Team cultural diversity and team innovation; what’s the deal?

(1)

Masterthesis Strategic Human Resources Leadership

Student: Femke Blankenburgh Student number: 4236882 Supervisor: dr. J.J.L.E. Bücker Second examiner: dr. C. Ossenkop Date: 16-07-2018

(2)

2

Abstract

This thesis provides further insights in the relationship between team cultural intelligence and team innovative work behavior. The effects of team leader cultural intelligence, team learning and team trust on this relationship are explored. Results show that team leader cultural intelligence and team learning positively affect team cultural intelligence and team innovative work behavior. Team trust moderates the relationships between team learning and team cultural intelligence and between team leader cultural intelligence and team cultural intelligence. Team cultural intelligence also has a positive effect on team innovative work behavior. No mediating effect of team cultural intelligence was found. These results are discussed in the light of the literature and qualitative findings.

(3)

3

Preface

Before you lies the master thesis: “Team cultural diversity and team innovation; what’s the deal?” This research concludes my master program Business Administration, specializing in Strategic Human Resources Leadership, at the Radboud University Nijmegen. Writing this master thesis has been a challenging and stressful process, with some challenges mostly regarding the collection of data. However in the end, I succeeded in finishing the research and I can say that I am proud of this achievement.

I think the subjects of team cultural intelligence and innovation are very relevant and important these days. As the world is globalizing, companies increasingly have to know how to deal with a diverse workforce in order to improve processes in multicultural teams. This is especially important as teams are the unit within organization where a lot of innovation takes place. Therefore it was a very interesting process to research the factors related to these subjects and seeing what relationships could be found and explained.

In the first place, I would like to thank dr. Joost Bücker for the support and guidance during this process, especially when I was having a hard time finding respondents. Furthermore, I would like to thank my second examiner, dr. Carolin Ossenkop for the feeback that helped me finalize this thesis. I also would like to thank all the people and organizations that participated in this research. Lastly, I would like to thank my parents, Remy and Mark for their support and encouragement at times when I did not feel confident about finishing this research.

I hope you enjoy reading this thesis.

Femke Blankenburgh

(4)

4

3.5 Research ethics ... 27 4. Results ... 29 4.1 Quantitative analysis ... 29 4.1.1 Data preparation ... 29 4.1.2 Data reduction ... 30 4.1.3 Assumptions ... 37 4.1.4 Analysis ... 38 4.2 Qualitative analysis ... 43 5. Conclusion ... 49 5.1 Quantitative conclusion ... 49 5.2 Qualitative conclusion ... 50 6. Discussion ... 53 6.1 Theoretical implications ... 53 6.2 Practical implications ... 57 6.3 Limitations ... 59

6.4 Recommendations for future research ... 60

References ... 62

Appendix A: Survey questions ... 68

Appendix B: Interview questions ... 76

Appendix C: Codebook ... 78

Appendix D: SPSS output ... 90

(5)

5

1. Introduction

1.1 Topic introduction

The world is changing, fast. As a result of globalization, companies are expanding their production and services to different parts of the world. This is happening at an immense pace (Resick et al., 2006) and it is becoming increasingly harder to find organizations which only serve domestic clients (Bücker & Korzilius, 2015). Globalization leads to more competition for organizations, as their competitors are now located worldwide. This means organizations need to achieve a competitive advantage. To do that, it is key that organizations find ways to innovate and stay ahead of the competition (Korzilius, Bücker & Beerlage, 2016). Organizations increasingly face a knowledge-based competitive environment (Doz, Santos & Williamson, 2001), which means they have to find innovative ways to be competitive. Innovation is also an important driver of growth, performance and valuation (Barsh et al., 2008). Innovative behavior of employees helps international organizations be successful in such a dynamic business environment (Yuan & Woodman, 2010).

Globalization leads to differences in the workforce as well. The cultural diversity of local workforces is increasing, which means employees and managers in many companies increasingly face and have to deal with intercultural differences (Andresen & Bergdolt, 2017). Having employees who understand, function in and manage their culturally diverse team(s) can lead to a competitive advantage for organizations, as these employees are rare, valuable and inimitable resources (Chen & Lin, 2013). To measure the extent to which individuals know how to deal with different cultures, the concept of cultural intelligence (CQ) was developed. CQ was first described by Earley (2002, p. 274) as: “a person’s capability to adapt effectively to a new cultural context.” Someone who is culturally intelligent is aware of other cultures and is able to relate to and work with people from different cultures (conference paper, Euram 2018 (oil industry)). Possessing higher levels of CQ enables understanding, adaptation, communication, and coordination in diverse settings (Johnson et al., 2006; Adair et al., 2013). However, to this point little research has been done about CQ on a team level. This is a gap in the literature, since teams have become one of the most important parts of organizations and team-based research regarding innovation has become a fruitful area of research (Chi et al., 2009). As teams are becoming increasingly multicultural, challenges can arise regarding communication, conflict and identity (Adair et al., 2013). Yet according to Zakaria, Amelinckx, and Wilemon (2004) potential advantages of global (virtual) teams include that they can create culturally synergistic solutions, enhance creativity and cohesiveness among team members and

(6)

6

promote a greater acceptance of new ideas. Therefore the need in diverse teams to interact effectively with people who have a different cultural background is essential to achieve innovation in teams (Thomas et al., 2015). Team CQ can be defined as a team’s abilities and knowledge that are necessary to function effectively and to adjust to different cultural interactions (conference paper, Euram 2018 (oil industry)). According to Chen and Lin (2013), culturally intelligent teams will outperform less culturally intelligent teams.

One of the factors that influence team behavior is the team leader. Team members are receptive to the content of the behavior of the team leader and how the team leader directs social processes (Tyler & Lind, 1992). This entails that team leader CQ could also influence the CQ in a team. Leaders who possess a higher level of CQ, are better at managing expectations of people with different cultural backgrounds and reducing exclusionary reactions that are likely to happen in cross-cultural situations (Rockstuhl et al., 2011).

Another factor that can influence team behavior is team learning, teams can learn as a group and adjust their behavior as a consequence. Team learning is a source of organizational competitive advantage (Barker & Neaily, 1999). It is a process that increases the collective intelligence of the team and can lead to the development of knowledge and solutions to complex problems (Senge, 2006; Kayes & Burnett, 2006). The quality of interpersonal relationships can improve through team learning (Zellmer-Bruhn & Gibson, 2006) and team members can learn how to deal with forces that hinder productive dialogue and discussion creatively through this process (Senge, 1994).

Team trust is also an important element in culturally diverse teams, since it enhances team creativity and a collaborative culture (Barczak et al., 2010). Trust leads individuals to expect that others will behave as expected and not be opportunistic (Barczak et al., 2010; Jarvenpaa et al., 1998). It is an essential factor for effective team processes and performance (Ashleigh & Prichard, 2012; Costa, 2003). In work teams, the ability to work together is increased by trust (Barczak et al., 2010). However, when teams have members with different perspectives (for example due to different cultures), there is greater potential for misunderstanding, conflict and miscommunication, which can decrease intragroup trust (Barczak et al., 2010).

