Synthesis Paper – Leadership Behavior

profilernixon@12345
1-s2.0-S1048984313001409-main.pdf

The Leadership Quarterly 25 (2014) 544–562

Contents lists available at ScienceDirect

The Leadership Quarterly

journal homepage: www.elsevier.com/locate/leaqua

Same difference? Exploring the differential mechanisms linking servant leadership and transformational leadership to follower outcomes

Dirk van Dierendonck ⁎, Daan Stam, Pieter Boersma, Ninotchka de Windt, Jorrit Alkema Rotterdam School of Management, Erasmus University, The Netherlands

a r t i c l e i n f o

⁎ Corresponding author at: Rotterdam School of Man fax: +31 4089015.

E-mail address: [email protected] (D. van

1048-9843/$ – see front matter © 2013 Elsevier Inc. A http://dx.doi.org/10.1016/j.leaqua.2013.11.014

a b s t r a c t

Article history: Received 26 July 2012 Received in revised form 26 September 2013 Accepted 18 November 2013 Available online 9 December 2013

Handling Editor: Shelly Dionne work engagement; however, the manner in which they exerted their influence differed. SL

This paper aimed to provide insights into the different mediating mechanisms through which servant leadership (SL) and transformational leadership (TFL) affect followers. We also investigated environmental uncertainty as a moderator of the effects of servant leadership and transformational leadership. Based on the results of two experimental studies and one field study, we concluded that both SL and TFL were related to organizational commitment and

worked primarily through follower need satisfaction, whereas TFL worked mainly through perceived leadership effectiveness. The moderating influence of uncertainty was inconsistent across the studies.

© 2013 Elsevier Inc. All rights reserved.

Keywords: Servant leadership Transformational leadership Engagement Commitment

1. Introduction

Leadership as a topic in management has generated an abundance of research over the past several decades. Although it would be going too far to suggest that leadership scholars agree on which behaviors and styles are optimal for leadership, it is clear that one style, more than any other, has been found to be effective: transformational leadership (TFL) (see the meta-analysis by Judge & Piccolo, 2004). However, as business environments change, leadership may face new challenges. One particularly important trend in this respect is the growing dependency on people in a knowledge-based economy, which makes attention to the needs of employees essential for long-term success (O'Leary, Lindholm, Whitford, & Freeman, 2002). Consequently, scholars have recently investigated a type of leadership that is particularly oriented to the needs of employees, known as servant leadership (SL), and although research on SL is in a relatively early stage, empirical findings regarding SL are promising (Van Dierendonck, 2011). However, several scholars have emphasized the considerable overlap between SL and TFL. An important aspect for research on SL is, therefore, investigating whether SL is actually different from TFL and, if so, how.

Several theoretical papers have argued that SL and TFL have different foci and may be suitable to different environments; TFL focuses on organizational effectiveness, whereas SL focuses on follower needs (Bass, 1985; Graham, 1991, 1995; Mayer, Bardes, & Piccolo, 2008; Van Dierendonck, 2011). Graham (1991) argued that where TFL emphasizes the leader's skills, hierarchical power relationships between leader and follower, visions for the organization, and especially performance, effort, and achieving the goals set out by the leader, SL emphasizes the humility and spirituality of leaders, mutual power, visions of a way of life for the leader and followers, emulation of the leader's service orientation, and the autonomy and moral development of followers. Graham (1995) added that while SL accomplishes OCB among followers by causing followers to reason in terms of universal

agement, Erasmus University, Burg Oudlaan 50, 3062 PS Rotterdam, The Netherlands. Tel.: +31 4089569;

Dierendonck).

ll rights reserved.

545D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

principles and justice, TFL does so by applying to utilitarian calculus and cost–benefit analysis for stakeholders. Smith, Montagno, and Kuzmenko (2004) also emphasized that while TFL focuses on change and organizational innovation and is especially effective in times of uncertainty, SL seems more oriented on preserving the status quo and focusing on individual people and is especially effective in time of stability (cf. Choi & Mai-Dalton, 1998). This suggests that SL and TFL affect outcomes through different processes (TFL through processes related to organizational effectiveness and SL through processes related to follower need satisfaction) and are effective under different circumstances (TFL would be especially effective under uncertainty, while SL would be especially effective under stability).

Unfortunately, however, the few empirical studies that have investigated these fundamental differences have gone no further than establishing the divergent validity of SL and TFL and demonstrating that SL explains unique variance in outcomes (such as follower commitment) beyond the effects of TFL (see, for instance, Barbuto & Wheeler, 2006; Liden, Wayne, Zhao, & Henderson, 2008; Peterson, Galvin, & Lange, 2012). Although these studies answer the question of whether SL and TFL are different, to test the fundamental differences between SL and TFL (i.e., how TFL and SL are different), there is a need for empirical research that goes beyond the correlation between SL and TFL and directly investigates the different underlying mechanisms through which these forms of leadership affect outcomes and the different boundary conditions for these effects. This is what the current research aims to accomplish.

In a series of three studies using a variety of research methods (experiments and survey research), we investigated the relationships between SL and TFL and between commitment and engagement. Importantly, we study whether the effects of TFL are mediated by followers' perception of leadership effectiveness and whether the effects of SL are mediated by followers' need satisfaction. Moreover, we test the moderating effect of times of uncertainty, specifically, whether the effects of TFL are stronger and those of SL are weaker in times of greater uncertainty.

By going beyond investigations of whether SL and TFL are different constructs and investigating directly how SL and TFL are different, the current research aims to contribute to the leadership literature in various ways. First, prior research has mainly focused on the discriminant validity of measures of SL and TFL. However, the fact that measures of two concepts can be discriminated does not imply that the theoretical concepts are different. The current research provides a first step to differentiating between SL and TFL based on theoretical models that specify mediating and moderating pathways of effects of SL and TFL and as such provides much-needed support for the notion that SL and TFL are indeed different theoretical concepts that work through different processes. Second, comparing the mechanisms underlying the effects of SL and TFL provides a much better understanding of why these leadership styles are effective. Consequently, this provides the field with insights into when and where the different leadership styles can be optimally effective (i.e., ideas for moderators to investigate). For instance, the finding that need satisfaction underlies the effectiveness of SL implies that in environments in which need satisfaction is a constraint and cannot possibly be expected to change, SL may not be very effective. Thus, an understanding of the differential mechanisms underlying SL and TFL can form the basis for theorizing about boundary conditions for these styles. In the current manuscript, the moderation of uncertainty is an example of such novel theorizing. Third, our research also has practical value. By showing that SL and TFL work through different processes, we provide information to managers and companies that may affect their decision to promote one style over the other. Organizations that emphasize need satisfaction may choose SL as their preferred leadership style for managers, while organizations that are more oriented to perceptions of effectiveness may prefer TFL. Our research provides the basis for more evidence-based decisions concerning management styles.

In the following, we first detail what TFL and SL refer to, and we subsequently develop a comprehensive conceptual model of the process through which these leadership styles affect the commitment and engagement of followers. Finally, we discuss the studies conducted.

2. Transformational leadership and servant leadership

TFL refers to a multidimensional leadership style that encourages followers to perform beyond expectations and emphasizes collective values and needs rather than followers' individual values and needs (Bass, 2005; Yukl, 1999). The different definitions of TFL have a common primary focus on organizational goals: transformational leaders inspire their followers to perform better for the sake of the organization. Rewards and praise are used to encourage a stronger focus on achieving high outcomes (Rafferty & Griffin, 2004). TFL theory discusses various elements (or dimensions) of leader behaviors. For instance, Bass (1985) includes inspirational motivation (communicating a stimulating vision), idealized influence (serving as a motivating role model), intellectual stimulation (stimulating followers to think outside of the box), and individualized consideration (an emphasis on followers' development). Rafferty and Griffin (2004) add personal recognition (recognizing the performance of followers) to these elements. Although often contrasted with transactional leadership (a leadership style that emphasizes the exchange relationship between leaders and followers and focuses on explaining and setting goals and providing rewards/punishment), TFL is not the opposite of transactional leadership but instead a leadership style that surpasses explanations, goal setting, and providing rewards for follower performance (Bass, 1985).

TFL is generally viewed as an effective leadership style, and studies show that TFL has many positive effects. For instance, TFL positively predicts work motivation (Shamir, House, & Arthur, 1993), employee satisfaction (Podsakoff, MacKenzie, Moorman, & Fetter, 1990), the number of accidents in warehouses (De Koster, Stam, & Balk, 2011), and innovative performance (Nederveen Pieterse, Van Knippenberg, Schippers, & Stam, 2010). For overviews of the effects of TFL, see Lowe and Kroeck (1996) and Bass and Riggio (2006).

546 D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

Notably, after decades in which TFL was the most widely studied leadership style, SL has been receiving increased attention in the leadership field (Van Dierendonck, 2011). The literature on SL advocates that servant leaders must primarily meet the needs of others (Greenleaf, 1977). Servant leaders focus on developing employees to their fullest potential in areas of task effectiveness, community stewardship, self-motivation, and future leadership capabilities, and they provide vision and gain credibility and trust from followers (Farling, Stone, & Winston, 1999). To generate the best performance in their followers, servant leaders rely on one-on-one communication to understand the abilities, needs, desires, goals, and potentials of their followers. With knowledge of each follower's unique characteristics and interests, leaders then assist followers to achieve their potential (Liden et al., 2008). Barbuto and Wheeler (2006) described SL as including an altruistic calling, which is the motivation of leaders to place others' needs and interests ahead of their own, and organizational stewardship, which orients others toward benefiting and serving the community.

Recent research on SL has shown promising results for its basic premises. For instance, Van Dierendonck and Nuijten (2011) found a relationship between SL and job satisfaction and work engagement (using various different samples, including high school staff, civil servants, and gas station employees). Other research has found relationships with trust and team performance (Schaubroeck, Lam, & Chunyan Peng, 2011), organizational citizenship behavior (Walumbwa, Hartnell, & Oke, 2010), team potency (Hu & Liden, 2011), and firm performance (Peterson et al., 2012). For an overview of the SL research, see Van Dierendonck (2011).

The descriptions of TFL and SL emphasize that there is considerable overlap between the two leadership styles. Both transformational and servant leaders are focused on their followers, extend leadership beyond simply setting and explaining task goals, provide visions for the future, and are generally positively correlated with various important outcome measures. However, we assert that there may be major differences in the way that they influence their followers and in the extent that their effectiveness is influenced by the environment. We will elaborate on these differences in the next section.

3. The differential influences of servant leadership and transformational leadership

As formulated by Stone, Russell, and Patterson (2004. P. 1) ‘The extent to which the leader is able to shift the primary focus of leadership from the organization to the follower is the distinguishing factor in classifying leaders as either transformational or servant leader’. Barbuto and Wheeler (2006) suggested that another important difference between SL and TFL is that SL focuses on a desire to serve and preparing others to serve as well, whereas TFL emphasizes a desire to lead and inspiring followers to perform well. Empirical research into these differences between SL and TFL has mostly been performed in the process of scale development. For instance, in their scale development efforts, Barbuto and Wheeler (2006), Liden et al. (2008), and Van Dierendonck and Nuijten (2011) were able to show that for their respective instruments, SL could be meaningfully and statistically distinguished from TFL.

Regretfully, none of these studies tests the fundamental theoretical differences between SL and TFL as proposed by early theorizing (see also Graham, 1991, 1995; Smith et al., 2004). In the following, we focus first on the different mechanisms underlying TFL and SL (i.e., attributions of leadership effectiveness and psychological needs) and how these affect outcomes (engagement and commitment of followers). We chose to focus on the above two mechanisms because they are suggested as the primary mechanisms that underlie the effects of TFL and charisma (in the case of perceptions of leadership effectiveness; Sy, Choi, & Johnson, 2014) and SL (in the case of psychological needs, Mayer, 2010) and as such should be the most distinguishing mechanisms. Next, we address how the environment (i.e., uncertainty) can differentially moderate the effects of TFL and SL. We chose uncertainty as a moderator because research suggests opposite moderation effects of uncertainty on the effects of TFL (positive moderation) and SL (negative moderation), thereby providing us with a moderator that optimally distinguishes between TFL and SL.

