تعدیل نقش نگرش زیردستان در رهبری تحول گرا و اثر بخشی: چشم انداز چند فرهنگی و چند سطح

We conducted the current multi-cultural, multi-level study with three objectives in mind. First, we examined the association of transformational leadership and leader effectiveness across two different cultures, the United States (U.S.) and Korea, with three followers' attitudes toward their leader (trust in the leader, loyalty, and value congruence) as moderators of this association. Second, we conceptualized followers' collectivistic orientation as a personal value, regardless of their nationality, and tested its moderating effect on the leadership–effectiveness association. Finally, we employed a multi-level approach throughout to examine whether transformational leadership processes, and in contrast individualized leadership, operated at different levels of analysis across different countries. Results from WABA and regression analyses indicated that transformational leadership operated at the individual level of analysis in both samples and was effective across different cultures. Specifically, followers' attitudes served as moderators of the transformational leadership–effectiveness association only in the U.S. sample, while collectivistic orientation had a significant moderating effect in both samples. Results from WABA also indicated that individualized leadership operated at the dyadic level for the U.S. sample and at the individual level for the Korean sample. Several theoretical and practical implications of the results are discussed.

We conducted the present study with several objectives in mind. First, we wanted to test whether transformational leadership had a positive effect on leadership effectiveness across two different cultures. Second, we also tested whether several attitudes followers held toward their leaders such as trust, loyalty, and value congruence, had a consistent moderating effect on leadership effectiveness across different cultures. Third, we examined whether the effect of transformational leadership would be facilitated when followers possessed collectivistic personal values. Finally, we incorporated a meso, multi-level framework in the current study (e.g., Hunt, 1991, Hunt, 2004, Yammarino and Dansereau, 2008 and Yammarino et al., 2005) and tested whether transformational leadership, and in contrast individualized leadership, operated at different levels of analysis across the two countries from which we collected data. While operating at the individual level of analysis, the results of regression analyses showed that transformational leadership had a positive effect on leader effectiveness in both samples, which supported the universality of transformational leadership as argued by Bass (1997) and Dorfman et al. (1997). Moreover, the regression coefficient was stronger in the Korean sample than in the U.S. sample, supporting Jung et al. (1995) and Pillai and Meindl (1998) who argued that collectivistic culture would enhance the effect of transformational leadership. The regression results also showed an interesting pattern for the moderation effects, also at the individual level of analysis. For the U.S. sample, transformational leadership, but not the subordinate attitudes, had a positive and significant effect on effectiveness directly. Moreover, trust interacted with transformational leadership and had a significant moderating effect, while value congruence had a marginally significant moderating effect, on leader effectiveness. In contrast, transformational leadership and all three subordinate attitudes had a positive and significant effect on leader effectiveness directly in the Korean sample. However, the interaction terms and moderating effects were not significant in this sample. The moderating effects of followers' attitudes toward their leader we found in the U.S. sample were consistent with several studies conducted in the past. Some scholars (Podsakoff et al., 1990) argued that these constructs needed to be considered mediators while other scholars (e.g., Dirks, 2000) took a position of viewing the variables as moderators. We believe, as discussed above, that one reason for these different views on followers' attitudinal variables in the leadership process might be that they indeed can serve as both mediators and moderators in practice. Therefore, we conducted a post hoc analysis and found no mediation effects, only moderation effects, for the variables of interest. What is more important, we believe, is to conduct a systematic set of studies and examine various situations so that we could identify different circumstances in which followers' attitudes serve as either moderators or mediators. For example, followers' attitudes are more likely to serve as mediators rather than moderators when leadership outcomes being measured require followers' extra-role behaviors such as OCB. Additionally, when followers' tasks require an extraordinary level of commitment, risk-taking behaviors, and/or intrinsic motivation such as developing new products or working on unconventional and risky projects, followers' attitudes could serve as mediators. And, as noted above, based on idiosyncratic credit leadership theory (Hollander, 1961), the length of time in relationships among leaders and followers may impact whether these constructs operate as mediators or moderators. Likewise, research design issues (e.g., lab versus field studies, or contrived versus real settings) might impact whether the obtained empirical results support mediation versus moderation in any particular study. Such systematic research efforts will allow both researchers and practitioners to shed more light on what and how certain leadership styles become more effective. Our results seem to suggest that the effect of transformational leadership on work outcomes is so profound and powerful in