1.2 Research problem

Globalization leads to organizations having to be increasingly innovative to create a competitive advantage. It also results in the workforce becoming more and more diverse and teams becoming increasingly multicultural. Therefore it is interesting to investigate the relationship

(7)

7

between cultural intelligence in multicultural teams and team innovative work behavior. Factors that seem important in this respect are team leader cultural intelligence, team trust and team learning behavior.

The aim of this study is to investigate how the variables of team leader cultural intelligence, team learning behavior, team cultural intelligence, team trust and team innovative work behavior are related. The research question is: How are team leader cultural intelligence, team

learning behavior, team cultural intelligence, team trust, and team innovative work behavior related in multicultural teams in multinationals?

1.3 Contribution

CQ is a relatively new construct and is receiving an increasing amount of attention in the literature. However, empirical studies on team cultural intelligence (team CQ) in multicultural teams are still scarce (Adair et al., 2013). More studies measuring CQ at a team level are necessary to better understand the construct. Furthermore, the direct effect of team CQ on team innovative work behavior (team IWB) has been studied (e.g. conference paper, Euram 2018 (oil industry)), but there are many factors that might influence this relationship. The theoretical contribution of this study is to fill this gap in the literature, by examining which factors are related to team CQ and team IWB.

The study also has a practical contribution, as it offers organizations tools to increase the team IWB behavior of their teams. By examining the factors that are related to team CQ and team IWB, organizations can understand how increased team CQ contributes to innovation. Actions can then be taken to enhance these processes.

1.4 Outline

Chapter 2 will contain a review of the literature that is already available regarding this subject. The theoretical framework will be used to formulate hypotheses and later as an analytical tool to analyze the findings. The hypotheses combined will result in a conceptual model, showing the presumed relationships between the different concepts. The next chapter discusses the research design. The data collection methods and methods of analysis will be elaborated upon, as well as the reliability and validity and some ethical considerations. The fourth chapter presents the findings of the study including an analysis of these findings. Then a chapter consisting of the conclusion that can be drawn based on the results will follow. In the last

(8)

8

chapter, a discussion of the results will be presented. The discussion will include theoretical and practical implications of the study, the limitations and suggestions for further research.

(9)

9

2. Theoretical framework

This chapter will provide an overview of relevant literature regarding cross-cultural effectiveness of teams in multinational organizations. First, the concepts of innovative work behavior and team cultural intelligence will be explained. Then, attention is payed to team leader cultural intelligence and team learning behavior. Furthermore, the concept of trust will be elaborated upon. Hypotheses are formulated and the chapter concludes with a conceptual model.

2.1 Innovative work behavior in teams

According to Cummings and O’Connell (1978), innovation is necessary for organizations to be effective. They define innovation as: ‘a subset of organizational change in which new products,

technologies, or structures are introduced with the objective of improving organizational effectiveness.’ (Cummings & O’Connell, 1978, p. 33). Innovation consists of several phases,

such as idea generation, adopting decisions and idea implementation (Hülsheger et al., 2009; Woodman et al., 1993; Damanpour & Schneider, 2006). Employees play a key role in innovation, as employee-driven innovation is a critical and effective resource for innovation (Høyrup, 2010). Employees acquire skills and knowledge about the organization which they can use to innovate (Høyrup, 2010). The innovative capacity of an organization can be considered to be located with its employees (Jiménez-Jiménez & Sanz-Valle, 2008) and its teams in their social capital (Tsai & Ghoshal, 1998; Landry et al., 2002). Therefore teams are an important part of organizations where much innovation takes place. Many organizations use employee-driven (non-R&D) innovation, but research regarding this topic is still limited (Høyrup, 2010).

Employee innovative work behavior can be described as: ‘the intentional creation,

introduction, and application of new ideas within a work role, group or organization, in order to benefit role performance, the group or the organization’ (Janssen, 2000, p. 288). Yuan and

Woodman (2010) state that developing, adopting, and implementing new ideas for products and work methods are components of employee innovative behavior. First, new ideas need to be thought of. Employees are likely to conceive new ideas by being in contact with the market, products and processes (Høyrup, 2010). Second, ideas need to be promoted to gain support and then they should be converted to products or services (Janssen, 2000). Organizations often hope to achieve innovation through encouraging teamwork (Chi et al., 2009). Innovation in teams can be achieved through effective cross-fertilization of ideas (Chi et al., 2009). This is due to

(10)

10

the fact that teams can integrate multiple perspectives and types of knowledge that lead to innovation.

Gibson and Vermeulen (2003) state that differences between team members can lead to higher combined cognitive capacity of the team, leading to more creativity, better information processing and decisions of a higher quality. Akdemir, Erdem and Polat (2010) similarly found that (cultural) diversity enhances creativity and innovation of responses. Interacting with people who are different from themselves enables employees to learn and innovate since they are exposed to different paradigms and perspectives and the cross-fertilization of ideas is encouraged (Van der Vegt & Bunderson, 2005). Niebuhr (2010) also states that cultural diversity can lead to new combinations of knowledge, resulting in (regional) innovation.

However, just the mere presence of members with diverse cultural backgrounds does not automatically lead to team members working together creatively (conference paper, Euram 2017 (dairy industry)). For instance, increased multiculturalism might be associated with problems regarding communication and identity and might lead to conflict (Adair, Hideg & Spence, 2013; Gibson & Vermeulen, 2003). Furthermore, employees who are a minority or dissimilar, might participate less in terms of communication, decision making and contribute less to group tasks (Hobman et al., 2004). This lack of involvement in the team might reduce the potential of making use of the advantages of diversity. This means is important for employees to know how to deal with different cultures.

2.2 Team cultural intelligence

To measure the extent to which individuals know how to deal with different cultures, the concept of cultural intelligence (CQ) was developed. CQ was first described by Earley (2002, p. 274) as: “a person’s capability to adapt effectively to a new cultural context.” The concept of CQ has been described as having multiple dimensions: metacognitive, cognitive (cultural knowledge), behavioral (flexibility) and motivational (motivation to adapt in cross-cultural settings) (Earley, 2002, p. 274; Earley & Ang, 2003). Employees, who possess a high level of CQ, are more perceptive in reading underlying motives, adjust their behavior more easily in new situations and are motivated to learn about new cultural surroundings and adjust to them (Earley & Ang, 2003, p. 191).

