Leadership and Team Values Shape Employee Engagement : Test of a Multilevel Moderated Mediation Model

The current research scrutinized how a leader’s communication and team value orientations interactively relate to employee engagement. The proposed model hypothesized that the impact of leadership on engagement would be mediated by followers’ trust toward a leader and this leadership-trust-engagement linkage would be moderated by team power distance and collectivism; in addition, employee voice behavior was examined as a behavioral manifestation of engagement. The results of multilevel structural equation modeling analyses with the data collected at a large electronics company in Japan (n = 638 members and 68 team leaders) revealed that transformational leadership was positively associated with employee trust and engagement when team power distance and collectivism were high, but not when those team values were low. Transactional leadership had negative effects on trust and engagement, regardless of team values. Finally, engagement was strongly positively associated with employee voice behavior. Theoretical and practical implications of these findings are discussed.


Introduction
The notion of employee engagement has received a great deal of attention over the last decade.Defined as "a positive, fulfilling, work-related state of mind that is characterized by vigor, dedication, and absorption" (Schaufeli, Salanova, Gonzales-Roma, & Bakker, 2002, p. 74), engagement is deemed to undergird desirable organizational behaviors (Cropanzano & Wright, 2001;Fleming, Coffman, & Harter, 2005;Luthans, 2002;Sonnentag, 2003).Highly engaged employees are so enthusiastic and pleasantly engrossed by their work that they lose track of the time, experiencing a state of flow (Csikszentmihalyi, 1998;May, Gilson, & Harter, 2004;Salanova, Bakker, & Llorens, 2006).Not only do engaged employees show positive attitudes but they also enact pro-organizational behaviors that are above and beyond prescribed duties (Bakker & Schaufeli, 2008;Kahn, 1990).At the same time, several reports point out that highly engaged workforce is hard to find, causing billion-dollar-scale loss of productivity, or so-called "engagement gap" problem (Bates, 2004;Kowalski, 2003;Richman, 2006).
Research has identified a number of individual-and group-level factors associated with engagement (see Albrecht, 2010, for a review).To date, however, integrative model of employee engagement that unifies those factors within a coherent framework has yet to be established.As Meyer (2008) put it, "to truly understand how engagement develops, we need more than a list of potential antecedents-we must be able to identify and explain the underlying mechanisms" (p.61).To address this limitation, the current research adopts Macey and Schneider's (2008) conceptual model and devises hypotheses by integrating relevant findings regarding team value orientation as a boundary condition.
Among the antecedent factors found in previous studies, communication and leadership style of the supervisor who works closely with employees seems notable (Sy, Cote, & Saavedra, 2005).Leaders can foster workers' job satisfaction, commitment, and willingness to "go the extra mile" by providing support and feedback (Bakker & Demerouti, 2007;Judge, Thoresen, Bono, & Patton, 2001;Mauno, Kinnunen, & Ruokolainen, 2007).Meyer (2008) posits that leaders' supportive communication enhances engagement because it helps employees to feel efficacy for the job, see the opportunities for self-improvement, and thereby develop intrinsic motivation (see also Gagné & Deci, 2005).Leaders can also promote members' sense of fulfillment and meaningfulness by arranging challenging environment and clarifying how their work adds to a bigger picture so that the members feel that they are contributing to something significant (Harter, Schmidt, & Hayes, 2002).Since efficacy, motivation, and perceptions of personal growth and meaningfulness are deemed indispensable for employees to engage themselves in the work (Csikszentmihalyi, 2003;Kahn, 1990Kahn, , 1992)), supervisors' communication as it relates to those perceptions provides an important antecedent factor of engagement.Macey and Schneider (2008) identify transformational leadership as the type of leadership behaviors that foster engagement.Transformational leadership is characterized by inspirational and support-oriented communication through ambitious goal setting, empowerment, and showing individualized concern (Burns, 1978;Podsakoff, MacKenzie, Moorman, & Fetter, 1990).Employees working with a transformational leader feel well-equipped with organizational resources necessary to achieve the set goal and even go beyond, and perceive that the leader cares about their personal success, growth, and well-being (Conger, 1999;Jung & Sosik, 2002).Under this style of leadership, team members not only strive for superior performance but also enact behaviors that are not necessarily in their personal best interest in order to benefit the team (Dvir, Eden, Avolio, & Shamir, 2002).
In addition, Macey and Schneider (2008) point out trust as a mediating factor that props up the inspiring effects of transformational leadership.To fully engage oneself in the work, employees need to trust that investment of their time, energy, and career opportunities will pay off in a meaningful way.As in any investment, however, the link between input (time and energy expended on work) and outcome (recognition and reward) involves uncertainty; to overcome this uncertainty, trust toward the leadership is crucially important.Because a transformational leader exhibits personalized care and concern for team members working under her or him, those members feel confident that their self-sacrifice and devotion will not be wasted.Without an established trust relationship, however, transformational leadership behaviors might not be effective because team members would not buy in the leader's encouragement to dare to challenge ambitious goals and hesitate to invest their limited personal resources (i.e., time, energy, and/or career opportunities).As such, trust provides a key mediator that connects supervisors' communication to employee engagement (Judge et al., 2001;Kahn, 1992).
At the same time, leadership does not manifest in a vacuum and not all workforces react in the same way to the given influence (Gong, Kim, Lee, & Zhu, 2013;Langford, 2009).Schaubroeck, Lam, and Cha (2007) found that the effect of leadership style on members' efficacy beliefs and performance is moderated by a team's power distance (PD) and collectivism orientations.High-PD and highly collectivistic teams readily appreciate leaders' influence, and thus, transformational leadership works more effectively in those teams than in low-PD or less collectivistic counterparts.Findings such as Schaubroeck et al.'s suggest that the way in which a leader's communication impacts members should vary by team value orientations.Thus, the current research focuses on team-level value orientations as a boundary condition of the effects of transformational leadership on employee engagement.
Finally, how engagement manifests as an organizational behavior needs to be explored.Most previous studies on engagement either presented a comprehensive review of the literature to clarify the conceptualization of the construct (Kahn, 1990;Macey & Schneider, 2008), or examined its operationalization vis-à-vis existing conceptually similar constructs (May et al., 2004;Rothbard, 2001).To address this limitation, I focus on employee voice, which is defined as a discretionary expression of work-related ideas with an intention to improve the functionality of one's work group (LePine & Van Dyne, 2001).This proactive and pro-organizational nature of voice is in line with the behavioral aspect of the engagement conceptualization (Macey & Schneider, 2008).Hence, it provides a useful window to examine how engagement relates to specific organizational behaviors.
Figure 1 presents an integrative model of leadership-trust-engagement-voice linkages qualified by the individualand group-level goal orientation.To explicate the conceptualization of each of those elements in the model and their theoretical associations, relevant literature is reviewed in turn below.Following this review, the results of an empirical study based on the data collected at a large Japanese electronics company are reported and the implications of the findings are discussed.

