Contingent reinforcement occurs when leaders arrange for followers to be praised, commended, or rewarded materially for successfully carrying out agreed-upon assignments—or to be reproved, reprimanded, or disciplined for failure to do so. Responsiveness by followers to a leader is heightened by their expectation that satisfactory performance will be rewarded and unsatisfactory performance will be punished materially, socially, or symbolically. According to Sims and Lorenzi (1992), effective leadership reinforces desired followers’ behavior and eliminates undesired follower behavior through providing or denying social, symbolic, and material rewards and punishments. According to Sims (1977), a leader is a manager of reinforcement contingencies. Connelly, Gaddis, and Helton-Fauth (2002) have suggested that leaders arouse positive emotions in followers by consistently applying contingent reinforcements and generating mutual liking. Negative emotions are aroused when leaders display a negative demeanor during punitive incidents.
The Iliad is filled with stories of contingent reinforcement. The heroes of the Iliad often engaged in contingent reward or punishment with men and the gods to gain compliance with their wishes. Achilles refused to fight further and went off to sulk over the arbitrary confiscation of Briseis, a girl captured from Achilles by the king, Agamemnon. Briseis was taken from Achilles to compensate Agamemnon for the loss of his own prize, Chryseis. To assuage the wrath of Apollo, public pressure was brought on Agamemnon to return Chryseis to her father, a priest of Apollo. Without Achilles, the Achaeans faced defeat. Agamemnon sent Odysseus and Ajax to lure Achilles back to take charge of the Achaean forces with promises by Agamemnon that Achilles would receive booty, land, and women if he achieved victory over the Trojans (Iliad, 1720/1943, Book IX).
For effective leadership, leaders need to let followers know what rewards are available and what behavior will be rewarded. They need to make clear what behavior will be rewarded—for example, quality rather than speed; lasting solutions not quick fixes; and creativity, not conformity. Reprimands should be directed against specific behaviors, not the person. They are given privately. They should be low-key, prompt, and unemotional. The leader should take responsibility for the reprimand. It should be accompanied by information about what behavior would be desirable (Howell & Costley, 2001, 2006).
Transformational leadership may transcend the satisfaction of self-interests; nevertheless, the dynamics of leadership-followership have most often been explained as a social exchange. The exchange is established and maintained if the benefits to both the leader and the followers outweigh the costs (Homans, 1958, 1961). This exchange is fair if “the leader gives things of value to followers such as a sense of direction, values, and recognition, and receives other things in return such as esteem and responsiveness” (Hollander, 1987, p. 16). Followers expect that the leader will enable them to achieve a favorable outcome, and they believe the exchange to be fair if the rewards are distributed equitably. However, they perceive the leader’s failure in this regard as unjust, particularly if the leader has not made an effort or is self-serving. “Fundamentally, there is a psychological contract between the leader and followers, which depends upon a variety of expectations and actions on both sides” (Hol-lander, 1987, p. 16).
The exchange or transaction cycle contains the following phases. The leader and the followers perceive each other as being potentially instrumental to the fulfillment of each other’s needs, say, for the completion of a task. If necessary, the leader clarifies what the followers must do (usually, they must successfully complete the task) to fulfill the transaction to obtain the material or psychic reward, or to satisfy their needs.1 The followers’ failure to comply may move the leader to take corrective action. The followers’ receipt of a reward or avoidance of punishment is contingent on their successful compliance with and completion of the task. Satisfaction of the leader’s need accrues from the followers’ success. Wilson, O’Hare, and Shipper (1989) expanded the description of the task cycle for the tasks performed by service and sales personnel (Table 15.1).
Table 15.1 The Task-Cycle Model
Blanchard and Johnson (1982) set out the following rules for contingent reinforcement in the “one-minute manager’s game plan”: First, obtain the subordinate’s agreement with the goal, including the appropriate behavior for achieving it. Check the subordinate’s behavior to see whether it matches the agreed-on goal. Second, if the goal is achieved, provide praise (contingent reward) as soon as possible that is specific to what the subordinate did right. Indicate how the subordinate’s action helps others and the organization. Third, if the subordinate’s performance fails to match expectations, deliver a reprimand (contingent punishment) as soon as possible after the failure. Despite the reprimand, it must be added, you should continue to think well of the subordinate, though not of this specific performance. Fourth, the subordinate’s success may call for setting a new goal; the subordinate’s failure may require a review and clarification of the old goal.
It takes at least two people for leadership to occur. Someone has to act, and someone else has to react. Whether the actions and reactions take place will depend on who the “someones” are and what their needs, competencies, and goals are. If one is perceived as instrumental to the other’s attainment of the goal, because of greater competence or power, the stage is set for an interaction and leadership to occur. Reinforcement leadership is often referred to as instrumental leadership.
In this transactional process, leaders are agents of reinforcement for the followers. At the same time, the followers’ compliance or noncompliance makes them agents for reinforcement of the leaders. “Leader and subordinate accept interconnected roles and responsibilities to reach designated goals. Directly or indirectly, leaders can provide rewards for progress toward such goals or for reaching them. Or, they can impose penalties for failure ranging from negative feedback to dismissal. … Contingent-positive reinforcement, reward if agreed-upon performance is achieved, reinforces the effort to maintain the desired … employee performance. Contingent-aversive reinforcement … signals the need to halt the decline in speed or accuracy of the employee’s performance, to modify or change the employee’s behavior. It signals the need for a reclarification of what needs to be done and how” (Bass, 1985a, pp. 121–122).
Adams, Instone, Prince, and Rice (1981) collected West Point cadets’ narratives about incidents of good or bad leadership that occurred during summer training. High on the list of good incidents was contingent-rewarding behavior. High on the list of bad incidents was punishing subordinates with or without provocation.
Rewarding by leaders, whether contingent on followers’ performance or noncontingent, was seen by 84 college students as exemplifying the good leader. Noncontingent punishment, but not contingent punishment, biased the students in the opposite direction (Korukonda & Hunt, 1989). Atwater, Dionne, Avolio, and Camobreco (1996) obtained 1,109 critical incidents along with leadership ratings for 286 cadet leaders at Virginia Military Institute (VMI), each rated by five subordinates. Of the four transactional reinforcers, only contingent reward was related to the effectiveness of leadership. Noncontingent reward was irrelevant. At VMI, noncontingent punishment such as reprimands, extra physical exercise, and assignment of extra duties was counterproductive (Arvey, Davis, & Nelson, 1984). Similar findings were obtained about the effects of noncontingent punishment on cohesiveness at the U.S. Naval Academy (Roush, 1991) and the U.S. Coast Guard Academy (Blake & Rotter, et al, 1993).
According to Peters and Waterman (1982, p. 123), the managers of the excellent companies they examined exerted a good deal of effort to provide positive reinforcement for the successful completion of tasks. They made a conscious effort to reinforce any action that was valuable to the organization. Subordinates’ satisfaction was increased the most, according to a field study by Reitz (1971), when their supervisors praised and rewarded them for acceptable performance (but reproved them for unacceptable work). For 231 subordinates surveyed by Klimoski and Hayes (1980), particularly strong associations were found between their satisfaction with their supervisors and the supervisors’ explicitness in giving instructions on what was required, support for efforts to perform effectively, and consistency toward the subordinates. The supervisors’ explicitness and consistency also had moderate effects on reducing role ambiguity and role conflict. The supervisors’ consistency contributed to some extent to the subordinates’ performance and knowledge of the job, although these appeared to be enhanced by the subordinates’ involvement in determining standards. The leaders’ behavior in the contingent-reward path-goal processes contributed to the subordinates’ efforts and performance by clarifying the subordinates’ expectations that a payoff would accrue to them as a consequence of their efforts. To a lesser extent, some of the contingent rewarding behaviors also contributed directly to the subordinates’ improved performance and satisfaction with supervision by reducing role ambiguities and role conflicts.
Effects on Satisfaction. Podsakoff and Schriesheim (1985) completed a comprehensive review of field studies of the effects on subordinates of supervisors’ contingent and noncontingent reinforcement.2 Among the general conclusions they reached was that subordinates were more satisfied with their situation if their leaders provided them with rewards (positive feedback) contingent on their performance.
Such satisfaction was not present if rewards were not contingent on their performance. “I feel well treated by my supervisor no matter what I do. In other words, it makes no difference how I perform; my supervisor will always arrange for me to be rewarded.” This noncontin-gent reward by the supervisor was not as satisfying as earning rewards for the subordinates’ good performance. Particularly dissatisfying to subordinates was noncontin-gent negative feedback—that is, not really being able to link reprimands with the behaviors that elicited them. This condition is likely to promote learned helplessness.
Field studies involved 17 large samples, with a median size of 275 cases, of financial managers, paramedics, engineers, white-collar employees of public utilities, registered nurses and nurses’ aides, administrators of nonprofit organizations, city and state government employees, hospital pharmacists, and mental health employees. For the 17 samples, a mean positive concurrent correlation close to .50 was found between the subordinates’ satisfaction with supervision and the extent to which the supervisors practiced contingent reward. The corresponding mean correlation was near zero between satisfaction with supervision and the practice of contingent punishment. Noncontingent reward had little relation to satisfaction with supervision. At the same time, noncontingent punishment was negatively correlated with satisfaction with supervision in these samples. The mean correlation here was –.32 for 11 of the samples in which the measures were available. Even after the same single source of variance was removed by removing the general factor of each individual’s responses to all questions (the subordinate raters described the leaders’ as well as their own satisfaction and performance), Podsakoff and Todor (1985) still found a positive association between contingent reinforcement leadership and a group’s cohesiveness, drive, and productivity. But Brown and Moshavi (2002) did not find that contingent rewarding of 440 faculty members by chairpersons in 70 academic departments enhanced the faculty members’ job satisfaction, expenditure of extra effort, or organizational effectiveness.
Effects on Performance. Considerable evidence from laboratory experiments and field studies has indicated that reinforcement leadership contributes to the effectiveness of subordinates and collegial groups. Spector and Suttell (1957) contrasted reinforcement leadership with authoritarian and democratic leadership of teams in producing correct plans in a controlled, laboratory setting. The reinforcing leader was encouraging and expressed approval of good solutions to problems every time the teams reached them. When incorrect planning occurred, the reinforcement leader suggested how it could be improved. The authoritarian leader made the groups’ decisions and did their planning. Under the democratic leader, the leader and the teams shared the responsibility for planning and decision making. The teams that received reinforcement leadership were the most productive, and members with little ability appeared to profit most from it. According to a survey of 526 subordinates and 543 peers describing the downward influence tactics of their 128 managers, least effective leadership involved using demands, threats, and persistent reminders; seeking the aid of others to persuade teams; and claiming the authority to legitimate requests. Moderate results were achieved with getting teams to think favorably of the leader before making a request, or making promises of later reward or reciprocating the favor later on. Most effective were rational persuasion and inspirational appeals to values, ideals, and aspirations.
For 11 of the samples in the Podsakoff and Schriesheim (1985) review, the mean concurrent correlation was .32 between contingent rewarding supervision and the effectiveness of the performance of subordinates and work groups. For 10 samples, the mean correlation was close to zero between contingent punishment by the leader and subordinates’ performance. Luthans and Kreitner (1975) did find that contingent penalization for unacceptable actions improved performance, but only when it was coupled with contingent reward for acceptable performance. The mean correlation was also close to zero between performance and noncontingent reward for eight samples and –.12 between noncontingent punishment and performance.
Material versus Nonmaterial Rewards. Pay for performance still remains a popular contingent reward in the United States, Britain, and elsewhere, with schedules set by the human resources staff. Supervisors may recommend pay raises in their performance reports of subordinates, but nonmaterial rewarding may be more important in motivating and maintaining performance. Although pay motivates people to seek and accept employment for better wages and salaries, leaders are likely to have even more discretionary effects on their subordinates in their use of nonmaterial positive feedback of praise and recognition for work well done (Kohn, 1993).
Generally, contingent rewarding correlated almost as highly as transformational leadership with outcomes in effectiveness and satisfaction. Among 19 female managers rated in transformational leadership by 38 subordinates, transformational leadership did not further account for performance of the subordinates rated by the managers beyond what was accounted for by contingent reward (Dionne, Yammarino, Comer, et al., 1996). Behavioral management, which included providing rewards contingent on performance, proved more effective at increasing productivity than did participative management, according to an experiment in a Russian textile factory by Welsh, Luthans, and Sommer (1993). According to a survey of 125 managers by Howell, Neufeld, and Avolio (2005), contingent rewarding was correlated with better communication—frankness (.41), two-way communications (.45), and careful transmitting (.22). Generally, creative performance and divergent thinking were increased by promises of reward contingent on performance (Eisenberger, Armeli, & Pretz, 1998). On the basis of his own experience in the Israeli Army, and after reviewing the literature, Berson (undated) noted that rewarding soldiers for the adequacy of their performance was especially relevant for their motivation.
In cross-lagged causal analyses of 206 employees, Greene (1976a) demonstrated that contingent rewarding by supervisors resulted in the improved subsequent performance and satisfaction of subordinates. The correlation was .48. The effect had some permanency. In a six-month longitudinal field study of 61 working MBA students, Sims (1977) corroborated Greene’s conclusion about how much contingent rewarding by supervisors contributed to the respondents’ subsequent performance. The correlation was .50. Szilagyi (1980b), in a three-month longitudinal study of 128 employees in a controller’s department, obtained a comparable coefficient of .48.
Szilagi’s (1980b) results were reversed or nil (–.23, –.12, –.07) for the effects of earlier contingent punishment by supervisors on the future performance of subordinates in Greene’s, Sims’s, and Szilagyi’s studies. Future punitive supervision appeared to be more a consequence of past poor performance of subordinates: the correlations for the three studies were –.40, –.04, and –.59. But future contingent reward was not consistently a consequence of subordinates’ earlier good performance: the correlations for the three studies were .26, .09, –.17.
Laboratory and field studies have indicated that although contingent reward leads to improved performance by subordinates, contingent punishment has less consistent effects. The performance of subordinates may be improved by a supervisor’s correction and by other forms of negative feedback (Arvey & Ivancevich, 1980). Sims and Szilagyi (1975) suggested that negative sanctions may improve performance by reducing role ambiguity. According to Podsakoff (1982), high-performing leaders use both contingent-positive and contingent-aversive reinforcement. O’Reilly and Weitz (1980) found that managers of high-performing units used negative reinforcers more quickly. Franke and Kaul (1978) reinterpreted the statistical inferences that could be drawn from the Hawthorne experiments (Roethlisberger, 1941) and concluded that managers’ tightening of discipline was an important reason for the increased productivity of workers—a change that was ignored by the original investigators. O’Reilly and Puffer (1983) asked 142 sales representatives to describe critical incidents of good and poor performance with and without supervisory reinforcement. They found that effort, satisfaction, and a sense of equity were enhanced as well by formal or informal reprimands for poor performance as by favorable recognition for good performance.