(11)

11

Since teams are increasingly seen as a vital part of organizations (Chi et al., 2009), it is important to look at CQ at a team level. Team CQ can be described as a (diverse) team’s abilities and knowledge that are necessary to function effectively and to adjust to different cultural interactions (conference paper, Euram 2018 (oil industry)). It is not just the individual CQ of all the team members combined, but it is based on the way the team behaves as a unit in culturally diverse situations (conference paper, Euram 2018 (oil industry)). Behaviour as a result of team dynamics is also part of the concept. Three dimensions of team CQ can be distinguished: team fusion, metacognition and openness to diversity (conference paper, Euram 2018 (oil industry)). Team fusion means that team members respect and recognize that there are differences in the team and that they use and maintain the specific qualities of those differences (Janssens & Brett, 2006). This means cultural differences and the unique perspectives that come with these differences are preserved and utilized to benefit the team. Team members will need to reject ethnocentrism, and recognize and respect different cultures while preserving their different cultural identities in order to achieve creative solutions (Janssens & Brett, 2006). Team fusion can be further divided into co-existence and meaningful participation (Bücker, Korzilius and Dijkman, in submission). Co-existence means that team members recognize, respect and accept the differences between them and intend to combine their differences in a way that maintains their unique qualities (Bücker, Korzilius and Dijkman, in submission). Meaningful participation means that team members use dialogue to extract information and make decisions (Janssens and Brett 2006). Team metacognition entails the team being able to reflect on the cultural knowledge and experiences that are available within the team (conference paper, Euram 2018 (oil industry)). The team is aware of the cultural knowledge they use during cross-cultural interactions and idea exchange and reflect on their knowledge and experiences. Lastly, openness to diversity can be described as: “awareness and acceptance of both similarities and differences that exist among team members” (conference paper, Euram 2018 (oil industry)). Openness to diversity can mean openness towards different levels of language proficiency, vocabulary and/or accents, openness towards different looks, openness towards different values, opinions and behavior and openness towards different types of information (Hobman et al., 2004; Lauring & Selmer, 2013).

The metacognitive dimension of team CQ means that culturally intelligent teams reflect on new insights (Crotty & Brett, 2012). This could result in idea generation, which is necessary for innovation. The openness to diversity dimension of team CQ could lead to team members embracing different viewpoint instead of rejecting them (Lauring & Selmer, 2013). This could

(12)

12

increase the creativity of idea generation. The fusion dimension of team CQ enables team members to respect differences in cultures and utilize the qualities of those diffeences (Janssens & Brett, 2006), enhancing team processes such as the generation of new insights.

Additionally, according to Chen and Lin (2013), culturally intelligent teams will outperform less culturally intelligent teams. This result can be explained by group cohesion, as cohesive groups outperform non-cohesive groups and need less hierarchical management (Sapsed et al., 2002). Cohesion can be described as to what extent team members work together to achieve the goals of the team (Mach et al., 2010). It occurs when team members have positive feelings toward each other and toward the whole team (Hansen et al., 2002). Group cohesiveness is influenced by factors regarding social interactions such as proximity, physical isolation, physical ties to workplace, high-required interactions, low noise level, free time at work etcetera (Sapsed et al., 2002). In general, factors such as background, ethnicity and social interactions determine attraction between individuals, as people are more attracted to people who are similar to themselves (Sapsed et al., 2002). Yet according to Lauring and Selmer (2013), teams that are open to diversity (one of the dimensions of team CQ) include other team members in their activities, whether they share demographic or other characteristics or not. Perceived openness to diversity of team members contributes to more cohesion in the team (Harrison et al., 2002; Van Dick et al., 2008). Therefore, team CQ can be expected to increase cohesion within the group. Cohesion is regarded one of the necessary conditions for IWB (e.g. Woodman et al., 1993). When there is more cohesion, team members are more likely to presume a psychologically safe environment. This frees the way for employees to explore new ways of acting and take risks (West & Wallace, 1991). Additionally, when team members feel strongly like they belong to the team and feel attached to other team members, there is likely to be more cooperation, interaction and exchange of ideas (Hülsheger et al., 2009). However, attention should be paid to the fact that too much cohesion can have a detrimental effect on innovation. Social cohesion can be seen as a factor that encourages employees to achieve consensus in such a way that they cannot or will not express divergent opinions (Mpeera Ntayi et al., 2010), thereby decreasing the number of creative ideas. In extreme cases, this can lead to ‘groupthink’, where group members try extremely hard to reach consensus (Turner et al., 1992).

Teams being more creative (idea generation) and implementing new ideas can lead to more innovation (Hülsheger et al., 2009). The Euram 2018 conference paper (oil industry) also indicates that team CQ leads to increased team IWB.

(13)

13

2.3 Team leader CQ

One of the factors possibly influencing team CQ and team IWB is team leader CQ. Team leaders can play a role in influencing attitudes and behaviors of their team members. According to Tyler and Lind (1992), team members are very receptive to team leader behavior. Not only the content of this behavior is important, but also how a leader directs social processes. Team leaders also play an important role in the meaning-making process. They play an important role in motivating team members to change or learn something challenging that they do not know yet (Kotter, 1990; Zaleznik, 1981). Edmondson (2003) also states that team leaders help their team members to frame and reframe their knowledge and experiences. This can lead to team members thinking about what they know and learning to adopt different perspectives.

As the workforce in many companies is increasingly becoming multicultural, cross-cultural competences are essential to have for leaders in modern organizations (Groves & Feyerherm, 2011). Global leaders need CQ to know how to deal with relationships within a global business (Alon & Higgins, 2005). They need to gain an understanding of the regional and ethnic cultural diversity in their companies and cross-cultural workers all over the world (Alon & Higgins, 2005). Groves and Feyerherm (2011) found that leader CQ has a stronger association with team performance of multicultural teams than emotional intelligence and other leadership competences. Rockstuhl et al. (2011) also state that CQ is a critical leadership competency in a globalizing world. According to Alon and Higgins (2005), low levels of CQ lead to conflict that could be avoided, delays, stereotyping and failure of leadership. In other words, team leaders need to develop CQ to be successful in globalizing organizations. Global team leaders who possess higher levels of CQ are likely to be more effective as cross-border leaders (Rockstuhl et al., 2011).

(Team) leaders who possess higher levels of CQ are likely to better understand what impact their own cultural background has (Offerman & Phan, 2002; Rockstuhl et al., 2011). They will become aware of their biases regarding behavior at work and of expectations they have of themselves and others. Additionally, having a higher level of CQ leads them to reflect upon cultural assumptions, cultural knowledge and to think about values, biases and expectations that might be involved in cross-cultural interactions (Offerman & Phan, 2002; Rockstuhl et al., 2011). Also, higher levels of CQ enable global leaders to use both their deeper understanding of themselves and others and their motivation and behavioral flexibility to adjust their

(14)

14

leadership style such that it is appropriate in a distinct cross-cultural situation (Offerman & Phan, 2002; Rockstuhl et al., 2011). Rockstuhl et al. (2011) state that leaders who possess a higher level of CQ are more skilled at managing expectations of people with different cultural backgrounds and reducing exclusionary reactions prone to happen in cross-cultural situations. Furthermore, team leader CQ increases awareness of verbal and nonverbal (cultural) cues and helps to overcome misunderstandings and miscommunications and reduces conflicts in teams (Groves & Feyerherm, 2011). (Team) leaders are supposed to motivate, influence and enable employees to increase organizational effectiveness and performance (House et al., 2002).

Furthermore, Ramsey et al. (2017) found that team leaders who possess higher levels of CQ are more likely to use a transformational leadership style. Leaders who practice a transformational leadership style set a clear vision, act as a role model to motivate and inspire their employees and build morale to accomplish common goals (Ramsey et al., 2017; Howell & Avolio, 1993). They want their employees to reach their full potential and try to achieve this by paying attention to the needs and motives of employees (Ramsey et al., 2017). Understanding these different needs and motives is easier for leaders who understand how to deal with different cultures. Ramsey et al. (2017) argue that team leaders using a transformational leadership style are better able to comprehend differences between cultures and can adjust their behavior appropriately. Several authors report that transformational leadership is positively related to innovation and therefore is an important antecedent of innovation (e.g. Rosing et al., 2011; Mumford et al., 2002; Elenkov & Manev, 2009; Jung et al., 2003).

