Leader–member exchange differentiation and team creativity: An investigation of nonlinearity

Abstract

Although substantial knowledge regarding the antecedents and outcomes of leader–member exchange (LMX) differentiation has been accumulated, numerous questions related to this topic remain underexplored. To enhance the understanding of LMX differentiation and team-focused outcomes, this study proposed that LMX differentiation has a curvilinear relationship with team creativity and that team LMX quality (represented by the LMX median in this study) moderates the association between these two variables. An investigation based on 59 teams from multiple Chinese companies was conducted. The results indicated that LMX differentiation has an inverted U-shaped relationship with team creativity, and LMX median moderates the inverted U-shaped relationship. Specifically, for teams with a low LMX median, the curvilinear relationship is stronger, whereas for teams with a high LMX median, the slope of the curve becomes nearly flat, thus losing the inverted-U effect. Theoretical and practical implications are discussed, and directions for future research are outlined.

Keywords

leader–member exchange LMX LMX differentiation LMX median team creativity

Introduction

With the advent of the fourth stage of the development of leader–member exchange (LMX) theory (Graen and Uhl-Bien, 1995), LMX differentiation―the development of differential relationships between a leader and the followers who report directly to the leader in a work group (Henderson et al., 2009; Liden et al., 2006)―has received increasing attention (Harris et al., 2014; Le Blanc and González-Romá, 2012; Liao et al., 2010; Nishii and Mayer, 2009; Stewart and Johnson, 2009). LMX differentiation represents the degree of variability in the quality of the exchange relationships (Erdogan and Liden, 2002). High LMX differentiation within a work group indicates a broad range of overall LMX quality, whereas low LMX differentiation implies a small range of overall LMX quality within a work group (Henderson et al., 2009). Much knowledge regarding the antecedents and outcomes of LMX differentiation has been accumulated (Harris et al., 2014; Ma and Qu, 2010; van Breukelen et al., 2012), and it has generally been argued that LMX differentiation affects individual- and team-level outcomes (Henderson et al., 2008; Liden et al., 2006).

Although substantial progress has been made in research on LMX differentiation, certain questions related to this topic remain underexplored. First, extant research has primarily focused on the effect of LMX differentiation on individual-level outcomes (e.g. employee performance, organisational citizenship behaviour (OCB) and psychological fulfillment). Less research has examined the effect of LMX differentiation on team-level outcomes (Nadioo et al., 2011). Second, the few studies on the role of LMX differentiation at the team level have yielded inconsistent results (Le Blanc and González-Romá, 2012). Some researchers propose that LMX differentiation positively affects team performance because the leader can strategically assign the most challenging tasks to the employees who are most capable of performing them successfully (Dansereau et al., 1975; Nadioo et al., 2011). However, emergent evidence suggests that LMX differentiation introduces relational boundaries, violates the principle of equality and disrupts team coordination, which will diminish team performance (Li and Liao, 2014). In light of these conflicting findings, Le Blanc and González-Romá (2012) called for future research to better understand LMX differentiation and its effect.

In response to their call, this study aims to examine the effect of LMX differentiation on team creativity. We focus on team creativity for two main reasons: (a) creativity is a widely studied behavior that is essential to organizational viability and success (Zhang et al., 2011; Zhou et al., 2009), and teams are increasingly valued for their potential to demonstrate creative performance (Leenders et al., 2003; Tsai et al., 2012); (b) LMX, as an important social exchange relationship embedded in a work team, may influence the context within which creativity can occur (Shalley and Gilson, 2004). LMX differentiation is the core element of the LMX model (Graen and Uhl-Bien, 1995; Henderson et al., 2009). However, it is surprising that existing research on leadership and team creativity has primarily focused on different leadership styles (e.g. transformational leadership, ethical leadership or abusive supervision) or the LMX model. Despite its importance, LMX differentiation has been notably absent from the literature. Given the inconsistent results regarding LMX differentiation and team-focused outcomes, we intend to use the too-much-of-a-good-thing (TMGT) metaprinciple (Pierce and Aguinis, 2013) as an overarching framework to address the subject. We argue that the relationship between LMX differentiation and team creativity may reach an inflection point, after which the relationship becomes negative. Furthermore, to answer the call for research exploring the relationship between LMX differentiation and median LMX quality (an indicator to depict a team’s score on LMX quality) as joint determinants of team outcomes (Henderson et al., 2009), we also examine the potential moderating role of the LMX median in the relationship between LMX differentiation and team creativity.

By examining whether LMX differentiation influences team creativity and if so how this effect occurs, we attempt to make two contributions to the literature. First, we make a contribution to the LMX differentiation and creativity literature. On the one hand, examining team creativity from the perspective of LMX differentiation extends current knowledge on the influence of leadership on team creativity. On the other hand, prior research has employed a linear framework when examining the effect of LMX differentiation on team-focused outcomes and obtained inconsistent results, which have puzzled scholars to some extent. The TMGT perspective allows for two possibilities, and considering an inverted U-shaped relationship between LMX differentiation and team creativity may help to explain the initially inconsistent effects of LMX differentiation on team-level outcomes observed in previous research. Second, to deepen the understanding of the potential boundary conditions related to this nonmonotonic association, we examine the moderating role of the LMX median in the curvilinear LMX differentiation-team creativity relation. We theorize that the LMX median diminishes the likelihood that a team will fall prey to the potentially debilitating negative consequences of LMX differentiation. This interaction analysis offers important insight into the effect of LMX-related constructs on team creativity.

