1
The effects of a climate for inclusion on the relationship
between faultlines and team performance
Business Studies – Leadership & management Track
August 15, 2014
Master thesis
Author: Jos Theo Dames
Student number: 10647791
Supervisor: Dr. C. Buengeler
2
Table of contents
Abstract 3 1. Introduction 4 2. Literature Review 8 2.1. Faultlines 92.1.1. The concept of faultlines 9
2.1.2. The effects of faultlines 10
2.1.3. Faultlines theories 11
2.1.4. Social category-based and informational category-based faultlines 15 2.2 The moderating role of a climate for inclusion 18
3. Methods 22
3.1. Research methods and data collection 22
3.2. Sample 23 3.3 Measures 25 3.3.1. Independent variable 25 3.3.2. Dependent variable 27 3.3.3. Moderating variable 28 3.3.4. Control variable 28 3.4. Data aggregation 29 4. Results 30 4.1. Descriptive Analysis 30 4.2 Hypothesis testing 31 5. Discussion 40 5.1. Discussion 40 5.2. Practical implications 45
5.3. Limitations and recommendations for future research 45
5.4. Conclusion 47
Acknowledgements 47
References 48
Appendices 57
Appendix A: Survey 57
Appendix B: Table 1. Means, standard deviations, correlations. 63 Appendix C: Table 2. Hierarchical Regression table. 64
3
Abstract
Following prior research, this study divides faultlines into two types: social category-based faultlines and informational category-based faultlines. An attempt to assign specific effects to the respective types of faultlinesdid not yield significant results. Furthermore, a climate for inclusion was proposed to moderate the relationships between the two respective faultlines types and team performance. A climate for inclusion did however not significantly moderate the relationship between social category-based faultlines and the team performance outcomes. It also did not significantly moderate the relationship between informational category-based faultlines and efficiency, overall achievement and team productivity. It did significantly moderate the relationship between informational category-based faultlines and quality, such that this relationship was more negative at high levels of a climate for inclusion than when a climate for inclusion was low.
4
1. Introduction
In today’s business environment, more and more organizations work with teams (Burke, Stagl, Klein, Goodwin, Salas & Halpin, 2006). Organizing the workforce in teams has become a popular phenomenon. With this increasing interest in teams comes an interest in team performance and how it can be affected. One way of influencing team performance that has been given great attention is the concept of diversity in teams (Thatcher & Patel, 2012). Van Knippenberg and Schippers (2007) describe diversity as “a characteristic of social grouping that reflects the degree to which objective or subjective differences exist between group members” (p 516). The authors state that suggested benefits of team diversity are improved decision-making, high-quality solutions, higher creativity and innovation, improved team performance. However, the very heterogeneity that creates the benefits of diverse teams also presents challenges for the team’s effectiveness (Horwitz & Horwitz, 2007; Van Knippenberg & Schippers, 2007).
Indeed, although diversity was found to influence team performance, research yielded inconsistent results as to how it influences team performance (Van Knippenberg, De Dreu & Homan, 2004; Meyer & Glenz, 2013). In the diversity literature, two perspectives are extensively researched: the social categorization perspective and the information/decision-making perspective. The social categorization perspective argues that similarities and differences among team members make individuals categorize themselves and others into groups, where team members with similar attributes are seen as ingroup members and team members with dissimilar attributes are seen as outgroup members. This perspective suggests that such processes be detrimental to team performance. By contrast the information/decision-making perspective argues that diverse teams will outperform homogeneous groups. The reasoning behind this
5
perspective is that diversely (as compared to homogeneously) composed teams have a broader range of knowledge, skills, abilities and perspectives to work with. This can lead to more creative and innovative ideas and solutions (Ancona & Caldwell, 1992; Bantel & Jackson, 1989; De Dreu & West, 2001). In short, the social categorization perspective suggests a negative effect of diversity on team performance due to the creation of sub-groups that divide the team and undermine collective functioning, and the information/decision-making perspective suggests a positive effect on team performance due to the broadened range of knowledge, skills, abilities and perspectives that heterogeneous groups are likely to possess (Triandis, Kurowski, & Gelfand, 1994; Van Knippenberg, De Dreu & Homan, 2004).
Trying to improve the understanding of the nature and effects of diversity has been the central theme in team diversity research during the past thirty years (Thatcher & Patel, 2012). However, the attempts to find a main effect for a specific type of diversity (e.g. gender diversity) have not yielded consistent results (Meyer & Glenz, 2013). This means diversity research has not yet developed a normative theory for how to approach specific types of diversity. So, researchers have been searching for ways to approach team diversity in a different way. One of the most interesting of these additions to the field of diversity research has been the conceptualization of faultlines (Lau and Murnighan, 1998).
Lau and Murnighan (1998) define faultlines as “hypothetical dividing lines that may split a group into sub-groups based on one or more attributes” (p 328). Faultline theory says that the alignment of diversity attributes can cause the form of sub-groups. Like diversity itself, faultines can have positive and negative effects on team performance. Most research on faultlines however, has focused on either trying to counter the negative aspects of faultlines (Rico, Molleman, Sánchez-Manzanares & Van der Vegt, 2007, Lau & Murnighan, 1998; Thacher et al.,
6
2003), or provide argumentation for why faultlines might be a good thing in teams (Nishii & Goncola, 2008; Gibson & Vermeulen, 2003).
As faultlines have both positive and negative aspects, I propose that instead of trying to prevent the emergence of faultlines or merely trying to deal with it’s negative aspects, teams should learn to live with faultlines and try to reap its potential benefits. Thus far, faultlines literature has not been able to do this. The current study will strive to find a way to simultaneously counter the negative effects of faultlines, and savor or increase the positive effects of faultlines. This study will use an idea by Bezrukova, Jehn, Zanutto and Thatcher (2009) to do this. In an attempt to link the specific effects of faultlines to specific types of faultlines, Bezrukova and collegues (2009) divided faultlines in two types: social category-based faultlines and informational category-based faultlines. This division is based on the nature of the faultlines. In this thesis, social category-based faultlines and informational category-based faultlines will be calculated on the basis of age and gender (social category-based characteristics) and level of education and tenure (informational category-based characteristics), respectively. These attributes were chosen based on their use in prior research (Jehn, Chadwick & Thatcher, 1997; Tsui, Egan & O’Reilly, 1992; Bezrukova et al., 2009).
