Chiao Da Managemenl Reνlew 均 1.32 No. 1.2012 pp.163-195
雙重層次轉換型領導、主管與部屬交
關條、與群體凝眾力對員工服務績效之
影響
Examining the Effects of Dual-level Transformational
Leadership,
LMX,
and Group Cohesiveness on
Employee
Service
Performance
翁良杰 I Liang-Chieh Weng 靜宜大學國際企業學系
Department ofIntemational Business, Providence University 賴蓋筑 Yi-ChuLai
國立中興大學企業管理學系
Department ofBusiness Adminis甘ation，National Chung-Hsing University
摘要:本研究指出 ，雙重層次之轉換型領導，分別通過主管與部屬交換關條 以及群體凝眾力對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 換關條、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凝眾力對 員工服務績效產生影響 。 此外，本研究亦指出，群體凝毅力對於主管與部屬 交換關靜、以及員工服務績效之間具有顯著正向干擾效果 。 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 。 關鍵字:雙重層次轉換型領導;主管與部屬交換關條群體凝盟主力 ;線性層 級模式分析
I Co叮esponding author: Department of Intemational Business, Providence University E-mail:Icweng@pu.edu.tw
The authors gratefully acknowledge the insightful comrnents of the two anonymous reviewers and the sponsorship ofthe National Science Council(NSC 98-24 I O-H-126-009)
164 Examining the Effects of Dual必νelTransformational Leadership, LMX,
and Group Cohesiveness on Employee Service Pelformance
Abstract: ln this study, we propose a multi-level theoretical 企amework within which we identi命 transformationalleadership at dual levels as the primary source of the positive influence on employee service performance through the mechanism of leader-member exchange (L卸1X) and group cohesiveness. Data are obtained from 23 branches of a large cornmercial bank in central Taiwan, with the samples collected from both managers and employees forming 228 manager-employee dyads, and thereby avoiding common method variance. Our results reveal that both levels of transformational leadership and LMX have significantly positive effects on employee service performance, with LMX also playing a mediating role between individual-focused transformational leadership and service performance. Group-focused transformational leadership a証ects employee service perfomlance through group cohesiveness. Interestingly, group cohesiveness is found to be an important moderator which also enhances the relationship between U.在X and service performance. Our study inc1udes a discussion ofthe theoretical and practical implications of our findings
Keywords: Dual-level Transformational leadership; Leader-member exchange;
Group cohesiveness; Hierarchical linear modeling
1.
Introduction
Excellent customer service has become one of the most important strategic aims for organizations to achieve competitive advantages (Hi缸， Ireland and Hoskisson, 2009); this is particularly the case for firms within the service industry,
of which banks are a typical 臥倒nple. Given the severity of modem day global economic challenges, improving service performance has become critical to survival within this particular industry. While Amazon is an exemplary on-line retailing company of providing superior customer service, Charles Schwab is an ace of customer service in the banking industry (McGregor, 2009). Since superior experience of customer service comes from direct interactions of employees, such as ba叫(tellers, with customers, we provide a theoretical framework to discuss the organizational and individual factors that might influence employee service
Chiao Da Management Review 均1.32 No. 1,2012 165
performance
Since excellent service performance is crucial for a firm to gain competitive advantages, we identi秒 transformational leadership as a critical factor in the motivation of superior service performance delivery amongst employees. There are
four distinct dimensions to transformational leadership, comprising of charismatic leadership, inspirational motivation, intellectual stimulation and individualized considerations (Bass
,
1985; Bass and Avolio,
1990). Leaders described astransformational will invariably concentrate their efforts on developing a vision and on persistence with regard to long-term goals, inspiring followers to pursue this vision and mobilizing the available resources to achieve such persistence.
There have been calls for a theoretical integration of the transformational leadership and LMX literatures (Gerstner and Day, 1997; Graen and Uhl-Bien,
1995). ln this study, we attempted such integration. Wang, Law, Hacke哎，Wangand Chen (2005) argued that the nature and qual i句 ofrelationships between leader and
follower are fundamental to linking leader behavior to follower response. Recent
studies have revealed that employees have a need not only for meaningful tasks at work, but also for meaningful relationships (Grant, 2008). Thus, the assumption has been that transformational leadership behaviors intluence follower service
performance through the quality ofthe leader-follower relationship. In line with this reasoning, we developed and tested a model in which L加IX mediates the relationship between transformational leadership behavior and employee service performance
A multi-Ievel approach has long been advocated by organizational scholars as the means of unveiling the richness and dynamics of social behavior across di仔erent organizational levels (Hi 哎， Beamish, Jackson and Mathi間， 2007). Several studies over recent years have advocated a combination of both individual-focused and group-focused transformational leadership research along with the cross-Ievel effects of contextual variabl的 (Graen and Uhl-Bien, 1995; Liao and Chuang, 2007). However, there appears to be a handful of studies in
which both individual-focused and group-focused transformational leadership perspectives on organizational outcomes are considered along with the cross-Ievel
166 Examining the Effects of Dual必νelTransformational Leadership, LMX,
and Group Cohesiveness on Employee Se川ce Pe"，舟，削
We set out in this study to respond to this gap in the literature by examining the extent to which the individual service performance is inf1uenced by group-focused transformational leadership through group cohesiveness, as well as the extent to which their service perfo口nanceis inf1uenced by individual-focused transformational leadership through LMX. We also examine the relationship between service performance and LMX, alongside the moderating role of group cohesiveness. We 由engo on to discuss dual levels of transformational leadership and LMX, and further discuss the moderating role of group cohesiveness.
2.
Theory Development and Hypotheses2.1.
Transformational Leadership and LMXTransformational leadership theory has been the most inf1uential leadership theory for decades (Judge and Piccolo, 2004). Abundant academic accumulations have been done. There are two major contrasting transformational leadership research streams amongst them, namely, leader-based and relationship-based approach (Graen and Uhl-Bien, 1995). Leader-based research examines a leader's behaviors have direct and significant effects on organizational performance outcomes related to followers, which is exemplified by most of the transformational leadership literature (Bass, 1985). Relationship-based research focuses on dyadic social ties between leader and follower that improve organizational outcomes. This approach is best exemplified by leadership-member exchange (LMX) theory (Graen and Uhl-Bien, 1995)
Past transformational leadership studies attempted to explain individual and organizational outcomes by identifying leader's specific behaviors (Cheng and Farh, 2001; Podsako缸~Mackenzie and Bommer, 1996). There have been calls for a theoretical integration of the transformational leadership and LMX theory (e.g.,
Gestner and Day, 1997; Graen and Uhl-Bien, 1995). To our best knowledge, only a handful studies have inc\uded both transformational leadership and LMX; however, only Wang et al.(2005) showed how transformational leadership and LMX are related to each other and to organizational citizenship behavior/task
Chiao Da Managemenl Review 均1.32 No. 1,2012 167
perfonnance. Wang et al.(2005) argued that a transforrnational lead巴r mobilize
social exchange relationshi阱， stimulating subordinates to surpass self-interests
More specifically, they provided solid results suggesting that the employee
organizational citizenship behavior/task perforrnance associated with
transforrnational leadership result from the LMX between subordinates and leader.
