Goal orientation diversity and team performance: The role
of transformational leadership
Abstract
This study aims to increase understanding of goal orientation diversity and its relationship with team performance. Learning goal orientation diversity and performance goal orientation diversity are the two forms of goal orientation diversity that will be examined in this study. Transformational leadership is expected to have a moderating influence on the negative relationship between goal orientation diversity and team performance such that this negative effect is reduced or even reversed. Information elaboration and conflict are hypothesized as the mediating mechanisms through which transformational leadership positively impacts on goal orientation diverse teams. Testing these mediated moderation hypotheses in a sample of 32 teams, was not found to be significant. Only main effects of learning goal orientation diversity and information elaboration were found.
Josephine Brakenhoff 10542094 MSc Thesis
Dr. C Buengeler and Dr. F Belschak Business Studies
2013-2014 Wordcount: 10.761
1
Table of Contents
Introduction 3
Goal Orientation Diversity and Team Performance 6
The Moderating Role of Transformational Leadership 6
The Mediating Role of Information Elaboration 11
The Mediating Role of Conflict 13
Method 14
Procedure and Sample 14
Missing Data Substitution 17
Justification for Aggregation 18
Measures 19
Goal Orientation Diversity 19
Transformational Leadership 20 Information elaboration 21 Conflict 21 Team performance 21 Task Performance 21 Team Innovation 21 Further variables 22 Control variables 22 Results 23 Discussion 37 Theoretical implications 39
2
Conclusion 42
Appendix 43
3
Introduction
Recently there have been marked increases in both the demands for innovation and the pressure to compete to be successful in business. Research has shown that teams provide more diverse solutions that are of a higher quality compared to individuals, and that teams have an increased ability to enhance the overall innovation and task performance within an organization (Jehn, Northcraft & Neale, 1999; Chen et al., 2007). Consequently, organizations actively pursued to shift their focus from individual-based structures towards team-based structures. The merits of team-based structures originate from the fact that teams have a larger pool of resources and knowledge that can be shared, especially when they are diverse (Kearney & Gebert, 2009). As performance is the key measure of success in business (Jehn et al., 1999), this study will focus on how diverse teams can deliver higher levels of team performance.
Nederveen-Pieterse, van Knippenberg and van Ginkel (2011) highlight the importance of goal orientation diversity and its influence on team performance and state, in line with Kozlowski and Bell (2003), that the way in which a team is composed influences how it will perform. Goal orientation diversity in a team, reflects the differences in preference with regard to the achievement of goals and is proven to have a great influence on team performance (Nederveen-Pieterse et al. 2011). Nevertheless, research on the relationship between diversity and team performance shows inconsistent results and its effects are not yet clearly understood. Some researchers claim that diversity can lead to higher task performance and team innovation, whereas others claim that diversity undermines team performance through provoking subgroups and intergroup biases conflict (Harrison & Klein, 2007; Kearney & Gebert, 2009; Williams & O’Reilly, 1998). Therefore, researchers suggest that moderators and mediators are required to help clarify the relationship between goal
4 orientation diversity and team performance (e.g. Jehn, Northcraft & Neale, 1999; Shin & Zhou, 2007; Kearney & Gebert, 2009).
According to Kearney and Gebert (2009) and Shin and Zhou (2007), transformational leadership is an effective instrument to foster performance of diverse teams. However, their research is based on demographic and informational diversity and does not take goal orientation diversity into account. Nevertheless, differences in goal orientation has a great effect on how team members collaborate on shared tasks (Nederveen-Pieterse et al., 2011). Nederveen-Pieterse and colleagues (2011) based their findings on the socially shared cognition theory and argued that when team members form different mental framework of a task. These different mental frameworks disrupt the process towards a shared mental framework and can result in implications in coordination and communication since team members put emphasis on different information when having different goals in mind. This in turn will influence the team performance (Nederveen-Pieterse et al., 2011).
In their review Van Knippenberg and Schippers (2007) suggest that due to social categorization, individuals are more likely to have better cooperation, more trust and easier interaction with the in-group than with the out-group. This could influence the information elaboration within a team and also could lead to conflict (Jehn, Northcraft & Neale, 1999; van Knippenberg, De Dreu and Homan (2004) ; Mannix & Neale, 2005).
Both information elaboration and conflict could mediate the relationship between goal orientation diversity and team performance. However, the influences of leadership, as suggested by Kearney and Gebert (2009), are not taken into account by the aforementioned researchers. In this study, information elaboration and conflict are hypothesized as the mediating mechanisms through which transformational leadership positively impacts on goal orientation diverse teams. In order to reach a higher level of team performance, it is essential to understand what can affect it (Stevens & Campion, 1994). Goal orientation diversity has
been found to negatively affect team performance
Understanding how diversity in goal orientation affects performance organizations.
Transformational leadership
successfully enhance the utilization of information and deliver higher performance (Kearney & Gebert, 2009)
the literature by focusing on the relationship between diversity in goal orientation and team performance wherein information elaboration and conflict are proposed as the mediating mechanisms through which transformational leadership
diverse teams (see Figure 1). By studying this relationship, this study following question: How can transformational
to deliver higher team performance?
Figure 1. Research Model
affect team performance (Nederveen-Pieterse et al., 2011). Understanding how diversity in goal orientation affects performance is therefore
ship is proposed to help goal orientation diverse teams to successfully enhance the utilization of information and reduce conflict so that
(Kearney & Gebert, 2009). Therefore, this study will contribute to on the relationship between diversity in goal orientation and team information elaboration and conflict are proposed as the mediating mechanisms through which transformational leadership has an impact on goal orientation
By studying this relationship, this study aims at address transformational leadership help goal orientation diverse teams to deliver higher team performance?
5 Pieterse et al., 2011). therefore valuable to
help goal orientation diverse teams to the team can will contribute to on the relationship between diversity in goal orientation and team information elaboration and conflict are proposed as the mediating has an impact on goal orientation addressing the ship help goal orientation diverse teams
6
Goal Orientation Diversity and Team Performance
Research extensively addresses the effects of demographic and informational diversity on team performance, however, research on goal orientation diversity lags far behind. Goal orientation is seen as trait-based, subjective mental frameworks and result from one’s perception and approach of a task and what important is to them in performance-relevant situations (Dweck, 1986; Nederveen-Pieterse et al., 2011). Goals motivate people to increase their effort and performance (Locke & Latham, 2006). Dweck and Leggett (1988) distinguish two forms of goal orientation based on standards which are used to estimate competence (Nederveen-Pieterse et al., 2011). These mental frameworks provide an insight into what drives people, how they approach tasks, how they react to others and what they value in performance-relevant situations (Nederveen-Pieterse et al., 2011; Harrison, Price & Bell, 1998). Learning goal orientation (LGO), also known as mastery orientation (DeGeest & Brown, 2011), is often described as “the desire to develop the self by acquiring new skills, mastering new situations and improving one’s competence” (Vandewalle, 1997, p. 1000). Individuals with high learning goal orientation tend to draw on intrapersonal standards in which they have the drive to improve past performances by encountering in deep-level information processing and have the drive to obtain new competencies. On the contrary, performance goal orientation (PGO) is driven by a focus on performance outcomes and the judgements by others (Vanderwalle, 1997; Nederveen-Pieterse et al., (2011). Individuals with high performance goal orientation have the feeling that when they are pursuing to fail on a task, their incompetence is proven. Hereby, the focus lays on external factors in which they compare themselves to others. Hence, they prefer to choose tasks that maximize their ability to show others their success (Nederveen-Pieterse et al., 2011). It is important to note that learning goal orientation and performance goal orientation are not mutually exclusive; individuals can demonstrate both orientations however, normally, one orientation will be
7 more dominant then the other (Button, Mathieu & Zajac, 1996; DeGeest & Brown, 2011). Given that individuals differ in their degrees of goal orientation, teams are likely composed of members with varying degrees of learning goal orientation and performance goal orientation. A team that consists of members with varying levels of goal orientation leads to a certain degree of goal orientation diversity. This diversity results in various process preferences towards a collective goal which may powerfully affect team performance (Nederveen-Pieterse et al., 2011).
