the tasks of a business enterprise, for example, are to develop products, processes, and procedures; create a demand for them; build loyalty to a brand; find a significant niche in the market; attract financing; develop personnel and new technology. Later, the tasks may require more attention to cost containment, divestment, maintaining the share of the market, and integrating new product lines with older ones. The founders who conceive and originate organizations that grow and thrive, for example, are more prone to take calculated risks than are their counterparts who must deal with old, declining organizations.8 Leadership that is predictable and that works within the established rules and procedures and under greater formal constraints is more characteristic of mature organizations than of those in the early stages of their development.
In the established organization in a stable environment, the executive must be able to serve as a transactional leader who works within the institutionalized values and systems. The executive leader serves as a steward, balancing out the strongly developed interests of the different constituencies that may be in conflict with each other. Much political compromise and accommodation are needed. During periods of gestation, birth, development, resolution, re-formation, and renewal, the executive’s tasks require more transformational leadership, greater persistence and effort, and greater inspirational and charismaticlike behavior.9 New strategies and commitments to new values and new systems are required. The cycle of development, maturity, and decline is not inevitable. Rather, intervention by transformational managers can provide for the renewal of an organization through reform or revolution (Normann, 1977).
Computerized Work. By 1990, more than 80% of office workers were using computers. In the developed world in 2006, it would have been difficult to find office workers operating without them. Office supervisors need to be familiar with computers and the complexities of computer systems. The same is true for workers and bluecollar supervisors as automation continues to increase in factories, mills, and warehouses. Public agencies are now among the biggest users for documentation and dissemination of information. In addition to the rapid expansion of technological capabilities are the concomitant changes in organizational arrangements and leadership. In the U.S. military, online instant communication makes it possible to link the U.S. president with the local combat commander on the battle site, bypassing many organizational echelons. Likewise, fax and e-mail make for instant communication and potential control across continents, oceans, and organizational levels. Increasingly, the centralization of decision making or overcontrol of the local commander is the result. Similarly, tighter reins by higher-level civilian executives are made possible by computer systems. Although the greatest emphasis has been placed on the impact of automation on stable, repetitive, processing systems, computer-assisted creative designing and engineering are now commonplace for production and service processes. Computerized planning and control models are also available for real-time planning and control. Word processing and office automation increase the productivity of individual typists. Nevertheless, supposedly to take advantage of the automation, the individualized dyadic relationship of the manager and the typist may be severed and replaced with an anonymous pool of typists or by eliminating most of the pool by having each manager using his own personal computer. First-line supervisors become more like area supervisors when advanced manufacturing computer technology or computer integrated manufacturing is put in place. Area supervisors are shop floor experts or engineering trained personnel with some shop experience who serve as consultants and coaches to the operators. Minimally, the new technology brings about changes in the roles of first-line supervisor and middle managers (Ettle, 1986).
Supervisory Relations. According to Kraemer and Danziger (1984), computerization at the workplace increases the closeness of the supervisor with those at the bottom of the hierarchy. The supervisors lose some of their potential for upward influence, while those who are higher up feel a greater sense of control. The pressure of time and the importance of deadlines increase. These changes enhance the possibilities of bureaucratic management. But the effects may be counterproductive. Applebaum (1982) noted that computer programming applied to the construction industry resulted in the reliance on authoritarian and mechanistic procedures. For a while, the outcome was inefficient and irrational; control of the work process was decreased instead of increased.
However, Kerr, Hill, and Broedling (1986) offered a more balanced perspective. They suggested that automation should have its biggest effects on workers whose work had been labor-intensive and closely supervised and required little discretion and judgment. Obviously, the supervisors of such workers would be responsible for more sophisticated equipment. Nevertheless, computerized systems directly give upper management detailed information on the individual worker, such as his or her error rate and missed deadlines. Upper-level managers can bypass the first-level supervisor to obtain such information. The supervisor will then be expected to explain the reasons for the worker’s deviation from the production plans. Further conflicts of interest arise for the first-level supervisor, who now faces demands from a specialized staff of systems analysts and programmers, along with those from the traditional line superiors, subordinates, and staff personnel: “first-line supervisors will have to be more technically proficient, as well as more highly skilled in human relations than their predecessors” (p. 114).