Wang et al.(2005) denoted that transforrnational leaders foster the forrnation
of high quality relationships with subordinates; while in a social exchange process, subordinates echo the leader by producing high-Ievel commitments to
organizational goals. Reporting structural equation analyses of data from multiple
organizations \ocated in a major city in northem China, Wang et al. (2005) wrote
that transforrnational leadership predicts LMX and further influences
organizational citizenship behavior/task performance. These results suggest that a
leader's charisma and individualized consideration both have dyad-Ievel influences which cause subordinates to reverberate in ways (such as extra role
behaviors) that further strengthen relational ties with the leader.
Prior studies advocate a mix of leader-based and relationship-based
transforrnational leadership research (Graen and Uhl-Bien, 1995; Podsakoff and MacKenz時， 1995). To advance this line of research, in the current study we aim
to largely extend Wang et al. 法(2005)research to bank service industry and shows
that leader- and relationship-based transforrnational leadership both exists. We
further corroborate the mediating role of LMX between individual-focused transforrnationalleadership and employee service perforrnance and identi秒 posltlve effects of individual-focused transfo口national leadership on employee service
perforrnance
2.2.
Individual- and Group-Focused Transformational LeadershipTransforrnational leadership theories have long been accused of ignoring the
group process (Nie\sen and Daniels, 2011), so some researchers have attempted to conc巴ptualize transforrnational leadership as a group-Ievel cons甘uct (e.g., Liao and Chuang, 2007; Wang and Howell, 2010; Wu, Tsui, and Kinicki, 2010). They argued that transforrnational leadership behaviors are directed at the who\e group
168 Examining the瓦fJects0/ Dua/-/eve/ 升。nsjOrmationa/ Leade丹hip，U.α; and Group Cohesiveness on Emp/oyee Service Performance
and, hence, result in a shared value and belief amongst followers. Thus,
group-level transforrnational leadership has been !inked to group variables (e.g.,
group identification and collective efficacy) and perforrnance variables (Bass,
Avolio, Jung and Berson, 2003; Schaubroe仗， Lam and C恤， 2007; Wu et al., 2010). However, two issues pertaining to this stream of research warrant further exammatlOn
First, past transforrnational leadership studies have typically focused on either the individual or the group level exclusively (Meng郎， Walter, Vogel and
Bruch, 2011). They do not investigate the e叮叮ts of transforrnational leadership on organizational outcomes at both levels concurrent旬，mainly because traditional multilevel methodology has not advanced to analyze the impact of transforrnational leadership at multiple levels (Lin, 2005). ln this current study, our multilevel methodology developed by Raudenbush, Bryk, Cheong and Congdon(2004) corroborates to exarnine the e質的tsof transforrnationalleadership at dual levels of analysis, which is widely adopted by recent studies (e.g., Liao and Chuang, 2007). To clarify this issue, we proposed a multiple-level transforrnational leadership model that divides transforrnational leadership into
the individual-focused level and the group-focused level. Liao and Chuang (2007)
have argued that individual-focused transforrnational leadership focus on
individual followers' needs and are expected to build strong ties between leader
and follower. The group-focused transforrnational leadership refers to the overall
pattem of leadership behavior, emphasizing the identity of the group and link the self-concept of followers to the shared values and be!iefs of the group. They also
refer to the group-focused transforrnational leadership may have a cross-Ievel, top-down effects on individual employee perforrnance. In this study, we inc1ude both levels of transforrnational leadership as predictors to employee service
perforrnance.
Second, transforrnational leadership studies needs to consider the joint effects of individual- and group-level factors on organizational outcomes, which is consistent with the contextual model (Liao and Chuang, 2007). Scott and Walker (1995) also called for further investigation on the moderation effects of contextual variables. Since groups are considered to be an effective solution for
ChiaoDa Manα'gemel1l Reνiew Vol. J2N，υ 1， 2012 169
organizations under severe competitions (Gilson and Shalley, 2004), this study echoes the research void in taking group cohesiveness as a moderation variable Group cohesiveness is an important contextual factor affects employee performance (George and Bettenhaus凹， 1990; Van Dyne, Curnmings and Parks,
1995) and warrant further examinations (Kidwell, Mossholder and Bennett, 1997)
It attracts and glues group members together to voluntarily participate group activities (Shaw, 1981)
To answer these calls, in this study, we integrate both individual- and group-focused transformational leadership and examine the extent to which the
transformational leadership created at the group level intluences the employee service performance. In addition, we will examine the moderating role of group cohesiveness, a group-Ievel phenomenon that members are attracted to the group to remain and actively participate in group activities, in the relationship between LMX and service performance (Geroge and Bettenhausen, 1990)
In the following section, we first test a hierarchical linear regression model that positions L卸1x as a mediator between individual-focused transformational leadership and employee service performance at individual level. At group level,
we propose to test group cohesiveness as mediator between group-focused
transformational leadership and employee service performance and as moderator
between LMX and individual service performance
2.3. Transformational Leadership and Employee Service
Performance
Previous transformational leadership studies have focused primarily on
leader b巴d昀ha盯叫V叫l叫or (仰引w仙a訂ang eωt al., 2005). Lea吋ders昀sl油啪h趾坤11叩1叩P t伽he昀leoωon凶s恥t心s pr戶rimaril抄y use s巴 Ifι:'c∞oncep叭t t由heOl句γa的s the means to expl旭ain the way戶sin wh划1ichleadership behavior
can ultimately transform the behavior of followers (Shamir, House and Arthur,
1993). Self-concept theory is the ways in which we see ourselves and our relationships with others are formed through our various interactions within the
working environment (Shamir et al., 1993). Sel仁concept theory explains the change of follower behavior in two ways. First of which is through social identification. Followers identify themselves as members of an honorable group,
170 Examining theξtfects 01 Duaιlevel Tranφrmational Leadershi泣，LMX, and Group Cohesiνeness on Employee Service Pe研ormance
whilst leaders will invariably use meaningful symbols, slogans, rites and rituals in order to strengthen the collective identity of the followers (Shamir, Zakay, Breinin and Popper, 1998)
The second way is transformational leadership behavior change the behavior of followers through the intemalization of co叩orate values. Since followers are a位racted by the vision, ideas and beliefs of a leader, they regard their work as an inseparable part of the work of the whole group (Deluga, 1994); they are proud of their membership within the group, and will normally generate high-level commitment towards it. Such high-Ievel commitment further establishes the followers' conception of ‘self', whilst also increasing self-e伍cacy (Shamir et al.,
1993). The two transformational processes refer to the ways in which the values and beliefs espoused by the leader can reinforce the conception of self amongst followers, with such followers internalizing the vision and values of their leader through their identification with the leader's organizational goals (Shamir et al., 1993)
When the need for provision of superior service performance is communicated as an important component of company's vision, and one which is advocated and communicated by the leader as an important organizational goal, the members of the organization are likely to p叮sueit with vigor (MacKenzie, Podsakoff and Rich,
2001). When transformational leaders intemalizethevision.values and beliefs into employees, they will be motivated to engage in high-level commi位ne肘， contribute
to their work, and strengthen their e宜orts to pursue overall service performance (Liao and Chuang, 2007). We therefore hypothesize that:
Hl: Individual-focused tranφrmationalleadership is positively related to
employee service p叫formance.