Based on this information, it is likely that when team leaders have a high level of CQ, they will encourage innovative work behavior in their teams, for example by removing misunderstandings and miscommunications. Therefore team leader CQ could be directly and positively related to team IWB. Furthermore, team leaders are likely to create more team fusion, as team members understand, respect and are aware of different cultures in their team while maintaining their own cultural identity. This means that the relationship between team leader CQ and team IWB might be mediated by team CQ. This literature review leads to the following hypotheses.

(15)

15

H1b: Team cultural intelligence positively and partly mediates the relationship between team leader CQ and team IWB.

2.4 Team learning behavior

Teams can learn as a group, which can enhance organizational performance and organizational learning (Bui et al., 2016). Team learning can be described as a process that aligns and develops the capability of the team to achieve the results it wants to achieve (Senge, 2006). According to Gibson and Vermeulen (2003), team learning is a process of multiple, interdependent team actions. They state that solutions (to problems) have to be found, chosen and implemented within teams. The process of team learning seems to exist of three actions: experimentation, reflective communication, and knowledge codification (Gibson & Vermeulen, 2003). According to Edmondson (1999), team learning consists of ongoing processes of actions and reflection upon those actions. These actions include asking questions and asking for feedback, experimentation, reflecting on results and discussing errors or unanticipated outcomes (Edmondson, 1999). Teams have to discuss differences in their opinions and test assumptions publicly, in order for the team to find holes in its plans and make appropriate adjustments (Edmondson, 1999).

Whether or not a team will engage in team learning behavior is influenced by the level of heterogeneity in the group (Gibson & Vermeulen, 2003). Moderate heterogeneity leads to less integration and more difficulties in communication and convergence, which makes it harder to exhibit group learning behavior. However, when there is very high heterogeneity, such that everyone is different, this leads to increased willingness to understand each other (Gibson & Vermeulen, 2003). This is important for multicultural teams in multinationals, as they will have different levels of heterogeneity in different teams.

When team members learn as a team, they achieve a higher level of collective intelligence than the sum of the intelligence of the individual team members (Senge, 2006). Through team learning, complex problems can be solved, knowledge can be created and performance can be improved (Kayes & Burnett, 2006). Team learning is also associated with increasing creativity (Hirst et al., 2009). Creativity can be seen as the first stage of innovation, since it encompasses the process of idea generation (Hülsheger et al., 2009). Creativity therefore is a sub process of innovation (Hülsheger et al., 2009; Egan, 2005), which indicates that team learning behavior can lead to increased IWB in teams. Team learning can also improve participation in innovation

(16)

16

(Bui et al., 2016; Molnar & Mulvihill, 2003, p. 172). Sarin and McDermott (2003) also state that innovation is a result of a team learning process.

Furthermore, according to Van den Bossche et al. (2006), mutual understanding and shared cognition are reached through interaction and discourse among team members. They consider this building of a common cognition to be team learning behavior. Mutual cognition can lead to team members learning from different viewpoints, as they learn and accept that different viewpoints exist and are legitimate (Van den Bossche et al., 2006). Also, team learning has been shown to have a positive effect on the quality of interpersonal relationships (Zellmer-Bruhn & Gibson, 2006). Team learning makes sure teams function effectively and that team members feel engaged in the team (Zellmer-Bruhn & Gibson, 2006). Senge (1994) states that through team learning, teams learn how to deal with the forces that hinder productive dialogue and discussion creatively. For these reasons, it is likely that team learning enables multicultural team members to understand and respect each other’s cultural viewpoints, increasing team metacognition, team fusion and perhaps openness to diversity. This increased team CQ makes it easier for employees to work together and develop, adopt and implement new ideas for products and work methods. For these reasons the following hypotheses have been developed.

H2a: Team learning behavior has a positive effect on team IWB.

H2b: Team cultural intelligence positively and partly mediates the relationship between team learning behavior and team IWB.

2.5 Trust

Trust leads individuals to expect that others will behave as expected and not be opportunistic (Barczak et al., 2010; Jarvenpaa et al., 1998). It is the confidence that someone has in the goodwill of another person and the expectation that they will behave in a beneficial way (Rockstuhl & Ng, 2008). Trust is an essential factor for effective team processes and performance (Ashleigh & Prichard, 2012; Costa, 2003). In work teams, the ability to work together is increased by trust (Barczak et al., 2010). When there is team trust, team members work together on achieving goals and providing added value (Mach et al., 2010). Trust also enables people to take risks (Jarvenpaa et al., 1998). Zand (1972) indicates that low trust between individuals leads to a spiral of reinforcement of low trust and that this can lead to a decrease in the effectiveness of problem solving. According to the conference paper Euram 2018 (oil industry), a lack of trust in teams can lead to poor communication, hostility, tension

(17)

17

and a lack of cooperation. On the other hand, when individuals do trust each other, this reinforcing pattern is constructive (Zand, 1972). Additionally, team trust enhances team creativity and a collaborative culture (Barczak et al., 2010). Creativity is one of the first stages of innovation as it is linked to idea generation (Hülsheger et al., 2009), which means that team trust can enhance IWB in teams.

Trust has consistently been found to have a positive influence on team cohesion (Mach et al., 2010; Calnan & Rowe, 2007; Hansen et al., 2002). According to exchange theory, trust is created through repeated exchanges between individuals (Ring & Van de Ven, 1992). Seeing someone act trustworthy makes individuals reciprocate this trustworthiness. This reinforcing effect of trust leads to more team cohesion (Mach et al., 2010). Being trustworthy, reliable and concerned about others can lead to team cohesion (Mach et al., 2010). This means trust can affect the relationship between team CQ and team IWB, since the team cohesion that is increased by team CQ will likely increase even more in teams where team members trust each other. Therefore, it is likely that team trust positively moderates the relationship between team CQ and team IWB. However, too much cohesion can have a detrimental effect on innovation as it can cause team members to think in the same way, thereby decreasing the number of creative ideas (Mpeera Ntayi et al., 2010). Therefore, this positive relationship is likely to exist when there is not too much cohesion.

For the reasons mentioned above, the last hypothesis is:

H3: Trust positively moderates the relationship between team CQ and team IWB.

2.6 Conceptual model Figure 1. Conceptual model

(18)

(19)

19

3. Methodology

In this chapter, the methods that are used to conduct the research will be discussed. The first paragraph introduces the research methods, after which the sample, data sources and measures will be discussed. Then the data analysis procedure is explained, to clarify how the results were retrieved. The fourth paragraph highlights the limitations of this research project and the last paragraph details how research ethics were taken into account.

