Transformational leader behaviors and substitutes for leadership as determinants of employee satisfaction, commitment, trust, and organizational citizenship behaviors

Transformational leader behaviors and substitutes for leadership as determinants of employee satisfaction, commitment, trust, and organizational citizenship behaviors

Philip M. Podsakoff

During the past decade and a half, two rather distinct lines of theory and research have emerged in an attempt to improve our ability to understand leadership effectiveness. One of these approaches focuses on the identification and examination of those leader behaviors that influence followers’ values and aspirations, activate their higher-order needs, and arouse them to transcend their own self-interests for the sake of the organization (Bass, 1985; Yukl, 1989a, 1989b). These transformational or charismatic behaviors are believed to augment the impact of transactional(1) forms of leader behavior on employee outcome variables, because “followers feel trust and respect toward the leader and they are motivated to do more than they are expected to do” (Yukl, 1989b, p. 272). Examples of this new focus on leadership include the work of Burns, Bass, House, and others (cf. Avolio & Bass, 1988; Bass, 1985; Bass, Avolio & Goodheim, 1987; Bass, Waidman, Avolio & Bebb, 1987; Bennis & Nanus, 1985; Boal & Bryson, 1988; Burns, 1978; Conger & Kanungo, 1987; House, 1977; House, Spangler & Woycke, 1991; House, Woycke & Fodor, 1988; Howell & Frost, 1989; Shamir, House & Arthur, 1993; Tichy & DeVanna, 1986). Although these approaches differ somewhat from each other, as noted by Podsakoff, MacKenzie, Moorman and Fetter (1990), the majority of them share the common perspective that by articulating a vision of the future of the organization, providing a model that is consistent with that vision, fostering the acceptance of group goals, and providing individualized support, effective leaders change the basic values, beliefs, and attitudes of followers so that they are willing to perform beyond the minimum levels specified by the organization.

Research on the transformational leadership paradigm has proven to be rather promising. For example, Bryman (1992) cites a variety of organizational studies demonstrating that transformational leader behaviors are positively related to employees’ satisfaction, self-reported effort, and job performance. Similar results have been reported in several field studies (cf. Avolio & Bass, 1988; Bass, Avolio & Goodheim, 1987; Bass, Waldman, Avolio & Bebb, 1987; Bennis & Nanus, 1985; Conger & Kanungo, 1987; House, Spangler & Woycke, 1991; House, Woycke & Fodor, 1988; Roberts, 1985; Trice & Beyer, 1986), from a variety of samples and organizational settings. In addition, in a laboratory study designed to examine the relative impact of directive leader behavior versus charismatic leadership behavior (which is considered by many to be a form of transformational leadership behavior), Howell and Frost (1989) found that charismatic leader behavior produced higher performance, greater satisfaction, and greater role clarity, than directive leader behavior.

Almost concurrently with the emergence of the transformational approach to leadership, but not necessarily related to it, has been an increased interest in Kerr and Jermier’s (1978) “substitutes for leadership” model. According to this approach, the key to improving leadership effectiveness is to identify the situational variables that can either substitute for, neutralize, or enhance the effects of a leader’s behavior. Included among variables that have been identified by Kerr and Jermier (1978) as potential substitutes for leadership are four subordinate characteristics (ability, experience, training, and knowledge; need for independence; professional orientation; and indifference to organizational rewards), three task characteristics (task feedback; routine, methodologically invariant tasks; intrinsically satisfying tasks), and six organizational characteristics (organizational formalization; organizational inflexibility; group cohesiveness; amount of advisory/staff support; rewards outside the leader’s control; and the degree of spatial distance between supervisors and subordinates). Unlike the transformational approach to leadership, which assumes that it is the leader’s transformational behavior that is the key to improving leadership effectiveness, the substitutes for leadership approach assumes that the real key to leadership effectiveness is to identify those important situational or contextual variables that may “substitute” for the leader’s behavior, so that the leader can adapt his or her behavior accordingly.

The substitutes model has attracted a great deal of research interest (cf. Howell & Dorfman, 1981, 1986; Jermier & Berkes, 1979; Kerr & Jermier, 1978; Podsakoff, MacKenzie & Fetter, 1993; Podsakoff, Niehoff, MacKenzie & Williams, 1993; Podsakoff, Todor, Grover & Huber, 1984; Sheridan, Vredenburgh & Abelson, 1984). However, the results of this research suggest that the substitutes variables behave somewhat differently than expected by Kerr and his colleagues (cf. Howell, Dorfman & Kerr, 1986). Consistent with other situational approaches to leadership, the basic assumption made by Kerr and his colleagues (cf. Howell et al., 1986) is that the substitutes for leadership variables have their primary effects on subordinate criterion variables through their interactions with the leader behaviors of interest. That is, the substitutes variables are predicted to moderate the relationships between leader behaviors and subordinate criterion variables. However, recent research (cf. Podsakoff, MacKenzie & Fetter, 1993; Podsakoff, Niehoff, MacKenzie & Williams, 1993) designed to test these predictions has not been all that supportive. For, in spite of the fact that the substitutes for leadership had a number of important main effects, and accounted for a large proportion of the variance in the criterion variables, relatively few of the substitutes had moderating effects consistent with those predicted by Howell et al., (1986).

Thus, the existing empirical evidence does not appear to be very supportive of the substitutes for leadership model, at least in terms of their moderating effects. This might suggest that they have little impact on the effects of leadership. However, such a conclusion is premature, because several studies (cf. Farh, Podsakoff & Cheng, 1987; Podsakoff, MacKenzie & Fetter, 1993; Podsakoff, Niehoff, MacKenzie & Williams, 1993) have found that the substitutes for leadership variables are significantly related to many forms of leader behavior (as well as to the criterion variables). This suggests that any structural model designed to examine the impact of a leader’s behavior on subordinate attitudes, role perceptions, and performance, that does not include both the substitutes for leadership, and the leader behaviors, is misspecified and will produce biased estimates of the effects of the leader’ s behavior, since the substitute variables are significantly correlated with the leader behaviors, and with the criterion variables. This implies that much of what we know about the impact of leader behavior on subordinate criterion variables could be inaccurate, due to the omission of the substitutes for leadership variables. Whether this is true or not depends upon the extent to which the substitute variables are correlated with the leadership behaviors of interest, and the criterion variables.

Potential Effects of Substitutes for Leadership on the Impact of Transformational Leader Behaviors

Interestingly, these two major streams of research have never been merged, even though there is good reason to expect that the substitute variables may influence the impact of transformational leader behaviors. In their original paper, Kerr and Jermier (1978) proposed a set of substitutes for leadership variables that they believed would generally moderate the impact of a leader’s behavior on subordinate criterion variables. Their first test of the model quite naturally focused on the dominant forms of leader behaviors extant at that time (e.g., relationship-oriented, supportive, and people-centered leader behaviors; and task-oriented, instrumental, and job-centered leader behaviors). However, from the beginning, Kerr and Jermier (1978) clearly intended their model to be more broadly applicable, “Since Table 1 is derived from previously conducted studies, substitutes are only suggested for the two leader behavior styles which dominate the research literature. The substitutes construct probably has much wider applicability, however, perhaps to…leadership in general.” (p. 378). Consistent with this interpretation, Farh, Podsakoff and Cheng (1987), Podsakoff, Todor, Grover and Huber (1984), Podsakoff, MacKenzie and Fetter (1993), and Podsakoff, Niehoff, MacKenzie and Williams (1993), applied the model to various forms of leader reward and punishment behavior, and Sheridan, Vredenburgh and Abelson (1984), applied it to decision making, informational, and interpersonal forms of leader behavior. Thus, the substitutes model was originally intended to be, and has been, applied to a broad range of leader behaviors.

However, the substitutes model has never been applied to transformational leadership, even though there is reason to suspect that the substitute variables may moderate the impact of some forms of transformational leadership behavior. For example, in their original specification of the model, Kerr and Jermier (1978) explicitly proposed that substitutes for leadership should moderate the impact of supportive leader behavior, which is regarded by many as a form of transformational leadership (cf. Avolio & Bass, 1988; Bass, 1985; Bass, Avolio & Goodheim, 1987; Bass, Waldman, Avolio & Bebb, 1987; Conger & Kanungo, 1987). This hypothesis subsequently found partial support in the work of Dobbins and Zaccaro (1986), who found that group cohesiveness moderates the impact of individualized support on employee satisfaction.

Similarly, one might expect the impact of high performance expectations, providing an appropriate model, and articulating a vision, to be moderated by group cohesiveness, because group members may set expectations for performance and appropriate behavior, and possibly even have a different vision than that of the leader. It is also possible that the impact of a leader’s high performance expectations, providing an appropriate model, and articulating a vision, may be moderated by the extent to which he or she is perceived to control important organizational rewards. Bandura’s (1977) research suggests that modeling the appropriate behavior will only be effective when a leader can reward subordinates for exhibiting the desired behavior. By a similar logic, followers may be less likely to move in the articulated direction, or meet high performance expectations, if they do not believe their leaders will be able to reward them for it. Task feedback might also be expected to have a similar effect on the impact of these behaviors, based on Locke and Latham’s (1990) review of a number of studies showing that feedback generally enhanced the effect that goal setting (high performance expectations) had on task performance.

The impact of fostering the acceptance of group goals and intellectual stimulation might also be moderated by some of the substitutes for leadership. For example, followers who like to work alone, and have a high need for independence, may be particularly immune to attempts on the part of a leader to foster the acceptance of group goals. In addition, continually pressuring subordinates to re-think the way they do their work (intellectual stimulation) may be quite effective for subordinates with a low need for independence, and quite irritating and ineffective for subordinates with a high need for independence. This may help explain why the effects of intellectual stimulation have been positive in some cases (Bass, 1985), and negative in others (Podsakoff, MacKenzie, Moorman & Fetter, 1990). Thus, it may be that the impact of each of the forms of transformational leader behavior previously identified in the leadership literature may be moderated by one or more of the substitutes for leadership.

Therefore, the purpose of the present study is to determine the potential effects that substitutes for leadership have on the relationships between transformational leader behaviors and followers’ attitudes, role perceptions, and performance. More specifically, we will examine: (a) the main effects of the transformational leader behaviors in the context of the effects of the substitutes for leadership, and (b) the moderating effects of the substitutes for leadership on the relationships between the transformational leader behaviors and the follower criterion variables. This will help us to determine whether the effects of transformational leadership are as context-free as many suppose them to be, and whether previous estimates of the effects of transformational leader behaviors on these criterion variables have been biased by the omission of the substitutes variables.