Other Changes in Supervising. Computerization shortcuts the need for management controls, for it can provide direct feedback from work performed on the task without any supervisory intervention. The self-managed employee can be provided by computer with continued feedback about his or her performance along with the information needed to improve it. In contrast to the findings of Kraemer and Danzinger (1984), Kerr, Hill, and Broedling (1986) inferred that first-line supervisors have moved farther from their subordinates and closer to middle managers. They reasoned that computerization reduces the time that supervisors need to spend in planning and scheduling work; documenting records and reports; engaging in coordination and control; organizing subordinates’ work; and maintaining quality and efficiency, safety and cleanliness, and machinery and equipment. Presumably, spans of control can be increased. Nonetheless, the first-line supervisor’s role does not seem to be diminished, for computerization facilitates his or her service as a boundary spanner; a maintainer of relations with other units; or a selector, trainer, and motivator of subordinates. Computerization makes it possible for first-line supervisors to operate more like middle managers, relative to the total organization, rather than remain oriented toward dealing mainly with subordinates (Hill & Kerr, 1984). Chapter 29 will detail how computerized networks have replaced human real ones.
Shift from Power to Achievement. The increasing technology and computerization of tasks suggests that in more technologically advanced organizations, the power motive may become less important than the need for achievement as a predictor of success in advancement to higher levels of management. As was noted in Chapter 8, McClelland (1975) demonstrated that individuals who scored high on the Thematic Apperception Test in the need for power and the inhibition of power tended to emerge as more successful leaders in a variety of situations. However, when McClelland and Boyatzis (1982) examined whether the leadership motive pattern was predictive of the long-term managerial success of technical and nontechnical managers at AT&T, the power motive pattern was related to significantly higher levels of advancement after 8 to 16 years with the company only by nontechnical managers. For technically trained and experienced managers, the need for achievement, rather than the power motive, predicted advancement into the next several echelons of the firm.
Beginning with Georgopolous, Mahoney, and Jones (1957) and delineated by M. G. Evans (1970a) and House (1971), path-goal theory stimulated the search for an explanation of how the nature of the group’s task systematically affects whether leader consideration, initiation of structure, or interplay makes a contribution to the group’s satisfaction and effectiveness. Rightfully, the theory has been modified on a continuing basis by experimental failures. According to T. R. Mitchell (1979), path-goal theory calls for the leader to provide subordinates with coaching, guidance, and the rewards necessary for satisfaction and effective performance necessitated by the subordinates’ abilities to meet the particular task requirements and attain the designated goals. Focus is on ways for the leader to influence subordinates’ perceptions of the clarity of the paths to goals and the desirability of the goals themselves. Leadership behavior that is best suited for increasing motivation depends on the subordinates’ personal characteristics and the demands of the task. Valued rewards should be awarded contingently on effective performance.10
Path-goal theory is an exchange theory of leadership. It attempts to explain why contingent reward works and how it influences the motivation and satisfaction of subordinates. In its earliest version by Georgopolous, Mahoney, and Jones (1957), it focused on the need for leaders to “point out the paths to successful effort” (Bass, 1965, p. 150). Leaders do so by “increasing personal payoffs to subordinates for workgoal attainment, and making the path to these payoffs easier to travel, by clarifying it, reducing roadblocks and pitfalls, and increasing the opportunities for personal satisfaction en route” (House, 1971, p. 324).
Path instrumentalities are the subordinate’s subjective estimates that his or her performance will lead to the accomplishment of the goal. Achievement of the goal will result in ends desired by the subordinate. The leader enhances the subordinate’s motivation, performance, and satisfaction by clarifying and enhancing path instrumentalities (Van Fleet & Yukl, 1986). This cognitive-perceptual explanation of path-goal theory can be matched in terms of operant behavior and reinforcement theory (Mawhinney & Ford, 1977).