3.1 Research strategy

The data was collected using a mixed method approach. Both quantitative and qualitative methods of data collection were be used to collect data from multicultural teams in different multinationals. This type of approach leads to data from different sources, which can be very complementary (Johnson & Onwuegbuzie, 2004). While the quantitative data can give a good idea about what relationships exist between variables, qualitative data uses statements about a phenomenon in reality (Johnson & Onwuegbuzie, 2004), which can clarify those relationships. The choice for multinationals was made based on the fact that multinationals usually contain (many) multicultural teams. The organizations that participated in this research were approached by dr. Bücker or through LinkedIn.

The first step in the collection of data was to send closed, structured and web-based surveys through e-mail to team members and team leaders of different multicultural teams in multinationals. The choice to use surveys as a data collection method was made since this method is suitable to gather information about many research units (teams), aspects and variables (Korzilius 2000, p. 9). This fits with the research question, as information about many variables and the relationships between them were investigated. Furthermore, this data collection method is the most suitable for a quantitative research, as a large amount of information can be analyzed to draw statistical conclusions. These conclusions can be used to answer the research question. The answers were recorded using 7-point Likert-scales, which resulted in an interval measurement level (Boone & Boone, 2012). This part of the research has a cross-sectional design, since the data was collected from the organizations at one point in time (Guest et al., 2012). Respondents were asked to rate themselves/their team on team CQ, trust, team leader CQ, learning behavior and team IWB. Furthermore, team IWB was rated by the team leader/supervisor. Employees rating themselves and their team could lead to respondents giving socially desirable answers and overestimation (Podsakoff & Organ, 1986). This means the survey items provoke responses that put the respondent in a favorable light instead of

(20)

20

responses of what they truly think (Podsakoff & Organ, 1986). Yet, the best way to know for example to what extent someone trusts a team member is to ask them. Therefore by using self-assessment, the most reliable information was collected on how people act and feel within their teams. A disadvantage of using surveys is that no follow-up questions could be asked to the respondents to find out why certain relationships exist. This was partly solved by also conducting interviews at another multinational.

When it appeared that the quantitative survey produced a relatively small number of respondents, it was decided to complement the quantitative study with a qualitative study. The opportunity arose to conduct in-depth interviews with a team of international product managers at another multinational. It was decided that this qualitative study could complement the quantitative study by finding general mechanisms that explain the effectiveness of and innovation within multicultural teams. The in-depth interviews were used achieve a fuller and deeper understanding of the respondents reasons, feelings, opinions and beliefs about these mechanisms (Legard et al., 2003). The mechanisms could then be used to further explain possibly found relationships in the quantitative study. The interviews were semi-structured, which makes it possible to find out how employees really experience certain situations (Barriball & While, 1994). Interview questions were formulated beforehand, but due to the semi-structured approach follow-up questions were also asked during the interviews to gather the information needed. The interviews were conducted in English as this is the main language used in the team and lasted on average about 45-50 minutes. The participants were informed by a short introduction about the goal of the research and the interview process. Furthermore, they were told that the collected data would be analyzed carefully and anonymously. The interviews were used to give more insight into the reasons behind possible relationships between the variables that were used in the quantitative study. This makes it easier to understand those relationships better. The interviews were returned to the interviewees for confirmation before the transcriptions were included as an appendix.

3.2 Sample, data sources and measures

3.2.1 Step 1 Quantitative study

3.2.1.1 Sample

The survey was distributed among several multinationals, to increase the diversity of the sample. The employees that filled out the survey had to be part of a multicultural team. Multicultural teams can be defined as: ‘A collection of individuals from different cultural

(21)

21

who see themselves and are seen by others as an intact social unity embedded in one or more larger social systems, and who manage their relationships across organizational boundaries and beyond’ (Halverson & Tirmizi, 2008, p. 5). One of the organizations was approached by

dr. Bücker and is a provider of technology and infrastructure for the energy industry based in The Hague. The company consists of multiple global teams that took part in the research. More team leaders working at different multinationals were approached via LinkedIn to participate and they were invited to forward the survey to the members of the team they lead and colleague team leaders at multinationals. The response rate could not be determined, as the it is not known how many team leaders actually forwarded the surveys and to how many people.

In the end, 48 team members and 26 team leaders filled out the survey. However, only the data of 36 team members and 20 team leaders was complete enough to use as the other respondents had not filled in more than 50% of the surveys. The employees that participated are all part of multicultural teams and the team leaders lead multicultural teams. In total, team members and team leaders combined, 87.5% of the respondents are male. Of the team members 83.3% is male and of the team leaders 95% is male. The average age of the team members is 40,6 years and the average age of the team leaders is 50,4 years. The team members are from 10 different nationalities and the team leaders from 5 different countries, but most team leaders and team members are European. 44.4% of the team members work in a team of 10 or more people. The majority of the team members (58.3%) has 10 or more years’ experience in working internationally. 65% of the team leaders has 10 or more years of experience with leading cross-cultural teams.

3.2.1.2 Data sources

The quantitative data was collected by distributing digital surveys among team members of multicultural teams in several multinationals. These team members rated their own team on the variables team CQ, trust and learning behavior. Team IWB and team leader CQ were rated by both team members and team leaders, as this provides a more objective image of team IWB and team leader CQ. Using different respondents to rate the dependent and independent variable also decreases the risk of common method bias (Podsakoff & Organ, 1986).

3.2.1.3 Measures

(22)

22

Team innovative work behavior

Team IWB can be described as intentionally creating, introducing and applying new ideas as a team, in order to benefit role performance, the group or the organization (Janssen, 2000). To measure team IWB, the scale developed by De Jong and Den Hartog (2010) to measure employee IWB will be used. This scale contains individual level items, but they will be adapted to the team level. This scale will contain 10 items (example item: “How often does the team search out new working methods, techniques or instruments?”). A 7-point Likert scale will be used ranging from 1 (never) to 7 (frequently).

Team cultural intelligence

Team CQ can be described as a (diverse) team’s abilities and knowledge that are necessary to function effectively and to adjust to different cultural interactions (conference paper, Euram 2018 (oil industry)). Team CQ will be measured using the scale developed by Bücker, Korzilius & Dijkman (in submission). This scale consists of 3 dimensions: Team Cultural metacognition, Fusion Teamwork and Openness to Diversity, with a total of 21 items. Example items include: “The team is conscious of the cultural knowledge it uses when interacting with people with different cultural backgrounds”, “The team uses a combination of norms or practices from different members’ cultures” and “The team enjoys doing jobs with people despite of language barriers”. A 7-point Likert scale will be used to record the answers, with answers ranging from 1 (completely disagree) to 7 (completely agree).

Team leader cultural intelligence

Team leader CQ will be measured using the scale developed by Groves and Feyerherm (2011). This 7-point Likert scale contains six items with answers ranging from 1 (strongly disagree) to 7 (strongly agree). Example items include: “The team leader has an understanding of the team members’ cultural backgrounds” and “The team leader has an understanding of the team members’ needs”.

Team learning behavior

Team learning can be seen as a process of reflection and action, team members testing assumptions and discussing differences of opinion openly inside the group to discover gaps in plans and make changes accordingly (Edmondson, 1999, p. 353). Examples of such activities are asking questions, seeking feedback, experimenting, reflecting on results and discussing errors or unexpected outcomes of actions (Gibson & Vermeulen, 2003; Edmondson, 1999).

