Inclusiveness Through Social Categorization

(1)

Inclusiveness Through Social Categorization

The role of Global Identity in Reducing Relational Conflicts and Enhancing Feedback Inquiry in

Multicultural Virtual Teams

Student: Joanne Chukundah Student number: 12097322

Submission Date: 30 June 2021

Program: Executive Programme in Management Studies – Leadership & Management track

Institution: Amsterdam Business School Thesis Supervisor: Maarten de Haas

(2)

2

Practical implications - Practitioners can use the insights derived from this study to effectively manage multicultural virtual teams in online global tech organizations by focusing on a global cultural approach and stimulating Global Identity behaviors in virtual teams. This can be done through facilitating social events, community building, development and the hiring selection process.

Originality/value – Uses the Social Categorization Theory and builds upon the Shared Global Identity model to demonstrate how disadvantages of cultural diversity in Virtual teams can be used as advantages through Global Identity. And how Perceived Proximity functions as a mediation mechanism in the relationship between Global Identity as predictor of Multicultural Virtual Team effectiveness.

(4)

4 1. Introduction

Over the last decades the development of Information and Communication Technologies (ICT) greatly impacts today’s way of working within organizations (Mysirlaki and Paraskeva, 2020). Teams no longer require face to face interaction when relying on ICT to communicate and coordinate work (Dulebohn and Hoch, 2017). As a result teams started to work in virtual settings benefitting from new opportunities for both employees and organization’s while no longer being limited by geographical dispersion and boundaries (Dulebohn and Hoch, 2017). Virtual Teams are defined as “groups of geographically and/or organizationally dispersed coworkers that are assembled using a combination of telecommunications and information technologies to accomplish a variety of critical tasks”

(Townsend, DeMarie and Hendrickson, 1998: 17). The coronavirus crisis (COVID-19) made working through Virtual Teams (VT’s) even more of a common phenomenon with global impact (Mysirlaki and Paraskeva, 2020). It forced companies to transform from traditional structures to networked organization’s working with VT’s in order to secure business continuity during the pandemic (Mysirlaki and Paraskeva, 2020). VT’s are beneficial for both organizations and employees providing them flexibility in structures, continuous productivity, knowledge sharing across boundaries, reduced workspace and commuting costs, a better work- life balance and increased levels of work satisfaction and commitment (Dulebohn and Hoch, 2017; Mysirlaki and Paraskeva, 2020). However research on VT’s also identified adverse effects of the lack of face-to face communication in the virtual relationship of team members and leaders such as hindered information flows (Mesmer-Magnus, DeChurch, Jimenez- Rodriguez, Wildman and Shuffler, 2011; Dulebohn and Hoch, 2017), psychological distance and isolation (Gajendran and Harrison, 2007; Dulebohn and Hoch, 2017), reduced team coordination (Penarroja, Orengo, Zornoza, 2013) and increased team conflict (Furumo, 2009;

Harush, Lisak and Glikson, 2018). Because team members in a virtual environment rely on computer-mediated cues instead of face-to-face communication, it seems that the virtual environment presents a substantially different context for interpreting and perceiving team members’ behavior (Charlier, Stewart, Greco, and Reeves, 2016). Additionally since VT’s are not limited geographical location, a typical characteristic of VT’s is the high level of multicultural diversity (Dulebohn and Hoch, 2017). Cultural diversity in teams has its benefits but also its downsides (Hogg, 2017). Interpreting and perceiving team members’ behavior virtually from geographically dispersed work locations in combination with high cultural diversity in VT’s could become a risk for the virtual team effectiveness (Thompson, 1967).

The combination of these two characteristics of VT’s and how to use these characteristics to

(5)

5 attenuate risks and enhance team effectiveness has only been researched recently by Harush, Lisak and Glikson (2018). In their study Global Identity was identified as a possible predictor of team effectiveness (Harush, Lisak and Glikson, 2018). There still is lack of understanding on the impact of Global Identity on emotional, cognitive and behavioral aspects of work such as psychological proximity, relational conflict and feedback seeking in organizational settings (Harush, Lisak and Glikson, 2018). This study scientifically researches the impact of Global Identity in the context of multicultural VT’s, which will help organizations and managers to learn how to use it as a strategic advantage (Mysirlaki and Paraskeva, 2020). It will especially help organizations who are considering to apply VT’s as the new post-pandemic norm. One of these organizations is organization X, which is used as case study. Organization X is an online auction platform in Europe and Asia for special objects, offering over 65.000 objects in more than 600 auctions each week across multiple categories such as collectables, art, design, jewelry, watches, classic cars and more. It was founded in 2008 and grew rapidly over the years with strong focus on international expansion and scaling up from a startup organization. The startup mentality is still very present in the organizational culture. Organization X currently has 707 employees working in multicultural virtual teams from disperse geographical locations in Western Europe and Asia. This makes organization X a good fit as a case study for this research on multicultural virtual teams.