This study will research the respective relationships of social category-based faultlines and informational category-based faultlines with team performance. Moreover, I suggest that a climate for inclusion (Nishii, 2013) could act as a moderator on these relationships. Nishii explains that a climate for inclusion “involves eliminating relational sources of bias by ensuring that identity group status is unrelated to access to resources, creating expectations and opportunities for heterogeneous individuals to
7
establish personalized cross-cutting ties, and integrating ideas across boundaries in joint problem solving” (2013, p.1754). These aspects of a climate for inclusion seem well fitted to positively influence the relationship between faultlines and team performance. Figure 1 depicts the research model of this research.
This thesis is structured as follows. First, an extensive literature review will be presented. Based on the literature review, the Hypotheses that can be seen in Figure 1 will be formulated. Subsequently, the methods of this research will be discussed. After this, the Hypotheses will be tested, and the results of the analyses will be presented. Finally, a discussion on the findings of this research and the limitations and recommendations for future research will conclude this thesis.
8
2. Literature review
2.1. Faultlines
2.1.1. The concept of faultlines
In an effort to extent the understanding of group dynamics, Lau and Murnighan (1998) argued that researchers should look beyond individual single dimensions of diversity. More dimensions of diversity can be present in a team, and these dimensions can align and subsequently affect the team. Focusing on a specific dimension of diversity thus is not enough to capture the effects of such converging diversity dimensions. In order to address this notion, Lau and Murnighan (1998) introduced their faultlines concept. It looks at the composition of teams on multiple diversity attributes such as age and gender, rather than at a single attribute such as gender. These diversity attributes can have a range of effects on the team. If different diversity attributes were to align, this could result in a strong division between sub-groups which can be detrimental to social integration and the information sharing process. When different diversity attributes cross-cut however, the boundaries between such sub-groups become weaker (Sawyer, Houlette & Yeagley, 2006).
The concept of Lau and Murnighan (1998) focuses on the alignment of diversity attributes. The authors define faultlines as “hypothetical dividing lines that may split a group into sub-groups based on one or more attributes” (p 328). The alignment of diversity attributes can cause the form of sub-groups. For example, a team where all team members with a Master’s degree have been with the company for less than 10 years, and members of the team that have only finished high school have been with the company for more than 30 years. Here, the
9
alignment of the diversity attributes level of education and tenure could create sub-groups within the team.
The example above explains faultline strength. The strength of the faultline is determined by how many attributes of diversity converge within the sub-groups, or in other words, how clear a group can be divided into two or more sub-groups that are internally quite similar. Another aspect of faultlines is faultline distance (Bezrukova et al., 2009). This faultline aspect shows how
far apart the sub-groups are on the respective diversity attributes. For example, a team with a sub-group with 20-year old men and a sub-group with 50-year old women has a higher faultline distance than a team with a sub-group with 20-year old men and a sub-group with 30-year old women. In this example, the former team has higher faultline distance because the sub-groups are further apart on the age attribute. Concluding, whereas faultline strength focuses on similarities within the groups, faultline distance emphasizes the differences between the sub-groups (Bezrukova et al., 2009; Zanutto, Bezrukova, & Jehn, 2011). Bezrukova et al. (2009) found that in order to accurately assess the faultlines present in the team, it is important to capture both faultline strength and faultline distance when measuring faultlines. Faultline distance can exacerbate the effects of faultline strength, and this interaction can be captured by a composite term that includes both measures (Bezrukova et al, 2009).
The definition of faultlines as ‘hypothetical dividing lines’ (Lau & Murnighan, 1998, p. 328) expresses the idea that faultlines in a team often represent dormant, or potential faultlines (Jehn & Bezrukova, 2010). Measuring dormant faultlines means calculating the probability that these dividing lines will actually emerge, based on demographic or other characteristics. Dormant faultlines can become active when sub-groups are actually perceived by the team members based on the respective diversity characteristics. This can be the result of a ‘faultline
10
trigger’, an occurence or a situation which turns the dormant faultline into an active one (Rink & Jehn, 2010).
2.1.2. The effects of faultlines
Faultlines research has delivered quite consistent results that point to a negative effect of strong faultlines on team outcomes (e.g. Lau & Murnighan, 1998; Thacher et al., 2003; Rico, Molleman, Sánchez-Manzanares & Van der Vegt, 2007; Homan, Van Knippenberg, Van Kleef & De Dreu 2007). These negative effects include lower trust between sub-groups, lower willingness to cooperate (Brewer, 1979; Brewer & Brown, 1998; Tajfel & Turner, 1986), less social integration (Rico, Molleman, Sánchez-Manzanares & Van der Vegt, 2007), lower levels of morale, increased conflict, and lowered team performance (Lau & Murnighan, 2005; Thatcher et al, 2003). Thatcher and Patel (2012) argue that unlike diversity research, faultlines research has been yielding consistent results. However, the general assumption that faultlines have a negative effect on team performance has been criticized (Bezrukova et al., 2009; Rico, Molleman, Sánchez-Manzanares & Van der Vegt, 2007).
As mentioned before, diversity research has shown that diversity can have both positive and negative effects on team performance. Multiple researchers have shown that the same is true for faultlines. Stronger faultlines can improve a team's creativity (Nishii & Goncola, 2008), improve its decision-making processes (Bezrukova et al., 2009) and increase the team learning (Gibson & Vermeulen, 2003). Thus, it can not be said that research of the effects of faultlines has yielded consistent results. In order to understand how faultlines work, and what effects they are likely to have, it is paramount to understand the underlying theories of faultlines. The following section will thus explore faultlines theory.
11 2.1.3. Faultlines theories
Understanding the theories explaining the emerge and effects of faultlines is vital if one wants to find a way to influence its relationship with team performance. First, the underlying theories of the emergence of faultlines will be discussed. The faultlines concept developed by Lau and Murnighan (1998) originally built on the social identity theory (Brewer, 2010) and the self-categorization theory (Turner, 1985). These theories argue that similarities and differences among team members make individuals categorize themselves and others into groups, where team members with similar attributes are seen as ingroup members and team members with dissimilar attributes are seen as outgroup members (Brewer, 2001; Turner, 1985). When looking at these arguments, it is useful to know on what basis people find individuals that are similar to themselves. It was found that people use characteristics that are salient to them in order to differentiate between similar and dissimilar individuals (e.g., demographic characteristics; Horwitz & Horwitz, 2007). This phenomenon is referred to as social categorization and is also used to explain the emergence of subgroups and the resulting outcomes in the diversity (cf. social categorization paradigm; Tajfel, Billig, Bundy & Flament, 1971), and seems to provide a partial explanation for the rise of faultlines in teams.