(23)

23

Team learning behavior was measured with a combination of the 7-item scale of Edmondson (1999) and the 9-item scale of Gibson and Vermeulen (2003) as both of these scales are used often in research about team learning (e.g. Hirst et al., 2009; Kostopoulos et al., 2011). Example items are: “We regularly take time to figure out ways to improve our team” and “People in this team often speak up to test assumptions about issues under discussion”. The items will be measured with a 7-point Likert scale, which ranges from 1 (completely disagree) to 7 (completely agree).

Team trust

Trust is defined as: “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party” (Mayer et al., 1995, p. 712). According to Costa (2003), this is one of the most cited definitions of trust. Team trust will be measured using the 21-item scale of Costa and Anderson (2011). These items measure trust on four indicators: propensity to trust (example item: “Most people in this team do not hesitate to help a person in need”), perceived trustworthiness (example item: “In this team people can rely on each other”), cooperative behaviors (example item: “In this team we work in a climate of cooperation”) and monitoring behaviors (example item: “In this team people watch each other very closely”). The responses will be recorded using a Likert scale, ranging from 1 (completely disagree) to 7 (completely agree). Since monitoring behavior is an indicator of distrust, these items will have to be reversed to construct the total team trust score.

Demographics

The survey ended with some demographic questions, to gain some background information about the team members. Questions were asked regarding their age, gender, ethnicity, which company they work at, how long they have worked at this company, which team(s) they are a part of or which team(s) they are leading, how long they have been a member of that/those team(s), team size, amount of cultural diversity in their team and how long they have been working internationally. This information could then be used to describe the sample and as control variables.

(24)

24

Control variables were used to see whether the relationships between the dependent and independent variable(s) were not caused by any other aspects than the items taken into consideration. Therefore by controlling for these variables, the focus is on the relationship under investigation.

The first control variable taken into account is team size, as this factor is associated with influencing group dynamics and group performance (Brewer & Kramer, 1986). When groups grow, problems regarding communication and coordination are more likely to arise (Blau, 1970) which could reduce the generation of new ideas and therefore team IWB. On the other hand when groups are very small (2 or 3 people), the lack of diversity and perspectives might also hinder innovation (Jackson, 1996). Group size will be measured in terms of number of people per team.

Another factor that is taken into account is gender. According to Apesteguia, Azmat and Iriberri (2010), there are differences between gender compositions in teams when it comes to economic performance. There are differences in risk attitudes, social preferences, and preferences over competitive environment, which can lead to different economic and social outcomes. Bear and Woolley (2011) state that mixed gender teams perform better than teams solely consisting of men. In measuring gender, 1 stands for male and 2 stands for female.

The level of cultural diversity of the teams is also taken into account. In teams that have to deal with cultural diversity innovation and creativity are stimulated (Akdemir, Erdem & Polat, 2010). However, when there is cultural diversity in a team it is also possible that people categorize themselves and team members into in-groups and out-groups or do not want to work with people who do not have similar values, beliefs and attitudes (Stahl et al., 2010). Working together is essential to generate new ideas together, which can lead to innovation. Therefore team cultural diversity could possibly have both a positive and a negative effect on team IWB. Cultural diversity is measured in percentages, where 1 = 0-25% culturally diverse, 2 = 26-50% culturally diverse, 1 = 51-75% culturally diverse and 1 = 75-100% culturally diverse.

Lastly, self-report measures can cause respondents to answer in a socially desirable way (Podsakoff & Organ, 1986). To control for this effect, a short form of the Marlowe-Crowne scale for social desirability was used (Crowne and Marlowe, 1960; Reynolds, 1982). This scale consisted of 13 items, with answering possibilities of 1 (true) or 2 (false).

(25)

25 3.2.2 Step 2 Qualitative study

3.2.2.1 Sample

5 in-depth interviews were conducted with members of a product management team at a technological multinational based in Arnhem. The product managers mostly work in their own business department, but also spend time as product management team to align business processes. The team consists of 10-12 product managers, so the sample consists of about half of the product managers. One of the HR managers at the company sent an e-mail to the product managers asking them to participate in this research. In the first instance 4 product managers responded, after which a reminder e-mail was sent. 1 more product manager replied in response to this e-mail. The interviews were conducted with 1 team member with the Dutch nationality and 4 team members with all different foreign nationalities. This means many culturally different perspectives were taken into account.

3.2.2.2 Data sources

The qualitative data was collected by conducting in-depth, semi-structured interviews with five different members of the same team. Due to the different perspectives, a general image of the team’s behavior could be created. The respondents were asked to give their own opinions and experiences.

3.2.2.3 Qualitative measures

To gain insight into mechanisms that play a role in multicultural teams regarding effectiveness and innovation, several questions were developed as guideline for the semi-structured interviews. The questions were mostly concerned with team processes and mechanisms, such as cultural challenges, team trust etcetera. Questions regarding specific constructs were developed using operationalizations of these constructs in the literature. The questions regarding innovation were based on the operationalization of innovation by Cummings and O’Connell (1978). The questions regarding team CQ were based on the operationalization of the team CQ construct by Bücker, Korzilus and Dijkman (in submission). The questions relating to team learning were developed based on the operationalization of team learning by Edmondson (1999) and Gibson and Vermeulen (2003). Lastly, the questions related to team trust and psychological safety were developed based on operationalizations of the team trust construct by Costa and Anderson (2011) and the psychological safety construct by Edmondson (1999). For the most part the researcher used these questions, however if another interesting topic came up, follow-up questions were asked about that topic and in subsequent interviews

(26)

26

questions were also asked about the topic. For example after the first interview, a question about competition between the product managers was added to the interview questions. A complete overview of the interview questions can be found in appendix B.

3.3 Data analysis method

3.3.1 Step 1 Quantitative data

The data from the surveys was analyzed using SPSS. The first step in the data analysis strategy is to check for missing data. The rule of thumb is that missing values need to be less than 10% (Field, 2013). When an item has missing values over 10% or a respondent has not filled in over 50% of the survey, action, such as deleting that particular respondent, will need to be taken (Field, 2013). However, deleting respondents can affect the reliability of the research, so the aim was to avoid missing data. This was for example achieved by distributing the survey digitally, as it forced the respondents to fill out answers to all questions. Therefore if the respondent completely filled out the survey, all questions were answered. After checking for missing data the descriptive statistics and correlations were examined, and the data was reduced by performing a factor analysis. Next, the assumptions for regression analysis (normality, linearity, homoscedasticity and absence of multicollinearity) were checked to see if regression analysis was appropriate. Finally the data was analyzed using regression analysis by using the SPSS plug-in PROCESS (Hayes, 2013), as this plug-in is a helpful tool to conduct mediation and moderation analysis.

3.3.2 Step 2 Qualitative data

The first step in the qualitative data analysis was to transcribe the interviews. These transcriptions were sent back to the respondents for them to check and complement. No modifications of the transcripts were received. To translate the data from the interviews to useful information, the interviews were coded. Since the intention was to find mechanisms that influence team performance in multicultural teams in the data, an inductive coding approach was used (Symon & Cassell, 2012). Every interesting piece of text relating to team processes or effectiveness was given an open code. These codes describe the selected codes and turn them into manageable building blocks. These open codes were then combined into axial codes, which identified some relationships between the codes and grouped them into similar categories. In the third phase, selective coding was used to identify core concepts consisting of these categories. A codebook was developed to make sure pieces of text with the same type of information were given the same code.