2. Literature review

A recent study on the effects of interpreting and perceiving team members’ behavior within VT’s, showed that social categorization as a process frequently leads to disruptive dynamics in the multicultural virtual team context (Harush, Lisak & Glikson, 2018). The study used the Social Identity Theory of Tajfel and Turner (1986) and the Self Categorization Theory of Turner, Hogg, Oakes, Reicher and Wetherell (1987) to research intergroup relations. Based on these ground theories, they identified Shared Global Identity as a possible predictor of Virtual Team Effectiveness. In their research on multicultural distributed teams they found evidence that highlighted the role of global identity in reducing relational conflict in virtual teams. In this effect, Perceived Proximity mediated the relation between Global Identity and Relational Conflict. Additionally they identified team interdependence as a moderator in the indirect relation between Global Identity and Relational Conflict through Perceived Proximity (Harush, Lisak and Glikson, 2018). However the research was conducted among multicultural distributed MBA student project teams (Harush, Lisak and Glikson, 2018). This particular study context may have limited the extent of the generalization of the findings for practitioners

(6)

6 in global organizations (Harush, Lisak and Glikson, 2018; Saunders and Lewis, 2018).

Therefore in order to validate the findings, this study will test this the Shared Global Identity model in organizational settings (Saunders and Lewis, 2018). In addition this study uses the Shared Global Identity model to explore the indirect effect of a shared global identity on another behavioral aspect of work in multicultural teams. One of the negative consequences of virtual communication is the lack of face-to-face interaction as this makes immediate feedback, as well as affective signals, more difficult to send, sustain, and receive (Hallowell, 1999). Face to face communication is considered the medium with the highest social presence and media richness (Gajendran and Harrison, 2007). A lack of it leads to negative effects on the interpersonal relationships between virtual team members (Gajendran and Harrison, 2007;

Dulebohn and Hoch, 2017) decreasing feedback seeking behavior from others and the utilization of feedback (Ashford and Cummings, 1983; Whitaker and Levy, 2012; Wu, Parker and Jong, de, 2014), in this study referred to as feedback inquiry and feedback seeking frequency (Linderbaum and Levy, 2010). This study tests the existing model and additionally uses the model to explore the indirect effect of a shared global identity on feedback inquiry and feedback frequency within VT’s, giving an answer on the research question: What is the effect of Global Identity on Relational Conflicts and Feedback Inquiry and Feedback Seeking Frequency in multicultural virtual teams?

2.1 The Self Categorization Theory and Social Identity Theory

Within the social psychology literature the Social Identity Theory of Tajfel and Turner (1986) and the Self Categorization Theory of Turner, Hogg, Oakes, Reicher and Wetherell (1987) have been highly influential on research of group processes and intergroup relations within a diverse group dynamics context (Hornsey, 2008). These two theories form the ground theories of the shared global identity model (Harush, Lisak and Glikson, 2018; Saunders and Lewis, 2018) and are interlinked in showing how relational conflicts are the result of the social categorization of one’s social environment into ingroups and outgroups (Tajfel and Turner, 1986; Turner, Hogg, Oakes, Reicher and Wetherell, 1987).

In the context of multicultural virtual teams, social categorization is the tendency of individuals to categorize team members in ingroups and outgroups (Turner, Hogg, Oakes, Reicher and Wetherell, 1987; Hogg, 2001). Ingroups consist of team members that the individual identifies him/herself with based on a perception of similarities with that particular team member (Guillaume Dawson, Otaye-Ebede, Woods and West, 2017; Harush, Lisak and

(7)

7 Glikson, 2018). Outgroups consist of team members that the individual finds it more difficult to identify with based on the lack of similarities with that particular team member (Guillaume Dawson, Otaye-Ebede, Woods and West, 2017; Harush, Lisak and Glikson, 2018). This social categorization behavior tends to be based on stereotyping that identifies sets of attributes that define and prescribe attitudes, feelings and behaviors to groups (Hogg, 2001). In the social context of an individual these characteristics are used to distinguish between different groups.