3.1. Attributions of leadership effectiveness

Perceptions of followers concerning their leaders' effectiveness play an important role in leadership (cf. Meindl, Ehrlich, & Dukerich, 1985). Attributions of leadership effectiveness are important because they provide followers with a sense of trust in their leader; individuals are more likely to follow leaders who they believe to be competent and effective. In this sense, the perception of leadership effectiveness represents the overall evaluation of leader more than any other possible mediating mechanism. We argue that TFL more strongly affects such attributions than SL. We base this proposition on several streams of research.

First, whereas TFL places the leader at the center of the leadership process, SL emphasizes the follower. TFL inherently has a charismatic component (Bass, 1985). By communicating a stimulating vision and acting as role models, transformational leaders often lead followers to attribute charisma and leadership effectiveness to leaders. It is through processes of vicarious learning and personal identification (Yukl, 1999) that followers of transformational leaders learn the norms and values of the organization. The high correlations between TFL and leader effectiveness in the meta-analysis of Judge and Piccolo (2004) are consistent with this reasoning. Moreover, Sy and colleagues (2014) recently showed a reciprocal relationship between charisma and leader effectiveness perceptions in a longitudinal study.

SL, on the other hand, is less leader-focused than TFL (cf. Graham, 1991, 1995). TFL puts the leader at the center of the group, while servant leaders will attribute successes to followers instead of themselves. When considering the various scales of SL, it is telling that many incorporate dimensions such as altruistic calling (Barbuto & Wheeler, 2006), putting subordinates first (Liden et

547D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

al., 2008), and standing back and humility (Van Dierendonck & Nuijten, 2011). Although such behaviors may be crucial to developing followers' feelings of competence and impact, they may come at the expense of being observed as extra-ordinary and a main source of performance. Thus, we suggest that TFL is highly visible and uses influence processes that rely on the leader taking the stage, whereas SL is less visible and uses influence processes that allow SL leaders to be more in the background.

Second, research on the romance of leadership (Meindl et al., 1985; Meindl & Ehrlich, 1987) argues that due to their notable position, leaders are generally attributed with more success (and failure) than would be predicted based on their behaviors. Interestingly, people even consider leader-attributed performance to be more positive than performance attributed to non-leader factors (Meindl & Ehrlich, 1987). In other words, leaders are accredited with more performance influence than is reasonable, and such performance is also observed as more positive than objectively appropriate. TFL uses this process to its advantage. By putting the leader center stage, the romantic image of leadership is optimally used to create an aura of effectiveness and charisma that is TFL's main source of influence. In other words, in addition to being very visible, group performance is even more highly attributed to TFL leaders due to this visibility. On the other hand, SL actively tries to remedy the notion of romance of leadership by showing humility, standing back, and generally emphasizing that it is the followers, rather than the servant leader, that are the cause of performance. To the extent then that they succeed, followers may therefore be less inclined to attribute the group's successes to the leader. Thus, SL is not only less visible than TFL, but this lack of visibility (combined with the leader's attribution of successes to followers) may also translate into SL leaders being less accredited for the group's successes. This leads to the following hypothesis:

Hypothesis 1. Transformational leadership is more strongly related to follower perceptions of leadership effectiveness than servant leadership.

3.2. Psychological needs

The fulfillment of basic psychological needs is a key determinant of health and well-being (Ryan & Deci, 2000). Many such psychological needs exist, but a dominant theory in the field, self-determination theory, suggests that there are three crucial elements of psychological needs for people (Ryan & Deci, 2000). Competence refers to effectively acting on and influencing one's environment. Autonomy is the experience of one's own will and initiative in one's own behavior. Relatedness refers to feelings of connection and belonging. The need for competence can be satisfied by offering optimal challenges and providing relevant feedback (Ryan & Brown, 2003). Autonomy is achieved through the informal control of employees by the leader and the perception of the freedom of choice on behalf of the individual. The need for relatedness is satisfied when the individual experiences warmth, acceptance, and care. The fulfillment of these needs can enhance an individual's intrinsic motivation and result in a sense of self-determination.

Concern for the needs of followers is more strongly emphasized in SL theory than in any other leadership theory (Mayer, 2010). Servant leaders invest time and energy to understand the needs of all individual followers and subsequently work to satisfy these needs. It is the core element of the famous quote by Greenleaf (1977) on what he considers the true test of servant leadership: ‘The difference manifests itself in the care taken by the servant-first to make sure that other people's highest priority needs are being served. The best test, and difficult to administer, is: Do those served grow as persons?’ There is also empirical support for the relationship between SL and need satisfaction (Mayer et al., 2008).

In TFL theory, personal attention for followers is aimed at enhancing innovation and creativity (Smith et al., 2004) to ultimately reach another – organizational – goal. In SL, personal attention is given for its own sake, to help followers grow as persons, and organizational goals are secondary (Greenleaf, 1977). Graham (1995) suggests that although TFL, similar to SL, encourages constructive participation, the level of moral development is primarily utilitarian and focused on the costs and benefits of the stakeholders. In contrast, SL comes from recognition of universal principles and a focus on justice and the greater whole. Indeed, while TFL incorporates elements related to a focus on individuals overall, the style strongly emphasizes the organization. This alignment is not problematic when the individuals' and organization's needs are in line, and organizational needs are deemed more important when the individuals' needs otherwise. Servant leaders would put the needs of followers first, even if they would conflict with their personal or the organization's interests. This leads to the following hypothesis:

Hypothesis 2. Servant leadership is more strongly related to the satisfaction of the psychological needs of followers than transformational leadership.

3.3. Organizational commitment and work engagement

Commitment can be broadly defined as “a force [that] binds an individual to a course of action that is of relevance to a particular target” (Meyer & Herscovitch, 2001, p. 301). Although different forms of commitment can be observed in studies of organizational commitment (i.e., affective, continuance, and normative commitment; see Allen & Meyer, 1990), affective commitment tends to be the most relevant form for predicting positive organizational behaviors (Mathieu & Zajac, 1990). This result has been confirmed by several meta-analyses with a broad range of outcomes, including job performance, turnover, absenteeism, and organizational citizenship behaviors (OCBs) (e.g., Mathieu & Zajac, 1990; Meyer, Stanley, Herscovitch, & Topolnytsky, 2002; Riketta & Van Dick, 2005). Affective organizational commitment can be observed as “an emotional attachment

548 D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

to, identification with, and involvement in the organization” (Meyer et al., 2002: 21). Similar to affective commitment, work engagement also refers to an attachment to, identification with, and involvement in an object or activity, but in this case, the object of this attachment is the work itself rather than the organization. Work engagement has been defined as consisting of three dimensions: dedication to work activities, absorption in work activities, and vigor in the pursuit of work activities (Schaufeli, Bakker, & Salanova, 2006).

Prior research has shown that TFL enhances organizational commitment (Avolio, Zhu, Koh, & Bhatia, 2004; Bono & Judge, 2003; Dvir, Eden, Avolio, & Shamir, 2002) and work engagement (Zhu, Avolio, & Walumbwa, 2009). Studies on SL have similarly suggested that SL fosters work engagement and affective commitment. For instance, research has shown that SL positively predicts work engagement (Van Dierendonck & Nuijten, 2011), while Liden et al. (2008) demonstrate that SL strongly predicts organizational commitment. This leads to the following hypothesis:

Hypothesis 3a. TFL and SL both positively influence organizational commitment and work engagement.

We maintain that two pathways through which SL and TFL affect commitment and engagement of followers are attributions of leader effectiveness and satisfaction of psychological needs. Perceptions of leader effectiveness provide followers with the feeling that their leaders and organizations are competent and worthy, and such positive evaluations may foster engagement and commitment on the part of followers. Indeed, research has shown that perception of effectiveness (and charisma) may positively affect such outcomes as commitment and engagement (cf. DeGroot, Kiker, & Cross, 2000). Furthermore, research on need satisfaction indicates that need satisfaction is positively related to organizational commitment (Brown, 1969; Hall, Schneider, & Nygren, 1970) and work engagement (Deci, Ryan, Gagne, Usunov, & Kornazheva, 2001; Van Den Broek, Vansteenkiste, De Witte, & Lens, 2008) because followers reciprocate the need satisfaction that they experience.

Combining these findings with the idea that SL and TFL influence perceptions of leadership effectiveness and satisfaction of the psychological needs of followers suggests that the latter may indeed be mediators of the relationship between SL and TFL as well as follower commitment and engagement. Considering that TFL is a strong predictor (and stronger than SL) of attributions of leadership effectiveness and that SL is a strong predictor (and stronger than TFL) of the need satisfaction of followers, we suggest that the mediation of attributions of leader effectiveness is especially strong for TFL, while the mediation of need satisfaction is especially strong for SL.

Hypothesis 3b. Attributions of leader effectiveness mediate the influence of TFL and SL on organizational commitment and work engagement, but the mediation of the influence of TFL is much stronger than the mediation of the influence of SL.

Hypothesis 3c. Followers' psychological need satisfaction mediates the influence of SL and TFL on organizational commitment and work engagement, but the mediation of the influence of SL is much stronger than the mediation of the influence of TFL.

3.4. Moderating effect of uncertainty

So far, we have argued that SL and TFL work through different processes (attributions of leadership effectiveness versus psychological need satisfaction), but that both affect important outcomes positively. These differential pathways are crucial to empirically studying the fundamental, theoretical differences between SL and TFL (how they work differently). However, understanding these different mechanisms is not only important for theory concerning SL and TFL but may also add important insight into the circumstances under which SL and TFL may be more (or less) effective. Here, we argue that, based on the differential mediating pathways of SL and TFL, an important moderator of the effects of SL and TFL is environmental uncertainty. Importantly, we argue that uncertainty strengthens the effects of TFL but weakens the effects of SL. Let us explain these ideas.

When times are uncertain, for instance, during economic crises, people experience high risk and turbulence (Antonakis, Avolio, & Sivasubramaniam, 2003), which may generate feelings of distress and uncertainty (Stubbart, 1987) and undermine people's feelings of safety (Lane & Klenke, 2004). A consequence of feelings of uncertainty and its associated negative affect is that people aim to reduce this uncertainty. There are several ways in which they may do so.

First, uncertainty identity theory (Hogg, 2007) argues that people use associations with social groups to reduce uncertainty. Groups provide safety, resources, and support; therefore, in periods of high uncertainty, people seek out groups to associate with or identify more strongly with groups they are attached to. Consequently, as followers tend to focus on groups and organizations more than themselves for comfort and safety in times of uncertainty, we argue that collective, organizational needs also become more pronounced for followers, while personal needs become less salient (cf. Hogg, 2007). Psychological need satisfaction therefore becomes more oriented toward the satisfaction of organizational needs (as opposed to personal needs).

Second, in times of uncertainty, the conventional routes of behavior become obsolete and ineffective; some even define crises and uncertainty as events that are ‘…characterized by ambiguity of cause, effect, and means of resolution, as well as by a belief that decisions must be made swiftly’ (Pearson & Clair, 1998, p. 60, see also, Dutton, 1986; Madera & Smith, 2009; Pearson & Mitroff, 1993). Leader effectiveness attributions will also be affected as people's implicit leadership models (ideas about the traits of an ideal leader; cf. Lord & Maher, 1991) change. As a consequence of experiencing uncertainty, people look for leaders who use

549D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

unconventional means to reach their goals and leaders who are determined and heroic because such leaders would be best able to guide them through uncertain times (cf. Choi & Mai-Dalton, 1998).