(27)

27

3.4 Reliability and validity

Reliability refers to the fact that results are roughly the same if the research is repeated, while validity means that the researcher actually measures what he wants to know (Field, 2013). To improve both the reliability and validity of the research, triangulation of methods was used, meaning that different data sources were used (Field, 2013). Both a survey and semi-structured interviews were used to gather information from respondents.

Reliability could not be guaranteed, as the quantitative sample was very small and the issues with generalizability can exist. However, the transcripts and SPSS output will be made accessible, so the process can be repeated by other researchers. Furthermore, many scales had a Cronbach’s α of more than 0.60 and many were close to 0.80, which means the scales are reliable. Also, the use semi-structured interviews increases the reliability of the research, since all respondents were asked the same questions (Bleijenbergh, 2013).

Triangulation of data sources and research methods ensured internal validity. For the survey, existing measuring scales were used to make sure the concepts were measured correctly. In the interviews, the respondents were instructed to ask for clarification if they did not understand a question. The question would then be explained as good as possible. Clarification was also asked if answers were not clear. The respondents were also asked to check and change the transcript of the interview if needed which improves the validity. Additionally, control variables were used to see if the effects still exist if the control variables were added. External validity could not be guaranteed due to the low number of respondents.

3.5 Research ethics

In designing this research, ethical considerations have been taken into account. In the first place, the respondent is a source of information and has been treated with care. For that reason, respondents to the survey were assured of complete anonymity in the accompanying e-mail, confirming that neither their personal names nor the names of their organizations will be disclosed. The participants in the interviews were also assured that the data they provided would be analyzed carefully and completely anonymously. Since the interests of the participant might be conflicting with the interests of the researcher and/or the organization and respondents might have a tendency to present themselves in a positive way, this might have helped to decrease the tendency to give socially desirable answers (Podsakoff et al., 2003). It also guaranteed that the interaction with the (potential) participants is ethical and that the team members can choose to participate voluntarily. The participants (both of the survey and in the interviews) were not

(28)

28

forced in any way whatsoever to participate. Furthermore, the purpose of the research, writing a Master thesis, was shared. The data collected will not be used for other purposes than conducting this research and the participating organizations will receive a copy of the final result (if desired). That way, the participants can see the research and results they contributed to. The organizations in this research were informed about the fact that this Master thesis will be read by supervisors and the fact that this Master thesis will be accessible in an online database. Throughout the research process, attention was paid to the ethical responsibility.

(29)

29

4. Results

This chapter describes the results of the study. The quantitative and qualitative results will be analyzed and discussed. A detailed overview of the SPSS output is included in appendix D and the coded interview transcripts can be found in appendix E.

4.1 Quantitative analysis

4.1.1 Data preparation

The data was prepared for regression analysis. In total, there were 48 team members and 26 team leaders who responded to the survey. However, the data of 12 team members and 6 team leaders was not complete enough to use as less than 50% of the survey was completed. Therefore this data was filtered out before commencing with any further analysis. Most team members who did not complete the survey did not fill out any answers at all. It is likely that they just wanted to see what the survey looked like, which means perhaps the process of filling out the survey could have been explained more thoroughly. The 4 team members who did start the survey stopped before the questions about team CQ or before the questions about team learning behavior. The team leaders who did not complete the survey also did not fill out any data, which means they were probably also more curious about the research than committed to participate. The sample size consisted of 56 respondents, which according to Field (2013, p. 314) is enough for regression analysis if a large effect size is expected. This was the case as in previous research, large effect sizes were found between the variables. For example in the Euram 2018 conference paper (oil industry) it was found that the relationship between team CQ and team IWB is significantly positive with a medium to large effect size.

Half of the team member indicated that the level of cultural diversity within their teams is between 0 and 25% (the percentage of team members that are from different countries). 65% of the team leaders also indicated that the level of cultural diversity within the team(s) they lead is between 0 and 25%. This could mean respondents could have found it difficult to answer the questions regarding cultural intelligence.

To begin the data analysis, the items that were formulated negatively (Team_CQ_14, Team_Trust_10, Team_Trust_11, Team_Trust_16 and Team_Trust_17) were reverse coded.

4.1.1.1 Missing data analysis

After removing the cases with too little response, a missing data analysis was performed (table 1-5 in appendix D). There was no significant missing data for any variable, which makes sense

(30)

30

since all the questions in the questionnaire were mandatory. Therefore if a respondent finished the questionnaire they have to have filled out all the questions.

4.1.1.2 Outliers

To check for outliers, the skewness and kurtosis of the variables were examined and are displayed in table 4.1. These values should be between -1.96 and 1.96 for small samples (Field, 2013) and in any case between -3.29 and 3.29. In this case, 3 variables have a higher z-core for skewness than -1.96 and 2 variables have a higher z-score for skewness than -3.29. However, looking at the histograms, it does not seem like the skewness and kurtosis are too big to do a regression analysis.

Table 4.1

Skewness and kurtosis

Skewness Z-score Kurtosis Z-score Total_TIWB -0.721 0.557

Total_TLCQ -4.589 8.387 Total_TL -2.224 3.444 Total_TCQ -6.229 13.469 Total_TT -1.321 1.478

Looking at the boxplots, there is one outlier for all of the variables. 4 out of 5 of these outliers (on TIWB, TCQ, TLCQ and TL) were caused by the same respondent. Examining the data leads to the conclusion that this respondent gave quite low scores on most items, but there is variation in the answers. Therefore this does not look like a response set. The answers are theoretically possible and these answers could hold relevant information. There are also no signs that the respondent did not understand the questions. Therefore the decision was made not to remove this data. The outlier on team trust was also examined, but no reason was found to exclude this data from the dataset. These answers are also theoretically possible and may hold interesting information.

4.1.2 Data reduction

In order to reduce the data, a factor analysis was performed. First, the Kaiser-Meyer-Olkin measure and Barlett’s Test of Sphericity were checked to see in factory analysis was appropriate. The Kaiser-Meyer-Olkin test is a measure of sampling adequacy which value needs

(31)

31

to be higher than 0.5 (Field, 2013). Bartlett's test of sphericity is used to examine the hypothesis that the variables are uncorrelated in the population and the significance level of this test should be < .05 (Field, 2013). For team IWB, The Kaiser-Meyer-Olkin is 0.90, for team CQ it is .77, for team leader CQ the value is 0.87, for team learning it is 0.74 and for team trust the Kaiser-Meyer-Olkin value is 0.47. The The Kaiser-Kaiser-Meyer-Olkin value is greater than 0.5 for almost all variables and Barlett’s Test of Sphericity is significant for all variables. Only for team trust the Kaiser-Meyer-Olkin measure is smaller than 0.5, however it is very close to 0.5 and Barlett’s Test of Sphericity is significant, so factor analysis is still considered an appropriate method of analysis for this dataset.