Figure 1. Overall theoretical model
Testing this integrative model empirically, the current research contributes to the development of the engagement literature.As noted earlier, the existing studies do explicate the construct of employee engagement and identify its related factors, but they do not necessarily present a fully unified view on the nomological network surrounding the construct.By exploring this underlying structure, the current research advances the theoretical understanding of the employee engagement phenomena.

Employee Engagement
Employee engagement is a multifaceted construct that encompasses not only "cool," cognitive perceptions of satisfaction and commitment but also "warm," affective elements such as excitement, passion, emotional attachment and identity (Harter et al., 2002;Langford, 2009;Macey & Schneider, 2008).Not only are engaged individuals dedicated to working hard but they also enjoy doing so; they willingly pursue goals that are above and beyond prescribed job duties and enthusiastically participate in extra-role activities to maximize their contributions and presence at work (Kahn, 1992).
Engagement is related to yet distinct from similar constructs, such as satisfaction and commitment (Meyer, Becker, & Vandenberghe, 2004).For example, satisfaction connotes satiation and adaptation to the status quo, whereas engagement involves passion, activation, and inclination toward potential changes (Macey & Schneider, 2008).Engagement also differs from job involvement, as the latter is a primarily cognitive concept representing the degree to which individuals relate themselves to the job and the work performed therein (Cooper-Hakim & Viswesvaran, 2005).Conversely, engagement has to do with one's emotional attachment as well as the determination to expend discretionary efforts to help the organization succeed and achieve something meaningful (Rothbard, 2001).In short, engagement entails emotions and actions, in addition to cognitions (May et al., 2004).
Employee engagement has been a hot topic in both academic and practitioner literatures over the past decade, as research has constantly shown that engagement is associated with superior business outcomes (Bates, 2004;Fleming et al., 2005).For example, Harter et al. (2002) found through their meta-analysis that business units where members are highly engaged with work demonstrate greater performance in terms of customer satisfaction, profitability, productivity, and turnover rate.It should therefore come as no surprise that top companies such as Apple and Google feature engagement as a key human resource management factor that signifies productive and loyal workforce (see Smythe, 2007).
Nonetheless, more research is needed to determine how engagement leads to the enactment of specific organizational behavior (i.e., engagement → communication) and also how a leader's communication fosters engagement (i.e., communication → engagement) (Richman, 2006).Engagement reflects mixed influences of one's work orientations, job attributes (e.g., challenge/excitement associated with the job), and communication with others at work (Bakker & Schaufeli, 2008;Robinson, Perryman, & Hayday, 2004).In particular, Macey and Schneider (2008) maintain that transformational leadership demonstrated by an immediate supervisor provides a key driver for the development of engagement.