In summary, uncertainty creates in followers an urge for leaders who emphasize the organization's needs and act unconventionally and heroically. We argue that TFL fits this picture perfectly. TFL, more so than SL, aims to transform followers' needs from primarily individual needs to primarily organizational needs (Bass, 1985; Shamir et al., 1993). As such, transformational leaders emphasize the organization to a great extent. Moreover, TFL emphasizes changes and innovation (Nederveen Pieterse et al., 2010), which closely resembles the use of unconventional means, and has a strong charismatic component that relates to the leader being extraordinary, exemplary, and heroic (Conger & Kanungo, 1987). Therefore, we argue that followers would experience strong need satisfaction and perceive transformational leaders to be especially effective when uncertainty is high.

Several researchers have indirectly confirmed this idea. For instance, various authors reason that TFL may be regarded as a necessity in times of crisis and uncertainty (Beyer & Browning, 1999; Conger & Kanungo, 1987; Weber, 1947), and some charismatic leadership theories (Beyer & Browning, 1999; Weber, 1947) even state that a crisis is a necessary condition for (charismatic) leaders to become successful. We do not propose such a strong relationship given that at least one study failed to show a moderating influence of the environment on organizational performance for CEO charisma (Agle, Nagarajan, Sonnenfeld, & Srinivasan, 2006), but we do maintain that TFL may be especially strongly related to perception of effectiveness during uncertain times. There is also empirical research that supports this position by showing that the success of charismatic/ transformational leaders is particularly likely in times of crisis and uncertainty with regard to subjective performance (De Hoogh et al., 2004) and profitability (Waldman, Ramirez, House, & Puranam, 2001). Thus, we argue that under higher uncertainty TFL, would be more strongly related to psychological need satisfaction and attributions of effectiveness, which subsequently translate into more positive outcomes.

Hypothesis 4. Higher environmental uncertainty strengthens the positive relationship between transformational leadership and a) psychological need satisfaction and b) follower attributions of leadership effectiveness, which subsequently increases followers' organizational commitment and work engagement.

We argue that the moderation of uncertainty is very different for SL. Specifically because followers tend to focus on groups and organizations more than themselves for comfort and safety in times of uncertainty, we argue that collective, organizational needs also become more pronounced for followers, while personal needs become less salient (cf. Hogg, 2007). Thus, during more uncertain times, the need for personal care in terms of a relationship-oriented leadership style, such as SL, becomes weaker (Cohen, Solomon, Maxfield, Pyszczynski, & Greenberg, 2004). Ironically then, as a consequence of its strong focus on the individual needs of the follower, SL may satisfy follower psychological needs under uncertainty less than under stability because followers' psychological needs shift towards group-oriented needs. Furthermore, SL's orientation on altruistic calling (Barbuto & Wheeler, 2006), putting subordinates first (Liden et al., 2008), and standing back and humility (Van Dierendonck & Nuijten, 2011) is far removed from the heroic, charismatic leader who uses unconventional means to achieve his or her goals. As a consequence, we believe that the leadership effectiveness attributions of SL may also decrease under higher uncertainty.

Although little empirical research investigated this issue, Humphreys (2005) uses a historical analysis to compare the leadership styles of Xenophon and Chief Joseph and provides support for this proposal. The underlying idea is that because of the altruistic and egalitaristic focus of SLs on the needs of followers over organizational success, there is less focus on renewal and change, which are necessary for managing uncertain situations.

Hypothesis 5. Higher environmental uncertainty weakens the positive relationship between servant leadership and a) psychological need satisfaction and b) follower attributions of leadership effectiveness, which subsequently reduces followers' organizational commitment and work engagement.

Fig. 1. Conceptual model.

550 D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

3.5. Overview of the studies

These hypotheses (see Fig. 1) were tested in two experimental studies and one field study. The first experimental study manipulates leadership (SL versus TFL) and organizational uncertainty (uncertainty versus stability) and measures the mediators in our model (perceived leader effectiveness and need satisfaction) and organizational commitment. The second experimental study contributes to the design control conditions (in the form of laissez-faire and transactional leadership, see study 2) and focuses on work engagement. Finally, the third study aims to replicate the findings of the experimental research in the field to improve the generalizability of the results of this research. Additionally, rather than focusing on organizational uncertainty, this study examines job uncertainty as a moderator.

4. Study 1

This first scenario study focuses on the differentiated influence of TFL and SL on organizational commitment through need satisfaction and leadership effectiveness. To test these influences, the participants were confronted with a scenario concerning a 44-year-old leader of a company, in which we manipulated the leadership style and environmental uncertainty and measured the participants' attributions of leader effectiveness, need satisfaction, and commitment to the company.1

4.1. Methods

4.1.1. Participants and design The participants of this study were 184 people from the network of one of the authors (83 females, 85 males, the rest of

unknown gender; mean age = 37.00 years, SD = 11.36). Colleagues at work as well as family members and friends were approached. A snowballing method (i.e., asking participants to suggest others that might be willing to participate) was used to recruit the remainder of the participants. Within this convenient sample, we made sure that all participants had experience working under a leader to establish that they could imagine themselves in the scenario provided. We used a 2 (leadership: SL versus TFL) × 2 (business environment: stable versus unstable) factorial design. The participants were randomly assigned to the different conditions, with 46 participants in each cell.

The participants were asked to complete a paper-and-pencil test. The test described the experimental manipulation, and then the participants were asked to imagine that they personally experienced the situation described in the manipulation. The company, ‘Live Long Innovations’, was described as a company that produced and distributed food supplements in the participants' place of residence.

4.1.2. Experimental manipulations

4.1.2.1. Leadership. Following De Cremer (2006), a manipulation was developed for SL and TFL. For SL, the following core elements, as described by Greenleaf (1977) and Van Dierendonck (2011), were included: ‘Your supervisor knows what you personally need. Your supervisor is modest, of integrity, honest, and authentic and shares his/her thoughts and feelings with you. Your supervisor is courageous, allows for mistakes, and provides freedom so you can develop your own abilities. Your supervisor shows great humanity and understanding of your position.’ The following core elements of TFL, as described by Podsakoff et al. (1990) and Rafferty and Griffin (2004), were included: ‘Your supervisor has the capacity to create a vision. Your supervisor communicates goals, values, purpose, and the importance of the organization's mission. Your supervisor examines new perspectives for solving problems and completing tasks. Your supervisor focuses on the development and mentoring of followers.’

4.1.2.2. Organizational uncertainty. The uncertainty of the external business environment was manipulated following Choi and Mai-Dalton (1998). The following text was provided for the certain, stable environment: ‘The company offers a broad range of products and is growing each year. The company has great research facilities, which develop new popular products and are expected to continue to do so for many years to come. Most of the products are protected by patents. You have a permanent contract with the company. The recession leaves the company untouched.’ The following text was provided for the uncertain environment: ‘The company offers a narrow range of products and is downsizing each year. The company has limited research facilities, so developing new popular products is a time-consuming process. Most of the products are not protected by patents. You have a temporary contract with the company. The recession influences the company.’

1 Originally in this study, gender was also incorporated as a control condition because both SL and TFL can be expected to be experienced as more feminine than masculine given that women are expected to be more understanding, helpful, sophisticated, and aware of the feelings of others (Schein, 1973). Particularly in times of uncertainty, the gender of the leader may influence the perception of leadership effectiveness, as exemplified in the glass cliff research of Ryan and Haslam (2007), who maintain that women are more likely to be chosen for leadership positions when the circumstances are risky and uncertain. Thus, in the scenarios, we informed the participants that the leader was either a man or woman as a manipulation of gender. However, we observed no main effects of gender manipulation, no two-way interaction effects of gender manipulation with other manipulations, and no three-way interaction between gender, leadership, and uncertainty manipulations. These results suggest that any findings for leadership and uncertainty generalize over leader gender. Because gender has no effect at all, we decided not to mention gender in the description of the study other than in this footnote to avoid complicating the study. Importantly, one could wonder whether the relative age between the proposed leader and participants could influence the results of this study. However, we observed that adding the age of the respondent to the analysis as a covariate did not significantly influence the results, and age was not significantly related to the outcome variables.

551D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

4.1.3. Dependent measures Unless otherwise indicated, all items were rated on a five-point Likert scale ranging from ‘strongly disagree’ to ‘strongly agree’.

4.1.3.1. Manipulation checks. We asked three questions from Ehrhart's (2004) one-dimensional measure of SL as a manipulation check for SL: “Your supervisor makes the personal development of employees a priority”, “Your supervisor makes you feel that you work with him/her instead of for him/her”, and “Your supervisor works hard to find new ways to help others be the best they can be”. Cronbach's alpha was .88. The items were chosen based on their ability to exemplify core aspects of SL.

We asked three questions from Podsakoff et al. (1990) and Rafferty and Griffin (2004) as a manipulation check for TFL: “Your supervisor has a clear understanding of where the company is going”, “Your supervisor challenges you to think about old problems in a new way”, and “Your supervisor mentions things that make you feel proud to be a part of this company”. Cronbach's alpha was .66. The items were obtained from three sub-dimensions of TFL: vision, intellectual stimulation, and inspirational communication. All items were included in Rafferty and Griffin's (2004) scale used in study 3. The correlation between the manipulation check items and the full scale was .95, which confirmed that the items represent the core variance of TFL.

The manipulation check for business environment was adopted from Choi and Mai-Dalton (1998): “The company is in a state of crisis”, “The performance of the company has seriously declined”, and “The business environment of the company is favorable” (reverse coded). Cronbach's alpha was. 88.

4.1.3.2. Need satisfaction. Based on studies by Gagne (2003) and Mayer et al. (2008), six items reflecting the three basic psychological needs described by self-determination theory were used to assess need satisfaction, e.g., “People at work tell me that I am good at what I do”, “People at work care about me”, and “I can influence the way I do my job”. Cronbach's alpha was .66.

4.1.3.3. Perceived effectiveness. Perceived leader effectiveness was measured using four items from De Hoogh, Den Hartog, and Koopman (2005), e.g., “Your supervisor is a competent manager” and “Your supervisor shows leadership skills”. Cronbach's alpha was .92.

4.1.3.4. Organizational commitment. Organizational commitment was measured with the nine-item shortened version of the Organization Commitment Questionnaire (OCQ) developed by Mowday, Steers, and Porter (1979), e.g., “I am proud to tell others that I am part of this organization” and “I really care about the fate of this organization”. Cronbach's alpha was .87.

4.2. Results

4.2.1. Manipulation checks In the SL condition, participants scored higher on the SL scale (M = 3.86; SD = .92) than in the TFL condition (M = 3.59;

SD = .73) (F(1,182) = 6.69, p = .01). In the TFL condition, participants scored higher on the TFL scale (M = 3.68; SD = .69) than in the SL condition (M = 3.38; SD = .77) (F(1,182) = 8.20, p = .005). No other effects were found. This result suggests that the leadership manipulation was successful.

For business environment, the score on the manipulation check was higher (indicating more instability) in the uncertain condition (M = 4.29; SD = .69) than in the certain, stable condition (M = 2.22; SD = .81) (F(1,182) = 348.63, p b .0001). No other effects were found. We can conclude that the manipulation for business environment was also successful.

4.2.2. Ratings of leadership effectiveness, follower need satisfaction and organizational commitment A multiple analysis of variance was used with the leadership and business environment manipulations and their interaction as

independent variables (see Table 1). Leadership effectiveness, organizational commitment and participant need satisfaction were the dependent variables. The homogeneity of variances was confirmed with the Levene test: F(3,180) = 1.268, p = .287 for leadership effectiveness, F(3,180) = 1.048, p = .373 for organizational commitment, and F(3,180) = .356, p = .785 for need satisfaction. The two-way interactions were not significant. The multivariate main effects were significant for leadership manipulation (F(3,178) = 7.78, p b .001) and business environment manipulation (F(3,178) = 9.59, p b .001).