This is referred to as depersonalization (Hogg, 2001). It is called depersonalization due to the fact that it perceives people as a match to the relevant ingroup or outgroup, rather than perceiving people as unique individuals (Turner, Hogg, Oakes, Reicher and Wetherell, 1987;

Hogg, 2001). This depersonalization also leads to self-categorization (by the individual themselves) based on the group normative of the ingroup of that particular individual (Hogg, 2001). In other words it changes self-perception and transforms this to the perception of matching attitudes, feelings and behaviors with the groups normative (Hogg, 2001). It changes the way someone thinks, feels and acts in order to create a social identity that better matches with the group normative (Hogg, 2001). Tajfel (1972, p.292) described social identity as “the individual's knowledge that he belongs to certain social groups together with some emotional and value significance to him of this group membership”. The social identity is self-evaluative and derives its value from the evaluative properties of the ingroup (Hogg, 2001). In turn these evaluative properties of the ingroup are used as social comparison with other groups (Hogg, 2001). It affects how the individual feels towards others and creates a tendency to behave more favorably toward others who belong to one’s ingroup and less favorably toward others that belong to outgroups (Tajfel and Turner, 1986; Harush, et al., 2018). This is called intergroup bias (Tajfel and Turner, 1986; Harush, et al. 2018) and involves a process of competition for positive identity and distinctiveness (Tajfel and Turner, 1986). Even within ingroups there’s a stronger liking to members that are perceived as more similar (Hogg, 2001). The reason why individuals tend to search for positive social identity through positive distinctive group behavior is triggered by a basic human need for positive self-esteem and reducing subjective uncertainty through group validation that can only be satisfied through social identification (Abrams and Hogg, 1988; Grieve and Hogg, 1999;).

2.2 Relational Conflict Literature

Another ground theory that the Shared Global Identity Model is based on, can be found in the literature of conflict. In review of early conflict literature a conflict is usually referred to

(8)

8 as a process in which one party perceives that its interests are being opposed or negatively affected by another party (Wall & Callister, 1995). In this definition conflict was considered a process that distracted a team from a task which had a negative effect on the team performance (Hackman & Morris 1975). Conflicts between team members can occur when they have to engage in an activity while they have diverging interests or when they differ in incompatible behavioral preferences (Rahim, 2002). Another cause for conflicts between team members is when they differ in attitudes, values, skills and goals that are experienced as exclusive by others or when team members are interdependent in the performance of functions or activities (Rahim, 2002). Throughout the literature different types of conflicts were identified such as task, relational, status and process conflict (Jehn 1995, 1997; Bendersky & Hays, 2012). By analyzing the effects of these types of conflicts it was concluded that each type has its own effects on group outcomes such as member satisfaction and team performance (Jehn 1995, 1997; Bendersky & Hays, 2012). Where all types have a negative effect on member satisfaction, task conflict on moderate levels can improve team performance, enhance decision quality and stimulate creativity (Jehn 1997; Hollenbeck, Ilgen, Sego, Hedlund, Major &

Phillips 1995). Contrary to task conflict, relational conflict has not any positive effects on performance outcomes (Jehn 1995, 1997). This type of conflict is motivated by a negative interpersonal relationship (Jehn 1995, 1997). The negativity can be caused by differences in values, preferences and/or priorities and results into incompatible disagreements along with the feeling of tension between team members (Jehn 1995, 1997).

2.3 Feedback Literature

The behavioral aspect that has been added in this study to the conceptual model of shared Global Identity, can be found in the literature of Feedback. The performance and motivation enhancing effect of feedback has been an accepted principle in the psychological research since the 1950’s and remains a topic until this day (Ashford & Cummings, 1983;

Chapanis, 1964; Vancouver & Morrison, 1995; Wu, Parker & de Jong, 2014). Feedback is considered to be a personal resource of an individual in fulfilling both performance and nonperformance goals (Ashford & Cummings, 1983). In modern literature it is considered as form of self-management in which employees take an active role regulating their personal performance, setting their own performance goals and monitoring their behaviors towards achievement of these goals (Vancouver & Morrison, 1995). It also includes rewarding themselves in case goals are achieved (Vancouver & Morrison, 1995). This study focuses on

(9)

9 feedback seeking behavior of individuals including feedback seeking inquiry and frequency (Vancouver & Morrison, 1995; Ashford & Tsui, 1991; Wu, Parker & de Jong, 2014). Feedback Inquiry entails directly asking someone for feedback to gather information to adjust one’s own goal-directed behavior, to assess their capabilities, to get a better understanding of their environment and to enhance their future effectiveness (Vancouver & Morrison, 1995).

