Description
Literature (from Latin litterae (plural); letter) is the art of written work. The word literature literally means "things made from letters." Literature is commonly classified as having two major forms—fiction & non-fiction—and two major techniques—poetry and prose.
HRM Study on Two Faces of Transformational Leadership: Empowerment and Dependency
Followers' identification with the leader and the organizational unit, dependence on the leader, and empowerment by the leader are often attributed to transformational leadership in organizations. However, these hypothesized outcomes have received very little attention in empirical studies. Using a sample of 888 bank employees working under 76 branch managers, the authors tested the relationships between transformational leadership and these outcomes. They found that transformational leadership was positively related to both followers' dependence and their empowerment and that personal identification mediated the relationship between transformational leadership and followers' dependence on the leader, whereas social identification mediated the relationship between transformational leadership and followers' empowerment. The authors discuss the implications of these findings to both theory and practice.
There is accumulating evidence that transformational leadership substantially influences employee motivation and performance (e.g., Barling, Weber, & Kelloway, 1996; Dvir, Eden, Avolio, & Shamir, 2002; Lowe, Kroeck, & Sivasubramaniam, 1996; Waldman, Ramirez, House, & Puranam, 2001). However, research on transformational leadership has left several questions unanswered (Yukl, 1998). The present study focuses on two of these questions. First, What are the processes by which transformational leadership exerts its influence on followers? As Yukl (1998) concluded, "A variety of different influence processes may be involved in transformational leadership" (p. 328). In the present study, we suggest that two of the mechanisms by which transformational leaders exert their influence on followers are the creation of personal identification with the leader and social identification with the work unit. The second question we focus on is, What are the possible effects of these different influence processes? Transformational leadership theory suggests that such leadership is likely to result in
Ronit Kark, Department of Psychology and Department of Sociology, Bar-Ilan University, Ramat Gan, Israel; Boas Shamir, Department of Anthropology and Sociology, Hebrew University of Jerusalem, Israel; Gilad Chen, School of Psychology, Georgia Institute of Technology. Ronit Kark is grateful for the financial support of the Israel Foundation Trustees, The Shaine Center for Research in Social Sciences, The Levi Eshkol Institute for Economic, Social and Political Research, and the Department of Psychology at the Hebrew University of Jerusalem. This article was written while Ronit Kark was a visiting scholar at the Institute for Research on Women and Gender and at the Department of Psychology, University of Michigan.
growth, independence, and empowerment of followers (Bass, 1985; Dvir et. al, 2002). However, influential leadership can also lead to the weakening of followers and their dependency on the leader (Howell, 1988). Dependence on the leader has several manifestations. It implies that the subordinate is limited in his or her ability to proceed with work and make decision without the leader's guidance. Psychologically it means that the subordinate's motivation and self-esteem depend on receiving recognition and approval from the leader. Empowerment, in contrast, connotes independence and autonomy. An empowered person is selfmotivated and believes in his or her ability to cope and perform successfully. In the present study we argue that personal identification with the leader and social identification with the work unit are intermediate outcomes that help explain why transformational leadership can result in two seemingly contradictory outcomes: followers' dependence and empowerment. Specifically, we propose that the relationship between transformational behaviors of leaders and followers' dependence is mediated by followers' personal identification with the leader whereas the relationship between transformational behaviors and followers' empowerment is mediated by followers' identification with the work group. The model of the relationships proposed in the present research is summarized in Figure 1.
Theory and Hypotheses Transformational Leadership and Followers' Identification
Transformational leadership has been presented in the literature as different from transactional leadership. Whereas transactional leadership was defined on the basis of the influence process
246
TWO FACES OF TRANSFORMATIONAL LEADERSHIP
247
Figure 1.
Summary of study hypotheses. H
Hypothesis.
underlying it, as an exchange of rewards for compliance, transformational leadership was defined on the basis of its effects, as transforming the values and priorities of followers and motivating them to perform beyond their expectations (Yukl, 1998). Researchers proposed that transformational leadership behaviors comprise four components: inspirational motivation, idealized influence, individualized consideration, and intellectual stimulation. The first two components represent the notion of "charisma" (Avolio, Bass, & Jung, 1999; Bass, 1985) and are similar to behaviors specified in theories of charismatic leadership (Conger & Kanungo, 1998). Inspirational motivation includes the creation and presentation of an attractive vision of the future, the use of symbols and emotional arguments, and the demonstration of optimism and enthusiasm. Idealized influence includes behaviors such as sacrificing for the benefit of the group, setting a personal example, and demonstrating high ethical standards. The third component, individualized consideration, includes providing support, encouragement, and coaching to followers. The fourth component, intellectual stimulation, involves behaviors that increase awareness of problems and challenge followers to view problems from new perspectives. Previous research has shown that these transformational behaviors are related to leadership effectiveness (Lowe et al., 1996) and high employee performance (Barling et al., 1996; Dvir et al., 2002). In the present study, consistent with previous research (e.g., Shamir, Zakay, Breinen, & Popper, 1998), we conceptualize transformational leadership as a group-level construct. Specifically, transformational leadership behaviors are directed at the leader's group and thus are a form of ambient stimuli (cf. Hackman, 1992) that influences the group as a whole, as well as individuals within the group. A review of transformational and charismatic leadership theories suggests that such leadership may achieve its effects partly through the creation of followers' identification with the leader (personal identification) and with the work group (social identification; Yukl, 1998). Thus, we conceptualize personal identifica-
tion and social identification as proximal individual-level outcomes through which transformational leadership (a group-level construct) influences more distal individual-level outcomes (i.e., dependence and empowerment). Personal identification. Borrowing from Pratt's (1998) definition of organizational identification, personal identification is evident when an individual's belief about a person (a leader) becomes self-referential or self-defining. According to early psychoanalytic theories of charismatic leadership (e.g., Kets de Vries, 1988), the influence of charismatic leaders on followers is based on personal identification with the leader, which in turn stems from a transference of earlier identifications with parents. More recent theories also highlight the role of identification with the leader. The charismatic leadership theory of Conger and Kanungo (1998) emphasizes personal identification as a central mechanism through which charismatic leaders influence their followers, suggesting that such influence is based on referent power. Similarly, Shamir, House, and Arthur (1993) posited role modeling as one of the major processes by which charismatic leaders influence followers. Role modeling implies a process by which followers mold their beliefs, feelings, and behavior according to those of the leader. In these theories, personal identification represents one of two (or both) modes of identification suggested by Pratt (1998): (a) evoking followers' self-concept in the recognition that they share similar values with the leader and (b) giving rise to followers' desire to change their self-concept so that their values and beliefs become more similar to those of the leader. The similarity and overlap between the charismatic components of transformational leadership (i.e., idealized influence and inspirational motivation) and leadership behaviors included in theories of charismatic leadership suggests that transformational leaders are likely to arouse personal identification among their followers. Social identification. Another underlying intermediate effect that might account for the exceptional impact of transformational leadership on followers' perceptions and behaviors is social identification. Social identification implies that an individual's belief
248
KARK, SHAMIR, AND CHEN
about a group (or an organization) is self-referential or selfdefining (Pratt, 1998). When individuals identify with a group, they base their self-concept and self-esteem partly on their belonging to the group, and group successes and failures are experienced as personal successes and failures (Mael & Ashforth, 1992). Ashforth and Mael (1989) were among the first to suggest that transformational leadership might influence followers' social identification. The influence of leaders on the social identification of followers is central to Shamir et al.'s (1993) motivational theory of charismatic leadership. They have suggested that the influence of charismatic and transformational leaders is based on their success in connecting followers' self-concept to the mission and to the group, such that followers' behavior for the sake of the group becomes self-expressive. Recent findings support this assertion, showing that leaders who raise followers' identification with the group increase followers' willingness to contribute to group objectives (Shamir et al., 1998; Shamir, Zakay, Breinin, & Popper, 2000). In sum, following the theoretical arguments presented above, we propose that transformational leadership is related to both personal and social identification of followers: Hypothesis 1a: Transformational leadership positively relates to followers' personal identification with the leader. Hypothesis 1b: Transformational leadership positively relates to followers' social identification with the group.
Transformational Leadership and Followers' Empowerment and Dependence
Identification with the leader and the group is important because it has perceptual, motivational, and behavioral consequences (Pratt, 1998). However, personal and social identification might each result in different outcomes (Howell, 1988; Shamir, 1991). In this research we focus on two important outcomes: followers' empowerment and dependence. Despite the centrality of the empowerment and dependence constructs in theories of charismatic and transformational leadership, there has been very little systematic research examining the relationships between transformational leadership and these outcomes. Empowerment. The impact of transformational leadership on followers' performance is often explained as stemming from followers' development and empowerment, which increase both their ability and their motivation (Bass, 1997). The empowerment of followers is often presented as one of the main features that distinguish such leadership from transactional leadership, which does not seek to empower the followers but merely to influence their behavior. To achieve these ends, transformational leadership includes empowering behaviors such as delegation of responsibility to followers, enhancing followers' capacity to think on their own, and encouraging them to come up with new and creative ideas (Dvir et. al., 2002). Empowerment has been defined as a broad and multidimensional concept (Spreitzer, 1995). In this study we adopt a narrower perspective and focus on three indicators of empowerment: selfefficacy, collective efficacy, and organization-based self-esteem (OBSE). Conger and Kanungo (1988) defined empowerment as the process of raising others' self-efficacy perceptions. Self-efficacy is
the individual's belief in his or her ability to successfully perform tasks (Bandura, 1986). According to Conger and Kanungo (1998), transformational leadership behaviors have clear empowering effects on followers in terms of raising their self-efficacy beliefs. Following Shamir et al. (1993), we extend the concept of empowerment suggested by Conger and Kanungo to include collectiveefficacy beliefs and OBSE. Collective efficacy has been conceptualized as being analogous to self-efficacy and is defined as individuals' beliefs that their work group can function effectively and perform its tasks successfully (Bandura, 1986). Identification with the group is often associated with the attribution of positive qualities to the group (Tajfel, 1982), and hence leader behaviors that increase social identification are likely to increase collectiveefficacy beliefs as well. In addition, Shamir et al. (1993) hypothesized that by having high expectations of followers, expressing the leader's belief in followers' abilities, and showing how the mission reflects followers' values, leaders also influence followers' self-esteem. Although Shamir et al. referred to global self-esteem, a more relevant construct is OBSE, which reflects the self-perceived value that individuals have of themselves as organization members acting within an organizational context (Pierce, Gardner, Cummings, &Dunham, 1989). According to Pierce et al. (1989), OBSE is influenced, among other things, by leaders' behaviors toward their subordinates. Members with high OBSE perceive themselves as important, influential, effective, and worthwhile in their organizational units. Furthermore, "employees that possess high levels of OBSE . . . may be described by their peers as being motivated, capable and empowered" (Gardner & Pierce, 1998, p. 50). This supports the inclusion of OBSE as an indicator of empowerment. According to the conceptualization of empowerment used by Spreitzer (1995), the indicators chosen in this study can be seen as related to two dimensions of empowerment: impact and competence. Impact, defined as the degree to which one believes he or she can influence outcomes at work, is evident in followers' OBSE perception of themselves as "making a difference" in the organization and being important and worthwhile (Pierce et al., 1989). The dimension of competence, defined as a belief in one's capacity to perform work activities with skill, is expressed in self-efficacy, collective efficacy, and to a lesser extent, in OBSE perception of efficiency. Dependence. Despite widespread reference to transformational behaviors that can contribute to the empowerment of followers, transformational leadership also includes charismatic behaviors that have been suggested to have the potential of creating dependence on the leader among followers. Theories of charismatic leadership propose that these types of leadership behaviors might result in followers who perceive the leader as extraordinary and exceptional and therefore become dependent on the leader for guidance and inspiration (Yukl, 1998). Psychoanalytic theories, which view the charismatic relationship as a regression to early childhood relationships with parents, imply that such leadership results in increased dependence on the leader (e.g., Kets de Vries, 1988). The potential of charismatic leadership for creating dependence is also recognized by more recent theories, for instance by Conger and Kanungo (1998), who claim that "what is unique in charismatic leadership in contrast to other leadership forms is the intensity of this identification and dependence" (p. 216). According to Conger and Kanungo, followers of charismatic leaders
TWO FACES OF TRANSFORMATIONAL LEADERSHIP
249
report that their self-esteem depends on the leader's evaluation and that their main motivation is to obtain recognition and approval from the leader. Furthermore, departure of the leader may result in followers' feelings of loss and distress (Shamir, 1991). Due to the similarity between charismatic leadership behaviors, as portrayed by charismatic leadership theories, and the charismatic components of transformational leadership, it seems possible that transformational leaders might also trigger feelings of dependency among followers. Following the considerations presented above, we suggest the following hypotheses: Hypothesis 2a: Transformational leadership positively relates to followers' dependence on the leader. Hypothesis 2b: Transformational leadership positively relates to followers' empowerment (self-efficacy, collective efficacy, and OBSE).