According to the similarity-attraction perspective, individuals seek others that are similar to themselves to work with; similar individuals are more likely to have enhanced levels of communication, collaboration and coordination since they share the same mental frameworks and therefore understand each other’s way of thinking and contributions (van Knippenberg et al., 2004; Nederveen_Pieterse et al., 2011). For example, individuals with a high degree of learning goal orientation may prefer to work with other individuals with a high degree of learning goal orientation because they have the same view and approach to tasks. Even though these homogeneous teams might initially appear to be an ideal approach for delivering higher team performance, the low degree of diversity could also reduce the degree to which novel perspectives and approaches are present, of task performance and team innovation, which in turn will, inadvertently lead to a lower team performance. In line with prior research (Nederveen-Pieterse et al., 2011), goal orientation diversity was conceptualized as separation since this reflects the differences in more stable characteristics, such as traits, values and orientations. Separation, as a form of diversity, could result in less knowledge sharing, more conflict and decreased team performance due to social categorization (Harrison & Klein, 2007; Williams & O’Reilley, 1998). Expected is, based on research of Nederveen-Pieterse and colleagues (2011) and Button and colleagues (1996) that diversity in
8 performance goal orientation will have a similar consequence as learning goal orientation, although being related to different mental frameworks.
With regard to diversity in teams, two perspectives have been highlighted in many studies (e.g. Jehn et al., 1999; van Knippenberg et al., 2004; Williams & O’Reilly, 1998). The social categorization perspective is based on the similarity attraction perspective in which one has preferences to work with ‘in-group’ members over ‘out-group’ members. In line with this, the social categorization perspective proposes that homogeneous groups outperform diverse groups due to their higher commitment and group cohesion, which results in less conflict and thus higher team performance (Jehn et al., 1999; Williams & O’Reilly). However, van Knippenberg and colleagues (2004) argued that the similarity-attraction perspective could lead to decreased levels of information elaboration in diverse teams. Conversely, the information decision perspective argues that diverse teams outperform homogeneous teams due to a broader pool of knowledge, skills and abilities (KSA) to work with, and therefore can have higher performance due to improved decision-making, higher quality solutions and increased creativity and innovation (Locke & Latham, 2006; van Knippenberg et al., 2004; Williams & O’Reilly, 1998). However, these two perspective have been studied in isolation for a long time without paying enough attention to the information processing of teams (van Knippenberg et al., 2004). Therefore, the categorization-elaboration model (CEM), an integrative framework is proposed wherein the information-decision perspective and the social categorization perspective are interacting (van Knippenberg et al., 2004). This interaction clearly points out that the relationship between diversity and team performance is be influenced by several moderators and mediators. Therefore, the following hypotheses are proposed:
9
Hypothesis 1b. Performance goal orientation diversity is negatively related to team
performance.
The Moderating Role of Transformational Leadership
Transformational leadership is proposed to positively influence the hypothesized negative effects of diversity in learning goal orientation and diversity in performance goal orientation on team performance. Transformational leadership has been shown to be especially successful in times of uncertainty, change and situations in which team members have to work together wherein all resources of the individual team members have to be fully utilized in order to reach higher team performance (Kearney & Gebert, 2009; Lim & Ployhart, 2004).
Transformational leadership has been described as “the achievement of extraordinary levels of follower motivation, admiration, commitment, respect, trust, dedication, loyalty and performance” due to the actions of the leader (den Hartog & Koopman, 2001, p. 173). Judge and Piccolo (2004) described in their meta-analysis that even though the transformational leadership theory has been revised several times, four dimensions, the four I’s, of this theory are commonly accepted. These four I’s represent idealized influence (charisma), inspirational motivation, intellectual stimulation and individualized consideration. Idealized influence, or charisma, enhances the way in which followers can identify themselves with their leader due to actions of the leader to serve as a role model. By ‘leading by example’, the leader enhances trust and confidence among followers and creates commitment to objectives. In line with this, the inspirational motivation of a leader inspires followers to exceed their performance by challenging them with high standards and creating strong feelings of purpose by articulating his or her vision. With regard to intellectual stimulation, a leader stimulates creativity by encouraging, challenging and empowering them to achieve their goals. The last ‘I’ represents
10 individualized consideration which focuses on leadership behaviour with regard to the attention to followers’ needs and concerns by acting as a role model, working with empathy and enhancing interpersonal connections with followers (Bass, 1985; Judge and Piccolo, 2004). In line with this, Podsakoff and colleagues (1990) examined transformational leadership behaviour in relationship to organizational citizenship behaviours. In their research they examined the following six leadership behaviours; identifying and articulating a vision, providing an appropriate model, fostering the acceptance of group goals, high performance expectations, providing individualized support and intellectual stimulation.
Transformational leadership is theorized to be able to help goal orientation diverse teams. Somech (2006) found in her empirical study that in functional diverse teams, participative leadership was positively related to team innovation through team reflection. Nederveen-Pieterse and colleagues (2011) found in line with this that team reflexivity was an effective moderator that reversed the negative effects of goal orientation diversity and Kearney and Gebert (2009) argued that transformational leadership positively affected the relation between diversity in age, nationality and education on team performance. Consistent with this reasoning, research shows that transformational leadership is an important influential factor when it comes to shaping the processes and outcomes of diverse teams (e.g. Kearney & Gebert, 2009; Picolo & Judge, 2004; Stahevsky & Kozlowski, 2006; Shin & Zhou, 2007; van Knippenburg et al., 2004). Shin and Zhou (2007) found that transformational leadership was an effective moderator in the relationship between educational diversity and team creativity. However, they argued that more research was required to determine the role of transformational leadership in shaping processes and performance of diverse teams. Therefore, the following hypotheses are proposed:
11
Hypothesis 2a. The negative relationship between learning goal orientation diversity and
team performance is moderated by transformational leadership, such that high levels of transformational leadership, when compared to low levels, reduce or reverse the negative effect of learning goal orientation diversity on team performance.