Reasons for Disciplinary Action. Contingent punishment is more likely to occur if the manager’s authority is challenged, or if the manager is inconvenienced by a trained and experienced subordinate’s failure to live up to expectations. For example, a subordinate might deviate from the norms, production might fall below agreed-on standards, or the quality of the product might become unacceptable (Rollinson, Hook, Foot, et al., 1996). The leader may call attention to the deviation. Being told of one’s failure to meet standards may be sufficient to provide aversive reinforcement for what was done wrong. Being told why can be particularly helpful to an inexperienced subordinate, especially if the message includes further clarification of the desired performance. The leader may administer penalties, such as fines or suspensions without pay. The leader may withdraw support, or the employee may be discharged.
Effects of Disciplinary Actions. Following initial interviews with a random sample of 100 hourly refinery employees, Arvey, Davis, and Nelson (1984) surveyed 526 hourly employees. The employees were asked to describe their perceptions of their immediate supervisors’ disciplinary behavior, as well as the factors that supervisors take into account when applying discipline. Of the employees, 20% reported that they had experienced disciplinary actions. The employees’ evaluation of the supervisors differed strongly, depending on how much and in what ways the supervisors used punishment. The more the supervisors used formal punishment (“Write you up for an infraction”) or informal punishment (“Yell at you”), the less satisfied their subordinates were with them. The employees’ satisfaction with their supervisors correlated –.46 with overly harsh discipline and –.50 with discipline lacking a clear explanation. On the other hand, satisfaction with the supervisors correlated .31 with the supervisors’ tendency to discipline people immediately after an infraction of the rules occurred, .60 with the tendency to maintain close and friendly relations with employees they had disciplined, and .32 with consistency in disciplining employees from person to person and situation to situation.
Justifications for Punishment. Future-oriented punishment aims to reduce or eliminate undesirable behaviors such as unsafe behavior. By itself, it does not reinforce the correct way to act, but it can avoid costly errors. Past-oriented punishment, by contrast, is retribution for past misdeeds, and it may only assuage the administrating leader and increase the emotional reactions of the follower. According to qualitative interviews with 77 managers from diverse organizations by Butterfield, Trevino, and Ball (1996), the interviewed managers generally were uncomfortable with punishing subordinates for poor performance and sought to avoid it. Many managers felt pressure against it from the subordinates as well as from the organization and work groups. These managers were aware that the emotional effects could spread much beyond what was intended. Nonetheless, the investigators concluded that punishment could be instrumental in improving subordinates’ behavior, attitudes, respect, and vicarious organizational learning. Many of the managers engaged in impression management when punishing, using justifications and excuses in acknowledgment of past mistakes. They considered the subordinates’ expectations, reactions, and acceptance, and their own past relations with the subordinates. To many of the managers, fairness and timeliness were less important than consistency, privacy, and the poor performance of the subordinates. Judging from the reactions of 421 graduate business students to scenarios, Fukami and Hopkins (1990) concluded that the severity of punishment meted out depended on the subordinates’ past performance and the amount of property damage and personal injury, not on whether the offending employee was a man or a woman.
Incorrect behavior may be only temporarily suppressed by disciplinary punishment alone, but when disciplinary punishment is coupled with advice on correct behavior, it is more likely to suppress the undesired behavior permanently (Arvey & Ivancevich, 1980). In a simulation experiment in which 182 undergraduates played roles as managers and poorly performing workers, the “managers’” advice added to corrective action increased the “workers” sense of interactional justice.
Content analyses by Atwater, Waldman, Carey, et al. (2001) of interviews with 123 disciplined persons and 46 observers disclosed some positive effects of the discipline. But respect for the discipliner was lost and negative attitudes developed toward the organization. The discipline was seen as unfair, particularly when it was applied for violations of informal rather than formal rules. Ball and Trevino (1990) asked 65 managers from different organizations and different kinds of positions to indicate whether they had been pleased or discouraged by a disciplinary event they had reported and whether justice had been served. As expected, the perception of justice depended on the appropriateness of the severity of the punishment (r = .31), the closeness of the punishment to the violation (r = .22) and the constructiveness of the punishment (r = .54). Managers were less discouraged when the punishment was perceived as just (r = –.21), when it was given in private (r = –.26), and when the manager perceived the relationship with the disciplined employee as remaining good (r = –.29).
Management by Exception. When leaders take corrective actions and intervene only when subordinates’ failures and deviations occur, they are practicing management by exception (MBE). In practicing MBE, some managers actively search for deviations and shortfalls. They set up standards and regularly monitor subordinates’ performance to see if the standards are being met. Other managers are more passive, asking no more than what is essential to get the work done (Hater & Bass, 1988). Leaders who practice MBE intervene only when something goes wrong. If a subordinate’s performance falls below some threshold, the leader feeds back information to the subordinate, at the emotionally mildest level, that the threshold has been crossed. The negative feedback may be accompanied by a contingent reward from the leader, in the form of reclarification and encouragement, At the other extreme, it may be accompanied by disapproval, a reprimand, a formal citation, suspension, or discharge. MBE is consistent with the cybernetics of negative feedback—feedback that signals the system to move back toward its steady-state base. The manager is alert for deviations and provides the negative feedback as needed.
A simulation of management control strategies used 182 students acting as managers and poorly performing employees. Punitive corrective actions and verbal reprimands were used repeatedly; but over time, the “managers” experimented in the search for effective management control (Gavin, Green, & Fairhurst, 1995) and the “employees’” sense of justice was affected. In most surveys of descriptions of industrial leaders by their colleagues and subordinates, MBE—especially if it is passive—does not contribute to positive appraisals of the leaders’ effectiveness or to satisfaction with the leaders (Bass, 1985a). However, it correlated .23 and .25 with the rated effectiveness of officers in the U.S. Army and U.S. Air Force and .29 and .31 with subordinates’ satisfaction with the officers’ leadership (Colby & Zak, 1988).
Particularly important to the effectiveness of leadership is whether MBE is active or passive. A leader who arranges to monitor errors and deviance that need to be corrected is said to be actively engaged in management by exception. A leader who waits to be informed about errors and deviance before taking action is said to be practicing passive management by exception. The correlations for 186 junior naval officers described by their subordinates as actively managing by exception were .50 and .54 with the subordinates’ evaluation of the effectiveness of the officers and satisfaction with them, and .28 and .22 with fitness reports and recommendations for early promotions from their superiors. The same correlations with passively managing by exception were, respectively, .11, .19, –.04, and –.05 (Yammarino & Bass, 1988).
Personal and Situational Modifiers. In interviews with 474 managers about their use of discipline, Beyer (1981) found that both personal and situational factors could account for the extent to which supervisors took disciplinary actions with problem employees. For example, managers who personally endorsed humane pragmatism—concern for both the employees’ welfare and the company’s productivity—were more likely to be constructive and confrontational with problem drinkers and less likely to take disciplinary action against these employees. At the same time, the judged seriousness of the drinking problem was the strongest determinant of whether disciplinary action, including issuing warnings in writing or ordering a suspension or discharge, would be taken. Union and management policies had systematic effects, as did the managers’ expectations that their constructive or disciplinary efforts would or would not please their higher authority. Unfortunately, the results were often different when the managers dealt with problem employees other than those who had drinking problems. Commentaries suggest that less serious disciplinary cases are handled personally by supervisors with a prescriptive autocratic style. More serious threats to the organization result in stronger disciplinary action. Additionally, whether the supervisor takes any disciplinary action depends on the subordinate’s length of service and on how much the subordinate’s violation inconveniences the supervisor, is a threat to the organization, and challenges authority. Also important are the supervisor’s training and the subordinate’s length of service (Hook, Rollinson, Foot, et al., 1996; Rollinson, Hook, Foot, et al., 1996). Observing coworkers who are being threatened or punished by supervisors can also provide vicarious reinforcement that promotes the productivity of the onlookers. Schnake (1986) hired students at $5 an hour ostensibly to perform clerical work, along with a confederate of the experimenter who made sure to perform more poorly than the student subjects. Threats to reduce the confederate’s wage to $3.50 and actually cutting the confederate’s pay resulted in significant increases in the output of the student subjects.
According to Zaleznik (1967), contingent reinforcement with an emphasis on the exchange should be practiced by managers who are interested in efficient processes rather than substantive ideas. It will be facilitated by the managers’ flexibility in using their powers to reward or punish to maintain or improve processes and organizational arrangements. Blanchard and Johnson (1982) advised that inexperienced subordinates may need to be told by their superior what they did right and how their superior feels about their work, and encouraged to continue. Experienced subordinates may need to be told what they did wrong without diminishing their self-esteem. Furthermore, a reprimand will work if it is timely and specific to the behavior involved, not to the person. It should focus on what was done wrong, how the supervisor feels about it, and the expectation that the subordinate’s performance is going to improve. The subordinate’s failure to appreciate what is expected calls for a clarification of requirements, rather than punitive actions. Failure that is due to a lack of knowledge calls for information, training, or transfer. Failure that is a result of insufficient challenge calls for enlarging the task, transfer, increased extrinsic incentives, or acceptance of reality by the subordinate. A lack of commitment requires that the employee’s involvement in the planning or control process be increased. An overload of goals necessitates the establishment of priorities.
According to Scott and Podsakoff (1982), if leadership and followership can be conceived of as behaviors, then the principles of operant reinforcement can be used to modify these behaviors. Leaders need to learn what reinforcers work with what kinds of followers. They also must be able to identify and specify the types of behavior by subordinates that will maximize the group’s performance. If this is not possible, they need to identify and specify the desired results and to keep up with changes that are required. Over time, a shift of control by the leader to self-control by the followers is salutary.
If extrinsic reinforcers are used, leaders must be able to link the reinforcement to the subordinates’ performance. Leaders need to be consistent in their administration of reinforcements. Punishment must be delivered quickly, and with reasons, in order to stamp out undesirable responses. Environmental events that sustain the dysfunctional behavior need to be eliminated. McElroy (1985) also argued that reinforcement is not received mindlessly. Rather, when reinforced, a subordinate will develop a reason for the reinforcement that may be the same as or different from the reason given by the leader. It is likely that the subordinate will attribute rewards for successful performance to his or her personal efforts but attribute penalties for failure to external causes. It is presumed that because of this attribution, penalties will have less effect on the subordinate’s subsequent behavior. Schedules of reinforcement will systematically affect causal attributes. For example, if one is rewarded only some of the time for good performance, one will tend to attribute the reward to effort or luck, rather than to ability.
Caveats. Caution is required in applying reinforcement theory that is based on animal research to human leader-subordinate relations. If animals are faced with choices X and Y and are reinforced 90% of the time for choosing X and 10% of the time for choosing Y, they will choose X 90% of the time. However, humans will make a cognitive leap, deducing that they will maximize their reward by choosing X 100% of the time. Animals will earn rewards only 81% of the time; humans will earn rewards 90% of the time (Mawhinney, 1982).
Using contingent reward or punishment is a tricky business. A 100% schedule of praise for every success will probably satiate the subordinates, lead them to discount the supervisor, and cause them to perceive the superior’s behavior as ingratiating. Continued praise in front of associates may create considerable feelings of discomfort and defensive feelings. Too infrequent a schedule of contingent praise may raise questions about the superior’s motivation. The subordinate’s expectation of an extrinsic reward that is contingent on his or her performance may reduce the intrinsic motivation to continue that performance (Deci, 1972). Thus in a field study of 48 health care technicians caring for handicapped and mentally retarded children, Jordan (1986) found that the intrinsic motivation of those whose rewards were contingent on performance decreased 21 or 22 months after the extrinsic incentive program started. The technicians’ intrinsic motivation increased if the rewards of the incentive program were not contingent on performance. Their satisfaction with the extrinsic reward of pay was not affected, however.
Mild punishment in private, in the form of suggested corrections by the superior of the subordinate’s behavior, may be salutary. Severe punishment, in the form of a suspension, may provoke hostility, resentment, and anxiety. Frequent contingent punishment may increase anxiety to the point where it interferes with the subordinate’s performance. Contingent punishment that is too infrequent may allow incorrect habits to become so entrenched that they are difficult to eliminate.
In a nationwide survey of 5,000 employees, almost half the respondents said that the managers at their firm were too lenient with poor performers (Anonymous, 1988a). Nonetheless, according to Veiga (1988), a majority of executives are in favor of candor in dealing with problem subordinates. Unwanted behaviors and outcomes need to be clarified. Both the leader and the subordinate need to acknowledge their responsibilities for the situation. Clear-cut expectations about what will be changed need to be established, and positive support for making a small new change needs to be given.
On the basis of theory and research, Arvey and Ivan-cevich (1980) offered a set of propositions that outline the conditions under which punishment is likely to be effective in organizational settings. They stated that punishment should be delivered as close as possible to the time of the subordinate’s undesired behavior, should be given by an otherwise close and friendly supervisor, and should be moderate in intensity. Furthermore, punishment should be accompanied by explanation. Alternatives to the undesired behavior should be available.
Contingent motivation remains a reality at the macro level. It was possible to eliminate the draft and change to an all-volunteer military. Opportunities for education, training, and advancement became important inducements to influence volunteers to enlist in the U.S. Armed Forces, Reserve, and National Guard. But at the time of this writing, as the insurrection in Iraq had heated up and threats to safety and security had increased, enlistment rates had fallen.
Bass (1985a, pp. 147, 149) presented a model of the most important linkages between contingent-positive and-aversive reinforcement by leaders and the resultant effort of their followers to comply and carry out the transaction:
Leaders … reward followers to encourage the followers’ acceptance of their work roles. Followers comply with the leaders’ directions … to gain the rewards promised by the leaders for such compliance. If the followers succeed, they earn material rewards, … satisfaction and enhanced self-esteem. … [In maintaining their] role behavior and … [in] renewing their efforts … [to] comply with what is expected of them … the leaders also clarify such expectations. Such clarification promotes followers’ understanding of their roles and builds followers’ confidence, further contributing to their compliance.