Leadership and Engagement
Transformational leadership, as conceptualized by Bass and Avolio (1994), entails a leader's charismatic charm, inspirational motivation (i.e., presenting an appealing vision with optimism and enthusiasm, encouraging followers to pursue ambitious goals with high standards), intellectual stimulation (i.e., challenging taken-for-granted assumptions, proposing new ways of thinking and doing things), and individualized consideration (i.e., providing advice, feedback, and coaching to attend to followers' personal needs, abilities, and aspirations) (see Bass, 1999, for a review).Transformational leaders inspire followers to shift focus from personal interest to the good of the collective and embolden them to venture to surpass self-ascribed limitations.Transformational leaders also support followers by providing organizational resources and acting as a role model (Podsakoff et al., 1990).
Transformational and transactional leaderships.A contrasting style of leadership is transactional leadership, which is based on the exchange relationship maintained through "carrot-and-stick," contingent reinforcement of followers (Bass & Avolio, 1990).Transactional leaders emphasize rules and regulations, communicate expectations about desired outcomes clearly, specify paths to achieve the agreed-upon objectives, monitor followers' performance and take corrective actions if necessary.They openly recognize successful members for superior performance but do not necessarily encourage followers to pursue greater responsibility for further growth or strive for goals that transcend one's self-interest (Bass, 1985(Bass, , 1999;;Eagly, Johannesen-Schmidt, & van Engen, 2003).
In terms of the current research, transformational leadership is considered conducive to developing employee engagement but transactional leadership is not.Transformational leaders articulate the value of pursuing a vision with passion and challenge followers to go beyond the status quo, while showing individualized concern for their well-being.Such inspirational and supportive communication fosters employees' enthusiasm, efficacy, and perceived sense of fulfillment and meaningfulness, which, in turn, lead them to a mental state of engagement (Macey & Schneider, 2008).In contrast, transactional leaders focus on followers' needs for immediate rewards, rather than appealing to higher-level motivations for personal growth, mastery, and contributing to something significant (see Maslow, 1954).They specify what needs to get done to achieve predetermined goals but do not encourage followers to exceed the expectations.Those messages cultivate reactive, rather than proactive, orientations and therefore negatively impact the development of employee engagement.Hence, the first hypothesis: Hypothesis 1a: Transformational leadership is positively associated with employee engagement.Hypothesis 1b: Transactional leadership is negatively associated with employee engagement.
Trust as a mediator of the leadership-engagement linkage.Robinson et al. (2004) posit that, essentially, engagement is characterized by a two-way relationship, which implies that followers are not just a passive receptor of the leadership influence.Rather, they actively process a leader's messages to determine if it is worth engaging themselves and investing their time, energy, and career opportunities (Maslach, Schaufeli, & Leiter, 2001).Note that these processes inherently involve some uncertainty because leaders could exploit the followers' investment without giving return.According to Macey and Schneider (2008), employees need trust toward the leadership to overcome this dilemma and believe that the efforts and time they expend on the work will be rewarded in a meaningful manner.This is consistent with uncertainty management theory (Lind & Van den Bos, 2002), which suggests that followers use communication and relational history with a leader as a heuristic device to assess her or his trustworthiness.They rely on this assessment to decide the degree to which they should follow directives issued by the respective leader (Lind, 2001;Van den Bos & Lind, 2002).
Transformational leaders set ambitious goals, show individualized care, and empower their subordinates (Podsakoff et al., 1990).Those inspirational and support-oriented behaviors give employees with the perception that they are well-equipped with resources necessary to achieve the set goal and even go beyond.Organizational members reflect over such experiences to build trust toward transformational leaders and thus developed trust, in turn, provides the psychological scaffold of strong engagement.Together, it stands to reason that trust should function as a mediator of the leadership-engagement association (Macey & Schneider, 2008).Stated in the form of a hypothesis: Hypothesis 2: Team members' perceptions of trust toward a leader mediate the association between the demonstrated leadership and employee engagement.
Team value orientation as a moderator of the leadership impact.Organizational members working together on a regular basis come to develop shared values, forming a distinctive "team culture" (Levine & Moreland, 1991;O'Reilly, Chatman, & Caldwell, 1991).This value orientation shapes unique workplace climate, a platform upon which leadership is both exercised and processed (Langford, 2009).Research indicates that the impact of leadership varies appreciably by those workplace norms and values (Den Hartog et al., 1999;Ehrhart & Klein, 2001;Scandura & Dorfman, 2004;Schaubroeck et al., 2007).
Among the range of values theories available from the extant literature (e.g., Hofstede, 1980;Schwartz, 1992), the current research focuses on power distance and collectivism.Power distance refers to the degree to which members of a group accept uneven allocation of power and resources (Hofstede, 1980).Collectivism involves two primary notions: the extent to which individuals regard the group's norms and obligations as superordinate to their personal needs, and the extent to which they seek to maintain relational harmony with other in-group members (Earley, 1993;Markus & Kitayama, 1991).
Those two values are highlighted for their relevance to communication and leadership in organizational contexts.By nature, organizations need to manage a fundamental paradox between inequality and solidarity-on one hand, organizational members must consent to differential distribution of status/power, resources, and opportunities to operate within hierarchy; on the other hand, leaders and followers need to work together to maintain group cohesiveness (Kabanoff, 1991).Power distance reflects organizational members' orientation toward the uneven allocation and inequality, whereas collectivism reflects the members' agreement to share common goals and relational identity.
With regard to the model to be tested in the current research (Figure 1), transformational leadership should have greater impact on employee engagement and trust for teams high on power distance (PD) and collectivism.Members of high-PD work teams tend to be attentive to the expectations of high-status members, i.e., leaders (Earley, 1999).They show greater respect for authority, and therefore, accept the support and feedback attempts of a leader more readily than their low-PD counterparts (Adler, 2008).Transformational leaders embody engagement through their own work as a role model and express optimism that their team can accomplish something significant, while attending to individual needs of team members.Such inspirational and supportive communication should promote engagement and generate trust among work team members, especially in high-PD teams.
Similarly, transformational leaders should be more influential in collectivistic teams.As noted above, collectivistic members give precedence to the group's goal over their individual desires (Eby & Dobbins, 1997).Inspirational approach of transformational leadership to transcend personal interest and contribute to the good of a team should be highly compatible with those value orientations.Collectivism is also associated with the needs for affiliation and relational harmony (Hui & Villareal, 1989;Markus & Kitayama, 1991).Thus, a transformational leader's communication style to promote collaboration and show concern for followers' needs should be appreciated especially in collectivistic teams and the team members would trust such leadership.These lines of reasoning suggest the following hypothesis: Hypothesis 3a: The positive associations of transformational leadership with engagement and trust are moderated by team power distance; the higher the power distance, the greater the magnitude of the linkages of transformational leadership with engagement and trust.
Hypothesis 3b: The positive associations of transformational leadership with engagement and trust are moderated by team collectivism; the higher the collectivism, the greater the magnitude of the linkages of transformational leadership with engagement and trust.
The way in which power distance and collectivism interact with the effects of transactional leadership is less clear.High-PD team members are sensitive to a leader's expectations (Earley, 1999), while transactional leaders are not only clear about what they expect but also transparent in terms of how they decide to allocate rewards among team members (Bass & Avolio, 1990).This clarity would help to garner the perceptions of fairness and certainty about one's liability within a team; yet, it seems rather unrelated to affective interpersonal trust, which undergirds the development of engagement (Macey & Schneider, 2008).Further, fear-and-favor-based approach of transactional leadership is argued to run counter to collectivistic values (e.g., harmony and cooperation) because rewarding superior performance inevitably differentiate team members; as a consequence, such contingent reinforcement might work to increase psychological distance between a leader and followers.Finally, transactional leaders focus on maximizing efficiency within the existing organizational framework, rather than exploring for innovations or going above and beyond prescribed job duties (Bass & Avolio, 1990).Because high-PD and collectivistic team members tend to emulate a leader's attitudes and behaviors (Earley, 1999;Eby & Dobbins, 1997), they might adopt those orientations to avoid challenges to the status quo or extra-role activities.Based on these arguments, the hypothesis below is proposed: Hypothesis 4: Transactional leadership is negatively associated with engagement and trust, and these associations are moderated by team power distance and collectivism; the higher the power distance and collectivism, the greater the magnitude of the linkages of transactional leadership with engagement and trust.Macey and Schneider's (2008) model suggests that engagement entails extra-role behaviors as it drives organizational members to willingly "go the extra mile" to maximize their contributions (see also Saks, 2006).There are several theoretical reasons that buttress this notion.First, engagement involves strong positive emotions, energy, and enthusiasm (Macey & Schneider, 2008).Such positive affects increase one's openness to new ideas and optimism toward novel challenges, which, in turn, bring about the willingness to expand behavioral repertoire (Mauno et al., 2007).Second, high levels of involvement and absorption associated with the state of engagement help individuals to be alert and spot otherwise unnoticeable opportunities to make improvement (Fredrickson & Branigan, 2005;Fredrickson & Losada, 2005;Isen, 2000).Third, Meyer (2008) points out from a self-determination theory perspective (Deci & Ryan, 1985) that engaged employees would autonomously regulate behavior and attempt to express their sense of self by making unique contributions (see also Gagné & Deci, 2005).