Hypothesis 2b. The negative relationship between performance goal orientation diversity and
team performance is moderated by transformational leadership, such that it is reduced or removed when levels of transformational leadership are high compared to low.
In line with this and the requests of Shin and Zhou (2007), it is necessary to understand how transformational leadership meets the demands of diverse teams (Bass and Riggio, 2006) and influences the diversity in goal orientation in order to increase team performance. As it will be argued in the following paragraphs, transformational leadership practices may help a diverse team to increase its information elaboration by encouraging deep-level processing. In addition, these practices also help to reduce conflict by creating strong feelings of purpose, commitment and trust among the team and consequently improve team performance (Den Hartog & Koopman, 2001) .
The Mediating Role of Information Elaboration
Van Knippenberg and colleagues (2004) highlight the inconsistent findings in the relationship between team diversity and team performance and propose that the social categorization perspective, combined and integrated with the information-decision perspective, could form a functional framework to address the inconsistencies in prior research. This categorization-elaboration model (CEM) states that the in-depth process of information elaboration is disrupted by the social categorization process. Hereby, intergroup biases occur due to more positive reactions towards the similar individuals in the in-group in
12 comparison compared to the individuals who are more different and belong to the out-group. Information elaboration in teams is a process in which team members share, discussing, and integrating knowledge and information, which is necessary to reach a certain goal or complete a task (Kearney & Gebert, 2009; van Knippenberg et al., 2004). Information elaboration has been shown to be crucial for diverse teams for solving problems, making decisions and is critical to accumulate all the informational resources available in a team to enhance performance (van Knippenberg et al., 2004).
Transformational leadership may help teams that are diverse in goal orientation to engage in more information elaboration. For instance, it cannot be expected that individuals who all possess different knowledge, skills and abilities to share this with their team members. Proposed is that transformational leadership can increase the exchange of information by fostering the acceptance of team objectives and articulating a strong vision (Podsakoff et al., 1990). Also by helping team members to feel involved and engaged, transformational leaders motivate their team members to share the information that is relevant for the task (Kearney & Gebert, 2009). In line with this, Shin and Zhou (2007) also argued that due to the charisma and inspiring ways that transformational leaders motivate their team members, collective optimism and information sharing will be enhanced. Simultaneously, by showing individualized support (Bass & Riggio, 2006; Podsakoff et al., 1990) transformational leaders fosters the individual feelings of uniqueness of the team members and encourage them to explore the positive sides of different perspectives within a team. This is in line with the findings of Kearney and Gebert (2009) who found evidence for information elaboration as a mediator on the relationship between team diversity and team performance. Hence, transformational leadership may partly undermine the negative effects of goal orientation diversity on team performance through stimulating information elaboration in these teams. Therefore, the following hypotheses are put forward:
13
Hypothesis 3a. Information elaboration mediates the moderating effect of transformational
leadership on the relationship between learning goal diversity and team performance.
Hypothesis 3b. Information elaboration mediates the moderating effect of transformational
leadership on the relationship between performance goal diversity and team performance.
The Mediating Role of Conflict
The second mechanism within the CEM is the social categorization perspective. This perspective states that intergroup bias can result from unfavourable social categorization and subsequent subgroup building. This can causes potential negative effects of diversity on team performance (van Knippenberg et al., 2004). Nevertheless, transformational leadership could help teams diverse in learning goal orientation and performance goal orientation to reduce conflict and actively stimulate the positive activities in these teams. By being an example to follow and fostering the acceptance of collective team goals, a transformational leader promotes cooperation among team members (Podsakoff et al., 1990). In line with this, Nederveen-Pieterse and colleagues (2011) found intervention opportunities for leaders by creating shared understanding through team reflexivity since this could reduce conflict and increase team performance. Therefore, transformational leaders aim to enhance this shared understanding to foster communication, coordination, teamwork and information elaboration in order to reduce conflict. Lower levels of conflict are namely beneficial to team performance since this enhances mutual trust, respect and open discussion norms which fosters team work and communication (Jehn & Mannix, 2001). In addition, De Dreu and Weingart (2003) argue in their meta-analysis that empirical evidence (Gladstein, 1984) confirms that conflict interferes team performance and results is decreased satisfaction among team members due to the tension and resentment within the team and the distraction of performing the task. Furthermore, Nishii (2013) also states that conflict could be influenced
14 by transformational leadership to create a certain context in which a team can deliver higher performance. In sum, it can be assumed that transformational leadership is able to reduce conflict in diverse teams, which increases team performance. This leads to the following hypotheses:
Hypothesis 4a. Conflict mediates the moderating effect of transformational leadership on the
relationship between learning goal diversity and team performance.
Hypothesis 4b. Conflict mediates the moderating effect of transformational leadership on the
relationship between performance goal diversity and team performance.
Method Procedure and Sample
After developing the conceptual model accompanied by corresponding hypothesis, a quantitative study was constructed. Current literature did not allow for the specific questions posed in the aforementioned hypothesis to be answered and as such, new data needed to be collected. The convenience sample technique (Saunders & Lewis, 2012) was used as, a higher response rate was expected when selecting organizations and teams from personal networks. These organizations and teams were more convenient to approach since either a friend or family member was working there and could function as a way to get into the organizations.
A survey was selected as the best method of data collection for this research due to its ability to obtain detailed information from both the team members and the supervisors (Saunders & Lewis, 2012). For the supervisor survey it was of great importance that the team composition matrix was completely filled in. Otherwise, no calculations of diversity could be made. Furthermore, since only established measures were used in the survey, the items can easily be compared across time and distance. Both factors are essential for this research
15 because participants came from different countries and different time-zones. A limitation of survey use is that less detailed information can be collected when compared to qualitative research approaches (Saunders & Lewis, 2012).
As this data collection was part of a broader data collection, other constructs than the ones used for this research were also included in the survey. An online survey (see appendix 1) using an advanced survey program (Qualtrics) was constructed that contained relevant measures of the study variables, one for team members and one for supervisors, each in English and Dutch. This survey was also made available in a paper version to accommodate various participant needs. All study variables except for team performance were obtained from the team members via the survey, whilst the performance measures were obtained from their direct supervisor; objective team performance ratings were not available in the respective organizations. Only measures that have been successfully used in prior research were used in the present study. The surveys were provided in English, however, considering that the data collection was being conducted in The Netherlands, it was also deemed important to offer a Dutch version of the survey. Ten teams filled in the English survey and twenty-two teams filled in the Dutch version. The original English survey was translated into Dutch by the author, which was then checked for consistency using Brislin’s (1970) back translation method. This method requires the Dutch version to be back translated into English by an individual not familiar with the English original and with high proficiency in both the Dutch and English language. This way, the translated version of the survey was found to be consistent with the original version, both versions are suggested to be equivalent to each other (Brislin, 1970). The length of both the supervisor and team member versions of the survey was tested beforehand. Both versions took approximately 5-10 and 10-15 minutes, respectively to complete. This information was presented to the participants in a cover letter along with information about consent, the purpose of the research, contact information and
16 data protection guidelines. Participation in this research was completely voluntary and all the answers were treated strictly confidential. Participants were also ensured that the results would only contain aggregated information over a large number of teams and team leaders and therefore would not contain any information on individuals, singular teams or their team leaders.
