[After making profuse apologies for disturbing the Minister, … the civil servant] indicated that some urgent matters has arisen as a result of the day’s business in the House, and the Minister must take certain decisions which would then be implemented in time for the question hour the next day. The civil servant … had brought the file so the Minister could study the questions, before [deciding]. But … all the Minister said was: ‘Where do I sign?’ A pen was produced by the civil servant, the places for signature were indicated, and with [deference and more apologies, he] … backed out of the room. Bailey, 1969, p. 73
Successful leaders influence their followers and bring about changes in their followers’ attitudes and behavior. In the same way, by accepting, modifying, or rejecting the influence, followers influence the leader’s subsequent behavior and attitudes. Sometimes, subordinates lead and superiors follow.
Leaders and followers matter to each other, as do the quality of relations between them. “Governments and laws cannot work without people in authority, and those in authority are powerless without support, cooperation, and obedience from the people they govern” (Sarsar & Stunkel, 1994). Leaders and followers are similar in many ways. The antitheses of leaders are not followers, but alienated, apathetic isolates and rejectees. Like good leaders, good followers are transformational, committed to principles and purposes beyond themselves. Like leaders, followers use both hard and soft tactics. Hegel suggested that to be a good leader, one needed to be a good follower. “The school for leadership is indeed follower-ship, a followership that is fully preserved within leadership, but transformed for having moved beyond itself” (Litzinger & Schaefer, 1982). Nevertheless, the study of followers has always played second fiddle to the study of leaders.
According to critical incidents obtained from 81 respondents, good leadership is distinguished from bad leadership by the quality of the leaders’ relations with the followers (Hollander & Kelly, 1990). Leaders need to be helped by their followers to understand the tasks and challenges they face. Followers need to learn to challenge their leaders while respecting the leaders’ authority (Hirschhorn, 1990). Leaders need to engage followers in satisfactory mutual pursuits. Followers give or withdraw support of their leaders (Hollander, 1997). The exchanges vary considerably in nature and amount. Different leaders have different relations with their followers; likewise, followers have different relations with their leaders. According to a survey in three British companies, the organizational culture affects what types of leaders are dominant, and the leaders in turn affect what types of followers are dominant (Brown & Thornborrow, 1996). Leaders command more attention, but followers affect and constrain what the leaders can do (Hollander, 1992). “Leadership is not what the leader does but what the leaders and collaborators do together to change organizations” (Rost, 1993, p. 92). Bennis (1999) agrees, stating that effective change in organizations requires an alliance between the leaders and the led.
This chapter examines the roles of both the leader and the led in the exchange that takes place between them. Specifically, it looks at the antecedents and consequences of the leader’s downward influence on the subordinate, the subordinate’s upward influence, and their mutual influence.
Leaders can be found who exert little discretionary influence on their followers. These leaders are the rule-governed, fully programmed administrator; the paper-pushing absentee supervisor; and the token officeholder without power. They are glorified doormen, whose behavior is almost fully determined by others. Other leaders use a lot of discretionary influence as disciplinarians, rule makers, active monitors, and instrumental purveyors of praise, reward, or penalties. They also use discretionary influence to be considerate, stimulating, informative, and inspirational. As a consequence of their influence, their followers can perform better and be more fully informed. These followers can avoid mistakes. Their interests may be enlarged. Their expectations may be developed, their preference for taking risks may be altered, and their satisfaction with their roles may be enhanced. The emphasis that followers believe their leaders place on self-guiding performance determines whether or not the followers persist at working on unsolvable problems (Brown, 2000). But followers fail to benefit from leaders’ successful influence when it is not effective in achieving goals of consequence.
In the formal organization, the performance of individual subordinates or groups depends on their energy, direction, competence, and motivation. They perform what is required to reach the objectives of their positions in the system. The leader may contribute to the adequacy of their performance by: (1) clarifying what is expected of the subordinates, particularly the purposes and objectives of their performance; (2) explaining how to meet such expectations; (3) spelling out the criteria for the evaluation of effective performance; (4) providing feedback on whether the individual subordinate or group is meeting the objectives; and (5) allocating rewards that are contingent on their meeting the objectives.
Effective leadership develops understanding and agreement about the leader’s and subordinate’s roles in this process. For instance, if the leader engages in management by objectives, leadership may take the form of periodic discussions between the leader and the subordinate. A review of past performance and obstacles to effectiveness is the basis for setting mutually acceptable objectives for the next period. Legitimacy for the roles of both the leader and the led is provided by organizational policies that declare and support the roles.
Influence tactics are behaviors designed to change another person’s values, attitudes, beliefs, or behaviors (Hughes, Ginnett, & Curphy, 1993). Yukl and Falbe (1990) used the Influence Behavior Questionnaire (IBQ) to measure nine tactics of influence upward, laterally, or downward, conceptually originated by Kipnis and Schmidt (1982) and assessed with the Profile of Organizational Influence Strategies (POIS).
The transactional leader may introduce a proposal with: (1) the tactic of ingratiation: “Mary, I know you can do this job;” (2) An exchange may be offered: “George, you can leave early if you get this work done;” (3) Another tactic is a personal appeal from the leader: “Bill, I need your help;” (4) Another influence tactic is pressure: “Jenny, if you want to stay out of trouble, you had better finish this now;” (5) A legitimating tactic is: “Carl, I am an attorney here to represent you;” (6) A coalition tactic is: “Eleanor, let’s join forces to confront the opposition.”
A transformational1 leader may use intellectual stimulation for rational persuasion: “Martha, considering the alternatives, it is the treatment with the least risk.” Inspirational appeal is another transformational tactic: “We must give until it hurts!” A tactic of considerate leadership is consultation: “Ann, before the decision, I need to know what you think about it.”
Although ingratiation, exchange, and personal appeal yielded task commitment from subordinates and pressure reduced it, the tactics that were actually effective were the transformational ones, according to a survey of 128 managers and their 526 subordinates from five companies in diverse industries (Yukl & Tracey, 1992). When “critical influence” incidents were gathered by 215 night MBA students, Yukl, Kim, and Falbe (1996) found that managers with lack of referent power used more pressure tactics (r = –.33), which tended to damage their relations with subordinates and peers. Managers with more referent power were less likely to use pressure tactics. Multiple regression analyses indicated that strong rational persuasion, inspirational appeal, and consultation were more likely to result in desired outcomes. Followers accord more discretionary opportunities to leaders who are seen as more competent, legitimate, and expert in solving problems (Lord & Maher, 1991). In a study of 78 leaders and their 156 subordinates, Tepper (1990) showed that less controlling downward influence tactics were displayed by follower-oriented than self-oriented leaders.
Hard tactics such as pressure and legitimizing are used when the user is powerful and resistance is expected; soft tactics such as ingratiation or personal appeal are used if the user lacks power or will personally benefit from a successful attempt (Hughes, Ginnett, & Curphy, 1993).
Feedback about a subordinate’s performance is the most common contingent reinforcement provided by a leader. Supervisory feedback often is required to improve the subordinate’s performance and can affect either the subordinate’s ability or his or her motivation to do the job (Locke, Latham, Saari, & Shaw, 1981; Payne & Hauty, 1955). Cook (1968) found that improvement in the attitudes and performance of managers who participated in a business simulation game was directly related to the frequency with which reports of their performance were fed back to them.
Rewarding when positive, feedback can be highly punitive when negative. However, negative feedback may quickly come to be interpreted as rewarding if it is seen as intended to be helpful and if it actually results in improved performance. Subordinates will see negative feedback about their failures as fair and accurate if the causes are attributed to bad luck or external circumstances rather than to their lack of ability or motivation, and if the feedback is about the task, not about them (Liden, Ferris, & Dienesch, 1988). Nevertheless, when subordinates suspect that they are doing poorly, they will seek feedback to short-circuit the buildup of negative feedback in order to uphold their self-esteem. They can arrange how and when feedback is given, and they can mitigate blame (Larson, 1989). Feedback can also be neutral and non-reinforcing, as when the superior merely acknowledges, without evaluating it, he or she has seen a subordinate’s behavior or has heard a subordinate’s statement.
According to a survey of 360 supervisor-subordinate dyads by Glynn, Larson, Fleenor, et al. (1985), supervisors differ from each other in the timeliness, specificity, frequency, and sensitivity of their feedback. But these four dimensions are highly intercorrelated. Furthermore, negative feedback that is prefaced by the Drake-Moberg (1986) sedative or palliative statements will be accepted by subordinates with less sense that they are being negatively reinforced (“It’s probably not necessary to say this, but …” or “I know it’s hard, but …”).
Types of Feedback. Supervisory feedback can range from a formal annual appraisal interview (see Chapter 35) to a grunt of acknowledgment of a message received or a pat on the back for a job well done. The grunt or pat has the advantage of occurring soon after the behavior about which the feedback is being given. The appraisal interview has the advantage of being systematic and of couching the feedback in the context of goals, needs, and plans for future action.
Impact of Feedback. It is important that subordinates accept and agree with the performance feedback their supervisors believe they are giving them (Ilgen, Fisher, & Taylor, 1979). Naturally, agreement and acceptance are more likely if the feedback is positive (Jacobs, Jacobs, Feldman, & Cavior, 1973). Agreement and acceptance are also more likely if the feedback is clear, convincing, credible, and frequent.2 The credibility of the superior’s feedback is likely to be enhanced if the subordinate believes the superior is highly knowledgeable about the subordinate’s job and has had sufficient opportunities to observe the subordinate’s performance.3 The context in which feedback is delivered will affect whether subordinates interpret it as positive or negative. For example, if feedback is solicited, it may be received more positively than if it is volunteered unexpectedly.
Pavett (1983) illustrated the positive effects of feedback on the performance of 203 staff nurses. Komaki, Collins, and Penn (1982) found feedback to have a positive impact beyond giving instructions. They monitored the safety performance of 200 employees over 46 weeks. After a baseline was established, safety rules were clarified at meetings in which considerable supervisory-subordinate interaction occurred. In contrast to the baseline record of performance, modest improvements occurred in two of four departments as a consequence of the clarification efforts. Then safety performance feedback graphs were introduced. It was this feedback that produced significant improvements in safety over the baseline. Whether the feedback was delivered once a week or twice a week did not seem to matter. According to a similarly designed investigation by Chhokar and Wallin (1984), similar effects of feedback on improvements in safety occurred among 58 employees over a 10-month period.
In addition to giving feedback, the leaders’ words and actions convey meaning. Their communications distinguish leaders who are successful and effective from those who are not. “Understanding and consensus … at an operational level are indispensable … to gain success [in] global competition” (Testa, 1998, p. 32). As was noted in earlier chapters, emergent leaders contribute strongly to the interactions in their groups. They initiate more ideas, express more opinions, and ask more questions than members who do not emerge as leaders (Bass, 1954a; Morris & Hackman, 1969). Elected and appointed leaders would be expected to do likewise in discussions with their groups, as well as in separate dyadic interactions with each of their team members (Watson, 1982).
In a survey, chief executive officers (CEOs) of large businesses ranked face-to-face communication as the most important source of their effectiveness (Anonymous, 1978). An intensive study of nine senior executives by S. Carlson (1951) over a four-week period noted that they spent approximately 80% of their time talking with others. A detailed study of four departmental-level managers also found that more than 80% of their time was spent in conversation. Zelko and Dance (1965) stated that when managers were asked how much of their work-day was spent in communicating, their replies ranged from about 88% to 99%, with most saying that it was above 90%.4
According to Baird (1980), the credibility of managers’ communications depends on their competence, esteem, personality, dynamism, character, and perceived intentions. Empirically, Klauss and Bass (1982) established strong positive linkages between the trustworthiness and informativeness of supervisors and their careful communications to subordinates; their two-way rather than one-way communications; and their attentive listening. In turn, the trustworthiness, informativeness, and care of supervisors contributed to their subordinates’ role clarity, satisfaction with their supervision, and to the effectiveness of their groups.
Hain (1972) reported that in four General Motors plants, productivity and profitability increased most in the plant that also showed the greatest improvement in communications. Similar parallel improvements in other General Motors plants were reported by Widgery and Tubbs (1975) and Tubbs and Widgery (1978). Hain and Tubbs (1974) found greater efficiency, fewer grievances, and lower absenteeism to be associated with employees’ ratings of the effectiveness of their supervisors’ communication. The effectiveness of supervisors’ communication was the best predictor of low grievance activity in still another General Motors automotive assembly plant (Tubbs & Porter, 1978). Such effective communication included agreement that supervisors were friendly, were easy to talk to, listened with interest, paid attention to what others said, were willing to listen to others’ problems, were receptive to ideas and suggestions, and showed how performance could be improved. These various communication behaviors are usually included in measures of leadership. It is not surprising that Klauss and Bass (1981) found correlations as high as .65 between the various communication styles of supervisors and the supervisors’ leadership styles as described by their subordinates.
Porter and Kaufman (1959) devised a scale for determining the extent to which supervisors described themselves as similar to top managers. Self-perceptions similar to those of top managers were associated with patterns of interaction that peers of the supervisors perceived to be similar to the interaction patterns of managers in top-level positions. Katz, Maccoby, and Morse (1950) noted that supervisors in an insurance firm tended to model their tendency to be coercive or participative on whether their bosses were coercive or permissive. Likewise, R. Cooper (1966) showed that workers tended to pattern their own task behavior after that of their supervisors. Task-oriented leaders supervised groups in which workers made fewer errors in their work and had lower rates of absenteeism and tardiness than was true of groups whose leaders did not have a task orientation. According to a study by Kern and Bahr (1974) of approximately 100 staff personnel in the Washington State Division of Parole, parole officers who interacted a lot with their supervisors used their superiors as models for the way they supervised their parolees. But such modeling did not occur when the parole officers interacted less frequently with their supervisors. H. M. Weiss (1977) studied 141 superior-subordinate pairs of leaders, obtaining from each member of the dyad a self-description of his or her supervisory behavior, along with the subordinate’s evaluation of the superior’s competence and success. These items were then correlated with the degree of similarity found in the self-descriptions of the superior-subordinate dyads. Weiss found that subordinates tended to choose for role models those superiors they saw as more competent and successful. Although transformational leaders are more likely to serve as models for subordinates who identify with them, transactional leaders may also serve as models
Adler (1982) found that the characteristics of superiors whom subordinates chose as models depended on the subordinates’ self-esteem. Among the subordinates of 66 Israeli heads of bank departments, those with high self-esteem were more likely to model themselves on the heads whom they perceived to have reward and coercive power; those subordinates with low self-esteem were more inclined to model themselves on heads whom they perceived to have referent power. Regardless of the subordinates’ self-esteem, modeling was also more apparent of heads who displayed more initiation and consideration in their leadership behavior. Behavioral contagion is less obvious modeling. A crude boss can spawn crude subordinates. The subordinates unconsciously adopt the boss’s spoken expressions, intonations, and peculiar nonverbal mannerisms.
Leaders as Cues. Followers may come to depend on their leader’s view of reality as their prime source of information and expectations. Beyond followers’ modeling of their leader’s behavior, Graen and Cashman (1975) noted that followers also enlarged their interests to match those of their leader more closely. Followers attempt to increase their esteem in the eyes of their leader to ingratiate themselves with the leader (E. E. Jones, 1964). Friedlander (1966b) asked members of a research and development organization to describe various aspects of members’ interaction and the group’s performance. The effectiveness of the group was associated with open discussion and with the leader’s suggestion of new approaches to problems. At the same time, a member’s influence on other members was associated with his or her influence on the leader. Members accepted the leader’s influence when policies were clear-cut and group tensions were low. The members tended to play their expected roles and discuss divergent ideas when the leader was oriented toward productivity and efficiency.
Daniels and Berkowitz (1963) experimentally varied supervisor-worker dependence for the attainment of goals, the degree of liking, and the time required for a supervisor to learn of a worker’s performance. They found that workers tried hardest under independent conditions when they believed that the supervisor would learn about their performance quickly. They also worked hardest when they had to depend on a supervisor whom they liked. Similarly, Katzell (1987) showed, from the results of a survey, how complex is the impact on the morale, involvement, and performance of employees of the extent to which supervisors help employees achieve intrinsic and extrinsic rewards. The extent to which they do so is linked to the extent to which they cue the employees by setting goals for them, maintaining normative standards, and preserving equity.