Therefore feedback inquiry can be considered a behavior that is highly influential in one’s ability to increase performance and therefore a behavior that organizations’ would like to stimulate (Vancouver & Morrison, 1995). An important characteristic of this behavior is the social component (Vancouver & Morrison, 1995; Wu, Parker & de Jong, 2014). Feedback Inquiry requires direct communication between two parties and is therefore affected by interpersonal relationships (Vancouver & Morrison, 1995). Feedback inquiry behavior is influenced by the potential costs of finding and receiving feedback compared to the potential benefit (Morrison & Cummings, 1992). One of the identified factors that influences feedback inquiry is the quality of the relationship between the seeker and the source (Vancouver &

Morrison, 1995). The better the relationship is, the more likely a feedback source will treat a feedback seeker in a sensitive manner and decrease potential cost or damage (Vancouver &

Morrison, 1995). Feedback seeking frequency refers to frequency that an individual engages in feedback seeking behaviors either through feedback inquiry or through monitoring (Ashford

& Tsui, 1991). In case of feedback inquiry one directly asks managers, colleagues or others for information on their performance (Ashford & Tsui, 1991). Feedback through monitoring happens through personal observations and comparing their situations and others’ behaviors performance (Ashford & Tsui, 1991; Wu, Parker & de Jong, 2014).

2.4 Hypothesis Development

Since VT’s are not limited in geographical location and often work beyond geographical borders, a typical characteristic of VT’s is the high level of cultural diversity (Hinds, Neeley & Cramton, 2014; Dulebohn and Hoch, 2017; Harush, Lisak and Glikson, 2018). The social categorization theory identified disruptive effects due to diversity within teams (Turner, Hogg, Oakes, Reicher and Wetherell, 1987). In case of multicultural virtual teams these disruptive effects tend to occur due to differences in values, behaviors and preferences based on local culture and nationalities (Hogg, 2001; Harush, Lisak and Glikson, 2018). Therefore it is more likely that team members from the same national culture will feel connected through the cultural similarities and create an ingroup based on nationality (Turner,

(10)

10 Hogg, Oakes, Reicher and Wetherell, 1987). As team members tend to act more favorable towards others in their ingroup, high quality relationships start to exist between those team members (Kamdar & Van Dyne, 2007). These high quality relationships are known for high levels of trust and reciprocity (Kamdar & Van Dyne, 2007). Based on this theory it is also more likely that team members with very different national cultures, will become part of the outgroup (Turner, et al., 1987). These intercultural differences create a potential risk on relational conflicts (Nurmi and Hinds, 2016; Jehn 1995, 1997). However a different approach using the social categorization theory is focusing on the binding factor within multicultural virtual teams which is their global character (Harush, et al. 2018; Turner, et al., 1987). In this approach one shifts from the principles inhibiting disruptive team dynamics to facilitating inclusive dynamics within cultural diverse teams (Harush, et al., 2018; Turner, et al., 1987). Focusing on the global component of cultural diverse teams shifts the focus from-cross cultural approach to a global cultural approach (Harush, et al., 2018; Turner, et al., 1987). If the global cultural approach in organizations is facilitated and encouraged, employees will more likely depersonalize themselves to a Global Identity and self-categorize themselves to the norm of this ingroup based on global similarities (Harush, et al., 2018; Turner, et al., 1987). As employees act more favorable to team members in the ingroup, it will lead to higher quality relationships with team members regardless their nationality (Harush, et al., 2018; Turner, et al., 1987 (Kamdar & Van Dyne, 2007). These relationships are known for high levels of trust based on the high quality exchanges, it will increase the reciprocity between members which could be referred to as Perceived Proximity (Wilson, O’Leary, Metiu, Jett, 2008; (Kamdar & Van Dyne, 2007).

Perceived Proximity is the psychological closeness someone feels towards other team members (Wilson, O’Leary, Metiu, Jett, 2008). The stronger the shared global identity, the higher the perceived proximity (Harush, Lisak and Glikson, 2018). Based on the in-group identity model (Tajfel and Turner, 1986) it can be claimed that the higher levels of identification with a global identity among multicultural virtual team members, the higher the levels of perceived proximity (Harush, et al., 2018). Therefore it can be hypothesized that:

H1: Global Identity of an employee in multicultural virtual teams will be positively related to perceived proximity.