Literature review and hypotheses

LMX differentiation and team creativity

Team creativity, as one type of team output, refers to ‘the product of novel and useful ideas concerning products, services, processes, and procedures by a group of employees working together’ (Shin and Zhou, 2007). Team output is dependent on individuals, and team members with greater ability can yield higher team outcomes (Barrick et al., 1998). Taggar (2002) demonstrated that aggregated peer ratings of group members’ creativity were predictive of externally rated group creativity. Team creativity is related to task type, and the ways in which team members contribute to team creativity vary with the task type. Pirola-Merlo and Mann (2004) distinguished between: (a) team creativity as an additive type of task, where each member’s creativity adds to that of the team; (b) team creativity resembling a disjunctive type of task, where the most creative ideas (which may be produced by a particular individual) are adopted by the team and determine team creativity; and (c) team creativity lying somewhat between these two types, where each member makes a contribution but the importance of the contribution to team creativity is weighted in some way (e.g. the most creative member’s contribution is the most important). Irrespective of the types considered, it is uncontroversial that individual creativity, as a critical premise, provides the raw material for novel and useful ideas. In other words, team creativity is influenced and to some extent determined by individual creativity. Therefore, to enhance team creativity, it is very important for team leaders to encourage every team member to perform creatively.

LMX has been demonstrated to positively relate to employees’ creativity (Shalley and Gilson, 2004; Volmer et al., 2012). However, in many situations, owing to limited resources and time, leaders cannot develop high-quality relationships with every team member. How these limited resources are allocated among team members to some extent determines the team’s achievements. Team creative performance is likely to be improved through effective task allocation, task variety and careful planning, as well as the coordination of diverse efforts (Brophy, 1998). Team members have different knowledge, skills and perspectives (Lovelace et al., 2001), and different configurations of members’ attributes within a team may influence team creativity (Bell, 2007). Different levels of fit between tasks and personnel will produce different outcomes. Highly creative members can help the team develop attitudes and behaviors such as persistence or tolerance of ambiguity in the work environment that are crucial to promoting team creativity (Somech and Drach-Zahavy, 2013). Apart from creative members, extraverted, conscientious and agreeable group members are likely to display more intragroup process behavior (i.e. team creativity-relevant processes, such as stimulating discussion, inspiring group members, keeping a team focused on a task and cooperating with other team members: Taggar, 2002). Therefore, when time and resources are restricted, leaders can assign these employees more challenging or more critical work tasks and give them more trust, more encouragement, high expectations and strong social emotional support (i.e. forming a high-quality LMX relationship). With respect to other members (e.g. low-ability or poorly performing subordinates), leaders may assign such members more menial work tasks and provide them with a lower level of support (Maslyn and Uhl-Bien, 2001). The more clearly leaders recognize subordinates’ competency, the better the team works (Liden et al., 2006), and the higher the subsequent team creativity. Based on these arguments, we propose that leaders can benefit from developing differentiated relationships with followers according to their different attributes (i.e. forming high- and low-quality exchanges with subordinates) and facilitate team creativity.

Although individual creativity is important for team creativity, team creativity is not entirely determined by individual creativity; rather, it emerges synergistically when members interact in certain ways. Groups perform best when they contain creative group members and effective team creativity-relevant processes. Taggar (2002) suggested that team creativity is not the simple sum of individual creativity, and emphasized the importance of team creativity-relevant process (e.g. inspiring group members to elevate their goals; providing feedback; organizing and coordinating contributions; and eliciting and appreciating different ideas, needs and viewpoints). Richter et al. (2012) highlighted the importance of teams as informational resources (shared ‘knowledge of who knows what’ and functional background diversity) for employee creativity. Differentiation can help optimize the distribution and utilization of time and related material resources; however, this does not imply that the higher the differentiation the better. The meta-theoretical principle of the TMGT effect suggests that an initially beneficial predictor variable reaches an inflection point after which its relationship with a certain outcome becomes negative (Pierce and Aguinis, 2013). High LMX differentiation may lead to in-group and out-group perceptions between group members in similarly high- and low-quality LMX relationships (Ford and Seers, 2006; Schyns, 2006) and engender interpersonal and emotional conflicts, which may absorb energy and distract team members from performing their tasks (Hooper and Martin, 2008). Li and Liao (2014) argued that a high level of LMX differentiation violates the principle of equality, disrupts interpersonal harmony and destroys team coordination. In such a situation, the presence of highly creative individuals may not ameliorate the negative effects of a scarcity of inspiration motivation, organization, coordination or individualized consideration (Taggar, 2002). Team failure may reside in not only particular members’ lack of ability but also the members’ collective failure to coordinate and synchronize their individual contributions (Zaccaro et al., 2001). The negative consequences of an emotional relationship among team members resulting from high LMX differentiation will harm team creativity-relevant processes, which will ultimately harm team creativity.

Following this line of inquiry, the present research assumes that the relationship between LMX differentiation and team creativity is neither positive nor negative. Instead, it proposes that an inverted U-shaped relationship exists between them. That is, LMX differentiation can optimize limited resources, and thus an increase in differentiation can be beneficial for team creativity, but not beyond a certain level. After the optimum level of differentiation, or the inflection point reflecting ‘too much of a good thing’ (Pierce and Aguinis, 2013) is reached, team creativity will begin to decline. Therefore, we propose the following hypothesis:

Hypothesis 1: LMX differentiation has an inverted U-shaped relationship with team creativity.

