本研究指出，雙重層次之轉換型領導，會分別透過不同層次主管與部屬 交換關係對不同層次員工服務績效產生影響。為避免共同方法變異，本研究 取樣自台中地區一間大型商業銀行，共計回收計 23 間分行，共組成有效的 228 組主管─成員對偶樣本資料。本研究發現，雙重層次之轉換型領導對於不 同層次員工服務績效均有顯著性之影響。其中，個人層次之轉換型領導透過 主管與部屬交換關係對員工服務績效產生影響。群體層次之轉換型領導透過 群體層次主管與部屬交換關係對群體層次員工服務績效產生影響。此外，本 研究亦指出，主管與部屬交換關係差異性對兩個層次主管與部屬交換關係以 及員工服務績效之間具有顯著負向干擾效果。本研究最後，亦提出理論與實 務上之管理意涵以及未來研究方向。

Conceptualizing superior customer service as behaviors exhibited by people who have direct interactions with customers in the workplace, we proposed a multilevel theoretical model that identified transformational leadership as the primary source, and the quality of leader-member exchange (LMX) as the underlying mechanism through which leadership exerts its positive influence on employee service performance, at both the individual and group levels. We also identified LMX differentiation within the group as a moderator between LMX quality and service performance. Data from 228 leader-member dyads of 23 branches in a large commercial bank revealed that both group-level transformational leadership and group LMX were meaningful group-level constructs that had significant positive effects on group service performance. More intriguingly, LMX differentiation played the role of a cross-level moderator that mitigated such relationship at the group level. The theoretical and practical implications of our findings are discussed in the service organizational context.

In this current study, we conceptualize superior customer service as behaviors exhibited by people who have direct interactions with customers in the workplace, proposing a multilevel theoretical model that identified transformational leadership as the primary source, and the quality of leader-member exchange (LMX) as the underlying mechanism through which leadership exerts its positive influence on employee service performance, at both the individual and group levels. We also identified LMX differentiation within the group as a moderator between LMX quality and service performance. Data from 228 leader-member dyads of 23 branches in a large commercial bank revealed that both group-level transformational leadership and group LMX were meaningful group-level constructs that had significant positive effects on group service performance. More intriguingly, LMX differentiation played the role of a cross-level moderator that mitigated such relationship at the group level. Our study has several managerial implications. First, our study indicates that transformational leaders need to form a set of norms, values, and ideology to motivate performance of group as a whole. The norms, values, and ideology providing group members with the information and understanding of their group reality are treated as the accepted behavioral standards upon which all group members are subsequently measured. The norms that are formed among group members can be regarded as taken-for-granted organizational routines. These norms can facilitate both the rapid fulfillment of tasks, and the desire for conformity among the members of the group. To drive group service performance, transformational leaders need to foster group identity by emphasizing the shared norms and values and the unique characteristics of the group. Second, transformational leaders must also be capable of forming good-quality LMX with subordinates. If they are to succeed in building up such relationships, transformational leaders will need to be able to sketch out the core value of organizational vision for their subordinates and to provide them with an effective link with their own conception of self through personalized role assignments. As a result of such work and non-work relationships, followers can obtain their perceived equity within the organization, and can then go on to further identify with the vision and values of their transformational leader. Thus, LMX is established, characterized by trust, loyalty and commitment, a form of social currency circulated in these social exchanges, within which subordinates feel some obligation to reciprocate through enhanced performance. Third, LMX differentiation plays a significant moderating role in the relationship between LMX and service performance at both levels. Our results suggest forming high-quality LMX relationship with a small portion of members may be disadvantages. Although high-quality LMX is positively related to service performance, leaders may often find themselves in a situation in which they cannot develop high-quality relationships with all members. Our results have implications for leaders with high LMX differentiation. If leaders realize that their relationships with followers are differentiated such that they are closer to some members, they should be aware that other members’ withdrawal behaviors will depends on the levels of LMX differentiation that exists in the group. The present study contributes to the literature on leadership and employee service performance by examining the mediating role of LMX using multisource, multi-level data in a cross-sectional design. In addition, the study extends the leadership literature both theoretically and empirically by demonstrating the existence of the group-level transformational leadership and group LMX constructs, the moderating role of LMX differentiation on the LMX-service performance relationship, and by proposing and testing a model that supports the integration of social exchange theory and social capital theory.