The majority of the teams (72%) were collected through convenience sampling, other teams were collected in an organization where the management had agreed to support the study. With regard to the online survey, the supervisors of the teams which agreed to participate received two emails. One email contained a link to the supervisor survey; another email contained information and a link to the team member survey to be sent to the supervisor’s direct reports. 8-10 days after the initial email was sent, an additional email was sent to the supervisors asking them to send a reminder email to all team members in an attempt to increase participation rates. By handing out the paper version of the survey, team members willing to complete the survey (participation was voluntary) filled in the survey immediately and a quick check could be performed to make sure that all team members and their supervisor had filled in the same team name. The survey was distributed according to the preferences of the participants, be that online or paper, or Dutch or English version. As previously discussed, potential differences in content between the two language versions of the survey were minimized. The Dutch survey was created to facilitate these varying preferences in order to increase the participation rates of both team members and supervisors. Most of the data (69%) was collected using the online version since this was more convenient for a lot of participants. Participants who did not had a personal computer available to them or preferred to participate offline were given a paper-based version (21
In total, 600 teams and their supervisors were invited to participate in the study. When only considering the data of the teams in which at least two team members and the supervisor
17 participated (as long as the supervisor also completed the team composition matrix) a total of 142 team members clustered in 32 teams with their 32 supervisors participated in the survey (participation rate: 5.3% ; participation rate of team members per team: 92%). These teams originated from 14 different organizations in both the private and public sector. For the 32 teams there was an average actual team size of 6.5 people per team (SD= 4.73). As indicated by the high standard deviation, there was a wide range in team size from 2 to 25 members. The average age of the participating team members was 27 (M = 30.8, SD = 10.4), ranging from 17 to 64. On average, participants worked in their organization for 5 years, whereof 3,5 years within the current team. In terms of gender percentages, for the supervisors there were almost double the amount of male supervisors than female: 35,7% women to 64,3% men, with a more equal split in the team members: 42,3% women and 54,2% men with 3.5% of the participants not indicating their gender. In terms of education level, 47,8% of the participants had a Master’s degree or higher, 17% had a Bachelor’s degree, 16,3% went to technical school and 12.8 % had finished high school (6.4% did not fill in their educational background).
Missing Data Substitution
First, to handle the missing data, all constructs were analyzed on their frequencies to check whether there were missing values and if it was appropriate to use hotdeck imputation. Since all constructs had missing values to a percentage of less than ten percent of the total, a hotdeck imputation was considered appropriate to use. The hotdeck imputation replaced the missing values with the values of a ‘donor’ variable in the data that matched the ‘donee’ by using available information to fill in the missing information. Before computing the hotdeck imputation, some items were recoded into the same variable since these were reverse coded. These were item 16 and item 19 of the transformational leadership scale (Podsakoff, 1990).
18 item 2 of the team innovation scale by Anderson and West (1998) and item 2 of the task interdependence scale of Van der Vegt and Janssen (2003). After the hotdeck imputation, the frequencies were analyzed to determine whether the hotdeck was implemented correctly; after the imputation there should be no more missing values.
Justification for Aggregation
To check whether it was justified to aggregate the individual-level constructs to the team level of analysis, rwg values were calculated using a tool for computing Interrater Agreement (IRA)
and Interrater Reliability (IRR) estimates for consensus composition constructs, prepared by Biemann, Cole and Voelpel (2012). This tool was used to estimate the ratio of the total variance between teams and the reliability of the individual team member ratings. For learning goal orientation and performance goal orientation, justification was not needed since these variables were used for calculating diversity (standard deviation), and the conceptualization of mean learning goal orientation and mean performance goal orientation (used as control variables) follow an additive instead of a consensus composition model (Chan, 1998). Conflict had values of .91 (rwg[J]), .20 (ICC[1]) and .50 (ICC[2]), information elaboration had values of .90 (rwg[J]), .25 (ICC[1]) and .58 (ICC[2]), job complexity .81 (rwg[J]), .39 (ICC[1]) and .73 (ICC[2]), task interdependence .90 (rwg[J]), .25 (ICC[1]) and .58
(ICC[2]). Since Rwg values were found to be above the acceptable point of .70 (Biemann et al.,
2012), even though some research see this cut-off criteria as completely random (e.g. LeBreton, James & Lindell, 2005), aggregation to the team-level was justified. However, even though the ICC[1] and the rwg show sufficient support to aggregate the data, the values of
ICC[2] are only modest; this could be explained by the small team sizes (LeBreton & Senter, 2008).
19
Measures
Goal orientation diversity. Goal orientation diversity was measured using a scale on
learning goal orientation and performance goal orientation. Learning Goal Orientation was measured using the same ten items as Button and colleagues (1996). These items were tested on a 5-point Likert scale that ranged from (1) “strongly disagree” to (5) “strongly agree”. High scores on this dimension reflect a high desire to learn new skills and to look beyond standard strategies. On the other hand, low levels of learning goal orientation reflects less concern with mastering certain skills and tasks. The scale consisted of items such as “The importance to do challenging work is important to me” and “When I fail to complete a difficult task, I plan to try harder the next time I work on it”.
Performance goal orientation was also assessed using ten items developed by Button and colleagues (1996). These items were also measured on the same 5-point Likert scale that ranged from (1) “strongly disagree” to (5) “strongly agree”. High levels of performance goal orientation reflect a strong drive to get positive evaluations and judgments from others or to avoid negative reactions and evaluations. A lower score on these items reflect less concern with making mistakes and less worry about what others think. The scale consisted of items like “I prefer to do things that I can do well rather than things that I do poorly” and “I’m happiest at work when I perform tasks on which I know that I won’t make any errors”.
Diversity was calculated in accordance with Harrison and Klein’s (2007) classification. Diversity can be conceptualized in three different ways, namely as separation, variety or disparity. Separation is often used in case of diversity in position or opinion within a team and is displayed on a horizontal continuum representing diversity in attitudes, values and beliefs (Harrison & Klein, 2007). Besides that, diversity in teams can also mean differences in category, which is done through calculating variety. Alternatively, diversity in teams could also represent disparity, which displays the differences in assets like salary and hierarchical
20 rank on a vertical continuum. The three forms of diversity do not differ much when the diversity is small, because all team members of the team have the same value, attitude, or belief (Harrison & Klein, 2007). However, when team members are maximally diverse in their attitudes, values and beliefs, the types of diversity differentiate strongly.