Organizational commitment was not significantly different between the leadership conditions, which was expected (F(1,180) = 1.79, p = .183). Leadership effectiveness was rated higher in the TFL condition (M = 3.67; SD = .78) than in the SL condition (M = 3.35; SD = .94) (F(1,180) = 6.74, p = .01), thus supporting Hypothesis 1. Participants rated their need satisfaction as higher in the SL condition (M = 3.44; SD = .52) than in the TFL condition (M = 3.18; SD = .53) (F(1,180) = 11.84, p = .001), thus supporting Hypothesis 2.

4.2.3. Mediation analysis The PROCESS tool developed by Hayes (2012) was used to detect the influence of the experimental conditions on

organizational commitment mediated by need satisfaction and leadership effectiveness. PROCESS is a computational tool that allows for the direct testing of both mediations with SPSS. Bootstrapping is used to test the indirect effects. Although there are different methods for testing mediating effects, bootstrapping is presently regarded as one of the most reliable (Preacher & Hayes, 2008). Bootstrapping is a resampling strategy for calculating an estimate of the population coefficient using multiple resamples of the sample data. In addition, bootstrapping provides confidence intervals around the estimated coefficient.

552 D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

The full model explained 30% of the variance of organizational commitment (F(3,180) = 25.34, p b .001). The leadership manipulation had a significant positive effect (.26 (SE = .08), p b .001) on need satisfaction (note that SL was coded 1 and TFL 0), and the effect of need satisfaction on organizational commitment was also positive and significant (.50 (SE = .08), p b .001). Moreover, the indirect effect of leadership manipulation on organizational commitment through need satisfaction was significant (effect size = .13 (SE = .04), 95% confidence interval: .05–.22). The leadership manipulation had a significant negative effect (−.32 (SE = .12), p = .013) on the perceptions of leadership effectiveness (note that SL was coded 1 and TFL 0), and the effect of perceptions of leadership effectiveness on organizational commitment was positive and significant (.18 (SE = .19), p b .001). Furthermore, the indirect effect of leadership manipulation on organizational commitment through leadership effectiveness was significantly related (effect size = −.06 (SE = .03), 95% confidence interval: −.01 to −.13). This result confirms the hypothesized mediating roles of psychological need satisfaction for SL and leadership effectiveness for TFL (Hypotheses 3a and 3b).

4.2.4. Moderation role of business environment The certain, stable business environment resulted in higher scores (M = 3.75; SD = .82) for leadership effectiveness than the

uncertain condition (M = 3.27; SD = .87) (F(1,180) = 15.43, p b .001). The stable business environment also resulted in higher scores (M = 3.41; SD = .56) for participants' estimations of their need satisfaction than the unstable condition (M = 3.22; SD = .51) (F(1,180) = 6.20, p = .01) and for the participants' estimation of their organizational commitment (M = 3.66, SD = .59 versus M = 3.45, SD = .63; F(1,180) = 21.99, p b .001). However, because no significant interaction effects were found, neither Hypothesis 4a nor 4b are supported.

4.3. Conclusion

The results of this initial study partially confirm our hypotheses. It appears that leaders who show TFL are perceived as more effective. Simultaneously, leaders who show SL are observed as being better at fulfilling the needs of their followers. Importantly, the perceptions of both leadership effectiveness and need satisfaction are strong predictors of organizational commitment. With regard to the environmental influence, we observed main effects on leadership effectiveness and need satisfaction but no interaction effects. This result indicates that the environment influences people's perceptions, but this influence is similar, in this particular sample, for SL and TFL. One reason for the absence of an effect of environmental uncertainty could be the nature of our sample, which was a snowball sample of relatively diverse individuals. Therefore, we aimed to replicate our findings in a more homogeneous sample in the next study (employees of one organization).

5. Study 2

Although the design of study 1 enabled us to compare the effects of TFL to SL, it did not allow us to assess the effects of TFL and SL independently because our design manipulated TFL versus SL but contained no other conditions. This is important, as the theory discussed earlier states that both TFL and SL, unlike other leadership styles, may affect perceptions of leadership effectiveness and need satisfaction (although TFL affects the former more strongly and the latter less strongly than SL). Therefore, in a second study, we used a design that compared SL and TFL to not only one another but also other leadership styles.

To find suitable leadership styles for comparison, we consulted Bass's (1985) full range of leadership model. This model emphasizes two styles in comparison to TFL (and SL): transactional leadership and laissez-faire leadership. Transactional leadership is similar to what DeRue and colleagues (2011) refer to as “task-oriented leadership”. Transactional leadership focuses primarily on rewarding and disciplining followers (Bass, 2005) and is reasonably effective in motivating followers to higher levels of in-role task performance. Transactional leadership motivates followers for in-role task performance extrinsically (through exchanges) and through clarifying work and work goals. As a consequence, transactional leadership is less likely to affect extra-role behaviors because it does not emphasize the building of relationships with the job and organization, such as affective commitment and job engagement (Bass, 2005). Therefore, we predict that transactional leadership, compared to TFL and SL, would be less strongly related to followers' perceptions of leadership effectiveness, need satisfaction, and work engagement.

Laissez-faire leadership is, according to Bass (2005), the most passive and ineffective type of leadership, which basically indicates the absence or avoidance of leadership. This implies that laissez-faire leaders do not make the necessary decisions, delay actions, and ignore responsibilities. These leaders leave their followers without guidance or correction. This type of leadership is similar to what DeRue and colleagues (2011) refer to as “passive leadership”. Research has observed many negative relationships between laissez-faire leadership with various outcomes (DeRue et al., 2011). We predict, therefore, that laissez-faire leadership, compared to TFL and SL and transactional leadership, would be less strongly related to followers' perceptions of leadership effectiveness, need satisfaction, and work engagement.

In summary, by comparing SL and TFL with one another and with laissez-faire and transactional leadership, we are able to assess the effects of SL and TFL relative to other leadership styles. This method creates a more complete and rigid replication test of our hypotheses, which is the aim of study 2. We focused on work engagement (Schaufeli et al., 2006) as the outcome variable in this study.

553D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

5.1. Methods

5.1.1. Participants and design The participants in this study were 200 employees (mostly nurses and doctors) working in a hospital (149 females, 51 males;

mean age 39.6 years old, SD = 11.5). All participants in this sample had experience working under a leader. We used a 4 (leadership: SL versus TFL versus transactional leadership versus laissez-faire leadership) × 2 (business environment: stable versus unstable) experimental design. The participants were randomly assigned to the different conditions with 25 participants per cell.

5.1.2. Procedure The participants were asked to complete a paper-and-pencil test similar to the one used in study 1. The test first described the

experimental manipulation. The participants were asked to imagine that they had personally experienced the described situation. ‘Stucco Healthcare’ was described as a company that produces and distributes medicine in the participants' place of residence.

5.1.3. Experimental manipulations

5.1.3.1. Leadership. The manipulations for SL and TFL were identical to those in study 1. For the transactional leader, the core elements (based on descriptions from Bass, 2005) were described as “Your supervisor clarifies your responsibilities to you, monitors your performance, and takes corrective actions if required. Your supervisor makes sure you meet certain standards despite your mistakes and failures. Your supervisor rewards you if your performance is satisfactory.”

The following text was provided (based on descriptions from Bass, 2005) for laissez-faire leadership: “Your supervisor takes no managing responsibility. Your supervisor is frequently absent and exhibits an absence of involvement during critical junctures. Your supervisor waits until problems become severe before attending to them and intervening.”

Similar to study 1, we again told participants that the leader was 44 years old.2

5.1.3.2. Business environment. The external business environment was manipulated in an identical manner to study 1.

5.1.4. Dependent measures

5.1.4.1. Manipulation checks. The manipulation checks for SL, TFL, and business environment were identical to those in study 1, and the Cronbach's alpha values were .86, .85, and .83, respectively.

Three questions from Podsakoff et al. (1990) were used as a manipulation check for transactional leadership: “Your supervisor always gives you feedback when you perform well”, “Your supervisor commends you when you do a better than average job”, and “Your supervisor acknowledges when you do a good job”. Cronbach's alpha was .88.

Three questions were formulated based on the research of Skogstad, Einarsen, Torsheim, Aasland, and Hetland (2007) as a manipulation check for laissez-faire leadership: “Your supervisor is not concerned with results”, “Your supervisor is absent when you need him or her”, and “Your supervisor postpones answering your questions”. Cronbach's alpha was .79.

5.1.4.2. Psychological need satisfaction. Items identical to those in study 1 were used. Cronbach's alpha was .68.

5.1.4.3. Perceived leadership effectiveness. Items identical to those in study 1 were used. Cronbach's alpha was .96.

5.1.4.4. Work engagement. The short nine-item version of the Utrecht Work Engagement Scale (UWES) was used to measure engagement (Schaufeli et al., 2006). The items were adjusted slightly to ensure that the focus of the answers would be directed at the situation at Stucco Healthcare, e.g., “At my work at Stucco Healthcare, I am very energetic” or “My job inspires me”. Cronbach's alpha was .91.

5.2. Results

5.2.1. Manipulation checks The highest average ratings on the SL scale occurred in the SL condition (M = 3.88; SD = .58). The overall test was significant

(F(3, 196) = 71.48, p b .001), as were the individual tests in comparison to the TFL condition (M = 3.59; SD = .81) (F(1,98) = 4.26, p = .04), transactional leadership condition (M = 3.00; SD = .80) (F(1,98) = 39.72, p b .001), and laissez-faire leadership condition (M = 1.85; SD = .80) (F(1,98) = 211.95, p b .001). These results are consistent with our intended manipulation.

The highest average ratings on the transformational scale were occurred in the TFL condition (M = 3.90; SD = .85). The overall test was significant (F(3, 196) = 55.00, p b .001), as were the individual tests in comparison to the SL condition (M = 3.32; SD = .84) (F(1,98) = 11.88, p = .001), transactional leadership condition (M = 3.01; SD = .85) (F(1,98) = 24.35, p b .001), and laissez-faire leadership condition (M = 1.91; SD = .64) (F(1,98) = 175.73, p b .001). These results are consistent with our intended manipulation.

2 Similar to study 1, we checked whether the age of the respondent influenced any of our results. Adding age as covariance did not significantly influence the results. Furthermore, age had no effect on the outcome variables.

554 D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

The highest average ratings on the transactional scale occurred in the transactional leadership condition (M = 3.76; SD = .81). The overall test was significant (F(3, 196) = 42.67, p b .001). However, the individual comparison tests showed that transactional leadership was rated approximately identical to the SL condition (M = 3.74; SD = .63; F(1,98) = .03, p = .87) and was slightly, but not significantly, lower in the TFL condition (M = 3.48; SD = .75) (F(1,98) = 3.21, p = .08). The difference with the laissez-faire condition was significant (M = 2.31; SD = .78; F(1,98) = 83.80, p b .001). These results are largely consistent with our intended manipulation. Specifically, we discussed that transactional leadership showed an overlap with TFL and SL, but transactional leadership missed those elements that made TFL and SL successful. Their overlap on this dimension, but not on measures of SL and TFL, confirms this idea.

The highest average ratings on the laissez-faire scale occurred in the laissez-faire leadership condition (M = 4.05; SD = .79). The overall test was significant (F(3, 196) = 76.30, p b .001), as were the individual tests in comparison to the SL condition (M = 2.12; SD = .73) (F(1,98) = 161.79, p b .001), TFL condition (M = 2.22; SD = .80) (F(1,98) = 133.45, p b .001), and transactional leadership condition (M = 2.19; SD = .72) (F(1,98) = 152.54, p b .001). These results are consistent with our intended manipulation.

Considering these findings and the fact that we did not find any other effects, we conclude that, overall, the manipulation for leadership was successful.

For business environment, the score was higher (indicating more instability) in the unstable condition (M = 3.77; SD = .78) than in the stable condition (M = 1.74; SD = .71) (F(1,198) = 369.93, p b .000). No other effects were found. Thus, the manipulation for business environment was also successful.