collectivistic cultures that it would be effective regardless of followers' attitudes toward their leader. Jung et al. (1995, p. 11) explained the profound effect that transformational leadership might have on followers' motivation by saying “since subordinates are highly motivated to accomplish organizational goals, willing to put extra effort, and share common goals… transformational leaders in collectivistic cultures are more likely to motivate followers to work for transcendental goals instead of immediate self-interests.” Bass (1997) also argued that transformational leadership may be far more pervasive in collectivistic societies than in individualistic ones for the same reason. In contrast, the significant moderating effect of trust and value congruence for the U.S. sample might suggest that the effect of transformational leadership on leader effectiveness would be greatly enhanced if followers establish a high level of trust and share a similar set of personal values. Given these arguments, we were surprised to find that transformational leadership and followers' attitudes toward their leader were highly correlated in the U.S. sample, whereas the correlations in the Korean sample were relatively low except for the transformational leadership-value congruence association (r = .20). We expected, based on prior work, to find a stronger association between transformational leadership and followers' attitudes in the Korea than the U.S. sample. These findings might have been due to the fact that each focal leader's leadership was evaluated by his/her superior. More specifically, the focal leaders in a more hierarchically-oriented country like Korea might have a stronger tendency and desire to create a positive perception about their leadership style toward their superior, which could result in different attitudes from their followers. Future research is certainly needed to draw a firmer conclusion on impression management and followers' attitudes toward their leader. The positive moderating effect of collectivistic orientation was also supported when we combined both samples and tested it as a personal value at the individual level of analysis. This result suggests that collectivistic followers, regardless of their cultural background or nationality, are more likely to accept a leader's challenge to put organizational objectives ahead of their personal ones, focus on teamwork, and embrace a collective vision and identity (Pillai & Meindl, 1998). Shamir et al. (Dvir and Shamir, 2003, Shamir, 1990 and Shamir et al., 1998) have argued that charismatic leadership effects on leader performance should be enhanced when followers hold collectivistic values since charismatic leaders tend to raise the salience of the collective identity in followers' self-concept. Given the current finding as well as theoretical arguments made by a number of scholars noted above, it seems safe to conclude that collectivistic orientation, either as a cultural value or personal disposition, would enhance the positive effect of transformational leadership on work outcomes. This conclusion has an important practical implication in terms of establishing a well-balanced and consistent approach between leadership style and organizational culture. Specifically, those leaders who try to practice transformational leadership might find it beneficial to choose a setting where followers hold more collectivistic and group-oriented values. Coupled with several decade-long efforts of team-based organizational restructuring processes in many industries (Katzenbach, 1997), top executives will find their transformational leadership creating a larger and more positive effect on leadership effectiveness when the organizational culture emphasizes collective responsibility and interdependent relationships (Pillai & Meindl, 1998). Stated alternatively, when top management tries to motivate followers through transformational leadership while the dominant organizational culture still emphasizes individualistic values, such a culture might create a boundary condition under which the potential effect of transformational leadership on followers' effectiveness and firm performance is limited. Finally, to better understand leader–subordinate relationships via alternative mechanisms as well as rule out effects that are simply artifacts of the samples, we tested whether individualized leadership, in contrast to transformational leadership, would operate at different levels of analysis across the two different national cultures. The results of WABA indicated (weak) whole dyadic effects for the U.S. sample but an individual differences effect for the Korean sample. While the dyadic effect made theoretical and intuitive sense given a strong individualistic cultural orientation among Americans, we were puzzled by the individual-level (rather than group-level) effect we found for the Korean sample. Two potential explanations are plausible. One, as some cross-cultural researchers have recently argued, due to a recent trend toward globalization, people in many Asian countries began embracing more individualistic and Western cultural values (Oyserman, Coon, & Kemmelmeier, 2002). Since we drew our samples from part-time MBA students, these leaders might have become even more Westernized than the overall population in Korea. As such, cross-cultural differences in management practices may no longer exist or do not make a meaningful difference. For example, Jung, Sosik, Chow, and Chen (2007) recently found that benevolent leader behavior, which was considered to be a fundamental and essential leader behavior in China by many cross-cultural leadership scholars (e.g., Tsui, Wang, Xin, Zhang, & Fu, 2003), had a negative influence on firm performance. Alternatively, as another potential explanation, it is possible that many of Korean participants