It is important to note that, in addition, this study will also examine the effects of transformational leader behaviors and substitutes for leadership on a more comprehensive set of performance measures. The majority of studies on transformational leadership behaviors have focused on the effects that these behaviors have on “in-role” performance, rather than on “extra-role” or “organizational citizenship behaviors” (OCBs, Organ, 1988). However, as noted by Podsakoff et al. (1990), even though the effects of transformational leadership on “in-role” performance are important, they may not be as important as the effects of transformational leadership on extra-role and/or citizenship behavior. Others have recognized this as well. Indeed, Boal and Bryson (1988, p. 11) argue that the essence of transformational leadership is that such leaders “…lift ordinary people to extraordinary heights”; Yukl (1989b, p. 272) argues that they cause subordinates to “…do more than they are expected to do”; Bass (1985) says that they get people to “perform beyond the level of expectations;” and House et al. (1988, p. 100) claim that these leaders motivate their subordinates to “perform above and beyond the call of duty.” Taken together, this suggests that transformational leadership may have a number of important effects on extra-role or organizational citizenship behaviors. The same is true of substitutes for leadership. Recent research by Podsakoff, Niehoff, MacKenzie and Williams (1993), and Podsakoff, MacKenzie and Fetter (1993) has documented the linkage between substitutes for leadership and OCBs or extra-role behavior. Consequently, organizational citizenship behaviors were included as a key criterion variable in this study.

Method

Sample

Measures of leader behaviors, job attitudes, and role perceptions for this study were collected from 1539 employees, and matching performance data were collected for 1200 of them from their managers.(2)

A portion of the data was previously published in a study by Podsakoff and MacKenzie (1994a), which examined the psychometric properties of the substitutes for leadership scales. However, that study did not include transformational leader behaviors or examine their impact on the criterion variables included in this study, nor did it examine the impact of substitutes for leadership on these criterion variables, or the interaction of the substitutes and transformational leader behaviors.

The majority of the respondents held white collar, managerial, and professional positions. The sample was drawn from multiple divisions of several large companies located throughout the U.S. and Canada. In an attempt to increase the variability of the substitutes measures, the sample was chosen from a wide range of industries (printing, automotive vehicles and parts, office furniture products, banks, diversified financial services, petroleum refining, chemical, pharmaceutical, electronics, photographic, information services, home appliances, pulp and paper, recycling, office products, computer services, electrical equipment, plastic products, and food industries), organizational levels (entry level through CEOs and presidents), and from companies of varying size (from 10 million dollars to Fortune 100 size), some of whom had both foreign and domestic operations. While the industries in which these organizations operate vary, all of these companies are comparable in that they are all large, publicly-held firms, consisting primarily of well educated, professional employees. More specifically, 63.4% of the subjects possessed a four-year college degree or higher, and 70.7% were either managerial or technical/professional employees. Almost forty percent (38%) were members of a professional association or society at the time of the survey. The respondents had an average age of 36.5 years, average company tenure of 11.3 years, and an average tenure of 2.9 years under their present supervisor.

Procedure

Survey questionnaires were administered to the respondents from each sample in their work settings during normal working hours, or respondents were allowed to take the survey home to complete if they so chose. Included with each packet was a letter from the researchers indicating the general nature of the survey, and assuring all respondents that their individual responses would remain anonymous. Also included with each survey was a stamped envelope, addressed directly to the researchers. The sample size of 1539 represents a return rate of 91%.

Predictor Variables

Transformational Leader Behaviors. Podsakoff et al.’s (1990) transformational leadership behavior inventory (TLI) was used to assess the leader behaviors measured in the study. This scale is designed to measure six key dimensions of transformational leadership that have been identified in the research literature (cf. Bass, 1985; Bennis & Nanus, 1985; Bradford & Cohen, 1984; Conger & Kanungo, 1987; House, 1977; Kouzes & Posner, 1987; Tichy & DeVanna, 1986). The dimensions are: articulating a vision, providing an appropriate model, fostering the acceptance of group goals, high performance expectations, providing individualized support, and intellectual stimulation. Previous research using this scale generally supported the hypothesized factor structure; however, three of the dimensions (articulating a vision, providing an appropriate model, and fostering the acceptance of group goals) were found to be highly intercorrelated. In view of the fact that Podsakoff et al.’s study is the only one that has empirically examined the scale’s properties, it is important to confirm the scale’s psychometric properties in the present study.

Substitutes for Leadership. The 13 substitutes for leadership constructs identified by Kerr and Jermier (1978) were measured with the 41-item scale developed by Podsakoff, MacKenzie and Fetter (1993). This scale has been shown to possess generally good psychometric properties, and to correlate with other variables in a manner that is consistent with its nomological net (cf. Podsakoff & MacKenzie, 1994a; Podsakoff, MacKenzie & Fetter, 1993).

Criterion Variables

Eleven criterion variables were examined in the current study. Five of these criterion variables (General Satisfaction, Organizational Commitment, Trust, Role Clarity, and Role Conflict) were self-report measures. The other six criterion variables (employee “in-role” performance, altruism, conscientiousness, courtesy, civic virtue, and sportsmanship) were behavioral measures provided by the supervisors of each of the respondents.

Employee (or Self-) Assessed Criterion Variables. General Satisfaction was measured with the 20-item short form of the Minnesota Satisfaction Questionnaire (MSQ; Weiss, Dawis, England & Lofquist, 1967). The MSQ has been shown to possess generally good psychometric properties (cf. Gillet & Schwab, 1975; Price & Mueller, 1986; Weiss et al., 1967), and correlates well with other measures of job satisfaction (cf. Gillet & Schwab, 1975; Wanous, 1974). Organizational Commitment was measured with the 15-item scale developed by Porter, Steers, Mowday, and Boulian (1974). This scale is designed to assess the relative strength of an employee’s identification with and involvement in the organization (Mowday, Steers & Porter, 1979). Previous research (Mathieu & Zajac, 1990; Porter et al., 1974; Mowday, Steers & Porter, 1979) has demonstrated that this 15-item scale possesses adequate psychometric properties. Trust in and loyalty to the leader was assessed with the six item scale used by Podsakoff et al. (1990). Their research has shown that: (a) all of the items load on the intended factor, (b) a one-factor model fit the data very well, and (c) the scale had a very acceptable internal consistency reliability (.90). Finally, shortened versions of Rizzo, House and Lirtzman’s (1970) scales were used to assess employees’ role clarity and role conflict. The role clarity and role conflict scales contained six and eight items, respectively. The reduced item versions of Rizzo et al.’s. (1970) scales have been widely used, and generally positive evidence exists on both their reliability and validity (cf. House, Schuler & Levanoni, 1983; Schuler, Aldag & Brief, 1977).

Manager Assessed Criterion Variables. The six behavioral measures provided by the respondents’ supervisors were intended to capture both “in-role” and “extra-role” aspects of subordinates’ performance. In-role performance was defined as those activities that an employee is expected to perform to meet the prescribed requirements of the job, and was measured with a 4-item scale developed by Williams (1989). This scale asked supervisors to rate the degree to which a subordinate fulfills the formal requirements of his or her job, and performs all essential job duties. In addition to measuring employees’ in-role performance, we also measured several “extra-role” or “organizational citizenship behaviors” (OCBs; Organ, 1988, 1990) using a modified version of the scales developed by Podsakoff and MacKenzie (1989). The items included in this scale measure all five of the “citizenship behavior” dimensions identified by Organ (1988), including altruism, courtesy, conscientiousness, civic virtue, and sportsmanship. Previous research by a number of researchers (cf. MacKenzie, Podsakoff & Fetter, 1991; Moorman, 1991, 1993; Moorman, Niehoff & Organ, 1993; Niehoff & Moorman, 1993; Podsakoff & MacKenzie, 1994b; Podsakoff, MacKenzie, Moorman & Fetter, 1990; Podsakoff, MacKenzie & Fetter, 1993; Tansky, 1993) has been very encouraging, and generally shows this scale to possess good validity and very acceptable levels of internal consistency reliability.

Seven-point Likert scales ranging from (1) “strongly disagree” to (7) “strongly agree” were utilized to assess all of the constructs measured in the present study, with the exception of the 20 MSQ scale items, which were assessed with the traditional five-point scales ranging from (1) “very dissatisfied” to (5) “very satisfied” used in prior research (Weiss et al, 1967).

Analytical Procedures

The data analysis was conducted in three major phases. First, we investigated the factor structure and reliability of the transformational leadership behavior inventory developed by Podsakoff et al. (1990). In the next phase of our analysis, we examined the aggregate effects of the set of transformational leader behaviors and leadership substitutes (i.e., individual, task, and organizational characteristics) on subordinate attitudes, role perceptions, and performance, to determine which groups of predictor variables had the greatest effects on each of the eleven criterion variables. Finally, we examined both the main and interactive effects of the transformational leader behaviors and 13 leadership substitutes on each of the criterion variables by using hierarchical moderated regression analysis procedures (Cohen & Cohen, 1983; Stone, 1988; Stone & Hollenbeck, 1984, 1989).

Results

Confirmatory Factor Analysis of the Transformational Leadership Behavior Inventory

Table 1 reports the completely standardized confirmatory factor loadings of the Transformational Leadership Behavior Inventory. As shown in this table, the [TABULAR DATA FOR TABLE 1 OMITTED] overall fit of the six-factor model to the data was quite good, even though the [[Chi].sup.2] (df) was 965.51 (194). Bentler’s (1990) comparative fit index (CFI) was .94, Bollen’s (1989) incremental fit index (IFI) was .94, Joreskog and Sorbom’s (1993) goodness of fit index (GFI) was .91, and Tucker and Lewis’s (1973) fit index (TLI) was .93. In addition, each of the hypothesized factor loadings was statistically significant at the .01 level, all of the items had completely standardized loadings of .60 or above, and Fornell and Larcker’s (1981) measure of the average amount of variance each latent factor accounted for in its indicators ([[Rho].sub.vc]) was quite large, ranging from 58% to 68% with an average of approximately 61%. Thus, there appeared to be good support for the hypothesized factor structure of the transformational leadership scale. However, this was evaluated further by testing whether any of the hypothesized factors could be combined – two, three, four, five or even six at a time – without significantly affecting the fit of the model. The results suggested that the hypothesized six factor model fit the data significantly better than any of these rival models.

Scale Means, Standard Deviations, Reliabilities, and Intercorrelations

The means, standard deviations, reliabilities, and intercorrelations of all of the variables used in the present study are reported in Table 2. An examination of this table indicates that the mean internal consistency reliability for the 30 constructs used in this study was a very respectable .82, and that the reliabilities for all of the constructs except two (need for independence, [Alpha] = .69; and rewards outside the leader’s control, [Alpha] = .67) reported in this table meet or exceed Nunnally’s (1978) recommended level of .70 for newly developed scales. The intercorrelations reported in Table 2 also indicate that several of the transformational leader behaviors included in this study are correlated with several of the substitutes for leadership. This underscores the importance of including both sets of variables in leadership research if the unique contributions of the leader behaviors and substitutes for leadership on the criterion variables are to be examined.