Group-focused transformational leadership, created by 仕ansformational leaders, is norms which come to be collectively shared, and which ultimately take on institutional status; this underpins the overall pa即m of transformational leadership as it exists in groups (Kitts and Chiang, 2008, Liao and Chuang, 2007). The norms that are formed amongst group members can be regarded as taken-for-granted organizational routines (Aldrich and Fiol, 1994) and behavior
Chiao Da Managemenl Reνiew Vol. 32 No. 1, 2012 171
scripts (Nooteboom, 1996). These norms can facilitate both task fulfillment (Scott,
2001) and conformity amongst group members (Tagger and Ellis, 2007)
There are two impo口ant elements for transformational leaders to construct organizational norms: cognitive legitimacy and sociaUpolitical approval (Aldrich and Fiol, 1994). First, cognitive legitimacy involves the adoption symbolic language and rhetoric, both of which have intention-framing purposes, aiming at transforming the beliefs of employees. Such techniques involve the use of words like
‘
us " as opposed to‘
me'; that is, transfonnationalleaders attempt to re-infuse new or radical ideas into socially accepted and approved conventional ideas to transfonn the values and be1iefs of their followers (Fiol, Harris and House, 1999).Second, sociaUpolitical approval refers to the techniques to construct ways of communications within an organization in order to eam leader's recognition. It often takes the form of storytelling (Barry and Elm白， 1997). Transformational leaders often use stories to set extemal criteria which are then, in tum, accepted and intemalized by group members (Aldrich and Fiol, 1994); that is, stories are used as metaphors and analogies to frame the real intentions of the transformational leader in order to familiarize the group members with formerly
unfamiliar extemal criteria (Lounsbury and Glynn, 2001)
Transformational leaders often use both cognitive legitimacy and sociaUpolitical approval to build shared organizational norms amongst group members; however, it is noted in prior studies that the creation of the overall pa出ms of 個nsformational leadership behavior requires a process of interaction between group members ( e.g., Bettenhausen and Murrigh閉， 1991; Liao and Chuang, 2007). Once these organizational norms become accepted by any single
group member, the processes of discussion and communication will determine
whether all other group members will ultirnately accept these same norms (Luria, 2008)
Organizational norms glue group members through sense-making process (Wei此， 1995). Sense司making process indicates that communication amongst
group members draws various confusing cues together within mu1tiple
environments and ultimately drive group members to proceed to a collective interpretation process. As a result, these diverse environments will gradual甘
172 Exalllining lhe E.lfecls of Dua/-/eνe/ 升。nsformalianα/Leadership, L品4α-Y.
and Group Cohe.臼S附間'SSl on Emp/oyee Service P訂戶川!Once become understandable to all group members as an organized system (Wei仗， 1995;
Wei此，Sutc1iffe and Obstfeld, 2005)
To summarize the discussion so far, we note that transformational leadership behavior gives rise to organizational norms shared amongst all group members; these norms (which are essentially institutions), can shape the behavior of group
members. However, powerful actors, such as transformational leaders, are needed to construct the shared norms (Schriesheim, 1980; Weick et 仗，2005). Following the
necessary process of behavior adaptation, through the recognition of the cognitive structures and the operation of the sense-making process, group members will finally come to accept these as taken-for-granted norms (Weick et al., 2005)
The above summary describes a process of transference 台om individual
transformational leadership behavior to shared collective norms amongst group members, norms which can be regarded as an ambient stimulus, which pervades the organizational atrnosphere and which is used to facilitate the process of task enforcement (Hackman, 1992; Liao and Chua嗯， 2007). Norms bond group
members together through their shared obligations within the working environment
(Liao and Chuang, 2007); that 時，these norms affect service performance through a
bi-directional process. Thus, we hypothesize
H2: Group-focused transformational leadership is positively related 10
employee service peυormance.
2.4.
Tbe Mediating Role of Group CobesivenessThe mediating role of group cohesiveness in the relationship between
group-focused transformational leadership and employee service performance is premised on the notion that high-Ievel group cohesiveness ref1ects an affective
bonding accompanied by mutual dependence amongst group members (Mudrack,
1989). Such a relationship develops from a predominantly transactional exchange into a social exchange as mutual trust and respect, which is developed by transformational leader and intemalized by employees as shared organizational norms and lead to higher identification with the uniqueness of leader value (Wech
Chiao Da Management Reνlew均1.32No. 1, 2012 173
organizational goals (Dutton, Dukerich and Harquail, 1994)
We argue that group-focused transfonnational leadership builds and nourishes high-quality group cohesiveness. Shared organizational norms developed by leader can align or re-align beljefs of group members, and reinforcing their beliefs, self-e佑 cacy，and task motivation (Conger and Kanungo, 1987; Shamir et al., 1993). Dvir and coauthors (2002) suggested that social bonding amongst group members mediate the effects of transformational leadership behaviors on follower performance. Mutual dependence among group members suggests high levels of interpersonal attractÎons and attachments (Wech
et al., 1998) and in tum leads to higher service perfonnance. We therefore hypothesize:
H3: Group cohesiveness mediates the relationship between group-focused transformationalleadership and employee service peφrmance. 2.5. The Mediating Role of LMX
As discussed earlier，因nsfonnational leaders tend to invoke the social and personal identification processes in their employees. We further reason that a c10se leader-member exchange relationsrup will emerge from these processes in wruch trust and respect are likely developed. 甘le theory of LMX (Graen and Uhl-Bien, 1995) highlights the importance of leader-member relationsrup in motivating employee perfonnance. Relationship buildillg involves mutual leaming and accommodation Transfonnational leaders tend to es帥lisha high-quali句socialexchange relationship with their subordinates (Wang et al., 2005) because their vision, inspiration, motivation, and inillvidualized consideration are likely to induce subordillates' endorsement of their values and goals, wruch are more of social rather than economic in nature. Consequently, a relational contract may be fonned between the leader and the subordinate (Rousseau, 1995). This relational con仕act escalates through a reciprocating process such that the better the social exchange relationsrup, the better the performance the subordinates will exllibit (Liden, Sparrowe and Wayne, 1997)
When personal identification with the leader is intemalized by followers, mutual tru仗，respect, and loyalty are eamed between the two parties. In the context
174 Examining the Effects of Dua/-/eve/7切問formationa/Leadership. L品α; and Group Cohesiveness on Emp/oyee Service Performance of a service organization where the value of superior customer service is s甘ongly promoted subordinates are likely to exhibit superior perforrnance to ensme the mam紀nanceof their high quality relationship with the leader, and to reciprocate the trust the leader has in them (Wang et al., 2005) by transferring it to establishing a meaningful relationship with customers. Stated altematively, achieving superior service perforrnance can be considered as a forrn of LMX currency circulated within
the social exchange relationships that exist between leaders and followers, as well as a way of showing reciprocity and obligations 仰Tang et al., 2005). Therefore, we
hypothesize:
H4: Leader-member exchange mediates the relationship between
individuaιjocused transformational leadership and employee
service peφrmance.
2.6.