If the followers fail to comply and the failure is attributed by the leaders [to a lack of] follower clarity and understanding, then the leaders will renew their clarification of what they expect. If the leaders attribute the failure of their followers to lack of motivation [and they wish to be positively reinforcing] the leaders will renew their promises of reward and confidence in the followers. … [Or if the leaders wish to use] aversive reinforcement such as reprimand, they [run the risk] of followers’ anxiety and withdrawal.
Leaders practicing contingent aversive reinforcement … will foster followers’ efforts to comply with the clarified standards to avoid negative consequences for failure. If followers succeed in complying, they avoid being aversively reinforced and may increase in self-esteem and self-reinforcement. If they fail and leaders attribute the failure to lack of clarity, ability, and understanding, the leaders will renew clarification and attempt to improve followers’ ability through training, thus increasing the likelihood of ultimate successful performance by followers. On the other hand, if aversively reinforcing leaders attribute followers’ failure to comply to [a] lack of follower motivation, they are likely to reprimand or threaten, possibly generating the unintended effects on followers of hostility, apathy, anxiety, and loss of self-esteem. In turn, there will be a reduction in self-reinforced effort and interference with the efforts of followers to comply.
Contingent Reward Behaviors. Klimoski and Hayes (1980) found four of six behaviors of supervisors that 231 surveyed subordinates thought were ways their supervisors provided contingent reward: (1) being explicit when giving instructions, (2) allowing the subordinates to be involved in determining standards of performance, (3) giving support for efforts to perform effectively, and (4) being consistent toward subordinates. When a supervisor behaved in these four ways, subordinates expected that their effort would lead to successful performance and that their successful performance would generate commensurate rewards. Two other seemingly relevant behaviors of supervisors—frequency of communications and frequency of reviews—were irrelevant to these subordinates’ expectations.
Yukl (1981) used content analysis to derive a category of “structuring reward contingencies”—the extent to which a leader rewards a subordinate’s effective performance with tangible benefits, such as a pay increase, promotion, more desirable assignments, a better work schedule, or more time off. Examples of contingent reward structuring included: “My supervisor established a new policy that any subordinate who brought in a new client would earn 10 percent of the contracted fee” and “My supervisor recommended a promotion for a subordinate with the best performance record in the group” (p. 122).
Leaders who give contingent rewards establish “contracts” with subordinates to set goals for each important aspect of a subordinate’s job, measure the subordinate’s progress toward reaching the goals, and provide concrete feedback. Yukl (1981, p. 123) cited these examples: “The supervisor held a meeting to discuss the sales quota for next month” and “My supervisor met with me for two hours to establish performance goals for the coming year and to develop action plans”
Using factor analyses of survey results, Sims (1977) found that contingent reward took two main forms: (1) praise for work well done; and (2) recommendations for pay increases, bonuses, and promotions. Also, there may be commendations for meritorious effort and honors for outstanding service. The first positive-reward behavior factor explained 37% of the variance among the supervisors. Items most highly correlated with this first factor included: “Your supervisor would show a great deal of interest if you suggested a new and better way of doing things.” “Your supervisor would give you special recognition if your work performance was especially good.” “Your supervisor would personally pay you a compliment if you did outstanding work” (p. 126). The second factor, which accounted for 18% of the variance, dealt with recommendations for promotion and advancement. Items highly correlated with the second factor included: “Your supervisor would see that you will eventually go as far as you would like to go in this organization, if your work is consistently above average.” “Your supervisor would recommend that you be promoted if your work was better than others who were otherwise equally qualified.” “Your supervisor would help you get a transfer if you asked for one” (p. 126). Bass and Avolio (1989) validated a scale of 10 items within the Multifactor Leadership Questionnaire (MLQ) to measure contingent reward in an exchange context that contained items such as: “Tells me what to do if I want to be rewarded for my efforts” and “There is close agreement between what I am expected to put into the group effort and what I can get out of it.”
Management by Exception and Contingent Punishment. Yukl and Van Fleet (1982) found numerous critical incidents of management by exception and contingent punitiveness among officers in Korea, such as: “Checks to see that tasks are accomplished satisfactorily.” “Shows concern about the appearance of cadets.” “Expresses disappointment that the unit did not perform better.” “Explains to a person why he is being disciplined.” Sims (1977) uncovered a factor of supervisory behavior that dealt with contingent punishment. The items include: “Your supervisor would get on you if your work was not as good as the work of others in your department” and “Your supervisor would give you a reprimand [written or verbal] if your work was consistently below acceptable standards.” As was noted earlier, a contingent-aversive or -punitive factor is involved in the practice of management by exception. Hater and Bass (1988) and Yammarino and Bass (1988) created factored scales of active and passive management by exception within the MLQ. The active practice of management by exception (MBEA) dealt with monitoring and searching for subordinates’ deviations from standards and making suitable corrections (“Checks to see if things are going along all right”). The passive practice of management by exception (MBEP) involved reacting only when deviations were brought to the manager’s attention (“Endorses the adage, ‘If it ain’t broke, don’t fix it’”). Compared with other ways of measuring leadership performance, these behavioral scales of contingent reinforcement by leaders tend to be relatively free of the usual biases of respondents to rate according to the social desirability of the behaviors (Podsakoff, Todor, Grover, & Huber, 1984) or the raters’ implicit theories of leadership (Bass & Avolio, 1988). MBEA tends to correlate with contingent rewarding whereas MBEP tends to correlate with laissez-faire leadership (Bass, Avolio, & Jung, 1999).
Scales of Contingent Reinforcement. The Contingency Questionnaire (Johnson, 1970; Reitz, 1971) assessed the extent to which subordinates perceive contingencies between their performance and how the organization responds to it. Three factors emerged: (1) supportive instrumentality—the contingency between the subordinate’s behaviors and a supportive or rewarding organizational response; (2) punitive instrumentality—the contingency between the subordinate’s undesirable behaviors and the administration of punitive events by the organization or the withdrawal of rewards; and (3) advancement instrumentality—the contingency between the subordinate’s behavior and the subordinate’s advancement in the organization. The three factors accounted for approximately 50% of the common variance in the scale of 20 items. House and Dessler (1974) developed four-item scales to measure leader-contingent approval behavior and leader-contingent disapproval behavior. Scales to assess noncontingent reward and noncontingent sanctioning were also added. Hunt and Schuler (1976); Fulk and Wendler (1982); and Podsakoff, Todor, and Skov (1982) likewise developed and evaluated the Leader Reward and Punishment Questionnaire, which had a total of 23 items to measure both contingent and noncontingent reinforcement by the leader. In typical items from the most recent report, supervisors were described in paraphrase as follows: Contingent reward—personally compliments me when I do outstanding work; Noncontingent reward—Praises me when I do poorly as when I do well; Contingent punishment—would reprimand me if my work was below standard: Noncontin-gent punishment—holds me accountable for things over which I have no control. Schriesheim, Hinken, and Tetrault (1988) corroborated the factorial validity of the four dimensions of contingent reward, noncontingent reward, contingent punishment, and noncontingent punishment with the responses of 176 aircraft controllers and 375 employees of a psychiatric hospital to a questionnaire in which they were asked to describe their supervisors.
In addition to the use by leaders of contingent reinforcement to develop and maintain effective follower compliance, contingent reinforcement processes also can explain who emerges as a leader. A variety of experiments have been conducted to show how reinforcement can increase individual members’ attempts to lead and their processes in leadership.
Who attempts and continues to lead and who attempts and continues to follow in newly formed groups without appointed leaders can be explained by the differential reinforcement experiences of the various members of the groups. To be successful in influencing others, one must attempt to lead. Such attempts to lead are increased by the positive reinforcement of success as a leader and are decreased by failure and aversive reinforcement. If one member’s attempts at leadership are inhibited by failure, the leadership attempts of others may increase. The positive reinforcement of one member can result in “crowding out” the other members. Contrived reward has been applied in experiments to test its efficacy in reinforcing attempts to lead the interaction process, as well as its efficacy in reinforcing the probability of successfully leading others.
Bavelas, Hastorf, Gross, and Kite (1965) assessed the amount of activity of each participant in initial discussions without an appointed leader. In a second session, only one particular member of each discussion group was reinforced for participation, by the flashing of a green light. Additionally, in some of these groups the participation of the remaining members was inhibited by the flashing of a red light. When other members were not inhibited by a red flashing light, those who were positively reinforced with the green flashing light significantly increased their talking (which is highly correlated with attempts to lead in initially leaderless discussions). But it was necessary to use positive reinforcement (the green flashing light) with the initially least active talker and to inhibit (with the red flashing light) the initially most active talker in order to increase the amount of talking by the originally least active talker. The least active talker’s gain was mostly at the expense of the originally most active talker.
Jaffee (1968) and Jaffee and Skaja (1968) reinforced the leadership acts of one member of pairs of experimental partners. The number of leadership attempts by the members who were reinforced by the appearance of a light significantly exceeded attempts by the nonrein-forced and control participants. The effects lasted over a test period of one week and were found to generalize to a different experimental situation. Zdep and Oakes (1967) likewise demonstrated that reinforcement of a group member’s leadership acts resulted in an increase in talking and leadership. In the same way, Binder, Wolin, and Terebinski (1965, 1966) and Wolin and Terebinski (1965) controlled the proportion of decisions as a leader for which each member of a group was reinforced. The resulting emergence as a leader fit a Markov mathematical model. Increases in leadership behavior were found to be responsive to changes in such rates of reinforcement.
Aiken (1965a) rewarded a group member in a second session who had talked the least in a preceding session. The other three members were punished for speaking in the second session. The rewarded member’s verbal output significantly increased in comparison with the output of unrewarded control group members, but punishment did not significantly decrease output. After the second session, the three unrewarded members of the experimental group rated the rewarded member higher than members of the control group rated each other in leadership, participation, and self-confidence.
In these experiments, mainly with school pupils and college students, group members whose leadership was reinforced with success rather than failure were more likely to increase their influence over the other members. Mausner (1954a) gave participants either positive, negative, or no reinforcement when they were working alone on a judgmental task. When combined in pairs, the non-reinforced partners had a significant tendency to shift their judgments to comply with those of their reinforced partners, whereas the reinforced participants did not change. The participants who had received positive reinforcement were significantly less influenced by their partners than were those who had received negative reinforcement. Thus Mausner demonstrated that the participants’ perception of the success of their partners gave them prestige and reinforced their expectation of future success. The participants who worked with previously successful partners were significantly more influenced by them than were those who worked with previously unsuccessful partners in a judgmental task. Banta and Nelson (1964) showed that for positively reinforced participants, the probability of having their suggestions adopted increased over 60 trials, whereas for negatively reinforced participants, the probability decreased. Kanareff and Lanzetta (1960) also found that the tendency to be imitated by others was related to the rate at which the imitated participant was positively reinforced. McClintock (1966) used paid accomplices who either supported, shifted their support of, or did not support the leaders of discussion groups. It was found that leaders evidenced more release of tension, more positive affect, and more task-oriented responses when they were supported than when they were not supported.
Cohen and Lindsley (1964) gave pairs of subjects a monetary reward for social responses or punished them, by blacking out the room, for individual (nonsocial) responses. They rewarded acts of leadership by opening a panel that permitted subjects to see each other briefly. Differential leadership responses were established by the differential reinforcement procedures. James and Lott (1964) rewarded some members of an experimental task group with six nickels, rewarded others with three nickels, and did not reward still others. Those who were rewarded with six nickels chose each other significantly more often than did those who were rewarded with three nickels or those who were not rewarded. G. Gardner (1956) studied groups of boys who were assigned fictitious scores for success in operating switches that activated various combinations of lights. The scores correlated .93 with their nomination as team captain. Observations of success appeared to reinforce the group’s perception of a member’s suitability for the leadership role. Katz, Blau, Brown, and Strodtbeck (1957) found that the greater the extent to which the leader’s suggestions were reinforced by the follower’s acceptance and support in the first session of a discussion task, the more the leadership was retained in later sessions. Likewise, York (1969) and Hastorf (1965) demonstrated that reinforcement of member’s behavior in a group significantly increased their leadership.
Hamblin, Miller, and Wiggins (1961) studied the effects on a group’s morale of the leader’s competence, reinforcement of the leader, and conflicting suggestions by group members. Reinforcement rate, perceptions of the leader’s competence, manipulation of the leader’s success, and the group’s morale were all highly and significantly intercorrelated. The members’ opposition to the leaders’ suggestions was reduced by reinforcement of the leader, the leader’s success and competence, and the group’s morale. Mausner and Bloch (1957) found that reinforcement of a partner and the partner’s prestige interacted in a judgmental task. That is, both the partner’s prestige and earlier reinforcement contributed to the partner’s influence in a condition of success but not when failure was contrived. Reinforcement did not overcome the effect of a partner’s failure to make correct judgments in a previous situation.
Caveats. Some qualifying considerations are in order. Reinforcements may have unintended and unwanted effects on attempted and successful leadership. Zdep (1969) assembled groups of four members in which one member scored either high or low in self-reported leadership (the Leadership Scale of the California Psychological Inventory). Under the experimenter’s reinforcement, the participation rate increased for high scorers but not for low scorers. As the high scorers’ rate of participation rose, the subsequent leadership ratings by the other members increased. But the experimenter’s reinforcement of one member built up expectations about that member among the others. Low scorers who failed to respond when given the reinforcement were rated as poorer leaders than low scorers who were not so reinforced. Aiken (1965b) also found that the use of reinforcement for operant conditioning of specific behaviors, such as giving suggestions and asking for contributions, resulted in marked changes in leadership style but produced a tense, anxious leader. This finding suggested that simple reinforcement technology, by itself, may be inadequate for leadership training. The shaping of one member’s responses must take into consideration the larger context of the interaction with other active participants.3
The superior-subordinate relationship has been conceived of as a social exchange or negotiated transaction. This conception led to a fully developed theory to explain the effects of leadership on the compliance of subordinates. Leader-member exchange implied an informally developed role—one that was negotiated between each individual group member and the leader (Graen, 1976).4 Coworkers may get involved in the role definition, but the leader, in particular, has a vested interest in the member’s role. The definition of a member’s role indicates what the member and the leader will expect the member to do. This was found with 62 new nonacademic employees on a college campus by Graen, Dansereau, Minami, and Cashman (1973). Given the leader’s control of reinforcements, the interpersonal exchange relationship of the member and leader is particularly important in shaping the member’s performance.