Falling Dominoes: Modeling or Alternating? Do strong leaders at one hierarchical level alternate with weak leaders at the level below them? Or does the style of leadership cascade from one management level to another through modeling and other processes? Do subordinates at each descending level below the boss imitate their boss, or do they complement their boss’s leadership with compliance to fit it? If A kicks B, will B kick C or will B become solicitous of C after being kicked by A? Do modeling and matching one’s superior make a difference, or is alteration from one level of supervision to the next more likely to be productive? The modeling of transactional leadership behavior is supported in a number of ways. When a manager’s boss rewards the manager for performance or allocates requested resources to him or her, some of these rewards and resources make it easier for the manager, in turn, to reward his or her subordinates for their performance. Discipline applied by the boss usually will require similar disciplinary action by the manager. Subordinates will be safer targets of the manager’s displaced hostility, which is sparked by the hostility of the boss toward the manager. The boss’s clarification of goals for the manager provides the means by which the manager clarifies goals for his or her subordinates. On the other hand, the threatening boss may create a manager who is attentive to rewarding subordinates in exchange for their support against the threatening boss. The inefficiencies of the boss who practices non-contingent reinforcement with the manager may result in the manager rejecting the boss’s style in favor of contingently reinforcing his or her subordinates. In the case of transformational leaders, J. M. Burns (1978) argued for the former point of view, but Tichy and Ulrich (1984) argued for the latter position. According to Burns, dedication, caring, and participation are multiplied outward from the leaders through their disciples; the leaders become the models to be imitated by successive expanding layers of followers. Tichy and Ulrich (1984) suggested that organizational changes envisioned by top management require lower-level managers to adopt leadership behavior supporting the practical implementation of their superiors’ vision.
Most evidence supports modeling rather than alternating. Bowers and Seashore (1966) found that leadership behavior patterns exhibited by executives in insurance agencies were reflected in similar behavioral patterns by the supervisors below them. The supervisors’ emphasis on the facilitation of goals and interaction with their subordinates was related to the extent to which the executives did the same. Similarly, Stogdill (1955) obtained data to indicate that participatory leadership at lower levels in an organization was dependent on its being practiced at higher levels. Ouchi and Maguire (1975) found that subordinates tended to use the same methods of control as their superiors in dealing with their own subordinates. Summarizing his studies of Japanese managers from the 1960s onward, Misumi (1985) reported that the supervisory style of a manager, with emphasis on either performance or maintenance, tended to be similar to the style found above and below the manager in the organizational hierarchy.
Bass, Waldman, Avolio, and Bebb (1987) collected self-rated and subordinate-rated leadership descriptions of second-line managers, their first-line supervisors, and their subordinates in New Zealand. A cascading effect of leadership behavior emerged. The amount of transformational and transactional leadership behavior observed at one level of management tended to be seen at the next lower level as well. The leadership patterns of subordinate-superior dyads tended to match each other. The correlation of the actual leadership observed among the levels was highest (r = .51) for the transactional exchange involved in providing contingent rewards.
To examine whether modeling of one’s superior made a difference in the performance of one’s group, Misumi (1985) calculated the capital growth rate in two Japanese banks with 25 and 54 branches, respectively. Pairs of high-producing and low-producing branches with similar socioeconomic characteristics were compared. First-level and second-level superiors were identified as being low or high in performance (P) and in maintenance (M) orientation. In one bank, 38% of the first-and second-level superiors matched each other in orientation in both the more productive and less productive branches; but in the second bank, 77% of the first-and second-level superiors in the less productive branches matched each other’s orientation and 33% did so in the productive branches. The probability of matching on a chance basis was 25%. If the matching was high in P and M at both levels, effects were salutary; but if the matching was low in P and M, effects were counterproductive.
The alternating approach was observed indirectly when the same performance by commanding officers of ships and their executive officers appeared to produce opposite effects. In a study of shipboard organizations, D. T. Campbell (1956) found that although the leadership scores and sociometric interaction scores of the commanding officers were positively correlated with measures of shipboard efficiency and morale, the leadership and sociometric interaction scores of the executive officers, the commanders’ closest aides, were negatively correlated with shipboard efficiency and morale.
Manz and Sims (1987) first studied self-leading and discussed its theory, constraints, and effectiveness (Manz & Sims, 1990). It could replace the leaders’ providing orders, promises, or objectives about what needed to be done. Power was to be shared with followers. Followers would provide more direction for themselves. Their commitment would be higher because they “owned” how they proceeded. The superleaders would model self-leadership for the followers, create positive thought patterns, and use contingent reinforcement of the self-leading individuals, teams, and cultures (Manz & Sims, 1991). Manz, Keating, and Donellon (1990) discussed the transition from traditional supervisor to superleadership of teams of self-leaders. Self-led teams and self-management have become popular practices and are discussed further in Chapter 26, with regard to what conditions further the success or failure of self-leadership and self-management.
The leader’s relations with followers can depend on his or her relations with others, such as superiors and peers. In many organizations, the CEO and the chairman of the board work closely together to make corporate decisions. Gronn (1999) presented the case of close coupling, for over 10 years, of the founder, J. R. Darling, and the first head, E. H. Montgomery, in leading Timbertop, an Australian school. Another prominent leadership couple is the platoon leader and the platoon sergeant. The platoon leader is usually a commissioned officer and the platoon sergeant, a noncom. The officer is in charge, but the sergeant often is more experienced. Together, they provide the leadership of the platoon. According to a study of 72 light infantry platoons in near-combat testing for 11 missions, when the relations between the leaders and sergeants were good rather than poor, the platoons were rated much higher in effectiveness by pairs of independent observers of the platoons in action (Bass, Avolio, Jung, & Berson, 1999).
Importance of Supervisors’ Influence with Higher-Ups. Pelz (1949, 1951, 1952) noted that when supervisors who had influence with their superiors took the side of their subordinates, the subordinates tended to feel more satisfied. But when a supervisor without such influence identified with the subordinates’ interests, the subordinates tended to be more dissatisfied. Closeness to subordinates and taking their side increased the subordinates’ job satisfaction only when the supervisors had enough influence with their superiors to provide conditions that could result in the fulfillment of the subordinates’ expectations. Jablin’s (1980) results concurred with Pelz’s findings, particularly for supportive supervisors with upward influence rather than nonsupportive supervisors with upward influence. Nahabetian (1969) found that in general, group members were better satisfied under leaders who had influence with their superiors than under leaders without such influence. Influential leaders were seen to facilitate the group’s task, whereas those without influence higher up were seen as hindering it. Ronken and Lawrence (1952) reported similar findings. Anderson, Tolson, Fields, et al. (1990) demonstrated the Pelz effect in a study of 195 nurses and 201 clerical employees who were more satisfied with their jobs and felt more upward control if they perceived that their superiors had more influence with higher authority.
Combined Effects of Multiple Hierarchical Levels. Misumi (1985) showed the combined effects of two layers of supervision on the performance of banking subordinates. Hill and Hunt (1970) observed that although the leadership behavior of supervisors one level removed from the employees was not related to the employees’ satisfaction, the combined behavior of first-and second-level supervisors did affect their satisfaction. Much initiative by both first-and second-level supervisors was significantly related to the employees’ satisfaction with their own esteem and autonomy. George (1995) found that the contingent reward behavior of leaders was correlated at successive levels of sales management.
As evidence of the systematic connections between subordinates, supervisors, and their superiors, Stogdill and Goode (1957) concluded that when supervisors interacted frequently with their superiors, their subordinates thought the leaders should spend more time than they did in interviewing personnel and in coordination. According to Stogdill and Haase (1957), more impersonal performance by superiors, such as inspection, kept subordinates away from superiors. When superiors spent little time in preparing procedures and much time in technical performance, their subordinates tended to interact with peers; but when superiors spent more than an average amount of time in supervision, subordinates tended to interact with them. A high rate of communicating and integrating behavior by superiors enlarged the total number of interactions initiated and received by their subordinates, increased reciprocated interactions within the subgroups, and decreased interactions with members outside the subgroups.
Stogdill (1955) studied the effects of interactions among three hierarchical levels of organization members—subordinates, supervisors, and the superiors of the supervisors. The supervisors tended to interact more with their own subordinates when their superiors interacted more with them. The interactions of these supervisors with members outside their own subgroups were affected by whether their superiors were the initiators or the recipients of interactions within their own subgroups of supervisors. If their superiors were initiators rather than recipients of interaction with the subgroup of supervisors, the supervisors tended to interact less frequently with members outside their own subgroups. When their superiors interacted frequently with members outside the subgroups of supervisors, the supervisors also interacted with members outside this unit, but the supervisors interacted less often with their superiors.
When superiors interacted with the supervisors’ peers, the supervisors tended to initiate more interactions with their superiors but received fewer interactions in return. In general, the superiors’ interactions with the supervisors induced similar patterns of interaction between the supervisors and their subordinates. The supervisors’ interactions with their superiors exerted the strongest effects in restricting the area of interaction of their subordinates.
In the previously mentioned study of three hierarchical levels of organization members (Bass, Waldman, Avolio, & Bebb, 1987), lower-level leaders who were seen by their subordinates as more charismatic, in turn, required less, not more charisma, in their superiors. It appears that charismatic leaders would rather not have a charismatic superior with whom they may have to compete.
Other Combinatory Effects. Stogdill and Goode (1957) observed also that leaders who interacted frequently with peers had subordinates who believed that their leaders had too little responsibility, were less active than they ought to be in representing their subordinates, and spent too much time in planning. When leaders interacted extensively with people outside their own units, their subordinates reported having to delegate and represent their groups too much. These followers also thought they ought to spend more time than they did in inspection, planning, and preparation. In a study of a large naval organization, Stogdill and Haase (1957) found that the more time superiors actually spent in highly personal interactions with others, such as interviewing personnel, the less their subordinates actually interacted with their peers and initiated interactions with members of other subgroups with whom their interactions were reciprocated. On the other hand, Greenberg and Barling (1999) reported that the more supervisors monitored subordinates, the more hostility and aggression were found among 136 full-time employees, particularly when the monitoring was seen as unfair workplace surveillance. In a survey of 116 employees, Fedor, Davis, Maslyn, et al. (2001) found that the sources of a supervisor’s power and the recipient’s self-esteem made a difference in how much negative feedback affected subordinate performance.
Explanations. The multilevel and falling-dominoes effects may be due to differential selection as well as modeling. Lower-level supervisors can be either self-selected, selected by their second-level manager, or selected by the organization. It is not a matter of chance that they may be stylistically compatible with their superior. It is also possible that certain leadership behaviors are reinforced by the norms of organizational subunits; therefore, the cascade effects may be due to the subculture of norms, beliefs, and values within which the leaders operate. In the same way, the environmental and technical demands in one subunit may generate common job requirements and therefore dictate the differential leadership observed and required at the different levels of the subunit. Thus Smith, Moscow, Berger, and Cooper (1969) found weak support for the hypothesis that under conditions of slow organizational change, good interpersonal relations between managers and superiors were associated with good relations between managers and subordinates. But strong support was found for the hypothesis that under rapid organizational change, good relations between managers and their superiors changed into poor relations between the managers and their subordinates. The greater need for rapid change put pressure on the managers to push their subordinates for better performance and faster response. Superiors encouraged the managers in this regard; subordinates were disturbed by it.
According to regression analysis, in eight organizations social distance and demographic divergence were important influences on 213 followers’ relations with their leader (Boccialetti, 1995). Followers come in three types, according to Boccialetti. Helpers show deference and comply with the leadership; independents distance themselves from the leadership and show less compliance; and rebels show divergence from the leader and are least compliant. Among other types of followers, moderate in compliance, are diplomats, partisans, and counselors. The types differ in how they put up with their leaders as authority figures, accept responsibility, seize initiatives, and stay informed. Barbuto (2000) formulated a framework for predicting follower compliance from the leader’s bases of power, the self-concept of the follower, the follower’s intrinsic and extrinsic motivation, and the follower’s zones of resistance to requests.
Heller and Van Til (1982, p. 405) argued that “leadership and followership are linked concepts; neither can be comprehended without understanding the other.” The compliance of followers is the mirror image of successful leadership. Just as successful leadership may be seen to influence the completion of tasks and socioemotional relations, so the compliance of followers can be seen as instrumental to the completion of tasks and both public and private socioemotional acceptance of the leadership effort. It also seems obvious that by their performance, subordinates control the nature of feedback from their superior (Jablin, 1980). In the same way, just as the leader can influence subordinates by initiatives and information, the subordinates can complete the process and influence their leaders by giving feedback to them. Hegarty (undated) demonstrated that feedback of subordinates’ ratings to supervisors resulted in positive changes in the supervisors’ behavior. The employees of the 58 supervisors in the experimental and control groups completed an information opinion survey. The survey results were used to prepare feedback reports for the experimental supervisors but not for the control supervisors. A second survey was conducted 10 weeks later to measure change. After adjusting for the initial scores, Hegarty found that all 17 measures of change shifted more in the expected direction in the experimental than the control supervisors, six significantly so. Such feedback has become common in many organizations as a means for improving the effectiveness of leaders and their operations.5
In the political arena, both immediate followers and grassroots citizens make a difference to what a leader can accomplish. Karl Rove, President George W. Bush’s close advisor, influenced the president as much as or more than the president has influenced Rove. President Bill Clinton’s Universal Health Plan of 1994 was defeated in Congress because of failure to gain public understanding and grassroots support. Active followers matter, as does their relationship to their leaders (Hollander, 1996). Followers increase their respect for a good leader and lose their respect for a bad one. A supportive, clearly communicating, and rewarding leader develops and strengthens the relationship with the follower and enhances the follower’s satisfaction and performance. Unsupportive, harsh, and demeaning leaders create poor relations with followers, who then withhold information, become passive, withdrawn, and discouraged, and consider quitting (Hollander, 1996).
Followers’ expectations affect the performance of their leaders. Followers’ perceptions of their leaders’ motives and actions constrain what their leaders can succeed in doing (Stewart, 1982b). Followers allow more discretionary opportunities for leaders who are seen to be more competent, legitimate, and expert in solving problems (Lord & Maher, 1991). From an analysis of a representative sample of 100 Swedish managers and their subordinates, Norrgren (1981a) found that the subordinates’ levels of education and aspirations affected the managers’ beliefs, intentions, and evaluations, particularly those of the younger managers. Furthermore, although older subordinates and subordinates with high aspirations and substantial seniority were most specifically favorable to-ward managers who had beliefs and intentions to allow the subordinates to participate in decisions, it was the younger subordinates who reacted most negatively to managers with beliefs and intentions that were opposed to such participation (Norrgren, 1981b).
The compliance of subordinates is not automatic; it depends on the active cultivation of the leader’s legitimacy. A leader’s management style is affected by how the subordinates respond to it. The subordinates can actively work to undermine it, or they can work hard to support it because it serves their own interests or the “greater good.” Self-interest, according to Biggart (1981), may have accounted for the strong loyalty of most of Ronald Reagan’s subordinates, despite his hands-off management style when he served as governor of California and later as president of the United States. Cabinet officers resign when they can no longer support the policies of a prime minister. As will be noted below, good followers go beyond their self-interests for the good of their group, organization, or society.
The follower’s influence on the leader also means that, contrary to popular notions, followership and leadership are highly similar, as are followers and leaders. Hollander and Webb (1955) showed that the same peers who are nominated as most desired leaders are also nominated as most desired followers. Nelson (1964a) found that among 72 men on a U.S. Antarctic expedition, the characteristics that made the men liked were about the same for the leaders as for the followers.
There are no sharp boundaries between the roles of leader and follower (J. M. Burns, 1978). Both roles always must be played in any group. Leaders cannot exist without followers; nor can followers exist without leaders. Moreover, leaders and followers exchange roles over time and in different settings. Many persons are leaders and followers at the same time. But the interaction between the leader and the follower is not symmetrical. K. M. Watson (1982) coded the antecedent acts of dyadic interactions of subordinates and leaders. When the leaders attempted to initiate structure, the subordinates were most likely to comply with deference. However, when the subordinates attempted to initiate structure, the most common reaction of the leaders was to resist by responding with efforts to try to structure the situation differently rather than to comply with the subordinates’ initiative (Boccialetti, 1996).
