6 Empowering Leadership and Team Creativity

The Roles of Team Learning Behavior, Team Creative Efficacy, and Team Task Complexity

Xiaomeng Zhang and Ho Kwong Kwan

Introduction

As one way of coping with rapid technological change and fierce global competition, many organizations rely heavily on employee creativity, which can substantially contribute to organizational innovation, productivity, and survival (Liu et al., 2017; Shalley, Zhou, & Oldham, 2004). To ignite the creative spark of employees, organizations are increasingly using team-based structures (Paulus, 2008). New ideas are proposed and pursued by work teams through interaction among team members, and such creative collaboration is shown to pay considerable dividends (Hülsheger, Anderson, & Salgado, 2009). Therefore, it is important to understand the conditions that contribute to team creativity (Shalley, Zhou, & Oldham, 2004), which refers to the production of novel and potentially useful ideas about products, services, processes, and procedures by a group of employees (Shin & Zhou, 2007).

A number of scholars theorize that the empowering behavior of team leaders plays a key role in giving team members latitude, thus stimulating creative action by the team (Amabile et al., 2004). Indeed, Zhang and Bartol (2010) theorized that empowering leadership addresses the underlying nature of creative work and they, for the first time, empirically found that empowering leadership has a positive influence on individual employees’ creativity. More recently, Zhang and Zhou (2014) also found that perceptions of empowering leadership are positively associated with employee creativity. Because organizations now rely heavily on team creativity to remain competitive, it is theoretically important and empirically necessary to extend the relationship between empowering leadership and creativity to a team level. Empowering leadership at the team level is defined as team leader behaviors that facilitate the sharing of power with and enhance the work motivation of team members (Srivastava, Bartol, & Locke, 2006). Moreover, Marks and colleagues’ team effectiveness model has theorized that team leader behaviors influence team effectiveness through the simultaneous mediating effects of team processes and emergent states (Marks, Mathieu, & Zaccaro, 2001). Therefore, a major purpose of this study is to build on and extend Amabile’s (1983, 1996) componential model of creativity and Amabile and Pratt’s (2016) dynamic componential theory of creativity to theorize a conceptual model that bridges empowering leadership with team creativity via the necessary mediation mechanisms.

Moreover, the contingency perspective of leadership (Liden & Antonakis, 2009; Yukl, 2013) holds that leadership is a social construct and cannot be fully understood when examined in isolation from the context in which it occurs. That is, a team leader’s role is salient only within a specific context (Zhang & Zhou, 2014). A few studies of creativity suggest that a promising direction for future research is to examine how leadership style interacts with other contextual variables in influencing the creative outcomes of teams (Wang, Kim, & Lee, 2016; Zhou & Hoever, 2014). Shalley and colleagues (2004) called for research into the interaction effects of supervisory style and task complexity—two key characteristics of the organizational context that have long been considered to be important contributors to creativity. Team task complexity refers to the extent to which a team task has a high level of information-processing requirements and few set procedures (Man & Lam, 2003). Based on the notion that a high level of team task complexity requires team leaders to play a greater role (Morgeson, 2005), we propose that empowering leadership has a greater impact on team learning behavior and team creative efficacy, and, ultimately, team creativity, when the level of task complexity is high.

In sum, this chapter makes a number of contributions. First, we extend both the leadership and the creativity literature by investigating the connection between empowering leadership and creativity in a field setting. Second, we consider team learning behavior and team creative efficacy (as the team process and the team emergent state, respectively) simultaneously to investigate their mediating effects on the empowering leadership-team creativity relationship. Third, we explore the moderating influence of team task complexity on the aforementioned relationships to identify the boundary condition for the effectiveness of empowering leadership.

Theory and Hypotheses

Empowering Leadership and Team Creativity

According to the studies of Arnold et al. (2000) and Srivastava, Bartol, and Locke (2006), there are five dimensions of empowering leader behavior: leading by example, which reflects a leader’s commitment to his or her own work and the work of team members for the achievement of better performance; coaching, which refers to actions that educate team members and help them to become more efficient and self-reliant; participative decision making, which encourages the sharing among team members of ideas and opinions in group decision making, and facilitates the recognition of valuable inputs from members; informing, which promotes the company-wide dissemination of information, resulting in team members who are more likely to understand the compelling mission and expectations of their leader, and thus to work cooperatively to fulfill collective goals; and showing concern, which includes the support and fair treatment of subordinates by a team leader in the form of trust, concern for their well-being, and willingness to help.