Identifications as Mediating Dependency and Empowerment
Hypotheses 2a and 2b stand in contrast to the common view that empowerment and dependence are opposite outcomes that cannot be produced by the same leadership. The effects of dependence and empowerment have often been used in attempts to distinguish "bad" or "negative" leadership from "good" or "positive" leadership (e.g., Conger & Kanungo, 1998; Howell, 1988). However, the arguments presented above suggest that transformational leadership might result in both dependency and empowerment of followers. The possibility of this dual effect raises the question of what the underlying influence mechanisms are that enable transformational leadership to have such divergent effects on followers. We suggest that the answer may lie, at least to some extent, not in the leader behavior but in the identifications that these behaviors evoke among followers, which in turn mediate the effects of leader behaviors on followers. Specifically, we suggest that personal identification with the leader helps explain the relationship between transformational leadership and dependency on the leader, whereas social identification helps explain the relationship between transformational leadership and followers' empowerment. A review of the literature offers a strong support for the relationship between personal identification and dependence. Howell (1988) was among the first to claim that personal identification with the leader is likely to lead to dependence, submissive loyalty, conformity, and "blind" obedience of the followers. Similarly, Conger and Kanungo (1998) stated, "Dependence stems in large part from a strong identification with the leader" (p. 216). Further, Shamir (1991) concluded a review of charismatic leadership theories asserting that "explanations that rely on personal identification with the leader imply that departure of the leader will result in a crisis, intense feelings of loss and severe orientation problems on the part of the followers" (p. 96). Therefore, we propose that personal identification mediates the influence of transformational leadership on dependence. In contrast, social identification with the work unit is not likely to result in followers' dependence on the leader. Social identification with the group has a larger potential to contribute to followers' empowerment, because the identification of the followers is focused on the group— of which they are an integral part. Thus,
the attribution of success to the group they belong to is likely, at least to some extent, to affect the self-perceptions of the individuals composing this group. In this article we view social identification as an individual-level construct. Social identity is part of the individual's self-concept and individuals differ in their level of identification with social entities like work groups or units. Research on group processes and intergroup relations suggests that people identify with groups in part because such identification increases their self-esteem and sense of efficacy (e.g., Alderfer, 1987; Hogg & Abrams, 1990). Additionally, as stated by Shamir (1990): "A person whose selfconcept is based in part on team, occupational, or organizational identities will participate in the activities of the collectivity because such participation clarifies and affirms his or her selfconcept" (p. 325). Such affirmation of self-concept can result in feelings of empowerment. Furthermore, social identity theory (Tajfel, 1982) suggests that identification with the group is often associated with the attribution of positive qualities to the group and increases the self-esteem of its members. Thus for Tajfel (1982), social identification includes not only the sense of belonging to a group but also the emotional value derived from this belonging. Extending the views presented above, one can argue that the more people identify with a group the more they are likely to experience certain psychological rewards, such as heightened feelings of empowerment. Specifically, we can expect that higher levels of social identification would be associated with higher levels of self-esteem and perceptions of self- and collective efficacy. Thus, transformational behaviors that enhance strong identification with the organizational unit are likely to empower followers by connecting them to a bigger and stronger entity, increasing their sense of self-worth and self-esteem, and raising their self- and collective-efficacy beliefs (Shamir et al., 1993, 1998). These considerations are captured in the following hypotheses: Hypothesis 3a: The relationship between transformational leadership and followers' dependence is mediated by personal identification with the leader. Hypothesis 3b: The relationship between transformational leadership and followers' empowerment is mediated by social identification with the organizational unit.
Method Sample and Data Collection
The study was carried out in a large Israeli banking organization and involved the leadership of midlevel managers of independent units. The sample included mostly bank branch managers, but in some large branches (over 40 employees) department managers with similar levels of autonomy and responsibility were included. The managers studied had between 8 and 30 subordinates (M 15). We decided not to include managers who had been in their roles for less than 6 months and employees who have been in the branches for less than 3 months to ensure sufficient acquain- tance of subordinates with the manager and to enable the development of identification, dependence, and empowerment. Average role tenure of the managers studied was 24 months (ranging from 6 months to 6 years). Data were collected by questionnaires from employees in 76 units. Thirty-eight units were managed by women. Nine hundred thirty-two employees participated in the study. Complete data were available from 888 employees,
250
KARK, SHAMIR, AND CHEN
working in a wide variety of roles within the branches (e.g., tellers, clerks, and investment consultants), representing a response rate of 89%. Sixtyseven percent of the participants were women, and participants' age varied from 21 years to 65 years.
Levels of Analysis
The outcome and mediating variables were examined as individual-level variables. Clearly, self-efficacy, OBSE, and personal identification are conceptually individual-level variables. Collective efficacy and social identification can both emerge as group-level variables, but they both originate at the individual level. Because we were interested in how leadership influences the origination of these outcomes, as opposed to their emergence to the group level, we examined them at the individual level of analysis (cf. Riggs & Knight, 1994). However, because evaluations of leadership and responses to leadership were collected from the same individuals by using the same questionnaire, relationships might be inflated because of common source and common method variance. To reduce this risk, we decided to treat leadership behaviors as a group-level variable. We based our analysis not on the relationships between leadership behaviors as perceived by each individual and his or her responses to these behaviors but rather on the relationships between the shared perception among members in a group regarding their leader's behaviors and their individual responses to these behaviors. That is, each bank branch was assigned a single transformational leadership, on the basis of the average leadership ratings within the branch. This is consistent with the theoretical argument that leaders direct many of their transformational behaviors to the entire group rather than to each individual and with the practice followed in other studies (e.g., Shamir et al., 1998).
Measures
Leader behaviors. We used 16 transformational leadership items from the short version of Bass and Avolio's Multifactor Leadership Questionnaire 5X (MLQ, Avolio et al., 1999), which included items of the following behavioral components: individualized consideration, intellectual stimulation, inspirational motivation, and idealized influence. The items measuring attributed charisma, which have been criticized for representing leadership impact and not leadership behavior (e.g., Yukl, 1998), were not included. Respondents were asked to mark the frequency with which the leader engages in each of the behaviors on a 5-point scale (ranging from 1 to 5; .93). All other measures used in this study were composed of a 7-point scale. Social identification. Eight items were adopted from identification measures developed by Mael and Ashforth (1992) and Shamir et al. (1998). The items focused on the branch or department to which the respondent belonged ( .80). A sample item is "I view the success of the branch as my own success." Personal identification. We used eight items that were similar to those used for the measurement of social identification, but in this case items focused on the manager of the branch or department ( .96). For instance, the item corresponding to the example item given above is "I view the success of the branch manager as my own success." Some of these items were previously used to measure identification with the leader by Shamir et al. (1998). Dependence on the leader. No published measure of dependence was found in the literature. An eight-item scale was developed for this study ( .84). A sample item is "Sometimes I find it difficult to do my job without the direction of the branch manager." Self-efficacy. We used a scale developed by Riggs and Knight (1994) to measure self-efficacy in the domain of work. The four positively worded items were chosen for this study ( .76). A sample item is "I have confidence in my ability to do my job." Collective efficacy. A nine-item measure was adapted from Riggs and Knight's (1994) measure of collective efficacy and from Guzzo, Yost, Campbell, and Shea's (1993) measure of group potency ( .87). The wording of the items from Guzzo et al. was changed to refer to the individual's perception of the unit efficacy, rather than to the shared perceptions of the group. A sample item is "The members of this department have excellent job skills." OBSE. The 10-item scale developed by Pierce et al. (1989) was used ( .93). A sample item is "I am an important part of this place."
Analysis Strategy
Random coefficient models (RCM), in conjunction with the mediation model standards described in Kenny, Kashy, and Bolger (1998), were used to test the hypotheses. Analyses were conducted using Version 3.0 of the Linear Mixed Effects program for S-PLUS and R written by Pinheiro and Bates (2000), which is statistically equivalent to other RCM programs, such as hierarchical linear modeling (HLM; see Chen & Bliese, 2002, and Jex & Bliese, 1999, for similar applications of the Linear Mixed Effects program for S-PLUS in multilevel research). RCM are well suited to analyses with predictors at different levels. Specifically, we can use RCM to regress individual-level dependent variables (e.g., dependence) on both individual-level predictors such as personal identification and group-level predictors such as average perceptions of transformational leadership. The RCM approach provides the correct parameter estimates and standard errors for these mixed, multilevel models (see Bliese, 2002; Bryk & Raudenbush, 1992). According to Kenny et al. (1998), four criteria need to be met to support mediated relationships. First, the independent variable (i.e., transformational leadership) needs to be related to the mediator (i.e., personal and social identification). Second, the independent variable needs to be related to the dependent variable. That is, transformational leadership needs to be related to dependence, OBSE, self-efficacy, and collective efficacy. Third, the mediators must be related to the dependent variables with the independent variable included in the model. Finally, the relationship between the independent variable and the criterion variable must disappear when controlling for the mediator variable. If the relationship between the independent and the dependent variable is reduced but remains significant in the presence of the mediator, there is evidence for partial mediation.
Construct Validity Evidence
To further examine the validity of our measures, we conducted confirmatory factor analyses in LISREL (Joreskog & Sorbom, 1993). These analyses compared the fit of a model in which measures of all seven factors (i.e., Transformational Leadership, Personal Identification, Social Identification, Dependence, Self-Efficacy, OBSE, and Collective Efficacy) were set to load on their respective factors with alternative, more constrained models in which certain factors (e.g., Transformational Leadership and Personal Identification) were set to load on a single factor. Results showed that the hypothesized seven-factor model fit the data significantly better than all alternative models, providing evidence for the validity and independence of our measures. More specific information regarding these analyses is available upon request from the authors.