Another theory that can help understand the emergence of faultlines is Brewer’s (1991)
Optimal Distinctiveness theory (ODT). Brewer argues that one’s social identity is formed through a delicate balance of a person’s need to be an individual and to be unique on the one hand, and a person’s need to be validated and to belong to a group on the other hand. This principle can indeed explain something about faultlines. The simultaneously acting needs of belongingness and uniqueness drive a person to want to belong to certain (sub-groups of) people
12
and make the person want to be different from other (sub-groups of) people, respectively. Together with the social identity and self-categorization, the ODT provides an explanation for why and how faultlines emerge in diverse teams.
Next, the underlying theories of the effects of faultlines are discussed. First of all, it was mentioned above that individuals see team members with similar attributes as ingroup members and team members with dissimilar attributes as outgroup members (Brewer, 2001; Turner, 1985). The similarity-attraction paradigm (Byrne, 1971) adds to the understanding of this phenomenon, as it explains that people are prone to individuals similar to themselves, and thus are more attracted to people that are similar to themselves than to people that are different to themselves. The Categorization Elaboration model (CEM) by Van Knippenberg, De Dreu, and Homan (2004) can be used to add to the understanding of the effects of faultlines. In their paper, Van Knippenberg et al. (2004) state among other things that when people are different in more than one diversity attribute, these attributes can align or cross-cut. Social categorization is more likely in teams where attributes align than in teams where attributes cross-cut. When multiple diversity dimensions align, the boundaries between sub-groups become stronger, as ingroup members are more similar to the individual and outgroup members are more dissimilar to the individual. Such a distinction between ingroup and outgroup members can result in intergroup bias. The concept of intergroup biases is explained by Van Knippenberg et al. (2004) as the tendency of people to react more positively towards ingroup members than they would to outgroup members. Intergroup bias is the cause for many of the negative effects associated with faultlines (Van Knippenberg et al., 2004). Intergroup bias is more likely to negatively affect the team when people feel threatened in their subgroup-related identities. Just as people like to have a positive image of the self, (sub-)groups like to have a positive image of the group identity. This
13
group identity may be challenged when other (sub-)groups are perceived to be higher in status, or when the (sub-)group is discriminated against (Van Knippenberg et al., 2004). Interestingly, the group identity can also become threatened when it is being suppressed by, for example, attempts to create an organization-wide shared identity (Hornsey & Hogg, 2000). Ergo, although such attempts aim to bring people together through a shared identity, it can paradoxically aggravate the effects of intergroup bias and divide the group even more.
The CEM also adds more understanding to the concept of social categorization and its effects. According to the model, whether or not the inter-member differences in a group will cause problems depends on the categories that people make for themselves. More specifically, the salience of the categorization determines the degree to which differences among team member cause problems. This salience of the categorization is explained by cognitive accessibility, the normative fit, and the comparative fit of the categorization. The cognitive accessibility of the categorization determines how easy it is for a person to place individuals in certain categories, and the ease with which that person can recall the categorization. The normative fit of the categorization says whether or not the categorization seems logical to the person. The comparative fit describes how similar the people in one group are and how dissimilar people are to people in the other groups. This theory about social categorization is important to understand when trying to influence the relationship between faultlines and team performance.
Another stream of research uses ‘distance theories’ to explain negative effects of faultlines. The social (Hraba, Hagendoorn & Hagendoorn, 1989), psychological (Jetten, Spears & Postmes, 2004) and cultural (Leong & Ward, 2000) distance theories explain that the more different sub-groups are from one another, the more ‘distant’ they become. According to these
14
distance theories, the ‘distance’ between sub-groups determines group outcomes to a certain degree (Thatcher & Patel, 2012), as large distances between sub-groups can negatively affect relations between these sub-groups (Jetten et al., 2004). More specifically, sub-groups become more distant from one another when they are further apart on certain diversity attributes. In faultline theory, the ‘distance’ of the faultline is established in the same manner.
The theories above help explain the negative effects of faultlines. But, as was mentioned earlier in this chapter, faultlines have been found to positively affect teams as well. The CEM proposes that diversity is positively related to the elaboration of task-relevant information and perspectives, which in turn positively affects group performance (Van Knippenberg et al., 2004). This assumption is based on the information/decision-making perspective (Ancona & Caldwell, 1992; Bantel & Jackson, 1989; De Dreu & West, 2001). This perspective views cognitive diversity as a cherished resource, and is sometimes referred to as the cognitive resource perspective (Bantel & Jackson, 1989; Bezrukova et al., 2009; Webber & Donahue, 2001). Here, the term cognitive diversity can be interpreted to mean diversity in, among other things, skills, knowledge and experiences. The CEM also proposes that diversity in teams is most likely to lead to elaboration of task-relevant information and perspective when the team has strong information processing and decision-making mechanisms.
The optimal distinctiveness theory (Brewer, 1991) can also explain some of the positive effects of faultlines. Nishii and Goncola (2008) use the theory to explain that faultlines can be positively associated with team outcomes, such as team creativity. The authors argue that the sub-groups that faultlines create have two positive consequences: the safety of the ingroup (fulfillment of validation and belonginess) makes people more confident to suggest ideas, and the interaction with dissimilar sub-groups (fulfillment of uniqueness) makes informational
category-15
based differences between people more salient, which should increase the creativity of the ideas. Similarly, Gibson and Vermeulen (2003) states that people feel more 'psychologically safe' when suggesting ideas from within the safety and with the support of a sub-group. The authors found the strength of sub-groups ("i.e., the degree of overlap in demographic characteristics", Gibson & Vermeulen, 2003, p. 204) was positively related to team learning. The authors even suggest that the emergence of sub-groups was a “healthy influence” on the teams (p.230).
These underlying theories of faultlines help understand the positive and negative effects of faultlines. However, if one wants to attempt to influence the effects of faultlines on team performance, it can be usefull to assess what the effects of the different aspects of faultlines are. In an attempt to link specific effects of faultlines to specific attributes of diversity, Bezrukova and collegues (2009) have made a distinction between two types of faultlines: social category-based faultlines and informational category-category-based faultlines.