Conceptually, through leading by example, empowering leaders demonstrate their own commitment to the work and to the team to achieve better performance, and thus act as role models for subordinate team members, who will try in different ways to achieve the team goals related to creative outcomes. Through coaching, empowering leaders boost the confidence and facilitate the ability of team members to challenge the status quo and increase the level of self-reliance and efficiency of the latter in accomplishing tasks. Enhanced team confidence and abilities will help the team develop new ideas, resulting in more creative outcomes (Amabile & Pratt, 2016).

Furthermore, by encouraging participation and enhancing perceived autonomy in decision making, empowering leaders enable team members to make decisions and implement actions without direct intervention by their leaders (Bass, 1985). As a result, team members are empowered to engage in a process in which they can explore diverse alternatives, which leads to higher levels of team creativity (Gilson & Shalley, 2004). By providing information and clarifying expectations, empowering leaders communicate motivational direction. Research shows that leaders elicit creativity by setting creativity-specific expectations (Shalley, Zhou, & Oldham, 2004; Shin & Zhou, 2007). Finally, empowering leaders demonstrate their concern for followers by showing them consideration, offering them support, and treating them fairly. These forms of behavior facilitate collaboration among team members and help them overcome the fear of challenging the status quo, which leads to greater team creativity (Hirst, van Knippenberg, & Zhou, 2009). Therefore, we posit the following hypothesis:

• Hypothesis 1: Empowering leadership is positively related to team creativity.

The Mediating Roles of Team Learning Behavior and Team Creative Efficacy

Amabile’s (1983, 1996) componential model of creativity and Amabile and Pratt’s (2016) dynamic componential theory of creativity identify the learning of domain-relevant skills and creativity-relevant skills as the key components necessary for creativity. Domain-relevant skills comprise an individual’s complete set of response possibilities from which the new response is to be synthesized. This component can be considered as a set of cognitive pathways for generating a solution or solving a problem. The larger the set, the more the numerous alternative responses will be available to create something new. Creativity-relevant skills include a cognitive-perceptual style that is characterized by “a facility in understanding complexities and an ability to break set during problem-solving” (Amabile, 1996: 88). That is, it is important for individuals to learn to break a perceptual and cognitive set and understand complexities of the task or situation and perceive it creatively to produce creative work. Therefore, the learning of the task and related subjects (domain-relevant skills) expands the scope of cognitive pathways, and the learning of cognitive set breaking abilities (creativity-relevant skills) enhances the flexibility with which cognitive pathways are explored (Amabile & Pratt, 2016).

Creativity scholars recently have argued in favor of the concept of creativity in teams as a learning process that energizes team member behaviors to produce creative outcomes (Wang, Kim, & Lee, 2016). This is because creativity involves generation of something new for which the necessary strategies often have yet to be learned (Hirst, Van Knippenberg, & Zhou, 2009). Team learning behavior is defined as an ongoing process of collective reflection and action through which team members interactively acquire, share, refine, and apply task-relevant knowledge and skills (Edmondson, 1999). This process is characterized by activities including asking questions, seeking different perspectives and feedback, challenging assumptions, discussing errors or unexpected outcomes, experimenting, evaluating alternatives, and reflecting on past actions (Van der Vegt & Bunderson, 2005). Following Edmondson (1999), we used the term “team learning behavior” to avoid confusion with the notion of learning outcome as team learning is sometimes also referred to knowledge being embedded within the team (e.g., Argote & Olivera, 1999), which represents a state that indicates what a team has learned at a given time (Cohendet & Steinmueller, 2000). In this study, team learning behavior “reflects an active set of team processes” (Mathieu et al., 2008: 431), thus can be viewed as an example of a “group action process” in Marks and colleagues’ (2001) team effective framework, during which teams conduct activities leading to goal accomplishment (Van der Vegt & Bunderson, 2005).

Team learning behavior is a collective process (Savelsbergh, van der Heijden, & Poell, 2009). The role of team discussion of and experimentation on the task in question and reflection on it is a prerequisite for a team’s ability to convert an individual’s skills and knowledge into team knowledge (Drach-Zahavy & Somech, 2001). In addition, team learning behavior may encourage social learning process such that team members engage in learning activities when they observe their peers engaging in the process and perceive learning as valuable and supported (Hirst, Van Knippenberg, & Zhou, 2009). When team members observe their teammates’ engagement in the learning processes (e.g., asking challenging questions, criticizing each others’ ideas, or evaluating alternatives) that can lead a team to achieve creative outcomes, they will engage in the similar learning activities as these processes create a context in which it is easier to learn and reduce the psychological risks related to learning (Hirst, Van Knippenberg, & Zhou, 2009).