In line with prior research (Nederveen-Pieterse et al., 2011) goal orientation diversity was conceptualized as separation since this reflects the differences in more stable characteristics, such as traits, values, and orientations. In the case of goal orientation diversity, team members differ with respect to motivational traits and their mental framework of the task. Harrison and Klein (2007) argued that, due to individual orientations, attitudes, and beliefs, there could be a variance in the level to which team members are spread along a horizontal continuum. According to these researchers, when all team members are either high or low on the continuum, diversity is low because a minimum separation occurs. In this case, it does not matter where on the continuum the team is positioned. On the other hand, when team members are spread along the continuum, diversity is high because of the occurring maximum separation. This occurs when tem members are evenly spread out over the continuum and have positions at both ends of the continuum. Due to the symmetric nature of this form of diversity, the separation between the individual team members can be defined by calculating the distance between them using the standard deviation. Hereby, the standard deviation, calculated by √[∑(Si - Smean)²/n], defines the spreading within a team.
Transformational leadership. Transformational leadership was assessed by using items
from Podsakoff and colleagues (1990). These items were measured using 23 items on a 7-point Likert scale ranging from (1) “strongly disagree” to (7) “strongly agree”. Team leaders who score high on this dimension have higher levels of transformational leadership compared
21 to leaders who score low on this dimension. The scale consisted of items like “Has a clear understanding of where we are going” and “Inspires others with his/her plans for the future”.
Information elaboration. Information elaboration was measured using the scale from
Kearney and Gebert (2009) containing four items which were answered on a 5-point Likert scale ranging from (1) strongly disagree” and (5) “strongly agree”. The scale consisted of items like “The members of this team complement each other by openly sharing their knowledge” and “The members of this team carefully consider the unique information provided by each individual team member”.
Conflict. Conflict was tested using the scale of Jehn and Mannix (2001) which
consisted of 9 items based on a 5-point scale ranging from (1) “none” to (5) “a lot”. The scale consisted of items like “How frequently do you have disagreements within your work team about the task of the project you are working on?” and “How often do you disagree about resource allocation in your work team?”.
Team performance
Team performance was measured using two different scales provided to the teams’ supervisor since the HR departments of the companies were unable to give this information.
Task performance. Task performance was measured using the 5-item scale of
Zellmer-Bruhn and Gibson (2006), which were measured on a 7-point scale ranging from (1) “very inaccurate” to (7) “very accurate”. The scale consisted of items like “This team fulfills its mission” and “This team meets the requirements set for it”.
Team innovation. Anderson and West (1998) scale was used to test team innovation.
Their 4-item scale was also measured on a 7-point scale ranging from (1) “far below average” to (7) “far above average”. The scale consisted of items like “This is an innovative team” and “Team members often implement new ideas to improve the quality of our products and
22 services”. At the end of the survey, a few questions were asked to collect general information about the participants’ demographic information.
Further variables
Besides the scales mentioned before, some other questions were also added to the survey for a broader research endeavor and as control variables. These scales consisted of items to measure belongingness and uniqueness (Shore et al., in preparation), basic psychological needs (Ryan & Deci, 2000), perceived insider status (Stamper & Masterson, 2002), diversity beliefs (Homan et al., 2007), leader inclusiveness (Nembhard & Edmondson, 2006), climate for inclusion (Nishii, 2013), LMX (Bauer & Green, 1996), job complexity (Morgeson & Humphrey, 2006) and task interdependence (Van der Vegt & Janssen, 2003). Besides task performance and team innovation two other performance scales were also included in the data collection. The scale of Kirkman and Rosen (1999) on team performance and a scale from Van der Vegt and Bunderson (2005) to measure overall performance. Lastly, participants were asked to provide demographic information, namely age, gender, nationality, culture level of education, team tenure, organizational tenure and whether the participant worked fulltime or part-time. Moreover, team leaders were asked to provide information on the actual team size as well as age, gender, nationality and team tenure of all team members.
Control variables. In this research, a number of control variables were added, namely
objective team size, job complexity, task interdependence, level of professionalism, and the mean value of learning goal orientation and performance goal orientation, respectively. These variables have been shown to be influential in prior team research (e.g. Chen et al., 2007; Jehn et al., 1999; Kearney & Gebert, 2009; Nederveen-Pieterse et al., 2011; Shin & Zhou, 2007). This way, the results were not influenced by any of these variables, however there are many other factors that might had a role in this setting. Objective team size was determined through
23 the supervisor survey wherein the supervisors completed a team composition matrix. Actual team size was used as a control variable since the collected teams varied in size. Job complexity was used as a control to ensure that the results were not influenced by the complexity of one’s job and was tested using a scale from Morgeson and Humphrey (2006) and task interdependence was measured using a scale from Van der Vegt and Jansen (2003). Furthermore, the teams were also diverse in terms of their career level. Some teams consisted of team members in director positions who all had a Masters degree or higher, whereas other teams only consisted of young students. Therefore, the variable ‘Career Level’ was created and a ‘0’ was given to teams which consisted of lower career levels and ‘1’ was given to teams that consisted of directors and similar positions.
Results
The main goal of this research was to examine the effect that transformational leadership had on the relationship between goal orientation diversity and team performance. Transformational leadership was hypothesized to enhance performance in learning goal orientation and performance goal orientation diverse teams through increased levels of information elaboration and decreased levels of conflict.
Correlations among the study variables were calculated (see Table 1). As the table 1 shows, a few significant correlations were found. Learning goal orientation was positively related to conflict. Furthermore, transformational leadership was positively related to information elaboration, task performance and team innovation and negatively related to conflict. Information elaboration was negatively related to conflict and positively related to task performance. As expected, the performance scales were also positively related to each other.