The Good Follower. James (1995) argued that we need to pay more attention to the good follower. Kelley (1988) enumerated many elements that go into being a good follower. These elements are similar in many respects to what makes a good leader. The differences between the effective leader and the effective follower are mainly in the different roles they play. Good followers are active, independent, critical thinkers who can manage themselves. They are committed to the organization and to persons, principles, or purposes beyond themselves. Their personal and organizational goals are aligned. They are competent and avoid obsolescence by pursuing continuing education and development. They disagree agreeably. They build credibility. They can move easily into the leadership role and return again to the role of follower. Campbell (2000) adds that the follower who is proactive is not only competent, interpersonally effective, and organizationally oriented but also has integrity and is enterprising. Such followers are good “organizational citizens” who willingly contribute to the firm’s effective functioning. They volunteer for assignments and put forth extra effort. They take the initiative to expand their role and take on new tasks. Albino (1999), a college president, agreed with Kelley (1992) that good followers are essential in higher education. Good followership is a discipline that requires going beyond self-interests. Effective leadership requires good followership. Effective leaders know how and when to follow. Laotzu suggested that to lead people, one must walk behind them. The effective leader follows the lead of the group (Rinne & Karl, 1990). According to 37 supervisors at retail stores and 58 bank supervisors, satisfactory subordinates (among a total of 274) helped to get things done.
One example is the turnaround in IBM in 1994 from a complacent manufacturer of mainframe computers and PCs that had been recognized by Fortune as a “most admired” company in the mid-1980s. In the three years preceding 1994 it had accumulated $15 billion in losses. David Grossman, a midlevel IBM programmer, was one of the first people to download the Mosaic browser and experience the Web. IBM was the highly advertised technology sponsor of the 1994 Winter Olympics, providing raw results on television. But when Grossman surfed the Web, he found that newcomer Sun Microsystems presented the results as a Sun contribution. Grossman drove to IBM headquarters with the necessary equipment to show the Internet to John Patrick, a senior marketing executive manager on IBM’s strategic task force. Grossman and an associate built a primitive intranet for IBM, and Patrick published a document extolling uses of the Web, from replacing paper communication to e-commerce, and giving every employee an e-mail address. The new CEO was highly supportive. By 1996, IBM’s Web site was able to earn $5 million from its e-commerce and launch into an effective changeover, becoming a consulting business (Hamel & Schonfeld, 2000).
The Susceptible Follower. According to Shamir and Howell (1999), followers who are more susceptible to the influence of a leader, particularly a directive and charismatic one, are likely to be unstable, uncertain, and inconsistent about their own self-concept. They do not have a clear, consistent self-concept that could guide their behavior. Furthermore, susceptible followers’ values and identities are congruent with those of their leader. They will also be more readily influenced by a leader who seemingly has the support of majority opinion.
Antitheses of Leaders. If not the followers, then who are the opposites of leaders? They are those barred from the process (for instance, the underage, who cannot vote). They are the isolates, the rejectees, and the anomic. They are those who exclude themselves from participation—the apathetic and the alienated. The apathetic may be too busy with other affairs or too busy just surviving. The alienated may reject and resist participation; they believe that the power to lead is in the hands of others, for the benefit of others. The anomic feel powerless, normless, and aimless, and see leaders indifferent to their needs (J. M. Burns, 1978). These nonleaders-nonfollowers, by their lack of involvement, can have a negative impact on their groups, organizations, and societies. Quoting deJouvenal (“A society of sheep must in time beget a government of wolves”), Gardner (1987b) noted that those who fail to follow but remain the antithesis of leadership invite the leader’s abuse of power.
Despite the asymmetry of the relationship, followers exert considerable upward influence on their leaders (Gabarro & Kotter, 1980). Followers can actively affect leaders in many important ways (Hollander, 1992). Upward feedback was provided by subordinates to their managers in a quasi-experimental Australian study. Compared with the managers’ initial performance and a control group, managerial performance was improved considerably six months later, particularly if the managers had an orientation toward the goal of learning (Heslin & Latham, 2004). Farmer and Maslyn (1999) confirmed three styles of upward influence first typed by Kipnis and Schmidt (1988): bystander, tactician, and shotgun.
Acts of moral courage by elected political leaders can be strongly pressured by constituencies, economic blocs, and organized letter writers (Paige, 1977). Reed (1996) argued that civic followers are actually leaders as they monitor the affairs of society and react accordingly. On tour, President Woodrow Wilson was first shocked, then braced, by hearing a voice from the crowd shout, “Attaboy, Woody” (Davies, 1963). After they have been elected, U.S. presidents may try to ignore their own campaign rhetoric or reinterpret it to fit their preferred policies. But when public opinion is sufficiently aroused, the presidents become responsive and usually adopt the opinion of the strong public majority as their own. In 1983, public opinion polls indicated that the public did not find much reason for placing U.S. Marines at risk in Lebanon in the exercise of a presidential policy that had been in effect for at least a year. Despite the continuing lack of public approval, days before the terrorist bombing of the Marine barracks President Reagan publicly expressed his opinion that we would never “cut and run.” Days after the bombing and the loss of life, which greatly aroused public opinion against the policy, the Marines were “redeployed offshore” (Gwertzman, 1983). President Clinton was sensitive to the results of daily public opinion polling. But like Reagan, President George W. Bush was less responsive to adverse domestic or foreign public opinion about American military forces and kept accusing the Democrats of wanting to “cut and run” from Iraq. Still, suddenly, in June 2006, Bush accepted a plan for a staged withdrawal from Iraq, which subsequently did not occur.
Upward influence is seen as an important contribution to organizational effectiveness (Gabarro, 1979) and a key to understanding organizational politics (Porter, Allen, & Angle, 1981). Subordinates have the responsibility to exert upward influence on their bosses for their mutual benefit and the benefit of the organization as a whole. Subordinates should challenge their superiors’ proposals and help them avoid mistakes, but subordinates can do so only if they have contributed to building trust between their superiors and themselves. The subordinates must provide adequate information and account for their performance in carrying out delegated assignments. Effective subordinates will actively invite review, support, and feedback from their superiors (Crockett, 1981).
The Tactics of Upward Influence. Upward influence tactics used by subordinates were identified and scaled by Kipnis, Schmidt, and Wilkinson (1980) as: assertiveness; reasoning; bargaining about the exchange of benefits; appealing to a higher authority; forming coalitions; or trying friendliness, ingratiation, and flattery.6 The list of items was refined and the subscales confirmed by Schriesheim and Hinkin (1988). For Schilit and Locke (1982), upward influence tactics included logical or rational presentation of ideas; informal exchange not related to performance, such as ingratiation or praising the superior; promising rewards or threatening sanctions; adhering to rules; manipulating matters so that the superior is unaware of being influenced; mobilizing coalitions of support among coworkers and higher-ups; and being persistent or assertive.
Supervisor-focused upward influence tactics by subordinates such as ingratiation and impression management aim to increase being liked by the supervisor. They result in receiving higher performance ratings from the supervisor (Wayne & Liden, 1995). Job-focused upward influence tactics are self-promoting and self-serving distortions to increase the appearance of competence (Dulebohn & Ferris, 1999).
Hard, Soft, or Persuasive? Like leaders, followers may use tactics of upward influence that are hard (assertiveness), soft (friendliness, ingratiation), and rational (persuasive). In a national sample of 225 mainly women employees of 60 branches of a nonprofit health and advocacy organization, Fedor, Goodman, and Maslyn (1992) found that hard tactics of upward influence were used more frequently by respondants higher in Machiavellianism; soft tactics were used by those higher in internal locus of control and self-monitoring. Followers were also more likely to use soft tactics with leaders perceived to have more control of rewards and punishments. Rational tactics were used more by followers if their leaders were valued and esteemed with referent power. Deluga (1991a) found that 102 bank employees high in self-monitoring used all three kinds of tactics for upward influence, particularly if they also were high in deceiving others. According to another survey by Deluga (1988), soft upward influence tactics such as reason and friendliness were used more frequently by 70 managers and employees from various departments when they rated their leaders as transformational. In still another survey by Deluga (1991b), 82 hospital personnel were shown to contribute to their managers’ felt interpersonal stress by using hard upward influence tactics (r = .72) and reduced such stress by using rational influence tactics (r = .52). Assertiveness, coalition, and bargaining were used less frequently. Epitropaki and Martin (2001) found that “to get their way with their manager” 271 employees in one service company and six manufacturing companies said they were more likely to use soft and rational tactics with a transformational manager and less likely to use hard tactics. They were also less likely to use hard or rational tactics with transactional leaders. According to Deluga and Souza (1991), 117 police officers reviewing scenarios were more likely to choose rational upward influence as more appropriate if the supervisor were transformational rather than transactional. Elsewhere, Deluga (1988) noted that when leaders were rated as transactional, fewer if any upward influence tactics were tried.
Impression Management. Followers practice impression management with defensive tactics such as trying to justify inadequate performance with excuses and apologies, or with proactive tactics such as self-promotion to appear competent and dependable so as to attract their boss’s attention or to obtain authorization for changes. They may also do personal favors for the boss and ingratiate themselves with the boss in other ways (Wayne & Ferris, 1990; Yukl, 1998). From 67 manager-subordinate pairs, Rao, Schmidt, and Murray (1995) obtained the managers’ appraisals of the subordinates and the subordinates’ styles of influence. Coalitions resulted in favorable impressions and appraisals of the subordinates, but subordinate ingratiation and assertiveness did not. Wayne and Liden (1995), in a study of 111 supervisor-subordinate dyads, found path coefficients of .40 and .33 between the subordinates’ supervisor-focused impression management and the supervisors’ liking of and perceived similarity to the subordinate. This in turn yielded a path coefficient of .21 for appraised subordinate performance when demographic similarity was taken into account.
Usages. All 123 salaried employees in a metal fabricating firm surveyed by Deluga (1986) used ingratiation, flattery, reason, and assertiveness much more frequently when trying to influence their superiors than they used bargaining and appeals to coworkers or to a higher authority. Deluga (1988) reported that when their first efforts to influence their transactional but not their transformational superiors did not work, 117 managers and employees of a northeastern manufacturing firm subsequently displayed significantly less friendliness and assertiveness toward their superiors and appealed less to a higher authority. According to Schmidt and Kipnis (1984), staff managers were more likely than line managers to attempt more upward influence to achieve individual goals.
Purposes. Subordinates’ goals will determine which tactics they use. To affect policy, subordinates will be more likely to use reasoning and the rational presentation of ideas (Kipnis, Schmidt, & Wilkinson, 1980; Schmidt & Kipnis, 1984). Mowday (1978) found that elementary school principals who were power motivated were more likely to attempt upward influence. In a study in India, the need for power determined the particular power strategies that 96 managers from the public sector said they used to influence their superiors. Most of the strategies they used were “soft and subtle” (Singh, Kumari, & Singh, 1988). To promote self-interests, to attain personal goals, and to obtain benefits, the Indian administrators were more likely to try ingratiation and promises for a satisfactory exchange with their superior based on their evaluation of the superior’s preferences. In looking at the specific reasons why subordinates attempt to ingratiate themselves with their bosses, Ralston (1985) argued that ingratiation was more likely to be used by Machiavellian subordinates with autocratic superiors in an ambiguous setting. In an experiment in India, Pandey and Bohra (1984) confirmed that subjects who acted as subordinates were likely to endorse ingratiating tactics with executives who liked to have admirers around them; who preferred employees who supported their views, praised their ideas, and conformed to their policies un-critically; who used subjective criteria in making decisions about employees’ requests for benefits; and who rewarded those whose behavior flattered and pleased them. In contrast, ingratiation was less likely to be suggested as a viable tactic if one was working for an executive who had the opposite tendencies.
Successful Upward Influence. Schilit and Locke (1982) asked employees and supervisors about more and less successful ways for subordinates to influence their superiors. Both subordinates and supervisors agreed that subordinates used logical presentations more than any other tactic in attempts at upward influence and agreed that the substance of the effort determined its success. But they disagreed on what caused attempts at upward influence to fail. Although the supervisors attributed failure to the substance of the attempt, subordinates attributed failure to the closed-mindedness of the supervisors. In an analysis of the diaries of 60 middle-level managers over a two-month period, Schilit (1987) found that the managers high in tested need for power thought they had more upward influence on strategic decisions. Such felt upward influence coincided with a high need for achievement, self-control, and experience with their superiors. As a whole, the managers tended to believe themselves to be successful in their attempts at upward influence, particularly in attempts involving less risky strategic decisions. Rice (1986) offered suggestions on how to cope with a difficult boss: subordinates should observe how others succeed in getting along with the boss; they should offer to be helpful; they need to keep track of their boss’s mood swings, discussing his or her mistreatment of them in private, when the boss is calm; subordinates should avoid seeking the boss’s approval when it is not required, but they should not dispute the boss’s legitimate authority even if they disagree with his or her judgments.
The Subordinate as a Source of Feedback. Considerable upward influence may occur in the course of the subordinate’s feedback to the superior. For effective superior-subordinate relations, feedback must flow both ways; a loop must be maintained. In their attitudes and behavior, leaders can hinder the flow or help close the loop from the subordinates back to them (Kaplan, Drath, & Kofodimos, 1985). Hegarty (1974) showed that subordinate ratings of supervisors could improve supervisors’ behavior. Reliability and validity of upward appraisals was likely to be higher than appraisal from a single boss, since multiple subordinates are available to rate each supervisor (Harris & Schaubroeck, 1988). Improvements in managerial performance were found over a five-year period for 252 managers who received feedback from their direct reports and discussed their previous year’s feedback, compared with managers who did not (Walker & Smither, 1999). Subordinate upward appraisals were found to be among the best predictors of managerial performance (McEvoy & Beatty, 1989).
Facilitators and Inhibitors. The willingness of 153 employees in a university affairs office to provide feedback to their supervisors was greater if they felt it was beneficial (r = .52), if they did not fear retaliation (r = .39), if they perceived organizational support (r = .39), if they felt it was appropriate to their role (r = .26), if they were knowledgeable about upward feedback (r = .23), and if they sought feedback themselves (r = .22; Smith, Kudisch, & Thibodeaux, 2000). At a research university, 411 of 1,135 faculty members who participated in anonymous upward appraisals were more confident about the accuracy of the ratings, felt they did not have to bias ratings, felt more secure in rating superiors, and perceived that more effects would accrue from the process (Westerman & Rosse, 1997).
Superiors prevent their subordinates’ criticism from reaching them by monopolizing conversations, by developing an abrasive style, by emphasizing their own power and status, and by “adopting the mantle of their office when interacting with lower level managers” (Bruce, 1986). Leaders who receive positive feedback do not seem to evaluate subordinates differently, but they do evaluate less active subordinates less favorably if the feedback is negative (Elgie, Hollander, & Rice, 1988). They avoid discussing issues until they are ready to make a decision. They distance themselves from subordinates and become isolated and insulated from feedback from below by selecting subordinates to bolster their own thoughts and feelings.
Channels of communication dry up as one moves upward in an organization. Information is increasingly filtered, as it must flow through more levels. To promote the upward flow of information, superiors need to reduce perceived differences in power, make themselves more accessible, and open informal channels (Bruce, 1986). They need to reward—not punish—messengers who bring bad news or disturbing opinions. They need to be aware that despite the greater pressure to communicate upward than downward, their subordinates are quite reluctant to risk their displeasure by being seen as critics of the current state of affairs or as bearers of unpleasant messages. Superiors can encourage systematic surveys of subordinates’ attitudes that are a formal medium for upward feedback. They can institute a policy, as did the British Army in World War II, that each commander must pay frequent visits to all the unit commanders at the level immediately below. They can practice “walk-around” management, in which managers visit employees’ work stations. This policy can promote upward feedback informally if the employees recognize that their superiors, particularly those at high levels in the organization, may feel out of touch with those at the lower levels and would appreciate receiving positive, supportive feedback as well as constructive criticism. The higher-level receivers must make it clear that they are open to accept feedback; the lower-level senders must be ready to provide feedback that is free of the distortion commonly observed in upward messages in which subordinates tell the superiors what they believe the superiors want to hear (Jablin, 1980).