5.2.2. Ratings of leadership effectiveness, follower need satisfaction and work engagement Multiple analysis of variance was used with the leadership manipulation, the business environment manipulation, and their

interaction as independent variables (see Table 2). Leadership effectiveness, need satisfaction, and work engagement were the dependent variables. The homogeneity of the variances was confirmed with the Levene test: F(7,192) = 1.213, p = .297 for leadership effectiveness; F(7,192) = .654, p = .710 for work engagement; F(7,192) = .773, p = .611 for psychological need satisfaction. The multivariate effect is significant for the main effect of leadership (F(9,576) = 17.205, p b .001), the main effect of uncertainty (F(3,190) = 10.157, p b .001), and the interaction between them (F(9,576) = 2.382, p = .012). Below, we describe the effects for each dependent variable separately.

5.2.2.1. Work engagement. The stable business environment provided higher scores (M = 3.56; SD = .66) for participants' estimations of work engagement than the unstable condition (M = 3.03; SD = .77) (F(1,192) = 29.844, p b .001). Work engagement was rated similarly in the SL (M = 3.50; SD = .73) and TFL conditions (M = 3.53; SD = .66), slightly lower (but not significant at the p b .05 level) in the transactional leadership condition (M = 3.26; SD = .76), and significantly lower in the laissez-faire leadership condition (M = 2.89; SD = .74) (F(3,192) = 4.427, p b .001). We observed no significant overall interaction effect for work engagement.

5.2.2.2. Perceptions of leader effectiveness. The above-mentioned multivariate two-way interaction effect of both conditions was significant (F(3,192) = 2.382, p = .012). This effect was qualified by a significant interaction on leadership effectiveness (F(3,190) = 3.72, p = .012). Fig. 2 depicts this interaction. The figure shows that leadership effectiveness is lowest for laissez-faire leadership (the overall main effect for leadership: F(3,192) = 67.51, p b .001) and that leadership effectiveness was highest in the TFL condition. These results support Hypothesis 1. There was also an overall main effect for uncertainty (F(3,192) = 8.925, p b .001), which signified that leadership effectiveness was perceived as lowest when people perceived themselves as being in uncertain circumstances. There was a significant difference in leadership effectiveness between the unstable and stable conditions for

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Laissez-fair Transactional Transformational Servant

Uncertain

Certain

Fig. 2. Interaction of leadership with environmental stability on leadership effectiveness (study 2).

555D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

transactional leadership (F(1,48) = 8.43, p = .006) and for SL (F(1,48) = 18.28, p b .001), whereas there was no significant difference between either condition for TFL (F(1,48) = 1.64, p = .207). This result shows that in this study, TFL is not perceived as more effective in uncertain situations than stable situations; therefore, no support for Hypothesis 4a was found. SL was perceived as more effective in stable (versus unstable) circumstances, which lends partial support for Hypothesis 4b.

5.2.2.3. Need satisfaction. The stable business environment provided higher scores (M = 3.25; SD = .58) for need satisfaction than the unstable condition (M = 3.05; SD = .57) (F(1,192) = 7.65, p = .006). Need satisfaction was rated highest in the SL condition (M = 3.48; SD = .47), followed by the TFL condition (M = 3.27; SD = .56), the transactional leadership condition (M = 3.03; SD = .61), and the laissez-faire leadership condition (M = 2.81; SD = .48) (F(3,192) = 15.86, p b .001). Post-hoc tests showed that the differences between all conditions were significant. These results support Hypothesis 2. The interaction effects were not significant, in contrast to Hypothesis 4b.

5.2.2.4. Mediation analysis. Finally, we tested the influence of SL and TFL against the laissez-faire condition (the control condition) with PROCESS (Hayes, 2012). For this test, only those respondents who were assigned to the appropriate conditions were included. The indirect effect of SL on work engagement was significant for need satisfaction (effect size = .38 (SE = .12), 95% confidence interval: .19–.63) and leadership effectiveness (effect size = .72 (SE = .17), 95% confidence interval: .43–1.08). The indirect effect of TFL on work engagement was significant for need satisfaction (effect size = .30 (SE = .09), 95% confidence interval: .15–.49) and leadership effectiveness (effect size = .48 (SE = .13), 95% confidence interval: .24–.76). This confirms our assumption that both leadership styles are independently related to work engagement.

However, to directly investigate the differential mediation effects of TFL and SL, we also conducted a mediation analysis similar to Study 1, testing the mediating role of need satisfaction and leadership effectiveness between SL and TFL and work engagement with PROCESS (Hayes, 2012). For this test, only those respondents that answered these two conditions were included. The full mediating model explained 46% of the variance in work engagement (F(3, 96) = 26.83, p b .001). The indirect effect through need satisfaction was positive (effect size = .09 (se = .05), 95% confidence interval .01 and .22). The indirect effect through leadership effectiveness is negative (effect size = −.07 (se = .07), 95% confidence interval .03 and −.27). This result confirms the hypothesized mediating role of psychological need satisfaction for SL and for that of leadership effectiveness for TFL (although it should be noted that, whereas the effect size is in the hypothesized direction and similar to study 1, the effect is only significant in a one-sided test).

5.3. Conclusion

The results of this study show that both SL and TFL are more influential overall than laissez-faire leadership for perceived leadership effectiveness, psychological need satisfaction, and work engagement. Similar to study 1, it appears that the effects of SL on work engagement are mediated by need satisfaction, whereas the effects of TFL are mediated by perceived leadership effectiveness. In contrast to study 1, we observed a significant interaction of leadership manipulation and organizational uncertainty. However, this interaction effect was only observed for perceptions of leadership effectiveness, not for need satisfaction. Furthermore, the results seemed to support the idea that the effect of SL is more pronounced in stable times than in uncertain times, while no difference was apparent for TFL.

6. Study 3

We next set out to ratify the main findings of both experimental studies with a field study. Combining the hypotheses with these findings, we presupposed that SL and TFL would be related to work engagement and organizational commitment through psychological need satisfaction and leadership effectiveness. Presently, we have only observed limited effects of (organizational) uncertainty (mostly main effects). Indeed, although it seems that SL is more effective in stable than uncertain environments, we have observed no evidence that SL fares worse than TFL in uncertain environments. Thus, our argument that, in uncertain environments, followers prefer leaders that they deem especially effective (i.e., TFL) over leaders that are especially good at satisfying their needs (i.e., SL) may not be completely valid. One reason for this result may be that environmental uncertainty is often related to job uncertainty (cf. Sverke & Hellgren, 2002). This is important because research shows that social support (i.e., the satisfaction of social needs) interacts with job uncertainty such that uncertainty is not negatively related to outcomes if social support is high (Sverke & Hellgren, 2002; Sverke, Hellgren, & Näswall, 2002). For the current study, in which we have argued that SL is more strongly related to the need satisfaction of employees, this would mean that under conditions of high job uncertainty, SL could be fairly effective relative to TFL. This would be counter to our argument for environmental uncertainty, leading to mixed findings. To investigate this issue in this study, we assess the moderating influence of job uncertainty on the effects of SL and TFL.

Using structural equation modeling, a model was tested in which SL and TFL were related to psychological need satisfaction and leadership effectiveness. Psychological need satisfaction and leadership effectiveness were related to work engagement and organizational commitment. Job uncertainty was tested as a moderator in the relationships between SL and TFL on the one hand and psychological need satisfaction and leadership effectiveness on the other hand.

556 D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

6.1. Methods

6.1.1. Participants and design The participants of this study were 200 individuals (84 females, 116 males; mean age of 42.18 years, SD = 9.59) employed as

support staff at a major university. All subjects voluntarily participated in a cross-sectional self-report survey.

6.1.2. Measures

6.1.2.1. SL. SL was measured with Ehrhart's (2004) 14-item measure of SL. Example items are “Your supervisor makes the personal development of employees a priority”, “Your supervisor makes you feel that you work with him/her instead of for him/her”, and “Your supervisor works hard to find new ways to help others be the best they can be”. Cronbach's alpha was .93. In studies 1 and 2, three items of Ehrhart's (2004) scale were used for the manipulation check of servant leadership. As an additional check of the validity of our choice of items, the average score of these same three items was correlated with the full scale used in this study. The resulting correlation was .91, which confirmed that the items chosen in studies 1 and 2 represent the core variance of SL as measured by Ehrhart's scale.

6.1.2.2. TFL. To assess TFL, the 15 items of Rafferty and Griffin (2004) were used. Example items are “Your supervisor challenges you to think about old problems in a new way” and “Your supervisor mentions things that make you feel proud to be a part of this company”. The full scale is used as an indication of TFL. Cronbach's alpha was .96.

6.1.2.3. Job uncertainty. Job uncertainty was measured with the four-item instrument of Mauno, Leskinen, and Kinnunen (2001). Example items are “I am worried about the possibility of being fired” and “I am certain my job will continue for a long time” (reverse coded). Cronbach's alpha was .86.

6.1.2.4. Need satisfaction. Need satisfaction was measured with nine items from Sheldon, Elliot, Kim, and Kasser (2001), which reflect the three basic psychological needs of self-determination theory. Sample items are “At work, I am successful at completing difficult tasks and projects”, “At work, I have a sense of contact with people who care for me, and whom I care for”, and “At work, I am free to do things my own way”. We chose a different scale from studies 1 and 2 to improve reliability. The content of the items was similar to those in studies 1 and 2. Cronbach's alpha was .89, thus supporting our decision.

6.1.2.5. Perceived leadership effectiveness. This scale was measured with items identical to those used in studies 1 and 2. Cronbach's alpha was .94.

6.1.2.6. Work engagement. The short nine-item scale of the UWES (Schaufeli et al., 2006) was used. See also study 2. Cronbach's alpha was .94.

6.1.2.7. Organizational commitment. The eight-item affective organizational commitment subscale from Allen and Meyer's (1990) instrument was used. Cronbach's alpha was .91.

6.2. Analysis

The hypotheses and findings formulated in the previous studies can be combined into an overall model, which allows us to test all of the variables together with structural equation modeling. This structural equation model consists of latent and manifest variables. The so-called measurement model specifies how the unobserved latent variables are measured in terms of the observed manifest variables. More specifically, the measurement model estimates the strength of the relationship between the corresponding latent and manifest variables. To be mathematically identified, each latent variable must be estimated by at least two manifest variables. Dividing items into parcels is the recommended practice if the goal is to study a variable at an overall level of generality, focus on the relations among the latent variables and reduce the level of nuisance and bias that may result from working with the separate items directly (Bandalos, 2002; Little, Cunningham, Shahar, & Wideman, 2002). The advantage of parcel-based modeling is that the measurement models are more parsimonious with fewer residuals that must be correlated and less possible dual loadings (Little et al., 2002). We mainly used three parcels, two if necessary, to allow for a similar number of items in each parcel within a scale. Namely, scales with odd-numbered items were divided into three parcels and scales with even-numbered scales into two parcels.

For TFL, psychological need satisfaction and work engagement, the items were divided among three parcels. Each of these measures consists of sub-dimensions. Therefore, a domain-specific approach was used, dividing the items from one sub-dimension over different parcels. The latent constructs of SL, job uncertainty, leadership effectiveness and organizational commitment were operationalized within the measurement model by evenly dividing the items of the scales (the so-called manifest variables) between two parcels. These measures are one-dimensional. Given our primary interest in the underlying latent construct and not in the relations among the items, items with high intercorrelations were divided between both parcels when possible.