The leader-follower transactional exchange is the essence of the dynamics of leadership, according to Hollander (1958, 1978). Hollander concluded that the leader and the follower enter into an exchange that begins with negotiation to establish what is being exchanged and whether it is satisfactory. The exchange results in the follower’s compliance for the leader’s assistance in pointing the way toward the attainment of mutual goals.
In real life, there is no experimenter to reinforce the attempts of one member rather than another to lead. On a playing field with similar personalities and a common situation, what, then, makes one member a leader and another a follower? According to Hollander, it is a matter of how much idiosyncrasy credit each member has previously earned. The interplay that determines who will lead and who will follow builds up idiosyncrasy credit for the emerging leader relative to the follower. Competence in the group’s task and conformity to the group’s norms earn idiosyncrasy credit for a member and, as a consequence, emergence as leader or success as an appointed leader. A buildup of such credit makes subsequent initiatives even more likely to be accepted and strengthens the member’s status as a leader. With idiosyncrasy credit, the emerging leader can deviate from the group’s norms in a way that is not permitted to those without such credit. Newcomers, for instance, are unlikely to have the necessary credit, although some may bring derivative credit with them from another group as a consequence of their reputation in that group. There is a complex interplay between the requisite competence and the requisite conformity in maintaining idiosyncrasy credit. Alvarez (1968) observed that a leader’s greater deviance from the norms was tolerated when it was accompanied by the group’s success. In a study of a simulated organization, Alvarez found that the leader lost credit at a slower rate than did the followers, but only in successful organizations. In unsuccessful organizations, the leader lost credit at a faster rate than the followers.
Consistent with the concept of idiosyncrasy credit, Zierden (1980) suggested that leading through the followers’ points of view is the way to achieve success and effectiveness as a leader. Litzinger and Schaefer (1982, p. 139) saw the fundamental conformity of leaders to their followers’ norms and values as being manifested in the way the Pope John Paul II called himself “The Servant of the Servants of God” and in Hegel’s conception of leadership: “Leadership is possible … not only on the condition that followership has been learned, but on the more radical condition that the leader has known … the travail of the follower; he must here and now incorporate within himself all that the follower is. The school for leadership is indeed followership, a followership that is fully preserved within leadership, but transformed for having moved beyond itself.” Litzinger and Schaefer also saw the idea that good followership precedes good leadership in the West Point faculty doctrine that leadership is developed by cadets first learning how to be good followers. Prior to gaining powerful office, many world leaders, such as Otto von Bismarck and George H. W. Bush, were seen as “faithful takers of orders.” To be elected, the first President Bush strongly espoused the norms of the majority of the electorate about law enforcement, the flag, taxes, education, and the environment. Once he was elected, his actions could deviate to some extent from these espousals without alienating too much of his majority constituency. President Bill Clinton consulted public opinion polls daily. President George W. Bush spoke of the political capital he earned by winning reelection in 2004, which would make it possible for him to launch new programs. But he failed to recognize that he could rapidly lose his credit with increasingly unpopular policies such as “staying the course” in Iraq. By February 2006, according to a Zogby-LeMoyne (2006) poll, only 23% of 944 American soldiers serving in Iraq supported staying as long as necessary. They agreed with polls of previous civilian majorities that the policy was wrong. But President George W. Bush said that to be successful, all he had to do was continue to ignore his critics and please the majority of voters who had elected him. By 2008, it was clear that a majority of citizens had no desire for an open-ended war in Iraq and wished for an exit strategy to be developed.
Limitations on Idiosyncrasy Credit. With idiosyncrasy credit, the leader can step out from the rest of the pack and suggest innovations that will bring about changes in the group. Nevertheless, although the leader is able to deviate from the group’s norm, he or she must continue to pay attention to the group’s norms, values, and standards in order to remain successful in influencing others. Holders of idiosyncrasy credit have other limitations on the deviance that they are permitted. For example, they have less latitude to deviate from particular role obligations (Hollander, 1961a, 1961b), although they may deviate from norms with less cost than other members. In effect, the leader’s freedom to deviate from general norms is exchanged for conforming more closely to the expectation others have about the requirements of the leader’s role.
A leader will lose much of his or her idiosyncrasy credit by an outrageous violation of the members’ expectations. In 2006 President George W. Bush got into trouble with his supporters as well as his critics and lost a lot of credit when, after five years of creating a climate of fear by maintaining that terrorism was the greatest continuing threat to national security and well-being, he gave immediate and strong support for a firm owned by an Arab country to take over the management of six U.S. seaports (Krugman, 2006). Ironically, the port deal was innocuous. Still, his policy of “staying the course” in Iraq became increasingly unpopular and resulted, in June 2006, in his accepting a plan for a phased withdrawal of American troops.
There is experimental evidence about possible limits of deviation from norms. Hollander (1961a) conducted an experiment in which the participants were given brief descriptions of persons who were to be ranked from high to low in accorded status. The participants were then given lists of behaviors that might be exhibited by the persons whom they had ranked. As the perceived status of those who were ranked increased, the participants expressed progressively less disapproval for “suggesting changes from group plans” and “discussing group concerns with outsiders.” However, “interrupting others to make comments” was increasingly disapproved for persons with increasing status. High-status members were permitted to innovate group changes but were not expected to interrupt low-status members.
Looking at the same issue, Michener and Lawler (1975) found that group members were more likely to endorse the continuance of leaders in their positions when the groups were positively reinforced by success in their efforts, when the members got more rewards than the leader, or when the leaders were not permanently fixed in the office. But the members were less enthusiastic about endorsing the continuance of leaders in office if the leaders were inconsistent in their competence or fairness. Polansky, Lippitt, and Redl (1950a) and Grosser, Polansky, and Lippitt (1951) showed that those group members who attempted to be more influential and those who were more successful in this attempt were more susceptible to contagious influence from other members. Compared with members who tried but failed to exert influence on the group, they were more likely to act in reference to the demands of the group situation than in terms of their own needs. Although they were able to react spontaneously to others and to initiate spontaneity in others, they also felt secure enough to resist direct attempts by others to influence them. They could maintain conformity to the steady state as well as enlarge the other members’ freedom to act.
In the social exchange, followers may exert influence and make demands on the leader. Negotiation may be required about what will be done and by whom. The negotiation may consider the distribution of effort and rewards between the leader and the followers, including need, justice, merit, equity, and fairness. Dyer, Lambert, and Tracy (1953) observed that in the more effective of two bomber wings (the wing in which the most successful leadership was likely), there were more favorable attitudes toward the method of allocating rewards and punishments. A satisfactory interchange between the leader and followers has been found to be enhanced by mutual trust (Deutsch, 1973), a sense of fairness and equity, mutual support, involvement, and wider latitudes of acceptance (Dansereau & Dumas, 1977; Graen, 1976). But the interdependence of the followers affected how important equity was the satisfactory exchange. According to Miller and Hamblin (1963), differential rewarding for relative achievement decreased a group’s productivity when the members were highly interdependent but not when they were less interdependent.
True equity would require that rewards be commensurate with contributions. But equity is relative. Members judge the equity of their costs and benefits in comparison with the costs and benefits of others whom they regard as similar to them (J. S. Adams, 1963). Negotiations may be between the leader and the group or between the leader and each individual member. Graen and Cash-man (1975) suggested that such individualized exchanges will be closer to the parties involved than an exchange between the leader and the group would be. Such closeness will increase acceptance of more responsible tasks by the individual members and also increase the leader’s assistance of the members. There will also be more support, sensitivity, and trust in the closer relationship.
How much can a leader depend on using contingent reinforcement? What limits it? What moderates its use and its impact on subordinates? Like most other concepts and practices in social science, contingent reinforcement is neither a universal panacea nor of trivial importance. Rather, the limitations to its use and effectiveness need to be detailed. For instance, subordinates may take shortcuts to complete the exchange of reward for compliance. Quality may be sacrificed to obtain rewards for quantity of output. Complicated piece-rate, reward, and bonus systems generate ambiguity and are likely to induce game playing and a fear of rate busting. Subordinates may react defensively. Reaction formation, withdrawal, or hostility may ensue. Also, the schedule of reinforcements—their timeliness, variability, and consistency—will have considerable effect on the inducements (Bass, 1985a).
Perceptions of injustice may outweigh cost-benefit considerations. Tyler, Rasinski, and McGraw (1985) found that the changes in the governmental benefits that 584 undergraduates and 300 Chicago residents personally received were less important to them than more abstract judgments of injustice when they endorsed their leaders. In the same way, federal managers regarded merit-based pay as inequitable and as undermining their agencies’ effectiveness. However, they believed that performance appraisal and performance standards contributed to supervisor-subordinate communications and to the planning of work, particularly if the appraisals and standards focused on the subordinates’ development (Gaertner & Gaertner, 1985). Silverman (1983) argued that the merit-pay contingent reward system in the federal government failed for numerous political, economic, and psychological reasons, including open-ended regulations, drifting implementation policies, overcomplication, inconsistent treatment of employees, the compression of managers’ pay, and simultaneous budget reductions. To this list could be added administrative blunders and unintended statutory provisions. Pearce, Stevenson, and Perry (1985) collected data two years before and two years after the introduction of a merit-pay program in the Social Security Administration and found no improvement in organizational performance.
As was already noted, contingent rewards for compliance may entail recommended increases in pay or advancement. If it is possible to make pay fully contingent on performance, such as by placing employees on a straight piece-rate or straight commission basis, productivity may rise as much as 30%, according to some studies. But in practice, for most work, there is likely to be little linkage between pay and performance, despite revealed wisdom to the contrary. Often when workers are placed on a piece-rate or straight commission basis, the work group exerts strong pressure to restrict all its members to conform to the same standard output so that all earn the same pay (Rothe, 1960, 1961). At the executive level, Barkema and Gomez-Mejia (1998) have argued for linking pay to performance, but within a larger framework of a long-term orientation and relative to a designated level of compensation. They recognize that in addition to performance, compensation should reflect market conditions, peer compensation, personal and positional factors, and the size of operations. Miscellaneous other factors need to be considered, such as the tax system, industry regulation, and the ownership structure.
In Japan, a familiar practice is to link pay and promotion primarily to age and seniority, not meritorious performance. In the United States, according to an interview study of a representative sample of 845 U.S. workers by Yankelovich and Immerwahr (1983), only 22% said there was a direct relationship between how hard they worked and how much they were paid. A similar representative national survey of 5,000 U.S. employees found that only 28% saw a correlation between the salaries and performance of workers (Anonymous, 1988b). In the first survey, 61 percent said they preferred a closer link between performance and pay.5 In the second survey, almost half the respondents said that managers at their firms were too lenient with underperformers and continued to reward them despite their continuing poor performance. Even when rewards have been found to be distributed according to merit, the link is modest. Lawler (1966c) surveyed 600 middle-and lower-level managers in several organizations and discovered only a small correlation between their pay and their performance ratings. In examining salary ranges at a single corporation, Patten (1968b) found no statistically significant evidence that employees whose performance was rated more highly were also paid better. Heneman (1973) reported one instance of a high correlation between performance and reward: .65 for a sample of 51 managers. In line with these contradictory findings, Kopelman and Reinharth (1982) found much variation in the correlation between performance and reward among 10 branches of a large financial organization.
In contracts, the compensation of chief executive officers (CEOs) is usually related to industry norms and comparable firms and depends not only on a CEO’s performance but also on the size and scope of the organization (Deckop, 1987). Firms going into bankruptcy may award the CEO a multimillion-dollar compensation because of contractual obligations, not the CEO’s performance. However, Gomez-Mejia, Tosi, and Hinkin (1987) found that the compensation of CEOs was linked to their performance if the firms were owner-controlled.
In simulations, recommended pay increases by 4,255 managers in an international survey correlated with meritorious performance, although they were modified extensively outside North America for other reasons, such as family considerations and personal circumstances (Bass, Burger, et al., 1979). Nonetheless, awarding more pay or less pay to subordinates seems to be a matter of the recommenders’ values. Bass (1968a) reported that students’ value orientations were related to their generosity in recommending pay increases for fictitious engineers. Students who were lower in intelligence and ability and those who scored high in social and service values, as opposed to theoretical and economic values, were more generous in recommending salary increases. In the international survey, Bass, Burger, et al. (1979) found that more successful managers recommended significantly greater salary increases for a fictitious average meritorious engineer, but they were no different from their slower-climbing counterparts in their salary recommendations for engineers on other grounds. In another simulation, Martin (1987) showed that recommendations increased for both meritorious performance and specialized expertise needed in particular circumstances.
Time and events can modify the influence of the same behavior by leaders that was effective earlier, and subordinates’ reactions to it. Similar behavior by the leader can be ineffective (Sleeth & Showalter, 2000). This is a case of negative transfer of an old response to a new stimulus. A breakdown can occur. A leader can lose control of a situation by ignoring the changes caused by an event. For example, suppose that a leader is accustomed to loosely monitoring an operation, but a new employee who needs close supervision enters the scene. The same leadership is now inadequate. A breakdown can occur in the operation.
In managerial in-basket experiments by Martin and Harder (1988), students’ use of simulated financial-related related rewards, such as profit sharing, office space, and company cars, was contingent on satisfactory performance; but simulated interpersonal rewards, such as help or friendliness, were distributed according to subordinates’ needs.
Managers may lack the necessary reward power required to deliver recommendations for pay increases. Those who can fulfill the self-interested expectations of their subordinates gain and maintain the reputation of being able to deliver pay, promotions, and recognition. Those who fail to deliver lose that reputation in the eyes of their subordinates and, therefore, no longer can be seriously seen as effective contingent reward leaders (Tsui, 1982). Worse yet, they may be seen as having become more punitive, and their legitimacy and influence will be reduced (Greene & Podsakoff, 1979, 1981). Subordinates’ resistance will increase according to Tepper, Schriesheim, et al. (1998), who asked a sample of 389 respondents to recall an instance within the last month in which they resisted and avoided doing something their supervisor asked them to do. Thirty-two scale items were extracted from the responses. These items formed two factors: (1) passive resistance (“I procrastinate, hoping that my supervisor will forget about it”; “I say OK but don’t get to it”; “I don’t do it in a timely manner”) and (2) active resistance (“I explain that the request will not yield the expected results”; “I explain that it should be done differently”; “I present logical reasons for doing the task differently or at a different time”). In multiple regression analyses for 133 subordinates of 50 managers in a mental health rehabilitation organization, scores for passive resistance were higher if they worked for abusive supervisors and were less conscientious on the NEOAC Personality Inventory. Active resistance was higher if they worked for higher-status managers and were more open and less agreeable on the personality inventory.