In addition to such motivated distortion, there is cognitive distortion because of the different meanings that superiors and subordinates attach to the same words and their different perceptions and broader views about the same issues (Smircich & Chesser, 1981). When upward feedback was provided to supervising officers in a police agency, the extent to which they reacted positively and took steps to improve depended on their organizational cynicism (Atwater, Waldman, Atwater, et al., 2000). When subordinate feedback is given as part of a 360-degree appraisal program, the upward feedback is perceived as useful if there is organizational support for it (Facteau, Facteau, Schoel, et al., 1998).
An axiom of learning theory and counseling is that a learner needs more guidance early in training and less guidance later in training. In the same way, the novice subordinate requires more direction and the mature subordinate requires less. Less can be delegated to the novice; more can be delegated to the mature subordinate. The subordinates’ readiness and competence determine how much and what kind of leadership can be efficacious.7
Readiness and Maturity. Followers strongly affect the likelihood of a leader’s success in influencing them as a consequence of whether they are ready for the leader. J. M. Burns (1978) suggested that followers could be ripe for mobilization by a leader or, on the contrary, imprisoned by fixed beliefs that would make it impossible to lead them. In the same vein, Heller (1969a) pointed out that some level of agreement about procedures, interests, and norms among followers is necessary for effective participatory leadership. Nie, Powell, and Prewitt (1969) found five sets of attitudes that were of consequence to such readiness among political followers in Germany, Italy, Mexico, the United States, and Britain: (1) a sense of duty; (2) information about politics; (3) a stake in political outcomes; (4) a sense of political efficiency; and (5) attentiveness to politics. According to L. L. Moore (1976), fully effective leadership tends to depend on mature followership. Moore developed a maturity index based on observations of verbal and nonverbal behavior of followers in task groups. Dimensions of maturity were perspective, position, awareness, activity level, dependence, motivation to achieve, ability and willingness to take responsibility, task-relevant education or experience, and variety of behavioral interests. The leader’s knowledge of the followers’ maturity can facilitate and modify the followers’ and the leader’s behavior.
One’s readiness to accept the suggestions of leaders will depend on one’s immediately preceding experiences. For instance, negative arguments that were introduced before an attempt to persuade students to volunteer for civil defense work were found more likely to reduce the success of the positive arguments than were the negative arguments that came later (Feierabend & Janis, 1954). Another antecedent to determining the readiness of subordinates for a subsequent attempt to lead them is the extent to which the attempted leadership fits with previous experiences. Weiss and Fine (undated) found that groups who were first subjected to failure and insult were more influenced by suggestions to be punitive; groups first subjected to rewarding experiences were more likely to respond to suggestions to be lenient. Also important is the followers’ sense of security—whether the followers have a general trust of others and are comfortable with others. The insecure are anxious and avoid attachments (Hazen & Shaver, 1987). Among 127 undergraduate management students who ranked themselves as followers in their small class groups, Berson and Yammarino (1998) found that on scales of implicit leadership, the secure students ranked their ideal leader high if the leader was charismatic and considerate. They ranked low a leader who tried to initiate structure but did not lead. Insecure students ranked their ideal leader as a nonleader low in consideration. Lord, Brown, and Frieberg (1999) suggested that the working self-concepts of followers are changed by the leader through the development of schemas and are influenced by the leader’s self-schema.
Effects on the Leader of the Follower’s Interest and Competence. What subordinates seek in their jobs affects what their supervisor can and will do (Maier, 1965). Jones, James, and Bruni (1975) showed that the degree to which trust and confidence of 112 engineering employees correlated with the behavior of their supervisors was affected by whether the employees were involved in their work. If the subordinates preferred to avoid risk, if they did not wish to become involved in the task, if they were uninterested in the task, and if their interest was of no relevance to getting the job done, leaders became directive. But if subordinates wanted to be involved and were interested in what happened, more supervisory leadership occurred to engage them. Thus when the British managers who were surveyed by Heller (1969a) saw a decision as being important to subordinates but not to the company, they used a high degree of power sharing. Conversely, when a decision was a matter of concern to the company but not to subordinates, they preferred the centralization of power.
If the time required for subordinates to participate in decision making is more expensive than the value of their contribution, effective supervisors will be more directive than participative (Tannenbaum & Massarik, 1950; Lowin, 1968). On the other hand, if the supervisors value the competence of their subordinates, they will be more participative (Likert, 1959). Thus Hsu and Newton (1974) found that supervisors of unskilled employees in a manufacturing plant were more directive than were supervisors of skilled employees in the same plant. Similar results were obtained experimentally by Lowin and Craig (1968) with students who thought themselves to be in part-time jobs. When the students thought they lacked competence for the tasks to be done, they were more appreciative of close, directive supervision than when they considered themselves competent. Likewise, Heller (1969a) reported that whenever managers considered their technical ability, decisiveness, and intelligence much greater than those of their subordinates, they were more likely to use autocratic decision methods. But when the subordinates were valued for their expertise, they were more likely to be invited to share in the decision process with their supervisors.
Ashour and En gland (1971) found, in experimental teams of supervisors and secretaries, that the perceived competence of followers was the major determinant in assignment of discretionary tasks. Supervisors allowed competent secretaries more discretion than they permitted incompetent secretaries. Another effect of the competence of subordinates on superiors was seen by Kim and Organ (1982) using an in-basket simulation. “Supervisors” were more likely to initiate noncontractual social exchanges with competent “subordinates.” Scandura, Graen, and Novak (1986) showed, in this regard, that supervisors were more willing to involve competent subordinates in nontrivial decisions, particularly subordinates with whom they had good relations. However, the competence of followers could produce resistance. Thus Mausner (1954a) showed that participants with a past history of positive reinforcement in a given type of judgment were less influenced by their partners in a group-judgment situation than were participants with a history of negative reinforcement. Nevertheless, effective followers are those who, despite their own competence, can avoid being overly resistant to influence from others because of the followers’ capacity and willingness to learn from others. They listen, discriminate, and are guided by others without feeling threatened or fearing loss of status (J. M. Burns, 1978).
The competence of followers affects how they view their leaders’ initiation and consideration. Cashman and Snyder (1980) reported systematic differences in the factor structure of 475 descriptions of supervisors by more competent and less competent subordinates. Snyder and Bruning (1985) found that the competence of 815 employees of federally funded social service organizations correlated .24 with their evaluations of the quality of the dyadic linkage with their superiors. The same correlation was obtained by Bass (undated) for 220 manufacturing employees who were sorted by their supervisors into “best” and “less than best” and the employees’ satisfaction with their supervision.
Subordinates’ competence may affect their ability to comply. Numerous studies have shown, as seems obvious, that the compliance of subordinates affects their superiors’ attitudes and behavior toward them. In a laboratory experiment, Price and Garland (1981) demonstrated that when the group members were low in competence, they were much more willing to comply with their leader’s suggestions. Nevertheless, in some situations, compliance can suffer because the member is unable to comply, rather than unwilling. Kessler (1968) hypothesized that highly rated subordinates in a governmental agency would be highly motivated to act in accord with their superiors’ expectations; instead, Kessler found that these subordinates were most independent. Consistent with this, Simmons (1968) observed that in a business game, students who were rated as very low, low, or average in performance complied more closely with the managers’ expectations than did those who were rated high in performance. Evidently, the competent, highly valued subordinates felt greater freedom than did the less highly valued subordinates to deviate from managers’ expectations.
Effects of Followers’ Good or Poor Performance. The quality of the subordinates’ performance—good or poor—has obvious effects on whether their leaders will be supportive or punitive toward them (Podsakoff & Schriesheim, 1958a, b). The evidence in this regard is overwhelming. Barrow (1976) and Herold (1977) found that leaders rewarded good performers and behaved more punitively toward poor performers. Farris and Lim (1969) agreed. They divided 200 male graduate management students into 50 groups to play a business discussion game. Some leaders, appointed at random, were told that they had high-producing groups, and other leaders were told they had low-producing groups. Leaders who were told that their groups were high producers were significantly more likely to be seen by their groups as sensitive, nonpunitive, maintaining high standards, exerting less pressure to produce, allowing freedom, and emphasizing teamwork than were leaders who were told their groups were low producers. The subordinates in the high-performance condition were better satisfied, felt they had more influence, and described their groups as more cohesive.
In a laboratory study, Lowin and Craig (1968) artificially established the level of subordinates’ performance as either high or low. “Poor” performing subordinates tended to elicit from their appointed leader close supervision with frequent directions and checking. The ideas of the “poor” subordinates were ignored. These subordinates were held closely to prescribed procedures, viewed by their leader as irresponsible, and treated with less consideration. The laboratory “supervisors” of the “poorly” performing subordinates criticized them more for their work and for taking unauthorized breaks, ordered them to return to work, and showed them less support than did supervisors of “high-producing” subordinates. In studying the behavior of 82 coaches of boys’ baseball teams, Curtis, Smith, and Smoll (1979) found that compared with the behavior of winning coaches, proportionally more of the behaviors of coaches on losing teams were reactions to players’ mistakes and misbehaviors. (Of course, players on losing teams usually make more mistakes.) But more important, players perceived the coaches of losing teams to be more punitive and less supportive than winning coaches. With teams of older boys, such perceptions by team members became more significant to their continuing performance as winners or losers.
Lanzetta and Hannah (1969) found that trainers were rewarding when trainees responded correctly, but the trainers were punitive when the trainees responded incorrectly. Bankart and Lanzetta (1970) found that reward and punishment were systematically related to observed performance; rewards were given for good performance and punishments for poor performance. Kipnis and Vanderveer (1971) concluded that “managers” in a simulated work setting rewarded superior performers more than average performers and average performers more than poor performers. Fodor (1973a) also reported that superior performers were given more rewards than average workers. Chow and Grusky (1980), in a laboratory study, observed that supervision was less close for productive workers than for unproductive workers. Sims and Manz (1984) studied laboratory subjects who served as the appointed supervisors of “subordinates” who actually were confederates of the experimenters. The subjects tended to provide positive verbal rewards for the “subordinates’” good performance and punitive and goal-setting verbal comments when the “subordinates’” performance was poor. More of the subjects’ supervisory activity dealt with clarifying how the task could be accomplished than with evaluations of the subordinates’ efforts.
Continued punishment may lead to a downward spiral in performance. Moreover, learned helplessness can occur. A particular area of poor performance will be dealt with even more punitively than usual if a subordinate is generally viewed as a poor performer (James & White, 1983). Overall performance will deteriorate further (Peterson, 1985a). However, there is some buffering against the downward spiral. Rothbart (1968) found that when the use of rewards in a prior trial in an experiment produced improved performance, the administrator had a strong tendency to use reward again. But if punishment on a previous trial led to increased performance, it was unlikely to be used in the next trial.
Effects of the Causes Supervisors Attribute to Subordinates’ Performance. Calder (1977) and Green and Mitchell (1979) proposed that to understand how a subordinate’s performance affects a supervisor’s reactions requires a determination of the cause of the subordinate’s good or poor performance as identified by the supervisor.8 Four causes were seen as possibilities: competence, effort, luck, and external causes within or outside the subordinate’s control. Competence and effort were causes internal to the subordinate; luck and uncontrollable causes were external. Knowlton and Mitchell (1980) demonstrated, with 40 undergraduates who were ostensibly supervising confederates of the experimenter, that if supervisors attributed the good performance of their subordinates to the subordinates’ efforts, they evaluated the subordinates more highly than ordinarily. But this did not happen if the good performance was attributed to the subordinates’ competence. Ilgen and Knowlton (1980) reached similar conclusions. Mitchell (1981) and Brown and Mitchell (1986) found that supervisors were more likely to blame subordinates for a poor performance and its consequences than for external causes. In doing so, the supervisors gave more weight to information about the subordinates’ lack of effort than to the subordinates’ lack of ability.
For supervisors in a role-playing simulation conducted by Liden and Green (1980), the poor performance of subordinates in missing a deadline was described as being due either to their taking an overextended lunch break (an internal cause) or to a delay in receiving materials (an external cause). The supervisors attributed taking too much time for lunch to the subordinates’ personal character and failure to take responsibility for their action. Supervisors were more intense in their punitive reactions to the subordinates who missed the deadline because they took a long lunch. The supervisors thought they should do something more punitive than just talking to the subordinate about the matter, but the subordinates thought that such a discussion was sufficient. The supervisors were less likely to carry out stated corrective policies when the failure was due to an external cause—the delay in receiving materials.
In similar studies, supervisors generally were found to be more punitive when they attributed poor performance to internal causes—subordinate ability or effort—rather than to external causes. If lack of effort was the diagnosed cause, punitive action was more likely than if lack of ability was seen as the cause. If lack of ability was regarded as the cause of the subordinate’s poor performance, then training or replacement of the subordinate was indicated (Mitchell & Wood, 1980; Wood & Mitchell, 1981). Employees who were described as succeeding because of their ability were seen to have more potential for top management positions than those who were thought to be successful because of their efforts, external conditions, or good luck (Heilman & Guzzo, 1978).
Brown and Mitchell (1986) reported that attributions of internal causes increased if poor performance was detected in just one of several employees. The employees did not have to be working on interdependent tasks for this to happen. However, in a laboratory experiment with undergraduates, Offerman, Schroyer, and Green (1986) showed that leaders advocated reward or punishment for actual success or failure rather than for the reasons they attributed for the subordinates’ performance. The leaders also increased their talking to subordinates whose success they attributed to luck, as well as to subordinates whose poor performance was attributed to a lack of effort. Attributions by leaders of the causes of followers’ failures also affected their talking and negative comments (Offerman, Schroyer, & Green, 1998). Heneman, Greenberger, et al. (1989) found that the reported quality of leader-subordinate exchanges by 188 supervisors in 37 organizations was increased by how much they attributed competence and effort to their subordinates’ performance, but not by how much they attributed the performance to luck or other external causes.
Effect of Supervisory Experience. In a laboratory experiment and in a field study with military officers, Mitchell and Kalb (1982) found that the superiors’ experience made a difference. Experienced supervisors were more likely to emphasize external causes than internal causes as the reason for the subordinates’ poor performance. Liden (1981) obtained similar findings for bank managers’ explanations of the poor performance of subordinates; again, more experience resulted in more external attributions. On the other hand, Mitchell and Kalb argued that if the subordinate had performed poorly on the same task before, if the subordinate also performed poorly on other tasks, and if coworkers were performing well on the task, supervisors would be expected to attribute the subordinate’s poor performance to the subordinate rather than to external causes, and to act accordingly. Gioia and Sims (1986) demonstrated with 24 experienced managers, each interacting with four different subordinates in simulated performance appraisal interviews, that the interviews resulted in a more lenient appraisal of the subordinates afterward. The managers attached less blame to their subordinates for failure and more credit for success after the meeting than before it.
Effects of Subordinates’ Accounts. If subordinates apologize or provide an explanation of the external causes for their poor performance, they can affect their supervisor’s interpretation and reaction. In an experiment with 109 nursing supervisors, Wood and Mitchell (1981) showed that although both tactics may work to reduce the supervisor’s blame of the subordinate, explanations were much more effective than were apologies. In another study, Mitchell and Liden (1982) established that a supervisor’s punitive action for a subordinate’s poor performance might be moderated by the subordinate’s social skills and popularity. Lenient reactions of supervisors, according to a study in Israel by Bizman and Fox (1984), may result because supervisors see good acts by subordinates—such as avoiding waste; not squandering resources; performing duties with precision and thoroughness; and making efficient and practical use of information—as being more stable in occurrence than bad acts by subordinates, such as setting up difficulties in planning and organizing work, creating unpleasantness, and not following supervisors’ instructions.