557D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

6.3. Results

Table 3 shows the mean values, standard deviations and intercorrelations of the variables included in study 3. Structural equation modeling was performed using Mplus 6 (Muthén & Muthén, 2010). Following Anderson and Gerbing (1988), we tested the adequacy of the measurement model before actually testing the relations in the latent variable model. In a sense, this is a confirmatory factor analysis of the model with all variables included which tests if the latent variables are indeed separate constructs. This is an important first step given the cross-sectional same source nature of the data. The procedure tests whether the measurement model is misspecified, in which case a fitting model cannot be observed. The full measurement model was tested, in which all latent variables were allowed to correlate. The fit was acceptable with relative fit indices of .96 for CFI, .95 for TLI and .04 for SRMR (χ2 = 234.38, df = 98, p b .000), which are generally observed as indicators for an adequate fit (Hu & Bentler, 1999) and allowed us to proceed with testing the hypothesized structural model.

In the hypothesized model, SL and TFL were allowed to intercorrelate, as were work engagement and organizational commitment. Psychological need satisfaction and leadership effectiveness were placed as mediators. Because we are interested in mediation, we ultimately want to conduct (bootstrapping) mediation analysis. However, within Mplus, bootstrapping can only be used for a latent variable model for testing possible mediators, not for interactions simultaneously. Therefore, we initially tested the mediation model without job uncertainty in the model. To provide a good insight into the differential influence of SL and TFL, first the model was tested for SL and TFL separately. This was followed by a test of the full mediation model including both leadership styles.

The model whereby SL was related to work engagement and organizational commitment through psychological needs and leadership effectiveness gave an excellent fit to the data (χ2 = 103.216, df = 47, p b .000: CFI = .97, TLI = .97, SRMR = .04). All relationships were significant except for organizational commitment on psychological need satisfaction (β = .14, p = .14). The standardized coefficients of SL with leadership effectiveness and psychological need satisfaction were .60 (p b .001) and.64 (p b .001), respectively.

The model whereby TFL was related to work engagement and organizational commitment through psychological needs and leadership effectiveness gave an excellent fit to the data as well (χ2 = 145.151, df = 58, p b .000: CFI = .97, TLI = .96, SRMR = .05). All relationships were significant. Most noteworthy for our hypothesized process were the standardized coefficients between TFL and leadership effectiveness (.85 (p b .001)) and psychological need satisfaction (.32 (p b .001)).

The full mediating model including both SL and TFL also fit the data well (χ2 = 196.106, df = 80, p b .000: CFI = .97, TLI = .96, SRMR = .04). Two main effects were not significant and therefore removed: TFL on need satisfaction and need satisfaction on organizational commitment. The resulting fit was good (χ2 = 200.66, df = 82, p b .000: CFI = .97, TLI = .96, SRMR = .06).

We then tested the mediating role of psychological need satisfaction and leadership effectiveness with bootstrapping. The standardized estimated indirect coefficient of SL to work engagement through psychological need satisfaction was .43 (p = .001; 95% confidence interval: .33–.59), and the indirect effect through perceived leadership effectiveness was .03 (p = .05; 95% confidence interval: −.01–.07). The standardized estimated indirect coefficient of SL to organizational commitment through perceived leadership effectiveness was .07 (p = .01; 95% confidence interval: −.01–.15). Because the relationship between need satisfaction and affective commitment is non-significant, we did not test this mediation path. These results support the idea of the mediating role of psychological need satisfaction and leadership effectiveness between SL and work engagement. The mediating role of leadership effectiveness is more limited for SL; although the indirect effect sizes were significant for work engagement and organizational commitment, the confidence interval included zero in both cases.

The standardized estimated indirect coefficient of TFL via perceived leadership effectiveness was .12 (p = .05; 95% confidence interval: .04–.23) for work engagement and .28 for organizational commitment (p b .001; 95% confidence interval: .18–.42). Because the relationship between TFL and need satisfaction is non-significant, we did not test the mediation path of TFL through need satisfaction. These results suggest that there is a mediating role of leadership effectiveness for TFL.

Subsequently, job uncertainty was placed as a moderator between SL and TFL on one side and need satisfaction and leadership effectiveness on the other. Testing the interaction with latent variables is allowed within MPlus; however, it should be noted that

Table 1 Means and standard deviations of the experimental conditions, study 1.

Overall Business environment

Stable Unstable

M SD M SD M SD

Organizational commitment Servant leadership 3.51 .67 3.70 .61 3.32 .68 Transformational leadership 3.39 .58 3.62 .57 3.17 .52 Total 3.45 .63 3.66 .59 3.25 .60

Leadership effectiveness Servant leadership 3.35 .94 3.63 .88 3.07 .91 Transformational leadership 3.67 .78 3.87 .75 3.46 .76 Total 3.51 .87 3.75 .82 3.27 .86

Need satisfaction Servant leadership 3.44 .52 3.52 .57 3.37 .46 Transformational leadership 3.18 .53 3.30 .52 3.06 .52 Total 3.31 .54 3.41 .56 3.22 .56

Note. N = 184.

Table 2 Means and standard deviations of the experimental conditions, study 2.

Overall Business environment

Stable Unstable

M SD M SD M SD

Work engagement Servant leadership 3.50 .73 3.86 .48 3.14 .77 Transformational leadership 3.53 .67 3.63 .67 3.44 .67 Transactional leadership 3.26 .76 3.60 .68 2.91 .68 Laissez-faire leadership 2.89 .74 3.13 .63 2.64 .77 Total 3.30 .77 3.56 .67 3.30 .77

Leadership effectiveness Servant leadership 3.41 .85 3.85 .69 2.97 .77 Transformational leadership 3.60 .97 3.77 .77 3.42 1.13 Transactional leadership 3.39 .84 3.07 .79 3.71 .77 Laissez-faire leadership 1.53 .85 1.44 .84 1.62 .86 Total 2.98 1.21 3.19 1.27 2.77 1.12

Need satisfaction Servant leadership 3.48 .47 3.55 .43 3.42 .50 Transformational leadership 3.28 .56 3.33 .63 3.21 .49 Transactional leadership 3.03 .61 3.27 .54 2.78 .59 Laissez-faire leadership 2.81 .48 2.85 .52 2.77 .43 Total 3.15 .59 3.25 .58 3.05 .57

Note. N = 200.

558 D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

the program gives limited fit indices and no standardized solution. We report the Akaike information criterion (AIC) and check the model for insignificant relations. Although this criterion cannot be used to test the absolute fit of a model, it can provide an indication of the relative fit of different models, with a lower score indicating a better fit.

The model including all four interactions resulted in an AIC fit of 7793.279. Only one of the interactions was significant, namely, TFL with job uncertainty on leadership effectiveness (the same interaction that was significant in study 2). To find the best fitting model, we removed the interactions that were not significant from the model. The resulting model had an AIC fit of 7727.377, and all remaining paths were significant (p b .05). The full model is depicted in Fig. 3. The residuals of the manifest model are not depicted for clarity. All coefficients are unstandardized.

Finally, we tested the mediated moderations depicted in Fig. 3 with SPSS and Hayes (2012) process syntax and bootstrapping to test the indirect effect. For organizational commitment, the full model was significant, with an R2 of 14% (F(2,197) = 16.32, p b .001). The main effects showed a significant positive relationship for leadership effectiveness (.35 (SE = .11), p = .0012) and no direct relationship for TFL (.03, (SE = .12), p = .8193). The indirect effects were significant for all three levels (mean ± SD) of job uncertainty (effect size for plus one SD = .38, confidence interval: .15–.63; for mean = .31, confidence interval: .13–.53; for minus one SD = .25, confidence interval: .23–.82). None of the slopes included zero, thus signifying an indirect effect in which the influence of TFL decreases with the increase of job uncertainty.

For work engagement, the full model was significant, with an R2 of 15% (F(2,197) = 17.28, p b .001). The main effects showed a positive significant relationship for leadership effectiveness (.30 (SE = .09), p = .0008) and no direct relationship for TFL (.02, (SE = .11), p = .8293). The indirect effects were significant for all three levels (mean and one plus/minus SD) of job uncertainty (effect size for minus one SD = .33, confidence interval: .07–.58; for mean = .27, confidence interval: .05–.48; for plus one SD = .22, confidence interval: .04–.40). Similar to organizational commitment, none of the slopes included zero for work engagement, thus signifying an indirect effect in which the influence of TFL through leadership effectiveness decreases with increasing job uncertainty.

6.4. Conclusion

This study provides a field replication of the experimental findings regarding the mediation effects of perceived leaders' effectiveness and need satisfaction for TFL and SL effects, respectively. Specifically, TFL is related most strongly to perceived

Table 3 Intercorrelations among variables in Study 3.

M SD 1 2 3 4 5

1. Servant leadership 4.38 1.09 2. Transformational leadership 4.46 1.27 .50⁎

3. Psychological need satisfaction 4.73 1.00 .61⁎ .29⁎

4. Leadership effectiveness 4.81 1.38 .59⁎ .80⁎ .38⁎

5. Work engagement 4.80 1.14 .49⁎ .32⁎ .74⁎ .39⁎

6. Organizational commitment 4.64 1.35 .25⁎ .31⁎ .29⁎ .38⁎ .55⁎

7. Job uncertainty 3.33 1.39 −.01 −.21⁎ .03 −.11 .09 .02

Note. N = 200. ⁎ p b .01.

Fig. 3. Latent variables empirical model, study 3 (unstandardized loadings are depicted; all paths are significant, p b .05).

559D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

leadership effectiveness, whereas SL is related most strongly to need satisfaction. Furthermore, the results show that although both leadership styles are related to work engagement and organizational commitment, the paths through which they work are different.

Notably, one result is not replicated. In study 1, we observed no moderating influence of environmental uncertainty, and in study 2, we observed a positive moderation of environmental uncertainty on the relationship between TFL and perceived leader effectiveness. This field study shows a moderation of job uncertainty on this relationship (TFL to perceived leader effectiveness) but in the opposite direction. We conclude that our findings regarding the moderation of environmental uncertainty are uncertain at best.

7. General discussion

The aim of this paper is to provide insights into the different mechanisms through which SL and TFL affect followers. Accounting for the results of the three studies, we can conclude that both styles are related to organizational commitment and work engagement, but the manner of their influence differs. SL works more through follower need satisfaction, whereas TFL works more through perceived leadership effectiveness. These findings have several important implications.

7.1. Theoretical implications

An initial contribution of this paper is that it provides insight into the crucial theoretical differential mechanisms underlying the effects of SL and TFL. The data presented here confirm their strong similarities and differences: namely, TFL behavior is more strongly related to being perceived/regarded as a leader, whereas SL behavior is more strongly related to an expected enhancement of the psychological needs of followers. Therefore, these results are the first empirical test of the ideas of Parolini, Patterson, and Winston (2009) and Van Dierendonck (2011), who suggested that relative to TFL, which is more focused on the organization, SL has a stronger focus on the individual. Our paper also confirms the enhancement of psychological needs as a crucial outcome of SL. Similar to earlier research (Mayer et al., 2008), this study shows that the enhancement of psychological needs is a central element of SL theory (Mayer, 2010) and one that differentiates SL from TFL. This research compares the fundamental theoretical bases of SL and TFL and finds crucial evidence for the notion that these bases are different.

The evidence that different processes underlie SL and TFL is important for several reasons. First, without evidence that SL is not equal to TFL, it cannot be seriously studied within the field of leadership. Specifically, TFL has been studied for decades, and many studies have confirmed its potency. However, SL is a newer area of study, and only a handful of empirical studies have been conducted to test the effects of SL. Considering the conceptual overlap with TFL, one strong critique of this stream of research is that it investigates what is previously known because SL is not substantially different from TFL. Prior research has tested whether measures of TFL and SL could be statistically distinguished, but such findings are not a sufficient basis to conclude that SL and TFL can be conceptually distinguished. Our study provides much stronger support of this conceptual differentiation, suggesting that there is indeed a reason to continue research on SL. In particular, servant leadership research investigating the interpersonal personal aspect of leadership in depth clearly has value without necessarily including transformational leadership as well.