The moderating role of the LMX median

As discussed above, the TMGT metaprinciple explains the curvilinear relationship between LMX differentiation and team creativity. Whether LMX differentiation has a positive or negative effect on team creativity depends on the extent of differentiation. When differentiation is within a certain range, it can lead to better role differentiation, but if it is too high, relational problems will arise. Here, implicit are the limitation on resources and the differences in distribution. The limitations on resources highlight the effect of LMX differentiation. Then, one noticeable question is whether the effect of differentiation (i.e. enhancing team creativity by providing in-group members with more resources and support) is salient when resources are rich and when most members have high exchange relationships with the leader. Liden et al. (2006) argued that when most members have high LMXs, they do not pay much attention to the distribution of resources. In other words, team members will not be particularly concerned about relationship differentiation. This implies that LMX quality is a contingent condition for the effect of LMX differentiation. To examine this viewpoint, we add LMX quality into the differentiation equation and propose that the effect of LMX differentiation on creativity depends on LMX quality. Specifically, we propose that the curvilinear relationship between LMX differentiation and team creativity is stronger for teams with a low LMX median.

Here, one notable issue is the measure of group LMX. The LMX mean and LMX median are typically considered to depict a team’s score on LMX quality. Certain researchers employ the former (e.g. Boies and Howell, 2006; Nishii and Mayer, 2009), whereas others employ the latter (e.g. Henderson et al., 2009; Le Blanc and González-Romá, 2012; Liden et al., 2006). Selecting between the two measures should be done with caution. Although the mean is widely used when calculating group-level variables in the literature, it is not applicable to all variables. Kozlowski and Klein (2000) argued that certain variables (e.g. team member abilities and demographic characteristics) can be aggregated to the group level to represent the overall tendency within the group, but agreement in these characteristics is not expected. Scholars may use the minimum, maximum, or measures of dispersion to calculate the aggregation of such characteristics, depending on the theory in question. Kozlowski and Klein (2000) contended that without a measure of agreement, the mean might not be an appropriate summary statistic because of its ambiguity and equifinality. As LMX is a variable for which within-group agreement is not expected and for which the median is a better measure than the mean or average, we used the median to represent the central tendency among LMX relationships within teams in the present study.

As discussed above, in teams with a high LMX median, most members receive resources, confidence and support from the leader, which can enhance their creativity. Their creativity consequently adds to team creativity. In these teams, members’ satisfaction with their relationships and with the support received from their leader may guarantee high levels of achievement regardless of differentiation (Graen and Uhl-Bien, 1995). The benefits of LMX differentiation then become marginal. In other words, in such teams, differentiation by the leader is not likely to elicit any additional effort from team members (Le Blanc and González-Romá, 2012). Furthermore, in these teams, because many members are considered the trusted assistants of the leader (Liden et al., 2006), status differentiation becomes unclear (Stewart and Johnson, 2009), which may minimize the likelihood of in- and out-group formation and reduce the risk of interpersonal conflicts. Therefore, we argue that the effect of LMX differentiation on team creativity becomes weaker when a team’s LMX median is high.

By contrast, when the LMX median is low, that is, when leaders do not or are unable to establish high LMXs with many members, LMX differentiation will have a stronger effect on team creativity. First, when the LMX median and LMX differentiation are low, that is, when the leader has low-quality LMX relationships with all followers (Liden et al., 2006), there is large relational distance between the team leader and the team members, and the team members will receive little support from the team leader. Creativity processes are highly unpredictable and replete with uncertainty, paradoxes and tensions (Baer, 2012; Rosing et al., 2011), and hence, performing creatively at work requires adequate resources (e.g. time, funding and assistance) from leaders or organizations to address difficulties and pressures. Positive leader behaviors (i.e. transformational leadership) enhance team creativity by providing support for creativity (Eisenbeiss et al., 2008). Low LMX differentiation combined with a low LMX median implies an absence of the resources and support necessary for creativity and, consequently, a low level of team creativity. When LMX differentiation is introduced into a team with a low LMX median, certain members (e.g. highly creative subordinates or those with strong needs for growth) can acquire relatively abundant resources, support and confidence from leaders to perform creatively. As we argued above, team creativity is determined by individual creativity, to some extent. When certain members’ creativity is improved through LMX differentiation, their creativity will consequently contribute to team creativity. However, when LMX differentiation is very high, it has the potential to generate interpersonal and emotional conflicts (Hooper and Martin, 2008), which may absorb energy and distract team members from performing their tasks. In an influential review article, Hülsheger et al. (2009) argued that interpersonal conflicts undermine team functioning to the extent that anger and frustration impede effective communication within the team and reduce team members’ receptiveness to one another’s ideas. In other words, the disparity in status between ‘in-group’ and‘out-group’ members resulting from a low LMX median combined with high LMX differentiation may place team members in a disadvantageous environment replete with envy, distrust and suspicion, which will ultimately impair team creativity. In summary, we propose the following hypothesis:

Hypothesis 2: The LMX median moderates the inverted U-shaped relationship between LMX differentiation and team creativity, such that the curvilinear relationship is stronger for teams with a low LMX median.

Methods

Data collection

We collected data from nine high-technology companies in China through our social network. These studied companies were engaged in software development (N = 6) and network technology (N = 3) in two cities (Beijing and Wuhan). The average age of the companies was 7.11 years (standard deviation = 3.14). We first obtained the permission and support of each company’s director to collect the data, and we then met with the companies’ personnel directors. The personnel directors helped us to contact team supervisors. Considering that team creativity is the criterion variable of the current study, we focused on research and development and marketing teams. After agreeing to participate, each team supervisor received information about the study’s objectives and procedure and then distributed the questionnaires to team members.

To reduce the potential problem of common method variance (Podsakoff et al., 2003), we collected data from different sources. Specifically, each subordinate completed a questionnaire regarding the relationship with his/her leader and his/her personal information, whereas supervisors rated team creativity. We used a code on the questionnaire to identify team membership. In addition, to avoid the problem of social desirability and to reduce respondents’ apprehension about the evaluation, we included an envelope with double-sided tape and a cover letter in each questionnaire explaining the purpose of the research and emphasizing that all responses would be anonymous and confidential. Furthermore, we noted that all competed questionnaires should be put into the envelope, which should be sealed.