Aggregate Effects on Criterion Variables

Given the preliminary evidence described above, our analysis shifted to an investigation of the impact of the transformational leader behaviors and substitutes for leadership on the eleven criterion variables. The first step was to assess the amount of variance in the criterion variables accounted for by the complete set of leader behaviors and substitutes. This was done by regressing each of the eleven criterion measures on the six leader behaviors and the 13 substitutes variables. The equation for this full model is as follows:

[Y.sub.i] = [[Alpha].sub.i] + [[Sigma].sub.j][[Beta].sub.ij][TLB.sub.j] + [[Sigma].sub.k][[Beta].sub.ik][SUBSUB.sub.k] + [[Sigma].sub.l][[Beta].sub.il][TASKSUB.sub.l] + [[Sigma].sub.m][[Beta].sub.im][ORGSUB.sub.m] + [Epsilon] (1)

where:

[Y.sub.i] = the ith criterion variable (i = 1 to 11)

[TLB.sub.j] = the jth transformational leadership behavior (j = 1 to 6)

[UBSUB.sub.k] = the kth subordinate substitute variable (l = 1 to 4)

[TASKSUB.sub.l] = the lth task substitute variable (m = 1 to 3)

[ORGSUB.sub.m] = the mth organizational substitute variable (n = 1 to 6)

[TABULAR DATA FOR TABLE 2 OMITTED]

Following the estimation of this model, the transformational leader behavior category was omitted from the full model to determine the total amount of variance in the criterion variables uniquely attributable to the set of leader behaviors included in the study. Next, each of the three substitutes for leadership categories (subordinate, task, and organizational characteristics) was omitted from the model sequentially so that we could determine: (a) the unique contribution of each set of substitutes to the amount of variance accounted for in the criterion variables, and (b) the total amount of variance in the criterion variables uniquely attributable to the complete set of substitute variables. Then, the amount of variance in each criterion variable shared by any combination of the three substitutes categories was derived by subtracting the unique contribution of each set from the total amount of variance attributable to all of the substitutes combined. Finally, the amount of variance in the criterion variables shared by the transformational leadership behavior category and the leadership substitutes categories was also examined. The results of the aggregate analyses are reported in Table 3. Each numerical entry in a column of this table represents the percent of variance ([R.sup.2]) attributable to the variables identified in the corresponding row of the table. For example, the first numerical entry in the table, under the column titled “General Satisfaction” indicates that transformational leadership behaviors accounted for 7% of the total variance in followers’ general satisfaction.

An examination of Table 3 reveals several interesting findings. First, consistent with the recent research reported by Podsakoff, Niehoff, MacKenzie and Williams (1993), and Podsakoff, MacKenzie and Fetter (1993), the amount of variance explained in general satisfaction (71%), organizational commitment (48%), trust (55%), role clarity (52%), and role conflict (28%), by the leader behaviors and substitutes for leadership is quite respectable when contrasted with prior research which has examined the same criterion variables (cf. Jackson & Schuler, 1985; Mowday et al., 1979; Rizzo et al., 1970). Of course, it is possible to argue that some of the variance explained in these criterion variables may be due to the fact that they were obtained from the same respondents that the leader behavior and substitute measures were obtained from and therefore may, in part, be due to common method variance (cf. Bagozzi, Yi & Phillips, 1991; Cote & Buckley, 1987; Podsakoff & Organ, 1986). However, the amount of variance explained in the behavioral criterion variables, which are not subject to any same-source bias, is also higher (with an average of approximately 11.5%) than the 3%-6% generally reported in much of the extant leadership literature.

Second, the data also indicate that, although transformational leadership behaviors account for more variance than substitutes for leadership in follower trust and courtesy, and about the same amount of variance in the case of “in-role” performance and altruism, leadership substitutes have a greater impact than these leader behaviors on the majority of follower attitudes, role perceptions, and behaviors. Thus, while transformational leadership behaviors generally account for more variance in follower trust (28% versus 5%) and courtesy (7% versus 4%), and about an equal amount of variance in the case of “in-role” performance (3% versus 5%) and altruism (2% versus 4%); leadership substitutes account for substantially more variance than transformational leadership behaviors in the case of follower [TABULAR DATA FOR TABLE 3 OMITTED] satisfaction (37% versus 7%), commitment (33% versus 1%), role clarity (29% versus 3%), role conflict (18% versus 3%), civic virtue (9% versus 1%), conscientiousness (6% versus 2%), and sportsmanship (5% versus 2%).

Third, with the exception of civic virtue, the percentage of variance shared between the combination of the transformational leader behaviors and substitutes for leadership, relative to the total amount of variance accounted for in the criterion variables, is quite noteworthy (ranging from 3% to 27%, with an average of approximately 10.8%). Moreover, the variance jointly explained by the leader behaviors and substitutes for leadership represents, on average, approximately one-third (33.7%) of the total variance in the follower criterion variables.

These findings suggest that, with the exception of employee trust, one generally loses little explanatory value by excluding the effects of variation in the transformational leadership behaviors in the case of follower attitudes and role perceptions. However, the same is not generally true in the case of employee performance. For example, as indicated in Table 3, the combination of the transformational leader behaviors and leadership substitutes accounted for a total of 48% of the variance accounted for in follower commitment, but only 1% is uniquely attributable to the transformational leader behaviors, while the remaining 47% is either attributable to the substitutes variables (33%), or shared (14%) by the leader behaviors and substitutes variables. Practically speaking, this means that even if variation in the leader behaviors was excluded from our analysis, all but 1% of the variance (which represents approximately 2% of the total of 48%) in employee commitment would be accounted for. Similarly, the exclusion of the transformational leader behaviors as predictors of followers’ role clarity, role conflict, or satisfaction, would result in accounting for all but approximately 3% (which represents less than 6% of the total of 52%), 3% (which represents less than 11% of the total of 28%), and 7% (which represents approximately 10% of the total of 71%), respectively, of the variance accounted for in these criterion variables. The one exception to this finding in the case of employee attitudes and perceptions is employee trust, where exclusion of the variation in the transformational leader behaviors would reduce the total amount of variance accounted for by 28% (which represents over 50% of the total of 55%).

However, in general, it is not safe to ignore the effects of leader behaviors on employee performance. For example, exclusion of variation attributable to the leader behaviors from employee courtesy would reduce the total proportion of variance accounted for from 15% to 8% (which represents almost 47% of the total variance accounted for in this criterion variable; 7/15 = 46.7%). Likewise, exclusion of variation attributable to the transformational leader behaviors from followers’ in-role performance, altruism, conscientiousness, and sportsmanship, would result in a reduction in the total percent of variance accounted for of approximately 27%, 20%, 18%, and 18%, respectively. The one exception to this finding is in the case of civic virtue, where exclusion of the variation attributable to the transformational leader behaviors would reduce the total amount of variance accounted for by a mere 1% (which represents about 9% of the total of 11%). Thus, in general, it appears that transformational leadership behaviors are relatively more critical to the determination of employee performance than they are to employee attitudes and perceptions.

Individual Effects

In the next phase of our analysis, we examined the individual effects of each of the six transformational leader behaviors and 13 substitutes on the 11 criterion variables in our study by regressing each of the criterion variables on the full set of leader behaviors and substitutes. This procedure permits one to examine the unique effects of each individual leader behavior and substitute variable while controlling for all of the other predictor variables. Table 4 reports the results of this phase of these analyses. The standardized regression coefficients ([Beta]’s) reported in this table provide an estimate of the relative effect of each leader behavior and substitute on the corresponding criterion variable.

An overall examination of Table 4 indicates that all of the transformational leadership behaviors, and all of the substitutes for leadership, significantly influenced one or more of the criterion variables. This table also indicates that the transformational leader behaviors, “organizational characteristics” and “task characteristics” categories had significant effects on all of the criterion variables (11 out of 11); and the “subordinate characteristics” category had significant effects on all but one of the criterion variables. In addition, the organizational characteristics category had the greatest number of effects (25 of the total of 89, or 28%), followed by transformational leader behaviors (24 of 89, or 27%), task characteristics (with 23 of 89, or 26%), and subordinate characteristics (17 of 89, or 19%). Finally, of all of the individual predictor variables, individualized support served as a predictor of the greatest number of criterion variables (10 out of 11), followed closely by task routinization (with 9 of 11), indifference to organizational rewards (which served as a predictor of 8 of the 11 criterion variables), organizational formalization (with 7 out of 11), and task feedback, intrinsically satisfying tasks, and group cohesiveness (each of which served as a predictor of 6 of the 11 criterion variables).

General Satisfaction. Five leader behaviors and six leadership substitutes had significant main effects on followers’ general satisfaction. Of the leader behaviors that were significant, individualized support ([Beta] = .12), providing an appropriate model ([Beta] = .10), and vision articulation ([Beta] = .09), had the strongest positive effects on general satisfaction. Fostering the acceptance of group goals also had a positive effect on satisfaction ([Beta] = .05), while high performance expectations had a negative effect on this criterion variable ([Beta] = -.05). These findings suggest that leaders who are supportive, provide an appropriate model, clarify their vision, foster common goals among their work groups, but do not convey overly ambitious expectations, have more satisfied employees than leaders who are not supportive, do not provide an appropriate model, do not clarify their vision or foster common goals within their group, or are perceived to demand too much from their subordinates.

Among the substitutes, intrinsically satisfying tasks had the strongest impact ([Beta] = .39) on general satisfaction, followed by indifference to organizational rewards ([Beta] = -.26), group cohesiveness ([Beta] = .14), organizational inflexibility ([Beta] = .03), and professional orientation ([Beta] = -.03). Thus, subordinates who perform satisfying [TABULAR DATA FOR TABLE 4 OMITTED] tasks and work in groups that are cohesive are more satisfied than subordinates who perform less satisfying tasks, or work in groups that are less cohesive. Somewhat surprisingly, subordinates who perceived their organization to be more inflexible were also more satisfied than subordinates who perceived their organization to be less inflexible. In addition, subordinates who are indifferent toward the rewards offered by their organization, or who are professionally oriented, are less Satisfied than subordinates who value the rewards offered by their organization, or are not professionally oriented. Taken as a whole, therefore, the above results suggest that subordinate satisfaction is determined somewhat more strongly by leadership substitutes than by leader behaviors.

