Researchers have come to the consensus that teams act as “complex adaptive” systems (Ilgen et al., 2005, p. 519), which suggests that team creativity is affected not only by the intrinsic features of the team but also by the properties and behaviors of its members as well as by the properties of the overarching social context to which they belong. Using an “Input-Mediators-Outcome-Input” (IMOI) model as a conceptual guide, Cirella et al. (2014) captures the set of factors that have been considered important for team creativity in past studies into individual-level inputs, team-level inputs, team-level mediators (processes), team-level mediators (emergent states), and properties of macro-social system.

Diversity between individuals in a team was observed as one of the most important individual-level inputs influencing team creativity. However, studies on team diversity and creativity yielded mixed results. Diversity is defined as “the distribution of differences among the members of a unit with respect to a common attribute” (Harrison and Klein, 2007, p. 1200). Milliken and Martins (1996) have listed 14 different attributes potentially affecting group outcomes on which individuals differ (e.g., race/ethnic background, nationality, and gender). Researchers have argued that diversity may represent a “double-edged sword” (Milliken and Martins, 1996), as it may have both positive as well as negative consequences (Shore et al., 2009). Based on an information processing approach (Dahlin et al., 2005), diversity can be beneficial in more diverse teams because more heterogeneous knowledge and perspectives are available to teams, which would potentially increase team creativity. However, similarity-attraction paradigm (Byrne, 1997), social identity theory and self-categorization processes (Turner and Tajfel, 1986) argue that diversity can hinder group process by limiting common understandings and shared experiences or by creating such a divergence of ideas and styles that detrimental conflict can result (Shin et al., 2012). As team diversity may be a covariate of shared leadership and team creativity, it was (although the direction of its influence may not clearly be predictable) therefore included in this study as a control variable.

Shared leadership has a potentially comprehensive influence on the contextual and structural characteristics of team innovation. This means multiple team members fulfill critical team leadership functions, collaboratively solve problems and collectively assume the responsibility for team outcomes. There is much evidence to support the positive effects of shared leadership on both team effectiveness (Muethel and Hoegl, 2013) and team performance (Mehra et al., 2006; Carson et al., 2007). Although the importance of shared leadership have been widely noted by scholars and practitioners, empirical examinations of the relationship between shared leadership and team creativity are very limited, among which, for example, Lee et al. (2015) showed that shared leadership positively contributed to team creativity in an e-learning environment. Wu and Cormican (2016) found that the density of a shared leadership network is positively related to team creativity in engineering design teams.

Past researchers have identified several environmental factors that are conducive to creativity: teams that are highly cooperative, teams that can self-determine procedures to perform tasks, organizations that have established a norm to actively share opinions, and so forth (Amabile, 2012). Shared leadership within a team precisely meets these conditions, such as autonomy for team members, support for cooperation among members with diverse expertise, and a team climate encouraging communication. Based on these studies, we proposed the following hypothesis:

Novelty and usefulness are frequently used as indicators of team creativity. Novelty means the degree to which the idea is unique from existing ideas. Usefulness indicates the degree of value offered by the idea within the organization and the broader domain in which it is embedded (Amabile, 1996; Berg, 2014). Previously, researchers have tended to assume that these two dimensions travel together in creative ideas (Oldham and Cummings, 1996; Shalley and Perry-Smith, 2001). However, increasing evidence suggests that novelty and usefulness are orthogonal dimensions (Ford and Gioia, 2000). Therefore, both dimensions were taken into consideration but examined separately in the current research.

Shared leadership is a group process in which leadership is distributed among and stems from team members. Vertical leadership emphasizes that a certain managerial role outside (and above) the work team possesses formal authority over the team and is responsible for team process and consequences (Druskat and Wheeler, 2003). Although vertical leaders continue to play a significant role in developing and maintaining shared leadership, shared leadership should also play an important role in explaining team effectiveness, especially when team tasks are highly interdependent, highly complex, or require great levels of creativity (Pearce and Bruce, 2004). Pearce and Sims (2002), Pearce and Bruce (2004) and Ensley et al. (2006) found that shared leadership is a more useful predictor of team outcomes than vertical leadership—in change management and virtual and new venture teams, respectively. However, this is not necessarily to downplay the relative role of vertical leadership. As noted by Hoch (2013), “shared leadership is not mutually exclusive to other leadership forms and behaviors, but can be engaged in simultaneously with other approaches such as vertical leadership.”

To eliminate the possible effect of vertical leadership on team creativity, the current study did not assign vertical leadership in the teams, following other researchers in the field (e.g., Small and Rentsch, 2010). In this way, we could focus on the function of shared leadership on team creativity and on the underline mechanism of the function.