Yukl (1981) enumerated the conditions that made it possible for managers to engage profitably in the contingent rewarding of their employees. As was already indicated, the managers needed substantial authority and discretion to administer tangible rewards to their subordinates. Their subordinates had to be dependent on them for access to the valued rewards. It was necessary for performance outcomes to be determined primarily by the subordinates’ efforts and skills, rather than by events beyond the subordinates’ control. It was also necessary to be able to measure the subordinates’ performance accurately. Finally, contingent reward was likely to have more impact if the work was “repetitive, boring, and tedious, rather than varied, interesting, and meaningful” (p. 141).
In the Soviet Union, workers were fined for failing to meet production quotas. Workers everywhere face suspensions for infractions of rules. A sufficient cause for dismissal may be simply displeasing a domineering boss. Feedback can be negative as well as positive. As indicated in earlier chapters, such contingent penalization, in general, is less effective than contingent reward. After observing 126 incidents of feedback, Balcazar, Hopkins, and Suarez (1985–1986) noted that feedback does not uniformly improve performance. In a survey by Ilgen, Fisher, and Taylor (1979), the role ambiguity of 527 Australian employees was reduced much more by positive rather than negative feedback about their performance. Although the effect of positive feedback is to reinforce behavior already displayed and hence to generate more of the same, for negative feedback to be effective, changes are required in performance as a consequence of changes in perception, motivation, or learning. Also, negative feedback may fail to work because of its unfavorable motivational impact on the subordinate’s self-image.6
Disciplinary actions may cause emotional reactions and a deterioration of supervisory-subordinate relations instead of their intended correctional effects, according to a survey of 177 firefighters in eight cities (Greer & Labig, 1987). The unintended results will occur if the administration of the actions is unpleasant, if poor relations already exist, and if the supervisor is judged to be inaccurate in his diagnosis and fails to represent adequate reasons for the disciplinary action.
Despite its potential to contribute to conflict and its relatively lower contributions to the improvement of subordinates, negative reinforcement is often used. There are many reasons for “accenting the negative,” despite the greater utility of emphasizing the positive and despite the decreased contribution of such leadership to the performance of subordinates. For instance, a supervisor may have a large span of control; as a consequence, the super-visor’s time may be fully occupied with monitoring deviations from the standard. Failure to pay attention to the deviations may invite disaster. A preoccupation with prospective failures will inhibit attention to the positive, particularly in the absence of clear goals, clear policies, and long-term objectives. Managers may lack the power to provide or recommend rewards. Faced with continuing demands for productivity, they will increase their tendency to use punishment, especially since managers report less difficulty identifying poor performance than good performance (Podsakoff, 1982).
Performance of subordinates that is below standard is more salient for a supervisor than performance that exceeds the standard (Larson, 1980). Thus subjects who were acting as supervisors were found, in a laboratory experiment by Fisher (1979), to require a smaller sample of work to evaluate a subordinate who was performing below average than to evaluate one who was performing above average. Larson (1980, p. 199) offered the following reasons for the greater sensitivity of supervisors to the failure rather than to the success of subordinates: “The criteria for minimally acceptable … performance are frequently more clearly defined than are the criteria for superior performance; poor performance often has a more significant impact upon the work group’s functioning than does superior performance.” Supervisors are likely to be more sensitive to their subordinates’ failure if how well the supervisors do their job depends on how well the sub-or dinates have completed assignments and if, as was noted earlier, the supervisors’ pay, recognition, or promotion depends on the subordinates’ performance (Larson, 1980).
Contingent reinforcement may fail because there is no clear differentiation of reinforcements for functionally adequate and functionally inadequate performance. Praise alone, without other accompanying rewards, may not work. Feedback may lack impact because goals have not been clearly set. In addition, to understand the effects of reinforcement leadership, according to Fedor, Buckley, and Eder (1990), the impact on a subordinate’s behavior of a supervisor’s attempts at differential reinforcement of the subordinate will depend on the subordinate’s perception of the supervisor’s intentions: “Does the supervisor want to dominate me, to focus my attention on unit standards, to support me, or to urge me to increase my productivity?” Many of the 220 subordinates in this study believed it was most likely that their supervisor wanted to dominate them rather than to support or assist them in achieving higher performance levels. This belief may be due, in part, to the tendency of subordinates to attribute their poor performance to situational causes like bad luck or difficulties with the problem, equipment, or supplies. As the observers of the subordinates’ behavior, the supervisors, especially if they are unfamiliar with the specific tasks involved, are more likely to attribute the poor performance to the subordinates’ lack of ability or a lack of effort. Mitchell and Kalb (1982) showed that if the supervisor had no experience with the subordinate’s job, he or she was less likely to accept situational reasons for the subordinate’s performance. Whether a supervisor likes or dislikes a subordinate also makes a difference. According to an experiment and field study by Dobbins and Russell (1986), although leaders may attribute the same reason for the poor performance of subordinates they like and subordinates they dislike, they will tend to act more punitively toward the subordinates they dislike and will be less inclined to punish the subordinates they like.
A Jackass Theory? Many scholars, consultants, and practitioners find contingent reinforcement a full and sufficient driving force for predicting, controlling, and understanding leader-follower relations. Many others are considerably less confident. Still others find that faults and failures override its successful applications. Levinson (1980) termed contingent reinforcement the “jackass theory” of leadership: When you, as a subordinate, are between a carrot and a stick, tempted by the carrot or threatened with the stick, it makes you feel like a jackass. Your self-esteem as a mature, conscientious, worthy individual is diminished. Furthermore, when you are between a carrot and a stick, you may see yourself in a contest that requires agility to reach the carrot or to avoid the stick. Or you may be immobilized by uncertainty, not knowing whether the benefits of the carrot are outweighed by the costs of being beaten by the stick. Feelings of being manipulated may outweigh responsiveness to the contingent rewards. Drawing from Dostoyevsky, Vice Admiral J. B. Stockdale (1981, p. 15) argued that people do not like to feel they are being programmed: “You cannot persuade [people] to act in their own self-interest all of the time. A good leader appreciates contrariness. … Some men all of the time and all men some of the time knowingly will do what is clearly to their disadvantage if only because they do not like to be suffocated by carrot-and-stick coercion. I will not be a piano key; I will not bow to the tyranny of reason.”
As Maccoby (1988) noted, with increasing technology and its servicing, employees are expected to work and think as mature persons. They are expected to make judgments; to solve problems; to develop good working relationships with colleagues, clients, and customers; and to be committed to and understand what they are doing. These expectations require a broader view of what motivates people to work. About half the sample studied by Maccoby regarded themselves as experts who got much of their reward from doing the work itself. Being helpful to others and maintaining good relationships are important to many employees. Although about a fifth of the sample were interested in material rewards, their self-development of a full career and personal life was more important to them.
Simple carrot-and-stick supervisory reinforcement clearly has considerable limitations. It will not work well in the absence of a trusting relationship between superiors and subordinates. Subordinates need to trust that their leaders can fulfill their promises. Leaders must be able to trust that their subordinates will comply as expected rather than take shortcuts or superficial actions to achieve promised rewards (Dwivedi, 1983). Many other forces constrain the superiors’ intentions when they use reinforcement to deal with their subordinates’ performance. These forces include information requirements, conflicting interests, biased perceptions, differential use of reward and punishment, and individual differences in reaction to contingent rewards.
Conflicts. Contingent reinforcement may fail to achieve the compliance of subordinates because intra-personal or interpersonal conflicts may be involved. For instance, the leader may establish dyadic contingent reinforcement with each member of a group, although the members need to reach decisions in a highly coordinated fashion. The members may suffer from pluralistic ignorance, each subordinate not knowing every other member’s views (Schanck, 1932). Or the members may find the costs greater than the promised benefits and, wishing to avoid confrontation with the leader, they may withdraw into anomie, apathy, or alienation. Followers may also rationalize, concluding that the leader’s demands on them are excessive or that the leader is incompetent. Ansbacher (1948) interviewed German prisoners of war during World War II, providing extensive evidence that although many continued to esteem Adolf Hitler and maintain confidence in him, they personally disagreed with him on many significant issues and interpreted his directions to suit their own needs. For example, Hitler demanded that Germany should never capitulate. But a majority of those who had confidence in him decided that he did not really want unnecessary bloodshed; therefore, they thought he would consider it his duty to end a lost war. And so, despite the fact that Hitler issued explicit orders that soldiers were to fight to the end, his supporters rationalized that it was all right to surrender.
Ephemeral or Negative Effects. Contingent reinforcement may succeed only in the short run, or sometimes not at all. In some circumstances, feedback about one’s job performance actually may be counterproductive and yield negative results, as Kiggundu (1983) reported for 138 employees at a head office. Greene’s (1979a) longitudinal study of new employees during their first nine months of work found that the subordinates’ performance, compliance, and satisfaction improved because of contingent rewards by the leader. Satisfaction accompanied the leader’s rewarding of subordinates for their good performance, and dissatisfaction emerged from the leader’s punishment of the subordinates for their poor performance. But the effects mainly occurred earlier in the subordinates’ employment. Contingent reinforcement may generate conflicting deleterious side effects that contribute to long-term failure. For example, subordinates may continue to accept the top managers’ direction but displace their aggression onto less powerful targets. Purcell (1953) found that packinghouse employees were more ready to blame their lower-status supervisors than their top managers for difficulties entailed in trying to comply with directions.
Biased Perceptions. Despite Peterson’s (1985b) conclusion that, in general, rewards for good performance tend to be effective, many errors are likely to be made in the application of contingent reward. Leaders’ performance evaluations are likely to be biased. Subordinates are likely to overestimate their own performance and that of their peers. Also, biased judgments may work both ways. Subordinates may inflate the attractiveness of an order to justify their compliance with it (Campbell, Dunnette, Lawler, & Weick, 1970). Furthermore, subordinates’ attributions of the reasons for the leader’s behavior will affect their satisfaction with it. And a leader whom subordinates judge to be willing but incompetent may be more forgivable than a leader who is judged to be competent but unwilling (Bass, 1982). The same leadership behavior—say, initiation of structure—has been seen more favorably by managers than by graduate students, although a more punitive emphasis on pushing for production was seen more unfavorably by both managers and students (Butterfield & Bartol, 1977).
Contingent Reinforcement May Be Avoided or Ignored. The use of contingent reinforcement is not essential for effective leadership or for a productive organization. Jacoby, Mazursky, Troutman, and Kuss (1984) found considerable avoidance of opportunities for obtaining feedback or ignoring it if it was given. In their simulation, they had 17 security analysts decide which common stocks represent a good purchase. During a sequence of four periods, the analysts were able to request feedback about their performance. The feedback provided results that could have predictive or explanatory value for the analysts. Some analysts did not request or consider any feedback, and other analysts were not consistent about doing so. The researchers found that the better-performing analysts were more likely than the poorer-performing analysts to ignore feedback that had no predictive or explanatory value.
Communes. When selected individuals and groups have a strong feeling of community, they have been able to maintain the high productivity that is associated with a strong work ethic without using reinforcement as the basis of their motivation to work. Communes depend even more strongly on trusting relationships than organizations where motivation to work is based on contingent reinforcement. In the Israeli kibbutz, “The underlying philosophy of the kibbutz insists that there must be no linkage between what a particular individual is able to contribute through his efforts and … what rewards he is entitled to. … In the kibbutz, everyone is entitled—as a human being and as a member of the community—to the comprehensive satisfaction of his needs … subject to limitations due to the economic capacity of the particular kibbutz. … This right is no way infringed if, because of health or for any other reason, he is unable to contribute to the kibbutz anything at all. In this respect the kibbutz constitutes a full-scale insurance system for every member. … [Although there may be some slackers] the large majority of kibbutz members generally perform at their best, and the prevailing differences in efficiency are the result of individual differences in ability, rhythm, [and] energy, … rather than of deficient motivation” (Tsur, 1983, pp. 24, 26). The young people who leave the kibbutz to work elsewhere tend to be highly praised by their employers because of the strong self-reinforcing work ethic of the kibbutz they continue to maintain.
Group Rather Than Individual Reward. There may be much more of this group effect in the U.S. industrial sector than one ordinarily supposes, which again points to the lesser importance of contingent reinforcement to the individual employee. For example, for the 71 managers and professionals of a manufacturing facility, Markham (1988) reported a correlation of –.03 between their rated performance and their merit raises within their own units, but .45 between the performance ratings of all the members of a unit and the merit raises they received. In this facility, group, not individual, performance was being rewarded.
The Augmentation Effect. Bass (1985a) theorized that contingent reward can be catalyzed by transformational leadership. Empirical support came from Waldman, Bass, and Einstein (1987) and Waldman, Bass, and Yammarino (1988) and indirectly from Schriesheim, Castro, Zhou, et al. (2006). The leadership of Ernest Shackleton, whose men overcame formidable obstacles in the Antarctic—or the leadership of Joan of Arc, Martin Luther King, Jr., Mahatma Gandhi, or Winston Churchill—can be perceived as leadership that added to any cost-benefit exchange a transformation of the followers’ needs from those at lower levels to higher-level concerns for achievement, glory, humanity, fortune, country, faith, or family, demanding excessive costs relative to tangible benefits. Self-interest was transcended; cost-benefit calculations were abandoned. It is difficult to conceive of the emotional response to the Ayatollah Khomeini in the rush to martyrdom during the Iranian-Iraqi war by the Iranian masses or the terrorist suicide bombers since then, merely as a social cost-benefit exchange between a leader and followers. On the positive side of reinforcement, appeals to self-interests alone will not result ultimately in leadership that is able to reward followers as much as they want. Continued attention to followers’ self-interests alone will not permit a group, organization, or society to operate optimally. A culture of cooperation is needed, as is trust in the benefit of optimal organizational outcomes. A prosocial charismatic leader who appeals to interests that transcend the individual member will be more likely to create such a culture than a leader who is limited to contingent reinforcement and the use of individual incentives (Miller, 1987). In the public sector, political leadership must appeal to both self-interests and shared values (Meier, 1988). In the early twenty-first century moral issues have become as important as bread-and-butter issues to American voters. The shared environment has become as powerful a political issue as self-interested economic gains. A 50-cent increase in the federal gasoline tax (from 18 cents) was supported by 70% of the American public as a response to automobile pollution and global warming.