Results Aggregation Analyses
Before testing the hypotheses, we first needed to justify the aggregation of individual perceptions of transformational leadership behaviors to the group level of analysis. Average rwg(j) across groups was .88, suggesting sufficient within-group agreement. In addition, ICC(1) was .25 and ICC(2) was .79, which provided
TWO FACES OF TRANSFORMATIONAL LEADERSHIP
251
sufficient evidence for between-group reliability. Finally, an analysis of variance indicated that individual perceptions of leadership significantly clustered by groups, F(76, 886) 4.85, p .01. These results provided sufficient statistical justification for aggregating individual perceptions of leadership to the group level (see Bliese, 2000).
Hypotheses Testing
Table 1 presents the individual-level descriptive statistics and zero-order correlations for the variables in the study. Relationships and significance tests associated with these variables should be viewed with caution until properly modeled in the RCM analyses, because the correlation table does not account for the fact that individual-level relationships might also be affected by the nonindependent nature of the data (Bliese, 2000). These caveats not withstanding, it is worth mentioning that the indicators of empowerment (self-efficacy, collective efficacy, and OBSE) were more highly related to each other (r .18 to .55) than to dependence (r .04 to .12). We address these findings in the Discussion section. Table 2 presents results of the RCM analyses following the steps suggested by Kenny et al. (1998) to test for mediated relationships. Note that the first step in these analyses also tested Hypotheses 1a and 1b, and the second step also tested Hypotheses 2a and 2b. Supporting Hypotheses 1a and 1b, results from the RCM analyses (see Table 2, Step 1) indicated that transformational leadership significantly and positively predicted both personal identification, coefficient 1.69, t(74) 12.46, p .01; and social identification, coefficient .45, t(74) 3.57, p .01. Also, supporting Hypothesis 2a, results from the second step (see Table 2) indicated that transformational leadership significantly predicted dependence, coefficient .49, t(74) 4.85, p .01. Additionally, lending support to Hypothesis 2b, transformational leadership significantly predicted OBSE, coefficient .18, t(74) 2.37, p .05, and collective efficacy, coefficient .35, t(74) 2.95, p .01, and marginally predicted self-efficacy, coefficient .10, t(74) 1.82, p .07. In order to gain support for Hypotheses 3a and 3b, the relationships between transformational leadership and dependence and transformational leadership and empowerment must disappear when including the identification variables in the equation. To ensure that personal and social identification predicted the outcomes differentially as predicted, we entered personal identifica-
tion, social identification, and transformational leadership simultaneously into the equations (see Table 2, Steps 3 and 4). Supporting Hypothesis 3a, personal identification positively and significantly predicted dependence, coefficient .31, t(810) 12.95, p .01, whereas the relationship between transformational leadership and dependence became nonsignificant, coefficient .07, t(74) 0.72, ns. Unexpectedly, social identification predicted dependence significantly and negatively, coefficient .21, t(810) 5.83, p .01, in this analysis. However, this unexpected finding seems to have been caused by the strong correlation between personal and social identification (i.e., due to multicollinearity). Indeed, when only social identification and transformational leadership were entered into the equation, transformational leadership significantly predicted dependence, coefficient .46, t(74) 4.50, p .01, whereas social identification did not significantly predict dependence, coefficient .05, t(811) 1.45, ns. In contrast, as hypothesized, when only personal identification and transformational leadership were entered into the equation, transformational leadership did not significantly predict dependence, coefficient .10, t(74) 0.96, ns, whereas personal identification significantly predicted dependence, coefficient .23, t(811) 11.52, p .01. These results support Hypothesis 3a, suggesting that personal identification but not social identification, mediates the relationship between transformational leadership and dependence. Tests of Hypothesis 3b involved the three indicators of empowerment—OBSE, self-efficacy, and collective efficacy. Results indicated that when entered simultaneously into the equation, both personal identification, coefficient .07, t(810) 3.93, p .01, and social identification, coefficient .20, t(810) 7.22, p .01, significantly predicted OBSE, whereas transformational leadership did not significantly predict OBSE, coefficient .03, t(74) 0.49, ns. As expected, social identification predicted OBSE to a greater extent than did personal identification. In addition, social identification significantly predicted selfefficacy, coefficient .16, t(810) 6.51, p .01, whereas personal identification, coefficient .02, t(810) 1.11, ns, and transformational leadership, coefficient .01, t(74) 0.17, ns, did not significantly predict self-efficacy. Likewise, social identification significantly predicted collective efficacy, coefficient .42, t(810) 13.12, p .01, whereas personal identification did not significantly predict collective efficacy, coefficient .00, t(810) 0.07, ns, and transformational leadership only marginally predicted collective efficacy, coefficient .17,
Table 1 Individual-Level Means, Standard Deviations, and Correlations
Variable 1. 2. 3. 4. 5. 6. Transformational leadership Personal identification Social identification Dependence Organization-based self-esteem Self-efficacy 888. All correlations M 3.55 4.54 5.54 2.33 6.00 6.26 SD 0.81 1.78 1.10 1.07 0.80 0.70 1 — .73 .43 .34 .23 .13 .01. 2 — .57 .40 .31 .19 3 4 5 6 7
— .07 .37 .29
— .04 .12
— .55
—7.
Collective efficacy
5.58
0.9
Note. N
.06 are significant, p
252
KARK, SHAMIR, AND CHEN
Table 2 Mediation of the Relationship of Transformational Leadership With Dependence and Empowerment by Personal and Social Identification
Dependent variable Independent variable Step 1 Transformational leadership Personal identification 1.69** (74) Dependent variable Dependence Step 2 Transformational leadership Steps 3 and 4 Transformational leadership Personal identification Social identification .49** (74) .07 (74) .31** (810) .21** (810) OBSE .18* (74) .03 (74) .07** (810) .20** (810) Self-efficacy .10† (74) .01 (74) .02 (810) .16** (810) Collective efficacy .35** (74) .17† (74) .00 (810) .42** (810) Social identification .45** (74)
Note. Results are unstandardized parameter estimates from random coefficient models analyses; numbers in parentheses are corresponding degrees of freedom. OBSE organization-based self-esteem. † p .10 (marginally significant). * p .05. ** p .01.
t(74) 1.70, p .09. These results support Hypothesis 3b, suggesting that social identification, but not personal identification, mediates the relationship between transformational leadership and empowerment. Figure 2 provides a summary of the results from the hypotheses tests. Note that the slight discrepancies between the results presented in Table 2 and Figure 2 stem from the fact that in the final analyses reported in Table 2 (Steps 3 and 4), both personal and social identification were included in analyses involving dependence and empowerment as the dependent variables. However, in the analyses reported in Figure 2, we included personal identifi-
cation (but not social identification) in analyses of dependence as the dependent variable, and social identification (but not personal identification) in analyses of the empowerment dimensions as the dependent variables. As shown in the figure, supporting Hypotheses 1a and 1b, transformational leadership positively and significantly predicted personal and social identification, respectively. Supporting Hypotheses 2a and 3a, personal identification predicted dependence over and above transformational leadership, whereas the significant relationship between transformational leadership and dependence became nonsignificant once controlling for personal identification. Finally, in support of Hypotheses 2b and 3b,
Figure 2. Summary of random coefficient modeling results. Numbers in parentheses are degrees of freedom. DV dependent variable. † p .10 (marginally significant). * p .05.
TWO FACES OF TRANSFORMATIONAL LEADERSHIP
253
social identification predicted the three empowerment variables over and above transformational leadership, whereas the significant relationships of transformational leadership with the empowerment variables became nonsignificant (when examining selfefficacy and OBSE) or marginally significant (when examin- ing collective efficacy) once controlling for social identification. In sum, results provided support for the notion that personal identification mediates the relationship between transformational leadership and dependence, whereas social identification mediates the relationship between transformational leadership and empowerment.
Discussion
The main goal of this study was to examine personal and social identification as possible mechanisms through which transformational leadership differentially relates to followers' dependence and empowerment. The results of this research offer several notable findings. First, as predicted, transformational leadership was found to be positively related to both personal identification with the leader and social identification with the work unit, thus providing evidence that transformational leaders are likely to exert their influence on followers by affecting their feelings of identification. Prima facie, it is interesting that transformational leadership was more strongly associated with personal identification with the leader than with social identification. However, this finding is not surprising because the latter is likely to be influenced by many other factors besides the leader's personality and behavior, such as the composition of the group, its history, and its cohesion. Although leaders can influence members' identification with the group, their impact on this identification is more limited than their impact on followers' personal identification. However, this finding suggests that prior assertions in the literature that portrayed transformational leaders as socialized leaders, assuming that their main avenue of influence may not rely on personal identification, should be reexamined. Second, although we have not covered the full range of factors that are included in some broad conceptions of empowerment (e.g., Spreitzer, 1995), results of this study confirm the hypothesized relationship between transformational leadership and some aspects of empowerment, represented by followers' self- and collectiveefficacy beliefs and their OBSE. More important, our results suggest that contrary to the common dichotomy in the literature (e.g., Howell, 1988), follower empowerment and follower dependence are not opposite to each other. Rather, our findings suggest that dependence and empowerment are independent of each other, as the correlations between dependence and the empowerment indicators were rather weak. Furthermore, because both dependence and empowerment were positively related to transformational leadership, our results suggest that, contrary to the common assumption that different types of leadership (e.g., charismatic vs. transformational) lead to followers' dependence or empowerment, the same type of leadership may be associated simultaneously with both empowerment and dependence. Consistent with this conclusion, results suggest that different mechanisms (personal or social identification) can, at least partially, account for these different outcomes (empowerment and dependence of followers) of transformational leadership. As expected, social identification with the work unit mediated the relationships between transformational
leadership and indicators of empowerment, whereas personal identification with the leader mediated the relationship between transformational leadership and dependence. It may be noted, in this regard, that some feminist relational perspectives (e.g., Fletcher, 1999; Miller & Stiver, 1997) challenge the commonly accepted distinction and opposition between dependence and empowerment, claiming that this distinction rests on traditional masculine models of relationships. Traditional models view the development of independence and self-reliance as including disengagement from significant others and as negating strong forms of relatedness, which include interdependence. In contrast, the feminist perspectives present a model of development and growth within a relationship, which emphasizes both dependence and independence as fluid states that are neither static nor achieved. Within such a model, a relationship may be characterized by both a sense of empowerment and, at the same time, at least some sense of interpersonal dependence. Apparently, the relationships between followers' dependence and empowerment and the relationships between transformational leadership and these states, as well as social and personal identification, are more complex than suggested by the literature. Further research should clarify these relationships by examining whether other leadership behaviors, particularly those associated with the "personalized charisma" (Howell, 1988) type may lead to dependence without empowerment and whether the mixture of empowerment and dependence varies between different stages of the follower-leader relationship. For instance, it is possible that in early stages of the relationship some dependence on the leader is a necessary condition for the leader's empowering effects, whereas in later stages the empowerment effects depend on the followers achieving independence from the leader and on their need for affirmation and recognition. It is also possible that followers will feel empowered in some aspects of their work (e.g., giving service to customers) and dependant on the leader's expertise and experience in other aspects of their work (e.g., decision-making processes). Furthermore, personality differences among followers (e.g., their tendencies to become dependent or empowered) and the ways in which they interact with the leader might affect the leaders and account for leaders' ability to influence in diverse ways (Dvir & Shamir, 2001). Moreover, the consequences of personal identification with the leader, social identification with the group, and dependency on the leader might be entirely different in the cases of moral and immoral leaders. In addition, it should be acknowledged that identifications, empowerment, and dependence do not depend only on the leader's behavior. For instance, Bion (1974) described the development of dependence perceptions in groups that do not depend entirely on the leader's actions. Guzzo and Shea (1992), among others, have drawn attention to the structure of the group task as a factor influencing group processes. This work suggests that high task interdependence might influence the level of social identification in groups. Future studies should look at the impact of the leader on followers' identifications, empowerment, and dependency against the background of other contextual factors that might affect these variables. We also encourage researchers to examine the broader conceptualization of the concept of empowerment (Spreitzer, 1995) in future research, in order to enable a more complete conclusion regarding the relationships among transformational leadership, identification, and empowerment.