2.1.4. Social category-based and informational category-based faultlines
Defending the assumption that faultlines do not necessarily result in only negative or only positive effects, Bezrukova et al. (2009) have proposed to extent the faultlines framework by arguing that faultlines can be divided into two types of faultlines: social category-based faultlines and informational category-based faultlines. The authors state that each of the two types of faultlines influences team performance in its own way. Their paper defends the notion that the
nature of the faultline should be taken into account when analyzing the concept. According to
the authors, faultlines can arise on the basis of social category characteristics and on the basis of informational category characteristics. Social category-based characteristics are individual attributes that are often salient, and that are not directly related to the job and its tasks, but
16
influence the team members’ behaviors and perceptions. Among others these individual attributes include age, gender, race, and nationality. Conversely, informational category-based characteristics are often less salient and are directly related to the tasks of the team members. They are often main contributors to the successful attainment of team goals. These characteristics are among others the level of education, tenure and area of expertise of an individual.
Bezrukova et al. (2009) have suggested that diversity in social category-based attributes can be the cause of negative effects on team performance. It should be said that social category diversity can also have positive effects on the team (e.g. Jehn, Northcraft & Neale, 1999) and that there are also negative team outcomes associated with diversity in informational characteristics (Cummings, Zhou & Oldham, 1993; Wagner, Pfeffer and O’Reilly, 1984). Van Knippenberg et al. (2004) suggest that every diversity attribute may lead to social categorization as well as informational elaboration. The authors argue that some social category-based attributes may not directly be seen as related to the task, but that they still may cause information elaboration. On the other hand, some information category-based attributes may not seem to cause social categorization, but the same attributes may still be used to categorize between in- and outgroups by certain groups.
These propositions are important to consider, and it is indeed possible that not all social category-based and informational category-based faultlines affect team performance in their same respective way. However, the negative effects of faultlines are mainly associated with the effects of social categorization. According to the CEM, social categorization is more likely to result in negative effects when the attributes on which basis people are categorized are more salient. As social category-based characteristics such as age and gender are prime examples of such salience, the literature seems to suggest a negative relation between social category-based
17
faultlines and team performance. Indeed, Bezrukova et al. (2009) found that social category-based faultlines were negatively associated with team performance. Therefore, also considering the suggestion by Bezrukova et al. (2009) that both faultline strength and faultline distance should be taken into account when measuring faultlines, Hypothesis 1 is formulated as follows:
H1: Social category-based faultlines in terms of strength and distance are negatively related to team performance.
According to the information/decision-making perspective and the cognitive resources perspective, diversity in informational category-based characteristics can result in positive effects on performance. It has been argued that, because such informational category-based characteristics directly contribute to the common goal, team members could well begin to value the differences between sub-groups and use those differences to increase team performance (Webber & Donahue, 2001). Diversity in such information category-based characteristics has been generally found to be associated with positive team outcomes such as higher levels of team innovation and effective decision-making processes (Bantel & Jackson, 1989; Bezrukova et al., 2009). Furthermore, the emergence of strong sub-groups can help teams increase team creativity (Nishii & Goncola, 2008) and team learning (Gibson & Vermeulen, 2003). Hence, the literature seems to suggest a positive relation between informational category-based faultlines and team performance. Therefore, again considering the suggestion by Bezrukova et al. (2009) that both faultline strength and faultline distance should be taken into account when measuring faultlines, Hypothesis 2 is formulated as follows:
18
H2: Information category-based faultlines in terms of strength and distance are positively related to team performance .
2.2. The moderating role of climate for inclusion
With the knowledge that faultlines can have both positive and negative effects on team performance - depending on the nature of the faultlines and the context of the situation - the goal should not be to merely find a way to counter the negative effects of faultlines, but also to savor and possibly increase the positive effect that faultlines can have on the team.
Accordingly, strategies are required that effectively reduce or eliminate the presumably negative effects of social category-based faultlines while optimizing potentially positive effects of informational category-based faultlines. Based on the theories of faultlines, various aspects should be included in the moderator to counter negative and/or increase positive effects of faultlines. It is important that the moderator would be able to adequately resolve and prevent such intergroup biases, as this is one of the main causes of the negative effects of faultlines. Another important concept with respect to moderation is cross-categorization. Cross-categorization takes place when sub-groups have a common denominator in the form of a diversity attribute (Sawyer, Houlette &Yeagley, 2006). For example, although sub-groups have emerged in a given group of people, all sub-groups share the fact that they include a woman. Brewer (2000) explains that it can reduce the intergroup bias that is caused by the differences among groups. Moreover, the crossing of sub-groups can results in team members from different sub-groups appreciating each other’s informational category-based differences and effectively using their collective cognitive capital through information elaboration (Homan, Van
19
Knippenberg, Van Kleef & De Dreu, 2007). Therefore, a good moderator would include some form of cross-categorization of faultlines (Homan, Van Knippenberg, Van Kleef & De Dreu). Faultlines research has focused on several distinct moderating factors that determine the effects of faultlines. In the research by Rico et al. (2007), for instance, task autonomy moderated the negative relationship between faultlines and decision quality and the negative relationship between faultlines and social integration (Rico, Molleman, Sánchez-Manzanares & Van der Vegt, 2007). Although the effect was only significant for high levels of task autonomy, the authors suggest that team task characteristics can possibly act as important moderators of the effects of faultlines on team performance. In another study, Jehn and Bezrukova (2010) used team identification to moderate the relationship between activated faultlines and group processes such that a strong team identity decreased the probability that activated faultlines caused negative effects on team processes.
However, because diversity research has failed to deliver consistent findings on the effects of diversity, Sawyer, Houlette and Yeagley (2006) suggest that these effects could be dependent on the social context of group processes. More specifically, it could be that the positive effects of diversity can only be reaped when the group actively engages in debate (Simons, Pelled, & Smith, 1999). Likewise, Van Knippenberg et al. (2004) propose that diversity is most likely to lead to the positive effects of informational elaboration if the team is highly motivated to process task-relevant information and perspectives. These propositions suggest that a good moderator would be able to set a certain social context.
In one of the first researches to have social context at the organizational level act as a moderator, Nishii (2013) used a climate for inclusion to moderate the relationship between conflict and gender-diverse groups, such that higher levels of a climate for inclusion decreased
20
conflict in gender-diverse groups. Among other things, a climate for inclusion is supposed to decrease interpersonal bias through fairly implemented employment and diversity-specific practices, as well as to introduce cross-categorization exercises and actively promote shared decision-making and information elaboration. The latter provides a social context in which the positive effects of diversity could be reaped (Sawyer, Houlette and Yeagley, 2006; Simons, Pelled, & Smith, 1999; Van Knippenberg et al., 2004). Moreover, a climate for inclusion does not try to form a shared identity, which could threaten the sub-group identities and lead to intergroup biases, but rather values the differences and creates a safe environment where individuals feel comfortable displaying their true self, and can easily identify. These aspects of the construct seem well fitted to positively influence a negative relationship between social category-based faultlines and team performance, as well as savor or increase a positive relationship between informational category-based faultlines. I therefore propose a climate for inclusion as a moderator of the relationship between faultlines and team performance.