Individual Differences. Contingent reinforcement may also be irrelevant if the subordinate attaches no importance or value to the reinforcement—if the subordinate is not interested in the rewards, promises of reward, or avoidance of punishment (Ilgen, Fisher, & Taylor, 1979). Subordinates’ indifference may be due to low expectations that satisfactory levels can be attained to merit the rewards or that the promises of reward will be kept (Larson, 1984).
Beyond the specialized circumstance of communal living or group reinforcement, one finds that individuals differ in their preference for external or self-reinforcement. In a survey of 339 utility company managers, Parsons, Herold, and Turlington (1981) found that the managers varied in their preference for external feedback—that is, in their agreement with the statement, “Even though I may think I have done a good job, I feel a lot more confident of it after someone else tells me so.” They also differed in their ability and tendency to reinforce themselves—that is, in agreeing that “If I have done something well, I know it without other people telling me so” and “As long as I think that I have done something well, I am not too concerned about how other people think I have done.” Bass (1967c) concluded that task-oriented subordinates and experienced subordinates are more likely to be self-reinforcing, whereas relations-oriented and self-oriented subordinates are more likely to be sensitive to both positive and negative reinforcement from others.
One is likely to see personally counterdependent followers working in opposition to what the leader intended with contingent rewards. Much of this reverse twist may also be seen in personally independent and even dependent followers. Among five styles of leadership—(1) directive, (2) manipulative, (3) consultative, (4) participative, and (5) delegative—displayed by large samples of managers, Bass, Valenzi, Farrow, and Solomon (1975) found manipulative leadership least satisfying to subordinates and least effective with them. But the managers’ superiors may not find fault with the manipulative potential of the managers in practicing contingent reinforcement. In fact, manipulative managers are more successful with higher-ups in that they earn salaries, after adjustment for their seniority, function, education, sex, and organization, that are higher than average (Farrow, Valenzi, & Bass, 1981).
Feedback is not always effective (Fedor, Buckley, & Eder, 1990). In fact, according to a meta-analysis by DiNisi & Kluger (2000), among 131 studies of feedback, 38% were ineffective in outcomes. Overall, the improvement in performance from feedback was less than half a standard deviation. Numerous factors constrain the use and impact of downward feedback, in which supervisory contingent reward is delivered. These factors include problems with the performance appraisal interviews, coping with unintentional consequences, and a reluctance to give feedback. For instance, among 508 air force personnel, lack of feedback was one of the highest concerns of leadership and management (Kunich & Lester, 1996). Even when given, it is not very helpful if it is insufficiently frequent, if it comes too late, if it is not based on evidence, if it is a one-way discourse instead of a dialogue, or if it is unclear or inaccurate. Rather, feedback should be timely, flexible, direct, specific, and descriptive (Hughes, Ginnet, & Curphy, 1993); and it should be about the subordinate’s performance, not the subordinate as a person: “Your report came in too late to be used,” not “You’re too slow” (Kluger & DiNisi, 1996). DiNisi and Kluger’s meta-analysis indicated that for improvement in performance, feedback must focus on the recipients’ performance, not on their self-concept. Furthermore, the recipient’s ego must not be threatened. Information on how to improve should be included; the performance of specific others should be excluded.
Contingent reinforcement may be constrained by the reluctance of some supervisors to give positive feedback. Komaki (1981) offered numerous reasons for this reluctance: time pressure, poor appraisal methods, doubts about the efficacy of positive reinforcement, lack of skill, and discomfort to the leader and the subordinate. Feldman (1986, p. 39) offered a number of reasons why some supervisors may avoid giving positive feedback: “Who among us has not found, much to our chagrin, that our praise, rewards, and favors often go unnoticed and un-thanked? … Consider the colleague who responds to subordinates’ efforts above and beyond the call of duty with no comment, as if those efforts were simply his due. Consider the supervisor who compliments a subordinate for a job well done, and receives in return a comment like ‘No thanks to you.’”
Subordinates may interpret feedback as a signal that the superior expects them to continue to perform at a high level. In such a case, the subordinates will regard the positive reinforcement as pressure to keep performing at the top of their capacity. Particularly when positive reinforcement is rare, subordinates may be ambivalent about receiving it for fear of becoming emotionally dependent on the good opinion of the supervisor, a good opinion that may easily be lost. Subordinates with low self-esteem will be jarred by the contrasting praise given to them and will evaluate it as only a momentary departure from their generally poor performance. Furthermore, the positive feedback that is given is often actually trivial and meaningless, or it may be interpreted as such. The supervisor who gives it frequently may be seen as politically motivated, manipulative, ingratiating, and insincere. Feldman (1986) agreed that positive feedback will be less favorably received when it is given frequently and for trivialities than when it is given only for unusually good performance.
Subordinates who are in conflict with their supervisors may react defensively to positive feedback and even interpret it as negative feedback or a precursor of negative feedback. “When receiving praise, some people may be waiting, like Pavlov’s dog, for the anticipated punishment to follow” (Feldman, 1986, p. 40).
Even if positive feedback is given, it may not provide the expected reinforcement. Subordinates may not believe that positive feedback from a supervisor is a consequence of their performance. Rather, they may perceive that the supervisor praises when he or she feels like doing so, not when the performance is good. Subordinates may also believe that positive reinforcement is too small or inequitable, which makes the subordinate angry because the subordinate believes that more of a reward should have been given for the performance. Finally, subordinates may feel a need to reciprocate and give unjustified, unwarranted positive feedback to the supervisor, and they will resent feeling pressured to do so.
Gaddis, Connelly, and Mumford (2004) set up an experiment to study the effects of a leader’s feedback that a subordinate had failed. Negative affect of the leader accompanying the feedback (as compared with positive affect) reduced the leader’s perceived effectiveness and lowered the quality of performance.
Supervisory negative feedback may be delayed, distorted, or avoided following poor subordinate performance whether or not the poor performance is attributed to lack of effort or failure of mastery of what needs to be done (Moss & Martinko, 1998).
Attitudes about Discipline, Reprimands, and Punishment. Although some supervisors and subordinates may find even positive feedback discomforting, many more are reluctant to give and receive negative feedback (Larson, 1986; Sims & Manz, 1984). An experiment by C. D. Fisher (1979) with 168 college students indicated that “supervisors” gave feedback sooner when the “subordinates’” performance was poor than when it was good; but Fisher believed that the results would be reversed in any long-term field study, since supervisors in real life are more likely to be reluctant to give negative feedback. It is one thing to say something negative to a stranger in a transient experience than to say it to someone with whom one will remain in close contact. Managers often are reluctant to discipline an employee because it is an unpleasant task for most of them. Discipline and friendship are difficult to maintain side by side (Harrison, 1982). Nevertheless, numerous factors contribute to supervisors’ willingness to use punishment. These factors include positive attitudes about punishment (O’Reilly & Weitz, 1980), supervisors’ independence of their subordinates’ performance (Ilgen, Mitchell, & Fredrickson, 1981), subordinates’ poor performance relative to coworkers’ (Ivancevich, 1983), and supervisors’ lack of reward power (Greene & Podsakoff, 1979/1981).
Salience of Poor Performance. It may take a considerable worsening of a subordinate’s performance for a supervisor to overcome his or her reluctance to give negative feedback (Larson, 1986) although, as was mentioned earlier, subordinates’ poor performance is more salient to supervisors than their subordinates’ good performance. The salience of the poor performance must outweigh how much the supervisor personally admires a subordinate. A supervisor may avoid giving negative feedback for poor performance to avoid risking a deterioration in personal relationships with subordinates. The supervisor may also be less likely to attribute poor performance to subordinates if relations with the subordinates are good (Bass, 1985a).
Use of Distortion. When supervisors are faced with poor performance that they attribute to the subordinates’ lack of ability, they tend to pull their punches. They distort their feedback and make it more positive than it should be (Fisher, 1979; Ilgen & Knowlton, 1980). Programmatic efforts to change the supervisor’s feedback behavior were not successful (Frank & Hackman, 1975; French, Kay, & Meyer, 1966). Some of this distortion may be due to the effort of the supervisor to act prosocially, to be more concerned with the subordinate’s feelings than with the supervisor’s or the organization’s best interests. To avoid causing distress to the subordinate, the supervisor may remain lenient and avoid giving needed negative feedback.
The reluctance to give negative feedback is abetted by supervisors’ tendency to search for excuses and justifications for the poor performance of subordinates. Gioia and Sims (1986) arranged for 24 experienced managers to handle a simulated performance appraisal interview. During the interviews, the managers tended to probe for excuses to justify the subordinates’ performance. After the interviews, the managers reduced their blame of the poorly performing subordinates.
Other Reasons. Various other reasons may also make the supervisor reluctant to use negative feedback. The supervisor may be highly dependent on the subordinate to attain their mutual objectives (Tjosvold, 1985) or may have limited authority to bring about a punishment (Beyer & Trice, 1984). Since negative reinforcement may cause the employee to quit (Parsons, Herold, & Leather-wood, 1985), the supervisor’s reluctance may be based on that concern. Political considerations may deter a supervisor from giving negative feedback for fear of the employee’s retaliation. An admonished employee may file a grievance for unfair discrimination or may complain to powerful friends in the organization. The supervisor’s ability and motivation to correct a problem employee are also decreased if the performance of most of the work-force is poor (Crawford, Thomas, & Fink, 1980).
Nevertheless, despite their reluctance, managers will become punitive with continuing poor performers and problem employees after asking questions and trying to solve the problems they believe are causing the poor performance (Fairhurst, Green, & Snavely, 1984). In the case of problem drinkers, for example, discipline by supervisors is widespread when the work context supports its use (Beyer & Trice, 1984). O’Reilly and Weitz (1980) found that supervisors who used informal warnings, formal warnings, and discharges led groups that performed better; but Kipnis (1976) concluded, from research evidence, that supervisors were most likely to use such sanctions when they did not expect to be successful in using their influence to improve their subordinates’ performance.
Improving the Positive Effects of Negative Feedback. Salespeople said they worked harder if they had a performance orientation but said they worked harder and smarter if they had a learning orientation (Sujan, Weitz, & Kumar, 1994). A review by Ilgen and Davis (2000) concluded that feedback was more likely to be effective if the recipients believed they had control and influence regarding the performance in question, if they were oriented toward a learning rather than a performance goal, and if they were self-efficacious. If they had a learning goal, failure was legitimate. If they had a performance goal, negative feedback, if not too harsh, could generate specific alternative ways of doing the task more effectively. “Mastery goals facilitate positive … strategies in the face of performance decrements” (Ilgen & Davis, 2000, p. 560).
Fedor, Buckley, and Eder (1990) found that subordinates perceived four possible intentions for supervisors to give them feedback: (1) to build relationships, (2) to provide self-evaluation, (3) to facilitate productivity, and (4) to serve the interests of the supervisor. The subordinates’ perceptions of a supervisor’s intentions may have an effect on their interpretation and acceptance of feedback. By expressing dissatisfaction with the process, subordinates, the intended receivers of feedback, may also add to the reluctance of supervisors to give it (Greller, 1978). They may discourage the supervisor from efforts to provide it by frequent rejections of the supervisor, displays of hostility, and displacement of the blame onto others. Many subordinates are likely to take personal responsibility for their successes but to attribute failures to external or situational causes, such as the difficulty of a task, bad luck, or a lack of cooperation and assistance from others (Greenwald, 1980; Sims & Gioia, 1984).
Interpretation and Acceptance. The contingency between performance and outcome may be more apparent to the subordinates when they are successful. Subordinates may expect success and tend to validate their expectations by reinterpreting as an actual success a failure that their supervisor sees. They may base their judgments of their performance on the desirability of the outcome to them personally, rather than on the explicit link between performance and outcome (Miller & Ross, 1975).
Negative feedback may be interpreted as helpful and therefore may be regarded as a contingent reward, rather than as a contingent punishment. It all depends on how the feedback is presented, how much it is embedded in sedating and palliating language, and how open the subordinates are to it. The subordinates may fail to “hear” negative feedback because they convert it into what they want to hear. Subordinates accept positive feedback more readily and recall negative feedback less accurately (Ilgen, Peterson, Martin, & Boeschen, 1981). Considerable discrepancy is found between the extent to which supervisors say they give feedback to their subordinates and the extent to which the subordinates say they receive such feedback. For example, in a survey of 178 managers and their subordinates by Mann and Dent (1954b), 82% of the managers said they gave “pats on the back” to their subordinates often, but only 13% of their subordinates agreed with them. Eighty percent of the managers said they gave subordinates sincere and thorough praise often, but only 14 percent of their subordinates agreed with them.
Context. Subordinates may ignore supervisors’ feedback because it is not backed up with evidence, because it is not consistent with feedback they have received from coworkers, because they judge it to be unimportant, or because they regard self-reinforcement as more relevant. Different working situations and levels of experience of the recipients of the feedback will determine which source of feedback the subordinates think is most useful and important (DeNisi, Randolph, & Blencoe, 1983). Greller and Herold (1975) obtained results indicating that a heterogeneous sample of working people attached more importance to feedback from a higher authority than from their peers; but Pavett (1983) noted that in a hospital, feedback from patients and coworkers had a greater impact than feedback from supervisors on nurses’ perceptions of the connection between delivering high-quality care and valued outcomes. In a different work setting, Ivancevich and McMahon (1982) reported that self-generated feedback about the attainment of goals did more than feedback from supervisors to improve the subordinates’ performance, intrinsic job satisfaction, and commitment to the organization. And according to Greller (1980), who surveyed 26 supervisors and 63 of their subordinates in the maintenance department of a metropolitan transit agency, the subordinates attached more importance than the supervisors to the feedback from the task itself, to their own comparisons with the work of others, and to coworkers’ comments about their work.
Insurance salespeople have fewer contacts with each other than do metropolitan transit employees, for whom feedback from coworkers is important. Brief and Hollen-beck (1985) found that among the 62 insurance salespeople they interviewed, the greatest amount of feedback on performance came from the supervisor and themselves, not from coworkers. Despite the general observation that we attribute our successes to ourselves and our failures to others and to external conditions, most of the contingent self-reinforcement of the insurance employees was negative, in the form of self-criticism; and it had detrimental effects on their performance.