did not have enough time as a group to interact with their leader and develop a homogeneous perception about their leader. The average length of tenure for the current job that the focal leaders had was only 1.7 years in the Korean sample. It was longer, 2.7 years, in the U.S. sample. Future research is certainly needed to test whether the length of interaction time and intensity of the interaction between the leader and followers in collectivistic cultures might allow them to establish their leadership perceptions at the whole group level as we expected. Moreover, we agree with Yammarino et al. (2005) and Yammarino and Dansereau (2008) who note the importance of explicitly including levels of analysis issues in theory/hypothesis development, measurement, data analysis, and inference drawing. Such an approach will greatly benefit future research on both transformational and individualized leadership. 4.1. Limitations and conclusions Although the current study found some interesting results, there are several limitations that may limit its generalizability to other settings, and thus readers need to exercise caution when interpreting our findings. First, our samples came from several external and offsite MBA programs in both countries. Although there are a number of studies that utilized MBA students as focal leaders (e.g., Sosik, 2005), they represent a group of “unusual” subjects in the sense that all of them were working on a master's degree in business administration at the time of their participation. Therefore, we would like to advise our readers to exercise caution against over-generalizing our findings here. The current findings may not be applicable to settings where employees are less educated or nonprofessionals. Second, in terms of our research design, since the focal leaders chose the three subordinates who evaluated leadership styles and their attitudes toward their leader, there is a possibility for sampling bias. Although we emphasized the developmental and non-evaluative nature of our survey, it was certainly possible for the focal leaders to select their favorite subordinates. Additionally, even though we collected our data from multiple sources as a way to decrease common-source bias, we relied on surveys exclusively, which might induce a common-method bias into our findings. Therefore, future research is needed with more independent-source and objective data (e.g., objective performance measures) before we can draw firmer conclusions. Third, although some of the more rigorous and conservative multi-level analyses (i.e., based on WABA) indicated that transformational leadership operated solely at the individual level, we nevertheless aggregated some measures to the focal (individual) leader level, using less rigorous yet acceptable aggregation criteria (i.e., based on ICC) to further test our substantive ideas. As such, these focal (individual) leader-level results should be viewed as tentative and subject to additional future research to clarify these effects. Fourth, the mean length of time that our subordinates worked for their focal leader was relatively short, especially in the Korean sample. Although we believe that an average of about 2 years for working relationships would be long enough to develop homogeneous perceptions regarding their leader, we are not sure that all subordinates who evaluated the same leader belonged to the same work team all of that time and interacted with their leader long enough. Therefore, future research might replicate our study and test the levels of analysis issues regarding individualized leadership in both countries while controlling the team effect with a longer history of interaction. Fifth, we advise readers to interpret our finding with caution because there might be a problem with measurement equivalence. Although there have been a number of studies on transformational leadership in various Asian settings (e.g., Dorfman et al., 1997, Shin and Zhou, 2003 and Zhu et al., 2005), and we went through a back-translation approach recommended by Brislin (1986) several times to make sure that our translation was done as carefully as possible, such procedures do not guarantee measurement equivalence across the two different cultures from which we drew our samples. Finally, while it could be considered beneficial in terms of external validity, because our samples came from a number of different companies in mainly high-tech industries, it would be interesting to examine our hypotheses in a single large company where organizational culture would be more consistent and employees would share common values. This approach could serve as another control or check on our findings and their generalizability. In conclusion, as the world becomes more globalized and managers have to develop a more global perspective, it is important to define some boundary conditions for which some types of leadership styles might become more or less effective (Dickson, Den Hartog, & Michelson, 2003). Our study found that several attitudes that followers developed toward their leader had a positive and significant moderating effect on leadership effectiveness only in the U.S. sample. Moreover, when we treated collectivistic orientation as a personal value at the individual level, it became a significant moderator across the two samples. These results demonstrate the importance of boundary conditions under which effects of these important moderators in leadership processes could vary. As we shed more light on different cultural, structural, and environmental boundary conditions and examine how they affect different leadership styles and outcomes at different levels of analysis, we will find ourselves “inside of the black box” and have a deeper knowledge of and insight on why and how some leaders create extraordinary levels of outcomes in organizations.