Martinko and Gardner (1987) traced different likely outcomes from the combined attributions of leaders and subordinates about the subordinates’ failure to perform adequately. If both the leader and the subordinate consider that the subordinate’s failure was due to bad luck, the leader will take no action and the subordinate will not change. If the leader attributes the subordinate’s failure to a lack of effort, if the subordinate does likewise, and if the leader reprimands the subordinate, the subordinate will increase his or her effort. However, if subordinates continue to perform poorly or if the leader continues to punish them, they will attribute the continued failure to external conditions, including the leader’s punitiveness.
Effects on Leaders of Compliance or Noncompliance. It is evident from the previous section that leaders tend to react to their subordinates’ compliance or non-compliance. As expected, Hinton and Barrow (1975) found that supervisors rewarded subordinates for compliance with the supervisors’ requests and were punitive when subordinates did not comply. Fodor (1974) demonstrated, also as expected, that leaders became more authoritarian when faced with a disparaging and disruptive subordinate. Naturally, they rated such a subordinate lower and gave the subordinate less pay than they did subordinates who did not engage in such behavior. In an experiment by Crowe, Bochner, and Clark (1972), male management students were asked to play the role of a leader. Each was confronted with accomplices of the experimenters who served as subordinates and acted either democratically or submissively. “Democratic” subordinates showed initiative by putting forth ideas and trying to set their own goals. “Submissive” subordinates avoided taking any initiative and asked for detailed instructions, which they followed without question. The student subjects’ leader behavior was affected accordingly. For example, the students were more autocratic with the submissive subordinates and more democratic with the democratic ones. Chow and Grusky (1980) found that in laboratory-simulated organizational settings, subordinates’ compliance generated employee-oriented supervision, whereas subordinate’s aggressiveness resulted in punitive supervision.
In a longitudinal field survey, Greene (1979a) reported that the subordinate’s compliance exerted considerable subsequent influence on the supervisor. If the subordinate was more compliant and performed better early on, the supervisor in the following three months displayed more considerate, supportive, and participative leadership. However, if the subordinate’s performance was poor early on, the supervisor initiated more structure and role clarification later. Poor subordinate performance, particularly noncompliance, generated the leader’s greater use of punishment three or six months later.
Better supervisors in real life tend to look for and to reward independent action in subordinates, whereas less effective supervisors tend to reward conformity and group action. The better supervisors are also less lenient (Kirchner, 1961; Kirchner & Reisberg, 1962).
Effects of Contrasting Subordinates. A supervisor’s reactions to a subordinate’s compliance appears to depend on the failure of other subordinates to comply. In a field study of the performance of clerical workers, Grey and Kipnis (1976) found that the supervisors’ evaluations of compliant workers were higher in work groups with more noncompliant workers. Evaluations of noncompliant workers tended to be lower in work groups with more compliant workers. Goodstadt and Kipnis (1970) and Kipnis and Vanderveer (1971) found that in the presence of a poorly performing subordinate with a bad attitude, subjects allocated nearly twice as many pay raises to complaint workers as they did when the poorly performing subordinate was seen as simply inept, rather than disparaging.
The concept of leader-member exchange (LMX) originated with the phenomenon that leaders of a group or work unit tend to perceive each member as in-group or out-group. The leaders pay more attention to the inner core and maintain different relations with the “inners” and “outers.” They “incorporate some members into the inner life of an organization but exclude others” (Spar-rowe & Liden, 1997). Leaders’ relations (LMXs) with the inner and outer circles differ. They evaluate members of the inner circle less critically (Duarte, Goodson, & Klich, 1994). According to LMX theory, the development of the members’ roles will be differentially reinforced because the leader does not have the time to give all members equal attention. Furthermore, the leader differentially values subordinates and fosters the success of those he or she values most (Chassie, 1984). The leader also initiates more exchanges with highly competent subordinates (Kim & Organ, 1982) and establishes a close relationship with only a few key members—the inner group. Subordinates who are similar to their leader in the leader’s perceived use of power, and in the subordinates’ use of negotiation, choose to enhance their closeness to the leader. Those not in the inner circle use distancing tactics with leaders seen as dissimilar, according to an analysis of 230 supervisor-subordinate dyads by Townsend and Jones (2000). For dealing with the outer group, the leader relies on formal authority (Graen, 1976). Hollander (1978), among others, observed that it is commonplace in groups and organizations for an inner clique to form with whom the leader has closer relations than with the rest of the group. The inner clique gets more attention and more approval and possibly more status but is expected to be more loyal and committed to the leader and the group. The envy of members of the outer circle is likely to be aroused, and the inner clique must share more blame for the leader’s failures. According to May-field and Mayfield (1998), the performance of members of the inner group is about 20% higher than that of members of the outer group.
Differences in Influence. Dansereau, Graen, and Haga (1975) interviewed 60 leaders and their individual subordinates four times over nine months. For members of the inner group, the leader-subordinate dyadic exchange was seen as a partnership that was characterized by reciprocal influence; extracontractual behavior; mutual trust, respect, and liking; and a sense of a common fate. For the members of the outer group, the exchange was characterized by downward influence, role-defined relations, and a sense of loosely coupled fates. The leader was seen as an overseer. The higher quality of the vertical-dyad linkage for the inner group than for the outer group was correlated with the leader’s greater attention to the inner group. The energy and effort of the inner group were greater; they had fewer job problems. Duchon, Green, and Tabor (1986) found that members of the inner group were more compatible with the leader and had better relations with the leader than did members of the outer group, although outer-group members were not necessarily less satisfied with their leader or with their influence. Graen and Schiemann (1978) obtained results indicating that leader-member dyads in the inner group agreed more on the meaning of shared experiences than did leader-member dyads of the outer circle. In a study of first-level supervisors and their immediate superiors by Liden and Graen (1980), the outer-group members reported spending less time in decision making and were less likely to volunteer for special assignments and extra work.
Effects. Where outcomes have been based on criteria other than self-reports of the leaders and the members, the findings about membership in the inner and outer circles have been mixed. For instance, when the performance of the inner and outer circles was measured objectively, the expected differences were not found among bank tellers or for enlisted airmen completing small-group tasks. Nor was there a significant correlation between the quality of the leader-subordinate relationship and the turnover of the bank tellers (Vecchio, 1982, 1985). Rosse and Kraut (1983) studied 433 managerial dyads and found that the exchange that gave subordinates considerable latitude in negotiations was positively correlated with the subordinates’ job satisfaction and was negatively related to their having job problems. But these researchers failed to confirm other predictions about job performance.
The quality of LMXs and productivity did improve as expected with special training among 106 governmental form-processing employees (Graen, Novak, & Sommerkamp, 1982). Also, for 45 supervisor-subordinate dyads in a business organization, Vecchio and Gobdel (1984) found that leader-member dyadic exchanges in the inner circle were of a higher quality and were associated with the members’ higher performance ratings, better actual job performance, lower tendency to quit, and greater satisfaction with supervision. Inner-group members become trusted lieutenants. For a sample of 83 administrators in a large midwestern hospital, Mael (1986) showed that such lieutenants had more latitude in negotiations in their dyadic exchange relationship with their leader and were more satisfied than were others with their job and their leader’s performance.
Graen and Scandura (1987) proposed a life-cycle model for a high-quality relationship of superior and subordinate. First, leader and member of a group or work unit test the attitudes and resources each has to offer the other and what to expect from each other. Then mutual trust, respect, and loyalty are developed. Finally, the relationship matures into mutual commitment. The relationship begins with the self-interest of transactional leadership and matures into the transcending self-interest of transformational leadership Although the relations may be different with each subordinate, the leader can maintain mutual trust and loyalty with all of them. The leader develops a different partnership with each subordinate. The dyadic relations grow into an organizational system (Graen & Uhl-Bien, 1991). Leadership becomes less important than dyadic relationships to the system (Graen & UhlBien, 1995). Schriesheim, Castro, and Cogliser (1997) found 182 journal publications, dissertations, and other research reports on LMX between 1972 and 1997, beginning with Graen, Dansereau, and Minami (1972).
To quantify the quality of the exchange, the leader rates the quality of the relationship with each member on a set of items. Each member does the same for the relationship with the leader. (See Scandura, Graen, & Novak, 1986, for an example.) Unfortunately, over the three decades of research and application of LMX, the number of items that have been the basis of LMX scores has ranged from 1 to 40, and the items have involved at least six kinds of content: liking, mutual support, trust, latitude, attention, and loyalty (Schriesheim, Castro, & Cogliser, 1997). In all, at least 16 different versions of LMX measurement have been used in reported studies. However, in 36 of 86 studies meta-analyzed by Gerstner and Day (1997), seven items were used. This version is increasingly the favorite. Although Scandura, Graen, and Novak (1986) found only a correlation of .24 between the LMX ratings of leaders and their subordinates, Deinisch and Liden (1986) noted that three dimensions were needed to account for LMX: contribution, loyalty, and affection. Graen and Wakabayashi (1992) obtained a different three-dimensional factor structure to describe the necessary content of LMX: respect, trust, and obligation. Liden & Maslyn (1998) created a multidimensional measurement of four factors (LMX-MDM): affect, loyalty, contribution, and professional respect. The average relationship for the leader (ALX) ordinarily is not as informative as LMX, nor predictive of various outcomes, since the leaders have different relations with different subordinates. Some relations are mainly role-defined and downward in influence; others involve reciprocal influence with mutual respect and trust. There may be less agreement than desired between a supervisor and a subordinate about the quality of their relationship, but according to a survey of 166 newly hired university employees and their direct supervisors, LMX appears to be reasonably stable. Employee LMX ratings made at the end of two weeks and six weeks, respectively, correlated .66 and .47 with their LMX ratings at the end of six months. For supervisors, the comparable correlations were .65 and .38 (Liden, Wayne, & Stilwell, 1993).
There is a need to obtain LMX from the point of view of the leader as well as the subordinate. Gerstner and Day (1997) found an average uncorrected correlation of .29 based on the leader as the only source of information about LMX and the subordinate as the only source of information. According to analyses of 422 matched supervisor-subordinate dyads by Ford and Greguras (2001), incremental validities were obtained beyond using only those from supervisor or from subordinate in predicting such criteria as subordinate job involvement by using both supervisor and subordinate ratings of LMX.
The “quality” of the leader-member exchange—the satisfaction of either or both parties with it—should be a determinant of subsequent outcomes. A high-quality LMX is a social exchange based on high trust, respect, support, mutual obligation, and latitude. A low-quality LMX is a social exchange that depends on a formal contract providing service for reward or avoidance of punishment (Dienesch & Liden, 1986). For 41 LMX dyads studied by Liden and Graen (1980), members with higher-quality relationships with the leader assumed more responsibility for their jobs, contributed more to their units, and were rated higher in performance than their colleagues who reported lower-quality relationships with their leader. According to Dienesch and Liden (1986), the quality of the exchange may be affected by the mutual trust of the leader and the member, mutual loyalty, their influence on each other, the competence of one or the other, the perceived equity of the exchange, and the interpersonal attraction of the leader and member.
Dimensions of Quality. Dienesch and Liden (1986) suggested three theoretically and methodologically appropriate dimensions that should be considered:
(a) Perceived contribution to the exchange—perception of the amount, direction, and quality of work-oriented activity each member puts forth toward the mutual goals (explicit or implicit) of the dyad; (b) Loyalty—the expression of public support for the goals and the personal character of the other member of the [leader-member exchange] dyad; … (c) Affect—the mutual affection members of the dyad have for each other based primarily on interpersonal attraction rather than work or professional values. (pp. 624–625)
Liden and Maslyn (1998) obtained three factors—contribution, loyalty, and affect—in validation studies of LMX with 302 working students and 249 employees, along with a fourth factor, professional respect. The three dimensions also appeared to be conceptually distinct, according to judges, but were empirically correlated (Dienesch, 1985). Each of the three dimensions would be expected to have a different impact. Mutually perceived contributions should result in the undertaking of more challenging assignments and joint efforts. Mutual loyalty should be reflected in more shared confidences. Interpersonal attraction should generate greater warmth in the workplace and emotional support for nonwork problems.
Following Sparrowe and Liden’s (1997) suggestion, Uhl-Bien and Maslyn (2000) identified positive and negative reciprocal relationships in 36 work groups with 29 managers and 280 subordinates in a sample from 1,100 employees in a division of an international firm. These relationships included immediacy (time for reciprocating in an exchange between manager and subordinate), equivalence (elements of equal value are exchanged), and motive (the exchange involves self-interests, mutual interests, and other interests). Negative reciprocity is characterized by extreme equivalence, extreme immediacy, and self-interest; balanced reciprocity by high equivalence, high immediacy, and mutual interest; and generalized reciprocity by low equivalence and low immediacy and concern for others (Sparrowe & Liden, 1997). The reciprocity of “strangers” ordinarily needs immediate equivalency and self-interest. The reciprocity of partners needs mutual interests and tolerates lengthy delays and only in-kind equivalence (Uhl-Bien & Maslyn, 2000).
Effects of the Quality of LMX. The quality of the leader-member exchange and the satisfaction of either or both parties with it should be determinants of subsequent outcomes of joint efforts. For example, for 41 such leader-member dyads studied by Liden and Graen (1980), members with higher-quality relationships with the leader assumed more responsiblity for their jobs, contributed more to their units, and were rated higher in performance than were their colleagues who reported lower-quality relationships with their leader. In a large firm when the leaders and subordinates were high rather than low in self-efficacy, LMX was found to be higher in quality among 56 leader-subordinate dyads. Performance was also better (Murphy and Ensher, 1999).
Gerstner and Day (1997) concluded from their meta-analysis that LMX quality was related to better job performance, higher overall satisfaction, more commitment, less role conflict, more role clarity, competence of leaders and followers, and less intention to quit.
Compared with low-quality LMX, high-quality LMX was often found, along with better task performance (Dansereau, Graen, & Haga, 1975), conducive to subordinate innovation (Basu, 1991). Murray and Markham (undated) obtained results with 25 managers paired with their 110 subordinates showing that high-quality LMX was related to performance effectiveness when there was agreement about this quality within each manager-subordinate dyad. Tierney, Farmer, and Graen (1999) found that among 191 nonclerical employees in R & D, those with higher LMX scores earned higher ratings for creativity from their supervisors, submitted more high-quality “invention-disclosures,” and did more research reports. (Results were moderated by the employees’ cognitive styles.) Deluga (1991) showed that among 376 managers, professionals, and white-and blue-collar personnel, higher LMX correlated with more leader effectiveness and greater satisfaction with the leadership. The Gerstner and Day (1997) meta-analysis reported a mean corrected correlation of .41 for 17 studies of LMX scores and organizational commitment, as did Green, Anderson, and Shivers (1996). Gerstner and Day also reported a mean corrected correlation of 2.35 for 10 studies of LMX and role conflict, .34 for 14 studies of LMX and role clarity, and only 2.13 for LMX and turnover. But mixed and insignificant results were found for LMX and turnover by Vecchio & Norris (1996).
Individual Differences. Kinicki and Vecchio (1994) reported a correlation of .32 between LMX and members’ locus of control. They also found a correlation of .48 of LMX with felt time pressure. Murphy and Ensher (1999) observed that among women supervisors, self-efficacy correlated with subordinates’ LMX. House and Aditya (1997) suggested that LMX would probably be lower among members with a low need for growth and authoritarianism, and with a desire for autonomy. Basu and Green (1995) found that good “organizational citizen” behavior was correlated with high-quality LMX. However, the meta-analysis by Gerstner and Day (1997) noted a lack of correlation between antecedents such as age, education, and sex and quality ratings of LMX. Member competence correlated only .12 with high-quality LMX in 13 studies. As might be expected, mutual liking by members was highly predictive of LMX (Engle & Lord, 1997).
Expectations and Effort. Liden, Wayne, and Stilwell (1993) demonstrated for 166 supervisor-subordinate dyads that mutual expectations of competence and future achievement correlated with high-quality LMX. Maslyn and Uhl-Bien (2001) showed with 232 manager-subordinate dyads that higher-quality LMX was reported by subordinates when the judged effort of the leader correlated with the subordinate’s own effort. But according to the leader, if LMX was high in quality as seen by the leader, and leader effort correlated with subordinate effort, leader effort could be lower with higher LMX. If one’s effort to develop quality relations with a partner was high, and the partner’s was low, the quality of the relationship would be low. In a Dutch firm, 170 employees with an orientation of mastery displayed higher job performance, innovation, and job satisfaction. In contrast, those employees with a performance orientation were the same or lower in these outcomes (Janssen & Yperen, 2004).