254
KARK, SHAMIR, AND CHEN
The findings of this study can also be interpreted from a perspective recently suggested by Lord, Brown, and Freiberg (1999) and by Kark and Shamir (2002). According to this perspective, leaders achieve their effects on followers by priming different aspects of their self-concepts. Lord et al. viewed the self-concept as having three aspects: a personal self, which is based on individual attributes and preferences; a relational self, which is based on close relations with specific others (such as a leader); and a collective self, which is based on group affiliations and identifications. Applying this perspective to our study, one can note that transformational leadership may prime both the relational self (evidenced by personal identification with the leader) and the collective self (evidenced by social identification with the unit). These two faces of transformational leadership have different consequences. Priming the relational self results in dependence on the leader or follower-leader interdependence, whereas priming the collective self results in empowerment (for a more comprehensive theoretical model, see Kark & Shamir, 2002). Further elaboration of transformational leadership theory and measurement is needed in order to identify specific leader behaviors that prime different aspects of followers' self-concepts. The results of this study are limited by its cross-sectional design and by the use of a single method of data collection. Although researchers have shown that common method bias is rarely strong enough to invalidate research findings (e.g., Doty & Glick, 1998), replications and extensions of our findings using experimental and longitudinal designs are needed. Furthermore, the specific banking context in which the research took place could also limit the generalization of the research conclusions. However, in this regard, it is interesting to note that bank culture is not very conducive to transformational leadership (e.g., it is not a highly ideological setting) and middle-level branch managers, who are constrained by bank procedures, are possibly limited in their ability to present the "full range" of transformational behaviors. In addition, branches are perhaps not the best foci for social identification, and perhaps stronger relationships between leadership behavior, social identification, and empowerment would have been found in situations in which the work unit and the nature of the work is more attractive and can arouse higher levels of identification. These considerations suggest that even stronger results than the ones found in the present study could be found in other settings more conducive to transformational leadership (e.g., military units). Despite its limitations, the study has made several tentative contributions to the literature on transformational leadership in organizations. In general, the results of the study provide, perhaps for the first time, fairly strong support for the hypothesized relationships between transformational leadership, followers' identification with the leader and the organizational unit, dependence on the leader, and certain indicators of followers' empowerment. These results strengthen recent experimental findings showing that leaders who received training in transformational leadership succeeded in obtaining higher levels of commitment, motivation and performance among followers (Barling et al., 1996; Dvir et al., 2002). Furthermore, support was also found for the hypothesized differential roles of personal and social identification in mediating the relationships between transformational leadership and followers' empowerment and dependence. Our results start to shed some light on the influence processes by which transformational leadersmay achieve their superior results. However, future research
should examine further whether followers' identification, dependence, and empowerment indeed mediate the relationships between transformational leadership and more distal effects on followers' behavior and organizational outcomes.
References
Alderfer, C. P. (1987). An intergroup perspective on group dynamics. In J. W. Lorsch (Ed.), Handbook of organizational behavior (pp. 190 222). Englewood Cliffs, NJ: Prentice Hall. Ashforth, B. E., & Mael, F. (1989). Social identity theory and the organization. Academy of Management Review, 14, 20 -39. Avolio, B. J., Bass, B. M., & Jung, D. I. (1999). Reexamining the components of transformational and transactional leadership using the multifactor leadership questionnaire. Journal of Occupational and Organizational Psychology, 72, 441- 462. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall. Barling, J., Weber, T., & Kelloway, E. K. (1996). Effects of transformational leadership training on attitudinal and financial outcomes: A field experiment. Journal of Applied Psychology, 81, 827- 832. Bass, B. M. (1985). Leadership and performance beyond expectation. New York: Free Press. Bass, B. M. (1997). Does the transactional-transformational leadership paradigm transcend organizational and national boundaries? American Psychologist, 52, 130 -139. Bion W. R. (1974). Experience in groups, and other papers. London: Tavistock. Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analyses. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 349 381). San Francisco: Jossey-Bass. Bliese, P. D. (2002). Using multilevel random coefficient modeling in organizational research. In F. Drasgow & N. Schmitt (Eds.), Advances in measurement and data analysis (pp. 401- 445). San Francisco: JosseyBass. Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage. Chen, G., & Bliese, P. D. (2002). The role of different levels of leadership in predicting self- and collective efficacy: Evidence for discontinuity. Journal of Applied Psychology, 87, 549 -556. Conger, J. A., & Kanungo, R. N. (1988). The empowerment process: Integrating theory and practice. Academy of Management Review, 3, 471- 482. Conger, J. A., & Kanungo, R. N. (1998). Charismatic leadership in organizations. Thousand Oaks, CA: Sage. Doty, D. H., & Glick, W. H. (1998). Common methods bias: Does common methods variance really bias results? Organizational Research Methods, 1, 374 - 406. Dvir, T., Eden, D., Avolio, B. J., & Shamir, B. (2002). Impact of transformational leadership on follower development and performance: A field experiment. Academy of Management Journal, 45, 735-744. Dvir, T., & Shamir, B. (2001, June). The role of follower characteristics in transformational leadership theory. Paper presented at the International Conference on Business, Honolulu, Hawaii. Fletcher, J. K. (1999). Disappearing acts: Gender, power, and relational practice at work. Cambridge, MA: MIT Press. Gardner, D. G., & Pierce, J. L. (1998). Self-esteem and self-efficacy within the organizational context: An empirical examination. Group and Organization Management, 23, 48 -70. Guzzo, R., & Shea, G. P. (1992). Group performance and intergroup relations in organizations. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 3, 2nd ed., pp. 269 -313). Palo Alto, CA: Consulting Psychologists Press, Inc.
TWO FACES OF TRANSFORMATIONAL LEADERSHIP Guzzo, R. A., Yost, P. R., Campbell, R. J., & Shea, G. P. (1993). Potency in groups: Articulating a construct. British Journal of Social Psychology, 32, 87-106. Hackman, J. R. (1992). Group influences on individuals in organizations. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 3, pp. 199 -267). Palo Alto, CA: Consulting Psychologists Press. Hogg, M. A., & Abrams, D. (1990). Social motivation, self-esteem and social identity. In D. Abrams & M. A. Hogg (Eds.), Social identity theory: Constructive and critical advances (pp. 28 - 47). New York: Harvester Wheatsheaf. Howell, J. M. (1988). Two faces of charisma: Socialized and personalized leadership in organizations. In J. A. Conger & R. N. Kanungo (Eds.), Charismatic leadership (pp. 213-236). San Francisco: Jossey-Bass. Jex, S. M., & Bliese, P. D. (1999). Efficacy beliefs as a moderator of the impact of work-related stressors: A multilevel study. Journal of Applied Psychology, 84, 349 -361. Joreskog, K., & Sorbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS command language. Hillsdale, NJ: Erlbaum. Kark, R., & Shamir, B. (2002). The dual effect of transformational leadership: Priming relational and collective selves and further effects on followers. In B. J. Avolio & F. J. Yammarino (Eds.), Transformational and charismatic leadership: The road ahead (Vol. 2, pp. 67-91). Amsterdam: JAI Press. Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., pp. 233-265). Boston: McGrawHill. Kets de Vries, M. F. (1988). Origins of charisma: Ties that bind the leader and the led. In J. A. Conger & R. N. Kanungo (Eds.), Charismatic leadership (pp. 237-252). San-Francisco: Jossey-Bass. Lord, R. G., Brown, D. J., & Freiberg, S. J. (1999). Understanding the dynamics of leadership: The role of follower self-concepts in the leader/ follower relationship. Organizational Behavior and Human DecisionProcess, 78, 167-203. Lowe, K. B., Kroeck, K. G., & Sivasubramaniam, N. (1996). Effectiveness of correlates of transformational and transactional leadership: A metaanalytic review of the MLQ literature. Leadership Quarterly, 7, 385425. Mael, F., &Ashforth, B. E. (1992). Alumni and their alma matter: A partial test of the reformulated model of organizational identification. Journal of Organizational Behavior, 13, 103-123.
255
Miller, J. B., & Stiver, I. P. (1997). The healing connection. Boston: Beacon Press. Pierce, J. L., Gardner, D. G., Cummings, L. L., & Dunham, R. B. (1989). Organization-based self-esteem: Construct definition, measurement, and validation. Academy of Management Journal, 32, 622- 648. Pinheiro, J. C., & Bates, D. M. (2000). Mixed-effects models in S and S-PLUS. New York: Springer-Verlag. Pratt, M. G. (1998). To be or not to be: Central questions in organizational identification. In D. A. Whetten & P. C. Godfrey (Eds.), Identity in organizations: Building theory through conversation (pp. 171-207). Thousand Oaks, CA: Sage. Riggs, M. L., & Knight, P. A. (1994). The impact of perceived group success-failure on motivational beliefs and attitudes. Journal of Applied Psychology, 79, 755-766. Shamir, B. (1990). Calculations, values, and identities: The sources of collectivistic work motivation. Human Relations, 43, 313-332. Shamir, B. (1991). The charismatic relationship: Alternative explanations and predictions. Leadership Quarterly, 2, 81-104. Shamir, B., House, R. J., & Arthur, M. B. (1993). The motivational effects of charismatic leadership: A self-concept based theory. Organization Science, 4, 577-593. Shamir, B., Zakay, E., Breinin, E & Popper, M. (1998). Correlates of charismatic leader behavior in military units: Subordinates' attitudes, unit characteristics, and superiors' appraisals of leader performance. Academy of Management Journal, 41, 387- 409. Shamir, B., Zakay, E., Breinin, E., & Popper, M. (2000). Leadership and social identification in military units. Journal of Applied Social Psychology, 30, 612- 640. Spreitzer, G. M. (1995). Individual empowerment in the workplace: Dimensions, measurement, validation. Academy of Management Jour- nal, 38, 1442-1465. Tajfel, H. (1982). Social identity and intergroup relations. Cambridge, England: Cambridge University Press. Waldman, D. A., Ramirez, G. G., House, R. J., & Puranam, P. (2001). Does leadership matter? CEO leadership attributes and profitability under conditions of perceived environmental uncertainty. Academy of Management Journal, 44, 134 -143. Yukl, G. (1998). Leadership in organizations. Englewood Cliffs, NJ: Prentice Hall.
doc_378500839.docx
Literature (from Latin litterae (plural); letter) is the art of written work. The word literature literally means "things made from letters." Literature is commonly classified as having two major forms—fiction & non-fiction—and two major techniques—poetry and prose.