Cross-Level Moderation Effects of Group CohesivenessCohesiveness is generally defined as the resultant of all forces acting on all the members to remain in the group (Cartwright, 1968). Group cohesiveness is
one of the essential concepts for understanding group dynamics (Zander, 1979) Theorists identify group cohesiveness as group spirit, interpersonal attraction, sense of belongingness (Mudrack, 1989), and the desire to stay in a group (Evans and Dion, 1992). 1n the current study, group cohesiveness is concentrated on
social cohesion, which would be a proper concept for examining the moderating effect in the person-context 企ameworkstudy (Shin and Park, 2009)
Group cohesiveness exercises moderating effects on employee service
perforrnance in two ways. First, social control theory (Hirschi and Stark, 1969; Shin and Park, 2009) focus on restraining or controlling factors that are broken
inside personalities. The theory demonstrates that individual behavior can be
restrained and focused to a certain degree if they belong to groups which have strong ties. 1n business situations, personal characteristics could be restrained and concentrated when group member belong to cohesive groups. Employees in
cohesive group would be influenced by other members due to s甘ong social ties; consequently, group cohesiveness would strengthen the LMX-service
Chiao Da Management Review 的/.32 No. 1,2012 175
performance relationship
Second, social capital theory. The term social capital is described as an investrnent in social relations with expected retums in the marketplace (Lin, 2001). Putnam (1993) also suggested that social capital would facilitate cooperation and increase mutual dependence in groups. It was assumed that members in highly-cohesive groups are willing to share their resource and cooperate with others due to mutual trust, respect, and obligations (Graen and Uhl-Bien, 1995); thus, group cohesiveness helps their members to achieve better performance Group performance will increase in such situations; hence, group cohesiveness would be a moderator in the LMX-service performance relationship
Higher levels of group cohesiveness strengthen interactions, comrnunication and mutual dependence among group members, whilst also enhancing 也E
willingness amongst employees to provide superior customer services. Group cohesiveness is thus regarded as a moderator between LMX and employee service performance. We therefore hypothesize that:
H5: Group cohesiveness acts as a moderator between LA在~and employee service performance; the greater the level of group cohesiveness
,
the stronger the pos;tive relationship between LMX and employee service performance.3.
Methods3.1.
Participants and ProceduresThe research setting in this study was 23 branches of Taiwan Cooperative
Ba叫< (TCB) in central Taiwan area. The bank, former Taiwan lndustrial Bank in
Japanese colonial period, was established in 1945 by Taiwan provincial govemrnent's integrating credit unions, farmers' associations, and fisherman 's associations. By the end of 2010, the capitalization stood at 1.4 billion USD. ln recent years however, with the increasing competition from diverse financial institutions, the bank has undergone significant changes and has shifted their financial services 企om co叩orate ba叫<Ï ng to personal banking and wealth
176 Examining the Effects of Dua/-/eνe/ Tran吶rmationalLeadership, LMX. and Group Cohesiveness on Employee Service Performance
management services. As a result, the bank has put paramount emphasis on customer service quality. For example, it has made great efforts in training its personal frnancial consultants internally and in strengthening customer relationship management externally. Each branch was regarded as a separate group because performance evaluation was branch-based in the bank. The research setting matches strong sample relevance (Sackett and Larson, 1990)
To avoid the common method bias (Podsakoff, Mackenzie, Lee and
Podsako缸 2003) ， we followed Scott and Bruce (1994) and divided the questionnaire into two parts: a manager questionnaire in which employee and group service performance were evaluated, and an employee questionnaire, in which their perception of transformational leadership and LMX were measured Data collection procedure was as follows.
Firstly, we contacted branch managers through telephone calls and e-mails, and invited them to distribute questionnaires. We dispatched questionnaires on a one-branch-one-package basis. Each package contained one copy of the manager questionnaire and 12 copies of the employee questionnaire. The manager was asked to evaluate up to 12 subordinates in the branch. One-manager-each-branch basis is used, 23 managers responded, and none of them out of the same branch. The criteria of selecting respondents were bank clerks who have direct contacts with customers. Judging 企om the scale of each bank, most first-line employees were chosen. A number code was used for each subordinate so that we could match the manager's evaluation data with that of the subordinate. The employees were not made aware that their managers were evaluating them. To ensure anonymity, no names were required in any part of the questionnaire and they were informed that all responses would be kept confidential. Finally, two phone reminders were sent: one after three weeks and the other after four weeks. To encourage participation, every participant was sent a little souvenir as a gesture of appreclatlon
With the bank's senior manager support, we obtained a high response rate: 85%. Our final usable sample comprised of 228 respondents from 23 branches of the commercial bank; the demographic characteristics of our study sample were as follows. The average group size in this study was ten persons (S.D. = 2.20),
Chiao Da M加αgemel1lReνiew Vol. 32 No. 1. 2012 177
which was wel1 above the minimum criterion of three (Carron and Spink, 1993)
Most of the study participants were female (65.4 percent), with a mean age of
40.64 (S.D.
=
8.12) and mean organization tenure of75.75 months (S.D.=
52.35)3.2.
Measures3.2.1. Service Performance
S巴rvice performance is assessed in this study using the Liao and Chuang
(2004) employee service performance scale, with rninor modifications being made
in order to accommodate our measures. Each of the managers was asked to rate
the service performance of twelve employees based upon a seven-item scale Examples of the statements included are
“
being friendly and helpful tocustomers" and "asking good questions and Iistening to find out what a customer
wants", with the response options ranging from 1 (strongly disagree) to 6 (strongly agree). The Cronbach'sα(reliability) for this scale was .92
3.2.2. lndividual-Focused Transformational .Leadership
This study adopts tbe Bass and Avolio (1990)
‘
multi-factor leadersh 巾，questionnaire (MLQ-form 5R) 的 the measure of transformational leadersh 巾，
using the Chinese version with minor revisions made for considerations of
relevance. The measure of transformational leadership comprises of four dimensions, ‘idealized influence', ‘inspirational motivation', ‘intellectual stimulation' and
‘
individualized considerations', with a six-item scale being used for each dimension. The transformational leadership scale in thjs study therefore comprises of 24 items, and includes statements such as“
talks to us about rus/her most lmpo口antvalues and beliefs" and “spends time teaching and coaching me",with the response options ranging from 1 (strongly disagree) to 6 (strongly agre巳)
The Cronbach'sαfor this scale was .97
3.3.3. Group Transformational Leadership
The measures of group transfonnational leadersh巾，which invo1ve the sbared
178 Examining the E.ffects o[ Duaιlevel Trans[ormational Leadershψ. u.α，
and Group Cohesiveness on Employee Service Per[ormance
1998), were aggregated from the 仕ansformational leadership measures. Prior to aggregation, the consensus indices, such as r"習(James，Demaree, and Wolf, 1984),
ICC(I) and ICC(2), were employed to justi命 the aggregation process The r"宮 score was .86, indicating that it was reasonable to apply a cross-leve1 analysis; the ICC(I) score was .40
,
which is higher than the benchmark suggested by Cohen (1988); and the ICC(2) score reached .94, which is also higher than the benchmark suggested by Klein and Kozlowski (2000).3.3.4. Leader-member Exchange
LMX is assessed in this study using the Graen and Uhl-Bien (1995) LMX scale; this is a seven-item scale (LMX-7), which inc1udes questions such as “how well does your leader understand your job problems and needs?" and “how well does your leader recognize your potential?"; the response options ranged from 1 (strongly disagree) to 6 (strongly agree), and the Cronbach'sαfor this scale was .93
3.3.5. Group Cohesiveness
Group cohesiveness is a referent-shift consensus construct (Chan, 1998;
Kozlowski and Klein, 2000). This is measured in the present study using the Dobbins and Zaccaro (1986) eight-item scale, which inc1udes statements such as
“
the members of my group get along well together" and“
there is little dissention in the group弋 withthe response options ranging from 1 (strongly disagree) to 6 (strongly agree). The Cronbach'sαfor this scale was .92.The lower-level (individual) evaluations obtained in this study were based upon prior consensus to either form, or shift to, a new construct, which was distinct from a construct original1y derived at individualleve1 (Ch血，1998). However, prior to such cross-leve1 analysis, there is a need to check the presence of group-level effects (Bliese, 2000). The r"官scorefor group cohesiveness in the present study was 0.71,
whilst the ICC(I) score was .29 and the ICC(2) score was .80. Since all of these values are consistent with the acceptable range of values suggested wi也inthe extant literature, these are applied as the measures of group cohesiveness in the present sωdy.