Second, understanding the differential pathways through which SL and TFL work is also important for future theorizing concerning these concepts. It may be especially conducive to finding moderators of the effects of TFL and SL. Specifically, it would seem that those factors that diminish or increase the role of leadership effectiveness perceptions would be especially important for the effectiveness of TFL. For instance, one could imagine situations in which employees need leaders only for their job-specific skills, which may reduce the importance of perceptions of leadership skills and effectiveness. At the same time, those factors that reduce or enhance the importance of psychological need satisfaction may be especially important for the effectiveness of SL. For

560 D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

instance, one could imagine situations in which psychological needs are largely addressed by the company, making the role of SL less important.

Third, and equally importantly, this study shows that perceptions of leadership effectiveness are only a portion of the equation and that need satisfaction is as important for fostering commitment and engagement among followers. This is particularly important because decades of dominance in leadership by TFL have led to the view in which the collective is emphasized over the individual. Our results show that this may not be conducive for optimal results. Because our research also shows that SL explains additional variance of leadership effectiveness in addition to TFL, respect for and a focus on the individual may be particularly important for leaders.

It is important to note here that we choose to investigate perceived leader effectiveness and need satisfaction as mediators rather than other potential mediators, such as organizational identification or self-efficacy (Shamir et al., 1993) because these two mediators emphasize the differential effects of SL and TFL. However, other mediators may also be important for understanding the relationships between SL, TFL, and their outcomes. Future research could emphasize the roles of other mediators for SL and TFL. For example, Van Dierendonck (2011) places the psychological climate in terms of fairness and trust as mediators between SL and follower outcomes. Here again, one could speculate on the potential additional value a combination of TFL and SL could have to more fully understand the impact of leadership. Our results can also be interpreted within the recent theoretical work by Sun (2013), who proposed what he calls servant-compartmentalized leaders who are able to balance the soft and hard aspects of organizational contexts. Sun (2013) explicitly presupposes that such servant leaders are able to make difficult decisions in situations of environmental uncertainty.

Finally, our paper also emphasizes the influence of an uncertain environment on the effectiveness of leadership. In our study, the results do not support our predictions that TFL is particularly important in times of uncertainty (Pawar & Eastman, 1997) and SL is particularly important in times of stability, as had been previously theorized (Graham, 1991, 1995; Smith et al., 2004). Rather, it appears that the most important influence is the main effect of uncertainty itself. Uncertainty leads to reduced employee well-being and the perception that the leader is less effective, which is in line with an earlier study by Pillai and Meindl (1998). To counterbalance this effect, what leaders do does not matter as long as they do something. Laissez-faire leadership in times of uncertainty is the least effective behavior a leader can display. Showing some type of leadership, whether transactional, transformational, or servant, is helpful. The specific nature of the leader behavior appears to play a lesser role. This finding is particularly important because it appears to indicate that servant leaders could be as effective as, if not sometimes better than, transformational leaders in terms of crisis leadership.

We could apply more specific theory to these findings. Indeed, whereas our experimental studies show only limited differences between the effectiveness of SL and TFL in uncertain versus stable environmental conditions (if anything, the results provide some indication that transformational leaders may perform slightly better in uncertain conditions), the results of the field survey study show that TFL is perceived as less effective with greater job uncertainty; a similar effect was not observed for SL. In other words, SL may be less sensitive to uncertainty in this case. This result is in direct contrast to the reasoning by such authors as Smith et al. (2004). An explanation for this result could be that when uncertainty affects one's organization (as in studies 1 and 2), transformational leaders may be more effective because they emphasize the needs of the organization rather than the individual. However, when environmental uncertainty affects the individual specifically (such as job uncertainty in study 3), SL may be particularly effective because it emphasizes individual needs. It shows that in uncertain circumstances, the interpersonal qualities of leadership may be as important, if not more. Although only an exploratory idea, this does open a new avenue of theorizing and research.

7.2. Strengths and limitations

Naturally, the construction of this paper is not without limitations. First, we used scenario experiments for two studies, which was necessary to investigate the causal relationships of interest. However, by using scenario studies in studies 1 and 2, the results may reflect the perception of the respondents more than their actual experience. Thus, although the internal validities of these studies were high, generalizability was somewhat absent. Fortunately, the cross-sectional field study confirmed the major findings of the experiments, which significantly strengthened the generalizability of the research.

Second, no behavioral data were included, and our field study was only cross-sectional. However, we employed an experimental design with a between-subjects design in two of the three studies, which are strengths that lend credence to the results provided that common-method variance, which is problematic for cross-sectional studies, is not a significant issue. It should also be noted that scenario experiments generally score relatively well on mundane realism. Nevertheless, even within the combined experimental design and cross-sectional study, it cannot be excluded that some type of reverse causality plays a role in our findings.

Finally, we observed inconclusive results for uncertainty. Although we are able to theoretically explain the differences in the findings (see above), they may be the result of using a variety of different measures that were not optimal. We would therefore welcome new research focusing on the interaction of leadership style and perceptions of uncertainty.

7.3. Conclusion

In conclusion, the studies reported in this paper were designed to enhance our understanding of the similarities and differences between servant and TFL behavior. The results confirmed both the overlap between styles and the different pathways through which leaders exert their influence on followers. We believe that this study provides a fruitful basis for studying SL in the future.

561D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

References

Agle, B. R., Nagarajan, N. J., Sonnenfeld, J. A., & Srinivasan, D. (2006). Does CEO charisma matter? An empirical analysis of the relationships among organizational performance, environmental uncertainty, and top management team perceptions of CEO charisma. Academy of Management Journal, 49, 161–174.

Allen, N. J., & Meyer, J. P. (1990). The measurement and antecedents of affective, continuance and normative commitment to the organization. Journal of Occupational Psychology, 63, 1–18.

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and two-step approach. Psychological Bulletin, 103, 411–423. Antonakis, J., Avolio, B. J., & Sivasubramaniam, N. (2003). Context and leadership: An examination of the nine-factor full-range leadership theory using the

Multifactor Leadership Questionnaire. The Leadership Quarterly, 14, 261–295. Avolio, B. J., Zhu, W., Koh, W., & Bhatia, P. (2004). Transformational leadership and organizational commitment: The mediating role of psychological

empowerment and the moderating role of structural distance. Journal of Organizational Behavior, 25, 951–968. Bandalos, D. L. (2002). The effects of item parceling on goodness-of-fit and parameter estimate bias in structural equation modeling. Structural Equation Modeling,

9, 78–102. Barbuto, J. E., & Wheeler, D. W. (2006). Scale development and construct clarification of servant leadership. Group & Organizational Management, 31, 300–326. Bass, B. M. (1985). Leadership and performance beyond expectations. Philadelphia: Lawrence Erlbaum Associates. Bass, B. M. (2005). Transformational leadership (second ed.)Philadelphia: Lawrence Erlbaum Associates. Bass, B. M., & Riggio, R. E. (2006). Transformational leadership (2nd ed.)Mahwah, NJ: Lawrence Erlbaum. Beyer, J. M., & Browning, L. D. (1999). Transforming an industry in crisis: Charisma, routinization, and supportive cultural leadership. The Leadership Quarterly, 10,

483–520. Bono, J. E., & Judge, T. A. (2003). Self-concordance at work: toward understanding the motivational effects of transformational leaders. Academy of Management

Journal, 46, 554–571. Brown, M. (1969). Identification and some conditions of organizational involvement. Administrative Science Quarterly, 14, 346–355. Choi, Y., & Mai-Dalton, R. R. (1998). On the leadership function of self-sacrifice. The Leadership Quarterly, 9, 475–501. Cohen, F., Solomon, S., Maxfield, M., Pyszczynski, T., & Greenberg, J. (2004). Fatal attraction: The effects of mortality salience on evaluations of charismatic,

task-oriented, and relationship oriented leaders. Psychological Science, 15, 846–851. Conger, J., & Kanungo, R. (1987). Toward a behavioral theory of charismatic leadership in organizational settings. Academy of Management Journal, 12, 637–647. De Cremer, D. (2006). Affective and motivational consequences of leader self-sacrifice: The moderating effect of autocratic leadership. Leadership Quarterly, 17, 79–93. De Hoogh, A. H. B., Den Hartog, D. N., & Koopman, P. L. (2005). Linking the big-five factors of personality to charismatic and transactional leadership: perceived

dynamic work environment as a moderator. Journal of Organizational Behavior, 26, 839–865. De Hoogh, A. H. B., den Hartog, D. N., Koopman, P. L., Thierrry, H., van den Berg, P. T., van der Weide, J. G., et al. (2004). Charismatic leadership, environmental

dynamism, and performance. European Journal of Work and Organizational Psychology, 13, 447–471. De Koster, R., Stam, D., & Balk, B. (2011). Accidents happen: The influence of safety-specific transformational leadership, safety consciousness, and hazard

reducing systems on warehouse accidents. Journal of Operations Management, 29, 753–765. Deci, E. L., Ryan, R. M., Gagne, M., Usunov, J., & Kornazheva, B. P. (2001). Need satisfaction, motivation, and well-being in the work organizations of a former

Eastern Bloc country. Personality and Social Psychology Bulletin, 27, 930–942. DeGroot, T., Kiker, D. S., & Cross, T. C. (2000). A meta-analysis to review organizational outcomes related to charismatic leadership. Canadian Journal of

Administrative Sciences, 17, 356–372. DeRue, D. S., Nahrgang, J. D., Wellman, N., & Humphrey, S. E. (2011). Trait and behavioral theories of leadership: An integration and meta-analytic test of the their

relative validity. Personnel Psychology, 64, 7–52. Dutton, J. E. (1986). The processing of crisis and non-crisis strategic issues. Journal of Management Studies, 23(5), 501–517. Dvir, T., Eden, D., Avolio, B., & Shamir, B. (2002). Impact of transformational leadership on follower development and performance: A field experiment. Academy of

Management Journal, 45, 735–744. Ehrhart, M. G. (2004). Leadership and procedural justice climate as antecedents of unit-level organizational citizenship behavior. Personnel Psychology, 57, 61–94. Farling, M. L., Stone, A. G., & Winston, B. E. (1999). Servant leadership: Setting the stage for empirical research. Journal for Leadership Studies, 6, 49–72. Gagne, M. (2003). Autonomy support and need satisfaction in the motivation and well-being of gymnasts. Journal of Applied Sport Psychology, 15, 372–390. Graham, J. W. (1991). Servant-leadership in organizations: Inspirational and moral. Leadership Quarterly, 2, 105–119. Graham, J. W. (1995). Leadership, moral development, and citizenship behavior. Business Ethics Quarterly, 5, 43–54. Greenleaf, R. K. (1977). Servant leadership: A journey into the nature of legitimate power and greatness. New York: Paulist Press. Hall, D. T., Schneider, B., & Nygren, H. T. (1970). Personal factors in organizational identification. Administrative Science Quarterly, 15, 176–189. Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling [White paper].

Retrieved from. http://www.afhayes.com/public/process2012.pdf Hogg, M. A. (2007). Uncertainty-identity theory. Advances in Experimental Social Psychology, 39, 69–126. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation

Modeling, 6, 1–55. Hu, J., & Liden, R. C. (2011). Antecedents of team potency and team effectiveness: An examination of goal and process clarity and servant leadership. Journal of

Applied Psychology, 94, 851–862. Humphreys, J. H. (2005). Contextual implications for transformational and servant leadership. A historical investigation. Management Decision, 43, 1410–1431. Judge, T. A., & Piccolo, R. F. (2004). Transformational and transactional leadership: A meta-analytic test of their relative validity. Journal of Applied Psychology, 89, 36–51. Lane, M. S., & Klenke, K. (2004). The ambiguity tolerance interface: A modified social cognitive model for leading under uncertainty. Journal of Leadership and

Organizational Studies, 10(3), 69–81. Liden, R. C., Wayne, S. J., Zhao, H., & Henderson, D. (2008). Servant leadership: Development of a multidimensional measure and multi-level assessment. The

Leadership Quarterly, 19, 161–177. Little, T. D., Cunningham, W. A., Shahar, G., & Wideman, K. F. (2002). To parcel or not to parcel: Exploring the question, weighing the merits. Structural Equation

Modeling, 9, 151–173. Lord, R., & Maher, K. (1991). Leadership and information processing: Linking perceptions and processes. Lowe, K. B., & Kroeck, K. G. (1996). Effectiveness correlates of transformational and transactional leadership: A meta-analytic overview of the MLQ literature. The

Leadership Quarterly, 7, 385–426. Madera, J. M., & Smith, D. B. (2009). The effects of leader negative emotions on evaluations of leadership in a crisis situation: The role of anger and sadness.The.