Engagement and Employee Voice Behavior
The current research focuses on voice as a behavioral manifestation of employee engagement.Voice refers to a discretionary expression of opinions and ideas with an intention to trigger constructive changes at work (LePine & Van Dyne, 2001;Van Dyne, Cummings, & McLean Parks, 1995).Although voice can be construed in various ways with distinct implications (e.g., Liang, Farh, & Farh, 2012;Matsunaga, in press), the current research draws on Van Dyne and LePine's (1998) conventional view that conceptualizes voice as an explicit, speak-up act of "making innovative suggestions for change and recommending modifications to standard procedures even when others disagree" (p.109).
Engagement should provide a strong driver for voice behavior.By definition, voice is a discretionary behavior conceptually located outside regular work routines (LePine & Van Dyne, 2001).Voicing employees therefore need enthusiasm and commitment, as well as vigor, to go beyond what is expected as their job duties-namely, engagement.Additionally, engagement is uniquely associated with voice due to the risk of "speaking up" (Burris, 2012;Morrison, 2011).Unlike other discretionary cooperative behaviors that are intended to enhance an organization's functionality (Organ, Podsakoff, & MacKenzie, 2006), voice is essentially a challenging act.That is, voicing employees question the status quo for the sake of improvement; such behaviors, however, might be misunderstood as efforts to elevate one's visibility at the expense of troubling others, undermine the credibility of authority, or disrupt the flow of business (Morrison & Milliken, 2000;Tepper, Duffy, Hoobler, & Ensley, 2004).To overcome those risks and venture to "speak up," employees must have a strong sense of ownership with their organization and determination to contribute to its success even in the face of misunderstanding or initial refusal; engagement is argued to provide the basis for such dedication.Therefore: Hypothesis 5: Employee engagement is positively associated with the enactment of voice behavior.