Organizational Commitment. Only one of the leader behaviors (vision articulation) had a significant main effect on organizational commitment ([Beta] = .10), while six substitutes for leadership (indifference to rewards, intrinsically satisfying tasks, routine tasks, organizational inflexibility, group cohesiveness, and rewards outside the leader’s control) influenced this criterion variable. Indifference to organizational rewards ([Beta] = -.46), rewards outside the leader’s control ([Beta] = -.06), and routine tasks ([Beta] = -.06) tend to decrease commitment, while intrinsically satisfying tasks ([Beta] = .21), organizational inflexibility ([Beta] = .10), and group cohesiveness ([Beta] = .06) tend to increase it. Thus, leadership substitutes, as opposed to leadership behaviors, appear to be the key determinants of employees’ commitment to the organization.

Trust. Only three leader behaviors, as opposed to five substitutes for leadership, had significant main effects on employee trust. However, two of the leader behaviors (providing an appropriate model, [Beta] = .33, and individualized support, [Beta] = .28), had the greatest impact on this criterion variable. Fostering the acceptance of group goals ([Beta] = .06) also had a positive impact on employee trust. These findings suggest that followers who perceive their leaders to provide appropriate models, be supportive, or foster the acceptance of group goals, tend to express more trust in their leaders than employees who perceive their leaders otherwise.

Of the substitutes variables, group cohesiveness ([Beta] = .10), task feedback ([Beta] = .08), and ability, experience, training, and knowledge ([Beta] = .04), were also found to be positively related to employee trust; while both indifference to organizational rewards ([Beta] = -.13) and organizational formalization ([Beta] = -.05) were found to be negatively related to this criterion variable. Thus, subordinates who perceive themselves to possess a high degree of ability, experience, training, or knowledge, work in more cohesive groups, receive a substantial amount of task feedback, value the organizations’ rewards, or do not perceive the organization to have too many rules and regulations, express more trust than subordinates who perceive themselves to have less ability, experience, training and knowledge, work in less cohesive work groups, receive little task feedback, are indifferent to organizational rewards, or feel the organization is overly formalized.

Role Clarity. As a whole, subordinates’ perceptions of role clarity are primarily influenced by leadership substitutes, rather than leadership behaviors. Only two leader behaviors, articulating a vision ([Beta] = .09) and individualized support ([Beta] = .10), affect subordinates’ role clarity. However, six leadership substitutes significantly influence this criterion variable. Among the substitutes, organizational formalization ([Beta] = .29) and task feedback ([Beta] = .25) had the strongest effects, followed by intrinsically satisfying tasks ([Beta] = .15), subordinates’ perception of their ability, experience, training, and knowledge ([Beta] = .14), organizational inflexibility ([Beta] = .07), and task routinization ([Beta] = .05). These results suggest that employees whose supervisors articulate an effective vision and provide support, who work in organizations that are more formalized or inflexible, who perform satisfying tasks or tasks that provide more feedback, who are members of more cohesive work groups, or who have high levels of ability, experience, training and knowledge, generally express more role clarity than employees who face the opposite conditions.

Role Conflict. Four leader behaviors and 11 substitutes for leadership had significant main effects on role conflict. Of the leader behaviors that had an effect, two (individualized support, [Beta] = -.11; and providing an appropriate model, [Beta] = -.09), were negatively related to role conflict, while the other two (intellectual stimulation, [Beta] = .14; and high performance expectations, [Beta] = .07) were positively related to this criterion variable. These findings suggest that leaders who are supportive, and provide an effective model, tend to reduce subordinates’ perceptions of role conflict; while leaders who have unreasonably high expectations, or who provide intellectual stimulation to their subordinates, tend to increase the subordinates’ perceptions of role conflict.

Among the 11 leadership substitutes that had effects on role conflict, advisory and staff support ([Beta] = .21), indifference to rewards ([Beta] = .15), spatial distance ([Beta] = .14), ability, experience, training, and knowledge ([Beta] = .11), and professional orientation ([Beta] = .05) were also found to increase role conflict, while task routinization ([Beta] = -.14), organizational inflexibility ([Beta] = -.13), intrinsically satisfying tasks ([Beta] = -.11), organizational formalization ([Beta] = -.07), group cohesiveness ([Beta] = -.06), and task feedback ([Beta] = -.06), were found to decrease it. Thus, as in the case of role clarity, employees’ perceptions of role conflict are primarily determined by substitutes for leadership, rather than leader behaviors.

In-Role Performance. Two of the transformational leader behaviors examined in this study had a significant main effect on employee in-role performance. Both individualized support ([Beta] = .10), and fostering the acceptance of group goals ([Beta] = .09) tended to increase followers’ in-role performance. Thus, followers who perceive their leaders to be more supportive, or to encourage the acceptance of group goals, generally perform at higher levels than followers who perceive their leaders to be less supportive, or not to encourage group goals.

In addition, five substitutes for leadership influenced this subordinate criterion variable. Routine tasks ([Beta] = -.09), indifference to organizational rewards ([Beta] = -.08), and rewards outside the leader’s control ([Beta] = -.06), all were negatively correlated with in-role performance; while need for independence ([Beta] = .07), and task feedback ([Beta] = .07) were positively related to this criterion variable. These findings suggest that subordinates who perform routine tasks, who are indifferent to the rewards they receive from the organization, perceive their leader to control few rewards, receive little feedback from the tasks they perform or who have a low need for independence, tend to exhibit less in-role performance than subordinates who do not perform routine tasks, value the rewards of their organization, perceive their leaders to control these rewards, receive substantial task feedback, or have a high need for independence.

Altruism. One leader behavior, and five substitutes for leadership had significant main effects on employee altruism. Both individualized support (Agb = .14) and intrinsically satisfying tasks ([Beta] = .10) were positively related to employee altruism; while task routinization ([Beta] = -.08), organizational formalization ([Beta] = -.06), advisory/staff support ([Beta] = -.06), and spatial distance ([Beta] = -.06), were negatively related to this criterion variable. These results suggest that subordinates who find their tasks intrinsically satisfying or who feel their leaders are supportive are more altruistic than subordinates who do not find their tasks to be intrinsically satisfying or do not feel that their leaders are supportive; while subordinates who perform routine tasks, receive advisory and staff support, perceive that the organization is more highly formalized, or are spatially removed from their leader, are less altruistic than their peers who face the opposite conditions. Thus, leadership substitutes were bigger determinants of employee altruism than transformational leadership behaviors.

Conscientiousness. Only one of the leader behaviors (individualized support) had any unique individual effects on subordinate conscientiousness; while four of the leadership substitutes had an effect on this criterion variable. Task routinization ([Beta] = -.12), indifference to organizational rewards ([Beta] = -.08), organizational formalization ([Beta] = -.06), and rewards outside the leader’s control ([Beta] = -.06), were all negatively related to employee conscientiousness; while individualized support ([Beta] = .11) was positively related to this form of citizenship behavior. Thus, employees who value organizational rewards and perceive that their leader controls them, perform less routine tasks, perceive their organizations to be less formalized, or perceive their leader to be supportive, tend to be more conscientious, than employees who do not value organizational rewards or do not perceive their leaders to control them, perform more routine tasks, perceive their organization to be more formalized, or perceive their leaders not to be supportive.

Sportsmanship. Two leader behaviors (articulating a vision, [Beta] = .12; individualized support, [Beta] = .09) were found to be positively related to employee sportsmanship. Subordinates who perceived their leaders to clearly articulate a vision of the future or be supportive, tended to exhibit more sportsmanship than subordinates who perceived their leaders not to exhibit these behaviors.

Six substitutes for leadership also had significant effects on subordinate sportsmanship. Four of these substitutes (ability, experience, training, and knowledge; indifference to rewards; routine tasks; and organizational formalization) were negatively related to employee sportsmanship; while the remaining two (intrinsically satisfying tasks and group cohesiveness) were positively related to this criterion variable. The results suggest that subordinates who possess more ability, experience, training, and knowledge, or are indifferent to organizational rewards, perform routine tasks or tasks that are less intrinsically satisfying, perceive their organizations to be more formalized, or work in groups that are less cohesive, are less likely to exhibit sportsmanship than subordinates who do not possess a lot of ability, experience, training, or knowledge, value organizational rewards, perform less routine or more intrinsically interesting tasks, perceive their organization to be less formalized, or work in more cohesive work groups. Of these main effects, vision articulation had the greatest impact ([Beta] = .12), followed by indifference to rewards ([Beta] = -.11), individualized support and intrinsically satisfying tasks (both with [Beta]’s = .09), group cohesiveness ([Beta] = .08), routine tasks ([Beta] = -.07), organizational formalization ([Beta] = -.06), and ability, experience, training, and knowledge ([Beta] = -.06). Thus, it appears that subordinate sportsmanship is influenced somewhat more by substitutes for leadership than by leader behaviors.

Courtesy. Only two of the transformational leadership behaviors, and three substitutes for leadership, had significant main effects on subordinates’ courtesy. However, one of the leader behaviors that did have an effect, individualized support, had the most substantial (positive) impact ([Beta] = .29) of any of the predictor variables. Somewhat surprisingly, high performance expectations also had a significant positive effect ([Beta] = .06), on this criterion variable.

Of the leadership substitutes, task routinization ([Beta] = -.10) and professional orientation ([Beta] = -.07) were negatively related to employee courtesy, and group cohesiveness ([Beta] = .09), was positively related to this form of citizenship behavior. Generally speaking, these findings suggest that employees who perceive their leaders to be supportive, or to have high performance expectations, as well as those employees who work in cohesive groups, perform less routine tasks, or are less professionally oriented, exhibit more courteousness to their peers than subordinates who perceive their leaders to be less supportive or have less demanding expectations, or who work in less cohesive work groups, perform routine tasks, or are more professionally oriented.

Civic Virtue. Although only one of the leader behaviors had any significant effects on civic virtue, six of the leadership substitutes influenced this criterion variable. The results indicate that employees who perceive their leader to provide individualized support, receive more task feedback, or who perceive themselves to have more ability, experience, training, and knowledge, are more likely to engage in civic virtue than employees who do not perceive their leader to be supportive, receive little task feedback, or who do not perceive that they have as much ability, experience, training, or knowledge.

In contrast, employees who perform routine tasks, perceive their organization to have highly formalized rules and regulations, do not believe that their leaders control rewards, or are indifferent to the rewards they do control, tend to exhibit less civic virtue than employees not faced with these conditions. Routine tasks had the greatest impact on civic virtue ([Beta] = -.18), followed by ability, experience, training, and knowledge ([Beta] = .12), organizational formalization ([Beta] = -.10), individualized support and task feedback (both with [Beta]’s = .08), and rewards outside the leader’s control, and indifference to rewards (both with [Beta]’s = -.08). Thus, civic virtue on the part of subordinates appears to be determined primarily by leadership substitutes, rather than leader behaviors.

Moderating Effects

The moderating effects of the substitutes were tested by examining the change in [R.sup.2] attributable to the leader behavior x substitute interaction terms added in the final stage of the hierarchical regression procedure. If the interaction term added to this stage of the regression analyses produces a significant [Delta][R.sup.2] (i.e., significantly increases the amount of variance explained in the criterion variable), then the substitute is identified as a moderator of the relationship between the leader behavior and criterion variable under investigation.