As mentioned intercultural teams have a higher risk on team conflicts, mistrust, fault lines, communication barriers, and disagreements on regulations, norms, expectations, and decision- making processes (Gibson & Gibbs, 2006; Jehn 1995, 1997; Harush, et al., 2018). This is due

(11)

11 to the fact that members differ more in perspectives regarding work expectations, procedures and decision making based on their cultural heritage (Harush, et al., 2018). What adds up to this effect is the reliance on virtual communication which creates greater potential for miscommunication (Gibson and Gibbs, 2006). But when team members experience a high level of proximity towards other team members in their multicultural virtual team, it is more likely that they will build high quality exchange relationships (Kamdar & Van Dyne, 2007; Wilson, O’Leary, Metiu, Jett, 2008). These relationships are known for higher levels of trust between members and in turn the high level of perceived proximity decreases the intercultural relational conflicts among members (Harush, et al., 2018; Kamdar & Van Dyne, 2007). Therefore it can be hypothesized that:

H2: Perceived Proximity of an employee in multicultural virtual teams will be negatively related to relational conflict.

Taking a global cultural approach instead of the cross-cultural approach shifts the focus to the binding factor within multicultural virtual teams (Harush, et al., 2018). This will lead to members re-categorizing themselves to the inclusive dynamics of the normative global ingroup (Tajfel and Turner, 1986). It changes the way someone thinks, feels and acts in order to create a social identity that better matches with the normative global ingroup (Hogg, 2001; Tajfel and Turner, 1986). This behavior will enhance the perceived proximity towards team members despite intercultural differences (Wilson, O’Leary, Metiu, Jett, 2008). In turn the higher levels of perceived proximity will create high quality exchange relationships between team members and will attenuate the risk on relational conflicts (Harush, et al., 2018; Jehn 1995, 1997).

Perceived Proximity in this context performs as a mediator between the relation of global identity and relational conflict (Harush, et al., 2018). Therefore it can be hypothesized that:

H3: Global identity of an employee in multicultural virtual teams will be negatively related to relational conflict through mediation of perceived proximity.

Researchers found evidence in a non-field study setting that the indirect relationship between Global Identity and Relational conflict is moderated by team interdependence as the attenuating effect on relational conflict appear under low rather than high team interdependence levels (Harush, Lisak and Glikson, 2018; Nurmi and Hinds, 2016).). Under high levels of team interdependence, members have more opportunities to interact (Guillaume, Brodbeck &

(12)

12 Riketta, 2012). Whenever team members need to work closely together they can identify bigger deep-level dissimilarities in values that are difficult to overcome (Harush, Lisak and Glikson, 2018). In these cases there’s more existence of dissimilarity among team members related to psychological characteristics, including personalities, cultural values and attitudes (Harush, Lisak and Glikson, 2018; Stahl, Mäkelä, Zander and Maznevski, 2010). This will weaken the attenuating effect of perceived proximity on relational conflict (Harush, Lisak and Glikson, 2018; Nurmi and Hinds, 2016). Therefore it can be hypothesized that:

H4: Global identity of an employee in multicultural virtual teams will be negatively related to relational conflict through mediation of perceived proximity. This effect will prevail under low, rather than high levels of team interdependence.

Researchers have found evidence that higher levels of virtual working (fulltime or structurally) exacerbates the negative impact on the quality of interpersonal interactions with others team members (Gajendran & Harrison, 2007; Lombard & Ditton, 1997). This is due to the fact that spending the majority of a work week away from a joined location implies fewer opportunities for rich face-to-face interactions with team members (Gajendran & Harrison, 2007). This will create fewer opportunities for spontaneous informal interactions (Mackie-Lewis, 1998). The interaction that takes place will mostly happen through digital channels, which is often experienced as lower social presence and increases the chances on miscommunication (Gibson and Gibbs, 2006; Short, Williams & Christie, 1976). This lower social presence leads to diminished perceptions of intimacy and immediacy, which are crucial to effective interpersonal communication (Lombard & Ditton, 1997). This ultimately can lead to psychological isolation (Dulebohn and Hoch, 2017) and decreased quality of relations with team members (Gejandran