Chiao Da Managemenl Reνlew 均1.32 No. 1, 2012 179
3.3.6. Control Variables
Controls are provided in this study for age, gender (0 = female; 1 = male),
tenure within the organization (months) and tenure with the current supervisor
(months). The last two controls, organizational tenure and tenure with current
supervisor, were ca1culated in order to avoid any potential confounding e缸ectson
the dependent variables (Kamdar and Van Dyne, 2007)
4.
ResultsThe means, standard deviations and variable correlations for the study
sample are presented in Table 1, from which we can see that most of the
correlations are within .00 to .55; the correlations between the variables are
therefore acceptable
Table 1
Means
,
Standard Deviations,
and CorrelationsVariables Mean SD 2 4
Individual-Ievel variables
I.Age 40.64 8.12
2. Organization tenure 75.75 52.35 77**
3.Tenure with c山Tentsupervlsor 50.95 47.57 46** 55**
4.LMX 4.71 80 -.04 -.08 。7
5.Employee Service performance 4.50 72 一02 。2 -.01 36**
6.1ndivid阻l-focused
4.38 83 -.22串串 -.21** 一10 29** 30**
transformationalleadership Group-Ievel variables
I.Group cohesiveness 4.64 43
2Gleraodueprs-hfom cused transihnnatIonal 4.37 57 90*' E
抖抖。1， * 抖 的 (two-tailed)
4.
1. Confirmatory Factor AnalysesWe conducted a set of confirmatory factor analyses (CFAs) to determine if our multi-item variables (i.e., Transformational leadership, LMX, and Group
180 Examining the EjJects of Duaιle叫 Transformational Leadersh帆 LMX. and Grolψ Cohesiνeness on Employee Service Pelformance proposed three-factor model displayed an acceptable fit (計 = 1282.肘，df= 699, p < .001, R恥1SEA = .08, CFI = .94, NNFl = .93). Tbe fit statistics for the hypothesized model were significantly better than a two-factor model (grouping
TFL and LMX) (計= 1676.42, df= 701, p < .001, RMSEA = .11, CFI = .侶， NNFI = .92, 6X2 = 394.37, df = 2, p < .001) or a one-factor model (χ 2014.62, df= 702,p < .001, RMSEA= .13, CFI = .92, NNFI = .91, 6X2 = 338.2,
df= l,p < .001)
4.2.
Hierarchicallinear modeling analysis‘
Hierarchical linear modeling' (HLM) analysis (Raudenbush, Bryk, Cheongand Congdon, 2004) is adopted in this study to test our hypothesized relationshi阱， with controls also being included for the employees' age
,
gender,
organization tenure and tenure with the current supe鬥 isor. Model 1 of Table 2 tests for the effects of both individual- and group-focused transformational leadership, with the resu1ts revealing the significant predictive ability of group transformational leadership with regard to service performance (1' = .34, p< .5); although less so,the effect of individual-focused transformational leadership is also found to be significant (1' = .13,p < .01). Support is therefore provided for both H1 and H2
The test resu1ts for the mediation effect of group cohesiveness between
group-focused transformational leadership and service performance are presented in Modell and 2 ofTable 2. This test follows the three-stage process proposed by Baron and Kenny (1986). In the first stage, group-focused transformational leadership must be shown to have a significantly positive correlation with service
performance; this result is already assured by the test of Hypothesis 1. In the
second stage, group-focused transformational leadership must be found to have a
significant1y positive association with group cohesiveness; this relationship is tested in the present study using ordinary least square (OLS) analysis, essentially
because this is seen as an appropriate way of assessing this effect at group level (Liao and Chua嗯， 2007). Our resu1ts reveal tbat group transformational leadership is indeed found to have such a positive association with group
Chiao Da Managemenl Review Vol. 32 No. 1, 2012 181
Table 2
Hierarchical Linear Modeling Results
Variables Employee Service Per臼rmance LMX
Modell Model2 Model3 Model4 (ModeI5)
Intcrcept γ00 4.55*** 4.55*** 4.51 *** 4.50*** 4.74***
I ndividual-Ievel variables
Age γ10 。。 。。 一 00 一00 。 l
Gender Î'2日 11* 一 11* -.09 I1 一08
Organization tenllre γ3日 。。 。。 。。 。。 -.00'
Tenllre with current supe鬥IS0r γ40 ' -.00* 一00* 。0** 。0** 。0*
LMX γ50 27** 28**
lndividual-focused transforrnational γ60 13** 14** IOt IOt 13*
Icadership
Group-Ievel variables
Group cohesivencss γ01 61 ** 39* 1.71 65 牢牢牢
Group-focused transforrnational γ02 34* -.05 06 1.13 一10 leadership
Group cohesivcness xGroup-focuscd γ0:1 5.33
transforrnational Icadership Cross-Ievel interaction variables
Group cohesiveness x LMX γ51 2 日*
δ2 36 35 32 33 74
700 。9** 。9 牢牢牢 。5 。5 09
T55 。。 。。
76fi 。7* 05* 12
Model Deviance 476.00 475.25 462.22 454.57 554.04
αn = 228 al individuallevel; n=23 al group level * * * p < .001, ** P < .01, * 戶<.05, t p < .1
bin all models, level 1 variables were group-mcan cenlered, excepl Gender "Oeviance is lhe叮leasurementof model fit. The smaller, the better lhe model fits
Given the high correlation found between group transforrnational leadership and group cohesiveness, as shown in Table 1
, we considered
it necessary in thisstudy to caπY out an additional test for potential multicollinearity (Hair, Anderson, Tatham and Black, 2006); however, the results of this test show that the variance inf1ation factor has a low degree of multicollinearity (VIF = 1)
Both group-focused transformational leadership and group cohesiveness are
included within the third stage of the mediation effect test. The results, as shown in Model 2 of Table 2, reveal that group cohesiveness has significant predictive
182 Examining the Effects 0/ Duaι /eve/7切，在(ormationa/Leadership, LMX,
and Group Cohesiveness on Emp/oyee Service Performance ability with regard to service performance (， 剃 ，p < .01), thereby providing
support for H3
A similar approach is followed to test the mediation e能ctof LMX between individual-focused transformational leadership and service performance. We also follow the three-stage process proposed by Baron and Kenny (1986). ln the first
stage, presented in Model 1 of Table 2, individual-focused transformational leadership is found to have significant predictive ability with regard to service
performance (， 矢口 ，p < .01), whilst in the second stage, presented in Model 5 of Table 2, individual-focused transformational leadership is also found to be positively associated with LMX (, = .13,p < .05)
ln the third stage, presented in Model 3 of Table 2, both individual-focused transformational leadership and LMX are included, with the results revealing an overall reduction in the effect of individual-focused 甘ansformationalleadership on
service performance (, = .10, p < .1); since this provides p訓ial support for the mediating role of Ll\在Xbetween individual-focused transformationalleadership and
service performance, partial support is also provided for H4
Further tests are undertaken in this study of the cross-Ievel moderation effects, with the results of the tests between LMX and service performance being presented in Model 4 of Table 2. ln order to avoid any potentially spurious
cross-Ievel moderation effects, we follow Hofmann and Gavin (1998) and Liao and Chuang (2007) to control for the interactions between groups (by inc1uding the product terms of botb group cohesiveness and group transformational leadership). The results reveal that the interaction effect is significant ( γ=.28，p < .05).