Controversy occurs when team members express their opposing ideas, opinions, conclusions, theories, and information that at least temporarily obstructs resolving an issue. Controversy within teams, when properly harnessed, could promote dialog and debate that stimulate innovation (Nonaka, 1990; Leonard and Straus, 1997; Leonard and Sensiper, 1998). Constructive controversy is defined as open-minded discussion of opposing perspectives for mutual benefit. The following are key components of constructive controversy: frankly expressing one's personal opinions, feeling uncertain instead of defensiveness about one's own positions, feeling eager to know and comprehend the opponent's arguments, viewing a situation or understanding of a concept from an alternate point-of-view, taking new and opposing information seriously, incorporating the arguments of the opponents into one's own ideas, and creating alternative solutions based on the more complete set of information (Tjosvold et al., 2009).

There are some facilitative interpersonal conditions that could promote constructive controversy, including cooperative goal interdependence, confirmation of personal competence, and collaborative influence (Tjosvold, 1985). Note that constructive controversy cannot be reduced to a sum of exchanged information; rather, ideas are continually built upon each other during the discussing process. It is through such mutual influence that people who hold different or even opposite views come to understand and appreciate one another. The result is idea generation, and ultimately, creative solutions. Shared leadership within a team is beneficial in that members become more involved in discussions and engage in deeper processing of ideas proposed by fellow members. Shared leadership also contributes to healthy criticism and comprehension of new information as well as to the integration of opposing ideas (Hoch, 2013). Hence, shared leadership promotes constructive controversy in teams.

In addition, constructive controversy has been found to enhance the quality of decision-making and team creativity. As diversity in ideas and opinions is almost unavoidable in decision-making or problem-solving within a team, an open and frank way to address conflicts is of great importance (Tjosvold, 1985). Constructive controversy guarantees a free and open discussion of diverse views. It also stimulates and activates team members' cognition and thinking (Tjosvold, 2008). Consequently, team members are more likely to promote creative decision-making, which leads to enhanced group-level creativity. Chen and Tjosvold (2002) reported an impact of constructive controversy on the innovation of teams. Alper et al. (2000) found that with more constructive controversy, the innovation of the team is higher. Open expression and frank communication of divergent opinions during the process alleviate possible tensions arising from dissents, facilitate idea integration among team members, give rise to transformational ideas and enhance team creativity (Alper et al., 2000; Chen et al., 2005). Based on this rationale, we expected that the increased constructive controversy triggered by shared leadership would enhance team creativity. Thus, we proposed the following hypothesis:

We proposed that shared leadership would influence team creativity through constructive controversy. However, this relationship may be moderated by other factors. Specifically, we zero in on the moderating role of team goal orientation.

Team goal orientation is defined as a consensus shared by team members about the goals of the team (Bunderson and Sutcliffe, 2003). As an emergent state, team-level goal orientation is subject to the influence of goal-relevant cues in the context, such as rewards and recognition (Maltarich et al., 2016), performance evaluation and organization policies (Druskat and Wheeler, 2003). For instance, in a social context where new ideas are valued, competency development is encouraged, and innovation is rewarded, a group climate in favor of learning orientation is more likely to form (Bunderson and Sutcliffe, 2003). By contrast, in an environment with competitive and persistent job evaluation and emphasis on performance-based rewards, a group climate favoring performance orientation is more likely to prevail (Dragoni, 2005).

For teams with differential goal orientations, the indirect effect of shared leadership on team creativity through constructive controversy may differ. Specifically, the constructive controversy triggered by shared leadership will promote team creativity when teams are led by learning goal orientation, which embodies group members' consensus of the goals of learning and competency development. According to Bunderson and Sutcliffe (2003), team-learning orientation determines the extent, range and intensity of learning behaviors in teams. Immersed in such a climate, the constructive controversy stimulated by shared leadership has beneficial effects: team members are encouraged to absorb instructive points from one another and persistently search for fresh ideas to deepen their understanding of the task and foster team competency. These efforts contribute to the promotion of team creativity in the objective sense.