In a recent article, Nishii (2013) introduced the novel construct within the diversity literature by proposing that teams should aspire to this ‘climate for inclusion’. According to the author, an inclusive climate consists of three dimensions. The first dimension emphasizes the importance of eliminating bias from organizational practices and establishing a positive climate for diversity. This involves a foundation of fairly implemented employment practices and the implementation of diversity-specific practices that help to eliminate bias. The second dimension is the integration of differences. Nishii states that an inclusive climate moves beyond just increasing diverse representation in the organization and implementing human resource practices that promote equality. It involves the interpersonal integration of diverse employees and is described as the degree to which employees perceive that they can be themselves and enact their
21
identity without negative consequences. The third and final dimension according to Nishii is inclusion in decision making. It entails the extent to which the diverse perspectives of the employees are actually sought out and integrated, even if this may shake up business as usual.. This conceptualization of a climate of inclusion by Nishii will be used in the current study.
Nishii’s (2013) work on a climate for inclusion remains the only published paper to date that describes the construct and its outcomes. However, it is rooted in the concept of ‘inclusion’, Inclusion has spawned empirical research that showed evidence of its advantages for diverse teams (Shore, Randel, Chung, Dean, Ehrhart & Singh, 2011). Positive outcomes of inclusion in teams include higher levels of career optimism among minority members, higher ratings of feeling respected and valued among team members, an increased ability of the team members to effectively work in cross-organizational collaborations, and the establishment of a climate where the individual team members can improve their skills.
Where social category-based faultlines mainly cause negative effects due to intergroup bias, a climate for inclusion aims to reduce such intergroup bias through fairly implemented employment and diversity-specific practices as well as prevent it by creating a non-threatening climate where individuals can safely express their true identity. Furthermore, a climate of inclusion implements cross-categorization practices, which should reduce the negative effects of the aligning social category-based attributes. Ergo, Hypothesis 3 is formulated as seen below.
H3: The relationship between social category-based faultlines in terms of strength and distance and team performance is moderated by a climate for inclusion such that
this relationship is less negative when climate for inclusion is high, and more negative when it is low.
22
are appreciated rather than suppressed. Also, an inclusive climate promotes inclusion for decisions, which creates a social context in which the positive aspects of diversity can flourish. in informational category-based characteristics. It creates a climate where people feel safe and dare to give input, contributing to the cognitive capital of the team. Finally, its cross-categorization practices can results in team members from different sub-groups appreciating each other’s informational category-based differences and effectively using their collective cognitive capital through information elaboration. Therefore, Hypothesis 4 was proposed as formulated below.
H4: The relationship between informational category-based faultlines in terms of strength and distance and team performance is moderated by a climate for inclusion
such that this relationship is more positive when climate for inclusion is high, and less positive when it is low.
3. Methods
3.1. Research method and data collection
The research method used for this thesis is the quantitative research method. The data is primary data and was collected with the use of a survey, which can be found in Appendix A. This method of collecting data was chosen as some concepts can best be captured by using surveys (McNabb, 2004). Moreover, it is a relatively low-cost option and it consumes relatively little time (Saunders, Lewis, & Thornhill, 2009). The survey was compiled in a team of researchers.
23
Likewise, the process of data collection was done with in this research team. The survey is structured and uses standardized, closed questions, specifically list questions, rating questions, quantity questions and matrix questions (Saunders et al., 2009). The survey was constructed using existing scales, which will be discussed in the ‘Measures’ section. A page on confidentiality was also added. The items or survey questions were subsequently entered into the online survey tool ‘Qualtrics’. Using this online tool, the survey could be send to many people at once in a time- and cost efficient manner. The survey was also printed out and taken to the respondents to increase participation numbers. The survey items were originally in English, so to make sure that the respondents understood the survey, the questions were also translated into Dutch. To minimize potential problems related to common-method variance ("i.e., variance that is attributable to the measurement method rather than to the constructs the measures represent" (Podsakoff, MacKenzie, Lee & Podsakoff, 2003, p.879)), multiple sources were used in this research to measure team performance (Podsakoff, MacKenzie & Podsakoff, 2012). The independent variables were not measured through survey items but were calculated based on objective team information. To measure the moderating variable climate for inclusion, only one scale is yet available, so a single source was used.
Because the conceptual model of this thesis is at the team level, data had to be collected on the team level as well. In a team with other researchers, I thus sought out teams willing to participate in our research. In the process of selecting teams to approach, the convenience sample technique was used (Saunders et al., 2009). This technique was chosen because the study needed a high response rate at the team-level.
24
The sample that is used for this thesis consists of teams. Every group of people that worked in the same team was counted as one unit. The population of the sample is every person that works in a team. Initially, only highly professional teams were approached, but this resulted in a lack of sufficient responses. Therefore, it was decided to approach different types of teams in different branches. The survey was sent to approximately 60 companies. Within those 60 companies, around 160 team leaders or warm contacts were asked to help motivate the team members to fill in the survey. The data was collected in a time period of nearly two and a half months. The survey was completed by a total of 409 individuals. However, of these 409 individuals, a few could not be used due to too many missing values. Unfortunately, there were many respondents that filled in the survey who could not be used in our sample because they were the only one, or one of the few team members from their team that filled in the survey. Only the teams that had response rates of above 75% of the team members were included in the data set, with the exception of one very large team of 25 members which had a response rate of 48% but had an absolutely high number of participants. This resulted in the fact that of the 409 individuals that had filled in the survey, 200 respondents could be used for the data set, resulting in a data set of 32 teams. The construct of faultlines requires a team to at least consist of three team members, and one of the 32 teams in the data set was a team of 2, so this team could not be used for the current research. Ergo, this thesis uses a sample of 31 teams and includes a total of 197 individuals.