We then follow the suggestion of Aiken and West (1991) to graphically observe the interaction effect; this is illustrated in Figure 1, wbich shows that when group cohesiveness is low, the influence ofLMX on service performance is
weak and slope is flat (ß = .14,p < .10). Conversely, wben group cohesiveness
is high, the influence of LMX on service performance is stronger and slope
becomes steeper (ß = .26, P < .001); that is, with high group cohesiveness, the effect of LMX is enhanced and there is a corresponding increase in service
Chiao Da Managemenl Review 均1.32 No. 1, 2012 183
Figure 1
The Interaction Effect of L恥fXand Group Cohesiveness on Employee
Service zuphJ AU 司 司 J3 、 4 司 I ‘ BEamE 』。℃ ω 已 ωu-Eω 吶 ωω 〉。一且 E 凶
t三士
/之三...-- . . . /之三...--low Group CohesivenessCohesiveness 一﹒-HighGroup 。 Low High LMX
5. Discussion
Three notable fmdings arise from the present s仙dy， each contributing to the
extant literature on employee service performance and transformational leadership
Firstly, both individual-and group-focused transformationalleadership are found to
have significantly positive effects on employee service performance. Secondly,
group-focused transformational leadership is found to exist in the form of shared organizational norms which directly affect employee service performance through group cohesiveness; individual也cusedtransformational leadership is found to be
甘anslated into employee service performance through the LMX relationship
Thirdly, the mediating role of group cohesiveness is found to 甜engthen the
relationship between LMX and employee service performance
Wbilst significant evidence presented in the prior studies show that
transformational leadership inspires organizational citizenship behavior amongst
employees, as well as task performance (Podsakoff et al., 1996; Wang et al., 2005)
The present study extends Wang et al., (2005) to service indus的， settings and
184 Examining Ihe EJJec/s of Dua/-/eνe/ Transformaliona/ LeadershiJ少，LA疵， and Group Cohesiveness on Emp/oyee Service Performance
This study corroborates the mediating effects of LMX between individual-focused transforrnational leadership-employee service perforrnance relationship and
identifies positive effects of individual-focused transforrnational leadership on employee service perforrnance.
As noted earlier, within the extant literature of 甘ansforrnational leadership,
various studies call for exarnination of the multi-level effects of the 甘ansforrnational
leadership behavior on organizational outcomes (Liao and Chung
,
2007; Wang and Howell, 2010). Our study excels prior studies by showing that group-focusedtransforrnational leadership and group cohesiveness are not only meaningful
constructs showing positive effects on employee service perforrnance, group cohesiveness also plays a moderation role in the relationship between LMX and superior employee service perforrnance.
By demonstrating the influence of group-focused transforrnationalleadership
on the individual service perforrnance, this current study provides a first step
towards the development of a more comprehensive model capable of effectively
identifying the underlying mechanisms of such influences. For example, our findings suggest that group-focused transforrnational leadership can promote
superior employee service perforrnance by creating a cohesive atmosphere within
the group as a whole; this could conceivably facilitate a more cooperative desire amongst the group members to engage in the greater sharing of information and
knowledge.
Our fmding of the LMX effects in 甘anslating into service perforrnance indicates a reciprocal process when employees are inspired by leader's vision,
motivation and individualized considerations. Furtherrnore, our findings that group cohesiveness plays a moderating role in the LMX-employee service
perforrnance relationship which thereby suggests that it has the effect of
strengthening this relationship adds yet another valuable piece to the theoretical puzzle of the way in which individual-focused transforrnational leadership influences employee service perforrnance whilst also providing suggestions with
regard to the circumstances under which the positive effects of LMX might be
augmented. The cross-Ievel effect of group cohesiveness adds complexity to the phenomenon of such cohesiveness at the individual level, thereby providing a
Chiao Da Managemenl Review 的 1.32 Nο 人 2012 185
more comprehensive illustration and understanding ofthe dynamics relating to the
ways in which individual- and group-level factors interact to influence the
emergence of important employee outcomes
Certain methodical merits of the present study should be noted. Firstly, in
response to the call for more multi-leve\ research, we adopt hierarchical linear regression analyses to examine both group cohesiveness and the cross-level effects of
transfonnationalleadership
,
at both group and individuallevels. Secondly,
in order toavoid the potential effects of comrnon method variance, as recommended by
Podsakoff et al. (2003), the data used in this study are obtained from many different
sources
5.1. Managerial Implications
The findings of this study have potentially valuable implications for general
management practice. Firstly, in addition to showing respect for subordinates and inspiring them to achieve their full potential, effective leaders must also be capable
of forming good-quality social exchange relationships with them. If they are to
succeed in building up such relationships, transformational leaders will need to be
able to sketch out the 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 tasks and relationshi阱， followers can obtain their
perceived equity within the organization (Dienesch and Liden, 1986), and can
then go on t。如此heridentify with the vision and values of their transformational leader. Thus, mutual exchange relationships are established, characterized by trust, loyalty and comrnitment, a fonn of social cuπency circulated in these social
exchanges, within which subordinates feel some obligation to reciprocate through
enhanced performance (Wang et al., 2005)
Secondly, since leaders cannot stay in their position forever, they have to
create shared organizational norms that will provide consistent and effective
cohesion within the organization. When building such organizational norms and
beliefs, leaders can use techniques involving symbolic language and rhetoric to transform the beliefs of subordinates, such as refe巾ng to 酒， instead of 'me.'
186 Examining the Effects of Dua/-/eve/ Transformational Leadership, LAα: and Group Cohesiveness on Employee Service Peφrmance
Leaders also use other techniques, such as story telIing, to avoid social political
approval and to in甘oduce extemal criteria into the processes and procedures that
already exist within the organization; these can also help group members to accept
and intemalize these norms and beIiefs
Thirdly, group cohesiveness is an atmosphere which is shared by alI group
members, within which there is increased knowledge sharing and cooperation
amongst the members
,
which ultimately strengthens the LMX-employee serviceperfonnance relationship. A high level of cohesiveness enhances the willingness of
employees to leverage resources in order to develop a s仕ong comrnitment to their
job. Many different activities can be used to improve this cohesion, such as annual
business group travel programs, cross-cultural 仕aining experiences for employees
or corporate adventure team-building programs (Tsai and Chi, 2008).