Procedure & Participants
Data for the current research were collected from 68 work teams in a large electronics company in Japan.The data collection project was approved by the company's HR director on the condition that they would receive the summary of key findings and implications.
To reduce potential common method biases (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003), data were collected from two different sources.Members reported on their leader's leadership and communication style, trust toward her/him, and their power distance and collectivism orientations, as well as their own engagement level.Team leaders reported on members' voice behavior.Members and team leaders in various business divisions of the company were invited to take an online survey.The invitation message emphasized that participation was voluntary and the responses would be confidential.
Among 817 members who received the research invitation, 638 completed the survey (member response rate = 78.1%;all team leaders who received the invitation completed the survey).Business divisions represented by these participants ranged from consumer products to big data analytics to educational infrastructure solutions.On average, a team included 9.38 members (SD = 2.67; min.= 4, max.= 16).Of the 638 team members, 48% were female, the average age was 25.81 years (SD = 2.34), and the average tenure at the company was 3.52 years (SD = 0.96).Of the 68 leaders, 21% were female, the average age was 36.56 years (SD = 3.39), and the average tenure was 10.32 years (SD = 3.44).The team-average age difference between members and leaders was 10.71 years (SD = 3.60).
Transformational/transactional leaderships.Team leader's transformational and transactional leaderships were measured using items taken from Bass and Avolio's (1997) Multifactor Leadership Questionnaire (MLQ) Form 5X.There are four components in MLQ 5X to represent transformational leadership-idealized influence (e.g., "Talks about the importance of the team's ethics and values"), inspirational motivation (e.g., "Emphasizes the importance of having a collective sense of mission"), intellectual stimulation (e.g., "Seeks different points of view when solving problems"), and individualized consideration (e.g., "Helps team members to develop their strengths").Members rated their team leader in terms of the degree to which he or she would generally exhibit these behaviors.
Because the current research did not pose any specific hypothesis that those individual elements of transformational leadership would be differentially associated with employee engagement or trust, they were combined into a composite 1-factor scale.This composite approach is consistent with previous empirical works (Bono & Judge, 2003;Jung & Sosik, 2002;Kark, Shamir, & Chen, 2003) and the theoretical framework of transformational leadership (Avolio, Bass, & Jung, 1999;Bass, 1999).Further, a confirmatory factor analysis (CFA) modeling a higher-order factor provided empirical support for the composite approach.A second-order CFA model which specified one higher-order factor combining latent factors representing the four components of transformational leadership showed a better fit than a 4-factor CFA model which specified four components as correlated yet distinct latent factors (the higher-order model's fit: χ 2 [146] = 543.56,p < .01,RMSEA = .03,CFI = .97,SRMR = .02;the 4-factor model's fit: χ 2 [144] = 606.89,p < .01,RMSEA = .05,CFI = .93,SRMR = .05);the difference between the two model's fit was statistically significant: Δχ 2 (2) = 63.33,p < .01.Taken together, these lines of reasoning and evidence justified the decision to combine four elements of transformational leadership into one scale.The reliability alpha of the composite scale was .93.
Transactional leadership was assessed using four items of contingent reward (e.g., "Rewards us when we do what we are supposed to do") in MLQ 5X (Bass & Avolio, 1997).The alpha was .80.
Trust.Members reported on the trust they feel toward their team leader using McAllister's (1995) five-item affect-based trust scale (e.g., "We have both made considerable emotional investments in our working relationship").The reliability alpha was .88.

Value orientations.
Power distance was measured using Earley and Erez's (1997) eight-item scale (e.g., "Employees should not express disagreements with their managers").The reliability alpha for power distance was .85.Collectivism was measured using three items taken from Earley and Erez's scale (e.g., "I would rather struggle through a personal problem by myself than discuss it with others" [reverse coded]).The reliability alpha for collectivism was .82.
Engagement.Team members reported on their own engagement level using a 12-item Gallup Workplace Audit (also known as Gallup Q 12 ; Gallup Organization, 1993Organization, -1998)).Q 12 corresponds to the psychological conditions of engagement specified by Kahn (1990) (e.g., "At work, I have the opportunity to do what I do best every day" and "The mission or purpose of my company makes me feel my job is important").The reliability alpha was .90.
Voice.Team leaders also rated each of the members' voice behavior using a modified version of Detert and Burris's (2007) three-item measure; the modification was made to turn the measure into an other-report scale for leaders to report on team members, rather than on themselves (e.g., "S/He gives me suggestions about how to make this work unit better, even if others disagree" and "S/He speaks up to me with ideas to address employees' needs and concerns").The reliability alpha was .81.
Control variables.I included several control variables at both individual and team levels.At the individual level, members' age, sex, tenure, and the age difference between the team leader and members were included as covariates in hypothesis tests reported in the following pages.At the team level, the team leader's age, sex, and tenure, as well as team size, were controlled in the analyses.

Missing Data
Overall, missing response rate was 2.4%, with no discernable patterns.Those missing data were accommodated using the full-information maximum likelihood method (FIML; Arbuckle, 1996) in the subsequent analyses.FIML has been shown to provide more accurate estimates as compared to other procedures such as regression-based imputations (see Enders & Bandalos, 2001).

Measurement Model Tests
A series of confirmatory factor analyses were performed using Mplus 4.2 (Muthén & Muthén, 2007) to scrutinize the factor structure of the measures used in the current research.The first model specified all factors separately as originally postulated and measured.A second model specified all variables to load onto one large latent factor.A third model specified all member-reported variables loading on one latent factor, and all leader-reported variables loading on another latent factor.Finally, a fourth model specified the transformational and transactional leadership items to load onto one "leadership" latent factor, power distance and collectivism items to load onto another "value orientation" factor, and the rest of the items to load onto separate factors.The original (i.e., separate-factor) model fit the data well: χ 2 (1350) = 3834.0,p < .01,RMSEA = .04,CFI = .96,SRMR = .03;in addition, as shown in Table 2, this model's fit was statistically significantly better than the alternative models.Thus, the decision was made to retain the originally formulated measurement model for the main analyses.c Items reported by work team members (i.e., transformational leadership, transactional leadership, trust, power distance, and collectivism) were specified to load onto one latent factor, while those reported by team leaders (i.e., engagement and voice) were specified to load onto another latent factor.d Two leadership factors (i.e., transformational and transactional leaderships) were modeled together as one latent "leadership" factor; likewise, two value orientation factors (i.e., power distance and collectivism) were modeled together as one latent factor.