The team size varied from 3 to 25 team members, with a mean of 6.6 and a standard deviation of .9. The age of individuals ranged from 18 to 64 with a mean of 31.7 7 and a standard deviation of .7. The sample was 57.9% female and 42,1% male. For 62.9% of the individuals in
25
the sample the highest level of completed education was a Master's, for 14.7% this was a bachelor's, for 18.3% this was a vocational or technical school and for 4.1% this was high school or an equivalent. The range of tenure of the individuals in the sample was from 1 month to 29.2 years, with an average of 3.9 years and a standard deviation of approximately 6 months (.46 year).
3.3. Measures
3.3.1. Independent variable
The independent variable in this research is faultlines. As discussed in the literature review, this thesis calculates two different types of faultlines values, one based on social categories and one based on information categories. This division between types of faultlines is an idea borrowed from the paper by Bezrukova, Jehn, Zanutto and Thatcher (2009), and this thesis also uses their measures for faultlines. Their measure is “adapted from the multivariate statistical clustering analysis literature, which provides statistics for estimating how well the variability within the group can be explained by the presence of different clusters within the group” (Bezrukova et al., 2009, p.41). The social category-based faultlines were calculated using the social category characteristics age and gender. The information category-based faultlines were calculated using the information category characteristics level of education and tenure.
The faultlines measure used for this research looks at the strength and at the distance of the faultline. Bezrukova and collegues (2009) use a composite term in their paper, which multiplies a faultline strength measure by a faultline distance measure. Faultline strength is measured by Fau, an algorithm developed by Thatcher, Jehn and Zanutto (2003):
26
“Where xijk denotes the value of the j-th characteristic of the ith member of subgroup k, x-j.
denotes the overall group mean of characteristic j, x-.jk denotes the mean of characteristic j in
subgroup k, and ngk denotes the number of members of the k-th subgroup (k = 1, 2) under split g”
(Thatcher et al., 2003, p. 226). A limitation here could be that this algorithm only measures faultlines that split the group into two subgroups. However, as Zanutto et al. (2011) argue, this is probably the most common type of situation and possibly the most important due to the fact that the faultlines are stronger. The Fau values can take on values from 0 to 1, where higher values mean stronger faultlines. The values of faultline strength for the teams in the sample of this thesis ranged from 0.42 to 1 for social category-based faultlines, and from 0.53 to 1 for information-based faultlines. Following Bezrukova et al. (2009), this thesis uses the following algorithm to measure faultline distance D:
“where the centroid (vector of means of each variable) for subgroup 1 =
and the centroid for group 2 = ” (Bezrukova et al., 2009, p. 41). The D values can take on values from 0 to infinity, where higher values indicate a larger distance between sub-groups. The values of faultline distance for the teams in the sample of this thesis
27
ranged from 0.14 to 3.75 for social category-based faultlines, and from 0.04 to 3.89 for information-based faultlines.
The actual calculation of the faultlines values was done using the open source statistical environment R. This program uses the algorithms for Fau and D to calculate faultline strength and faultline distance, respectively. In the paper by Meyer and Glentz (2013), it is explained how to use the the open source statistical environment R in this respect. When all values had been calculated, the Fau and D from social category-based faultlines were multiplied to create a composite term that serves as the measure of social category-based faultlines, the Fau and D from informational category-based faultlines were multiplied to create a composite term that serves as the measure of informational category-based faultlines.
3.3.2. Dependent variable
The dependent variable in this model is team performance. This was done using multiple sources to minimize potential problems with common-method bias. Supervisor ratings for efficiency, quality and overall achievement of the team were collected using three single-item performance measures developed by Ancona and Caldwell (1992). These measures were chosen on the basis of Van der Vegt and Bunderson (2005), whose research covered similar topics and used these measures for team performance. Van der Vegt and Bunderson looked at team identification as a moderator for the relationship between team diversity and team performance. The response set of the scale by Ancona and Caldwell (1992) ranged from 1 (far below average) to 7 (far above average). The Cronbach’s alpha was the same for all three scales: 0.99.
In addition, a multi-item scale was used, as this allows for more fine-grained measures of team performance, and because measures with several items have higher reliabilities. Following
28
previous research (Kirkman and Rosen, 1999; Kirkman & Shapiro, 2001; Kirkman, Tesluk & Rosen, 2001) the 6-item scale for team productivity was adopted from Kirkman and Rosen (1999). The scale assesses team performance on how productive the team is. The 6-items (e.g. “This teams meets or exceeds its goals”) had a response set of 1 to 7, where 1 = Strongly disagree and 7 = Strongly agree. This scale had a Cronbach’s alpha of 0.92.
3.3.3. Moderating variable
The moderating variable in this research is climate for inclusion developed by Nishii (2013). It was measured using Nishii’s (2013) 15-item scale (e.g. “In this team, everyone’s ideas for how to do things better are given serious consideration”). The 15-item scale is a shortened version of the original 31-item scale used by Nishii to measure a climate for inclusion. As Nishii notes that “shortening the scale did not noticeably reduce the coefficient alphas for any of the sub-dimensions” (p 1762), the shortened version should suffice. The response set for this scale ranges from 1 to 5, where 1 = Strongly disagree and 5 = Strongly agree. The Cronbach’s alpha for this scale was 0.86.
3.3.4. Control variables
To control for factors that have been shown to influence team outcomes in prior research (Shalley, Gilson & Blum, 2009; Katz-Navon & Erez, 2005; Brewer & Kramer, 1986), three control variables are taken into account: job complexity, task interdependence and team size. The data set consists of a broad range of team types, and one team may be more professional or perform on a higher level than the other. Moreover, job complexity has been found to affect team performance (Shalley, Gilson & Blum, 2009). To account for this difference between teams in the data set, the variable job complexity is controlled for. To measure job
29
complexity, the scale developed by Morgeson and Humphrey (2006) is used. The scale consists of 4 items (e.g. “The tasks on the job are simple and uncomplicated”) and the response set ranges from 1 to 5, where 1 = Strongly disagree and 5 = Strongly agree. The Cronbach’s alpha for this variable was 0.90.
The second control variable is the task interdependence of the team. This research focuses on faultlines, a construct which mainly affects teams and team members that need to work together in order to complete tasks. So, the degree of task interdependence is a variable that needs to be considered when testing the Hypotheses. Task interdependence is measured with the scale that was used by Van der Vegt and Janssen (2003). The scale has 5 items (e.g. “I need information and advice from my colleagues to perform my job well”) and has a Likert scale ranging from 1 to 7, where 1 = Completely disagree and 7 = Completely agree. The Cronbach’s alpha for this variable was 0.77.