Extensive evidence indicates that a team leader’s behavior is a salient factor influencing whether and the extent to which team members pay attention to what other members say and do, and the ways in which they work together to solve problems (Srivastava, Bartol, & Locke, 2006). Empowering leader behaviors such as coaching help team members take collective and appropriate action and encourage them to exchange information and solve problems together (Dovey, 2002). In the process, they actively share ideas and utilize different opinions to improve overall team effectiveness. Coaching-oriented leaders help teams to become self-reliant by encouraging team members to build strong relationships with each other and work collaboratively to achieve collective goals (Arnold et al., 2000). Similarly, team leaders facilitate participative decision making by encouraging team members to openly express ideas and suggestions, which can ultimately influence decisions that affect the team and provide opportunities for team members to collaboratively evaluate each other’s suggestions (Locke, Alavi, & Wagner, 1997).

In addition, by conveying a clear compelling mission, empowering leaders provide teams with access to information and a collective goal, which can reduce insecurity and defensiveness among team members (Edmondson, 1999). Members then feel psychologically safe to freely challenge the assumptions underlying each other’s ideas and perspectives, and openly discuss differences of opinion. Empowering leaders show their respect for team members by recognizing their individual inputs to the team, which encourages the latter to contribute their knowledge and skills, provides them with the opportunity to learn from each other and facilitates the collaborative learning process (Srivastava, Bartol, & Locke, 2006). These empowering leader behaviors collectively encourage team members to engage in learning by increasing the knowledge and information available to team members, creating a safe and fair environment in which it is easier for team members to help one another and learn collaboratively (Hirst, Van Knippenberg, & Zhou, 2009). In sum, empowering leadership encourages team members to engage in team learning behavior, which in turn, promotes team creativity.

Another necessary building block for facilitating creativity, according to Amabile’s componential theory, is task motivation. Efficacy beliefs have long been regarded as one of the most important determinants of task motivation (Bandura, 1986). Self-efficacy nourishes intrinsic motivation by enhancing the perceptions of self-competence (Bandura, 1986), thus creative self-efficacy is argued to reflect intrinsic motivation to engage in creative activities (Gong, Huang, & Farh, 2009). Empirical evidence indicates that creative self-efficacy is a crucial factor that drives employee creativity (Gong, Huang, & Farh, 2009; Tierney & Farmer, 2011).

Team efficacy is an essential motivational factor that influences goal setting and commitment among team members, their initiation of and persistence in activities, and how they deal with challenging situations (Bandura, 1986, 1997). Previous studies have emphasized the importance of team efficacy to team performance (Lindsley, Brass, & Thomas, 1995). A meta-analysis of 67 empirical studies indicated that team efficacy is positively related to team performance (Gully et al., 2002). Whereas team efficacy refers to a team’s collective belief in its ability to perform a task successfully (Bandura, 1997), team creative efficacy concerns a team’s shared belief in its ability to perform a specific task—producing new and useful ideas. Research suggests that a domain-specific efficacy belief (i.e., team creative efficacy) is necessary in predicting a specific outcome (i.e., team creativity) (Shin & Zhou, 2007).

Studies suggest that team leaders often strongly influence the efficacy perception of team members (Kozlowski et al., 1996). Thus, we argue that empowering leadership plays a key role in enhancing team creative efficacy because it is likely to influence the four sources of efficacy beliefs identified by Bandura (1986) observational learning, verbal persuasion, enactive mastery, and physiological arousal. Specifically, by showing commitment to the work and to the team itself, empowering leaders act as role models for subordinate team members; that is, team members learn to commit to their work through observational learning. This makes team members more confident in their team’s ability to challenge the status quo and develop new ideas. By disseminating information throughout the company and clarifying expectations, empowering leaders can verbally persuade team members to coordinate their efforts to achieve collective goals such as team creativity. Kirkman and Rosen (1999) suggested that giving information about where an organization is going can help motivate team members to engage in actions that are aligned with collective goals.

In addition, by showing care and consideration for team members, empowering leaders encourage them to perform tasks without fear or anxiety, which increases the level of collective efficacy and can inspire team members to try different things. Research shows that support from leaders encourages team members to be more creative (Hülsheger, Anderson, & Salgado, 2009). In addition, such consideration and care can help team members understand that the leader values and appreciates their efforts to achieve collective goals, which causes physiological arousal among team members and strengthens their shared creative efficacy belief. Lastly, through coaching and the delegation of authority, empowering leaders help team members become more efficient and self-dependent in accomplishing tasks and help the team act autonomously. Team members are then likely to have successful enactive mastery experiences, thereby increasing the level of team efficacy related to creative outcomes. Thus, we offer the following hypothesis:

• Hypothesis 2: Team learning behavior and team creative efficacy mediate the positive relationship between empowering leadership and team creativity.