Non-Independence Analyses
To evaluate the appropriateness to specify a multilevel model (i.e., members nested within work teams), between-team and within-team agreement in individual responses was examined using two intraclass correlations (ICCs) recommended by James (1982).ICC(1) indexes the degree to which individuals' responses agree with one another within a given group, whereas ICC(2) indicates how well groups can be differentiated on the variable of interest.Following James (1982), ICC(1) ranging from .00 to .50 and ICC(2) exceeding .70 were considered acceptable to justify multilevel modeling.As shown in Table 3, the current research's data satisfied these conditions for all variables of interest, providing empirical support to take a multilevel approach for hypothesis tests.

Hypothesis Tests
To examine the hypotheses of the current research and test the proposed model (Figure 1) within a coherent framework, a series of multilevel structural equation modeling (ML-SEM) analyses were performed using Mplus 4.2 (Muthén & Muthén, 2007).One model featured transformational leadership and examined the paths associated with it, whereas another model highlighted transactional leadership; this approach was taken to keep the given model's structure consistent with the hypotheses advanced in the current research while maintaining the analysis parsimony to minimize Type I errors.
To further reduce the computational complexity, the item parceling procedure for structural equation modeling was applied (Little, Cunningham, Shahar, & Widaman, 2002).With the parceling procedure, all items used as the indicators of a given latent variable are aggregated into a single composite score, or parcel, and this parcel is used as a sole indicator of the latent variable.To avoid the underidentification problem (i.e., df < 0), the factor loading of the parcel is fixed at 1.0 and the error variance at (1 − α) × s 2 , where α represents the scale reliability and s 2 the observed variance of the variable of interest (see Bollen, 1989).Past research has demonstrated that parceling provides a useful technique to reduce the complexity of a structural model and stabilize the computation processes without compromising the accuracy of model fit and parameter estimation (Bandalos, 2002;Little et al., 2002).
Hypothesis 1 predicted that transformational leadership would be positively associated with employee engagement, whereas transactional leadership would have a negative association.As shown in Table 4, transformational leadership showed a small yet statistically significant positive association with engagement (β = .10,p < .05);transactional leadership also showed a significantly negative association (β = −.16,p < .01)as expected.The 95% confidence interval (CI) of the relationship between transformational leadership and engagement was (.02, .18),whereas the 95% CI of the relationship between transactional leadership was (−.22, −.09).Thus, Hypothesis 1 was supported.Theories of leadership (Bass & Avolio, 1994;Burns, 1978) posit that transformational leaders are morally uplifting, articulate ambitious goals, and emphasize collaboration; those leaders also show individualized concerns for members and provide support and feedback so that followers feel valued, perceive empowerment, and see how their work relates to the vision of their organization (Dvir et al., 2002;Hamstra, Sassenberg, Van Yperen, & Wisse, 2014;May et al., 2004).Such an inspiring yet supportive communication style of a transformational leader generates the feeling of trust, sense of meaningfulness, and psychological safety among followers (Jung & Avolio, 2000;Nemanich & Vera, 2009).According to Kahn (1990Kahn ( , 1992)), those feelings of trust, meaningfulness, and safety provide an indispensable psychological foundation of engagement, which requires employees to invest a great deal of time, efforts, and career opportunities.Because supervisors can always at least potentially exploit those investments without providing rewards in return (Whitener, Brodt, Korsgaard, & Werner, 1998), employees struggle with the uncertainty about whether or not their engagement will pay off in a meaningful manner.The current findings suggest that trust toward a leader-a belief that s/he will not unduly take advantage of one's vulnerability-fostered by transformational leadership helps employees to overcome the ambient dilemma of potential exploitation in the leader-member relationship.
On the other hand, transactional leadership capitalizes on the contingent exchange relationship through the administration of rewards and punishment (Avolio et al., 1999).Transactional leaders clarify their expectations and recognize high-performers.This clarity and transparency might help generate the perceptions of procedural justice among followers (Lind, 2001;Walumbwa, Wu, & Orwa, 2008).Nonetheless, the contingent exchange style of transactional leadership is argued to hamper the affective trust relationship between a leader and followers, because it heightens the risk of failure and disturbs psychological safety (Walumbwa & Schaubroeck, 2009).Such disturbance of safety and trust explains the negative effect of transactional leadership on employee engagement found in the current research.
It should be noted, however, that transformational and transactional leaderships are not always mutually exclusive and good leaders should exercise both styles (e.g., an inspiring leader communicates her or his expectations clearly and recognizes high-performers openly while attending followers' concerns and providing individualized support).Further, research indicates that the effects of leadership styles vary by organizational cultures, leader-member relationships, and individual attributes of involved parties (Parr, Hunter, & Ligon, 2013;Van Dierendonck, Stam, Boersma, de Windt, & Alkema, in press).For example, Howell and Hall-Merenda (1999) found that transformational leadership is more effective when followers feel psychologically close to their leader, which is consistent with the current findings of the mediating effect of trust on the relationship between leadership and engagement.Thus, although not directly examined in the current research, future studies should investigate how different styles of leadership interact to affect followers' affection, cognition, and behavior, and also explore individual-and group-level factors that moderate the impact of differential leadership styles.
Along with this line of speculations, the direct effects of transformational and transactional leaderships-which remained significant even after the mediation effect of trust was controlled (Table 4)-imply that there should be other mechanisms that convey the residual impact of leadership on engagement.Among a range of factors, the theoretical reasoning of the current research suggests that uncertainty should play a key role.As repeatedly alluded in this manuscript, the decision to engage oneself in the work involves uncertainty on the side of followers because of the potentially exploitative nature of the leader-member relationship (Whitener et al., 1998).In other words, uncertainty of whether one's engagement will be meaningfully rewarded may provide a complementary explanatory mechanism about the relationship between leadership and engagement (see Figure 4).Given the trust-mediated structure found in the current research (Figure 4A), there are several possibilities about how uncertainty might mediate the relationship between leadership and engagement.One possibility is that uncertainty provides another mediator to form a dual-process model (Figure 4B).This model is in line with several uncertainty-related theories such as problematic integration theory (Babrow, 2007;Babrow & Matthias, 2009), which posits that individuals draw on multiple distinct psychological mechanisms as they manage uncertainty.Another possibility is that trust and uncertainty not only mediate the impact of leadership but they also interact with one another to generate synergistic effects (Figure 4C).This model implies that some moderator variables differentiate the first-stage effects of leadership on trust and uncertainty-a pattern consistent with the moderation effects of team value orientations found in the current research.Finally, it is conceivable that leadership and/or trust might work as a "buffer" to mitigate the negative impact of uncertainty (Figure 4D).As noted above, uncertainty is deemed as a major hindrance of engagement development (Whitener et al., 1998).Thus, if leadership and trust mitigate the effect of uncertainty, they should indirectly enhance employee engagement.Future research should build on the current findings and systematically examine these theoretical possibilities to clarify the relationships among leadership, trust, uncertainty, and engagement.