The Leader’s Role. Leaders can affect a subordinate’s efforts in several ways in the path-goal process. They can clarify the subordinate’s role, that is, state what they expect the subordinate to do. They can make the rewards to the subordinate more dependent on his or her satisfactory performance. They can increase the size and value of the rewards (M. G. Evans, 1970a). Specific leadership behaviors can contribute to their followers’ attainment of the goal by coaching, providing direction, and support to followers, alleviating their boredom and frustration with work, and fostering their expectations that their efforts will result in the successful completion of the task (Fiedler & House, 1988).
But early on, House and Mitchell (1974) recognized that path-goal leadership, as such, was needed and useful only in certain circumstances. The leader needs to complement only what is missing in a situation to enhance the followers’ motivation, satisfaction, and performance. What is missing is determined by the environment, the task, and the competence and motivation of the followers (Fiedler & House, 1988). Thus, the followers’ productivity is enhanced if the leader provides needed structure to clarify means and ends if they are missing or unclear to the followers. This contribution of the leader is particularly apparent in jazz ensembles. A deviation-counteracting loop is observed in jazz that involves the leader’s interpretation, criticism, and adjustments of the ensemble. The need for such correction is increased by the variety of selections the ensemble plays, the availability to them of new musical numbers, the difficulty of the numbers played, and the lack of rehearsal time (Voyer & Faulkner, 1986).
Given clear tasks and roles, the supervisor contributes to continued productivity by consideration, support, and attention to the followers’ personal and interpersonal needs for satisfying relationships (Fiedler & House, 1988). If what is missing can be supplied in other ways by the organization through policies, regulations, improved communications, channels of information, contingent reward schemes, and counseling services, such substitutes for the leadership11 may result in the same outcomes that would have been expected from the appropriate leadership.
The Path-Goal Linkage. The exchange involved in path-goal theory is seen when subordinates perceive high productivity to be an easy “path” to attain personal goals and, as a consequence, the subordinates are productive. Directive leadership increases the promise of reward to the followers for their performance and makes the paths to their goals clearer and easier. Accordingly, such directive leadership is needed only if the task is complex, difficult, or ambiguous and its goals are unclear. Whether the followers are self-reinforcing and have a great need for autonomy, growth, achievement, or affection will also make a difference. On the other hand, if subordinates are faced with simple but boring tasks, a leader may do better by being supportive and considerate rather than directive. Too much motivation among followers, evidenced by a state of high anxiety, may call for calming support from the leader rather than any talk about contingent (uncertain) rewards, which will increase such anxiety. Supportive confidence building may be required rather than more drive (Yukl, 1981).
Translated into experiments, considerate leadership behavior or supportive relations-oriented leadership was expected to correlate more highly with satisfaction and productivity in structured than in unstructured situations. The leader’s initiation of structure was expected to correlate more highly with satisfaction and productivity in unstructured than in structured situations. These expectations fit with the conventional wisdom that “chaos is the midwife of dictatorship” (Durant, 1957).
Supportive Results. Reviews of the empirical literature by House and Mitchell (1974) and Schriesheim and Kerr (1974) tended to confirm the theory, as did a meta-analysis by Indvik (1986a) involving 87 empirical tests. In addition, House and Dessler (1974) demonstrated that, as predicted, the available task structure generally determined whether the initiating structure and consideration by the leader would contribute to the followers’ satisfaction, positive expectations, and role clarity. Earlier, House (1971) found support a posteriori in several studies cited in earlier chapters.12 In specific a priori tests of the theory, House found, as expected, that the satisfaction of followers was associated with the extent to which the leader’s initiating structure reduced role ambiguity. Likewise, Meheut and Siegel (1973) observed that the leader’s initiation that was role clarifying was positively related to the followers’ satisfaction with management by objectives. A more complicated finding was that the more autonomous the followers, the more the leader’s initiation of structure correlated with followers’ satisfaction but the less the leader’s initiation correlated with the followers’ performance. At the same time, as the scope of the subordinates’ task decreased, the leader’s consideration correlated more with the followers’ satisfaction and performance. Also supportive were direct tests of the theory by Dessler (1973), who found that with the leader’s consideration held constant, the leader’s initiation of structure correlated less with the followers’ satisfaction and role clarity as the ambiguity of the task decreased.