We contribute to the literature on transformational leadership and employee service performance by examining the mediating role of group cohesiveness and
the leader-member exchange relationship using cross-sectional, multi-sourced,
multi-Ievel data. Our study extends both the theoretical and empirical literature on
leadership by demonstrating the existence and effects of group transformational
leadership constructs, and by proposing and testing a model which supports the
integration of 甘ansformationalleadershipIiterature and social exchange theory.
5.2.
Limitations and Suggestions for Future ResearchWe concIude with a few Iirnitations and suggestions for future research. Firstly, to control for undesirable environmental facto眩， such as industry or organizational
structu時，our sample was obtained from 23 branches of a large comrnercial bank.
The choice of research settings is essential for exarnining theoretical 企amework.
Since Taiwanese banks are facing the pressure for increasing corporate performance
after the enforcement of govemment's fmancial liberaliz泌的n poli句， they are
engaged in service quality irnprovement actions. As a result, this sample bank is most suited for our study and is also a Iirnitation. The constructs such as
transformational leadership, LMX, and group cohesiveness are now important
characteristics in banking service indus虹y. However, there may be other factors,
Chiao Da Management Review 的 l. 32 No. 1,2012 187
service perfonnance, but which are not inc1uded in the present study; therefore,
future studies should attempt to explore these
Secondly, no infonnation was collected in the present study on psychological
process variables; therefore, future studies should aim to examine these in an
attempt to further explain the ways in which transfonnational leadership inspires
employees to provide superior service perfonnance. Thirdly, our evaluation of
service perfonnance reflects a managerial perspectiv巴，whereas it is suggested that such perfonnance evaluations should also be obtained directly from customers (Liao and Chuang
,
2007).The study by Liao and Chuang (2007), which focused on a franchised hair salon in Taiwan, involved a sample of customers who were regular visitors to the salon; thus, this made them very accessible. This study used a managerial perspective essentially because bank customers do not attend on a regular basis; this
makes access to them very costly. Nevertheless, future studies should make some
a仕empt to collect data on service perfo口nance 企om the perspective of bank
customers
Finally, although the data collected for this study comprised of infonnation
obtained from multiple sources, it was still undertaken using a cross-sectional
design; this c1early limits our ability to identif扯 any causal influences. Future
studies may elect to adopt a longitudinal design in order to examine the ways in which the relationship dynamics between 前ansfonnational leadership,
leader-member exchange and group cohesiveness change over time, and to 虹Y to
trace their causallinks.
In conc1usion, the present study contributes to the Iiterature on
transfonnational leadership and employee service performance by examining the
mediating role of leader-member exchange relationship using multisource,
multi-Ievel data in a cross-sectional design. ln addition, the study extends the
leadership literature both theoretically and empirically by demons甘ating the
existence of the group-focused transfonnational leadership and by proposing and
testing a model that supports the cross-level mediation/moderation effects of group
l88 Examining the EjJects of Duaιleνel Transformational Leadership, LAα; and Group Cohesiνeness on Employee Service Perjórmance
6. References
Aiken, L. S. and West, S. G. (1991), Multiple Regression: Testing and Interpreting Interactions, Newbury Park, CA: Sage.
Aldrich, H. and Fiol, C. M. (1994),“Fools Rush ln? The Institutional Context of Industry Creation," Academy of Management Review, 19(4),645-670 Baron, R. M. and Kenny, D. A. (1986), “The Moderator-Mediator Variable
Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Consideration," Journal of Personality and Social Psychology, 51 (6), 1173-1182.
Ba釘y， D. and Elmes, M. (1997),“Strategy Retold: Toward A Narrative View of Strategic Discourse," Academy of Management Review, 22(2), 429-452.
Bass, B. M. (1985), Leadership and Performance Beyond Expectations, New York, NY: Free Press.
Bass, B. M. and Avolio, B. 1. (1990), The Multifactor Leadershψ Questionnaire，
Palo Alto, CA: Consulting Psychologists Press
Bass, B. M., Avolio, B. 1., Jung, D. I. and Berson, Y. (2003),“Predicting Unit Performance by Assessing Transformational and Transactional Leadership,"
Journal of Applied Psychology, 88(2), 207-217
Bettenhausen, K. L. and Murrighan, J. K. (1991),“The Deve10pment of An
Intragroup Norm and the Effects of lnterpersona1 and Structura1 Challenges," Administrative Science Quarter鈔， 36(1)， 20-35
Bliese, P. D.(2000), “Within-Group Agreement, Non-lndependence, and
Reliability: Implications for Data Aggregation and Analysis," In K. 1. K.lein and S. W. J. Kozlowski (Eds.), Multilevel Theo吵" Research, and Methods in Organizations. San Francisco, CA: Jossey-Bass, 349-381.
Carron, A.V and Spink, K. S. (1993),“Team Building in an Exercise Setting,"
The Sport Psychologist, 7(1), 8-18.
Cartwright, D. (1968),“The Nature of Group Cohesiveness," In D. Cartwright and A. Zander (Eds.)
,
Group Dynamics: Research and Theory. 3rd ed.,
New York, NY: Harper & Row, 91-109.Chiao Da Managemenl Review Vo/. 32 No. 1,2012 189
Domain at Different Levels of Analysis: a Typology of Composition Models," Journal of Applied Psychology， 的 (2)，234-246.
Cheng, B. S. and Farh, J. L. (2001),“Social Orientation in Chinese Societies: A
Comparison of Employees from Taiwan and Chinese Mainland." Chinese
Journal of Psychology，的 (2)， 207-221.(ln Chinese)
Cohen, J. (1988), Statistica/ Power Analys的 for the Behavioral Sciences, 2nd e仗，
Hillsdale
,
NJ: Lawrence EribaumConger, J. A. and Kanungo, R. A. (1987), “Towards a Behavioral Theory of
Charismatic Leadership in Organizational Settings," Academy of
Management Review, 12(4),637-647.
Deluga, R. J. (1994),“Supervisor Trust Building, Leader-Member Exchange and Organizational Citizenship Behavior," Journal of Occupational and
Organizational Psych%gy，的 (4)， 315-326
Dienesch, R. M. and Liden, R. C. (1986),
Leadership: a Critique and Fu的er Development," Acαdemy of Management Review, 11(3),618-634.
Dobbins, G. H. and Zaccaro, S. 1. (1986),“The E仔ects of Group Cohesiveness and Leader Behavior on Subordinate Satisfaction," Group and Organization
Studies, II (3),203-219.
Dutton, J. E., Dukerich, J. M. and Harquail, C. V. (1994),“Organizational Images and Member Identification," Administrative Science Quarterly, 39(2), 239悔263
Dv汀， T., D. Eden, Avolio, B. J. and Shamir, B. (2002), “Impact of
Transformational Leadership on Follower Development and Performance A Field Experiment," Academy of Management Journal, 45(4), 735-744
Evans, C. R. and Dion, K. L. (1992),“Group Cohesion and Performance: A Meta
Analysis," Small Group Research, 22(7), 175-186.