As discussed by Podsakoff, MacKenzie and Fetter (1993), for a variable to qualify as a neutralizer, both the main effect of the leader behavior and the interaction term must be significant, and they must have different signs. To qualify as an enhancer, both the leader behavior main effect and the interaction term must be significant, with the same signs. Finally, to qualify as a substitute: (a) the leader behavior must have a significant main effect; (b) the potential substitute variable must weaken the relationship between the leader behavior and the criterion variable (i.e., the interaction must be significant and it must have a different sign than the leader behavior main effect); and (c) the substitute must have a significant main effect on the criterion variable in the same direction as the leader behavior’s main effect. Only when conditions (a), (b), and (c) are met can it be said that the variable both weakens the impact of the leader’s behavior on the criterion variable and also replaces, or “substitutes” for it.

Graphically, the criteria specified above suggest that neutralizers may take any of the forms indicated in Figure 1. The figure shows six possible types of neutralizing effects, which vary according to whether the main effect of the leader behavior on the criterion variable is positive or negative, and whether the main effect of the neutralizer variable is positive, nonsignificant, or negative. Since substitutes are special types of neutralizing variables, they are also shown in this figure. Figure 1 (a) shows the classic substitution effect diagrammed by Howell et al. (1986, p. 93), where the effect of the leader’s behavior on the criterion variable is positive, this effect is weakened by the moderator variable, and it is replaced by a positive main effect of the moderator on the criterion variable. However, Figure 1 (f) also meets the statistical requirements for a substitution effect in that the moderator variable both weakens the impact of the leader’s behavior, and also replaces it. Examples of moderators that meet Howell et al.’s (1986) statistical requirements for enhancers are shown in Figure 2.

Consistent with the recommendations of Stone (1988), and Stone and Hollenbeck (1984, 1989), follow-up analyses were conducted in those cases in which a significant [Delta][R.sup.2] value was obtained for the interaction term by: (1) calculating separate regression equations (Y values) at values one standard deviation above and below the mean of the moderator variable, (2) plotting these results graphically to develop a visual picture of the nature of the moderating relationship, and (3) comparing the plots with Figures 1 and 2 to determine which of the interactions were consistent with the criteria specified by Howell et al. (1986).

The moderating effects of the substitutes on the leader behavior-criterion variable relationships are provided in Table 5. The values reported in this table are the standardized regression (Beta) weights for those interactions between the leader behaviors and substitutes that were found to be significant. Also reported (in parenthesis) is an alphanumeric character next to each interaction, which indicates the nature of the moderating relationship that was obtained in the follow-up split groups regression analyses. These characters refer to the basic patterns that are displayed in Figures 1 and 2.

As an aid to interpreting the information provided in Table 5, it may be helpful to examine a few of the interactions displayed in this table. For example, as shown by the first numerical entry under the “General Satisfaction” column of this table, intrinsically satisfying tasks served as a significant moderator of the relationship between a leader’s articulation of a vision and general satisfaction ([Beta] = .43). Next to this coefficient is the alphanumeric character (2c), which indicates the plot for this interaction looks like Figure 2(c). This figure indicates that articulating a vision has a significant positive impact on general satisfaction when subordinates perform a task that is more intrinsically satisfying, but no impact when their task is less intrinsically satisfying. Given that vision articulation also was found to have a significant positive main effect on employee satisfaction (see Table 4), professional orientation is an enhancer of the positive effects of vision articulation on employee satisfaction, according to Howell et al.’s (1986) statistical criteria.

As another example, notice that the third entry in the employee “Conscientiousness” column ([Beta] = -.86) has an (nc) symbol next to it. This symbol means that even though group cohesiveness moderated the effect of providing an appropriate model on employee sportsmanship, the nature of this interaction was not classifiable using the Howell et al. (1986) criteria. This situation arose whenever the leader behavior of interest interacted with the moderator variable, but: (a) the leader behavior did not have a main effect on the criterion variable itself, or (b) the effect of the leader behavior on the criterion variable was not significant at either the high or the low level of the moderator variable. Although such interactions indicate that the substitute of interest does, indeed, moderate the relationships between the leader behavior and criterion variables of interest, they do not fit Howell et al.’s [TABULAR DATA FOR TABLE 5 OMITTED] (1986) classification scheme, because the leader behavior of interest does not have a main or simple main effect on the criterion variable.

Finally, notice that the last numerical entry in the “General Satisfaction” column ([Beta] = .27) has an (ncc) symbol next to it. This symbol means that even though the moderating effect of group cohesiveness on the relationship between intellectual stimulation and general satisfaction was not classifiable using Howell et al.’s (1986) criteria, the interaction between the leader behavior and the substitute variable took the form of a “classic” moderating effect; that is, one in which the slope of the relationship between the leader behavior and criterion variable was significantly positive at one level of the moderator, but significantly negative at the other level of the moderator. Although these “classic” interactions do not fit readily into the classification scheme developed by Howell et al. (1986), they are nevertheless potentially important to leaders, since they indicate conditions under which the effects of the leader behavior may change the most dramatically. For this reason, we have decided to identify them explicitly.

With this in mind, several general comments regarding the overall pattern of moderating effects in Table 5 are in order. First, it bears noting that of the 858 possible interaction effects (6 leader behaviors x 13 substitutes variables x 11 criterion variables = 858 possible interactions) examined in this study, 69 significant moderators were actually observed. This represents about 8% of the possible cases examined, which is somewhat higher than would be expected by chance. However, of the 69 moderating effects detected, only 18 (or 2.1% of the 858 examined) met the criteria for “substitutes,” “neutralizers,” or “enhancers” specified by Howell et al. (1986). Of the remaining 51 interaction effects detected: (a) 4 (5.8% of the 69) represented “classical” interaction effects, in which the relationship between the leader behavior of interest and the criterion variable were significantly positive at one level of the moderator, and significantly negative at the other level of the moderator, (b) 21 (30.4% of the 69) were cases in which the overall relationship between the leader behavior and the criterion variable was significant at one level of .the moderator variable, but not at the other level of the moderator, and (c) 26 (37.8% of the 69) were cases in which the relationship between the leader behavior and the criterion variable was slightly positive (but not significant) at one level of the moderator, and slightly negative (but not significant) at the other level of the moderator.

Second, it appears that, on average, the substitutes tended to moderate the effects of the leader behaviors more on the self-report criterion variables than they did on the behavioral criterion variables. Approximately two-thirds (65.2%) of the 69 moderating effects, and every one of the 18 effects that met the criteria for “substitutes,” “neutralizers,” or “enhancers” specified by Howell et al. (1986), involved the self-report criterion measures (i.e., satisfaction, commitment, trust, role clarity and role conflict). Indeed, none of the moderating effects involving the behavioral criterion variables met Howell et al.’s (1986) criteria. Of the self-report criterion measures, trust was associated with the greatest number of moderating effects (17), followed by organizational commitment (13), role clarity (7) and general satisfaction and role conflict (both with 4 each). Of the performance measures, in-role performance (with a total of 7), sportsmanship (with 5), courtesy and civic virtue (with 4 each), were associated with the greatest number of moderating effects, and altruism with the least (just 1).

Third, the effects of some of the transformational leadership behaviors on the criterion measures were influenced by more moderators than others. The two leader behaviors that were moderated the most often were articulating a vision and fostering the acceptance of group goals (each with 15 of the 69 interactions that occurred or, when taken together, approximately 43% of the total), followed closely by providing an appropriate model and intellectual stimulation (each with 12 of the 69 interactions, or about 17%), and high performance expectations (with 10 of the 69 interactions, or about 14%). On the other hand, the leader behavior that was moderated the least was individualized support (with only 5 of the 69 interactions, or about 7.2%).

Finally, although all of the 13 substitutes moderated the impact of at least one leader behavior, some substitutes produced far more moderating effects than others. Organizational formalization was the most frequent moderator (accounting for 9 of the 69 moderated relationships), followed by intrinsically satisfying tasks (accounting for 8 of the 69 moderated relationships), group cohesiveness (accounting for 7), and professional orientation, need for independence, and spatial distance (each accounting for 6 of the 69 moderated relationships). Thus, it would appear that the most important category of moderators in our sample relate to organizational characteristics, and that of these, organizational formalization, group cohesiveness, and spatial distance, were the most important individual moderators.

Moderators of articulating a vision. As indicated in Table 5, 15 (10.5%) of the possible 143 interactions were significant in the case of the relationships between articulating a vision and the 11 criterion variables. Follow-up analyses of these interactions indicated that eleven of these interactions did not fit the criteria specified by Howell et al. (1986), and therefore were not classifiable in their system. Interestingly, advisory/staff support served as a “classic” moderator of the relationship between articulating a vision and employees’ trust in their leader. More specifically, the relationship between articulating a vision and trust was significantly positive at low levels of advisory/staff support, but was significantly negative at high levels of advisory/staff support.

Of the remaining four interactions: (a) intrinsically satisfying tasks served as an enhancer of the positive relationship between articulating a vision and general satisfaction, (b) group cohesiveness served as substitute for the positive relationship between vision articulation and organizational commitment, and (c) ability, experience, training, and knowledge served as a substitute for and need for independence served as a neutralizer of the positive relationship between vision articulation and role clarity. Taken together, these findings indicate that about one-fourth (4 of 15 or 27%) of the significant interactions observed between articulating a vision and the substitutes variables are consistent with the criteria specified by Howell et al. (1986).

Moderators of providing an appropriate model. Twelve (8.4%) of the possible 143 interactions between providing an appropriate model and the 13 substitutes for leadership on the 11 criterion variables were found to be significant. However, as indicated in Table 5, only six of these interactions were consistent with the criteria specified by Howell et al. (1986). An examination of the follow-up regression analyses indicated that: (a) advisory/staff support served as an enhancer of the positive relationship between providing an appropriate model and employees’ general satisfaction, and (b) professional orientation, indifference to organizational rewards, and rewards outside the leader’s control served as enhancers, while organizational inflexibility served as a neutralizer, and task feedback was a substitute of the positive relationship between providing an appropriate model and employees’ trust in their leader.

Moderators of fostering the acceptance of group goals. Fifteen (10.5%) of the 143 interactions between fostering the acceptance of group goals and the 13 substitutes for leadership examined in this study were found to be significant. Of these interactions, only two (1.4% of the total possible) were found to be consistent with Howell et al.’s (1986) criteria. Routine tasks served as a substitute and spatial distance served as an enhancer of the positive relationship between fostering the acceptance of group goals and employees’ trust in their leader. In addition, both ability, experience, training, and knowledge, and need for independence served as “classic” moderators of the relationship between fostering the acceptance of group goals and role clarity. The relationship between fostering the acceptance of group goals and role clarity was significantly positive when subordinates perceived they possessed high levels of ability, experience, training, and knowledge, or had a high need for independence; but was significantly negative when subordinates perceived that they possessed low levels of ability, experience, training, and knowledge, or had a low need for independence.