HRM Study on Two Faces of Transformational Leadership: Empowerment and Dependency
Followers' identification with the leader and the organizational unit, dependence on the leader, and empowerment by the leader are often attributed to transformational leadership in organizations. However, these hypothesized outcomes have received very little attention in empirical studies. Using a sample of 888 bank employees working under 76 branch managers, the authors tested the relationships between transformational leadership and these outcomes. They found that transformational leadership was positively related to both followers' dependence and their empowerment and that personal identification mediated the relationship between transformational leadership and followers' dependence on the leader, whereas social identification mediated the relationship between transformational leadership and followers' empowerment. The authors discuss the implications of these findings to both theory and practice.
There is accumulating evidence that transformational leadership substantially influences employee motivation and performance (e.g., Barling, Weber, & Kelloway, 1996; Dvir, Eden, Avolio, & Shamir, 2002; Lowe, Kroeck, & Sivasubramaniam, 1996; Waldman, Ramirez, House, & Puranam, 2001). However, research on transformational leadership has left several questions unanswered (Yukl, 1998). The present study focuses on two of these questions. First, What are the processes by which transformational leadership exerts its influence on followers? As Yukl (1998) concluded, "A variety of different influence processes may be involved in transformational leadership" (p. 328). In the present study, we suggest that two of the mechanisms by which transformational leaders exert their influence on followers are the creation of personal identification with the leader and social identification with the work unit. The second question we focus on is, What are the possible effects of these different influence processes? Transformational leadership theory suggests that such leadership is likely to result in
Ronit Kark, Department of Psychology and Department of Sociology, Bar-Ilan University, Ramat Gan, Israel; Boas Shamir, Department of Anthropology and Sociology, Hebrew University of Jerusalem, Israel; Gilad Chen, School of Psychology, Georgia Institute of Technology. Ronit Kark is grateful for the financial support of the Israel Foundation Trustees, The Shaine Center for Research in Social Sciences, The Levi Eshkol Institute for Economic, Social and Political Research, and the Department of Psychology at the Hebrew University of Jerusalem. This article was written while Ronit Kark was a visiting scholar at the Institute for Research on Women and Gender and at the Department of Psychology, University of Michigan.
growth, independence, and empowerment of followers (Bass, 1985; Dvir et. al, 2002). However, influential leadership can also lead to the weakening of followers and their dependency on the leader (Howell, 1988). Dependence on the leader has several manifestations. It implies that the subordinate is limited in his or her ability to proceed with work and make decision without the leader's guidance. Psychologically it means that the subordinate's motivation and self-esteem depend on receiving recognition and approval from the leader. Empowerment, in contrast, connotes independence and autonomy. An empowered person is selfmotivated and believes in his or her ability to cope and perform successfully. In the present study we argue that personal identification with the leader and social identification with the work unit are intermediate outcomes that help explain why transformational leadership can result in two seemingly contradictory outcomes: followers' dependence and empowerment. Specifically, we propose that the relationship between transformational behaviors of leaders and followers' dependence is mediated by followers' personal identification with the leader whereas the relationship between transformational behaviors and followers' empowerment is mediated by followers' identification with the work group. The model of the relationships proposed in the present research is summarized in Figure 1.
Theory and Hypotheses Transformational Leadership and Followers' Identification
Transformational leadership has been presented in the literature as different from transactional leadership. Whereas transactional leadership was defined on the basis of the influence process
246
TWO FACES OF TRANSFORMATIONAL LEADERSHIP
247
Figure 1.
Summary of study hypotheses. H
Hypothesis.
underlying it, as an exchange of rewards for compliance, transformational leadership was defined on the basis of its effects, as transforming the values and priorities of followers and motivating them to perform beyond their expectations (Yukl, 1998). Researchers proposed that transformational leadership behaviors comprise four components: inspirational motivation, idealized influence, individualized consideration, and intellectual stimulation. The first two components represent the notion of "charisma" (Avolio, Bass, & Jung, 1999; Bass, 1985) and are similar to behaviors specified in theories of charismatic leadership (Conger & Kanungo, 1998). Inspirational motivation includes the creation and presentation of an attractive vision of the future, the use of symbols and emotional arguments, and the demonstration of optimism and enthusiasm. Idealized influence includes behaviors such as sacrificing for the benefit of the group, setting a personal example, and demonstrating high ethical standards. The third component, individualized consideration, includes providing support, encouragement, and coaching to followers. The fourth component, intellectual stimulation, involves behaviors that increase awareness of problems and challenge followers to view problems from new perspectives. Previous research has shown that these transformational behaviors are related to leadership effectiveness (Lowe et al., 1996) and high employee performance (Barling et al., 1996; Dvir et al., 2002). In the present study, consistent with previous research (e.g., Shamir, Zakay, Breinen, & Popper, 1998), we conceptualize transformational leadership as a group-level construct. Specifically, transformational leadership behaviors are directed at the leader's group and thus are a form of ambient stimuli (cf. Hackman, 1992) that influences the group as a whole, as well as individuals within the group. A review of transformational and charismatic leadership theories suggests that such leadership may achieve its effects partly through the creation of followers' identification with the leader (personal identification) and with the work group (social identification; Yukl, 1998). Thus, we conceptualize personal identifica-
tion and social identification as proximal individual-level outcomes through which transformational leadership (a group-level construct) influences more distal individual-level outcomes (i.e., dependence and empowerment). Personal identification. Borrowing from Pratt's (1998) definition of organizational identification, personal identification is evident when an individual's belief about a person (a leader) becomes self-referential or self-defining. According to early psychoanalytic theories of charismatic leadership (e.g., Kets de Vries, 1988), the influence of charismatic leaders on followers is based on personal identification with the leader, which in turn stems from a transference of earlier identifications with parents. More recent theories also highlight the role of identification with the leader. The charismatic leadership theory of Conger and Kanungo (1998) emphasizes personal identification as a central mechanism through which charismatic leaders influence their followers, suggesting that such influence is based on referent power. Similarly, Shamir, House, and Arthur (1993) posited role modeling as one of the major processes by which charismatic leaders influence followers. Role modeling implies a process by which followers mold their beliefs, feelings, and behavior according to those of the leader. In these theories, personal identification represents one of two (or both) modes of identification suggested by Pratt (1998): (a) evoking followers' self-concept in the recognition that they share similar values with the leader and (b) giving rise to followers' desire to change their self-concept so that their values and beliefs become more similar to those of the leader. The similarity and overlap between the charismatic components of transformational leadership (i.e., idealized influence and inspirational motivation) and leadership behaviors included in theories of charismatic leadership suggests that transformational leaders are likely to arouse personal identification among their followers. Social identification. Another underlying intermediate effect that might account for the exceptional impact of transformational leadership on followers' perceptions and behaviors is social identification. Social identification implies that an individual's belief
248
KARK, SHAMIR, AND CHEN
about a group (or an organization) is self-referential or selfdefining (Pratt, 1998). When individuals identify with a group, they base their self-concept and self-esteem partly on their belonging to the group, and group successes and failures are experienced as personal successes and failures (Mael & Ashforth, 1992). Ashforth and Mael (1989) were among the first to suggest that transformational leadership might influence followers' social identification. The influence of leaders on the social identification of followers is central to Shamir et al.'s (1993) motivational theory of charismatic leadership. They have suggested that the influence of charismatic and transformational leaders is based on their success in connecting followers' self-concept to the mission and to the group, such that followers' behavior for the sake of the group becomes self-expressive. Recent findings support this assertion, showing that leaders who raise followers' identification with the group increase followers' willingness to contribute to group objectives (Shamir et al., 1998; Shamir, Zakay, Breinin, & Popper, 2000). In sum, following the theoretical arguments presented above, we propose that transformational leadership is related to both personal and social identification of followers: Hypothesis 1a: Transformational leadership positively relates to followers' personal identification with the leader. Hypothesis 1b: Transformational leadership positively relates to followers' social identification with the group.
Transformational Leadership and Followers' Empowerment and Dependence
Identification with the leader and the group is important because it has perceptual, motivational, and behavioral consequences (Pratt, 1998). However, personal and social identification might each result in different outcomes (Howell, 1988; Shamir, 1991). In this research we focus on two important outcomes: followers' empowerment and dependence. Despite the centrality of the empowerment and dependence constructs in theories of charismatic and transformational leadership, there has been very little systematic research examining the relationships between transformational leadership and these outcomes. Empowerment. The impact of transformational leadership on followers' performance is often explained as stemming from followers' development and empowerment, which increase both their ability and their motivation (Bass, 1997). The empowerment of followers is often presented as one of the main features that distinguish such leadership from transactional leadership, which does not seek to empower the followers but merely to influence their behavior. To achieve these ends, transformational leadership includes empowering behaviors such as delegation of responsibility to followers, enhancing followers' capacity to think on their own, and encouraging them to come up with new and creative ideas (Dvir et. al., 2002). Empowerment has been defined as a broad and multidimensional concept (Spreitzer, 1995). In this study we adopt a narrower perspective and focus on three indicators of empowerment: selfefficacy, collective efficacy, and organization-based self-esteem (OBSE). Conger and Kanungo (1988) defined empowerment as the process of raising others' self-efficacy perceptions. Self-efficacy is
the individual's belief in his or her ability to successfully perform tasks (Bandura, 1986). According to Conger and Kanungo (1998), transformational leadership behaviors have clear empowering effects on followers in terms of raising their self-efficacy beliefs. Following Shamir et al. (1993), we extend the concept of empowerment suggested by Conger and Kanungo to include collectiveefficacy beliefs and OBSE. Collective efficacy has been conceptualized as being analogous to self-efficacy and is defined as individuals' beliefs that their work group can function effectively and perform its tasks successfully (Bandura, 1986). Identification with the group is often associated with the attribution of positive qualities to the group (Tajfel, 1982), and hence leader behaviors that increase social identification are likely to increase collectiveefficacy beliefs as well. In addition, Shamir et al. (1993) hypothesized that by having high expectations of followers, expressing the leader's belief in followers' abilities, and showing how the mission reflects followers' values, leaders also influence followers' self-esteem. Although Shamir et al. referred to global self-esteem, a more relevant construct is OBSE, which reflects the self-perceived value that individuals have of themselves as organization members acting within an organizational context (Pierce, Gardner, Cummings, &Dunham, 1989). According to Pierce et al. (1989), OBSE is influenced, among other things, by leaders' behaviors toward their subordinates. Members with high OBSE perceive themselves as important, influential, effective, and worthwhile in their organizational units. Furthermore, "employees that possess high levels of OBSE . . . may be described by their peers as being motivated, capable and empowered" (Gardner & Pierce, 1998, p. 50). This supports the inclusion of OBSE as an indicator of empowerment. According to the conceptualization of empowerment used by Spreitzer (1995), the indicators chosen in this study can be seen as related to two dimensions of empowerment: impact and competence. Impact, defined as the degree to which one believes he or she can influence outcomes at work, is evident in followers' OBSE perception of themselves as "making a difference" in the organization and being important and worthwhile (Pierce et al., 1989). The dimension of competence, defined as a belief in one's capacity to perform work activities with skill, is expressed in self-efficacy, collective efficacy, and to a lesser extent, in OBSE perception of efficiency. Dependence. Despite widespread reference to transformational behaviors that can contribute to the empowerment of followers, transformational leadership also includes charismatic behaviors that have been suggested to have the potential of creating dependence on the leader among followers. Theories of charismatic leadership propose that these types of leadership behaviors might result in followers who perceive the leader as extraordinary and exceptional and therefore become dependent on the leader for guidance and inspiration (Yukl, 1998). Psychoanalytic theories, which view the charismatic relationship as a regression to early childhood relationships with parents, imply that such leadership results in increased dependence on the leader (e.g., Kets de Vries, 1988). The potential of charismatic leadership for creating dependence is also recognized by more recent theories, for instance by Conger and Kanungo (1998), who claim that "what is unique in charismatic leadership in contrast to other leadership forms is the intensity of this identification and dependence" (p. 216). According to Conger and Kanungo, followers of charismatic leaders
TWO FACES OF TRANSFORMATIONAL LEADERSHIP
249
report that their self-esteem depends on the leader's evaluation and that their main motivation is to obtain recognition and approval from the leader. Furthermore, departure of the leader may result in followers' feelings of loss and distress (Shamir, 1991). Due to the similarity between charismatic leadership behaviors, as portrayed by charismatic leadership theories, and the charismatic components of transformational leadership, it seems possible that transformational leaders might also trigger feelings of dependency among followers. Following the considerations presented above, we suggest the following hypotheses: Hypothesis 2a: Transformational leadership positively relates to followers' dependence on the leader. Hypothesis 2b: Transformational leadership positively relates to followers' empowerment (self-efficacy, collective efficacy, and OBSE).