4.1.2.1 Factor analysis team leader CQ

The initial analysis shows that there is 1 factor which explains 62.109% of the variance. No items were removed, as all items had a factor loading of more than 0.30.

Table 4.2

Rotated factor loadings team leader CQ

Item Factor 1 Team_Leader_CQ_1 0.837 Team_Leader_CQ_2 0.772 Team_Leader_CQ_3 0.750 Team_Leader_CQ_4 0.776 Team_Leader_CQ_5 0.803 Team_Leader_CQ_6 0.788

The Cronbach’s α of this scale is 0.874, which is higher than 0.85. Therefore the scale is reliable. Cronbach’s α could not be increased by deleting one of the items from the scale.

4.1.2.2 Factor analysis team learning behavior

After the initial analysis, there were 3 factors explaining 67.704% of the variance. However there were multiple cross-loaders which had to be removed. These items were (in order of removal): Team_Learning_14, Team_Learning_2, Team_Learning_7 and Team_Learning_10. After these items were deleted, the 3 factors remained and they explain 70.048% of the variance. Table 4.3 shows the factor structure.

(32)

32

Table 4.3

Rotated factor loadings team learning behavior

Item Factor 1 Factor 2 Factor 3

Team_Learning_1 0.733 Team_Learning_3 0.769 Team_Learning_4 0.762 Team_Learning_5 0.783 Team_Learning_6 0.898 Team_Learning_8 0.797 Team_Learning_9 0.790 Team_Learning_11 0.742 Team_Learning_12 0.729 Team_Learning_13 0.811 Team_Learning_15 0.862

Factor 1 consists of 8 items and has a Cronbach’s α of 0.908, which is quite high. The internal consistency could not be improved by deleting one of the items. This factor was named TL Attitude (TLAttitude), as all items relate to an attitude that is open to learning.

Factor 2 consists of 2 items and has a Cronbach’s α of 0.408. This is very low and since the factor only consists of 2 items, no item can be deleted to improve Cronbach’s α. Therefore the items will be taken into account in the analysis as separate items. Team_learning 3 is named TL conflict (TLConflict) and Team_learning 4 is named TL Rrole model (TLRolemodel).

Factor 3 only consists of 1 item (Team_Learning_6), which means no reliability analysis could be done. This factor is named TL External, as is concerns team members collecting information from external sources.

4.1.2.3 Factor analysis team cultural intelligence

The initial analysis of team CQ points to 5 different factors underlying team CQ These 5 factors have an Eigenvalue greater than 1 and explain 77.603% of the variance together. After the first rotation, some cross-loaders were discovered. Items Team_CQ_12 and Team_CQ_11 were removed first, as they had the lowest difference between two factor scores (0.014 and 0.020 respectively). After their deletion, 4 factors remained. There were still some cross-loaders, so

(33)

33

the items: Team_CQ_5, Team_CQ_13 and Team_CQ_15 also had to be deleted. The 4 factors explain 76.198% of the total variance. The structure that remained is shown in table 4.4.

Table 4.4

Rotated factor loadings team CQ

Item Factor 1 Factor 2 Factor 3 Factor 4

Team_CQ_1 0.774 Team_CQ_2 0.749 Team_CQ_3 0.846 Team_CQ_4 0.819 Team_CQ_6 0.696 Team_CQ_7 0.634 Team_CQ_8 0.801 Team_CQ_9 0.784 Team_CQ_10 0.770 Team_CQ_14 0.597 Team_CQ_16 0.710 Team_CQ_17 0.885 Team_CQ_18 0.845 Team_CQ_19 0.848 Team_CQ_20 0.827

(34)

34

For factor 1, the internal consistency was very high with a Cronbach’s α of 0.906. This number could not be increased by deleting one of the items. This factor was named TCQ Fusion, since the items Team_CQ_19 and Team_CQ_20 also can be related to the co-existence dimension of team fusion (TCQFusion).

Factor 2 also has a high Cronbach´s α of 0.831, but this internal consistency could be improved to 0.887 if Team_CQ_14_R was deleted. This is an increase of more than 0.005 and for that reason it was decided to delete Team_CQ_14. After deleting this item, Cronbach’s α could be increased even further, to 0.901, by deleting item Team_CQ_6. As this is also quite a big increase in Cronbach’s α, this item was also deleted. Factor 2 therefore consists of Team_CQ_1, team_CQ_2 and Team_CQ_3. This factor was named TCQ Metacognition, because all items are about the knowledge of cultural knowledge (TCQMeta).

Factor 3 has a Cronbach’s α of 0.870, which is also high enough. The internal consistency could not be improved by removing one of the items. This factor consists of Team_CQ_16, Team_CQ_17 and Team_CQ_18 and is named TCQ Openness, since all items are about openness to diversity (TCQOpen).

Lastly, factor 4 has a Cronbach’s α of 0.712 which is not a very high Cronbach’s α, but it is acceptable (Field, 2013). No items could be deleted to increase Cronbach’s α, as the factor only consists of 2 items. Therefore factor 4 consists of Team_CQ_4 and Team_CQ_9 and is named TCQ Interaction as these items can be related to the meaningful participation dimension of team fusion (TCQInteract).

4.1.2.4 Factor analysis team trust

After the initial analysis, 7 factors appear to explain 78.566% of the total variance of team trust and have an Eigenvalue that is greater than 1. However, since there were some cross-loaders (with a difference between factor loadings that was smaller than 0.20), those had to be removed. After removing the factors: Team_Trust_9, Team_Trust_13, Team_Trust_16, Team_Trust_5, and Team_Trust_18, a structure with 5 factors that explain 71.526% of the variance was left. The factor structure is shown in table 4.5.

(35)

35

Table 4.5

Rotated factor loadings team trust

Item Factor 1 Factor 2 Factor 3 Factor 4 Factor 5

Team_Trust_1 0.700 Team_Trust_2 0.777 Team_Trust_3 0.868 Team_Trust_4 0.833 Team_Trust_6 0.667 Team_Trust_7 0.726 Team_Trust_8 0.827 Team_Trust_10 0.785 Team_Trust_11 0.846 Team_Trust_12 0.697 Team_Trust_14 0.813 Team_Trust_15 0.745 Team_Trust_17 0.654 Team_Trust_19 0.657 Team_Trust_20 0.865 Team_Trust_21 0.905

Factor 1 has a Cronbach’s α of 0.879, which is quite high. This number could not be increased by removing one of the items. The factor consists of items Team_Trust_2, Team_Trust_6, Team_Trust_7, Team_Trust_8, Team_Trust_12, Team_Trust_14 and Team_Trust_15 and was named TT cooperation, as all items are related to cooperation (TTCoop).

Factor 2 has a Cronbach’s α of 0.798. This number could be increased to 0.822 by deleting item Team_Trust_19. As this increase is bigger than 0.005, it was decided to delete this item from the scale. Item Team_Trust_19 is quite similar to item Team_Trust_21, so from a theoretical perspective this removal is also justified. The factor consists of the items Team_Trust_20 and Team_Trust 21 and was named TT monitoring (TTMon).

Factor 3 has a Cronbach’s α of 0.694. This is not very high (>0.80), but it is an acceptable level (>0.60). Since the factor only consists of 2 items (Team_Trust_10 and