Moderators of high performance expectations. Ten (7%) of the interactions between high performance expectations and the substitutes for leadership variables were found to be significant; three of these (three of the 143 possible, or 2.1%) met Howell et al.’s criteria. Spatial distance served as an enhancer of the positive relationship between high performance expectations and employees’ perceptions or role conflict, while task feedback served as a neutralizer of this relationship. Thus, task feedback tended to assuage the positive effects of high performance expectations on role conflict, while spatial distance from the leader tended to magnify them. In addition, advisory/staff support served as a neutralizer of the positive relationship between high performance expectations and employee’s trust in their leader.

Moderators of individualized support. There were only five moderators (3.5%) of the effects of individualized support on the 11 criterion variables. Of these moderators, two (or 1.4% of the 143 possible) were consistent with the criteria specified by Howell et al. (1986). Indifference to organizational rewards served as a neutralizer of the positive relationship between individualized support and employee trust, and routine tasks served as a neutralizer of the negative relationship between individualized support and employee perceptions of role conflict.

Moderators of intellectual stimulation. There were 12 moderators (8.4%) of the effects of leader intellectual stimulation on the subordinate criteria variables; but only one of these (less than 1% of the total possible) met the criteria specified by Howell et al. (1986). Subordinates’ professional orientation served as a neutralizer of the positive relationship between intellectual stimulation and role conflict; indicating that the relationship between intellectual stimulation and role conflict was positive for subordinates who lacked a professional orientation, but was not significant for subordinates who were professionally oriented. In addition, one of the moderating effects took the form of a “classic” interaction. At low levels of group cohesiveness, intellectual stimulation was negatively related to general satisfaction, while at high levels intellectual stimulation was positively related to this criterion variable.

Discussion

The primary objective of the study reported here was to explore the relative contributions of transformational leader behaviors and substitutes for leadership to employee attitudes, role perceptions, and “in-role” and “extra-role” performance. This was accomplished by examining the aggregate effects of the leader behaviors and substitutes for leadership on the subordinate criterion variables, as well as their individual main and interactive effects. In discussing the overall pattern of results, we first examined the aggregate effects of the leader behaviors and the substitutes for leadership on employee attitudes, role perceptions and performance. Next, we examined the unique effects of the six transformational leader behaviors and the 13 substitutes for leadership on the eleven criterion variables. Finally, we examined the moderating effects of the leadership substitutes on the leader behavior-criterion variable relationships.

Aggregate Effects on Criterion Variables

An examination of the aggregate effects of the transformational leader behaviors and leadership substitutes on subordinates’ attitudes, role perceptions, and performance produced a number of interesting findings. First, the data reported in Table 3 indicate that adding the 13 substitutes for leadership to the six transformational leader behaviors included in this study significantly improved the proportion of variance accounted for in the 11 criterion variables. Second, the data also indicate that although the transformational leader behaviors contributed to the variance in every one of the criterion variables, with the exception of employee trust and courtesy, the amount of variance accounted for in the remaining (nine) criterion variables by the substitutes variables, was always greater than the amount of variance accounted for by the leader behaviors. This was especially true in the case of the employee attitudes (other than trust), and role perceptions. Third, with a few exceptions, a substantial portion of the variance explained in employees’ attitudes and role perceptions is shared by both of the leader behavior and substitute categories.

Taken together, these findings suggest the importance of including both the transformational leader behaviors and substitutes for leadership if one wishes to obtain a comprehensive understanding of the antecedents of the employee criterion variables examined in this study. Indeed, the results of the aggregate analysis indicate that because of the large proportion of the shared variance between the transformational leader behaviors and substitutes for leadership, it is essential to include the substitutes variables in any test of the effects of transformational leadership. Otherwise, the parameter estimates obtained will be biased due to misspecification of the regression model.

Individual Effects on Criterion Variables

As indicated in Table 4, one transformational leader behavior (individualized support) appears to be a particularly important determinant of employee attitudes, role perceptions and behaviors. Individualized support was found to be positively related to employees’ satisfaction, trust in their leader, role clarity, in-role performance, altruism, conscientiousness, sportsmanship, courtesy, and civic virtue; and negatively related to employees’ perceptions of role conflict. These effects suggest that employees who perceive their leaders to provide individualized support generally trust their leaders more, and are better sports, more satisfied, productive, altruistic, conscientious, courteous, experience more role clarity and less role conflict, and exhibit more civic virtue, than are employees who perceive their leaders to provide less support.

According to the propositions of many transformational leadership approaches (cf. Bass, 1985; Bennis & Nanus, 1985; Boal & Bryson, 1988; Burns, 1978; Conger & Kanungo, 1987; House, 1977; Tichy & DeVanna, 1986), leaders who articulate a vision should have positive effects on employee attitudes, role clarity, and extra-role behaviors. Consistent with these expectations, articulating a vision was found to be positively related to employees’ general satisfaction, organizational commitment, role clarity, and one dimension of organizational citizenship behaviors (sportsmanship). Somewhat surprisingly, however, articulating a vision was not related to employee trust, or to other dimensions of employees’ citizenship behaviors.

Leaders who were perceived by their subordinates to provide an appropriate model tended to have functional effects on employee attitudes and role perceptions, but no effects on employee behaviors. Providing an appropriate model was positively related to employee satisfaction and trust in one’s leader, and negatively related to subordinates’ perceptions of role conflict. Thus, leaders who provide appropriate models increase employees’ trust and satisfaction, and decrease the amount of conflict they experience in their jobs.

Leaders who communicate high performance expectations, and those who provide intellectual stimulation, had some effects on employee outcome variables; but with the exception of the positive effect of high performance expectations on courtesy, the effects that they did have were not functional to the organization. Both high performance expectations and intellectual stimulation tended to increase role conflict. In addition, high performance expectations tended to decrease employees’ general satisfaction. Although these findings are contrary to the predictions of a variety of transformational models of leadership, they are not totally unanticipated. For example, Podsakoff et al. (1990) reported a negative relationship between intellectual stimulation and employee trust and satisfaction, and a negative relationship between high performance expectations and employee trust. They noted that one possible explanation of the dysfunctional effects of intellectual stimulation might be due to the fact that, “Although intellectual stimulation may produce desirable effects in the long run, it may be that in the short run, leaders who continually urge or exhort followers to search for new and better methods of doing things create ambiguity, conflict, or other forms of stress in the minds of those followers.” (p. 135). Thus, it is possible that when leaders provide too much intellectual stimulation to their subordinates, they also may increase the levels of role conflict their subordinates experience.

The reason for the negative relationship between high performance expectations and employee satisfaction, and the positive relationship between this form of transformational leader behavior and employee perceptions of role conflict, are not as straightforward. However, House (1977) has suggested that in order to be effective, high performance expectations must not only be communicated to followers in an unambiguous manner, but the leader must also let followers know that the leader has confidence in their ability to meet those expectations. Otherwise, it may appear that the leader is continually expressing a desire for higher and higher levels of performance on the followers’ part, without really expecting them to be able to meet these higher expectations. Unfortunately, although our measure taps the leader’s high expectations, it does not appear to tap his/her expressions of confidence that the followers can meet these expectations. Thus, it is possible that the dysfunctional consequences of high performance expectations observed in this study would be alleviated in future research, if the measure of high performance expectations was broadened to include expressions of confidence by the leader in the followers’ ability to meet these expectations.

In addition to the leader behaviors, all of the substitutes for leadership were found to independently influence at least one criterion variable. The pattern of main effects indicates that, of the subordinate characteristics, indifference to organizational rewards had the greatest number of effects (eight), followed by ability, experience, training, and knowledge (with five), and professional orientation (with three). The effects of both professional orientation and indifference to organizational rewards were generally dysfunctional. Professional orientation tended to be associated with reduced levels of employee satisfaction, and courtesy, but increased role conflict. Similarly, employees’ indifference to organizational rewards tended to reduce employees’ satisfaction, commitment, trust in their leader, in-role performance, conscientiousness, sportsmanship, and civic virtue; and also increased role conflict. However, the effects of employees’ ability, experience, training, and knowledge, were somewhat more complex. On the one hand, employees who perceived that they possess more ability, experience, training, and knowledge, tended to express more trust in their leader and role clarity, and exhibit more civic virtue than employees who do not perceive they possess as much ability, experience, training, and knowledge. On the other hand, higher levels on this predictor variable also tended to result in more role conflict on the part of employees, and reduced sportsmanship. Taken together, these findings appear to strongly reinforce the notion that professionally oriented employees present significant challenges to organizational leaders (cf. Von Glinow, 1988), and that, in order to be effective, leaders must identify rewards that employees value.

Of the task characteristics, routine tasks were found to have the greatest number of significant main effects (with nine), followed by both task feedback and intrinsically satisfying tasks (with six each). Task feedback and intrinsically satisfying tasks primarily affected employee attitudes and perceptions, but task routinization influenced all of the behavioral criterion variables, as well as employee attitudes and role perceptions. Task feedback and intrinsically satisfying tasks generally had functional effects on subordinate criterion variables. Both of these predictor variables tended to increase employee satisfaction, and role clarity, and decrease employees’ perceptions of role conflict. In addition, task feedback tended to increase employees’ trust in their leader and their in-role performance and civic virtue; and intrinsically satisfying tasks tended to increase organizational commitment, altruism, and sportsmanship.

However, the effects of task routinization were generally negative. For example, although higher levels of task routinization tended to increase role clarity, and decrease role conflict, employees with more routinized tasks also engaged in substantially less in-role and citizenship behaviors. These latter findings relating task routinization to citizenship behaviors appear to support Farh, Podsakoff and Organ’s (1990) conceptual arguments regarding the potential importance of job characteristics as determinants of organizational citizenship behaviors, and suggest that additional attention to task properties may prove valuable in future OCB research.

Finally, in the case of organizational characteristics, organizational formalization had the greatest number of effects (with seven), followed by group cohesiveness (with six), and organizational inflexibility and rewards outside the leader’s control (with four each). Interestingly, organizational formalization had many of the same effects as task routinization. For example, like routine tasks, organizational formalization increased employees’ perceptions of role clarity, and decreased role conflict, altruism, conscientiousness, sportsmanship, and civic virtue. In addition, organizational formalization also decreased employees’ trust in their leader. One possible explanation for the similarity in the results regarding the effects of these two predictor variables on the citizenship behaviors relates to the fact that when tasks are relatively routinized, and the rules and regulations of the organization are relatively formal in nature, employees may perceive their behaviors to be relatively prescribed for them, and they do not feel that they can go beyond their well defined areas of responsibility. Of course, another possible explanation is that, when employees feel that their behavior becomes over-prescribed by the organization and represents a threat to their independence, they may resist these threats by choosing not to do what may benefit the organization. Brehm (1966), for example, has noted that such forms of psychological reactance are one of the main responses to threats to one’s independence and freedom.