Team value influences.
Another set of findings obtained in the current research concerns the moderation effects of team value orientations.Examinations through Edwards and Lambert's (2007) moderated path analytic procedure revealed that transformational leadership have a significant and positive effect on engagement via trust when team power distance (PD) is high, but not when team PD is low (Figure 2A).Team collectivism also showed such moderation effects (Figure 2B).In contrast, neither team PD nor team collectivism showed statistically significant moderation effects on the relationship between transactional leadership and employee engagement via trust (Figures 3A & 3B).
The positive moderation effects of team power distance (PD) on the impact of transformational leadership may reflect that the members of high-PD teams readily pay attention and respect to the authority figure (i.e., leader).
Prior research has shown that the individuals who are high on power distance accept hierarchy and embrace the leader's influence more fully than their low-PD counterparts (Den Hartog et al., 1999;Hofstede, 1980;Schaubroeck et al., 2007).Thus, teams composed of high-PD members provide a favorable climate for transformational leaders to exercise their inspiring influences and foster team members' dedication and engagement (Ehrhart & Klein, 2001;Langford, 2009).
Conversely, the positive moderation effects of team collectivism may be better explained in terms of the team-focused communication style of transformational leadership.Transformational leaders encourage followers to transcend personal ego for the good of the collective and emphasize the value of collaboration (Bass, 1998;Bono & Judge, 2003).This communication style is compatible with the orientations of collectivistic groups, where members prioritize the group's goal over their own and seek to maintain group harmony (Eby & Dobbins, 1997;Hofstede, 1980).Shamir, House, and Arthur's (1993) self-concept based theory suggests that the decision to follow a leader is an active process, in which individuals carefully examine if their leader genuinely represents the values they consider are important (see also Ashforth & Mael, 1989).It stands to reason that members of collectivistic teams determine to willingly accept a transformational leader's influence in view of the team-centered values presented through her or his communication; this consenting attitude of followers, in turn, enhances the positive impact of transformational leadership on trust and work engagement.
These lines of reasoning provide an account for why team values did not moderate the effects of transactional leadership.As for power distance, members of high-PD teams should be more or less equally susceptible to the influence of transactional as well as transformational leaderships.As discussed earlier, however, transactional leadership capitalizes on the contingent exchange relationship, which would not necessarily help promote trust and engagement.As for collectivism, note that the values represented by transactional leadership (e.g., efficiency and competitiveness) are at odds against those of collectivism.This incongruence might drive collectivistic team members psychologically away from transactional leaders (Shamir et al., 1993) and, as a result, transactional leadership and team collectivism do not synergize to engender positive interaction effects on those members' engagement.
Engagement-voice linkage.Additionally, the linkages found in the current research between engagement, value orientations, and employee voice should merit a separate discussion.Voice behavior, or speaking up of constructive ideas to improve a work group's functionality, has been focused in the literature as an important organizational behavior (Van Dyne & LePine, 1998).At the same time, voice is difficult for employees to enact because it essentially challenges the status quo and calls into question the validity of the current practice, procedure, and/or personnel (Crant, 2000;Morrison, 2011).
On this front, the strong positive effect of engagement on voice (β = .47,p < .01)found in this study deserves particular attention.Engaged employees are not only highly satisfied but they also show strong commitment and emotional involvement (Harter et al., 2002;Kahn, 1992).It is posited that this involvement helps employees to overcome the fear of speaking up and initiate difficult communicative endeavor to verbalize the problems of their work group (Rees, Alfes, & Gatenby, 2013).
The ways in which value orientations affect voice seem more complex.On one hand, power distance (PD) and collectivism amplify the positive effect of transformational leadership on engagement, and therefore, indirectly foster voice behavior.On the other hand, PD had a small yet significant negative effect on voice at the individual level (β = −.16,p < .01).High-PD individuals willingly accept the notion of hierarchy, and thus, it is difficult for them to enact voice, which by nature casts doubt on the validity of organizationally authorized practices.The current research suggests that transformational leaders can leverage those ambivalent effects of PD and collectivism by capitalizing on followers' engagement and challenging them to surpass personal interests to contribute to the benefit of the group.