The Moderating Role of Team Task Complexity

The contingency perspective of leadership (Liden & Antonakis, 2009; Yukl, 2013) holds that a team leader’s influence depends on certain conditions. Leadership is found to have a greater impact on team processes, team emergent states, and team outcomes when the context requires a greater role to be played by team leaders (Yukl, 2013). Team behavior and the need for leadership depend on the demands of the task (McGrath, 1984). Under simple and routine task conditions, teams have less need for leadership because of the familiarity of the task environment (Zaccaro, Rittman, & Marks, 2001). However, it is important for team leaders to facilitate the team to function effectively when tasks are complex, creating novel environments for which teams do not have preexisting responses (Marks, Zaccaro, & Mathieu, 2000).

Team task complexity refers to the extent of the difficulty and/or complexity of the team task, characterized by more information-processing requirements and few set procedures, which require the use of a variety of complex and high-level skills (Man & Lam, 2003; Pelled, Eisenhardt, & Xin, 1999). According to Wood (1986), compared with simple tasks, complex ones contain more distinct informational cues that must be processed, placing substantial cognitive demands in terms of comprehension and execution on the people carrying them out, and requiring task doers to adapt more frequently to changes in the process of task execution. In contrast, simple and routine tasks create a relatively familiar and consistent task environment and thus place minimum cognitive demands on task doers, who develop routines that specify precise actions and whose information processing and behavior become automatic (Gersick & Hackman, 1990).

When the team task is complex, discussing competing perspectives is essential for team members to identify appropriate strategies to assess the situation and solve the problem (Jehn, Northcraft, & Neale, 1999). Also, there is a greater need for team members to pool their expertise and resources, exchange opinions, and challenge each other openly to accomplish the task (Pelled, Eisenhardt, & Xin, 1999; Wood, 1986); hence, higher levels of empowering leadership are required. Such leadership, as noted, provides team members with considerable latitude and facilitates interactions among them in the accomplishment of tasks. In contrast, when tasks are simple/routine and well-defined teams have less need for empowering leadership. This is because when the level of team task complexity is low, team members only need to follow standard procedures, the discussion of work methods and the exchange of opinions may not be necessary or essential, as the tasks are generally familiar and performed the same way each time (Jehn, Northcraft, & Neale, 1999). As a result, team leaders’ empowering leadership may be less effective in encouraging team members to engage in team learning behavior and team creative efficacy. Accordingly,

• Hypothesis 3a: Team task complexity moderates the relationship between empowering leadership and team learning behavior in such a way that empowering leadership has a stronger and positive relationship with team learning behavior when the level of team task complexity is high, while empowering leadership has a weaker impact on team learning behavior when the level of team task complexity is low.
• Hypothesis 3b: Team task complexity moderates the relationship between empowering leadership and team creative efficacy in such a way that empowering leadership has a stronger and positive relationship with team creative efficacy when the level of team task complexity is high, while empowering leadership has a weaker impact on team creative efficacy when the level of team task complexity is low.

In addition, based on the notion that team learning behavior and team creative efficacy mediate the relationship between empowering leadership and team creativity, it is logical to argue that the enhanced team learning behavior and team creative efficacy resulting from the positive interaction between empowering leadership and team task complexity will in turn contribute to team creativity. Hence, we advance a mediated moderation model and propose:

• Hypothesis 4: The interaction between empowering leadership and team task complexity is mediated by team learning behavior and team creative efficacy, as related to team creativity.

Method

Sample and Procedures

We collected the data from 102 R&D teams in three Chinese information technology (IT) companies. The participants were professional-level employees who were members on these teams and the corresponding external team leaders. Team members’ tasks are in areas including cell phone design, software development, and integrated network system design. With the assistance of the human resource managers from the three companies, a list of 113 teams, randomly selected, was compiled. These were intact teams in which members worked interactively to develop new products and services. All 866 team members (the number of team members ranged from 5 to 14 with an average of 8) and 113 external team leaders were invited to participate in the survey. Team leaders are the external leaders of these teams. They are not part of the team, as all of them supervise more than one team (ranging from two to four teams). We randomly chose one team for each team leader to prevent the design effect (Kaiser et al., 2006).

Separate questionnaires were administered to team leaders and team members. The survey questionnaires were coded before being distributed, and the human resource departments assisted in the recording of the identification numbers to match team leader and team member responses. Following the commonly used back translation procedure, the scales were translated from English into Chinese and then back translated into English by two independent bilingual individuals to ensure equivalency of meaning (Brislin, 1980). The respondents were informed that the survey aimed to examine human resource practices and assured of the confidentiality of their data. Each respondent placed his or her completed survey in a sealed envelope and returned it to a box in the human resource department in each company.