Fiol, C. M., Haηis ， D. and House, R. (1999),
for Ef宜fe郎cω叫tingSocial Cαha叩nge，"Leadership Quart.仿er~砂y，10(3),449-482 Geroge, J. M. and Bettenhausen, K. (1990),“Understanding Prosocial Behavior,
Sales Performance, and Turnover: A Group-Level Analysis in A Service
190 Examining the Effects of Dual必νelTransformational Leadership. LMX.
and Group Cohesiν'eness on Employee Service Pe柚rmance
Gestner, C. R. and Day, D. V. (1997),“Meta-Analysis Review of Leader司Member
Exchange Theory: Correlation and Construct Issues," Journal of Applied Psychology, 82(6), 827-844.
Gilson, L. L. and Shalley, C. E. (2004),“A Little Creativity Goes A Long Way: An
Examination of Teams' Engagement in Creative Processes," Journal of Management, 30(4), 453-470.
Graen, G. B. and Uhl-Bien, M. (1995), “Relationship-Based Approach to Leadership: Development of Leader-Member Exchange (LMX) Theory of Leadership Over 25 Years: Applying A Multi-Level Multi-Domain Perspective," Leadersh加 Quarter鈔， 6(2) ， 219-247.
Grant, A. (2008), “Does Intrinsic Motivation Fuel the Prosocial Fire?
Motivational Synergy in Predicting Persistence, Perforrnance, and Productivity," Journal of Applied Psychology, 93(1), 48-58.
Hackman,1. R. (1992),“Group Influences on Individuals in Organizations," In M. D. Dunnette and L. M. Hough (Eds.)," Handbook of lndustrial Organizational Psychology, Vol. 3, Palo Alto, CA: Consulting
Psychologists Press, 1455-1525
Hair J. F., Anderson R., Tatham R. L. and Black W. C. (2006), Multivariate Data
Analysis, Upper Saddle River, NJ: Prentice Hall.
Hirschi, T. and Stark, R. (1969),“Hellfire and Delinquency," Social Problems
17(2),202-213.
Hi哎， M., Beamish, P., Jackson, S. and Mathi凹， J. (2007),“Building Theoretical
and Empirical Bridges Across Levels: Multilevel Research in
Management," Academy of Management Journal, 50(6), 1385-1399
Hi前， M. A., Ireland, R. D. and Hoskisson, R. E. (2009), The Management of Strategy: Concepts and Cases, 8th ed. New York, NY: Thomson Leaming. Hofi:nann, D. A. and Gavin, M.B. (1998),“Centering Decisions in Hierarchical
Linear Models: lmplications for Research in Organizations," Journal of Managemef呵，24(5)，623-64 1.
James, L. R., Demaree, R. G. and Wolf, G. (1984),“Estimating Within-Group Interater Reliability with and Without Response Bias," Journal of Applied Psychology, 69(1), 85-98.
Chiao Da Management Review均 l.32 No. 1,2012 191
Judge, T. A. and Piccolo, R. F. (2004), “Transfonnational and Transactional Leadership: A Meta-Analytic Test of Their Relative Validity," Journal of Applied Psychology，的(5)，755-768
Kamdar, D. and Van Dyne, L. (2007),‘'The Joint Effects of Personality and Workplace Social Exchange Relationships in Predicting Task Perfonnance and Citizenship Perfonnance," Journal of Applied P5ychology, 92(5),
1286-1298
Kidwell, R. E., Mossholder, K. W. and Bennett, N. (199們7η)，
Organizational Citizenship Be咄ha盯Vlωor仁: A Mu叫11t“ile盯V巴el Analysis Using Work
G 臼
ωrou恥p戶sa削n吋d Indi吋1吋凶di圳V圳idωu閻1泌al叮s丸U，Jf川"γJJ，捌Oω仰ur叫叩叫ntω叫alof且蜘必n呵1叫呵G句!ge仰mη1e削叫n1t，23(6), 775物
Kitts, J. A. and Chiang, Y. S.(2008), Norms. ln V. N. Parrillo (Ed.), Encyclopedia
ofSocial Problems, Thousand Oaks, CA: Sage Publications, 100-101. Klein, K. 1. and Kozlowski, S. W. 1. (2000), Multilevel Theory, Research, and
Methods in Organizations, San Francisco, CA: Jossey-Bass.
Kozlowski, S. W. J. and Klein, K. J. (2000),“A Multileve\ Approach to Theory
and Research in Organizations: Contextu訓， Temporal, and Emergent Processes," in K. 1. Klein and S. W. J. Kozlowski (Eds.), Multilevel Theo吵， Research, and Methods in Organizations, San Francisco, CA: Jossey-Bass, 3-90
Liao, H. and Chuang, A. (2004), “A Multilevel lnvestigation of Factors Inf1uencing Employee Service Perfonnance and Customer Outcomes,"
Academy of Management Journal，的(1)，41-58
Liao, H. and Chuang, A. (2007),“Transfonning Service Employees and Climate A Multilevel Multisource Examination of Transfonnational Leadership in Building Long-Tenn Service Relationship丸" Journal of Applied Psychology,92(4), 1006-1019
Liden, R. c., Sparrowe, R. T. and Wayne, S. 1. (1997), “Leader-Member Exchange Theory: The Past and Potential for the Future," ln. G. R. Ferris (Ed.), Research in Personnel and Human Resources Managemel呵， Vol.的，
Greenwich, CT: JAI Press, 47-119.
Lin, C. C. (2005)
“
A Cross-Level Examinations of Organizational Citizenship Behavior: an Analysis Using Hierarchical Linear Modeling," Journal of192 Examining Ihe Effecls of Dual-Ievel Transformalional Leaders峙" LMX, and Group Cohesiveness on Employee Service Peiformance
Management, 22(4), 503-524.(in Chinese)
Lin, N. (2001), Social Capital: A Theory of Social Structure and Action, New York, NY: Cambridge University Press
Lounsbury, M. and G1ynn, M.A. (ρ20ω01叫)，
Legitimacy, and the Acquisition of Resources," Strategic Management Journal, 22(6-7), 545-564
Luria, G. (2008),“Climate Strength: How Leaders form Consensus," Leadershψ Quarter紗， 19(1)，42-53.
MacKenzie, S. B., Podsako宜: P. M. and Rich, G. A. (2001),“Transformational and Transactiona1 Leadership and Salesperson Performance," Journal of the Academy of Marketing Science
,
29(2),
115-134.McGregor, J. (2009),“When Service Means Survival," Business Week, March,
2(4121),26-33
Menges, J., Wa1ter, F., Voge1, B., and Bruch, H. (2011) “Transformationa1 Leadership Climate: Performance Linkages. Mechanisms, and Boundary Conditions at the Organizationa1 Leve1," Leadersh伊 Quarter鈔， 22(5),
893-909.
Mudrack, P. E. (1989),“Group Cohesiveness and Productivity: A C10ser Look,"
Human Relations, 42(9), 771-785
Nielsen, K. and Danie1s, K. (2012), “Does Shared and Differentiated Transformationa1 Leadership Predict Follower' Working Conditions and Well-Being?" Leadersh伊 Quarter紗， 23(3)，383-398
Nooteboom, B. (1996) “Innovation, Leaming and Industria1 Organization," Cambridge Journal of Econom悶，23(2)， 127-150.
Podsako缸; P. M. and MacKenzie, S. B. (1995),“An Examination of Substitutes for Leadership within a Leve1s-of-Ana1ysis Framework," Leadership
Quarter鈔，6(3), 289-328.
Podsako缸 P. M., MacKenzie, S. B. and Bommer, W. H. (1996),