A total of 73 teams participated in the survey. Considering the importance of the team level in the present study, we eliminated teams in which the majority of members did not respond and included teams in which two out of three members responded; teams with fewer than three members were excluded from the analysis. The average response rate within the teams was 75.0 percent. As a result of these criteria, the number of teams that were eligible for inclusion in the study was 59, consisting of 276 employees. Actual team sizes ranged between 4 and 12, with a mean of 6.36. Among the final sample, the team supervisors were primarily men (66.1%) with an average age of 35.22 years. On average, each team supervisor had worked for the company for 6.13 years, and the range of tenure spanned from one to 14 years. Team supervisors were embedded within their teams as direct leaders of and active participants in their respective teams. Team members had worked with their teams for an average of 2.41 years, with the range of tenure spanning from 0.5 to nine years, and they were between 22 and 45 years old, with an average age of 29.41 years. Most team members had a college education. The average employee–supervisor relationship length was 2.38 years. Finally, 61.2 percent of team members were men.

Measures

All the scales we used in the survey were originally written in English and were translated into Chinese. To ensure equivalence of the measures written in Chinese and English, we used a standard translation and back-translation procedure (Brislin, 1970). Unless otherwise indicated, the response options for all the measures were the same, ranging from 1, ‘strongly disagree’, to 6, ‘strongly agree’.

Team LMX quality

We measured LMX quality by using an adapted version of the LMX-7 scale (Scandura and Graen, 1984), which is widely used in the LMX literature (Dulebohn et al., 2012; Nishii and Mayer, 2009). Sample items are ‘My supervisor understands my problems and needs’ and ‘I feel that my supervisor recognizes my potential’. Cronbach’s alpha for this scale was .89. As discussed above, LMX is a variable for which within-group agreement is not expected, and the median is a better measure than the mean or average. Thus, we used the median to represent the central tendency among LMX relationships within teams.

LMX differentiation

Consistent with previous research, we calculated LMX differentiation by using the within-group standard deviation (Erdogan and Bauer, 2010; Harris et al., 2014; Henderson et al., 2008). A higher within-team standard deviation indicates higher differentiation.

Team creativity

Consistent with previous studies on team creativity (e.g. Gilson and Shalley, 2004; Tu, 2009), we adapted the following three items from Zhou and George’s (2001) measure of individual creativity to the team level: ‘Team comes up with new and practical ideas in solving problems’, ‘Our team easily develops new ways and procedures related to the task’, and ‘Confronting problems, our team generates creative solutions’. Cronbach’s alpha was .91. Team creativity was reported by team leaders, who are viewed as reliable sources of team information (Akgün et al., 2008; De Dreu and West, 2001; Pirola-Merlo and Mann, 2004).

Control variables

Considering that larger teams have been shown to be more likely to obtain resources for creativity (Stewart, 2006), we controlled for team size (i.e. total number of team members reported by leaders). In addition, because team diversity or heterogeneity may influence team creativity (e.g. Harrison et al., 2002; Shin and Zhou, 2007; Van der Vegt and Janssen, 2003), we controlled for teams’ demographic heterogeneity (e.g. education, tenure). Heterogeneity in tenure was calculated by using the coefficient of variation (standard deviation divided by the mean), and heterogeneity in education was computed by using Blau’s (1977) heterogeneity index, which is widely used in the heterogeneity literature (e.g. Baer et al., 2008; Polzer et al., 2002; Van der Vegt and Bunderson, 2005). Finally, organizational function was also controlled for (research and development = 0, marketing = 1).

Analytical strategy

Because the data pertained to the team level, we applied hierarchical multiple regression analyses to detect the main and interaction effects of LMX differentiation and the LMX median on team creativity. In the first step, all the control variables and LMX differentiation were entered into the equation. To assess the curvilinearity of the relationship, we added quadratic LMX differentiation into the model in the second step. Finally, to evaluate the moderating role of the LMX median on the relationship, we included the LMX median and its interaction with quadratic LMX differentiation into the model in the third step.

Results

Table 1 shows the descriptive statistics, reliabilities and correlations for the variables.

Table 1.

Team-level descriptive statistics, reliabilities and correlations among measures.

Variable	Mean	SD	1	2	3	4	5	6	7
1. Team size	6.36	1.75
2. Education heterogeneity	.38	.16	−.01
3. Team tenure heterogeneity	.37	.19	−.09	−.38
4. Organizational function	.46	.50	−.10	−.12	.16
5. LMX differentiation	.62	.34	.01	.05	.06	.18
6. LMX median	4.42	.65	.13	−.02	−.27*	.27	−.06
7. Team creativity	4.45	1.08	.06	−.04	−.11	.07	−.08	.45*	(.91)

N = 59 at team level. Reliabilities of the scales are noted in the diagonals. LMX = leader–member exchange; SD = standard deviation. *p < .05; two-tailed tests.

Curvilinear relationship between LMX differentiation and team creativity

Hypothesis 1 predicts that LMX differentiation has a curvilinear relationship with team creativity. As shown for Model 2 in Table 2, LMX differentiation is not significantly related to team creativity (β = .18, not significant), whereas quadratic LMX differentiation hasa negative effect on team creativity (β = –.39, p < .05). This pattern of results (a nonsignificant linear effect in the presence of a negative and significant curvilinear effect for centered explanatory variables) indicates that an inverted U-shaped relationship exists between LMX differentiation and team creativity (see Aiken and West, 1991), which lends support to Hypothesis 1 (see Figure 1). That is, the relationship between LMX differentiation and team creativity shows a positive, or upward, trend at lower levels of LMX differentiation and a negative, or downward, trend at higher levels of LMX differentiation.

Table 2.