Communications. A total of 537 students viewed videotapes of the facial expressions of leaders giving simulated performance appraisal feedback to subordinates. The participant rated the likely LMX quality of the consequences of positive and negative communications and coordinated actors’ facial expressions. Positive performance evaluations and “message-congruent” facial expression yielded ratings of higher-quality LMX (Ashkanasy & Newcombe, 2000). Kacmar, Witt, and Zivnuska (2001) conducted three experiments showing that subordinates in a high-LMX relationship who communicated frequently with their supervisor received higher performance ratings. Subordinates in low-LMX relationships interacted less frequently with their supervisor. Those with high LMX scores also were more likely to communicate by telephone and face-to-face than by memos and e-mail. When LMX relationships were poor according to the supervisors, the subordinates made more use of memos and e-mail than telephone and face-to-face communications with the supervisors. Communications between leader and subordinate increase in frequency and quality if LMX is high rather than low (Crouch & Yetton, 1988).
Tactics. The quality of the exchange relation between a leader and followers (LMX) affected the upward tactics used by the followers. In a sample of 376 employees, when the exchange relation with their leader was lower in quality, they were more likely to use hard tactics: assertiveness, cultivation of relations with higher authority, and coalition with colleagues (Deluga & Perry, 1991). Upward influence tactics such as ingratiation and impression management by subordinates added to the positive contribution of LMX to performance and commitment in an analysis of 150 manager-subordinate dyads of 75 managers. Each manager was coupled with one highly performing and one poorly performing subordinate (Schriesheim, Castro, & Yammarino, 1994).
Similarity. Similarity of attitudes and education of supervisor and subordinate correlated with LMX among dyads composed from 223 subordinates and their 58 supervisors (Basu & Green, 1995). Similarity of cognitive style correlated with LMX for 142 pairs of managers and their subordinates in two British manufacturing organizations (Allison, Armstrong, & Hayes, 2001). Likewise, similarity of ratings about supervisors’ power and subordinates’ negotiation in 230 supervisor-subordinate dyads correlated .77 with LMX according to the subordinates and .23 with LMX according to the supervisors (Townsend & Jones, 2000). Ashkanasy and O’Connor (1997) obtained similar findings for 160 members of 30 Australian work groups between LMX and the congruence of five personal values. For 42 dyads of different racial or ethnic background and 67 of mixed gender, LMX correlated .71 with perceived similarity of subordinates’ and leaders’ values and attitudes, .31 with speed of forming the relationship, and .33 with the similarity in age of the leader and member The same investigators obtained a correlation between LMX and speed in forming the relationship among dyads of 83 sales representatives and their 13 managers (Hepperlen & Reiter-Palmon, 2000, 2001). In all, in their meta-analysis, Gerstner and Day (1997) reported that for 28 studies, LMX correlated .34 with agreement by the leader and member.
Person-Organization Fit. To assess how closely schoolteachers’ personal attitudes fit those of their school, Erdogan, Kraimer, and Liden (2002) calculated the correlation of the individual culture profiles of 524 teachers with the organizational profiles of their 30 schools. The fits were uncorrelated with LMX of teacher and supervisor. However, when LMX was low, job and career satisfaction remained high if the fit between teacher and school was good. When LMX was high, job and career satisfaction could remain high even if the fit between teacher and school profile was low: LMX could make up for the lack of fit in personal attitudes with the school norms.
Health. For 312 retail and health care managers and professionals, Rose and Nelson (1998) found that LMX correlated positively with job satisfaction (.63) and negatively with job burnout (–.39), medical problems, personal impatience and irritability (–.24).
“Organizational Citizenship” Behavior. Manogran and Conlon (1993) and many others have found that high-quality LMX is correlated with organizational citizenship behavior (OCB). A meta-analysis by Hackett, Farh, and Song (2003) found an average correlation of .29 between LMX and OCB. High-quality LMX was also associated with more rapid movement up the corporate ladder (Wakabayashi, Graen, Graen, et al. 1988) and with more favorable perceptions of the organizational climate (Kozlowski & Doherty, 1989). But contrary to expectations, felt empowerment did not mediate the correlations of LMX and organizational commitment and other such outcomes, according to a survey of 337 employees and their immediate superiors (Liden, Wayne, & Sparrowe, 2000). The meta-analysis by Gerstner and Day (1997) reported for 10 and 30 studies, respectively, a mean corrected correlation of .37 of LMX and performance ratings according to leaders and .29 according to members. Mean satisfaction with supervision correlated .74 with LMX in 25 studies. Effects last and remain strong (Crouch & Yetton, 1988; Mayfield & Mayfield, 1998).
Leadership Style. The quality of LMX was higher if there was a match between leaders’ and subordinates’ styles of implicit leadership (Engle & Lord, 1997). Kuhnert and Lewis (1987) contrasted the quality of leader-member relationships in terms of three levels of the leader’s perspective—how meaning is made from the regularities of social interactions. At the lowest level, leaders are transactional and self-interested. At the middle level, they focus on mutual rewarding, looking out for both their own and their subordinates’ interests; there is positively reinforcing interaction between leaders and subordinates. The highest level is transformational: the interaction goes beyond the self-interests of both parties. Howell and Hall-Merenda (1999) reported positive correlations of LMX with both transformational and contingent reward leadership but not management by exception. Transformational, especially charismatic and individually considerate leaders, had higher-quality exchanges (Deluga, 1991; Deluga & Perry, 1991). Deluga (1992) found that LMX correlated with individualized consideration, a component of transformational leadership. Schriesheim, Neider, and Scandura (1998) found that when LMX was higher in 106 dyads, delegation by leaders was higher, as was subordinate satisfaction and performance. They also showed, for 183 dyads of managers, that rated performance, promotion, and salary progress were greater with higher LMX scores—independent of mentoring, which also contributed to the criteria of managerial success. In India, Aryee, Tan, and Budhwar (2002) obtained results for 161 subordinates and their 40 supervisors indicating that if LMX was of high quality, the correlation with willingness to initiate actions was higher (.36), particularly when subordinates felt a sense of autonomy.
Given the continuing problems with the meaning of LMX, the variations in the content of its measurements, the variations in its effectiveness outcomes, biased design flaws, the need to control for individual differences, and the need to consider individual differences and numerous mediating variables, many questions about LMX still require answering, although methods and measurements are converging. House and Aditya (1997) concluded that LMX and performance outcomes may reflect common biases. The results depend as much on the characteristics and behavior of the subordinates as of the leaders. Thus LMX may be a better explanation of the development of superior-subordinate relationships than of the effects of the leaders on their subordinates. Also, LMX theory may be about dyadic relationships and their subjective consequences rather than about leadership effects on followers. The positive relations of LMX and organizational citizenship behavior may be over-stated. One question not fully answered concerns the level of analysis.
The Case for Group Analysis. Some investigators remain unconvinced that much is added by substituting dyadic relationships for the group’s average relationship with the leader. For example, Fujii (1977) found that analyses of leader-subordinate relations in an experimental setting based on dyads did not differ greatly from analyses based on mean group results. Proponents of leader-group exchange argue that most measures remain reliable and valid even though they are couched in terms of leader-group rather than leader-member relationships (for example, “He shows the group what to do,” not “He shows me what to do”; Schriesheim, 1979a). Variations among members within a group are merely individual differences in describing the same leadership behavior (Nachman, Dansereau, & Naughton, 1985). In a study of public utility employees, Schriesheim (1979a) found little difference in the correlations of descriptions of leader-member relations and outcomes, compared with correlations of leader-group descriptions and these same variables. Schriesheim concluded that distinguishing between leader-member and leader-group measurements, although theoretically meaningful, may have little prac tical utility; however, he favored continuing to use available leader-member descriptions rather than leader-group descriptions for theoretical purposes.
The Case for Dyadic Analysis. Graen (1978) and Graen and Cashman (1975), among others, provided evidence to support the importance of studying the dyadic exchange relationship (LMX) instead of the average exchange (ALX). Graen, Liden, and Hoel (1982) showed that a group analysis alone might fail to detect important outcomes. For example, in an information systems department, although the average way a leader related to the group did not predict the turnover of group members, the quality of the dyadic leader-member exchange did predict the tendency of 20 out of 48 individual members to quit. Ferris (1985) replicated this finding and showed again that the differences of the dyadic-exchange quality scores of the individual members within the groups predicted turnover. But the quality of the average exchange relationships of the groups and the leaders did not.
Graen, Orris, and Johnson (1973) found that supervisors established effective dyadic exchange relationships with some newly hired employees and established ineffective relationships with others. Early on, the supervisors thought they could predict which newcomers would stay and which would leave within a few months, and they fulfilled their prophecy by acting on it. They invested most of their time and energy in the development of those who they thought would stay. The newcomers who were expected to leave early remained unclear about what their supervisors expected of them, whereas those who worked for the same supervisors but were expected to remain were quite clear after the first week of employment. These results could not have been captured by analyzing only the supervisor’s relation to the group of newly hired employees. Crouch and Yetton (1988) looked at 323 manager-subordinate dyads in Australia in which 78 managers were involved. The same manager maintained a different relationship with subordinates, depending on the subordinates’ performance. High-performing subordinates whose experience with the managers was friendly had more task contact with the managers; low performers had less contact and experienced less friendliness.
In the previously mentioned nine-month study of a management hierarchy by Dansereau, Graen, and Haga (1975), most of the dyadic relationships were new at the beginning of the study. The same manager developed different dyadic relations with different subordinates. These relations ranged from mentor-protégé to overseerpeon. The quality of each dyadic exchange relationship remained stable from the first month onward and could forecast career outcomes. Graen and Cashman (1975) showed over a nine-month period of study that the different quality of the dyadic relationship of different subordinates with the same superior determined the role assignments of the subordinates. Subordinates who had a high-quality relationship with the manager carried out less routine activities. They had more responsible administrative discretionary opportunities. They also had more resources at their disposal. The dyadic exchange was not just in the perceptions of the superior and subordinate who were directly involved. Peer observers tended to agree with the particular superior and subordinate about the quality of that relationship.
This is not an either-or question. Both analyses are needed. Dyads can be examined within groups. They may be viewed as independent of the group of which they are a part. Most supervisors have learned to praise subordinates in public and to reprove them in private. We should expect that the extent to which a leader praises subordinates will be seen in the same way by the subordinates (whichever description is used), since praise is likely to be a public affair. In this case, the average member of the group of employees will provide a description of the leader that is equal in accuracy to the description provided member by member. But to describe the leader’s contingent punishment, dyadic analysis should prove more fruitful, since such contingent punishment is more likely to be a private one-on-one exchange. In some circumstances, the effects on the group are likely to be more important; in others, individual exchanges will be more important. Much more multiplex examinations are required to take into account both possibilities as well as their statistical interaction (Dansereau, Alutto, Markham, & Dumas, 1982).
In an effort to examine both group and individual effects, Dansereau, Alutto, and Yammarino (1984) created a theory and methodology for analyzing variance in leadership along with covariance in leadership outcomes of leader-member relations and their effects for the dyad, the group, and higher organizational levels of analysis. Dansereau, Alutto, and Yammarino’s DETECT computer program facilitated an examination of what had heretofore not been fully analyzed—variances and covariances within the same group and in a set of groups as opposed to variance and covariances among these groups. Dansereau, Alutto, and Yammarino’s strategy and analysis program determined whether the relationships were a function of: (1) different styles of leaders described by their average differences in perceptions by their subordinates; (2) leadership processes that occur only within groups; or (3) individualized differences in perceptions of the same leader’s behavior.
Along with demonstrating that inner and outer circles of subordinates are likely to be working for the same superior, researchers have built a case for focusing on LMX, in contrast to the average exchange relationship between the leader and the group (Graen, 1978; Graen & Cash-man, 1975).
Graen, Cashman, Ginsburgh, and Schiemann (1977) showed that managers who developed higher-quality dyadic linkages with their bosses produced greater resources for their subordinates than did those managers who developed lower-quality linkages. Liden and Graen (1980) obtained the same results with first-level supervisors. Some supervisors appeared to collaborate with their bosses on unstructured tasks and to receive appropriate resources in return; others who were led by the same bosses did not collaborate or obtain desired resources.
Katerberg and Hom (1981) correlated the satisfaction and role perceptions of 672 National Guardsmen and their leadership descriptions of their first sergeants and unit commanders. Between-units leadership effects were based on the usual correlations between scores averaged for each unit. Within-unit effects were obtained by means of a hierarchical regression analysis in which unit means were entered as the first step and the remainder were then regarded as being due to within-group effects. Considerable increases in explanatory covariance emerged when the within-group component was added to the leadership-satisfaction outcomes, despite the fact that the instrument used to assess leadership, the Leadership Behavior Description Questionnaire, was couched in terms of group, not individual members. This preliminary effort was followed shortly by the work, mentioned on page 427, of Dansereau, Alutto, and Yammarino (1984).
Members’ ratings of their leader or of outcomes within a group may be similar but differ from the ratings of the average members of other groups about their own leaders and outcomes. These whole-group differences lie between and not within groups. There are intergroup differences in responses, coupled with the homogeneity of responses within the groups. This condition fits with the average-member-exchange approach to studying leadership based on the average member of each group. For instance, the reinforcing style used by a leader with all the group members correlated with the leader’s sense of security and personality (Hinton & Barrow, 1976).
Another possibility is that in a group-parts condition, individuals’ responses differ from one another within a group, but their average does not differ from the average responses provided by the members of other groups. Differences lie within groups, not between groups. There are intragroup differences, and there is heterogeneity among the members within the groups. The LMX approach is necessary to study the leadership and outcomes.
A third alternative is that the reports of the individuals may differ reliably and consistently both within or between the groups. This condition is equivocal, for the ratings are codeterminable from both sources. Individual implicit theories may dominate an equivocal condition.
A fourth possibility is that members may respond not as a group but only independently; thus their identity as a member of a particular group or the set of groups under analysis may be irrelevant. Here the results of interest are inexplicable, accounted for neither by systematic differences among individual member differences nor by the groups. There are no reliable differences among individuals within or between groups, and thus a null case is demonstrated.
Dyadic and Group Possibilities. Yammarino (1995) pointed out four types of supervisor-subordinate dyads: (1) Dyads within groups can be examined to see if they are different in a group, if the same differences in supervisor-subordinate relations appear in other groups but aggregated, and if the supervisors show up the same across the groups. (2) Independent dyads are not influenced by which group they are in. The supervisor-subordinate relationships differ from each other without reference to their group association. (3) Independence within dyads occur if the supervisors and subordinates differ from each other regardless of their group membership. (4) Cross-level dyads occur when each leader and subordinate in a dyad are in agreement but dyads within groups differ from each other. Independent dyads become dyads within groups.
Some WABA Results. Variance theory is tested empirically by within-and-between analysis (WABA) of variance and covariance. Yammarino and Naughton (1987) completed within-and-between analyses of the variances in LBDQ initiation and consideration scores (WABA I) of 70 members of a law enforcement organization on a university campus. Each member of the different hierarchically arranged units described his or her own immediate superiors. The investigators also completed corresponding analyses of within-and-between analyses of the covariances in the leadership-outcomes descriptions (WABA II). Outcomes included satisfaction with supervision, satisfaction with rewards, supervisory control, adequacy of communications, efforts of subordinates, and stress on subordinates. The results for the WABA I analyses of variance indicated the existence of an equivocal or codeterminable condition. There was significant variance both within and between groups; also, consistent individual differences in perceptions were at work here, rather than just differences between leaders. On the other hand, the WABA II analyses of covariance generally supported the conclusion for 10 of the 28 relationships that leader-outcome differences between whole groups dominated. This supported the acceptance of the average leadership style or whole-group approach. For example, members’ satisfaction with rewards and the leader’s initiation of structure correlated .67 among groups led by the different leaders and only .09 for members within groups led by the same leader. It did not matter if the descriptions were correlated in group terms (“The leader rewards the group if …”) or in individual terms (“The leader rewards me if …”). In the case of six variables, an equivocal condition was seen. For example, satisfaction with supervision emerged as codeterminable; there were covarying differences between leaders and outcomes as well as among members’ descriptions of their leader within the groups. The correlations between initiation and consideration were particularly likely to emerge as equivocal in interpretation. Finally, in the case of 12 variables, such as leadership and communications, the results were inexplicable. There was no statistically significant between-or-within covariance.