Views about the Supervisor. Subordinates may rationalize feedback as being due to a supervisor’s need to complain, the supervisor’s failure to understand the situation, or the supervisor’s boss’s malevolence. According to Coy (1982), who conducted a laboratory experiment with undergraduates, “subordinates” were readier to accept feedback from a leader they perceived as competent rather than incompetent, particularly if the leader had a strong need for achievement. If they perceived the leader as incompetent, they would seek alternative sources of task-relevant information.
Self-Esteem. Subordinates’ self-esteem will moderate their request for feedback and their acceptance of it. Those with high self-esteem set higher goals, perform at a higher level, and experience more positive affect when performing well than subordinates with low self-esteem (Taylor & Slania, 1981). Those with high self-esteem seek less information from others in making decisions and forming judgments. They are more self-reinforcing. Those with low self-esteem tend to be more dependent on social and environmental feedback (Weiss, 1977) and are more responsive to it. Inexperienced personnel prefer to receive feedback from their superiors, but experienced personnel prefer to be more self-reinforcing (Hillery & Wexley, 1974). However, in contrast to reinforcement only from the supervisor, a combination of both self-feedback and supervisory feedback, in which the subordinate first prepares a self-appraisal and then meets to compare it with the superior’s feedback, appears to reduce the subordinate’s defensiveness and to bring about greater subsequent improvement in the subordinate’s performance (Bassett & Meyer, 1968). Whether the subordinate seeks feedback will also be affected by his or her role ambiguity and tolerance of ambiguity (Ashford & Cummings, 1981). Whether a subordinate will improve in performance after receiving negative feedback from the supervisor will depend on the subordinate’s self-esteem as well as the supervisor’s sources of power (Fedor, Davis, Maslyn, et al., 2001).
Problems inherent in the feedback itself may cause it to fail to work as intended. In place of the summary ratings many supervisors use, Deets and Tyler (1986) introduced detailed feedback about the specific initiation and consideration behaviors7 of supervisors at Xerox, along with interim evaluations of the supervisors’ progress toward specific personal and professional goals that were given closer to the events involved. They found a marked improvement in attitudes toward the appraisal system and toward teamwork.
The accuracy, amount, distinctiveness, consistency, and method of giving feedback (oral, written, or face-to-face) may also matter (Duncan & Bruwelheide, 1985–1986). But in a study of 360 pairs of managers and subordinates from more than 50 organizations in three countries, Larson, Glynn, Fleenor, and Scontrino (1986) found it impossible to single out the specificity of feedback from other positive features of good-quality feedback, such as its timeliness, frequency, and sensitivity. However, Quaglieri and Carnazza (1985) found that most subordinates paid attention to multiple aspects of feedback, including its accuracy, timeliness, attractiveness, and specificity, as well as to the trustworthiness, ex-pertise, and power of the source. Furthermore, these and other features of feedback could be experimentally manipulated. Earley (1986a) completed a field experiment in which 60 subjects believed they were in a working situation processing subscriptions. Positive feedback increased their expectations of successful performance more than did negative feedback, as expected. The subjects’ performance was better than anticipated if the feedback was specific rather than general. In turn, specific feedback resulted in more planning by the subordinates, which contributed to the subordinates’ better performance. However, the supervisors had less of an impact on the subordinates’ performance than self-generated feedback retrieved from computer storage, particularly if the positive or negative feedback was general. Only when the supervisors’ feedback was specific and negative did its effect match that of self-generated feedback. Mediating these effects was the extent to which the subordinates trusted the computer feedback more than they trusted the supervisors’ feedback.
The credibility of the source of feedback is important. Feedback will be more effective if the leader is respected and trusted, if the subordinate regards the feedback as useful, and if it is delivered with consideration, tact, and support Another important consideration is whether feedback from one source is discordant with feedback from other sources (Liden & Mitchell, 1985). When one colleague finds fault with a member’s behavior, the impact is much less if other colleagues do not agree.
Coping with Unintended Consequences. Contingent negative feedback may have unintended consequences. Reprimands may serve not only to increase clarity about what is undesirable behavior and to correct it (Reitz, 1971), but they can also generate unproductive anxiety. In turn, this anxiety can result in a variety of dysfunctional behaviors, such as reaction formation, guilt, and hostility, particularly in highly motivated subordinates who are already overloaded or under stress and who may interpret well-intentioned negative feedback as a personal attack (Bass, 1985a).
Komaki (1982) and Sims and Gioia (1984) suggested ways to cope with the unintended consequences and potential shortcomings of feedback processes and make better use of them. According to 60 professionals interviewed by Komaki, to increase the effective utilization of feedback the long-term benefits of reinforcement need to be emphasized. More positive feedback should be used, but it should not be limited just to delivering praise; it should include connections to the performance involved. Reinforcement schedules have to be adapted to fit busy schedules. Standards of performance and of performance appraisals need to be clarified.
To reduce self-serving bias among subordinates and to increase the subordinates’ acceptance of negative feedback, Sims and Gioia proposed that the supervisor needs to recognize and expect the biased reaction to negative feedback for poor performance. The subordinate’s acceptance of negative feedback should be positively reinforced, and efforts to cover up mistakes should be rejected. Emphasis should be placed on determining the causes of and remedies for the poor performance.
Contingent reinforcement may occur at the time the supervisor observes the subordinate’s good or poor performance or in a performance appraisal interview. An annual appraisal review seems to be the norm in the United States. However, more frequent appraisal interviews are suggested for poor performers (Kay, Meyer, & French, 1965) and for subordinates in nonroutine jobs in which goals are set, progress is evaluated, and then goals are reset (Cummings & Schwab, 1973, 1978; McConkie, 1979). Numerous problems can impair the efficacy of the feedback in this interview. Reinforcement theory suggests that the feedback will be most reinforcing if it is given as soon as possible after the occurrence of the subordinate’s good or poor performance. Yet the delay caused by waiting to provide the feedback in a formal interview increases the reliability of the evidence to be discussed by the supervisor and the subordinate. Such a discussion can also take into account broader systematic issues and longer-range goals and plans than could feedback given immediately after an incident. Since the supervisor and subordinate can both be better organized to give and receive the feedback, a more orderly, calm, and thoughtful exchange is possible than in the heat of the immediate operations. Of course, nothing precludes both timely feedback and feedback for an extensive period given during an appraisal interview in which the subordinates’ performance during the six months or year preceding the interview is reviewed. In this formal meeting of the superior and the subordinate, the subordinate’s performance during a preceding period is reviewed, evaluations are made, and suggestions for improvement are advanced. The meeting can involve the superior’s and subordinate’s sharing of information or it can be directed by the supervisor. The interview can be structured in a variety of ways around goals and their attainment or around traits and behavior.
Although it is an organized effort to provide positive reinforcement for good performance and negative feedback for poor performance, the performance appraisal interview may fail to improve the subordinate’s subsequent performance, for numerous reasons. First, the formal feedback process tends to be constrained in use and impact by, for example, the supervisor’s lack of familiarity with the subordinate’s job. Second, the superior may not encourage the subordinate’s participation and two-way communication in the interview (Cederblom, 1982). Other reasons for ineffective feedback during a performance appraisal include a lack of supervisory support, a lack of goals, an unfocused discussion, and a failure to separate administrative and developmental feedback (Ivancevich, 1982). The performance appraisal interview may also fail to have an impact if no reward is linked to it (Burke, Weitzel, & Weir, 1978).
The lack of connection between performance appraisals and performance may be due to a variety of other goals that managers have in mind when they evaluate their subordinates (Longnecker, Sims, & Gioia, 1987). Supervisors may avoid discouraging subordinates or confronting them with their inadequacies, may seek favors from the subordinates, or may discourage a top-performing employee from demanding a commensurate increase in salary. Thus appraisals may be unrelated to performance even in the face of available objective evidence about the performance. For example, Leana (1984) discovered that the periodic evaluations of 98 claims adjustors by their 44 supervisors were not correlated with such objective measures of performance as the settlement ratio and the average cost of the claims that were settled.
Lawler, Mohrman, and Resnick (1984) conducted a survey of over 300 supervisor-subordinate dyads regarding the extent to which the performance appraisal interviews fulfilled their purposes. They found that although 82% of the supervisors said that all important matters had been discussed, most subordinates thought there had been less than a full discussion of their development, work, planning, or pay. Furthermore, they thought that not enough attention was being paid to documenting their performance. On the other hand, supervisors in general were unenthusiastic about discussing decisions about pay or documenting subordinates’ performance in these interviews. Presumably, both these matters would be extremely important for leaders to cover as means of contingent reinforcement (Mohrman & Lawler, 1983). Consistent with these findings was the survey by Greller (1980), which showed that supervisors actually overestimated the importance of the feedback they provided to subordinates and underestimated the value their subordinates place on feedback from sources under their own control such as the task and comparisons with the work of colleagues.
Many supervisors avoid conducting appraisal interviews. When forced by organizational policy, they often hold perfunctory general discussions with subordinates, which the subordinates do not consider to be the formal feedback of a performance appraisal interview (Meyer, Kay, & French, 1965). The failure to conduct high-quality performance appraisal interviews with optimal frequency may militate against these interviews’ serving to provide effective contingent reinforcement. Most employees seek more such feedback opportunities than they receive (Ash-ford & Cummings, 1983).
Frequent appraisals, however, particularly negative appraisals, may generate more resistance and defensiveness in subordinates (Burke, Weitzel, & Weir, 1978). If the feedback is not what the subordinate expected, it may arouse resistance (Bernstein & Lecomte, 1979), or the subordinate may consider it less accurate (Taylor, 1981). If the feedback is not as positive as expected, it may reduce the subordinate’s commitment to the organization (Pearce & Porter, 1986). Decreased commitment is likely to be common, since subordinates ordinarily give themselves higher evaluations than their superiors do (Bradley, 1978). Furthermore, subordinates’ reactions to the performance appraisal interview will depend on their opinions about the supervisor’s credibility, knowledge, and trustworthiness, which parallel their reactions to informal feedback from their supervisor, as was mentioned earlier. Ilgen, Fisher, and Taylor (1979) noted that subordinates’ judgments of their supervisors’ credibility as an appraiser will influence how the subordinates react in performance appraisal interviews. If the subordinates think their supervisor is knowledgeable about their job and performance, they will regard the interview as fair and accurate, but those who judge their supervisor to be less knowledgeable will think the interview is unfair and inaccurate (Landy, Barnes, & Murphy, 1978). And, as Ilgen, Peterson, Martin, and Boeschen (1981) observed, subordinates who trust their supervisor will be more satisfied with the appraisal interview process than subordinates who do not trust their supervisor.
Herold and Greller (1977) added a number of other characteristics of feedback that recipients can discern in the performance appraisal interview, such as whether the feedback is helpful or unhelpful, clear or ambiguous, important or unimportant, and formal or informal. The utility of the feedback appeared to be most strongly related to its frequency, helpfulness, and importance. In the interviews, ratings and discussions that focus on the past performance of the subordinates appear to be the most satisfactory to both superiors and subordinates (Fletcher & Williams, 1976; Mount, 1984). The subordinates’ participation in the development of the rating forms also seems to help (Landy & Farr, 1980). Agreement on the meaning of scores and terms, such as “good” and “satisfactory,” will help as well (Taylor, Fisher, & Ilgen, 1984). Supervisors who begin the interviews with positive feedback before giving negative feedback are seen as more accurate by the subordinates than supervisors who begin with the negatives before presenting the positive aspects of the subordinates’ performance (Stone, Gueutal, & MacIntosh, 1984).
Fox (1987–1988) enumerated guidelines for an improved performance appraisal system. First, credit should be given for good form. Second, emphasis should be on work-relevant behaviors that are known to be desirable and productive, rather than on outcomes that are often beyond the employee’s total control. These behaviors should be determined from critical-incidents surveys of the occurrence of highly effective and highly ineffective job events. Third, the appraisals should be from multiple sources, not just from the boss. Fourth, ratings should be accompanied by evidence. According to Fox, improvement in an employee’s performance would stem from an ongoing defining of behaviors that need to be changed, setting goals for change, and administering feedback and other reinforcements on appropriate schedules.
Following a comprehensive review of the literature, Biddle and Fisher (1987) described an ideal procedure for promoting such contingent reinforcement from performance appraisal interviews. First, there would be a preperformance meeting in which the supervisor and subordinate reached an understanding of exactly how the subordinate’s future performance was to be assessed. If rating scales were to be used, the meaning of each dimension and anchor point would be clarified. A performance monitoring system would be set up to ensure that good data were available for the appraisal. The sources of feedback would be considered, and the supervisor would arrange to collect data, as needed, from these sources, such as peers and clients, as well as to gather his or her own data. The subordinate would also keep track of his or her performance and accomplishments through self-monitoring. Between interviews, the supervisor would give more frequent informal feedback and coaching, so that the formal performance appraisal interview would not bring up much that was unexpected for the subordinate. The timeliness of the frequent informal feedback would allow the subordinate to learn from and correct his or her errors immediately. It also would afford opportunities to clear up differences in attributions for the same specific incidents of performance. In sum, the actual performance appraisal interview would begin with a self-appraisal, followed by the supervisor’s sharing of his or her feedback with the subordinate as well as feedback collected from the other sources. Discussion should then center on how the subordinate might be able to do things differently and what additional support the superior might provide. Finally, new goals for the upcoming period would be discussed and established.
The perceived and actual descriptions of the quality of the exchange, the causes of the exchange, and the consequences of the exchange may be obscured or magnified by the implicit theories of leadership that are held by the leaders and subordinates who are involved in the relationship.
Implicit Theories of Leadership as Cognitive Frameworks. The implicit theories we hold about leadership and its antecedents and consequences join with our explicitly expressed attitudes, beliefs, values, and ideologies about leadership8 to color our judgments of specific leaders we describe or evaluate. Implicit theories are cognitive frameworks or categorization systems that are in use during information processing to encode, process, and recall specific events and behavior. An implicit theory can also be conceived of as the personalized factor structure we use for information processing.