Schriesheim and DeNisi (1981) studied how the variety of tasks in a job, feedback, and dealing with others moderated the impact of the initiation of structure on satisfaction with supervision among 110 employees who were working in a medium-size plant and among 205 employees of a medium-size manufacturer. The variety of tasks was expected to require more initiation of structure for employees’ satisfaction with supervision. Such initiation would be redundant with routine jobs (House & Dessler, 1974) and when feedback was already structured and followers dealt with others a lot. The results confirmed the moderating effect of all three task variables.
Mixed and Nonsupportive Results. Szilagyi and Sims (1974) obtained data from 53 administrative, 240 professional, 117 technical, and 231 service personnel at multiple levels of occupational skills in a hospital. Although the results supported path-goal propositions concerning the demands of the task and the relationship between the leader’s initiating structure and the followers’ satisfaction, they failed to do so for the relationship between the leader’s initiation and the followers’ performance. Similarly, Stinson and Johnson (1975) tested hypotheses derived from the path-goal theory of leadership that the correlations between the leader’s initiating structure and satisfaction and role clarity are more positive under conditions of low task structure, low task repetitiveness, and high task autonomy than under high task structure, high task repetitiveness, and low task autonomy. The leader’s consideration and the subordinates’ satisfaction and role clarity were expected to be more positively related under structured, repetitive, dependent conditions than under unstructured, unrepetitive, autonomous conditions. The subjects were military officers, civil service personnel, and project engineers. The results were consistent with path-goal theory with respect to consideration but tended to be contrary to the theory regarding the initiation of structure.
Contrary to path-goal theory, the leader’s consideration was still generally found to result in the higher satisfaction of followers, regardless of the characteristics of the task (Johns, 1978; Miles & Petty, 1977). Thus, J. F. Schriesheim and C. A. Schriesheim (1980) surveyed 290 managerial and clinical employees in nine different jobs at five levels in the operations divisions of a large public utility. Contrary to path-goal theory, they found that regardless of the task structure, organizational level, or type of job, supportive (considerate) leadership explained 63% of the variance in the followers’ satisfaction with their supervisors after instrumental leadership (initiation) was partialled out. But, as predicted, instrumental leadership (after supportive leadership was partialed out) contributed 17% to accounting for the variance in role clarity. However, again this covariance was unmoderated by the task structure, organizational level, or type of job. Likewise, Seers and Graen (1984) found that without reference to leadership, performance and satisfaction outcomes directly depended on the characteristics of the task, as well as on the followers’ need for growth. In the same way, the satisfactory quality of the leader-followers relationship independently added to the prediction of the outcomes of performance and satisfaction without reference to the task and the followers’ needs, which were also related to the outcomes.
Wolcott (1984) tested path-goal predictions for library supervisors and the performance of their reference librarians and catalogers. Contrary to path-goal predictions, the initiation of more structure contributed to better performance when the task structure was already high than when it was low. The librarians’ high educational level and low need for independence were seen to be possible explanations for the results.
More initiation of structure is likely to be seen when the group task is already structured (Bass, Valenzi, Farrow, & Solomon, 1975). It may be that a leader’s initiation may be easier to accomplish if subordinates’ tasks are already structured but actually may be useful when tasks are unstructured. Generally, initiating structure still frequently increases tensions, especially when consideration is low (Miles & Petty, 1977; Schriesheim & Murphy, 1976) and when the initiation of structure measure continues to contain coercive, threatening items, along with direction and order giving. In turn, this linkage of direction and coercion is a consequence of dependence on empirical rather than conceptual analyses for developing measurements. Although autocrats tend to want to structure situations, conceptually one can give orders without being a threatening autocrat. For more than 1,300 subordinates of their 340 managers, the correlation between being coercive and being directive was only .38 (Farrow, Valenzi, and Bass, 1980).