T ab le 1: M ean s, S tan d a r d D ev iat ion s, C or re lat ion s V ar ia bl es M S D 1 2 3 4 5 6 7 8 9 10 11 12 13 1. C ar ee r le v el 0, 38 0, 492 1 2. J ob c om pl e xi ty 2. 41 .69 -. 50*** (. 90) 3. T a sk i nt er de p e nde n ce 3. 86 .46 .27 -. 40** (. 77) 4. T e am si ze 6. 50 4. 73 .03 -. 16 .18 1 5. M ea n l ea rni ng g oa l or ie nt at ion 3. 96 .33 .18 -. 54*** .49*** .25 (. 81) 6. M ea n p er for m anc e g oa l or ie nt at ion 3. 53 .29 -. 12 .24 -. 30 -. 28 -. 05 (. 75) 7. D iv er si ty l ea rni ng g oa l or ie nt at ion .61 .15 -. 07 .14 -. 02 -. 15 -. 16 .29 (. 81) 8. D iv er si ty pe rf or m an ce g oa l or ie nt at ion .83 .17 -. 21 .23 .16 -. 11 .04 -. 29 .24 (. 75) 9. T ra ns for m at ion al l e ad er shi p 5. 17 .63 .44** .32 .43** -. 04 .21 -. 16 -. 19 -. 26 (. 92) 10. I nt or m at ion e la bor a ti on 3. 88 .51 .37** -. 35** .51*** -. 05 .19 -. 15 .01 .05 .80*** (. 82) 11. C on fl ic t 2. 41 .50 -. 01 -. 19 -. 23 .24 .39** .25 -. 07 -. 28 -. 36** -. 44** (. 83) 12. T a sk pe rf or m an ce 5. 80 .56 .22 -. 17 -. 01 -. 39** .04 .18 .27 -. 23 .36** .36** -. 08 (. 84) 13. T e am i nnov a ti on 4. 86 .95 .24 -. 24 .03 .01 .13 .20 .19 .05 .37** .34 -. 11 .51*** (. 68) N = 32 , v a lu es a t th e di ag ona l re pr es ent i nt er n a l cons is te n cie s. * C or re la ti on i s si g ni fi ca nt a t th e 0 .10 l ev e l (2 -t ai le d) . ** C or re la ti on i s si g ni fi ca nt a t the 0 .05 l ev e l (2-ta il ed) . *** C or re la ti on i s si g ni fi ca nt a t th e 0 .01 l ev e l (2-ta il ed) . 24
Hierarchical linear regression analyses were conducted to test the study hypotheses. All variables (except team performance) were standardized. To test Hypothesis 1a which states that learning goal orientation diversity is negatively related to team performance and 1b that performance goal orientation diversity is negatively related to team performance, several steps were taken. Firstly, the control variables, job complexity, task interdependence, team size, professionalism, and the mean scores of both diversity forms were entered in the regression. Secondly, diversity (standard deviation) in performance goal orientation and diversity in learning goal orientation were added. Learning goal orientation diversity was found to be positively related to task performance (β = .41) at the .05 level of significance and team innovation (β = .38) at the .10 level of significance. Interestingly, however, diversity in learning goal orientation had a positive relationship to task performance and team innovation instead of the expected negative relationship. Therefore, Hypothesis 1a is rejected. With regard to performance goal orientation diversity, no main effects were found for task performance and team innovation and thus Hypothesis 1b is rejected.
Hypothesis 2a proposed that the negative relationship between learning goal orientation diversity and team performance is moderated by transformational leadership, such that high levels of transformational leadership, when compared to low levels, reduce or reverse the negative effect of learning goal orientation diversity on team performance. Whereas hypothesis 2b proposes that the negative relationship between performance goal orientation diversity and team performance is moderated by transformational leadership, such that high levels of transformational leadership, when compared to low levels, reduce or reverse the negative effect of performance goal orientation diversity on team performance. Interactions between transformational leadership and diversity in learning goal orientation as well as diversity in performance goal orientation, respectively, were calculated. In the second step, transformational leadership was included in addition to check whether an indirect with
26 team performance could be found and afterwards, the interaction terms were added to the regression analysis. However, no significant moderation effect was found for learning goal orientation in relationship to task performance and team. For performance goal orientation diversity also no significant moderation effect was found in relationship to task performance and team innovation.Interactions were plotted in accordance with Dawson (2014) (see Figure 2, 3, 4, & 5).
Tab le 2 . H ie rar c hi c al r egr e ss ion T as k P e rf or m an c e. R ² = . 24 f or s te p 1; ∆ R ² = . 22** f or s te p 2; ∆ R ² = . 00 f or S te p 3; ∆ R ² = . 48 for S te p 4 . *S ig ni fi c ant a t th e 0 .10 l e v el ( 2-ta il e d) . **S ig ni fi c ant a t th e 0 .05 l e v el ( 2-ta il e d) . ***S ig ni fi c ant a t th e 0 .0 1 le v el ( 2-ta il e d) . T as k P er for m an c e S te p 1 S te p 2 S te p 3 S te p 4 V a ri abl e B S E β T B S E β T B S E β t B S E β t Int er c ep t 5. 724 .130 43. 903*** 5. 779 .120 48. 106*** 5. 779 .127 45. 452*** 5. 763 .129 44. 845*** Job com pl exi ty -. 097 .136 -. 173 -. 713 -. 021 .132 -. 037 -. 155 -. 021 .139 -. 037 -. 148 .059 .153 .105 .389 Ta sk int e rde p end e nc e -. 025 .121 -. 004 -. 205 -. 113 .120 -. 202 -. 942 -. 113 .128 -. 201 -. 880 -. 137 .139 -. 243 -. 983 T ea m si ze -. 220 .105 -. 391 -2. 085** -. 226 .101 -. 402 -2. 230** -. 226 .106 -. 402 -2. 123** -. 245 .109 -. 435 -2. 238** M e an P G O .069 .110 .123 .629 -. 062 .121 -. 110 -. 509 -. 061 .138 -. 109 -. 443 -. 145 .156 -. 258 -. 930 M e an L G O .026 .132 .047 .200 .120 .126 .214 .955 .121 .136 .214 .888 .142 .172 .252 .826 C ar ee r le v el .185 .234 .162 .791 .040 .222 .035 .179 .041 .242 .036 .171 .051 .246 .044 .205 Di v er si ty L G O .229 .103 .408 2. 233** .229 .118 .407 1. 935* .244 .120 .434 2. 029* D iv er si ty i n P G O -. 152 .121 -. 271 -1. 261 -. 151 .140 -. 268 -1. 082 -. 258 .165 -. 459 -1. 564 T ra ns for m a ti on a l le a d er shi p .197 .114 .351 1. 723* .196 .137 .349 1. 434 -. 007 .216 -. 012 -. 031 Int er a ct ion L G O .000 .107 .000 .001 -. 065 .126 -. 132 -. 519 Int er a ct ion P G O .003 .113 .006 .029 .002 .114 .003 .015 Inf or m at ion el a bor at ion .291 .221 .518 1. 316 C onf li c t .033 .163 .058 .201 27
28 T ab le 3 . H ie rar c hi c al r egr e ss ion T e am I nnov at io n. T eam I n n ovat ion S te p 1 te p 2 S te p 3 S te p 4 V ar ia bl e B S E β T B S E β t B S E β t B S E β t Int e rc ep t 4. 722 .237 19. 922*** 4. 819 .217 22. 173*** 4. 808 .229 21. 025*** 4. 803 .242 19. 877*** Job com pl exi ty -. 106 .248 -. 111 -. 426 -. 129 .240 .135 -. 537 -. 128 .250 -. 135 -. 514 -. 096 .287 -. 101 -. 334 Ta sk int er de p e nd enc e -. 182 .219 -. 191 -. 830 -. 446 .218 .468 -2. 050** -. 432 .230 -. 452 -1. 874* -. 428 .262 -. 449 -1. 637 T e am si ze -. 076 .192 -. 080 -. 297 .025 .183 .027 .139 .029 .192 .030 .151 .018 .206 .019 .087 M ea n P G O -. 209 .200 -. 219 -1. 048 -. 261 .219 -. 274 -1. 195 -. 293 .248 -. 307 -1. 181 -. 333 .294 -. 349 -1. 133 M ea n L G O .135 .239 .142 .564 .198 .228 .208 .868 .175 .244 .184 .718 .158 .323 .165 .489 C ar ee r le v el .358 .426 .185 .841 .102 .401 .052 .253 .105 .435 .054 .241 .100 .462 .052 .217 Di v er si ty L G O .366 .186 .384 1. 968* .397 .213 .416 1. 863* .406 .226 .426 1. 792* D iv er si ty i n P G O .121 .219 .127 .552 .113 .251 .119 .451 .081 .310 .085 .262 T ra ns for m at ion al le ad er shi p .497 .207 .521 2. 398** .521 .246 .547 2. 121** .463 .406 .486 1. 141 Int e ra ct ion L G O -. 082 .193 -. 097 -. 423 -. 117 .237 -. 139 -. 492 Int e ra ct ion P G O .021 .204 .023 .105 .021 .214 .022 .097 Inf or m a ti on el abor at ion .116 .416 .122 .279 C onf li ct .052 .306 .055 .171 R ² = . 13 f or s te p 1; ∆ R ² = . 26** f or s te p 2; ∆ R ² = . 01 f or S te p 3; ∆ R ² = . 00 for S te p 4 . *S ig ni fi c ant a t th e 0 .10 l e v el ( 2-ta il e d) . **S ig ni fi c ant a t th e 0 .05 l e v el ( 2-ta il e d) . ***S ig ni fi c ant a t th e 0 .0 1 le v el ( 2-ta il e d) .