The third control variable in this research is team size. The size of the teams in the data set differs greatly, ranging from 3 to 25, and this could have consequences on the relationship between faultlines and team performance. Research has shown that team size can affect group dynamics (Brewer & Kramer, 1986). Therefore, the team leaders were asked to fill in a matrix that was attached to the survey, so that individual demographic attributes and team level information such as team size could be collected. Hence, team size can be controlled for.
3.4. Data aggregation
The level of analysis of this thesis is at the team level. However, the moderating variable a climate for inclusion and the control variables job complexity and task interdependence were surveyed on an individual level. The individual team members that participated in the survey
30
were asked to rate their team on certain attributes, after which those individual scores were averaged to create a mean score for each team. To check whether this aggregation to the team level was justified, the degree of agreement between the team member within the team was assessed by calculating mean rwg[J] values (James, Demaree, & Wolf, 1984). Furthermore, an
intraclass correlation coefficient (ICC[1]) analysis was done to check whether there was sufficient variance within and between units of analysis (Bliese, 2000; Biemann, Cole, & Voelpel, 2012) and values for the reliability of team member’s average ratings (ICC[2]) were calculated (Biemann, Cole, & Voelpel, 2012). These values were .98 (rwg[J]), .23 (ICC[1]), p =
.001, and .55 (ICC[2]) for a climate for inclusion, .81 (rwg[J]), .39 (ICC[1]), p < .001, and .73
(ICC[2]) for job complexity, and .90 (rwg[J]), .25 (ICC[1]), p < .001, and .58 (ICC[2]) for task
interdependence. Therefore, agreement was acceptable and aggregation to the team level was justified (James et al., 1984; Biemann, Cole, & Voelpel, 2012).
4. Results
4.1. Descriptive analysis
The means, standard deviations and correlations of the variables are given in Table 1. A larger version of Table 1 is also attached in Appendix B. As can be seen from the Table, there are some significant correlations. Efficiency positively correlates with quality (r = .512, p < 0.01), overall achievement (r = .617, p < 0.01) and team productivity (r = .697, p < 0.01). Next to its aforementioned correlation with efficiency, quality positively correlates to overall achievement
31
(r = .689, p < 0.01). Overall achievement only corresponds with efficiency and quality (mentioned above). Team productivity in turn only correlates to efficiency (mentioned above). The two independent variables social- and informational category-based faultlines also significantly correlate with each other (r = .533, p < 0.01).
Table 1. Means, Standard Deviations, Correlations
4.2. Hypothesis testing
The hypothesized relationships between the independent variables social- and informational category-based faultlines, the moderating variable climate for inclusion and the dependent variables efficiency, quality, overall achievement and team productivity are tested through hierarchical regression analyses. The independent variables, the moderating variable and the control variables were all standardized before computing the product terms and performing the regression analyses. The control variables job complexity, task interdependence and team size were included in Step 1 of the hierarchical regression. In Step 2 the main effect variables of
32
social category-based faultlines (H1) and informational category-based faultlines (H2) were added. Subsequently, Step 3 also included the two-way interactions between social category-based faultlines and climate for inclusion (H3) as well as informational category-category-based faultlines and climate for inclusion (H4). Results are presented in Table 2. A larger version of Table 2 is also attached in Appendix C.
Table 2. Hypothesis testing using hierarchical regression analysis
Hypothesis 1 posited that social category-based faultlines in terms of strength and
distance are negatively associated with team performance. The results of the analysis show that social category-based faultlines do not significantly affect efficiency (β = .509, s.e. = .250, t = 2.038, p = .053), quality (β = .354, s.e. =.211, t = 1.678, p = .106), overall achievement (β = .267, s.e. = .183, t = 1.459, p = .158) and team productivity (β = .134, s.e. = .223, t = .602, p = .553). Thus, Hypothesis 1 was rejected.
Hypothesis 2 predicted that informational category-based faultlines in terms of strength
and distance positively affect team performance. The analysis revealed that informational category-based faultlines did not significantly influence efficiency (β = -.082, s.e. = .251, t =
33
-.325, p = .748), quality (β = -.122, s.e. = 2.12, t = -.573, p = .572), overall achievement (β = -.278, s.e. = .184, t = -1.509, p = .144) and team productivity (β = -.253, s.e. = .225, t = -1.127,
p = .271). Hence, Hypothesis 2 was also rejected.
Hypothesis 3 stated that the relationship between social category-based faultlines in terms of strength and distance and team performance is moderated by a climate for inclusion such that this relationship is less negative when climate for inclusion is high, and more negative when it is low. The results of the analysis revealed that the moderating effect of climate for inclusion was in the expected direction, but not significant. Figure 2 to 5 plot the moderating effect of climate for inclusion on the relationship between social category-based faultlines and the various performance outcomes. Climate for inclusion affected the relationship between social category-based faultlines and efficiency (β = .671, s.e. = .471, t = 1.424, p = .169), quality (β = .632, s.e. = .372, t = 1.699, p = .103), overall achievement (β = .497, s.e. = .347, t = 1.429, p = .167) and team productivity (β = .488, s.e. = .430, t = 1.133, p = .269), such that this relationship is less negative when climate for inclusion is high, and more negative when it is low. However, none of these effects were significant.
To further analyze the interaction between social category-based faultlines and climate for inclusion in predicting team performance, following Aiken and West (1991), a simple slope analysis was performed. The simple slopes were tested for low (-1 S.D.) and high (+1 S.D.) levels of climate for inclusion. At low levels of climate for inclusion, social category-based faultlines negatively affected performance outcomes. Insignificant negative effects were observed on efficiency (β = -0.304, s.e. = .587, t = -0.518, p = .608), quality (β = -.468, s.e. = .463, t = -1.012, p = .321), overall achievement (β = -.322, s.e. = .432, t = -.745, p = .463) and team productivity (β = -.454, s.e. = .537, t = -.846, p = .405). At high levels of climate for
34
inclusion, social category-based faultlines positively affected the performance outcomes. Significant positive effects were found on efficiency (β = 1.038, s.e. = .490, t = 2.119, p = .043) and quality (β = .796, s.e. = .387, t = 2.055, p = .049). Additionally, the positive effects of social category-based faultlines on overall achievement (β = .672 s.e. = .362, t = 1.857, p = .074) and team productivity (β = .522, s.e. = .447, t = 1.167, p = .253) were non-significant Based on the results of the hierarchical regression analysis, Hypothesis 3 was rejected.