Hierarchical regression results: The main and interactive effects of LMX differentiation and LMX median on team creativity.

Variables	Team creativity
	Model 1	Model 2	Model 3
1. Intercept	4.65**	4.76**	4.47**
2. Control and independent
Team size	.05	.04	.08
Education heterogeneity	−.08	−.06	−.02
Team tenure heterogeneity	−.15	−.13	−.05
Organizational function	.10	.14	.01
LMX differentiation	−.09	.18	.25
LMX differentiation²		−.39*	−.33*
LMX median			.20
LMX differentiation × LMX median			−.17
LMX differentiation² × LMX median			.41*
Overall R²	.04	.11	.36
∆R ²	.01	.07*	.25*

N = 59 at team level. LMX = leader–member exchange. **p < .01; *p. < .05; two-tailed tests.

Figure 1.

The curvilinear relationship between LMX differentiation and team creativity.

The contingent effect of the LMX median. Hypothesis 2 predicts that the LMX median moderates the curvilinear relationship between LMX differentiation and team creativity. To test this hypothesis, we introduced the LMX median as a moderator in the regression equation in Model 3 in Table 2. The results indicate that the LMX median significantly interacts with quadratic LMX differentiation (β = .41, p < .05). As the interaction plot in Figure 2 shows, the relation between LMX differentiation and team creativity follows an inverted U-shaped function for teams with a low LMX median, but becomes almost flat, indicating the lack of an inverted-U effect, for teams with a high LMX median. Overall, these results provide support for Hypothesis 2.

Figure 2.

The interactive effect of LMX differentiation and LMX median on team creativity.

To further analyze this quadratic interaction effect, we tested simple slopes of the regression curves. As suggested by Aiken and West (1991), we estimated simple slopes at three levels of LMX differentiation: low (one standard deviation below the maximum of the regression curve), intermediate (maximum of the regression curve), and high (one standard deviation above the maximum of the regression curve). The results indicate that when the LMX median is low the simple slope of the regression curve: has a positive, significant value for low differentiation (β = .76, p < .05); does not significantly differ from zero for intermediate differentiation (β = .37, p > .05); and has a significant negative value for high differentiation (β = –.03, p < .05). When the LMX median is high, the simple slopes of the regression line do not significantly differ from zero (ps >.05) at low, intermediate, or high levels of differentiation. Overall, these results provide support for Hypothesis 2.

Discussion

The purpose of this study was to examine the relationship between LMX differentiation and team creativity. Specifically, we examined whether the relationship between LMX differentiation and team creativity could be characterized by an inverted U-shape function and whether the LMX median moderates this relationship. Our results suggest that the relationship between LMX differentiation and team creativity is nonmonotonic and that an inflection point indeed exists. Furthermore, in accordance with our hypothesis, we found that the LMX median moderates the squared LMX differentiation term for team creativity, with LMX differentiation exerting a stronger effect on team creativity in teams with a low LMX median. By contrast, for teams with a high LMX median, the slope of the curve became nearly flat, thus indicating the lack of inverted-U effect. The findings of this research have some interesting theoretical and practical implications.

Theoretical implications

The study has significant implications for the LMX and creativity literature. First, and perhaps most importantly, our research demonstrates that LMX differentiation, as a team-level construct, has a curvilinear relationship with team creativity. Past work examining the influence of leadership on team creativity has primarily concentrated on different leadership styles or the LMX model, and little is known about how leaders’ differentiated treatment of subordinates affects team creativity. Van Breukelen et al. (2012) argued that attention should not focus exclusively on the general leadership style of the leader toward the team as a whole but that leaders’ specific behaviors toward various followers should also be considered. Furthermore, previous research has assumed a linear relationship between LMX differentiation and team-focused outcomes (e.g. Liden et al., 2006; Naidoo et al., 2011) but has failed to find a direct positive or negative relationship between LMX differentiation and team outcomes (Boies and Howell, 2006; Liden et al., 2006). We explicitly proposed and empirically tested the TMGT framework to explain the relationship between LMX differentiation and team creativity. Our results demonstrate that the relationship between LMX differentiation and team creativity is curvilinear and that teams are likely to exhibit the highest creativity at intermediate levels of differentiation. This finding contrasts with previous studies that contend that LMX differentiation has no main effect on team outcomes (Boies and Howell, 2006; Liden et al., 2006). Our findings suggest that focusing on a linear relationship will not capture the essence of LMX differentiation; instead, a more complex understanding of how LMX differentiation influences team-level outcomes should be developed. Therefore, our findings advance the understanding of the influence of leadership on creativity and enrich the literature on the construct of LMX differentiation.

Second, in an effort to derive a contingent condition of the curvilinear relationship, we examined the moderating effect of the LMX median on this relationship. In support of our argument, we found that the LMX median moderates the curvilinear relationship between LMX differentiation and team creativity. Furthermore, consistent with prior findings, we found that LMX differentiation has a stronger effect on team creativity in teams with a low LMX median. The difference between our results and those of prior studies is that we found that the association between LMX differentiation and team creativity can be described by an inverted-U shape. We further found that in teams with a high LMX median, the slope of the curve became nearly flat, thus losing the inverted-U effect. An explanation for this result may lie in the buffering effect of the LMX median. In teams with a high LMX median, as most members are treated as trusted assistants, they are not sensitive to differentiation. In addition, high LMX quality reduces the risk of interpersonal conflicts, and thus teams devote their energy to performing tasks. This suggests that LMX differentiation is not necessarily a concern if LMX quality is high. LMX differentiation combined with the LMX median creates an important team-level context (Henderson et al., 2009) that may shape team outcomes (Le Blanc and González-Romá, 2012). Although scholars have begun to examine such interaction effects, the conflicting findings obtained thus far have been puzzling. The analysis of the quadratic term interaction offers important insight into the effect of LMX-related constructs on team creativity and takes a step toward resolving the disputes concerning the interaction effect of LMX differentiation and team LMX quality.