To reduce the potential common method bias, three waves of data collection from two different sources took place over eight months (Podsakoff et al., 2003). In wave one, we administered questionnaires to all 866 team members, in which respondents were asked to give their demographic information and perceptions of empowering leadership and team task complexity. We received usable survey responses from 729 team members affiliated with 110 teams, with a response rate of 84.2%. Four months later, in wave two, questionnaires were distributed to the targeted 729 team members asking them to evaluate their team learning behavior and team creative efficacy. As 10 team members had either changed their team affiliation or left the company, only 719 team members received the questionnaires. We received usable survey responses from 628 team members affiliated with 107 teams, with a response rate of 87.3%. After deleting those teams with fewer than three team members, 105 teams with 624 team members remained in the wave-two sample. All remaining participants were affiliated with the same team across times one and two. Finally, four months after the second-wave survey, in wave three, questionnaires were distributed to the 105 external team leaders asking them to rate team creativity and provide information on their own and team demographics. We confirmed with the human resource departments in all three companies that none of the teams had had a change in leader across times one, two, and three. One hundred and two team leaders returned the questionnaires. After matching leader-team pairs and deleting eight individuals who had either changed their team affiliation or left the company during the last four months, the final sample of this study consisted of 102 teams, which included 102 team leaders and 598 team members. All participants in the final sample, including team leaders and members, were affiliated with the same team across times one, two, and three.

Among the 102 teams, the average number of team members was 7.65 (s.d. = 2.07) and the average duration of the team (team age) was 7.58 years (s.d. = 2.26). Among the 102 external team leaders, 80.4 % were men, the average age was 36.06 years (s.d. = 5.82), and all held a bachelor’s degree or above (51.0%, bachelor’s degree; 49.0%, postgraduate degree). Among the 598 team members, 65.1% were men, the average age was 32.89 years (s.d. = 4.98), and most held a bachelor’s degree or above (2.5%, high school degree; 29.1%, associate’s degree; 36.6%, bachelor’s degree; 31.8%, postgraduate degree).

Measures

Six-point Likert-type scales ranging from 1 “strongly disagree” to 6 “strongly agree” were used for all measures in this study, except for team creativity.

Empowering Leadership

Arnold et al. (2000) constructed and empirically validated a scale for measuring empowering leadership in teams that has five dimensions: leading by example, participative decision making, coaching, informing, and showing concern/interacting with the team. Following Srivastava, Bartol, and Locke (2006), we adopted three items for each of the five dimensions as listed in Arnold et al. (2000). Confirmatory factor analysis (CFA) showed that the fit indices for a single second-order factor fell within a good range: χ² (85) = 154.04, p ≤ 0.01; comparative fit index (CFI) = 0.96, Tucker-Lewis Index (TLI) = 0.96; and root mean square error of approximation (RMSEA) = 0.09.

Team Learning Behavior

A four-item scale developed by Van der Vegt and Bunderson (2005) was used to measure team learning behavior.

Team Creative Efficacy

We used the three-item scale developed by Shin and Zhou (2007) to measure team creative efficacy.

Team Task Complexity

Following Gladstein (1984) and Man and Lam (2003), team task complexity was measured using a three-item scale.

Team Creativity

We used the four-item scale developed by Shin and Zhou (2007) to measure team creativity in R&D settings. The team leader completed the scale, which ranged from 1 “poor” to 6 “very much.”

Control Variables

Following previous research (e.g., Hülsheger, Anderson, & Salgado, 2009; Shalley, Zhou, & Oldham, 2004; Shin & Zhou, 2007), we controlled for several key variables that are likely to be associated with team processes, team emergent states, and team outcomes. First, we controlled for team age diversity, gender diversity, educational background diversity, and team member mean education level. Data on age (years), gender (0 = male; 1 = female), educational background (1 = Arts; 2 = Law; 3 = Humanities; 4 = Sciences; 5 = Engineering; 6 = Agriculture; 7 = Medicine; 8 = Business and Economics), and education level (1 = high school degree or below; 2 = associate’s degree; 3 = bachelor’s degree; 4 = postgraduate degree) were obtained from each team member. Following Allison (1978) and Cannella, Park, and Lee (2008), we measured age diversity using the coefficient of its variation (standard deviation divided by the mean). As gender and educational background are categorical variables, following Shin and Zhou (2007) and Wiersema and Bantel (1992), we measured gender and educational background diversity using Blau’s (1977) index of heterogeneity calculated as 1-Σ Si², where Si is the portion of a team’s members in the ith category. The team member mean education level was obtained by averaging the education level scores of team members.

Second, we controlled for team size (the number of members in a team), team age (years since the team was formed), and team interdependence to parcel out their potential influence on the team process, the emergent state, and the outcome (Hülsheger, Anderson, & Salgado, 2009). Team interdependence is a defining team characteristic (Barrick, Bradley, & Colbert, 2007). As a team-level input (Mathieu et al., 2008), team interdependence describes the extent to which team members cooperate and work interactively to accomplish team tasks (Steward & Barrick, 2000). It consists of task, goal, and outcome interdependence (Campion, Medsker, & Higgs, 1993), each of which is a key determinant of team effectiveness (for a review, see Gully et al., 2002). Team interdependence was measured using a nine-item scale developed by Campion and colleagues (1993).