However, if this reactance explanation is accurate, one probably would assume that employees’ perceptions of organizational inflexibility would also decrease citizenship behavior, since inflexible work rules and procedures might be expected to restrict employees’ perceptions of personal freedom and independence of action. Contrary to this expectation, organizational inflexibility actually increased employees’ general satisfaction and commitment to the organization, and had no effects on employee in-role or extra-role behaviors. These effects are difficult to reconcile with those produced by organizational formalization. However, one possible explanation relates to the perceptions of employees regarding the legitimacy of the organizational rules and procedures. Building on the work of Katz and Kahn (1966), Baum and Youngblood (1975) predicted that organizational control policies that are perceived as legitimate, and compulsory for all employees, would be more effective than laissez-faire policies that are noncompulsory. Consistent with their expectations, they found that organizational control policies that were perceived as legitimate, and compulsory for all, produced more functional effects than did noncompulsory policies.

Finally, given the generally functional effects of group cohesiveness found in this study, it would appear that the recent rejuvenation of interest among researchers in work teams and “team building” is justified. Group cohesiveness was positively related to employees’ general satisfaction, commitment, trust, sportsmanship, and courtesy, and negatively related to employees’ perceptions of role conflict; suggesting that cohesive groups generally improve employees’ attitudes, role perceptions, and citizenship behaviors.

Moderating Effects of the Substitutes Variables

An examination of Table 5 indicates that only four variables moderated the impact of a transformational leadership behavior on a criterion variable in a manner consistent with the definition of a substitute [Figures 1 (a) or (f)], as specified by Howell et al. (1986). Group cohesiveness substituted for the impact of articulating a vision on organizational commitment; task feedback substituted for the impact of providing an appropriate model on employees’ trust in their leader; routine tasks substituted for fostering the acceptance of group goals on employees’ trust in their leader; and ability, experience, training, and knowledge, substituted for articulating a vision on employees’ perceptions of role clarity; all in the manner shown in Figure 1 (a). If replicated by future research, this would suggest that articulating a vision may be less important when followers have more ability, experience, training, and knowledge, or work in a cohesive group; providing an appropriate model may be less important when followers work on tasks that give them a high degree of feedback; and fostering the acceptance of group goals may be less important when followers perform routine tasks.

An examination of Table 5 also indicates that six variables met the criteria of a neutralizer, as specified by Howell et al. (1986). Almost half (three) of these neutralized the effect of a leader behavior on role conflict. Task feedback served as a partial neutralizer of the positive effect of high performance expectations on role conflict; professional orientation served as a full neutralizer of the positive effect of intellectual stimulation on role conflict; and routine tasks served as a full neutralizer of the negative effect of individualized support on role conflict. (The difference between “full” and “partial” neutralizers is that “full” neutralizers completely eliminate the significant positive or negative effect of the leader behavior, while “partial” neutralizers simply reduce the impact of the leader behavior, although its effect remains significant.) Of the three remaining neutralizing effects observed, need for independence served as a full neutralizer of the positive effect of articulating a vision on role clarity; indifference to organizational rewards served as a partial neutralizer of the positive effect of individualized support on employees’ trust in their leader; and organizational inflexibility served as a partial neutralizer of the positive effect of providing an appropriate model on employees’ trust in their leader.

Eight variables moderated the relationship between transformational leader behaviors and subordinate criterion variables in a manner consistent with Howell et al.’s (1986) definition of an enhancer. The majority of these effects (six) took the form specified by Figure 2(b). Interestingly, half of these relationships occurred when trust in one’s leader served as the criterion variable, and four of the interactions resulted from cases where providing an appropriate model was the leader behavior of interest. Rewards outside the leader control, employees’ indifference to organizational rewards, and professional orientation, all served as partial enhancers of the positive relationship between providing an appropriate model and subordinates’ trust in their leaders. In addition, advisory/staff support served as a full enhancer of the positive relationship between providing an appropriate model and general satisfaction; intrinsically satisfying tasks served as a full enhancer of the positive relationship between articulating a vision and general satisfaction; spatial distance served as a full enhancer of the positive relationship between fostering the acceptance of group goals and subordinates’ trust in their leader; and spatial distance served as a full enhancer of the positive relationship between high performance expectations and employees’ perceptions of role conflict.

Finally, an examination of Table 5 also indicates that 51 (73.9%) of the 69 significant interactions did not meet the requirements for substitutes, neutralizers, or enhancers specified by Howell et al.(1986), primarily because their corresponding leader behavior main effects were not significant.

Conclusions

The results of the present study do not provide strong evidence for the moderating effects of Kerr and Jermier’s (1978) substitutes for leadership on the relationships between transformational leader behaviors and subordinate criterion variables. Although the number of interaction effects observed (69, or 8% of a possible 858) was somewhat greater than one would expect by chance alone, only 18 (or approximately 2.1% of the total) of these interactions behaved in a manner consistent with that specified by Kerr and his colleagues (cf. Howell et al., 1986). The vast majority of the remaining interactions were ones in which the leader behavior of interest had no significant overall effect on the subordinate criterion variables. Thus, strictly speaking, the results provided little or no support for the universal prediction that substitutes for leadership moderate the impact of transformational leadership on the job attitudes, role perceptions, or performance of subordinates.

However, it would be a mistake to conclude from this that it is safe to ignore the substitutes when examining the effects of transformational leader behaviors. We believe that this conclusion would be misguided because several of the substitutes for leadership examined in this study were significantly correlated with both the transformational leadership behaviors and the criterion variables. Functionally, this means that any structural model that examines the effects of these leader behaviors on the subordinate criterion variables that does not include the substitutes variables would be misspecified. Thus, to ignore the substitutes variables when examining the effects of transformational leadership behaviors would lead to biased parameter estimates of the effects of these leader behaviors on the subordinate criterion variables, and generally obscure their importance relative to these “contextual” variables. This implies that much of what we know about the impact of transformational leadership on employee attitudes, perceptions, and performance, may be subject to qualification.

In addition, the fact that the substitutes for leadership and transformational leader behaviors are correlated with each other suggests that future research ought to investigate why this is the case. There are three reasons why the substitutes variables might be correlated with leader behaviors. First, it is possible that the relationships between the substitutes and the leader behaviors might be spurious, and be caused by other unrecognized factors. Second, it is possible that the substitutes influence or constrain a leader’s behavior in some way. A final possibility, noted by Kerr (1977), Howell et al. (1990), and Podsakoff, MacKenzie, and Fetter (1993), among others, is that leaders can influence the substitutes variables. If so, it is possible that managers actually have a stronger impact than previously suggested because they can influence subordinates not only directly through their behavior, but also by shaping the context in which the subordinates work.

Finally, the findings of this study suggest that leaders need to have a better understanding of those contextual variables that influence subordinate attitudes, role perceptions, and performance, and how to influence these contextual variables. An examination of the results of this study, when taken together with the findings reported previously by Podsakoff, Niehoff, MacKenzie and Williams (1993) and Podsakoff, MacKenzie and Fetter (1993), suggests that some substitutes that may be particularly good candidates in this regard include indifference to organizational reward, routine tasks, intrinsically satisfying tasks, organizational formalization, and group cohesiveness, because these substitutes generally have had several effects across all three of these studies. Thus, future research should focus additional attention on the effects of these substitutes, as well the things that leaders can do to influence them.

Notes

1. According to Burns (1978), transactional leadership behavior stems from the exchange process between leaders and their subordinates wherein a leader provides rewards in return for the subordinate’s effort.

2. A portion of the data was previously published in a study by Podsakoff and MacKenzie (1994a), which examined the psychometric properties of the substitutes for leadership scales. However, that study did not include transformational leader behaviors or examine their impact on the criterion variables included in this study, nor did it examine the impact of substitutes for leadership on these criterion variables, or the interaction of the substitutes and transformational leader behaviors.

References

Avolio, B.J. & Bass, B.M. (1988). Transformational leadership, charisma, and beyond. Pp. 29-49 in J.G. Hunt, B.R. Baliga, H.P. Dachler & C.A. Schriesheim (Eds.), Emerging leadership vistas. Lexington, MA: Lexington Books.

Bagozzi, R.P., Yi, Y. & Phillips, L.W. (1991). Assessing the construct validity in organizational research. Administrative Science Quarterly, 36: 421-458.

Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall.

Bass, B.M. (1985). Leadership and performance beyond expectation. New York: Free Press.

Bass, B.M., Avolio, B.J. & Goodheim, L. (1987). Biography and the assessment of transformational leadership at the world class level. Journal of Management, 13: 7-19.

Bass, B.M., Waldman, D.A., Avolio, B.J. & Bebb, M. (1987). Transformational leadership and the falling dominoes effect. Group and Organization Studies, 12: 73-87.

Baum, J.F. & Youngblood, S.A. (1975). Impact of an organizational control policy on absenteeism, performance, and satisfaction. Journal of Applied Psychology, 60: 688-694.

Bennis, W. & Nanus, B. (1985). Leaders: The strategies for taking charge. New York: Harper & Row.

Bentler, P.M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107: 238-246.

Boal, K.B. & Bryson, J.M. (1988). Charismatic leadership: A phenomenological and structural approach. Pp. 528 in J.G. Hunt, B.R. Baliga, H.P Dachler & C.A. Schriesheim (Eds.), Emerging leadership vistas. Lexington, MA: Lexington Books.

Bollen, K.A. (1989). A new incremental fit index for general structural equation models. Sociological Methods and Research, 17:303-316.

Bradford, D.L. & Cohen, A.R. (1984). Managing for excellence: The guide to developing high performance in contemporary organizations. New York: Wiley.

Brehm, J.W. (1966). A theory of psychological reactance. New York: Academic Press.

Bryman, A. (1992). Charisma and leadership in organizations. London: Sage.

Burns, J.M. (1978). Leadership. New York: Harper & Row.

Cohen, J. & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum.

Conger, J.A. & Kanungo, R.N. (1987). Toward a behavioral theory of charismatic leadership in organizational settings. Academy of Management Review, 12: 637-647.

Cote, J.A. & Buckley, R. (1987). Estimating trait, method, and error variance: Generalizing across 70 construct validation studies. Journal of Marketing Research, 24:315-318.

Dobbins, G.H. & Zaccaro, S.J. (1986). The effects of group cohesion and leader behavior on subordinate satisfaction. Group and Organizational Studies, 11: 203-219.