Practical implications

Our findings have important implications for organizations and managers. First, the curvilinear effect of LMX differentiation suggests that LMX differentiation may be a double-edged sword, which must be handled with care. Most leaders develop differentiated relationships with different members in modern organizations (van Breukelen et al., 2006). Such differentiation can help leaders to enhance team creativity. However, according to our findings, LMX differentiation and team creativity are related in a curvilinear fashion, and thus managers should be cautious of creating high differentiation within work groups. As the proverbs state, ‘too much can be worse than too little’ and ‘everything in moderation; nothing in excess’. High differentiation will run the risk of stimulating intragroup interpersonal conflicts, which will in turn harm group cohesion, group cooperation, and ultimately efforts to increase team creativity.

Second, by examining the boundary conditions and contextual effects underlying the negative consequences of otherwise beneficial states such as LMX differentiation, our study can inform and help practitioners understand how to mitigate the negative consequences of LMX differentiation. The current research suggests that the LMX median has a moderating effect on the LMX differentiation-team creativity association. Thus, leaders should establish relative good relationships with most team members and strive for a balanced state of differentiation to enhance team creativity.

Limitations and future research

The current study has several limitations that indicate directions for future research. First, the study used a cross-sectional design, and thus we cannot draw strong causal inferences. Future research might employ longitudinal designs, whereby measures of LMX and team creativity are collected over several periods to permit an examination of the causal relationship between LMX differentiation and team creativity. In addition, our sample contained 59 teams. This small sample size provided low statistical power for hypothesis testing; however, the results we obtained supported all of the study’s hypotheses. Future research may expand the range of the sample and enhance the statistical power of results.

Second, our research builds on emerging theoretical and empirical evidence, suggesting that LMX differentiation has a curvilinear effect on team creativity. This perspective substantially differs from previous linear research. Future research may continue to build on these findings and examine whether the nonlinear effects we uncovered extend to other team outcomes (e.g. team performance). This different perspective may enrich the understanding of LMX differentiation and add to the burgeoning literature examining the effect of LMX differentiation on team-level outcomes.

Third, we propose that the observed decreases in team creativity may have resulted from interpersonal conflicts. However, we do not empirically evaluate such conflicts as a mediator. It will be important for future researchers to explore the mechanisms underlying the reduced creativity of highly differentiated teams. In addition to potential mediators, it is plausible that other moderators may also influence the point at which the benefits of LMX differentiation diminish. To gain a deeper understanding of the effect of LMX differentiation on team creativity or other team-focused outcomes, future research should explore such potential moderators.

Fourth, we considered a Chinese sample, which might limit the generalizability of our findings to other cultural contexts. China is a country with unique social, economic and cultural characteristics (Tsui and Lau, 2002), and our findings may, to some extent, be culturally specific. For example, high power distance and collectivism (Hofstede, 2001) may influence the development and outcomes of LMX (Ma and Qu, 2010). Thus, it would be worthwhile to replicate the investigation of this study in different cultures.

Finally, we have demonstrated that a moderate level of LMX differentiation is optimal for team creativity. Unfortunately, beyond describing such a level as moderate, neither we nor our predecessors in LMX research have been able to clearly define a moderate level of differentiation. To provide more useful guidelines for practitioners, future research may devote effort to exploring the point at which differentiation is beneficial and the point at which it is not.

Conclusion

In response to recent calls in the LMX literature for research on group-level issues related to LMX (Hogg et al., 2003; Yukl, 2002), this study examined the relationship between LMX differentiation and team creativity. Drawing on the TMGT metaprinciple, we conducted an analysis of how differentiation influences team creativity. The findings reveal that: (a) LMX differentiation has an inverted U-shaped effect on team creativity; and (b) the LMX median moderates the association between LMX differentiation and team creativity. These results suggest that the relationship between LMX differentiation and team outcomes is more complex than previous research has suggested, and highlight the importance of adopting a nonlinear perspective for elucidating these complex problems.

Footnotes

Acknowledgements

The authors are indebted to the associate editor Samuel Aryee and three anonymous reviewers for their valuable comments that have greatly improved this article. Special thanks are also given to the nine companies in China for their help in data collection.

Funding

This work was supported by grants from the National Natural Science Foundation of China (nos. 71272158, 71172009 and 71421061)

Yan Li is a researcher at School of Economics and Management, Tsinghua University, China. She received her PhD in Human Resource Development from Renmin University of China. Her research interest includes leadership, team creativity and organization innovation. Her research has been published in journals such as The Leadership Quarterly and Advances in Psychological Science. [Email: chilli666@163.com]

Fengying Fu is an associate professor at College of Applied Arts and Science of Beijing Union University, China. She received her PhD in Philosophy from Renmin University of China. Her research interests include management philosophy and cross-cultural studies. [Email: ffy2003@sina.com]

Jian-Min Sun is a professor at School of Labor and Human Resources, Renmin University of China. He received his PhD in Psychology from Beijing Normal University, China. His research interests include leadership, creativity and human resource management. His research has been published in journals such as The Leadership Quarterly, Journal of Experimental Social Psychology and International Journal of Human Resource Management. [Email: chinajms@126.com]

Baiyin Yang received his PhD in Human Resource Development from the University of Georgia, USA. He is COSCO Chair Professor and Chair of Department of Leadership and Organization Management, Tsinghua University, China. He also holds the Cheung Kong Scholar Professorship awarded by the Ministry of Education, People’s Republic of China. His research interests include adult and organizational learning, employee creativity and organization innovation, leadership and organization development. His research has been published in journals such as Human Resource Management Review, International Journal of Human Resource Management and Journal of Business Research. [Email: yangby@sem.tsinghua.edu.cn]

References

Aiken

West

(1991) Multiple Regression: Testing and Interpreting Interactions. Newbury Park, CA: SAGE.