Practical Implications
The current findings suggest that high-PD and collectivistic team members supervised by a transformational leader are most engaged and likely to enact voice behavior.This insight brings at least two managerial implications.First, companies should implement training programs designed to enhance transformational leadership skills.Although indirect, promoting transformational leadership should help develop a leader's trust relationship with followers and thereby foster their engagement and voice.
Second, leaders who wish to promote engagement by drawing on transformational leadership should understand that the impact of leadership is contingent upon team contexts.Most leadership trainings are leader-centric (House & Aditya, 1997;Riggio, 2008).The current findings suggest, however, that leaders should be able to identify and shape followers' value orientations to maximize the impact of their leadership (Chatman & Cha, 2003).Specifically, leaders need skills to uphold a healthy level of respect for their authority and promote team-centered, collaboration-focused values among followers to maintain team contexts that are conducive to transformational leadership.At the same time, the idea of fostering team power distance needs cautious evaluation, given its negative impact on voice (i.e., high power distance would increase the unwillingness of employees to share ideas, causing a loss of valuable upward feedback; Botero & Van Dyne, 2009).Future studies should explore the conditions that differentiate when team values lead to desirable follower outcomes from when they do not.

Limitations & Future Directions
The current findings should be interpreted with its limitations in mind.First, the correlational, cross-sectional design of the study prohibits a conclusive statement about the causal direction of the links among leadership, trust, engagement, voice, and team values.It may be possible, for example, that individuals who actively express opinions become trusted by their supervisor (voice → trust), or leaders feel comfortable to present ambitious goals to such engaged followers (engagement → transformational leadership).Future studies should draw on the follower-centric perspective of leadership (Salanova et al., 2006;Tee, Ashkanasy, & Paulsen, 2013) to explore these reverse possibilities.
Second, this study did not examine how employees' engagement affects their work team climate.Nonetheless, prior research indicates that engagement is contagious.If a certain member is highly engaged, her/his passion, enthusiasm, and drive for high performance transfer to the surrounding members (Bakker, 2011;Bakker & Xanthopoulou, 2009).Such contagion effects can be both direct and indirect.Engaged workers directly enhance their colleagues' engagement by acting as a role model and stimulating them.Engaged workers also indirectly promote others' engagement through "job crafting" and inducing positive changes in the workplace (Bakker, 2010;Petrou, Demerouti, Peeters, Schaufeli, & Hetland, 2012); thus improved work environment, in turn, fosters other members' engagement levels.Future studies should scrutinize the mechanism underlying such mutually-reinforcing processes between engagement and work team contexts to further advance the understanding of engagement development.
Third, this study examined the impact of a leader's communication style holistically in the light of transformational and transactional leaderships.This holistic approach was taken by design for the current research to be true to Macey and Schneider's (2008) model and test its key tenets associated with transformational leadership.Nonetheless, examination of more specific communication behaviors would reveal how particular leadership style impacts employees and affects their engagement.
Fourth, sampling conducted in only one culture (i.e., Japan) should be noted as an important limitation.Thus, although the current research drew on culturally grounded value constructs (i.e., power distance and collectivism; Hofstede, 1980), how culture influences the relationships among leadership, engagement, and employee voice remains to be addressed through future cross-cultural examinations.
Finally, the current research has examined engagement in terms of one's general mindset toward work.Different conceptualizations and operationalizations could have resulted in different findings.For example, distinguishing broad organizational/work engagement from specific job/task engagement would reveal interesting patterns that are left unexplored in the current research (see Macey & Schneider, 2008, for a discussion of various levels of conceptualization of engagement).

Conclusion
As the demands for highly engaged workforce increase (Kowalski, 2003;Richman, 2006), organizations face a pressing need to foster engagement among their employees.The current findings suggest that leadership and team values interactively shape employee engagement.In this sense, engagement is a communicative and relational phenomenon, which reflects the leader's communication style and work team contexts.Leaders should be mindful about how their communication is received by team members and thereby strategically utilize leadership skills to promote employee engagement.

.
n = 638 team members and 68 teams.PD = Power Distance.All parameter estimates (βs) are standardized.More detailed results are available from the author upon request.* p < .05. ** p < .01. Figure

Figure 4 .
Figure 4. Theoretical models of the relationships among leadership, trust, uncertainty and engagement

Table 1 .
Means, Standard Deviations, and Correlations

Table 2 .
Comparison of measurement models * p < .05. ** p < .01.a Items were specified to load onto seven separate factors (i.e., transformational leadership, transactional leadership, trust, power distance, collectivism, engagement, and voice) as originally postulated in the current research.b All items were specified to load onto one latent factor.

Table 4 .
Full-Information maximum-likelihood multilevel structural equation modeling results a