In a first study, Greene (1979a) showed that, as expected, instrumental (structuring) leadership behavior was correlated positively with the satisfaction and performance of 119 engineers, scientists, or technicians if they faced tasks with little structure. But such instrumental leadership was negatively correlated with satisfaction and minimally with performance when the tasks were more structured. Considerate or supportive leadership, as expected from the theory, increased the correlation with intrinsic satisfaction (but not with performance or extrinsic satisfaction) as the task structure increased. In a second study, Greene (1979b) tested several assumptions about causation that underlie the theory. The findings supported the theory, except, again, for the hypotheses concerning the followers’ performance. Downey, Sheridan, and Slocum (1975) found only partial support for the path-goal predictions, and J. P. Siegel (1973) and Szilagyi and Sims (1974b) found none. Dessler and Valenzi (1977) failed to find moderator effects across supervisory levels. T. R. Mitchell (1979) concluded that the findings were stronger for the consideration hypothesis than for the structuring hypothesis and stronger for satisfaction as a criterion than for performance.
Indvik (1985, 1986a) completed a meta-analysis of 48 path-goal studies involving 11,862 respondents. Task structure, as such, was measured directly in some of the studies. In the rest of the studies, low job level was accepted as an indicator of high task structure, as was large organizational size. As predicted, when structure was absent from the work environment, directive, structuring leadership behavior contributed to the intrinsic motivation of followers, their satisfaction with the leader, and their overall satisfaction, but, surprisingly, it failed to add to role clarity, as such. However, contrary to predictions, directive, structuring leadership contributed to the subordinates’ performance when the structure was high but not when the structure was low.
Considerate, supportive leadership behavior in a highly structured work setting did enhance motivation, satisfaction, performance, and role clarity, as predicted. In a related meta-analytic report, Indvik (1986b) concluded that participative leadership provided the most overall satisfaction to subordinates who preferred and experienced a low task structure. Furthermore, when the task structure was high, achievement-oriented leadership behavior was related to increased intrinsic satisfaction among subordinates but decreased extrinsic satisfaction and performance for those subordinates with a high need for achievement.
Johns (1978) suggested that much more is “missing” that the leader may supply efficaciously than just task structure, as measured by House and Dessler’s (1974) scale about the extent to which tasks are simple, repetitive, and unambiguous. Johns (1978) argued for using a broader measurement of the scope of a job for determining what could be missing from it. Such a measurement would be Hackman and Oldham’s (1975) index based on variety, autonomy, the significance of the job, feedback from the job, and the identity that the job provides to its occupant. Johns (1978) found that Sims, Szilagyi, and Keller’s (1976) Job Characteristics Inventory, which measures the scope of a job, provided a measure to moderate more consistently the relationships between leadership behavior and the satisfaction of subordinates than task structure alone could do.
Coercive versus Noncoercive Initiation of Structure. The measures of leadership behavior are obtained most often from the Leader Behavior Description Question-naire—Form XII (LBDQ-XII) and less from the Supervisory Behavior Description Questionnaire (SBDQ).13 Schriesheim and Von Glinow (1977) first noted that path-goal predictions of job satisfaction were less likely to be supported when a more coercive measure, such as the SBDQ initiation of structure, had been used.14 Schriesheim and Von Glinow then demonstrated, with 230 maintenance workers, that if a coercion-loaded scale was used, reverse results were obtained for the path-goal predictions for job satisfaction. But when coercion-free scales (the LBDQ and the LBDQ-XII) or items from them were employed, path-goal predictions were confirmed if task structure and role clarity were used to moderate the relationship between the leader’s consideration and initiation of structure and the followers’ job satisfaction.
Conditions Affect What Leaders Can Do. A second source of contradictory findings results from the fact, as noted above, that leaders tend to be more directive when it is easier for them to do so, such as when roles are clear, conditions are structured, and jobs are routine (Bass, Valenzi, Farrow, & Solomon, 1975). But such structuring would seem to be redundant for subordinates’ productivity when conditions are already structured. Rather, it would seem that such direction is needed more when conditions are unstructured, for in such unstructured situations, it might be argued that the group wants some direction from the leader, not just the leader’s sympathy. Nevertheless, Indvik’s (1985, 1986a) previously mentioned meta-analysis proved otherwise: directive, structuring leadership contributed to followers’ performance when structure was high, not when it was low.