Fig. 2. Relationship between learning goal orientation diversity on task performance under
high and low levels of transformational leadership.
Fig. 3. Relationship between learning goal orientation diversity on team innovation under
high and low levels of transformational leadership. 1 2 3 4 5 6 7 Low Diversity in performance goal orientation High Diversity in performance goal orientation T as k p er for m an ce Low Transformational leadership High Transformational leadership 1 2 3 4 5 6 7 Low Diversity in performance goal orientation High Diversity in performance goal orientation T eam i n n ovat ion Low Transformational leadership High Transformational leadership 29
30
Fig. 4. Relationship between performance goal orientation diversity and task performance
under high and low levels of transformational leadership.
Fig. 5. Relationship between performance goal orientation diversity and team innovation
under high and low levels of transformational leadership. 1 2 3 4 5 6 7 Low Diversity in performance goal orientation High Diversity in performance goal orientation Ta sk p er fo rm a n ce Low Transformational leadership High Transformational leadership 1 2 3 4 5 6 7 Low Diversity in performance goal orientation High Diversity in performance goal orientation Te a m i n n o v a ti o n Low Transformational leadership High Transformational leadership
31 Since no moderating effect of transformational leadership was found on both learning goal orientation diversity or performance goal orientation diversity in relationship to task performance and team innovation, Hypotheses 2a and 2b are rejected. Interestingly, however, even though no moderation effect was found, transformational leadership did have a main effect on task performance (β = .35) and team innovation (β = .52) at the .10 level of significance.
The fourth step in the hierarchical regression was mediation testing. Mediation was tested in accordance with the stepwise method of Baron and Kenny (1986). After checking that the independent variable is related to the dependent variable, Baron and Kenny (1986) recommend that the relationship between the independent variable and the mediators is tested. To test Hypothesis 3a which proposes that information elaboration mediates the moderating effect of transformational leadership on the relation between learning goal orientation diversity and team performance and Hypothesis 3b which proposed that information elaboration mediates the moderating effect of transformational leadership on the relationship between performance goal orientation diversity and team performance, a regression analysis was conducted wherein the interaction of diversity in learning goal orientation diversity and the interaction of performance goal orientation diversity were taken as independent variables and information elaboration as dependent variable. The interaction effect of diversity in performance goal orientation (β = .01) was not significantly related to information elaboration, however the interaction effect of learning goal orientation was positively related (β = .23) to information elaboration at the .10 level of significance. Even though an interaction effect of performance goal orientation diversity was not found, diversity in performance goal orientation was related (β = .36) to information elaboration at the .05 level of significance and
32 transformational leadership was also positively related (β = .84) to information elaboration at the .01 level of significance.
To test Hypothesis 4a, which proposes that conflict mediated the moderating effect of transformational leadership on the relationship between learning goal orientation diversity and team performance and Hypothesis 4b, which proposes that conflict mediates the moderating effect of transformational leadership on the relationship between performance goal orientation diversity and team performance, a regression analysis was conducted wherein the interaction of diversity in learning goal orientation diversity and the interaction of performance goal orientation diversity were taken as independent variables and information elaboration as dependent variable. The interaction of diversity in learning goal orientation and the interaction of diversity in performance goal orientation were both not significant. However, diversity in performance goal orientation was negatively related (β = -.32) to conflict at the .10 level of significance and transformational leadership was also negatively related (β = -.47) to conflict at the .01 level of significance.
Tab le 4 . R e gr e ss ion r e su lt s w it h i n for m at ion e lab or at ion a s de p e nde n t v ar iabl e . In for m at ion E lab or at ion S te p 1 S te p 2 S te p 3 V ar ia bl e B S E β t B S E β t B S E β t Int er ce p t -. 141 .214 .660 .030 .138 .215 .058 .135 .431 Job c om pl e xi ty .-156 .223 -. 156 -. 701 -. 270 .152 -. 270 -1. 780* -. 272 .147 -. 272 -1. 846* T as k i nt er de p e nde n ce .498 .198 .498 2. 519** .145 .138 .145 1. 050 .120 .136 .120 .880 Te am si ze -. 131 .173 -. 131 -. 761 .060 .116 .060 .512 .053 .113 .053 .473 M ea n L e ar ni ng g oa l or ie nt at ion -. 137 .215 -. 137 -. 636 -. 204 .145 -. 204 -1. 412 -. 148 .144 -. 148 -1. 029 M ea n P er for m a n ce g oa l or ie nt at ion .018 .180 .018 .099 .185 .139 .185 1. 334 .271 .146 .271 1. 852* C a re er l ev e l .376 .384 .185 .979 -. 079 .254 -. 039 -. 310 -. 055 .257 -. 027 -. 216 D iv e rs it y i n le ar ni ng g oa l or ie nt a ti on .038 .118 .038 .324 -. 044 .126 -. 044 -. 347 D iv e rs it y i n p e rf or m anc e g oa l or ie n ta ti on .356 .138 .356 2. 575** .399 .148 .399 2. 696*** T ra ns for m at ion al l e ad er shi p .839 .131 .839 6. 398*** .758 .145 .758 5. 226*** Int er ac ti on l ea rni ng g oa l or ie nt at ion di v er si ty .200 .114 .226 1. 753* Int er ac ti on pe rf or m a n ce g oa l or ie n ta ti on di v er si ty .006 .120 .006 .046 R ² = . 36* f or s te p 1; ∆ R ² = . 42*** f or s te p 2; ∆ R ² = . 03*** f or S te p 3 . *S ig ni fi c ant a t th e 0 .10 l e v el ( 2-ta il e d) . **S ig ni fi c ant a t th e 0 .05 l e v el ( 2-ta il e d) . ***S ig ni fi c ant a t th e 0 .0 1 le v el ( 2-ta il e d) . 33
34 T ab le 5 . R e gr e ss ion r e su lt s w it h c onf li c t as d e p e n de nt v a ri abl e . C on fl ic t S te p 1 S te p 2 S te p 3 V ar ia bl e B S E β t B S E β t B S E β t Int er ce p t .012 .196 .059 .073 .183 .398 .041 .183 .224 Job c om pl e xi ty .132 .205 .132 .645 .016 .202 .016 .079 .018 .200 .018 .091 T as k i nt er de p e nde n ce .510 .181 .510 2. 811*** .300 .183 .300 1. 633 .330 .185 .330 1. 785* Te am si ze .237 .158 .237 1. 499 .101 .155 .11 .657 .094 .154 .094 .613 M ea n L e ar ni ng g oa l or ie nt at ion .527 .198 .527 2. 667*** .604 .192 .604 3. 141*** .667 .196 .667 3. 407*** M ea n P er for m a n ce g oa l or ie nt at ion .219 .165 .219 1. 326 .062 .184 .062 .338 .158 .199 .158 .793 C a re er l ev e l .031 .183 .015 .087 .194 .338 .. 96 .575 .216 .349 .106 .618 D iv e rs it y i n le ar ni ng g oa l or ie nt a ti on .017 .157 .017 .107 .074 .171 .074 .436 D iv e rs it y i n p e rf or m anc e g oa l or ie n ta ti on .323 .184 .323 1. 754* .278 .201 .278 1. 383 T ra ns for m at ion al l e ad er shi p .468 .172 .468 2. 685*** .556 .197 .556 2. 820*** Int er ac ti on l ea rni ng g oa l or ie nt at ion di v er si ty .225 .155 .255 1. 453 Int er ac ti on pe rf or m a n ce g oa l or ie n ta ti on di v er si ty .003 .163 .003 .017 R ² = . 33*** f or s te p 1; ∆ R ² = . 11 *** f or s te p 2; ∆ R ² = . 01*** f or S te p 3 . *S ig ni fi c ant a t th e 0 .10 l e v el ( 2-ta il e d) . **S ig ni fi c ant a t th e 0 .05 l e v el ( 2-ta il e d) . ***S ig ni fi c ant a t th e 0 .0 1 le v el ( 2-ta il e d) .