Table 16.1 Correlations between Leader-Members and Leader-Group Satisfaction and Rewards
| Initiation | Consideration | |||
| Member's Satisfaction with Rewards | Leader-Group Relation Described | Leader-Member Relation Described | Leader-Group Relation Described | Leader-Member Relation Described |
| Between Leaders | .67 | .57 | .42 | .30 |
| Within the Same Leaders | .09 | .03 | .27 | .26 |
SOURCE: Yammarino and Naughton (1987).
Table 16.1 illustrates the importance of both a dyadic and group account of leader-outcome relationships based on members’ descriptions. We need to ask the same questions about such relationships in two ways, dyadic and group. We need to examine both within-and-between covariances. As is seen in Table 16.1, satisfaction is correlated modestly with rewards and with considerate leadership (.26 to .42), regardless of the analytic component (between or within leaders) or type of description (leader-group or leader-member). Differences between leaders in initiation correlate with the average members’ satisfaction with rewards. The same results apply for all the members, regardless of their group. However, the leaders’ perceived use of initiation correlates with satisfaction only between the leaders’ groups, not within the leaders’ groups.
In the replication of the Graen, Liden, and Hoel (1982) study, it was concluded that the group effect was also of consequence when 81 registered nurses who quit or remained described their supervisors, although the dyadic effect on turnover was still the stronger (Ferris, 1985). However, Yammarino, Spangler, and Dubinsky (1998) analyzed individual, dyadic, and group results for 111 subordinates and their 34 superiors in a sales organization, who formed 111 dyads in 34 groups. The Multi-factor Leadership Questionnaire (MLQ) was used by the respondents to measure transformational leadership and contingent reward along with satisfaction with job and leadership. WABA I and WABA II results were of consequence for the MLQ and satisfaction, but only at the individual level, not the dyadic or group levels. In a sample of 285 employees of a county library system, work-unit size, cohesiveness, organizational climate, and leader power all predicted the quality of leader-member exchanges (LMX), but WABA also found individual differences—that is, effects at the level of the individual employee (Cogliser & Schriesheim, 2000). A sample of 83 sales personnel and their supervisors in a chain of retail stores rated their relationships. According to WABA, in comparison with effects at the group and individual levels, differences between the 83 dyads were most significant. But for a second sample of 116 dyads composed of insurance sales personnel and their supervisors in 37 organizations, the most significant differences were at the level of individuals (Yammarino & Dubinsky, 1992).
Although the group, dyadic, or individual approach may be appropriate in some instances, only analyses that take all the possibilities into account are likely to explain more fully what is happening. Schriesheim, Castro, and Yammarino (2000) examined the quality of LMX ratings of 75 managers, each paired with a high-performing and a low-performing subordinate. Generally, individual differences appeared more important than the group or dyadic effects implied by the span of supervision, efforts at upward influence, and commitment of the managers or subordinates.
The preceding research examined the processes through which leaders affect their followers or the processes through which followers affect their leaders. In fact, of course, both processes are occurring simultaneously and may be discussed as two-way dynamics. For example, Graen (1976) examined the dyadic relation between the leader and each member of the group. He noted that the dyads formed a behavioral interdependence between the respective roles. This interdependence included the dyadic “partnership” engaged in reciprocal influence, its extracontractual behavioral exchanges, its role-defined relations, and its loosely coupled fates (Dansereau, Graen, & Haga, 1975). Graen and Schiemann (1978) observed that the manager’s and subordinate’s agreement about the relationship was higher if the quality of their behavioral interdependencies was high. Such interdependence of the manager and the subordinate is necessary for the leader to give and the subordinate to receive opportunities for growth on the job (Graen & Scandura, 1986). Subordinates with high need for growth are more likely to collaborate in a closer and better-quality relationship with superiors on tasks that provide opportunities for growth.
But behavioral interdependence has its dark side. A superior and subordinate can share a delusion—a folie à deux—when both have a strong need to be dependent and are in an isolated position, say, on a space exploration mission. The delusion is likely to occur in more everyday settings if mutual trust and closeness are lacking. The subordinate may have the need to depend on the superior, but this need is frustrated if he or she cannot depend on the superior. In turn, the superior may want to depend on the subordinate but feels it unwise to do so. The world becomes dangerous, and the danger is compounded by the superior’s and subordinate’s lack of trust in each other. Each can trust only a few people, but not each other. The superior and subordinate can elaborate their mistrust of each other into a common fantasy about an organizational betrayal of both of them. Thus they become trapped by their shared delusion (Kets de Vries, 1984).
Chapters 17 through 20 will present consistent styles of leadership that describe how leaders relate in the same way to all their followers. But according to Dansereau (1995), when previously the variations in a leader’s behavior toward different individuals in a group were treated as statistical error, WABA theory and WABA allowed for a leader to regard each subordinate as an independent, unique individual. To be successful, the leadership efforts must vary within and between work groups (Dansereau, Yammarino, Markham, et al., 1998). Perceptions of the same leader differ from one subordinate to another. Nevertheless, if the leader supports the individual subordinate in exchange for satisfactory performance, the leader will reinforce the subordinate’s sense of self-worth The knowledge, skills, and abilities of both the supervisor and the subordinate set the stage for individualized leadership. Individualized leadership is expected to correlate positively with LMX (Yammarino & Dansereau, 2002). This result was found by Paul, Scheyns, and Rigotti (2001).
Psychological contracts are the expectations and obligations of the leaders and the led toward each other, their unit, and the organization (Argyris, 1960; Schein, 1965). The concept was accepted but little empirical work on it was completed until the 1990s.
Building on a 13-nation survey of psychological contracts by Rousseau (1995) and Rousseau and Schalk (2000), Van den Brande, Janssens, et al. (2002) analyzed survey data from structured interviews with 1,106 Belgian employees from public, private, for-profit, and nonprofit agencies. Formal written legal contracts and statutes also covered their employment. The employees’ expected entitlements generated five factors: long-term involvement; tangibility, clarity, and transparency of obligations and rights; treatment as a person, not an economic resource; carefulness about arrangements; and equal treatment of all employees. The employees’ expected obligations factored into: loyalty; open attitude; high personal investment; flexibility and tolerance for internal organizational changes; and respect for authority. On the basis of their high and low entitlement and obligation factor scores, employees formed the following six clusters of types of contracts:
1. Contractual loyalty—high long-term involvement, equal treatment, loyalty, and low personal investment. The employee perceives loyalty in his or her contract. He or she is a poorly educated, low-paid blue-collar employee or civil servant in a large organization with a personal investment and flexibility but lower expectations about the operational job.
2. Contractual instrumentality—high expectations but low obligations. The employee is a poorly educated, low-paid white-collar or blue-collar employee in a medium-size organization with reasonable job opportunities elsewhere.
3. Contractual investing—higher employee. The psychological contract is as perceived by a highly paid executive in senior management in a small company.
4. Contract strength—high on all dimensions. The psychological contract is that of a young, highly educated, highly paid white-collar employee or executive, holding a professional or senior management position in a small or medium-size organization and readily marketable.
5. Contract weakness—low scores on all dimensions of expected entitlements and obligations. Reasonably employable elsewhere with an average education and jobs at all hierarchical levels. The employee perceives a weak contract and is average in pay for his or her level in the hierarchy.
6. Contractual lack of attachment—low scores on all dimensions, particularly involvement and loyalty. These psychological contracts were typical of young, highly educated, highly paid executives or white-collar employees in professional or senior management jobs in small or medium-size organizations.
Rousseau (1995) suggested that psychological contracts may appear as: (1) Transactional: Short-term, economically focused, specific and clear about expected performance and rewards, as in seasonal work; (2) Transitional: Short-term, uncertain, and ambiguous with unspecified mutual obligations, rewards, and performance expectations, as in jobs in radical reorganizations, mergers, and takeovers; (3) Balanced: Long-term, economically focused, with clear and specific terms, as among structured work team members. Psychological contracts may be also said to be in balance when the obligations of employees and employers are at relatively the same level (Shore & Barksdale, 1998); or (4) Relational: Long-term, ambiguous, and lacking tangibility, as in a family business.
As might be expected, personality affects the type of psychological contracts formed. For example, individuals with high self-esteem would be expected to form balanced psychological contracts; individuals with low self-esteem, transitional psychological contracts (Raja, Ntalianis, & Johns, 2002). Among 210 graduate business students with two years of work experience, varying in attachment style, those who were avoidant were less likely than those secure or anxious in attachment to ask others for contractual information. The secure seek more information about various aspects of their psychological contract than do others and are most likely to believe their employers will fulfill their obligations. The insecure, avoidant, and anxious are most pessimistic about the probability that employers will fulfill obligations. If violations occur, those with a secure attachment style are most likely to speak out against contractual violations by higher authority. Again, the insecure are less likely to do so (Bendapudi, Bendapudi, & Ballam, 2002).
Crouch (1987) and Zahn and Wolfe (1981) started with different assumptions and used different methods to construct different models to describe how mutual influence develops between superiors and subordinates. Crouch (1987) assumed that manager-subordinate relationships stem from the influence that the subordinates can exert in solving unstructured problems that they and the manager face. Over time, the model predicts the emergence of stable relationships. Either high-or low-performing groups will be developed to deal with outside demands on the groups. Zahn and Wolf (1981) provided Markov mathematical states-events models to deduce the continuing interplay of the superior and the subordinate, showing how their emerging task and interactional relationships can predict future outcomes. Unexpected results appeared. Contrary to Crouch, the model predicted long-term behavior to be highly variable and versatile, rather than stable. The mutual biases of the superior and the subordinate were most salient in effect; both were likely to contribute to the cyclical maintenance of their relationships.
Considerable empirical evidence is available to provide a cross-sectional picture of the two-way impact—the combined effects of the leader’s and the follower’s competence and compatibility, the interplay of the leader’s style and the follower’s personality and motivation, and the combined effects of the leader’s and follower’s concerns for the task and for their relationship with each other. Pratt and Liambalvo (1982) showed that an understanding of the leadership of accounting auditors in working with their auditing assistants required consideration of aspects of the auditor, the assistants, and their interplay. Outcomes depended on the personal dominance of the auditors, the intolerance of ambiguity by the assistants, and the match between the perceived complexity of the task for the auditor and the assistants and the assistants’ job experience.
Rao (1982) conducted an experiment in which the leader could reward the subordinate. In one setting, the leader was to receive a bonus based on the subordinate’s performance. The results were the leader’s increased rewarding of the subordinate, large improvements in the subordinate’s performance, and greater stability and cooperativeness between the leader and the subordinate. These results were different from those in the situation in which the leader was not offered any bonus for the subordinate’s good performance.
The two-way nature of leadership and followership was investigated by Herold (1974), who carried out a double-substitution laboratory experiment in which both leaders’ and subordinates’ actual behaviors were intercepted and substituted with fully programmed supervisory and subordinate behaviors. Thirty-two groups of three persons each, consisting of one leader and two subordinates, were balanced in treatment so that the subordinates were powerful in half the groups and the leaders were powerful in the other half. The experimental task consisted of proofreading manuscripts and finding errors. The leader of each group received a “good product” from one of the subordinates and a “bad product” from the other. Within each group, one subordinate received a punitive communication following the performance, while the other received a supportive communication. Herold found strong mutual effects on the attitudes of the leaders and followers toward each other and toward the situation. Whether the leader received a good or bad product from a subordinate strongly affected what the leader did, but whether the subordinate received a punitive or a supportive communication from the leader had somewhat less of an impact on the subordinate. The leader’s power was important in its effects on subordinates, but the subordinates’ power was not significant in affecting the leader’s behavior.
This lack of symmetry was also seen by G. J. Palmer (1962a, 1962b), who studied leaders and followers in a difficult task situation. Palmer concluded that follower-ship in such a situation is explained largely by the lack of ability. Leadership, on the other hand, requires a more complex explanation involving both individual and situational variables. For instance, the joint competence of the leader and the subordinate has to be considered. This conclusion was confirmed by Rohde (1958), who observed different combinations of qualified and unqualified leaders and followers under four conditions of reward and punishment. The leader’s ability to perform the task was found to be more highly related to the group’s performance than to the group’s motivation. Not unexpectedly, the poorest performance was exhibited by groups with both unqualified leaders and unqualified followers. Almost equally poor were groups with qualified leaders and two unqualified followers.
Blades (1976) studied the joint effects of leaders’ and subordinates’ competence on their group’s performance in servicing U.S. Army mess halls. Ratings of the group’s performance by inspectors were correlated with ratings by 102 mess hall personnel of the intelligence, ability, and behavior of the leader. Blades found that the subordinates’ intelligence, ability to perform the task, and motivation correlated positively with their performance only under participative management and when the subordinates were highly motivated. The leader’s intelligence correlated positively with the group’s performance only under directive management with highly motivated leaders and subordinates. The leader’s ability to perform the task and the leader’s motivation correlated positively with the group’s performance only with directive supervision of highly motivated subordinates. Blades concluded that competent, motivated subordinates can be led best if they are allowed to participate in the decision-making process, and that competent, motivated leaders do best with a directive style if their subordinates’ competence is limited.
The leader and the follower depend on each other because each stimulates and reinforces the other’s behavior. The leader initiates, questions, or proposes; the follower complies, resists, or ignores. Leadership and follower-ship are mutual activities of influence and counterinfluence. Leaders and followers both give and both receive benefits. The relationship is maintained by this social exchange and this mutual influence (Hollander, 1978). As Hollander and Julian (1969) noted, when leaders are effective, they give something and get something in return. This transactional approach to leadership involves trading of benefits. The leader provides a benefit by directing the group toward desirable results. In return, the followers provide the leader with status and the privileges of authority, influence, and prestige. However, the leader may demand from the followers what they regard as an excessive expenditure of energy. The followers’ compliance may be tempered if desired outcomes do not match the perceived effort required. As members become less involved in the group’s success, they complain less about obstacles to such success but more about demands for expending their time and energy (Willerman, 1954). Bass, Gaier, and Flint (1956) observed that ROTC cadets who were strongly motivated to enter advanced training complained about the difficulty of the test used to screen applicants, whereas those whose motivation was low complained about having to take the test rather than about its difficulty.
Intermember Expectations. Much of the differentiation of members’ roles and the emergence of leadership in an informal group comes about as a result of the mutual reinforcement of intermember expectations. Because of their initiative, interaction, and contributions to the group’s task, some members reinforce the expectation that they will be more likely than other members to establish conditions that promote progress in the task, members’ freedom and acceptance, and the group’s cohesiveness. Other members, by compliance, reinforce the expectation that whoever has started and succeeded with it should continue in the leadership role; that is, “Don’t change horses in midstream.” Similarly, members build up expectations regarding the contributions that they are to make. The reactions of other members confirm the expectation that they are (or are not) to continue in the same role. The role system and the status structure of a group are determined by a set of such mutually reinforced intermember expectations. Thus in a verbal learning experiment, Bachrach, Candland, and Gibson (1961) observed that the members of the group differentially reinforced the behavior of other members. Such differential reinforcement accounts, in large part, for individual differences in role specialization and conformity to norms.
Seeking Feedback. Some people seek feedback; others shy away from it; still others ignore it when it is given to them. Subordinates have been found more likely to seek feedback if their leaders were perceived to be supportive, esteemed, expert, and transformational (Steelman, Williams, & Levy, 1996; Williams, Miller, Steelman, et al., 1999). Also, they were more likely to seek feedback from supervisors who were thought to be accessible, who clarified roles, and who had constructive intentions (Vancouver & Morrison, 1995; Levy, Miller, & Cobar, 2000). In addition to confirming many of these effects, Thibodeaux and Kudisch (2000) showed that feedback was more likely to be sought if the quality of LMX was high. But Dunford and Williams (1999) found that nurse supervisors were less likely to ask for feedback from the nurses in 200 nurse-supervisor pairs when there was stronger agreement between the nurses’ and the supervisors’ ratings of the nurses.