Identifications as Mediating Dependency and Empowerment
Hypotheses 2a and 2b stand in contrast to the common view that empowerment and dependence are opposite outcomes that cannot be produced by the same leadership. The effects of dependence and empowerment have often been used in attempts to distinguish "bad" or "negative" leadership from "good" or "positive" leadership (e.g., Conger & Kanungo, 1998; Howell, 1988). However, the arguments presented above suggest that transformational leadership might result in both dependency and empowerment of followers. The possibility of this dual effect raises the question of what the underlying influence mechanisms are that enable transformational leadership to have such divergent effects on followers. We suggest that the answer may lie, at least to some extent, not in the leader behavior but in the identifications that these behaviors evoke among followers, which in turn mediate the effects of leader behaviors on followers. Specifically, we suggest that personal identification with the leader helps explain the relationship between transformational leadership and dependency on the leader, whereas social identification helps explain the relationship between transformational leadership and followers' empowerment. A review of the literature offers a strong support for the relationship between personal identification and dependence. Howell (1988) was among the first to claim that personal identification with the leader is likely to lead to dependence, submissive loyalty, conformity, and "blind" obedience of the followers. Similarly, Conger and Kanungo (1998) stated, "Dependence stems in large part from a strong identification with the leader" (p. 216). Further, Shamir (1991) concluded a review of charismatic leadership theories asserting that "explanations that rely on personal identification with the leader imply that departure of the leader will result in a crisis, intense feelings of loss and severe orientation problems on the part of the followers" (p. 96). Therefore, we propose that personal identification mediates the influence of transformational leadership on dependence. In contrast, social identification with the work unit is not likely to result in followers' dependence on the leader. Social identification with the group has a larger potential to contribute to followers' empowerment, because the identification of the followers is focused on the group— of which they are an integral part. Thus,
the attribution of success to the group they belong to is likely, at least to some extent, to affect the self-perceptions of the individuals composing this group. In this article we view social identification as an individual-level construct. Social identity is part of the individual's self-concept and individuals differ in their level of identification with social entities like work groups or units. Research on group processes and intergroup relations suggests that people identify with groups in part because such identification increases their self-esteem and sense of efficacy (e.g., Alderfer, 1987; Hogg & Abrams, 1990). Additionally, as stated by Shamir (1990): "A person whose selfconcept is based in part on team, occupational, or organizational identities will participate in the activities of the collectivity because such participation clarifies and affirms his or her selfconcept" (p. 325). Such affirmation of self-concept can result in feelings of empowerment. Furthermore, social identity theory (Tajfel, 1982) suggests that identification with the group is often associated with the attribution of positive qualities to the group and increases the self-esteem of its members. Thus for Tajfel (1982), social identification includes not only the sense of belonging to a group but also the emotional value derived from this belonging. Extending the views presented above, one can argue that the more people identify with a group the more they are likely to experience certain psychological rewards, such as heightened feelings of empowerment. Specifically, we can expect that higher levels of social identification would be associated with higher levels of self-esteem and perceptions of self- and collective efficacy. Thus, transformational behaviors that enhance strong identification with the organizational unit are likely to empower followers by connecting them to a bigger and stronger entity, increasing their sense of self-worth and self-esteem, and raising their self- and collective-efficacy beliefs (Shamir et al., 1993, 1998). These considerations are captured in the following hypotheses: Hypothesis 3a: The relationship between transformational leadership and followers' dependence is mediated by personal identification with the leader. Hypothesis 3b: The relationship between transformational leadership and followers' empowerment is mediated by social identification with the organizational unit.
Method Sample and Data Collection
The study was carried out in a large Israeli banking organization and involved the leadership of midlevel managers of independent units. The sample included mostly bank branch managers, but in some large branches (over 40 employees) department managers with similar levels of autonomy and responsibility were included. The managers studied had between 8 and 30 subordinates (M 15). We decided not to include managers who had been in their roles for less than 6 months and employees who have been in the branches for less than 3 months to ensure sufficient acquain- tance of subordinates with the manager and to enable the development of identification, dependence, and empowerment. Average role tenure of the managers studied was 24 months (ranging from 6 months to 6 years). Data were collected by questionnaires from employees in 76 units. Thirty-eight units were managed by women. Nine hundred thirty-two employees participated in the study. Complete data were available from 888 employees,
250
KARK, SHAMIR, AND CHEN
working in a wide variety of roles within the branches (e.g., tellers, clerks, and investment consultants), representing a response rate of 89%. Sixtyseven percent of the participants were women, and participants' age varied from 21 years to 65 years.
Levels of Analysis
The outcome and mediating variables were examined as individual-level variables. Clearly, self-efficacy, OBSE, and personal identification are conceptually individual-level variables. Collective efficacy and social identification can both emerge as group-level variables, but they both originate at the individual level. Because we were interested in how leadership influences the origination of these outcomes, as opposed to their emergence to the group level, we examined them at the individual level of analysis (cf. Riggs & Knight, 1994). However, because evaluations of leadership and responses to leadership were collected from the same individuals by using the same questionnaire, relationships might be inflated because of common source and common method variance. To reduce this risk, we decided to treat leadership behaviors as a group-level variable. We based our analysis not on the relationships between leadership behaviors as perceived by each individual and his or her responses to these behaviors but rather on the relationships between the shared perception among members in a group regarding their leader's behaviors and their individual responses to these behaviors. That is, each bank branch was assigned a single transformational leadership, on the basis of the average leadership ratings within the branch. This is consistent with the theoretical argument that leaders direct many of their transformational behaviors to the entire group rather than to each individual and with the practice followed in other studies (e.g., Shamir et al., 1998).
Measures
Leader behaviors. We used 16 transformational leadership items from the short version of Bass and Avolio's Multifactor Leadership Questionnaire 5X (MLQ, Avolio et al., 1999), which included items of the following behavioral components: individualized consideration, intellectual stimulation, inspirational motivation, and idealized influence. The items measuring attributed charisma, which have been criticized for representing leadership impact and not leadership behavior (e.g., Yukl, 1998), were not included. Respondents were asked to mark the frequency with which the leader engages in each of the behaviors on a 5-point scale (ranging from 1 to 5; .93). All other measures used in this study were composed of a 7-point scale. Social identification. Eight items were adopted from identification measures developed by Mael and Ashforth (1992) and Shamir et al. (1998). The items focused on the branch or department to which the respondent belonged ( .80). A sample item is "I view the success of the branch as my own success." Personal identification. We used eight items that were similar to those used for the measurement of social identification, but in this case items focused on the manager of the branch or department ( .96). For instance, the item corresponding to the example item given above is "I view the success of the branch manager as my own success." Some of these items were previously used to measure identification with the leader by Shamir et al. (1998). Dependence on the leader. No published measure of dependence was found in the literature. An eight-item scale was developed for this study ( .84). A sample item is "Sometimes I find it difficult to do my job without the direction of the branch manager." Self-efficacy. We used a scale developed by Riggs and Knight (1994) to measure self-efficacy in the domain of work. The four positively worded items were chosen for this study ( .76). A sample item is "I have confidence in my ability to do my job." Collective efficacy. A nine-item measure was adapted from Riggs and Knight's (1994) measure of collective efficacy and from Guzzo, Yost, Campbell, and Shea's (1993) measure of group potency ( .87). The wording of the items from Guzzo et al. was changed to refer to the individual's perception of the unit efficacy, rather than to the shared perceptions of the group. A sample item is "The members of this department have excellent job skills." OBSE. The 10-item scale developed by Pierce et al. (1989) was used ( .93). A sample item is "I am an important part of this place."
Analysis Strategy
Random coefficient models (RCM), in conjunction with the mediation model standards described in Kenny, Kashy, and Bolger (1998), were used to test the hypotheses. Analyses were conducted using Version 3.0 of the Linear Mixed Effects program for S-PLUS and R written by Pinheiro and Bates (2000), which is statistically equivalent to other RCM programs, such as hierarchical linear modeling (HLM; see Chen & Bliese, 2002, and Jex & Bliese, 1999, for similar applications of the Linear Mixed Effects program for S-PLUS in multilevel research). RCM are well suited to analyses with predictors at different levels. Specifically, we can use RCM to regress individual-level dependent variables (e.g., dependence) on both individual-level predictors such as personal identification and group-level predictors such as average perceptions of transformational leadership. The RCM approach provides the correct parameter estimates and standard errors for these mixed, multilevel models (see Bliese, 2002; Bryk & Raudenbush, 1992). According to Kenny et al. (1998), four criteria need to be met to support mediated relationships. First, the independent variable (i.e., transformational leadership) needs to be related to the mediator (i.e., personal and social identification). Second, the independent variable needs to be related to the dependent variable. That is, transformational leadership needs to be related to dependence, OBSE, self-efficacy, and collective efficacy. Third, the mediators must be related to the dependent variables with the independent variable included in the model. Finally, the relationship between the independent variable and the criterion variable must disappear when controlling for the mediator variable. If the relationship between the independent and the dependent variable is reduced but remains significant in the presence of the mediator, there is evidence for partial mediation.
Construct Validity Evidence
To further examine the validity of our measures, we conducted confirmatory factor analyses in LISREL (Joreskog & Sorbom, 1993). These analyses compared the fit of a model in which measures of all seven factors (i.e., Transformational Leadership, Personal Identification, Social Identification, Dependence, Self-Efficacy, OBSE, and Collective Efficacy) were set to load on their respective factors with alternative, more constrained models in which certain factors (e.g., Transformational Leadership and Personal Identification) were set to load on a single factor. Results showed that the hypothesized seven-factor model fit the data significantly better than all alternative models, providing evidence for the validity and independence of our measures. More specific information regarding these analyses is available upon request from the authors.