Farh, J.L., Podsakoff, P.M. & Cheng, B.S. (1987). Culture-free leadership effectiveness versus moderators of leadership behavior: An extension and test of Kerr and Jermier’s “substitutes for leadership” model in Taiwan. Journal of International Business Studies, 18: 43-60.

Farh, J.L., Podsakoff, P.M. & Organ, D.W. (1990). Accounting for organizational citizenship behavior: Leader fairness and task scope versus satisfaction. Journal of Management, 16: 705-721.

Fornell, C. & Larcker, D.F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 28: 39-50.

Gillet, B.J. & Schwab, D.P. (1975). Convergent and discriminant validities of the corresponding Job Description Index and Minnesota Satisfaction Questionnaire scales. Journal of Applied Psychology, 60:313-317.

House, R.J. (1977). A 1976 theory of charismatic leadership. Pp. 000-000 in J.G. Hunt & L.L. Larson (Eds.), Leadership: The cutting edge. Carbondale, IL: Southern Illinois University Press.

House, R.J., Schuler, R.S. & Levanoni, E. (1983). Role conflict and ambiguity scales: Reality or artifacts? Journal of Applied Psychology, 68: 334-337.

House, R.J., Spangler, W.D. & Woycke, J. (1991). Personality and charisma in the U.S. presidency: A psychological theory of leadership effectiveness. Administrative Science Quarterly, 36: 364-396.

House, R.J., Woycke, J. & Fodor, E.M. (1988) Perceived behavior and effectiveness of charismatic and non-charismatic U.S. Presidents. Pp. 000-000 in J. Conger & R. Kanungo (Eds.), Charismatic leadership and management. San Francisco: Jossey-Bass.

Howell, J.P., Bowen, D.E., Dorfman, P.W., Kerr, S. & Podsakoff, P.M. (1990). Substitutes for leadership: Effective alternatives to ineffective leadership. Organizational Dynamics, (Summer): 20-38.

Howell, J.P. & Dorfman, P.W. (1981). Substitutes for leadership: Test of a construct. Academy of Management Journal, 24:714-728.

Howell, J.P. & Dorfman, P.W. (1986). Leadership and substitutes for leadership among professional and non professional workers. Journal of Applied Behavioral Science, 22: 29-46.

Howell, J.P., Dorfman, P.W. & Kerr, S. (1986). Moderator variables in leadership research. Academy of Management Review, 11: 88-102.

Howell, J.M. & Frost, P.J. (1989). A laboratory study of charismatic leadership. Organizational Behavior and Human Decision processes, 43: 243-269.

Jackson, S.E. & Schuler, R.S. (1985). A meta-analysis and conceptual critique of research on role ambiguity and role conflict in work settings. Organizational Behavior and Human Decision Processes, 36: 16-78.

Jermier, J.M. & Berkes, L. (1979). Leader behavior in a police command bureaucracy: A closer look at a quasimilitary model. Administrative Science Quarterly, 24: 1-23.

Joreskog, K.G. & Sorbom, D. (1993). LISREL VIII: Analysis of linear structural relationships by maximum likelihood, instrumental variables and least squares methods, 8th ed. Morresville, IN: Scientific Software.

Katz, D. & Kahn, R.L. (1966). The social psychology of organizations. New York: Wiley.

Kerr, S. (1977). Substitutes for leadership: Some implications for organizational design. Organization and Administrative Sciences, 8:135-146.

Kerr, S. & Jermier, J.M. (1978). Substitutes for leadership: Their meaning and measurement. Organizational Behavior and Human Performance, 22: 375-403.

Kouzes, J.M. & Posner, B.Z. (1987). The leadership challenge. San Francisco: Jossey-Bass.

Locke, E.A. & Latham, G.P. (1990). A theory of goal setting & task performance. Englewood Cliffs, NJ: Prentice-Hall.

MacKenzie, S.B., Podsakoff, P.M. & Fetter, R. (1991). Organizational citizenship behavior and objective productivity as determinants of managerial evaluations of salespersons’ performance. Organizational Behavior and Human Decision Processes, 50: 123-150.

Mathieu, J.E. & Zajac, D.M. (1990). A review and meta-analysis of the antecedents, correlates, and consequences of organizational commitment. Psychological Bulletin, 108:171-194.

Moorman, R.H. (1991). Relationship between organizational justice and organizational citizenship behaviors: Do fairness perceptions influence employee citizenship? Journal of Applied Psychology, 76: 845-855.

Moorman, R.H. (1993). The influence of cognitive and affective based job satisfaction measures on the relationship between satisfaction and organizational citizenship behavior. Human Relations, 46: 759-775.

Moorman, R.H., Niehoff, B.P. & Organ, D.W. (1993). Treating employees fairly and organizational citizenship behavior: Sorting the effects of job satisfaction, organizational commitment, and procedural justice. Employee Responsibilities and Rights Journal, 6: 209-225.

Mowday, R.T., Steers, R.M. & Porter, L.W. (1979). The measurement of organizational commitment. Journal of Vocational Behavior, 14: 224-247.

Niehoff, B.P. & Moorman, R.H. (1993). Justice as a mediator of the relationship between methods of monitoring and organizational citizenship behavior. Academy of Management Journal, 36: 527-556.

Nunnally, J. (1978). Psychometric theory, 2nd ed. New York: McGraw-Hill.

Organ, D.W. (1988). Organizational citizenship behavior: The good soldier syndrome. Lexington, MA: D.C. Heath.

Organ, D.W. (1990). The motivational basis of organizational citizenship behavior. Pp. 43-72 in B.M. Staw & L.L. Cummings (Eds.), Research in organizational behavior, Vol. 12. Greenwich, CT: JAI Press.

Podsakoff, P.M. & MacKenzie, S.B. (1989). A second generation measure of organizational citizenship behavior. Indiana University.

Podsakoff, P.M. & MacKenzie, S.B. (1994a). An examination of the psychometric properties and the nomological validity of some revised and reduced “substitutes for leadership” scales. Journal of Applied Psychology, 79: 702-713.

Podsakoff, P.M. & MacKenzie, S.B. (1994b). Organizational citizenship behavior and sales unit effectiveness. Journal of Marketing Research, 31:351-363.

Podsakoff, P.M., MacKenzie, S.B., Moorman, R.H. & Fetter, R. (1990). Transformational leader behaviors and their effects on followers’ trust in leader, satisfaction, and organizational citizenship behaviors. Leadership Quarterly, 1: 107-142.

Podsakoff, P.M., MacKenzie, S.B. & Fetter, R. (1993). Substitutes for leadership and the management of professionals. Leadership Quarterly, 4: 1-44.

Podsakoff, P.M., Niehoff, B.P., MacKenzie, S.B. & Williams, M.L. (1993). Do substitutes for leadership really substitute for leadership? An empirical examination of Kerr and Jermier’s situational leadership model. Organizational Behavior and Human Decision Processes, 54: 1-44.

Podsakoff, P.M. & Organ, D.W. (1986). Self reports in organizational research: Problems and prospects. Journal of Management, 12:531-544.

Podsakoff, P.M., Todor, W.D., Grover, R.A. & Huber, V.L. (1984). Situational moderators of leader reward and punishment behavior: Fact or fiction? Organizational Behavior and Human Performance, 34:21-63.

Porter, L.W., Steers, R.M., Mowday, R.T. & Boulian, P.V. (1974). Organizational commitment, job satisfaction, and turnover among psychiatric technicians. Journal of Applied Psychology, 59: 603-609.

Price, J.L. & Mueller, C.W. (1986). Handbook of organizational measurement. Marshfield, MA: Pittman.

Rizzo, J.R., House, R.J. & Lirtzman, S.I. (1970). Role conflict and ambiguity in complex organizations. Administrative Science Quarterly, 15: 150-163.

Roberts, N.C. (1985). Transforming leadership: A process of collective action. Human Relations, 38: 1023-1046.

Schuler, R.S., Aldag, R.J. & Brief, A.P. (1977). Role conflict and ambiguity: A scale analysis. Organizational Behavior and Human Performance, 20:111-128.

Shamir, B., House, R.J. & Arthur, M.B. (1993). The motivational effects of charismatic leadership: A self concept based theory. Organization Science, 4: 1-17.

Sheridan, J., Vredenburgh, D. & Abelson, M. (1984). Contextual models of leadership influence in hospital units. Academy of Management Journal, 27: 57-78.

Stone, E.F. (1988). Moderator variables in research: A review and analysis of conceptual and methodological issues. Pp. 191-229 in K.M. Rowland & G.R. Ferris (Eds.), Research in personnel and human resources management, Vol. 6. Greenwich, CT: JAI Press.

Stone, E.F. & Hollenbeck, J.R. (1984). Some issues associated with the use of moderated regression. Organizational Behavior and Human Performance, 34:195-213.

Stone, E.F. & Hollenbeck, J.R. (1989). Clarifying some controversial issues surrounding statistical procedures for detecting moderator variables: Empirical evidence and related matters. Journal of Applied Psychology, 74: 3-10.

Tansky, J.W. (1993). Justice and organizational citizenship behavior: What is the relationship? Employee Responsibilities and Rights Journal, 6: 195-207.

Tichy, N. & DeVanna, M. (1986). The transformational leader. New York: Wiley.

Trice, H.M. & Beyer, J.M. (1986). Charisma and its routinization in two social movement organizations. Pp. 113164 in B.M. Staw & L.L. Cummings (Eds.), Research in Organizational Behavior, Vol. 8. Greenwich, CT: JAI Press.

Tucker, L.R. & Lewis, C. (1973). The reliability coefficient for maximum-likelihood factor analysis. Psychometrika, 38: 1-10.

Von Glinow, M.A. (1988). The new professionals: Managing today’s high-tech employees. Cambridge, MA.: Ballinger.

Wanous, J.P. (1974). A causal-correlational analysis of the job satisfaction and performance relationship. Journal of Applied Psychology, 59:139-144.

Weiss, D.J. Dawis, R.V., England, G.W. & Lofquist, L.H. (1967). Manual for the Minnesota Satisfaction Questionnaire. Minnesota Studies in Vocational Rehabilitation: XXII. Minneapolis: University of Minnesota, Industrial Relations Center Work Adjustment Project.

Williams, L.J. (1989). Affective and non-affective components of job satisfaction and organizational commitment as determinants as organizational citizenship and in-role behaviors. Unpublished doctoral dissertation. Bloomington, IN: Indiana University.

Yukl, G.A. (1989a). Leadership in organizations, 2nd ed. Englewood Cliffs, NJ: Prentice Hall.

Yukl, G.A. (1989b). Managerial leadership: A review of theory and research. Yearly Review of Management, 15:251-289.

COPYRIGHT 1996 JAI Press, Inc.

COPYRIGHT 2004 Gale Group