Figure 2 to 5: Interaction of climate for inclusion and social category-based faultlines in predicting the various performance outcomes.
36
Hypothesis 4 predicted that the relationship between informational category-based
faultlines in terms of strength and distance and team performance is moderated by a climate for inclusion such that this relationship is more positive when climate for inclusion is high, and less positive when it is low. The hierarchical regression analysis shows that the moderating effect of climate for inclusion is not in the expected direction, but also not significant. Figure 6 to 9 graphically depict the interaction of climate for inclusion and informational category-based faultlines in predicting the various performance outcomes. Climate for inclusion does not affect the relationship between social category-based faultlines and efficiency (β = -.291, s.e. = .267, t = -1.090, p = .288), overall achievement (β = -.172, s.e. = .197, t = -.873, p = .392) and team productivity (β = -.200, s.e. = .244, t = -.820, p = .421). There is a significant moderating effect of climate for inclusion on the relationship between informational category-based faultlines and
37
quality (β = -.455, s.e. = .211, t = -2.160, p = .042) in the opposite direction from what was hypothesized.
Again, simple slope tests were performed to further investigate the moderating effect (Aiken & West, 1991), where simple slopes were tested for low (-1 S.D.) and high (+1 S.D.) levels of climate for inclusion. Low levels of climate for inclusion resulted in insignificant positive relationships between informational category-based faultlines and efficiency (β = .302,
s.e. = .492, t = .614, p = .544), quality (β = .536, s.e. = .390, t = 1.375, p = .180) and team
productivity (β = .007, s.e. = .449, t = .016, p = .988). Furthermore, an insignificant negative effect was observed on overall achievement (β = -.065, s.e. = .363, t = -.179, p = .859). High levels of climate for inclusion created insignificant negative relationships between informational category-based faultlines and efficiency (β = -.280, s.e. = .279, t = -1.003, p = .325), quality (β = -.374, s.e. = .219, t = -1.707, p = .100), overall achievement (β = -.409, s.e. = .210, t = -1.950, p = .062) and team productivity (β = -.393, s.e. = .257, t = 1.530, p = .138). Based on the results of the hierarchical regression analysis, Hypothesis 4 was also rejected.
38
Figure 6 to 9: Interaction of climate for inclusion and information category-based faultlines in predicting the various performance outcomes.
40
5. Discussion
5.1. Discussion
Faultlines have been found to both negatively and positively affect team performance. This paradox lead the current author to try to find an optimal way of living with faultlines in the team. In an attempt to attribute the positive and negative effects of faultlines to specific aspects of the construct, following Bezrukova et al. (2009), a distinction was made between social category-based faultlines and informational category-category-based faultlines.
Hypothesis 1 predicted that social category-based faultlines in terms of strength and distance would be negatively related to team performance. This Hypothesis was rejected however, on the basis that none of the four performance outcomes were significantly affected by social category-based faultlines. Thus, it is not possible to predict any effect based on these analyses.
Similarly, Hypothesis 2 was rejected. It proposed that information category-based faultlines in terms of strength and distance would be positively related to team performance. It was rejected on the basis that none of the analyzed effects of informational category-based faultlines with team performance were significant.
A possible explanation for the lack of a significant effect of faultlines on team performance may lie in the difference between aligning diversity attributes and cross-cutting diversity attributes. Faultlines have their effects on team performance when diversity attributes of individuals align in the team. However, diversity attributes could also cross-cut, for example in teams where people are assigned to work in specific sub-groups. This cross-categorization
41
would decrease the effects of faultlines (Sawyer, Houlette & Yeagley, 2006). The calculated faultlines values would indicate strong faultlines however, because they calculate the chance that the existing diversity attributes in a team align and cause faultlines. In the future, this problem could be controlled for by including a survey scale measuring the degree of perceived cross-categorization in the team.
Another explanation for the insignificant results of Hypothesis 1 and 2 can be found in the difference between dormant and active faultlines. The social category-based fautlines that were calculated for this research were based on the objective social category-based characteristics (age and gender) of the individuals in the teams. However, the consequent faultline values were merely probabilities, and remained hypothetical. This means that nothing could be said of whether or not these faultlines were actually perceived by the team members. According to Rink and Jehn (2010), dormant faultlines can become active when a 'faultline trigger' activates the faultlines. It is possible that there were teams in the sample that had high levels of dormant faultlines, but that their faultlines were not activated and not actually perceived by the teams. A possible remedy for this problem could be the use of a 'perceived faultlines' scale that measures the extent to which faultlines are actually perceived by team members.
The results of this thesis can further be discussed along the propositions of the Categorization-Elaboration model (Van Knippenberg et al., 2004). Among other things, this model argues that every diversity attribute could elicit social categorization processes as well as information elaboration processes, and that all forms of diversity can have positive and negative effects. So, these authors would argue that dividing diversity attributes into either social category-based or informational category based characteristics may not be very effective. The results of the current study failed to assign specific effects to these two types of diversity, so the
42
propositions of the CEM may explain some of this.
Furthermore, the CEM also states that the social categorization process of individuals is based on the salience of the categories. In this study, the diversity attributes age, gender, level of education and tenure were researched. Some of these attributes might not have resulted in very salient categories for the team members in the sample, in which case the alignment of these attributes would be less detrimental to team performance (Van Knippenberg et al., 2004). The effects of variables outside this research should not be discounted. The variable 'time' may be a relevant influence on the relationships in this study. As time passes and the length of the collaboration of the team members becomes longer, group dynamics might change. Harrison, Price and Bell (1998) found that the passing of time helps reduce the negative effects of the surface level (salient) attributes that social category-based faultlines are based on. Conversely, the deep level attributes (not easy to detect, less salient) that informational category-based faultlines are based on would start affecting teams more strongly and diminishing group cohesion with the passing of time. Thus, it is possible that the passing of time moderated the effects of faultlines on team performance.
The main goal of this thesis was to test the moderating effect of a climate for inclusion on the relationship between faultlines and team performance, in an attempt to reduce the negative effects of faultlines while unlocking its potential positive effects. Specifically, Hypothesis 3 predicted that the relationship between social category-based faultlines in terms of strength and distance and team performance would be moderated by a climate for inclusion such that this relationship is less negative when climate for inclusion is high, and more negative when it is low. This Hypothesis was rejected as none of the p-values were significant. Plotting the relationships revealed that albeit non-significant, the interaction effect was disordinal, low levels of climate