Step 3 of the Baron and Kenny (1986) approach is to check whether the mediator has a significant effect on the dependent variable (see Table 6.). Consequently, a regression analysis was conducted wherein the team performance indicators were regressed on information elaboration and conflict in addition to the previously reported predictor variables. It was found that only information elaboration was significantly positively related to task performance (β = .47) on the .05 level of significance.
Table 6. Regression results of information elaboration and conflict as independent variables.
Variables Team performance Task performance Team innovation Overall performance Intercept 5.880*** 6.659*** 6.208*** 6.245*** Job complexity .079 -.148 -.280 -.127 Task interdependence -.287 -.067 -.202 -.151 Teamsize -.006 -.049** -.003 -.036 Career level .669 .177 .321 .310 R² .108 .226 .086 .109 ∆R² .108 .226 .086 .109 Intercept 3.590*** 4.820*** 5.217*** 5.328 Job complexity .176 -.071 -.260 -.105 Task interdependence -.491 -.246 -.580 -.448 Teamsize -.002 -.045** .011 -.026 Career level .562 .086 .168 .189 Information elaboration .600 .503** .677 .550 Conflict .221 .167 -.106 -.061 R² .164 .336 .187 .204 ∆R² .056 .110 .101 .095
*Significant at the 0.10 level (2-tailed). **Significant at the 0.05 level (2-tailed). ***Significant at the 0.01 level (2-tailed).
Sobel’s test was conducted to determine whether there was a significant mediated moderation effect. The reason behind this is that Sobel’s test is based on the normality assumption and the small sample sizes of this research could therefore be a problem. In line with this, no significant results were found after conducting the Sobel’s test for information elaboration within the relationship of the interaction of learning goal orientation diversity on task performance and team innovation. Similarly for the interaction of performance goal 35
36 orientation diversity, information elaboration was not significantly related to the performance measures. Conflict also provided similar results: no significant mediation was found in the relationship between the interaction in learning goal orientation diversity and task performance or team innovation. Finally, conflict was not a mediator of the relationship between the interaction of performance goal orientation diversity and the performance measures. Furthermore, conflict also did not mediate the relationship between the interaction of learning goal orientation diversity and the performance measures.
To analyze the relationships between the variables using a non-parametric resampling approach, bootstrapping was used by using the process tool in accordance to Preacher and Hayes (2004) on task performance and team innovation. These two performance measures were chosen since these measures showed some effects that need further research. This test was done since it has some advantages over the Sobel test such as increased power. Especially with the small sample size of this research where normality assumptions might fail, this test is very useful. According to Hayes (2009), by using bootstrapping the 32 teams of this research are taken as a representation of the total population. Bootstrapping generates the data like it is a perfect representation of the whole population. The two main purposes of bootstrapping are to deal with distributions that are not necessarily normal and to estimate parameters that are not easy to determine through analysis. Bootstrapping is thus recommended in cases like this where analysis cannot provide a solution or when this solution is dependent on weak assumptions (Edward & Lambert, 2007). When checking bootstrap intervals, an interval of 90% is chosen to test for the significance. It is justified to use a 90% bootstrap interval since this research had a shortage of results. Therefore, it is justified for indirect effects to bootstrap at the 90% interval. Hereby, zero should not be in the range of the confidence interval provided by bootstrapping to have evidence for mediation moderation. Essentially, the question is whether there is a possibility, within 90% confidence, that the true indirect effect
37 would be zero. This would mean that there is no mediation. By analyzing the interaction effect of goal orientation diversity on task performance and team innovation at values of information elaboration as mediator, it could be concluded that the effect of both the interactions to task performance and team innovation were not mediated by this variable on a confidence level of 90%. Hence, since no mediated moderation effect was found, hypothesis 3a and 3b are rejected. Conflict did not mediate the relationship between the interaction of learning goal orientation diversity and the interaction of performance goal orientation diversity on the two performance measures and thus hypothesis 4a and 4b could not be supported.
Discussion
To keep pace with the demand for innovation and creativity, organizations shifted their focus towards team-based structures as opposed to individual-based structures (Jehn, Northcraft & Neale, 1999; Chen et al., 2007). To achieve this, it is assumed that more diverse teams have a larger pool of resources and that if they all share their knowledge, higher levels of team performance were expected. However, goal orientation diversity could also result in conflict and disrupted information sharing which will lower their performance (e.g. Harrison & Klein, 2007; Kearney & Gebert, 2009; Williams & O’Reilly, 1998). This study focused on the role of transformational leadership as a moderator on the relationship between learning goal orientation diversity and performance goal orientation diversity wherein information elaboration and conflict were considered to mediate the moderated effect. Interestingly, results showed that learning goal orientation diversity had a positive main effect on both task performance and team innovation. Probably, the two forms of diversity in goal orientation differ in the way teams process information. Teams diverse in performance goal orientation are probably more related to the preferences in the way or strategy to get results instead of