Attributes of Leaders and Subordinates. The leader’s personality also needs to be taken into account in the structured situation, given Farrow and Bass’s (1977) finding that highly directive leaders tend primarily to be satisfied authoritarians. In addition, the subordinates’ personalities need to be considered. Griffin (1979) proposed a set of prescriptions combining path-goal theory and the subordinates’ need for achievement and self-actualization. Griffin called for achievement-oriented, consultative leadership for self-actualizing followers with “big” jobs. But for self-actualizers in routine jobs of little scope, supportive leadership (consideration without consultation) was required. For “big” jobs performed by occupants who are uninterested in self-actualization, directive leadership (structuring without threat) was seen as most needed. For occupants of routine jobs who have no need for self-actualization, maintenance leadership behavior (management by exception) was suggested.
Schriesheim and Schriesheim (1980) added other subordinate variables that are likely to act as path-goal moderators of the leader-outcome relationships. These variables included the followers’ authoritarianism, ability, training, need for affiliation, and experience relative to the demands of the task and their internalization of professional norms and standards. Similarly, Abdel-Halim (1981) found that the subordinates’ locus of control (internal or external) had important effects on the path-goal leader-outcome relationships associated with the ambiguity of the role and the complexity of the job.
Algattan (1985) examined the extent to which the scope of the followers’ task, strength of the need for growth, and locus of control moderated leader-outcome relationships for two periods, two months apart. At each time period, if the subordinates’ locus of control was external, the scope of their tasks and the strength of their need for growth increased the extent to which both participative and directive leadership contributed to their satisfaction and performance. But if the subordinates’ locus of control was internal, task-oriented leadership, as such, was of more importance to their satisfaction and performance. However, a cross-lagged analysis of the correlations for the two time periods failed to support the existence of casuality in the relationships.
Craig (1983) attempted to show the importance of subordinates’ self-esteem to path-goal leader-outcome relationships but failed to find the expected interactions. Wolcott (1984) found no effect on the relationships from differences in the subordinates’ need for independence. Keller (1987) argued that the discomfort of role ambiguity may differ from one follower to another. Some people who may want to clarify and structure their roles themselves are unlikely to be enthusiastic about a leader who initiates structure even if the task is unstructured or ambiguous. Followers with high levels of education, such as R & D professionals, who may have internalized professional norms that provide them with role clarity, may not need or want the leader to initiate structure. Some followers may actually enjoy the unstructured nature of a task; they may have a low need for clarity and prefer to create their own structure. For Keller, compared to task structure, the followers’ need or lack of need for clarity was seen to be a more important moderator of the correlations between the leader’s initiation of structure and the subordinates’ satisfaction and performance.
In a survey of 477 professionals employed in four R & D organizations, Keller (1988) employed Rizzo, House, and Lirtzman’s (1970) role ambiguity scale to measure the subordinate’s perceived task clarity, as well as Ivancevich and Donnelly’s (1974) scale to measure the subordinate’s felt need for clarity on the job. He found that the need for clarity had a moderating effect on the initiation of structure-satisfaction relationship for both concurrent data and data gathered one year later. The higher a subordinate’s felt need for clarity, the stronger was the relationship between the leader’s initiation of structure and the subordinate’s job satisfaction. The subordinate’s need for clarity was similarly found to moderate the initiation-performance relationship in the largest of the R & D organizations. But, as proposed, the actual clarity of the task for the subordinates, as such, failed to serve as a moderator for these leader-outcome relationships. In the same way, Kroll and Pringle (1985) failed to find the expected effects of the leader’s directiveness on the satisfaction of 43 middle managers in marketing. Kroll and Pringle explained the results by noting that managers rated the ambiguous situation as a positive experience, particularly if they judged the amount of direction they received to be the amount they actually desired.