Results Aggregation Analyses
Before testing the hypotheses, we first needed to justify the aggregation of individual perceptions of transformational leadership behaviors to the group level of analysis. Average rwg(j) across groups was .88, suggesting sufficient within-group agreement. In addition, ICC(1) was .25 and ICC(2) was .79, which provided
TWO FACES OF TRANSFORMATIONAL LEADERSHIP
251
sufficient evidence for between-group reliability. Finally, an analysis of variance indicated that individual perceptions of leadership significantly clustered by groups, F(76, 886) 4.85, p .01. These results provided sufficient statistical justification for aggregating individual perceptions of leadership to the group level (see Bliese, 2000).
Hypotheses Testing
Table 1 presents the individual-level descriptive statistics and zero-order correlations for the variables in the study. Relationships and significance tests associated with these variables should be viewed with caution until properly modeled in the RCM analyses, because the correlation table does not account for the fact that individual-level relationships might also be affected by the nonindependent nature of the data (Bliese, 2000). These caveats not withstanding, it is worth mentioning that the indicators of empowerment (self-efficacy, collective efficacy, and OBSE) were more highly related to each other (r .18 to .55) than to dependence (r .04 to .12). We address these findings in the Discussion section. Table 2 presents results of the RCM analyses following the steps suggested by Kenny et al. (1998) to test for mediated relationships. Note that the first step in these analyses also tested Hypotheses 1a and 1b, and the second step also tested Hypotheses 2a and 2b. Supporting Hypotheses 1a and 1b, results from the RCM analyses (see Table 2, Step 1) indicated that transformational leadership significantly and positively predicted both personal identification, coefficient 1.69, t(74) 12.46, p .01; and social identification, coefficient .45, t(74) 3.57, p .01. Also, supporting Hypothesis 2a, results from the second step (see Table 2) indicated that transformational leadership significantly predicted dependence, coefficient .49, t(74) 4.85, p .01. Additionally, lending support to Hypothesis 2b, transformational leadership significantly predicted OBSE, coefficient .18, t(74) 2.37, p .05, and collective efficacy, coefficient .35, t(74) 2.95, p .01, and marginally predicted self-efficacy, coefficient .10, t(74) 1.82, p .07. In order to gain support for Hypotheses 3a and 3b, the relationships between transformational leadership and dependence and transformational leadership and empowerment must disappear when including the identification variables in the equation. To ensure that personal and social identification predicted the outcomes differentially as predicted, we entered personal identifica-
tion, social identification, and transformational leadership simultaneously into the equations (see Table 2, Steps 3 and 4). Supporting Hypothesis 3a, personal identification positively and significantly predicted dependence, coefficient .31, t(810) 12.95, p .01, whereas the relationship between transformational leadership and dependence became nonsignificant, coefficient .07, t(74) 0.72, ns. Unexpectedly, social identification predicted dependence significantly and negatively, coefficient .21, t(810) 5.83, p .01, in this analysis. However, this unexpected finding seems to have been caused by the strong correlation between personal and social identification (i.e., due to multicollinearity). Indeed, when only social identification and transformational leadership were entered into the equation, transformational leadership significantly predicted dependence, coefficient .46, t(74) 4.50, p .01, whereas social identification did not significantly predict dependence, coefficient .05, t(811) 1.45, ns. In contrast, as hypothesized, when only personal identification and transformational leadership were entered into the equation, transformational leadership did not significantly predict dependence, coefficient .10, t(74) 0.96, ns, whereas personal identification significantly predicted dependence, coefficient .23, t(811) 11.52, p .01. These results support Hypothesis 3a, suggesting that personal identification but not social identification, mediates the relationship between transformational leadership and dependence. Tests of Hypothesis 3b involved the three indicators of empowerment—OBSE, self-efficacy, and collective efficacy. Results indicated that when entered simultaneously into the equation, both personal identification, coefficient .07, t(810) 3.93, p .01, and social identification, coefficient .20, t(810) 7.22, p .01, significantly predicted OBSE, whereas transformational leadership did not significantly predict OBSE, coefficient .03, t(74) 0.49, ns. As expected, social identification predicted OBSE to a greater extent than did personal identification. In addition, social identification significantly predicted selfefficacy, coefficient .16, t(810) 6.51, p .01, whereas personal identification, coefficient .02, t(810) 1.11, ns, and transformational leadership, coefficient .01, t(74) 0.17, ns, did not significantly predict self-efficacy. Likewise, social identification significantly predicted collective efficacy, coefficient .42, t(810) 13.12, p .01, whereas personal identification did not significantly predict collective efficacy, coefficient .00, t(810) 0.07, ns, and transformational leadership only marginally predicted collective efficacy, coefficient .17,
Table 1 Individual-Level Means, Standard Deviations, and Correlations
Variable 1. 2. 3. 4. 5. 6. Transformational leadership Personal identification Social identification Dependence Organization-based self-esteem Self-efficacy 888. All correlations M 3.55 4.54 5.54 2.33 6.00 6.26 SD 0.81 1.78 1.10 1.07 0.80 0.70 1 — .73 .43 .34 .23 .13 .01. 2 — .57 .40 .31 .19 3 4 5 6 7
— .07 .37 .29
— .04 .12
— .55
—7.
Collective efficacy
5.58
0.9
Note. N
.06 are significant, p
252
KARK, SHAMIR, AND CHEN
Table 2 Mediation of the Relationship of Transformational Leadership With Dependence and Empowerment by Personal and Social Identification
Dependent variable Independent variable Step 1 Transformational leadership Personal identification 1.69** (74) Dependent variable Dependence Step 2 Transformational leadership Steps 3 and 4 Transformational leadership Personal identification Social identification .49** (74) .07 (74) .31** (810) .21** (810) OBSE .18* (74) .03 (74) .07** (810) .20** (810) Self-efficacy .10† (74) .01 (74) .02 (810) .16** (810) Collective efficacy .35** (74) .17† (74) .00 (810) .42** (810) Social identification .45** (74)
Note. Results are unstandardized parameter estimates from random coefficient models analyses; numbers in parentheses are corresponding degrees of freedom. OBSE organization-based self-esteem. † p .10 (marginally significant). * p .05. ** p .01.
t(74) 1.70, p .09. These results support Hypothesis 3b, suggesting that social identification, but not personal identification, mediates the relationship between transformational leadership and empowerment. Figure 2 provides a summary of the results from the hypotheses tests. Note that the slight discrepancies between the results presented in Table 2 and Figure 2 stem from the fact that in the final analyses reported in Table 2 (Steps 3 and 4), both personal and social identification were included in analyses involving dependence and empowerment as the dependent variables. However, in the analyses reported in Figure 2, we included personal identifi-
cation (but not social identification) in analyses of dependence as the dependent variable, and social identification (but not personal identification) in analyses of the empowerment dimensions as the dependent variables. As shown in the figure, supporting Hypotheses 1a and 1b, transformational leadership positively and significantly predicted personal and social identification, respectively. Supporting Hypotheses 2a and 3a, personal identification predicted dependence over and above transformational leadership, whereas the significant relationship between transformational leadership and dependence became nonsignificant once controlling for personal identification. Finally, in support of Hypotheses 2b and 3b,
Figure 2. Summary of random coefficient modeling results. Numbers in parentheses are degrees of freedom. DV dependent variable. † p .10 (marginally significant). * p .05.
TWO FACES OF TRANSFORMATIONAL LEADERSHIP
253
social identification predicted the three empowerment variables over and above transformational leadership, whereas the significant relationships of transformational leadership with the empowerment variables became nonsignificant (when examining selfefficacy and OBSE) or marginally significant (when examin- ing collective efficacy) once controlling for social identification. In sum, results provided support for the notion that personal identification mediates the relationship between transformational leadership and dependence, whereas social identification mediates the relationship between transformational leadership and empowerment.
Discussion
The main goal of this study was to examine personal and social identification as possible mechanisms through which transformational leadership differentially relates to followers' dependence and empowerment. The results of this research offer several notable findings. First, as predicted, transformational leadership was found to be positively related to both personal identification with the leader and social identification with the work unit, thus providing evidence that transformational leaders are likely to exert their influence on followers by affecting their feelings of identification. Prima facie, it is interesting that transformational leadership was more strongly associated with personal identification with the leader than with social identification. However, this finding is not surprising because the latter is likely to be influenced by many other factors besides the leader's personality and behavior, such as the composition of the group, its history, and its cohesion. Although leaders can influence members' identification with the group, their impact on this identification is more limited than their impact on followers' personal identification. However, this finding suggests that prior assertions in the literature that portrayed transformational leaders as socialized leaders, assuming that their main avenue of influence may not rely on personal identification, should be reexamined. Second, although we have not covered the full range of factors that are included in some broad conceptions of empowerment (e.g., Spreitzer, 1995), results of this study confirm the hypothesized relationship between transformational leadership and some aspects of empowerment, represented by followers' self- and collectiveefficacy beliefs and their OBSE. More important, our results suggest that contrary to the common dichotomy in the literature (e.g., Howell, 1988), follower empowerment and follower dependence are not opposite to each other. Rather, our findings suggest that dependence and empowerment are independent of each other, as the correlations between dependence and the empowerment indicators were rather weak. Furthermore, because both dependence and empowerment were positively related to transformational leadership, our results suggest that, contrary to the common assumption that different types of leadership (e.g., charismatic vs. transformational) lead to followers' dependence or empowerment, the same type of leadership may be associated simultaneously with both empowerment and dependence. Consistent with this conclusion, results suggest that different mechanisms (personal or social identification) can, at least partially, account for these different outcomes (empowerment and dependence of followers) of transformational leadership. As expected, social identification with the work unit mediated the relationships between transformational
leadership and indicators of empowerment, whereas personal identification with the leader mediated the relationship between transformational leadership and dependence. It may be noted, in this regard, that some feminist relational perspectives (e.g., Fletcher, 1999; Miller & Stiver, 1997) challenge the commonly accepted distinction and opposition between dependence and empowerment, claiming that this distinction rests on traditional masculine models of relationships. Traditional models view the development of independence and self-reliance as including disengagement from significant others and as negating strong forms of relatedness, which include interdependence. In contrast, the feminist perspectives present a model of development and growth within a relationship, which emphasizes both dependence and independence as fluid states that are neither static nor achieved. Within such a model, a relationship may be characterized by both a sense of empowerment and, at the same time, at least some sense of interpersonal dependence. Apparently, the relationships between followers' dependence and empowerment and the relationships between transformational leadership and these states, as well as social and personal identification, are more complex than suggested by the literature. Further research should clarify these relationships by examining whether other leadership behaviors, particularly those associated with the "personalized charisma" (Howell, 1988) type may lead to dependence without empowerment and whether the mixture of empowerment and dependence varies between different stages of the follower-leader relationship. For instance, it is possible that in early stages of the relationship some dependence on the leader is a necessary condition for the leader's empowering effects, whereas in later stages the empowerment effects depend on the followers achieving independence from the leader and on their need for affirmation and recognition. It is also possible that followers will feel empowered in some aspects of their work (e.g., giving service to customers) and dependant on the leader's expertise and experience in other aspects of their work (e.g., decision-making processes). Furthermore, personality differences among followers (e.g., their tendencies to become dependent or empowered) and the ways in which they interact with the leader might affect the leaders and account for leaders' ability to influence in diverse ways (Dvir & Shamir, 2001). Moreover, the consequences of personal identification with the leader, social identification with the group, and dependency on the leader might be entirely different in the cases of moral and immoral leaders. In addition, it should be acknowledged that identifications, empowerment, and dependence do not depend only on the leader's behavior. For instance, Bion (1974) described the development of dependence perceptions in groups that do not depend entirely on the leader's actions. Guzzo and Shea (1992), among others, have drawn attention to the structure of the group task as a factor influencing group processes. This work suggests that high task interdependence might influence the level of social identification in groups. Future studies should look at the impact of the leader on followers' identifications, empowerment, and dependency against the background of other contextual factors that might affect these variables. We also encourage researchers to examine the broader conceptualization of the concept of empowerment (Spreitzer, 1995) in future research, in order to enable a more complete conclusion regarding the relationships among transformational leadership, identification, and empowerment.