In the case of performance orientation, the indirect effect of shared leadership on team creativity through constructive controversy may be weaker because such teams add more weight to the performance indicators than the task per se, tend to regard mistakes as evaluative threats (Martocchio and Frink, 1994), give up more easily when facing obstacles (Button et al., 1996), and contribute the least effort (Mangos and Steele-Johnson, 2001). As a result, the constructive controversy goes in the wrong direction (i.e., focusing on external performance appraisal instead of the ongoing task), and it is difficult to boost creative idea generation. Thus, we proposed the following hypothesis:

Study 1 aimed to examine the positive impact of shared leadership on team creativity using the measurement of real-world task teams (Hypothesis 1). Study 2 further tapped into the mediating mechanism of constructive controversy in another wave of task teams (Hypothesis 2). In Study 3, we manipulated team goal orientation experimentally and tested its moderating effect on the relationship among shared leadership, constructive controversy and team creativity (Hypothesis 3). Taken together, we propose a moderated-mediation model (see Figure 1). The project was reviewed and approved by the Academic Ethics Committee of the School of Psychology at Beijing Normal University (approval number: 2015068) before being conducted.

This study contributes to the literature by providing theorists and researchers with a better understanding of the nature of shared leadership and its effect on team creativity. The current research is a reply to the appeal that more research attention should be devoted to examining the moderating and mediating variables of the shared leadership and team outcome (Pearce and Conger, 2002). Through our research, we deepen the understanding of how shared leadership increases team creativity and under what conditions this mechanism works better. Moreover, relatively few studies have been conducted on team level creativity (Lee et al., 2015). By considering the novelty and usefulness dimension of creativity separately, the current research is a valuable endeavor in exploring how different components of team creativity may respond differently to the shared leadership process of a team.

Firstly, researchers have not yet agreed on how shared leadership influences team innovation. Our finding that shared leadership enhanced team creativity by facilitating constructive controversy provides an explanation of the underlying mechanism. Song et al. (2015) investigated how shared leadership influences team performance and team creativity, respectively. In their study, shared leadership significantly influenced team performance through the cognitive mechanism of information exchange. However, the anticipated mediating role of information exchange between shared leadership and team creativity was not confirmed. We contend that a clashing and merging process of ideas and knowledge among members beyond mere information exchange is necessary for achieving a high level of team creativity. This core characteristic of team creativity is captured by the concept of constructive controversy. Constructive controversy emphasizes that team members openly discuss their opposite opinions, focus on cooperation and interdependence in pursuing collective goals, confirm the positions of their opponents based on their own evidence, and give more attention to the generation of new ideas based on divergent information and perspectives instead of being anxious to win the debate (Tjosvold, 1985). What makes constructive controversy different from information sharing is that during the process there is not only information exchange but also mutual inspiration and support of different perspectives. It is through such a thorough mutual understanding and appreciation process that new ideas and solutions are generated as the result of a collective endeavor.

Secondly, we noted a boundary condition for the established relationship between shared leadership and team creativity. Although shared leadership can spur the integration of different ideas and perspectives in a team (i.e., constructive controversy), its positive effects may be suppressed when the team is performance goal oriented. This suppression occurs probably because when the team places too much emphasis on performance, members become more concerned with the evaluations of others, longing for appreciation and avoiding failure. Conversely, learning goal-oriented teams focus on developing competency, engage in more active learning behaviors, and strive to generate more new solutions, thus ensuring the positive role of shared leadership and constructive controversy.

Thirdly, we found that the mediation mechanism between shared leadership and team creativity and its boundary condition hold only for the novelty component of creativity, not for the usefulness dimension. This finding is consistent with the existing literature. More and more scholars recognized that it may be rare for ideas to be observed as high in both novelty and usefulness (Rietzschel et al., 2010; Mueller et al., 2012) and suggested that novelty and usefulness probably are motivated by different psychological processes (Rietzschel et al., 2010; Grant and Berry, 2011; Mueller et al., 2012; Berg, 2014). For example, individuals generate useful solutions under conditions when they are experiencing cognitive closure (Miron-Spektor and Beenen, 2015), or when they are anxious to reduce uncertainty by drawing on existing practices (Janssen and Van Yperen, 2004; Mueller et al., 2012). In contrast, the generation of novel ideas is most likely when individuals are intrinsically motivated (Grant and Berry, 2011), cognitively flexible (Miron-Spektor and Beenen, 2015), eager to learn new domains, and feel safe to take risks (Hirst et al., 2009). These conditions are exactly what shared leadership and learning goal orientation could jointly offer. Past research on shared leadership and team creativity has tended to examine novelty and usefulness together in one overall creativity construct (Oldham and Cummings, 1996; Shalley and Perry-Smith, 2001). By exploring novelty and usefulness separately, the current study deepens our understanding as to how each aspect of team creativity benefits from shared leadership. Future research could also explore mechanisms under which the usefulness dimension of team creativity could be improved.

Team creativity is of great significance for the success of organizations. The findings of our study thus may have useful implications to management practitioners.