A Comprehensive Study. Using data from a survey of 467 nonacademic staff at a university, Indvik (1988) completed tests of 17 hypotheses that involved directive, supportive, participative, and achievement-oriented leadership behavior, plus the expectancies that increased effort would improve performance and that improved performance would yield valued outcomes. Also measured were intrinsic satisfaction with work, extrinsic satisfaction with pay and promotion, and satisfaction with one’s superior. The subordinates’ performance was appraised by their superiors. Indvik examined the task structure, norms of the work group, and organizational formalization as situational moderators of the relations between superiors’ leadership behavior and subordinate outcomes. Personal subordinate moderators included the need for achievement and preference for environmental structure. Hierarchical stepwise regression analyses15 provided support for seven of the 17 hypotheses tested. Moderators that had significant effects included the subordinates’ preference for structure and need for achievement. However, Indvik concluded that generally, because of its low reliability, the subordinates’ preference for structure had a weak moderating effect on the relations of leadership behavior to subordinate outcomes. Directive and achievement-oriented leadership behaviors were too highly correlated with each other to be distinguishable. Indvik recommended that future studies should measure transformational leadership behavior instead of the transactional leadership behavior on which path-goal research has concentrated, for it is likely that transformational leadership behavior is more sensitive to task structure and the characteristics of subordinates.
Integration of Findings. Neider and Schriesheim (1988) constructed a comprehensive path-goal model, shown in Figure 27.1, that attempts to incorporate many of the consistent findings about the process and the variables of consequence.
In the model, the manager stimulates the subordinate’s effort by offering valued rewards and linking them to the subordinate’s effort and performance. How much the effort yields high performance depends on the subordinate’s knowledge, skills, and abilities, as well as on the absence of hurdles to performing the job. The rewards received by the subordinate, if valued and equitable, create satisfaction and encourage the subordinate to remain on the job.
By 1993, Wofford and Liska had located 120 studies that tested aspects of path-goal theory, not counting reviews and expositions. Jermier (1996) explained the continuing popularity of the theory as being that it rejected a single best way to lead and focused attention on situational contigencies. It placed more emphasis on the effects on leaders and followers of dyadic than group relationships. Although in some instances its derived propositions were sometimes invalidated, they remained interesting and stimulating to scholarly research. Raw empirical publication in leadership was no longer acceptable or convincing without guidance from theory and hypothesis testing. Path-goal theory provided a more sophisticated way to examine leadership processes.
Caveat. Despite a considerable amount of general empirical support for it, path-goal theory is complex, which makes it difficult to test the theory’s deduced relationships. Furthermore, too much rigor may be required of such tests, and sampling and measurements may be inadequate to meet the requirements (Yukl & Van Fleet, 1986). It is not surprising that a wide array of empirical results, sometimes contradictory, have emerged from the surveys and experiments that tested various propositions derived from path-goal theory.
Figure 27.1 An Integrated Path-Goal Perspective of Motivation and Leadership
SOURCE: Neider and Schriesheim (1988).
The leader’s and subordinates’ competence, personality, and autonomy and the requirements of the task systematically moderate how different leadership behaviors affect what happens in the group. The characteristics of the task that make a difference include its structure, clarity, provisions for the subordinate’s autonomy and use of discretion, routineness, variety, complexity, difficulty, interdependencies, automation, multiplicity of solutions to problems, and task requirements for cooperation. As the task is different at different phases in a group’s development and at different organizational levels, so is the requisite leadership. Socioemotional and task requirements call for sociotechnical designs of the workplace.
Path-goal theory has been widely tested and modified to account for the impact of the task on optimum leader-subordinate relations. Currently, it suggests that to obtain subordinates’ effective performance and satisfaction, the leader must provide structure if it is missing and must supply rewards that are contingent on the adequate performance of the subordinates. To do so, leaders must clarify the desirability of the goals for the subordinates, a role seemingly suited for transformational leaders. But their efficacy in doing so will depend on such personal characteristics as the subordinates’ need for clarity. However, the lack of clarity and ambiguity of the situation may also be sources of stress—to be examined next.