Both the leader and the led are affected in their exchange relationship by the implicit theories of leadership they carry around in their heads. If they both believe that leadership is mainly a matter of striking a deal for the payment for services rendered and monitoring the arrangements, then they are likely to judge that contingent reinforcement explains the motivation and performance of the leader and the subordinate. But if they believe that leadership is mainly an inspirational process, then they will judge contingent reinforcement as being of little or no consequence. Thus Eden and Leviatan (1975) noted that people have preconceptions about what constitutes appropriate behavior by leaders. Subordinates have expectations regarding the kind and amount of leadership behavior that is proper for given situations (L. R. Anderson, 1966b), as do the leaders themselves (Yukl, 1971). These preconceptions are important, according to Pfeffer and Salancik (1975), who obtained results showing that the leadership behavior of focal supervisors was affected by the expectations of both their superiors and their subordinates. In addition, subordinates attach their own individual specific value-laden meanings to the actions of their superiors (McConkie, 1984), whereas leaders may regard their performance as leaders to be a matter of their traits, their power, or the way they act.
Birnbaum (1987c) collected the definitions of leadership provided in interviews with 32 college and university presidents. Almost all of the presidents thought that leadership involved power, influence, and ways of behaving. About a quarter talked as well about traits of consequence, situational determinants, or the symbolism of leadership. Content analyses of the interviews suggested that about a third of the presidents believed that the goals of leadership stemmed from the mission of the institution; a few thought that the goals came from followers; and the rest thought that the goals came from the leaders themselves. Presidents saw themselves as either: (1) the heads of a bureaucracy; (2) collegial in their relations with the staff, faculty, and students; (3) political manipulators; or (4) occupants of a highly symbolic office. Some of the presidents had a single frame of reference about leadership, and about a quarter had multiple frames of reference (Bensimon, 1987). According to Bolman and Deal (1984), those with multiple frames of reference were better equipped to deal with the complexities of the modern college community.
Effects on Raters of Leadership. Implicit theories of leadership facilitate the assimilation of specific events and behaviors into a collective interpretation. They help raters to categorize their stimulus environment into less complex classifications. Implicit theories make it possible to simplify complex information by enabling raters to process that information automatically, as well as to add missing details not observed in the actual behavior of the leader (Lord, Foti, & DeVader, 1984). The raters’ application of their own implicit theory to observed events affects the truth of their observations. The result is that their responses to the survey questions are closer to their own implicit theories than to the actual events they observed. The impact of these implicit theories is greatest when raters are asked to recall the behavior and events they observed sometime in the past.
Lord, Binning, Rush, and Thomas (1978) estimated that up to 40% of the variance in ratings of leadership can be accounted for by implicit theories. Thus instead of describing the actual behavior of the individual leader who is being judged, we respond in a way that is highly biased by our preconceptions about how leaders in general are supposed to behave. Support for this proposition first emerged when Eden and Leviatan (1975) applied D. J. Schneider’s (1973) Implicit Personality theory to the study of leadership. Staw and Ross (1980) and McElroy (1982) provided additional support for the idea that the implicit conceptions we use in attributing what we observe to leadership seriously alter our attributions. Thus an exchange may take place between a supervisor and a subordinate. The subordinate is actually complying with the supervisor’s request to gain a recommendation for promotion from the supervisor. But the subordinate has an implicit theory that says leaders are inspiring. The subordinate reports she is complying as a consequence of the way the supervisor inspired her to work with him. Thus an actual transactional exchange is reported as a transformational event. Or the reverse may occur. A subordinate is actually inspired to transcend her own interests. However, she holds an implicit theory that you get what you give, so she interprets her appraisal by the supervisor as compliance because she wants to be recommended by him for a promotion.
Effects on Leaders’ Behavior. Supervisors are also clearly affected by the implicit theories of leadership they hold. Neubauer (1982) found, in a sample of 90 German supervisors, that implicit leadership theories are characterized by a belief in participative management and acceptance of informal leader-member relations for workers who live up to the supervisors’ expectations. But for workers who fail to live up to expectations, not very helpful implicit theories emphasize the limiting of such workers to simple activities, the need for supervisory control, and the prevention of the development of informal leader-member relations.
Implicit theories of leadership can be manipulated in advance to get raters to assess the same observed leadership behavior differently. If we are told that a leader is effective and believe the leader to be effective in carrying out his or her assignment, this will influence how much and what type of leadership we judge to exist in the subsequent performance we observe. In a laboratory study by Gioia and Sims (1985), subjects viewed videotapes of contingently reinforcing leadership behavior by managers in action. When managers were presented to the subjects as effective leaders before the subjects saw the tapes, the subjects described the managers subsequently as being significantly higher in initiating structure than the managers who were presented to the subjects as ineffective before the same tapes were viewed. Similar results were reported by Rush, Thomas, and Lord (1977).
Other conditions can be arranged to increase the effect of implicit theories of leadership. Thus the more ambiguous a leader’s actual behavior, the more implicit theories will affect how we describe or evaluate it (DeNisi & Pritchard, 1978). Also, implicit theories of leadership will have more of an impact in less structured situations (Gioia & Sims, 1985). Halo effects can also be connected with implicit theories (Nathan & Alexander, 1985). In addition, subordinates are likely to fall back on their implicit theories of leadership when they lack information about a situation (Schriesheim & DeNisi, 1978).
Implicit theories of leadership need to take into account the causal expectations that leaders have about themselves and their followers, as well as the complementary theories held by the followers. The extent to which leaders will modify their reactions to the good and poor performance of subordinates, depending on whether they attribute the performance to the subordinates’ competence or motivation, to external causes, or to luck, has already been addressed. The leaders are also likely to respond according to their own beliefs about the value of promises of reward for compliance or threats of punishment for noncompliance. As a consequence, one supervisor may praise a subordinate for an effort to reach desired standards that failed because of uncontrollable obstacles, and another supervisor may reprove the same subordinate for the failure despite the obstacles.
Larson (1980) suggested that supervisors vary in the implicit theories they have about feedback in general, as well as their theories about the consequences of feedback to specific subordinates. Some supervisors may think that reprimands invite retaliation and that praise lacks credibility. Disapproval is more likely if rationality, objectivity, and certainty are regarded as more important than adaptability; if security is valued as more important than affiliation; and if the leaders see themselves as aiming for homogeneity, regularity, standardization, safety, or consolidation (Quinn & Hall, 1983). Subordinates will use various tactics to try to influence their leaders, according to the subordinates’ beliefs about the likely effects on the leader of reason, ingratiation, excuses, and coalitions, as well as the implications for the subordinate’s self-image involving his or her integrity and self-esteem. The organizational role the subordinate believes he or she expected to play is also important.
Phillips and Lord (1981) concluded that implicit theories of leadership could best be understood in terms of cognitive categorization processes. (“In my head, Joe comes across as a dynamic leader. Dynamic leaders are a category containing ‘such and such attributes.’ Therefore, Joe has these attributes.”) The researchers demonstrated this experimentally by using 128 undergraduates to view one or another of two videotapes of a four-person problem-solving group in which the leader’s salience and the group’s performance were manipulated. Furthermore, Lord, Foti, and Phillips (1982) argued that implicit theories reflect the cognitive categories used to distinguish leaders from nonleaders. Such categories can be applied to explaining which information is connected to designated labels of leadership (Cronshaw & Lord, 1987; Lord, Foti, & Phillips, 1982).
For most people from the same culture, a common set of categories fits the image of what the typical leader is like. These categories describe the “prototypical” leader. A prototypical leader is actively participating; an “anti-prototypical” leader remains withdrawn. Perceiving someone as a leader involves a relatively simple categorization of the stimulus person as a leader or a nonleader. Such a categorical judgment, according to Rosch (1975), is made on the basis of the similarity between the stimulus person and the “prototype” of the category. The prototype (in this definition) is an abstract representation of the most representative features of members of the category.
Foti, Fraser, and Lord (1982) asked students to rate the extent to which phrases like “bright” or “sides with the average citizen” (taken from a Gallup poll) fit their image of a leader, a political leader, an effective leader, or an effective political leader. The students’ different prototypical categorizations helped considerably to account for changes in Gallup poll results of public support for President Jimmy Carter at different times during his administration. If Jimmy Carter acted more like the students’ prototype of an effective political leader, he received more support from the public. The mere social desirability of the categorizations was not as predictive.
People simplify rather than describe or evaluate their leaders in terms of many specifications. People form mental images of categories of leaders. As was noted before, they rapidly judge whether their leader falls into these categories according to their beliefs about the leader and how they see the leader acting (Phillips & Lord, 1981, 1982). Lord, Foti, and DeVader (1984) conducted three studies that used 263 undergraduates to specify the internal structure of these leadership categories and how they relate to prototypicality. These studies showed how properties of the categories can be used to facilitate information processing, such as recalling information about a leader, and to explain simplified perceptions about leadership.
In the first study, different students were each given five minutes to generate attributes of a particular type of leader or nonleader in 11 different situations: (1) business, (2) education, (3) finance, (4) labor, (5) politics, (6) the mass media, (7) the military, (8) minorities, (9) religion, (10) sports, and (11) at the world level. Master lists of attributes of the frequency of mentions of specific clusters of traits were then assembled. Intelligence was seen as highest in “family resemblance,” that is, attributed generally to leaders in almost all of the 11 situations. Other categories with relatively high family resemblance for leaders included honesty, outgoing, understanding, verbal skills, aggressiveness, determined, and industrious. Most of the other 59 attributes, such as caring, authoritarian, and decisive, were limited to just one or two types of leaders.
In the second study, a different set of students judged how much these attributes were “prototypical” for leadership. The students rated the categories according to how well they “fit my image of a leader or a nonleader.” Attributes “fitting the image of leader” included intelligence, honesty, verbal skills, determined, informed, strong character, believable, concerned, goal-oriented, and disciplined. The family resemblance of these attributes correlated .40 with the attributes of the prototypical leader. A correlation of .42 was found between the speed of reaction in judging these attributes and their fitting the image of the prototypical leader. Such attributes of prototypicality were thus linked to implicit theories as reactions “off the top of the head.”
In the third study, Lord, Foti, and DeVader presented vignettes about a district store manager. One vignette ascribed to the manager four or five attributes of the prototypical leader, such as “provides information” and “talks frequently.” Another vignette ascribed to the manager neutral attributes, such as “seeks information” or “explains actions.” Still another vignette ascribed to the manager “antiprototypical” attributes, such as “admits mistakes” and “withholds rewards.” Those who read the vignette containing the prototypical, rather than the neutral or antiprototypical, attributes perceived that the manager displayed more leadership and made much more of a contribution to the store’s effectiveness and to the successful merchandising of a new product. The investigators inferred that attributes of causes and responsibility were a retrospective rationalization of events in which “leadership” was the central construct stimulated by the few prototypical attributes embedded in the vignette. However, Lord and Alliger (1985) did not find that prototypicality weightings of the 12 functional leadership behaviors of members of small task groups had much impact on the frequent correlation between the sheer frequency of activity of group members and the leadership ratings they received.
Prototypicality and Contingent Reinforcement. Bass and Avolio (1988) asked 87 part-time MBA students to describe their full-time supervisors’ transactional and transformational leadership behavior with the Multifactor Leadership Questionnaire. They also had the students complete the prototypicality ratings developed by Lord and his associates about the attributes that “fit my image of a leader.” They found that contingent reinforcement was less highly correlated with the prototypicality of the leader than was charismatic leadership, individualized consideration, or intellectual stimulation. When the prototypicality ratings were statistically controlled by partial correlation, only charismatic leadership remained correlated significantly with effectiveness of operations (.34) and continued satisfaction with the leadership (.68). The partial correlation with contingent reinforcement was reduced to close to zero. Bass and Avolio concluded that the positive contribution that contingent rewarding by a leader is often seen to make to perceived effectiveness may be accounted for by the image of leadership held by the raters. The same may be true for the transformational effects of intellectual stimulation and individualized consideration. However, the charismatic effect cannot be accounted for in this way. Complicating these results may be the extent to which the followers’ motivation affects their image of the ideal leader.
Imagined Ideal Leader and the Followers’ Needs. Singer and Singer (1986) reported that among male undergraduates in New Zealand, the extent to which contingent-reinforcement leadership was desired in an imagined ideal leader did not depend on the men’s motivation. But those who had a higher need for affiliation were more likely to prefer transforming, charismatic, and considerate leaders. Nonconformists were more likely to prefer transforming intellectually stimulating leaders.
Leadership can be understood as a transaction or exchange of material, social, and psychological benefits. In a fair and profitable exchange, the benefits to both the leader and the follower exceed their costs. In the transactional process, leaders and followers reinforce each other’s behavior with either reward or punishment—preferably reward, and preferably reward that is contingent on fulfilling the transacted role arrangements. But the exchange may be less rewarding; it can involve management by exception or punitive discipline. In the dynamics of the exchange relationship, competence and early acceptance of the group’s norms are important, as a follower builds the idiosyncrasy credit of a member, making it more likely that his or her attempts to lead will be successful. With enough idiosyncrasy credit, one can begin to stand out in the group and be successful in introducing new ideas. But those with credit need to avoid violating the norms and expectations of the group.
Leaders’ rewarding and punishing may be contingent or noncontingent, depending on followers’ performance. Management by exception may be active or passive. Leaders may treat each follower differently or in the same way. Within-group and between-group analyses can show how much leaders vary in these respects.
Reinforcement can be manipulated in experiments to determine who emerges as a leader. Reinforcement can be used to increase different members’ attempts to lead and to affect whether their attempts will be successful in influencing other members. To understand many individual and organizational outcomes of leadership, it is necessary to focus attention on the dyadic leader-member exchange, for the same leader is likely to have different expectations and reactions about his or her different subordinates. These different perceptions and behaviors of leaders result in different levels of performance by subordinates—who are working for the same leader—that cannot be captured by looking only at leader-group relationships. However, this is not an either-or matter.
The combination of dyadic and group effects can be teased out.
Contingent reward may be effective for leaders in many situations, but it has limitations. Rewards for performance and disciplinary actions for failures may not work as expected, for numerous reasons ranging from the leader’s lack of control over what the followers are seeking to the overriding impact of group norms. The carrot-or-stick approach may make the subordinate feel denigrated and less than an adult person. What can be accomplished with feedback depends on its timeliness, accuracy, attractiveness, and judged importance, as well as the trustworthiness, expertise, and authority of the source. Negative feedback is often distorted by both the leader who sends it and the subordinate who receives it. Many other factors related to superiors and subordinates will moderate its effects and generate unintended consequences. Systematically affecting the meaning, interpretation, and understanding of the leader’s efforts to provide contingent reinforcement will be the leader’s and followers’ implicit theories about leadership. Important here are the prototypical leaders that fit the followers’ images of what leaders are like. Leaders and followers have been likened to two sides of the same coin. The behavior of each can serve to reinforce the other.