Third, as our sample was taken from three companies, we created two company dummies and added them as control variables in the regression model to rule out the potential company effect. Finally, we controlled for leader education level and uncertainty avoidance orientation to rule out the potential confounding effect of these leader differences on team learning behavior, team creative efficacy, and team creativity. Uncertainty avoidance orientation reflects the willingness of people to take risks and accept uncertainty (Hofstede, 1980), and thus may influence a team leader’s willingness to support team members to engage in creative activities. Uncertainty avoidance was measured using a five-item scale developed by Dorfman and Howell (1988).

Results

Preliminary Analyses

Attrition Analysis

Because data were collected from team members in two waves, we followed Goodman and Blum’s (1996) approach to test for systematic differences in responses between the first and the second wave of data collection. The results indicated that there were no significant mean differences among these variables.

Confirmatory Factor Analysis

As the data on empowering leadership, team task complexity, team learning behavior, team creative efficacy, and team interdependence were collected from the same source (team member), we conducted CFA at the team level to evaluate the distinctiveness of these variables. For each of the single dimensional constructs, we created a first-order construct using the aggregated items as indicators. For empowering leadership, we first created five first-order constructs and each construct consists of three aggregated items. We then created a second-order construct representing empowering leadership and linked it to the five first-order constructs. All indicator means at the team level were then included in the CFA model. We first examined a five-factor model, which included empowering leadership, team task complexity, team learning behavior, team creative efficacy, and team interdependence. The model yielded an acceptable fit to the data: χ² (509) = 764.60, p ≤ 0.01; CFI = 0.94; TLI = 0.94; RMSEA = 0.07; and Akaike’s information criterion (AIC) = 936.60. In addition, all the factor loadings were significant, which indicates convergent validity.

The discriminant validity of the five constructs was tested by contrasting the five-factor model with one-factor and four-factor models. The one-factor model was obtained by loading all items measured into a “grand” latent factor. The four-factor model was obtained by combining team learning behavior and team creative efficacy into one-factor while keeping the other three constructs separate. We combined these two variables because correlation analysis showed that their correlation was the highest among the five constructs (r = 0.52, p ≤ 0.01). The one-factor and four-factor models yielded a poor fit to the data: χ² (527) = 3095.43, p ≤ 0.01; CFI = 0.41; TLI = 0.38; RMSEA = 0.22; and AIC = 3231.43 for the former and χ² (513) = 986.62, p ≤ 0.01; CFI = 0.89; TLI = 0.88; RMSEA = 0.10; and AIC = 1150.62 for the latter model. Chi-square difference and AIC difference tests indicated that the hypothesized five-factor model fit the data better than either the one-factor model (Δχ² (18) = 2330.83, p ≤ 0.01; ΔAIC = 2294.83) or the four-factor model (Δχ² (4) = 222.02, p ≤ 0.01; ΔAIC = 214.02). Thus, the discriminant validity of these constructs was confirmed.

Level of Analysis and Data Aggregation

The hypothesized model was tested at the team level. Although the scales comprised team-level items, we empirically examined the appropriateness of aggregating the responses of individual team members by determining the level of inter-member agreement (r_wg) and the intra-class correlation coefficients, ICC (1) and ICC (2) (Bliese, 2000; James, Demaree, & Wolf, 1984). The results are presented in Table 6.1.

As shown in Table 6.1, all teams had r_wg values equal or greater than 0.70 for empowering leadership (100%), with a median r_wg of 0.94, for team task complexity (100%), with a median r_wg of 0.93, and for team task interdependence (100%), with a median r_wg of 0.95. All teams except one had r_wg values equal or greater than 0.70 for team learning behavior (99%), with a median r_wg of 0.94, and all teams except two had r_wg values equal or greater than 0.70 for team creative efficacy (98.1%), with a median r_wg of 0.95. The levels of inter-member agreement within teams thus justify our aggregation of individual-level data. In addition, the results showed that the ICC (1) values ranged from 0.15 to 0.25 (0.15 for empowering leadership, 0.18 for team task complexity, 0.18 for team learning behavior, 0.24 for team creative efficacy, and 0.25 for team interdependence, values greater than the conventional cutoff value of 0.05), and all were significant, indicating that between-team variance was greater than within-team variance (James, 1982). Finally, the ICC (2) values were 0.50 for empowering leadership, 0.57 for team task complexity, 0.56 for team learning behavior, 0.64 for team creative efficacy, and 0.66 for team interdependence. All of these values are higher than the conventional cutoff value of 0.50 (James, 1982). These results indicate that it was appropriate to average the scores provided by team members to obtain team-level scores.