In the United States, information exchange is more direct and explicit. In East Asian cultures, it is more indirect and implicit, with more nonverbal nuances and symbolism (Hall, 1976). More cost in seeking feedback was seen by 167 Taiwanese university students than by 76 American counterparts; the Taiwanese noted greater effort and loss of face (Kung & Steelman, 2003). The norm is unlikely to encourage the Asian student to seek feedback (Ashford & Northcraft, 1992). Ashford and Tsui (1991) conducted a study of the feedback-seeking behavior of 387 managers from their superiors, peers, and subordinates. Some of the managers regulated their own performance by seeking negative feedback (“Tell me what I did wrong”) and tended to increase the accuracy of their understanding of how their work was evaluated by the three sources. In turn, the three sources formed better opinions of the effectiveness of these negative feedback–seeking managers. But the sources tended to downgrade managers who sought positive feedback.
Differing Leader Treatment of Subordinates. Leaders do not deal with each of their subordinates in the same way. Leaders give more directions to followers who are performing poorly and monitor them more closely. The leaders are more considerate, cooperative, and tolerant when dealing with able subordinates (Podsakoff & Schriesheim, 1985). Leader-follower interaction and mutual support are seen among political leaders and their followers. The leaders activate and mobilize support for their objectives, control the communications media, and make use of intermediary opinion leaders. The followers may be actively interested and ready to be motivated to engage in political activity. Multilayered networks of leaders and followers are formed. Those below yield an aggregate of public opinion to support the aims of those at the top. The importance of such followers’ support for leaders was seen by Pepinsky, Hemphill, and Shevitz (1958), who studied experimental groups working on construction problems. When the leaders were made to believe that they were accepted by the members, the groups were more productive and exhibited high degrees of satisfaction and participation. When the leaders were made to feel that they were rejected, the groups made fewer and poorer decisions and exhibited less participation and satisfaction.
Key Interactions. Komaki and Citera (1990) conducted an experiment with 60 “manager-subordinate” pairs consisting of 160 female undergraduate students. Managers either were only monitors of subordinates’ performance or only directed the subordinates. Afterward, in comparison with the “directing manager” condition, managers who monitored spent much more time talking about the consequences rather than their own performance. Their subordinates spent more time talking about their own performance. The same interaction was stimulated in the next trial.
Effects of Language. More than direct cost-benefit calculations may lie behind an interaction between the leader and the followers. Drake and Moberg (1986) suggested that two adjustments in language can be made to change the cost-benefit calculations by a follower who is considering an attempt by a leader to influence him or her into compliance. First, sedating language, such as semantic indirectness, can be used in the leader’s attempt to keep the follower from considering the costs of compliance. For example, instead of giving a direct order, the leader can hint at the need for action or prompt the follower by saying, “It is getting warm in here” instead of, “Please turn down the heat.” Other sedative language includes teases and questioning: “I wonder what that’s all about?”
Second, the linguistic form may palliate the follower to protect him or her from loss of respect that might occur if a more straightforward order is used to correct a follower or to reverse a decision. Numerous positive polite strategies may be used as palliatives, such as expressing admiration, claiming common viewpoints, displaying concern for the follower, and desiring cooperation. With palliatives, the attempt to influence may be accompanied by efforts to reduce the difference in power between the leader and the follower. Instead of the direct order, “Check the switch again,” the leader suggests “You may want to try to check the switch again.” Other palliation includes disclaimers, hedging, appeals to suspend judgment, excuses, and justifications.
Under some circumstances, clarification of what needs to be done may depend on the subordinates’ participation with the leader in agreeing on the tasks and setting of the group’s goals. Critical experiments described by Locke, Latham, and Erez (1987) suggested that such participation by subordinates in setting goals is important to their understanding, commitment, and productivity if the leader is curt and unsupportive. The subordinates’ participation in goal setting is not necessary if the leader assigns goals in a friendly and supportive manner. This may explain why Dossett, Cella, Greenberg, and Adrian (1983) found in a laboratory study with 40 undergraduates that although, as was expected, productivity on a clerical task was highest when a supervisor set high, clear, standards for them and gave them friendly support, the subordinates’ participation in setting the goals did not contribute to their productivity by itself.
The failure of participation in setting goals to make a difference also may be explained by the character of the participation. Neider (1980) demonstrated, in a study of performance in retail stores, that positive outcomes were obtained from such participation only when the participation process clarified the linkage of effort and performance, and only when the rewards given for high performance were valued. The effects of clarification also depend on the competence of the leader. Podsakoff, Todor, and Schuler (1983) found that expert leaders who work at setting and clarifying goals did decrease their subordinates’ sense of role ambiguity. But the reverse was true for leaders who were inept; their efforts resulted in increasing their subordinates’ sense of role ambiguity.
In the previously mentioned longitudinal survey by Greene (1979b) of the mutual effects of leaders and subordinates, 60 subordinates from five manufacturing organizations, were queried. These subordinates had joined the management of their firms within the past three months and had never before worked for their immediate superiors. Greene assumed that because these subordinates were newly appointed, their expectations were unclear and they were not yet fully compliant. The subordinates described their leaders’ behavior and indicated their own expectations, compliance, performance, and satisfaction. The data were gathered three times at three-month intervals to permit path analyses. Consistent with path-goal theory (see Chapter 27), early role clarification by a leader directly resulted in greater satisfaction of and compliance by the subordinates three and six months later. Role clarification by the leader also was seen to enhance the subordinates’ expectations that their greater effort would produce better performance.
Compatibility. It seems obvious that if leaders and their subordinates, individually or in groups, share the same approaches, values, and attitudes, they will be more satisfied with their relationship and experience less conflict and more mutual support. Such positive associations between the compatibility of the values of superiors and subordinates and satisfaction with their relationships were obtained by Duchon, Green, and Tabor (1986) in a field study of 49 Junior Achievement companies. Avolio and Howell (1993) showed in regression analysis of 76 senior executives and their 237 immediate direct reports that a match in personality (locus of control, innovation, and risk taking) contributed to the performance of their units.
Fuji (1977) conducted an experiment in which the independent variables were manipulated by simulating a division of a greeting card company. Eighty paid male volunteers participated as work-group leaders or workers in this unit. The followers’ performance, based on merit ratings by the leaders but not on objective performance, was positively related to greater interpersonal compatibility between the leaders and the followers. Extrinsic satisfaction increased with leader-member compatibility. The relationship between leader-member compatibility and intrinsic and overall satisfaction was moderated by the amount of cooperation required by the task and the amount of experience with the task. Compatibility was positively related to relations-oriented leadership behavior, but less so to task-oriented leadership behavior.
McLachlan (1974) reported a study of group therapy with 94 alcoholic inpatients in which patients and therapists were matched for conceptual compatibility on the basis of Hunt’s Paragraph Completion Test. Matched pairs of patients and therapists were positively associated with outcomes as evaluated by staff ratings 12 to 16 months later. Beutler, Jobe, and Elkins (1974) also investigated the effects of patient-therapist matching on attitudes. Matched attitudes were associated with self-rated improvement (although some dissimilarity of attitudes resulted in more attitudinal changes in patients). Similarly, Steiner and Dobbins (undated) conducted a laboratory study of 111 management students and their ostensible subordinates in which high and low intrinsic and extrinsic work-related values were the manipulated bases of matched and mismatched “superiors” and “subordinates.” When superiors and subordinates matched each other in high extrinsic or high intrinsic work values, the superiors were more likely to attribute the subordinates’ past good performance to internal rather than external causes.
Pulakos and Wexley (1983) examined 171 manager-subordinate dyads composed of the same and opposite sexes. Sex as such was not as important as perceived compatibility. The managers’ leadership, particularly their facilitation and support of the work, was appraised higher by their subordinates when both the managers and the subordinates felt they were “similar kinds of people.” Subordinates also were rated more favorably by their managers if they were perceived as similar. Conversely, lower performance appraisals occurred when the managers and subordinates were mutually perceived as dissimilar. When the subordinates saw themselves as similar to their managers, the managers gave higher dependability ratings to their female than to their male subordinates. When subordinates said they were dissimilar to their managers, managers rated both males and females uniformly lower.
Actual Similarity. The actual similarity between managers and their subordinates tends to correlate with the managers’ appraisals of their subordinates’ performance (Miles, 1964a; Senger, 1971); with the subordinates’ evaluations of their managers; and with the subordinates’ satisfaction with their managers (Weiss, 1977). Subordinates who perceive their superiors’ attitudes toward work more accurately are rated more highly by their superiors and are more satisfied with their superiors (Greene, 1972; Howard, 1968; Labovitz, 1972).
Subordinates are more satisfied in general if their descriptions of their manager’s attitudes more accurately match the manager’s self-description. In turn, subordinates receive higher evaluations from their manager if the manager’s description of the subordinates’ attitudes matches the subordinates’ self-descriptions (Wexley, Alexander, Greenawalt, & Couch, 1980). As noted above, Avolio and Howell (1993) found that unit performance and satisfaction were higher when Canadian senior executives were matched rather than mismatched in locus of control, innovation, and risk taking.
Using Exercise Supervise (Bass, 1975c), approximately 3,500 managers in training workshops in 12 countries enacted one of three supervisory roles—authoritarian, persuasive, or participative—or one of three subordinate roles: highly involved, moderately involved, or uninvolved. Meetings took place between dyads composed of one supervisor and one subordinate. The three tasks were to decide which of 25 traits are most and least characteristic of lower, middle, and top managers. A Latin-Square design was completed so that data about all nine supervisory-subordinate combinations of role play emerged. As might be expected, the uninvolved, apathetic subordinate was least preferred by supervisors as a whole, and the participative supervisor was most preferred by subordinates as a whole. The fastest and most easily completed interactions took place between the authoritarian supervisors and the uninvolved, apathetic subordinates. Dissatisfaction of subordinates and supervisors tended to be greatest for highly involved subordinates meeting with authoritarian supervisors; satisfaction tended to be greatest for highly involved subordinates meeting with participative supervisors. The interaction effects of persuasive supervision and subordinates at the three different levels of involvement fell in between in satisfaction and speed of decision (Bass, Burger, et al., 1979; Thiagarajan & Deep, 1970).
Sales, Levanoni, and Saleh (1984) contrasted the reactions of 226 intrinsically and extrinsically motivated clerical, technical, and professional employees to two types of supervision. General supervision provided broad specifications of goals and gave subordinates the means of achieving them. Close supervision involved giving detailed instructions at each stage in the process of completing a task and checking to see that instructions were carried out. The investigators found that intrinsically oriented employees expressed more satisfaction with their supervisors under general supervision than under close supervision; extrinsically oriented employees indicated the opposite. Steiner and Dobbins (undated) experimented with the extrinsic and intrinsic work values of the subordinates of management students who were acting as their supervisors. Subordinates with high intrinsic work values were given more autonomy and more challenging assignments by their “supervisors” than were the other subordinates. W. W. Burke (1965) studied student leaders in 24 groups participating in an interfraternity contest. Each group performed a clerical and organizational task, as well as a decision-making task. Regardless of the task, followers with high need-achievement scores, working with relations-oriented leaders, reported more tension than followers with high need-achievement scores working with task-oriented leaders. Those with low need-achievement scores, working with task-oriented leaders, reported more tension than those working with relations-oriented leaders. The more relaxed situation occurred when the leader’s personality met the needs of the group of followers. Misumi and Seki (1971) showed that Japanese subjects with low need for achievement demonstrated relatively high performance but low satisfaction with a task-oriented leader in experimental tasks. Subjects with high need for achievement were most productive and most satisfied under leaders with high task and relations orientation and performed worst under leaders whose task and relations orientations were low. Misumi (1985) also noted that high-anxiety subjects did better under a relations-oriented leader than under a task-oriented leader.
Superiors were viewed as less active and directive by hospital employees who were more concerned with power, achievement, and independence (Niebuhr, Bedeian, & Armenakis, 1980). Engineering personnel with high self-esteem who were in dyads with superiors of long duration saw more initiation and consideration in their superiors’ leadership behavior (Niebuhr & Davis, 1984). Greer (1960–1961) reported that authoritarian infantrymen and airmen worked better under authoritarian leaders whereas egalitarian men did better when led by egalitarians. A group’s performance was positively related to followers’ perceptions of the leader as a problem solver and the extent to which the leader met the followers’ expectations. Weed, Mitchell, and Moffitt (1976) concluded, from a laboratory study, that supervision that was structuring generated higher performance by subordinates if the subordinates were higher in dogmatism. Considerate supervision generated higher performance in subordinates who were lower in dogmatism.
Concerns for Task or Relationships. Back (1948) analyzed the interactions of leaders and followers in two discussion groups—work-centered and “emotionally toned.” When the leaders emphasized work performance more than did the followers in the first period, the followers increased their work responses in the next period. When the leaders emphasized friendliness more than did the followers in the first period, the followers exhibited more friendly responses in the next period. The emotionally toned group devoted more time in the first period to establishing stable intermember relations and a group structure. The work-oriented group spent more time in strengthening goal-directed activities. In the next period, the emotionally toned group spent more time maintaining participation among members and with leaders. In the work-oriented group, the leader tended to lose importance, whereas interactions between members were strengthened. Bass and Dunteman (1963) studied sensitivity training groups that were composed of participants who were homogeneous in relations (interaction) orientation, task orientation, and self-orientation. Relations-oriented groups tended to be most satisfied with their highly relations-oriented leaders. The leaders who emerged in task-oriented groups tended to be even more highly task-oriented than the average member. However, the emergent leaders of the self-oriented groups were relatively low in self-orientation. Stimpson and Bass (1964) found that compared with task-oriented and self-oriented participants, relations-oriented participants in problem-solving groups made it more difficult for their work partners to attempt and succeed as leaders, to reach agreement, and to feel responsible.
Bass (1967c) observed task-oriented and relations-oriented oriented followers under coercive and persuasive styles of leadership. He found that task-oriented followers produced greater quantities of work under persuasive leadership and that relations-oriented followers produced more work under coercive leadership. The followers’ satisfaction with the task was significantly higher under a directive relations-oriented leader than under a participative relations-oriented leader. Conversely, their satisfaction was lower under a directive than under a persuasive task-oriented leader.
In their interaction with followers, leaders show what needs to be done, provide feedback to subordinates on how well it is done, communicate needed information, and act as models for their subordinates. They mediate between and connect their followers to their superiors. They are a source of feedback and serve as models for their subordinates. They need to have influence with those above them as well as those below them in the hierarchical organization.
Leadership and followership are reciprocal. Followers have an impact on their leaders. To be a good leader, one needs to be a good follower. Isolates, not followers, are the opposite of leaders. Followers are an important source of feedback to leaders. Followers’ readiness, maturity, interests, competence, and compliance affect their leaders’ performance. Followers’ effects on leaders also depend on the leaders’ and followers’ explanations for the followers’ dispositions and performance. Among the interactions of importance to performance are the leader’s style and the followers’ motivation and personality. Another important effect is due to the interaction of the task and relations orientations of the leaders and their followers. Groups operate more successfully when the task, the leader’s personality, and the follower’s personality are compatible. For example, groups with task-oriented leaders perform better than do those with person-oriented leaders when followers are also task-oriented. Followers whose task orientation is low experience less tension under person-oriented leaders.
Leaders may treat each follower in a group the same or differently, depending on whether followers are members of the inner or outer circle. It is important to examine the quality of the dyadic leader-member exchange (LMX) as well as the leader-group exchange. Analysis of variance within groups may be as important as analysis of variance between groups.
Followers affect, to a considerable extent, what their leaders may do and can do. The exertion of this upward influence involves many different tactics ranging from ingratiation to establishing coalitions of influence. Important feedback can flow upward.
Still to be determined is whether the exchange relationship between the leader and subordinates becomes stable or remains variable in the long term.