Akgün

Dayan

Di Benedetto

(2008) New product development team intelligence: Antecedents and consequences. Information & Management 45: 221–226.

Baer

(2012) Putting creativity to work: The implementation of creative ideas in organizations. Academy of Management Journal 55(5): 1102–1119.

Baer

Oldham

Jacobsohn

Hollingshead

(2008) The personality composition of teams and creativity: The moderating role of team creative confidence. Journal of Creative Behavior 42(4): 255–282.

Barrick

Stewart

Neubert

Mount

(1998) Relating member ability and personality to work-team process and team effectiveness. Journal of Applied Psychology 83(3): 377–391.

Bell

(2007) Deep-level composition variables and predictors of team performance a meta-analysis. Journal of Applied Psychology 92: 595–615.

Blau

(1977) Inequality and Heterogeneity. New York: Free Press.

Boies

Howell

(2006) Leader-member exchange in teams: An examination of the interaction between relationship differentiation and mean LMX in explaining team-level outcomes. The Leadership Quarterly 17: 246–257.

Brislin

(1970) Back-translation for cross-cultural research. Journal of Cross-Cultural Psychology 1: 185-216.

10.

Brophy

(1998) Classroom management as socializing students into clearly articulated roles. Journal of Classroom Interaction 33(1): 1–4.

11.

Dansereau

Graen

Haga

(1975) A vertical dyad linkage approach to leadership within formal organizations. Organizational Behavior and Human Performance 13: 46–78.

12.

De Dreu

CKW

West

(2001) Minority dissent and team innovation: The importance of participation in decision making. Journal of Applied Psychology 86: 1191–1201.

13.

Dulebohn

Bommer

Liden

. (2012) A mate-analysis of antecedents and consequences of leader-member exchange: Integrating the past with an eye toward the future. Journal of Management 38(6): 1715–1759.

14.

Eisenbeiss

Van Knippenberg

Boerner

(2008) Transformational leadership and team innovation: Integrating team climate principles. Journal of Applied Psychology 93(6): 1438–1446.

15.

Erdogan

Bauer

(2010) Differentiated leader-member exchanges: The buffering role of justice climate. Journal of Applied Psychology 95(6): 1104–1120.

16.

Erdogan

Liden

(2002) Social exchanges in the workplace: A review of recent developments and future research directions in leader–member exchange theory. In: Neider

Schriesheim

(eds) Leadership. Greenwich, CT: Information Age, 65–114.

17.

Ford

Seers

(2006) Relational leadership and team climates: Pitting differentiation versus agreement. Leadership Quarterly 17(3): 258–270.

18.

Gilson

Shalley

(2004) A little creativity goes a long way: An examination of teams’ engagement in creative processes. Journal of Management 30: 453–470.

19.

Graen

Uhl-Bien

(1995) Development of leader-member exchange (LMX) theory of leadership over 25 years: Applying a multi-level multi-domain perspective. The Leadership Quarterly 6: 219–247.

20.

Harris

Kirkman

(2014) Leader-member exchange (LMX) in context: How LMX differentiation and LMX relational separation attenuate LMX’s influence on OCB and turnover intention. The Leadership Quarterly 25: 314–328.

21.

Harrison

Price

Gavin

Florey

(2002) Time, teams, and task performance: Changing effects of surface- and deep-level diversity on group functioning. Academy of Management Journal 45: 1029–1045.

22.

Henderson

Wayne

Shore

. (2008) Leader-member exchange, differentiation, and psychological contract fulfillment: A multilevel examination. Journal of Applied Psychology 93(6): 1208–1219.

23.

Henderson

Liden

Glibkowski

Chaudhry

(2009) LMX differentiation: A multilevel review and examination of its antecedents and outcomes. The Leadership Quarterly 20: 517–534.

24.

Hofstede

(2001) Culture’s Consequences. Beverly Hills, CA: SAGE.

25.

Hogg

Martin

Weeden

(2003) Leader–member relations and social identity. In: van Knippenberg

Hogg

(eds) Leadership and Power: Identity Processes in Groups and Organizations. London: SAGE, 18–33.

26.

Hooper

Martin

(2008) Beyond personal leader-member exchange (LMX) quality: The effects of perceived LMX variability on employee reactions. Leadership Quarterly 19: 20–30.

27.

Hülsheger

Anderson

Salgado

(2009) Team-level predictors of innovation at work: A comprehensive meta-analysis spanning three decades of research. Journal of Applied Psychology 94: 1128–1145.

28.

Kozlowski

Klein

(2000) A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In: Klein

Kozlowski

(eds) Multilevel Theory, Research, and Methods in Organizations: Foundations, Extensions, and New Directions. San Francisco: Jossey-Bass, 3–90.

29.

Le Blanc

González-Romá

(2012) A team level investigation of the relationship between leader-member exchange (LMX) differentiation, and commitment and performance. Leadership Quarterly 23: 534–544.

30.

Leenders

RTHAJ

van Engelen

JML

Kratzer

(2003) Virtuality, communication, and new product team creativity: A social network perspective. Journal of Engineering and Technology Management 20(1): 69–92.

31.

Liao

(2014) How do leader-member exchange quality and differentiation affect performance in teams? An integrated multilevel dual process model. Journal of Applied Psychology 99(5): 847–866.

32.