Leadership Quarterly, 20(2), 103–114. Mathieu, J. E., & Zajac, D. M. (1990). A review and meta-analysis of the antecedents, correlates, and consequences of organizational commitment. Psychological

Bulletin, 108, 171–194. Mauno, S., Leskinen, E., & Kinnunen, U. (2001). Multi-wave, multi-variable models of job insecurity: applying different scales in studying the stability of job

insecurity. Journal of Organizational Behavior, 22, 919–937. Mayer, D. (2010). Servant leadership and follower need satisfaction: Where do we go from here? In D. van Dierendonck, & K. Patterson (Eds.), Servant leadership.

Developments in theory and practice (pp. 147–154). Hampshire, UK: Palgrave Macmillan. Mayer, D. M., Bardes, M., & Piccolo, R. F. (2008). Do servant–leaders help satisfy follower needs? An organizational justice perspective. European Journal of Work

and Organizational Psychology, 17, 180–197. Meindl, R., & Ehrlich, S. B. (1987). The romance of leadership and the evaluation of organizational performanceAcademy of Management Journal, 30, 91–109. Meindl, J. R., Ehrlich, S. B., & Dukerich, J. M. (1985). The romance of leadership. Administrative Science Quarterly, 30, 78–102. Meyer, J. P., & Herscovitch, L. (2001). Commitment in the workplace: Toward a general model. Human Resource Management Review, 11, 299–326.

562 D. van Dierendonck et al. / The Leadership Quarterly 25 (2014) 544–562

Meyer, J. P., Stanley, D. J., Herscovitch, L., & Topolnytsky, L. (2002). Affective, continuance, and normative commitment to the organization: A meta-analysis of antecedents, correlates and consequences. Journal of Vocational Behavior, 61, 20–52.

Mowday, R. T., Steers, R. M., & Porter, L. W. (1979). The measurement of organizational commitment. Journal of Vocational Behavior, 24, 224–247. Muthén, L. K., & Muthén, B. O. (2010). MPlus user's guide (6th ed.)Los Angeles: Muthén & Muthén. Nederveen Pieterse, A., Van Knippenberg, D., Schippers, M., & Stam, D. (2010). Transformational and transactional leadership and innovative behavior: The

moderating role of empowerment. Journal of Organizational Behavior, 31(4), 609–623. O'Leary, B. S., Lindholm, M. L., Whitford, R. A., & Freeman, S. E. (2002). Selecting the best and brightest: Leveraging human capital. Human Resource Management,

41, 325–340. Parolini, J., Patterson, K., & Winston, B. (2009). Distinguishing between transformational and servant leadership. Leadership and Organization Development Journal, 30,

274–291. Pawar, B. S., & Eastman, K. K. (1997). The nature and implications of contextual influences on transformational leadership: A conceptual examination. Academy of

Management Review, 22, 80–109. Pearson, C. M., & Clair, J. A. (1998). Reframing crisis management. Academy of management review, 23(1), 59–76. Pearson, C. M., & Mitroff, I. I. (1993). From crisis prone to crisis prepared: A framework for crisis management. The academy of management executive, 7(1), 48–59. Peterson, S. J., Galvin, B. M., & Lange, D. (2012). CEO servant leadership: Exploring executive characteristics and firm performance. Personnel Psychology, 65,

565–596. Pillai, R., & Meindl, J. R. (1998). Context and charisma: A “meso” level examination of the relationship of organic structure, collectivism, and crisis to charismatic

leadership. Journal of Management, 24, 643–671. Podsakoff, P. M., MacKenzie, Moorman, R. H., & Fetter, R. (1990). Transformational leader behaviours and their effects on follower's trust in leader, satisfaction,

and citizen behaviours. The Leadership Quarterly, 1, 107–143. Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior

Research Methods, 40, 879–891. Rafferty, A. E., & Griffin, M. A. (2004). Dimensions of transformational leadership: Conceptual and empirical extensions. The Leadership Quarterly, 15, 329–354. Riketta, M., & Van Dick, R. (2005). Foci of attachment in organizations: A meta-analytic comparison of the strengths and correlates of workgroup versus

organizational identification and commitment. Journal of Vocational Behavior, 67, 490–510. Ryan, R. M., & Brown, K. W. (2003). Why we don't need self-esteem: On fundamental needs, contingent love, and mindfulness. Psychological Inquiry, 14, 71–76. Ryan, R. M., & Deci, E. L. (2000). The darker and brighter sides of human existence: Basic psychological needs as a unifying concept. Psychological Inquiry, 11, 319–338. Ryan, M. K., & Haslam, S. A. (2007). The glass cliff: Exploring the dynamics surrounding the appointment of women in precarious leadership positions. Academy of

Management Review, 32, 549–572. Schaubroeck, J., Lam, S. S. K., & Chunyan Peng, A. (2011). Cognition-based and affect-based trust as mediators of leader behavior influences of team performance.

Journal of Applied Psychology, 96, 863–871. Schaufeli, W. B., Bakker, A. B., & Salanova, M. (2006). The measurement of work engagement with a short questionnaire. Educational and Psychological

Measurement, 66, 701–716. Schein, V. E. (1973). The relationship between sex role stereotypes and requisite management characteristics. Journal of Applied Psychology, 57, 95–105. Shamir, B., House, R., & Arthur, M. (1993). The motivational effects of charismatic leadership: A self-concept based theory. Organizational Science, 4, 577–594. Sheldon, K. M., Elliot, A. J., Kim, Y., & Kasser, T. (2001). What is satisfying about satisfying events? Testing 10 candidate psychological needs. Journal of Personality

and Social Psychology, 80, 325–339. Skogstad, A., Einarsen, S., Torsheim, T., Aasland, M. S., & Hetland, H. (2007). The destructiveness of laissez-faire leadership behavior. Journal of Occupational Health

Psychology, 12, 80–92. Smith, B. N., Montagno, R. V., & Kuzmenko, T. N. (2004). Transformational and servant leadership: Content and contextual comparisons. Journal of Leadership and

Organizational Studies, 10, 80–91. Stone, A. G., Russell, R. F., & Patterson, K. (2004). Transformational versus servant leadership: A difference in leader focus. Leadership and Organization

Development Journal, 25, 349–361. Stubbart, C. I. (1987). Improving the quality of crisis thinking. Columbia Journal of World Business, 22, 89–99. Sun, P. Y. T. (2013). The servant identity: Influences on the cognition and behavior of servant leaders. The Leadership Quarterly, 24, 544–557. Sverke, M., & Hellgren, J. (2002). The nature of job insecurity: Understanding employment uncertainty on the brink of a new millennium. Applied Psychology,

51(1), 23–42. Sverke, M., Hellgren, J., & Näswall, K. (2002). No security: a meta-analysis and review of job insecurity and its consequences. Journal of occupational health

psychology, 7(3), 242. Sy, T., Choi, J. N., & Johnson, S. K. (2014). Reciprocal interaction between group perceptions of leader charisma and group mood through mood contagion.

Leadership Quarterly, 24, 463–476. Van Den Broek, A., Vansteenkiste, M., De Witte, H., & Lens, W. (2008). Explaining the relationships between job characteristics, burnout and engagement: The role

of basic psychological need satisfaction. Work and Stress, 22, 277–294. Van Dierendonck, D. (2011). Servant leadership: A review and synthesis. Journal of Management, 37, 1228–1261. Van Dierendonck, D., & Nuijten, I. (2011). The servant leadership survey: Development and validation of a multidimensional measure. Journal of Business and

Psychology, 26, 249–267. Waldman, D. A., Ramirez, G. G., House, R. J., & Puranam, P. (2001). Does leadership matter? CEO leadership attributes and profitability under conditions of

perceived environmental uncertainty. Academy of Management Journal, 1, 134–143. Walumbwa, F. O., Hartnell, C. A., & Oke, A. (2010). Servant Leadership, procedural justice climate, service climate, employee attitudes, and organizational

citizenship behavior: A cross-level investigation. Journal of Applied Psychology, 95, 517–529. Weber, M. (1947). The theory of social and economic organization. Glencoe, IL: The Free Press. Yukl, G. (1999). An evaluation of conceptual weaknesses in transformational and charismatic leadership theories. The Leadership Quarterly, 10, 285–305. Zhu, W., Avolio, B. J., & Walumbwa, F. O. (2009). Moderating role of follower characteristics with transformational leadership and follower engagement. Group &

Organization Management, 34, 590–619.

  • Same difference? Exploring the differential mechanisms linking servant leadership and transformational leadership to follower outcomes
    • 1. Introduction
    • 2. Transformational leadership and servant leadership
    • 3. The differential influences of servant leadership and transformational leadership
      • 3.1. Attributions of leadership effectiveness
      • 3.2. Psychological needs
      • 3.3. Organizational commitment and work engagement
      • 3.4. Moderating effect of uncertainty
      • 3.5. Overview of the studies
    • 4. Study 1
      • 4.1. Methods
        • 4.1.1. Participants and design
        • 4.1.2. Experimental manipulations
          • 4.1.2.1. Leadership
          • 4.1.2.2. Organizational uncertainty
        • 4.1.3. Dependent measures
          • 4.1.3.1. Manipulation checks
          • 4.1.3.2. Need satisfaction
          • 4.1.3.3. Perceived effectiveness
          • 4.1.3.4. Organizational commitment
      • 4.2. Results
        • 4.2.1. Manipulation checks
        • 4.2.2. Ratings of leadership effectiveness, follower need satisfaction and organizational commitment
        • 4.2.3. Mediation analysis
        • 4.2.4. Moderation role of business environment
      • 4.3. Conclusion
    • 5. Study 2
      • 5.1. Methods
        • 5.1.1. Participants and design
        • 5.1.2. Procedure
        • 5.1.3. Experimental manipulations
          • 5.1.3.1. Leadership
          • 5.1.3.2. Business environment
        • 5.1.4. Dependent measures
          • 5.1.4.1. Manipulation checks
          • 5.1.4.2. Psychological need satisfaction
          • 5.1.4.3. Perceived leadership effectiveness
          • 5.1.4.4. Work engagement
      • 5.2. Results
        • 5.2.1. Manipulation checks
        • 5.2.2. Ratings of leadership effectiveness, follower need satisfaction and work engagement
          • 5.2.2.1. Work engagement
          • 5.2.2.2. Perceptions of leader effectiveness
          • 5.2.2.3. Need satisfaction
          • 5.2.2.4. Mediation analysis
      • 5.3. Conclusion
    • 6. Study 3
      • 6.1. Methods
        • 6.1.1. Participants and design
        • 6.1.2. Measures
          • 6.1.2.1. SL
          • 6.1.2.2. TFL
          • 6.1.2.3. Job uncertainty
          • 6.1.2.4. Need satisfaction
          • 6.1.2.5. Perceived leadership effectiveness
          • 6.1.2.6. Work engagement
          • 6.1.2.7. Organizational commitment
      • 6.2. Analysis
      • 6.3. Results
      • 6.4. Conclusion
    • 7. General discussion
      • 7.1. Theoretical implications
      • 7.2. Strengths and limitations
      • 7.3. Conclusion
    • References