254
KARK, SHAMIR, AND CHEN
The findings of this study can also be interpreted from a perspective recently suggested by Lord, Brown, and Freiberg (1999) and by Kark and Shamir (2002). According to this perspective, leaders achieve their effects on followers by priming different aspects of their self-concepts. Lord et al. viewed the self-concept as having three aspects: a personal self, which is based on individual attributes and preferences; a relational self, which is based on close relations with specific others (such as a leader); and a collective self, which is based on group affiliations and identifications. Applying this perspective to our study, one can note that transformational leadership may prime both the relational self (evidenced by personal identification with the leader) and the collective self (evidenced by social identification with the unit). These two faces of transformational leadership have different consequences. Priming the relational self results in dependence on the leader or follower-leader interdependence, whereas priming the collective self results in empowerment (for a more comprehensive theoretical model, see Kark & Shamir, 2002). Further elaboration of transformational leadership theory and measurement is needed in order to identify specific leader behaviors that prime different aspects of followers' self-concepts. The results of this study are limited by its cross-sectional design and by the use of a single method of data collection. Although researchers have shown that common method bias is rarely strong enough to invalidate research findings (e.g., Doty & Glick, 1998), replications and extensions of our findings using experimental and longitudinal designs are needed. Furthermore, the specific banking context in which the research took place could also limit the generalization of the research conclusions. However, in this regard, it is interesting to note that bank culture is not very conducive to transformational leadership (e.g., it is not a highly ideological setting) and middle-level branch managers, who are constrained by bank procedures, are possibly limited in their ability to present the "full range" of transformational behaviors. In addition, branches are perhaps not the best foci for social identification, and perhaps stronger relationships between leadership behavior, social identification, and empowerment would have been found in situations in which the work unit and the nature of the work is more attractive and can arouse higher levels of identification. These considerations suggest that even stronger results than the ones found in the present study could be found in other settings more conducive to transformational leadership (e.g., military units). Despite its limitations, the study has made several tentative contributions to the literature on transformational leadership in organizations. In general, the results of the study provide, perhaps for the first time, fairly strong support for the hypothesized relationships between transformational leadership, followers' identification with the leader and the organizational unit, dependence on the leader, and certain indicators of followers' empowerment. These results strengthen recent experimental findings showing that leaders who received training in transformational leadership succeeded in obtaining higher levels of commitment, motivation and performance among followers (Barling et al., 1996; Dvir et al., 2002). Furthermore, support was also found for the hypothesized differential roles of personal and social identification in mediating the relationships between transformational leadership and followers' empowerment and dependence. Our results start to shed some light on the influence processes by which transformational leadersmay achieve their superior results. However, future research
should examine further whether followers' identification, dependence, and empowerment indeed mediate the relationships between transformational leadership and more distal effects on followers' behavior and organizational outcomes.
References
Alderfer, C. P. (1987). An intergroup perspective on group dynamics. In J. W. Lorsch (Ed.), Handbook of organizational behavior (pp. 190 222). Englewood Cliffs, NJ: Prentice Hall. Ashforth, B. E., & Mael, F. (1989). Social identity theory and the organization. Academy of Management Review, 14, 20 -39. Avolio, B. J., Bass, B. M., & Jung, D. I. (1999). Reexamining the components of transformational and transactional leadership using the multifactor leadership questionnaire. Journal of Occupational and Organizational Psychology, 72, 441- 462. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall. Barling, J., Weber, T., & Kelloway, E. K. (1996). Effects of transformational leadership training on attitudinal and financial outcomes: A field experiment. Journal of Applied Psychology, 81, 827- 832. Bass, B. M. (1985). Leadership and performance beyond expectation. New York: Free Press. Bass, B. M. (1997). Does the transactional-transformational leadership paradigm transcend organizational and national boundaries? American Psychologist, 52, 130 -139. Bion W. R. (1974). Experience in groups, and other papers. London: Tavistock. Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analyses. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 349 381). San Francisco: Jossey-Bass. Bliese, P. D. (2002). Using multilevel random coefficient modeling in organizational research. In F. Drasgow & N. Schmitt (Eds.), Advances in measurement and data analysis (pp. 401- 445). San Francisco: JosseyBass. Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage. Chen, G., & Bliese, P. D. (2002). The role of different levels of leadership in predicting self- and collective efficacy: Evidence for discontinuity. Journal of Applied Psychology, 87, 549 -556. Conger, J. A., & Kanungo, R. N. (1988). The empowerment process: Integrating theory and practice. Academy of Management Review, 3, 471- 482. Conger, J. A., & Kanungo, R. N. (1998). Charismatic leadership in organizations. Thousand Oaks, CA: Sage. Doty, D. H., & Glick, W. H. (1998). Common methods bias: Does common methods variance really bias results? Organizational Research Methods, 1, 374 - 406. Dvir, T., Eden, D., Avolio, B. J., & Shamir, B. (2002). Impact of transformational leadership on follower development and performance: A field experiment. Academy of Management Journal, 45, 735-744. Dvir, T., & Shamir, B. (2001, June). The role of follower characteristics in transformational leadership theory. Paper presented at the International Conference on Business, Honolulu, Hawaii. Fletcher, J. K. (1999). Disappearing acts: Gender, power, and relational practice at work. Cambridge, MA: MIT Press. Gardner, D. G., & Pierce, J. L. (1998). Self-esteem and self-efficacy within the organizational context: An empirical examination. Group and Organization Management, 23, 48 -70. Guzzo, R., & Shea, G. P. (1992). Group performance and intergroup relations in organizations. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 3, 2nd ed., pp. 269 -313). Palo Alto, CA: Consulting Psychologists Press, Inc.
TWO FACES OF TRANSFORMATIONAL LEADERSHIP Guzzo, R. A., Yost, P. R., Campbell, R. J., & Shea, G. P. (1993). Potency in groups: Articulating a construct. British Journal of Social Psychology, 32, 87-106. Hackman, J. R. (1992). Group influences on individuals in organizations. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 3, pp. 199 -267). Palo Alto, CA: Consulting Psychologists Press. Hogg, M. A., & Abrams, D. (1990). Social motivation, self-esteem and social identity. In D. Abrams & M. A. Hogg (Eds.), Social identity theory: Constructive and critical advances (pp. 28 - 47). New York: Harvester Wheatsheaf. Howell, J. M. (1988). Two faces of charisma: Socialized and personalized leadership in organizations. In J. A. Conger & R. N. Kanungo (Eds.), Charismatic leadership (pp. 213-236). San Francisco: Jossey-Bass. Jex, S. M., & Bliese, P. D. (1999). Efficacy beliefs as a moderator of the impact of work-related stressors: A multilevel study. Journal of Applied Psychology, 84, 349 -361. Joreskog, K., & Sorbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS command language. Hillsdale, NJ: Erlbaum. Kark, R., & Shamir, B. (2002). The dual effect of transformational leadership: Priming relational and collective selves and further effects on followers. In B. J. Avolio & F. J. Yammarino (Eds.), Transformational and charismatic leadership: The road ahead (Vol. 2, pp. 67-91). Amsterdam: JAI Press. Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., pp. 233-265). Boston: McGrawHill. Kets de Vries, M. F. (1988). Origins of charisma: Ties that bind the leader and the led. In J. A. Conger & R. N. Kanungo (Eds.), Charismatic leadership (pp. 237-252). San-Francisco: Jossey-Bass. Lord, R. G., Brown, D. J., & Freiberg, S. J. (1999). Understanding the dynamics of leadership: The role of follower self-concepts in the leader/ follower relationship. Organizational Behavior and Human DecisionProcess, 78, 167-203. Lowe, K. B., Kroeck, K. G., & Sivasubramaniam, N. (1996). Effectiveness of correlates of transformational and transactional leadership: A metaanalytic review of the MLQ literature. Leadership Quarterly, 7, 385425. Mael, F., &Ashforth, B. E. (1992). Alumni and their alma matter: A partial test of the reformulated model of organizational identification. Journal of Organizational Behavior, 13, 103-123.
255
Miller, J. B., & Stiver, I. P. (1997). The healing connection. Boston: Beacon Press. Pierce, J. L., Gardner, D. G., Cummings, L. L., & Dunham, R. B. (1989). Organization-based self-esteem: Construct definition, measurement, and validation. Academy of Management Journal, 32, 622- 648. Pinheiro, J. C., & Bates, D. M. (2000). Mixed-effects models in S and S-PLUS. New York: Springer-Verlag. Pratt, M. G. (1998). To be or not to be: Central questions in organizational identification. In D. A. Whetten & P. C. Godfrey (Eds.), Identity in organizations: Building theory through conversation (pp. 171-207). Thousand Oaks, CA: Sage. Riggs, M. L., & Knight, P. A. (1994). The impact of perceived group success-failure on motivational beliefs and attitudes. Journal of Applied Psychology, 79, 755-766. Shamir, B. (1990). Calculations, values, and identities: The sources of collectivistic work motivation. Human Relations, 43, 313-332. Shamir, B. (1991). The charismatic relationship: Alternative explanations and predictions. Leadership Quarterly, 2, 81-104. Shamir, B., House, R. J., & Arthur, M. B. (1993). The motivational effects of charismatic leadership: A self-concept based theory. Organization Science, 4, 577-593. Shamir, B., Zakay, E., Breinin, E & Popper, M. (1998). Correlates of charismatic leader behavior in military units: Subordinates' attitudes, unit characteristics, and superiors' appraisals of leader performance. Academy of Management Journal, 41, 387- 409. Shamir, B., Zakay, E., Breinin, E., & Popper, M. (2000). Leadership and social identification in military units. Journal of Applied Social Psychology, 30, 612- 640. Spreitzer, G. M. (1995). Individual empowerment in the workplace: Dimensions, measurement, validation. Academy of Management Jour- nal, 38, 1442-1465. Tajfel, H. (1982). Social identity and intergroup relations. Cambridge, England: Cambridge University Press. Waldman, D. A., Ramirez, G. G., House, R. J., & Puranam, P. (2001). Does leadership matter? CEO leadership attributes and profitability under conditions of perceived environmental uncertainty. Academy of Management Journal, 44, 134 -143. Yukl, G. (1998). Leadership in organizations. Englewood Cliffs, NJ: Prentice Hall.
doc_378500839.docx