Descriptive Statistics

Table 6.2 presents the means, standard deviations, zero-order Pearson correlations, and scale reliabilities for all key variables in this study.

Bracketed values on the diagonal are the Cronbach’s alpha value of each scale at the team level.

Hypothesis Testing

We conducted hierarchical multiple regression analysis (Baron & Kenny, 1986) to test Hypotheses 1 and 2. As shown in Table 6.3 (see Model 10), empowering leadership was positively related to team creativity (β = 0.24, p ≤ 0.05, ∆R² = 0.05); thus, Hypothesis 1 was supported.

Hypothesis 2 predicted that team learning behavior and team creativity efficacy mediate the positive relationship between empowering leadership and team creativity. The results in Table 6.3 show that (1) empowering leadership was positively related to team learning behavior (β = 0.43, p ≤ 0.01, ∆R² = 0.15, Model 2) and team creative efficacy (β = 0.37, p ≤ 0.01, ∆R² = 0.11, Model 6), (2) empowering leadership was significantly related to team creativity (β = 0.24, p ≤ 0.05, ∆R² = 0.05, Model 10), (3) team learning behavior (β = 0.28, p ≤ 0.05, ∆R² = 0.14, Model 11) and team creative efficacy (β = 0.25, p ≤ 0.05, ∆R² = 0.14, Model 11) were positively related to team creativity, and (4) the relationship between empowering leadership and team creativity became nonsignificant (β = 0.02, n.s., Model 11) when both mediators were present. Thus, team learning behaviors and team creative self-efficacy were full mediators in the relationship between empowering leadership and team creativity, thus supporting Hypothesis 2.

To examine the moderating and mediated moderating effects predicted by Hypotheses 3 and 4, respectively, we adopted the moderated causal steps approach to regression analysis (Muller, Judd, & Yzerbyt, 2005). All interaction variables were mean centered to minimize multicollinearity (Aiken & West, 1991).

Hypotheses 3a and 3b predicted that team task complexity moderates the relationship between empowering leadership and team learning behavior, and between empowering leadership and team creative efficacy, respectively. As shown in Table 6.3, the interaction between empowering leadership and team task complexity was positively related to team learning behavior (β = 0.29, p ≤ 0.05, ∆R² = 0.05, Model 4) and team creative efficacy (β = 0.24, p ≤ 0.05, ∆R² = 0.04, Model 8). We plotted the interaction using Aiken and West’s (1991) procedure: compute slopes one standard deviation above and below the mean of the moderating variable. Figures 6.1 and 6.2 show that the interaction patterns are consistent with the hypothesized relationships. Specifically, the slope test results show that empowering leadership had a positive and stronger effect on team learning behavior when the level of team task complexity was high (β = 0.66, p ≤ 0.01), but its effect on team learning behavior was weaker when the level of team task complexity was low (β = 0.13, n.s.). Similarly, empowering leadership had a positive and stronger effect on team creative efficacy when the level of team task complexity was high (β = 0.56, p ≤ 0.01), but its effect on team creative efficacy was weaker when the level of team task complexity was low (β = 0.12, n.s.). In addition, we performed moderated path analysis (Edwards & Lambert, 2007), bootstrapping 1,000 samples to compute bias-corrected confidence intervals. The results presented in Table 6.4 support first-stage moderating effects (∆β = 0.80, p ≤ 0.05 for team learning behavior; ∆β = 0.69, p ≤ 0.05 for team creative efficacy). Thus, Hypotheses 3a and 3b were supported.

The mediated moderations were tested further based on Muller and colleagues’ (2005) procedures. Supporting Hypothesis 4, the results showed that (1) empowering leadership × team task complexity was significantly related to both mediators as indicated earlier for Hypotheses 3a and 3b; (2) empowering leadership × team task complexity was significantly related to team creativity (β = 0.26, p ≤ 0.05, ∆R² = 0.04, Model 13); (3) after controlling for the interactions between team task complexity with the two mediators and other predictors, team learning behavior (β = 0.26, p ≤ 0.05, ∆R² = 0.12, Model 14) and team creative efficacy (β = 0.25, p ≤ 0.05, ∆R² = 0.12, Model 14) were positively related to team creativity; and (4) the interaction effect of empowering leadership and team task complexity on team creativity became nonsignificant after entering the two mediators and controlling for the interactions between team task complexity with the two mediators and other predictors (β = 0.12, n.s., Model 14). Therefore, the interactive effect of empowering leadership and team task complexity on team creativity was fully mediated by team learning behavior and team creative efficacy. We plotted the interactive effect of empowering leadership and team task complexity on team creativity and conducted simple slope tests. Figure 6.3 shows that the interaction pattern is consistent with our arguments; that is, empowering leadership had a positive and stronger effect on team creativity when the level of team task complexity was high (β = 0.35, p ≤ 0.01), but its effect on team creativity was weaker when the level of team task complexity was low (β = −0.01, n.s.).