Liao

Liu

Loi

(2010) Looking the both sides of the social exchange coin: A social cognitive perspective on the joint effects of relationship quality and differentiation on creativity. Academy of Management Journal 53(5): 1090–1109.

33.

Liden

Erdogan

Wayne

Sparrowe

(2006) Leader-member exchange, differentiation, and task interdependence: Implications for individual and group performance. Journal of Organizational Behavior 27: 723–746.

34.

Lovelace

Shapiro

Weingart

(2001) Maximizing cross-functional new product teams’ innovativeness and constraint adherence: A conflict communications perspective. Academy of Management Journal 44(4): 779–793.

35.

(2010) Differentiation in leader-member exchange: A hierarchical linear modeling approach. The Leadership Quarterly 21: 733–744.

36.

Maslyn

Uhl-Bien

(2001) Leader-member exchange and its dimension: Effects of self-effort and other’s effort on relationship quality. Journal of Applied Psychology 86(4): 697–708.

37.

Naidoo

Scherbaum

Goldstein

Graen

(2011) A longitudinal examination of the effects of LMX, ability, and differentiation on team performance. Journal of Business Psychology 26: 347–357.

38.

Nishii

Mayer

(2009) Do inclusive leaders help to reduce turnover in diverse groups? The moderating role of leader-member exchange in the diversity to turnover relationship. Journal of Applied Psychology 94(6): 1412–1426.

39.

Pierce

Aguinis

(2013) The too-much-of-a-good-thing effects in management. Journal of Management 39(2): 313–338.

40.

Pirola-Merlo

Mann

(2004) The relationship between individual creativity and team creativity: Aggregating across people and time. Journal of Organizational Behavior 25: 235–257.

41.

Podsakoff

MacKenzie

Lee

Podsakoff

(2003) Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology 88: 879–903.

42.

Polzer

Milton

Swann

(2002) Capitalizing on diversity: Interpersonal congruence in small work groups. Administrative Science Quarterly 47: 296–324.

43.

Richter

Hirst

van Knippenberg

Baier

(2012) Creative self-efficacy and individual creativity in team contexts: Cross-level interactions with team informational resources. Journal of Applied Psychology 97(6): 1282–1290.

44.

Rosing

Frese

Bausch

(2011) Explaining the heterogeneity of the leadership-innovation relationship: Ambidextrous leadership. Leadership Quarterly 22: 956–974.

45.

Scandura

Graen

(1984) Moderating effects of initial leader–member exchange status on the effects of a leadership intervention. Journal of Applied Psychology 69: 428–436.

46.

Schyns

(2006) Are group consensus in leader-member exchange and shared work values related to organizational outcomes? Small Group Research 37: 20–35.

47.

Shalley

Gilson

(2004) What leaders need to know: A review of social and contextual factors that can foster or hinder creativity. Leadership Quarterly 15: 33–53.

48.

Shin

Zhou

(2007) When is educational specialization heterogeneity related to creativity in research and development teams? Transformational leadership as a moderator. Journal of Applied Psychology 92: 1709–1721.

49.

Somech

Drach-Zahavy

(2013) Translating team creativity to innovation implementation: The role of team composition and climate for innovation. Journal of Management 39(3): 684–708.

50.

Sparrowe

Soetjipto

Kraimer

(2006) Do leaders’ influence tactics relate to members’ helping behavior? It depends on the quality of the relationship. Academy of Management Journal 49: 1194–1208.

51.

Stewart

Johnson

(2009) Leader-member exchange as a moderator of the relationship between work group diversity and team performance. Group & Organization Management 34(5): 507–535.

52.

Taggar

(2002) Individual creativity and group ability to utilize individual creative resources: A multilevel model. Academy of Management Journal 45: 315–330.

53.

Tsai

W C

Chi

Garndey

Fung

(2012) Positive group affective tone and team creativity: Negative group affective tone and team trust as boundary conditions. Journal of Organizational Behavior 33: 638–656.

54.

Tsui

Lau

(2002) The Management of Enterprises in the People’s Republic of China. Boston: Kluwer Academic Press.

55.

(2009) A multilevel investigation of factors influencing creativity in NPD teams. Industrial Marketing Management 38: 119–126.

56.

van Breukelen

Schyns

Le Blanc

(2006) Leader-member exchange theory and research: Accomplishments and future challenges. Leadership 2(3): 295–316.

57.

van Breukelen

van der Leeden

Wesselius

Hoes

(2012) Differential treatment within sports teams, leader-member (coach-player) exchange equality, team atmosphere, and team performance. Journal of Organizational Behavior 33: 43–63.

58.

van der Vegt

Bunderson

(2005) Learning and performance in multi-disciplinary teams: The importance of collective team identification. Academy of Management Journal 48: 532–547.

59.

van der Vegt

Janssen

(2003) Joint impact of interdependence and group diversity on innovation. Journal of Management 29: 729–751.

60.

Volmer

Spurk

Niessen

(2012) Leader-member exchange (LMX), job autonomy, and creative work involvement. Leadership Quarterly 23: 456–465.

61.

Yukl

(2002) Leadership in Organizations (5th edn). Englewood Cliffs, NJ: Prentice-Hall.

62.

Zaccaro

Rittman

Marks

(2001) Team leadership. Leadership Quarterly 12: 451–483.

63.

Zhang

Tsui

Wang

(2011) Leadership behaviors and group creativity in Chinese organizations: The role of group processes. The Leadership Quarterly 22: 851–862.

64.

Zhou

George

(2001) When job dissatisfaction leads to creativity: Encouraging the expression of voice. Academy of Management Journal 44: 682–696.

65.

Zhou

Shin

Brass

Choi

Zhang

(2009) Social networks, personal values, and creativity: Evidence for curvilinear and interaction effects. Journal of Applied Psychology 94(6): 1544–1552.