Firstly, team creativity could be promoted by advocating shared leadership in teams. Demands for team creativity, ranging from designing products to innovating working ideas, are ubiquitous in contemporary organizations. If we could create certain conditions to facilitate the development of shared leadership, team novelty and competitiveness of the organization would improve accordingly. As suggested by Pearce and Bruce (2004), training systems can be used to develop the shared leadership skills of both vertical leaders and team members; reward systems can be used to promote shared leadership in teams; and cultural systems can be used to emphasize the significance of shared leadership.

Secondly, as constructive controversy plays a critical role in the effect of shared leadership on team creativity, a group atmosphere of freely expressed dissent is worthy of advocacy. To achieve this state, managers could emphasize cooperative goals in teams. By doing so, members would be more motivated to take into account the rational parts of other's thoughts, seeking verification and elaboration of different opinions rather than antagonism.

Thirdly, managers should pay close attention to team goal orientations and create a benign climate for learning. For example, situational cues, group norms, organization policies, and cultural channels can be employed to convey the value of learning treasured by the organization. Only in such conditions can the advantage of shared leadership transfer to team creativity through active interactions. Several managerial behaviors are likely to foster team learning goal orientation. First, individual differences in learning orientation should be considered when selecting team members. However, putting high learning-oriented individuals on the team is not enough. What is even more important is for managers to provide support and encouragement to alleviate employee fear and anxiety that may arise from the uncertainty of creative endeavors. When teams experience failure, managers could provide educational coaching emphasis on what precious lessons the teams have learned during the process and what could have been done to make the result better. This support and coaching should nurture a the learning orientation of the team and boost team novelty, as suggested by the current research.

Several limitations of the current research should be addressed in future studies.

Firstly, as mentioned earlier, team diversity is an important individual-level input influencing team creativity, although past research on team diversity has yielded mixed results (Williams and O'Reilly, 1998; Jackson and Joshi, 2004; Harrison and Humphrey, 2010). Shin et al. (2012) suggest that diversity should have conceptual relevance to the outcome variables, that is, be task-related. To focus on the mechanism of how shared leadership influences team creativity, the current study did not test the impact of team diversity. Future research could take task-related team diversity into consideration and test how team diversity influences the relationship between shared leadership and team creativity. We propose that when task-related team diversity is high, shared leadership would have a stronger influence on team creativity because teams could benefit more from shared leadership when the competencies of their members differ from each other.

Secondly, to highlight the important function of shared leadership on team creativity does not mean that vertical leadership is dispensable. The results of this study provide robust evidence for the value of shared leadership. Findings in this regard can not only help establish the legitimacy of shared leadership research but also provide fruitful future research directions on the potential interplay between shared and vertical leadership. Future studies could explore how one utilizes both vertical and shared leadership to leverage the capabilities of knowledge workers in innovative teams.

Thirdly, the current study was conducted in China. The generalizability of the research results to other cultures needs further evidence. Some cultural dimensions may shape the relationship between shared leadership and team creativity. Take power distance, for example. China is a high-power distance culture. When shared leadership is advocated in teams embedded in a high-power distance culture, team creativity may benefit more, compared with those teams in low-power distance settings. The reason is because vertical leadership is a rather dominant role in high-power distance culture, and team members normally accept the authority of vertical leadership without questioning. When shared leadership was introduced in such teams, the wisdom and valuable input of team members could be released and team creativity could be boosted substantially. However, things may be different in a low-power distance culture. Further research could be conducted in different cultures to testify to the findings of the current research. Additionally, collecting data from teams in organizations will also contribute to the external validity of the findings of the current study.

Another limitation is that we manipulated team's performance orientation following the instruction of Miron-Spektor and Beenen (2015), which is widely used in the field of team research. The instruction introduced a competitive framework that could possibly generate anxiety or tension in team members and therefore might have a negative impact on the level of team creativity. Future research could make additional effort to manipulate performance orientation without inducing potential confounding affect (e.g., anxiety) in teams.

Finally, in our research, we manipulated two types of team goal orientations (i.e., learning and performance orientation), which have also been confirmed in the extant literature at the group level. Theories and practice of goal orientation at the individual level, however, provided a distinction between 3 goal orientations, namely, learning, performance-prove and performance-avoid orientations (Harackiewicz et al., 2002). Despite their common concern with the task, individuals with a performance-prove orientation have a strong desire to prove their ability, whereas those with a performance-avoid orientation mainly think about avoiding failure (Elliot and Church, 1997). A distinction between the two performance orientations has been proposed yet has rarely been studied empirically (Druskat and Wheeler, 2003). We believe that the three-dimensional frame of goal orientation provides another avenue for future researchers to deepen their understanding of its moderating mechanism in the constructive controversy-mediated relationship between shared leadership and team creativity.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.