To provide more robust support for our results, we performed moderated path analysis (Edwards & Lambert, 2007), bootstrapping 1,000 samples to compute bias-corrected confidence intervals. An indirect effect is considered to be statistically significant if the 95% confidence interval excludes zero. Our results, which are summarized in Table 6.4, show that the size of the indirect effect of team learning behavior and team creative efficacy was 0.42 and 0.35, respectively. The 95% confidence intervals computed using the bootstrap estimates exclude zero. Therefore, the indirect effects of team learning behavior and team creative efficacy were significant, providing further support for Hypotheses 2 and 4.

Discussion

We theorized, and found, that empowering leadership positively influences creativity in R&D teams through team learning behavior and team creative efficacy. In addition, when the level of team task complexity is high, empowering leadership is found to have a stronger, positive relationship with team learning behavior and team creative efficacy, which in turn, are both positively related to team creativity.

Theoretical Implications

Our research makes three distinct contributions. First, we have built and empirically tested a conceptual model that integrates empowering leadership theory with creativity theories at the team level. Our results empirically demonstrate for the first time that the empowering behavior of team leaders effectively promotes creativity in R&D teams. Due to the changes in communication technology and the social dynamics of a new generation of young professionals, it has been argued that most teams today are to some degree virtual (Kirkman & Mathieu, 2005). As a result, delegating authority to and enhancing the motivation of team members who interact with each other virtually might be different from the influence of empowering leadership on R&D teams discussed in this study. Because the theoretical framework built in this study can be broadly applied to other job categories (e.g., service), it is thus interesting to extend the current research model to other types of team to investigate.

Second, building on Amabile’s (1983, 1996) componential model and Amabile and Pratt’s (2016) dynamic componential theory of creativity, our study contributes to the empowering leadership and creativity literatures by theorizing and empirically testing that both team learning behavior (the team process) and team creative efficacy (the team emergent state), representing a team’s learning of domain-relevant and creativity-relevant skills and team intrinsic motivation, are necessary building blocks for enhancing team creativity. In this way, we extend Amabile’s solid componential model from individual employees to teams, thus building a theoretical framework that help bridge empowering leadership with creativity in teams.

Finally, based on the contingency perspective of leadership (Yukl, 2013), we further established the relationship between empowering leadership and team creativity by identifying the boundary condition. Our results are consistent with and provide empirical support for the contingency perspective. That is, leadership has a greater impact on team processes, emergent states, and outcomes when teams face demanding situations (e.g., high level of task complexity), which require team leaders to play a greater role (Morgeson, 2005).

Managerial Implications

Our theoretical model and empirical findings have important implications for managers. As an increasing number of organizations rely on team creativity to maintain competitive advantages in the face of global competition (Shin & Zhou, 2007), it is important for companies to keep in mind that the empowering behavior of team leaders can promote team creativity. Empowering leadership not only facilitates interaction and cooperation among team members in acquiring task-relevant knowledge (i.e., team learning behavior) but also creates favorable conditions that promote the shared belief of team members in their ability to produce something new and useful (i.e., team creative efficacy). Thus, those organizations and teams whose survival and competitiveness depend extensively on creative outcomes may benefit from leader selection and training to ensure that team leaders demonstrate empowering behaviors including leading by example, coaching, facilitating participative decision making, informing, and showing concern.

The study findings also indicate that managers need to consider task complexity when empowering teams. Specifically, managers may find that empowering behaviors are more effective in facilitating the team learning process and boosting team creative efficacy when teams face complex, as opposed to simple, tasks. Highly complex tasks require a high level of team collaboration to refine task-relevant knowledge and increase the inherent interest of team members in the task. Thus, when the level of team task complexity is high, leaders may need to engage in more empowering behaviors to help teams build stronger collaborations and to enhance the shared belief of team members in their creative capabilities.

Conclusion

This study integrates empowering leadership and creativity theories to further develop and test a theoretical model that connects empowering leadership and team creativity. Including both team learning behavior (the team process) and team creative efficacy (the team emergent state) in the model, we find that they fully mediate the influence of empowering leadership on team creativity. Furthermore, the moderating effect of team task complexity helps identify a critical boundary condition for the aforementioned relations. We hope that our study will stimulate future research to advance the theoretical understanding of how empowering leadership can promote team creativity in the workplace.