& Harrison, 2007). Therefore there is a potential risk of lower quality relationships between team members in virtual teams (Gejandran & Harrison, 2007). Feedback Inquiry requires direct communication between two parties and is therefore affected by interpersonal relationships (Vancouver & Morrison, 1995). The better the quality of the interpersonal relationship, the more likely a feedback source will treat a feedback seeker in a sensitive manner and decrease potential cost or damage (Vancouver & Morrison, 1995). As previously mentioned, higher levels of perceived proximity towards other team members in virtual teams, makes it more likely that a high quality exchange relationships will be build, this creates higher levels of trust and reciprocity (Kamdar & Van Dyne, 2007; Wilson, O’Leary, Metiu, Jett, 2008). Therefore it can be argued that high levels of perceived proximity towards team members in virtual teams

(13)

13 will despite the geographical distance, lead to higher levels of Feedback Inquiry (Linderbaum and Levy, 2010). Additionally the reduced experienced risk on potential cost or damage of seeking feedback due to the perceived proximity will increase the frequency that someone seeks for feedback (Wilson, O’Leary, Metiu, Jett, 2008). Therefore it can be hypothesized that:

H5a: Perceived Proximity of an employee in a multicultural virtual team will be positively related to feedback inquiry.

H5b: Perceived Proximity of an employee in multicultural virtual teams will be positively related to feedback seeking frequency.

The context of the Self Categorization Theory and Social Identity theory is also applicable to feedback inquiry and feedback seeking frequency (Tajfel and Turner, 1986;

Linderbaum and Levy, 2010). As previously mentioned, disruptive effects tend to occur in cultural diverse teams due to differences in values, behaviors and preferences based on local culture and nationalities (Tajfel and Turner, 1986). Therefore it is more likely that team members with very different national cultures, will become part of the outgroup which will lead to lower quality of relationships (Tajfel and Turner, 1986). The lower the quality of the interpersonal relationship, the higher the risk on potential cost or damage for the feedback seeker (Vancouver & Morrison, 1995). This will impact the feedback inquiry and feedback frequency (Ashford & Tsui, 1991; Wu, Parker & de Jong, 2014). The opposite is also the case, with a global culture approach, team members in cultural diverse teams will experience a higher level of perceived proximity which will lead to higher level of feedback inquiry and frequency (Wilson, O’Leary, Metiu, Jett, 2008; Ashford & Tsui, 1991; Wu, Parker & de Jong, 2014). We argue that in this relationship Perceived Proximity operates as a mediator between Global Identity and Feedback Inquiry and Feedback Frequency in cultural diverse virtual teams (Wilson, O’Leary, Metiu, Jett, 2008; Ashford & Tsui, 1991; Wu, Parker & de Jong, 2014).

Therefore it can be hypothesized that:

H6a: Global Identity of an employee in a multicultural virtual team will be positively related to Feedback Inquiry through mediation of Perceived Proximity.

H6b: Global Identity of an employee in a multicultural virtual team will be positively related to Feedback Seeking Frequency through mediation of Perceived Proximity.

(14)

14 Under high levels of team interdependence, members have more opportunities to interact (Guillaume et al., 2012). Whenever team members need to work closely together they can identify bigger deep-level dissimilarities in values that are difficult to overcome (Harush, Lisak and Glikson, 2018). In these cases there’s more existence of dissimilarity among team members related to psychological characteristics, including personalities, cultural values and attitudes (Harush, Lisak and Glikson, 2018; Stahl, Mäkelä, Zander and Maznevski, 2010) which effects the relations negatively (Jehn 1995, 1997). Additionally the lack of face-to-face communication in virtual teams can create communication and coordination difficulties in case of a high degree in reciprocal interdependence (Thompson, 1967). This will hamper performance and hinder collaboration, because interactions are mediated by relatively lower bandwidth channels (Bell & Kozlowski, 2002; Bordia, 1997). Over time this will also negatively affect the relationship between the team members (Jehn 1995, 1997). Negative effects on the relationship between team members in turn increases the risk on potential cost or damage for the feedback seeker (Vancouver & Morrison, 1995) which will affect the feedback inquiry and feedback frequency negatively (Ashford & Tsui, 1991; Wu, Parker & de Jong, 2014). Therefore it can be hypothesized that the indirect relationship between Global Identity and Feedback Inquiry and Feedback Seeking Frequency is moderated by team interdependence as the positive effect on Feedback Inquiry and Feedback Seeking Frequency will appear under low rather than high team interdependence levels.

H7a: Global Identity of an employee in a multicultural virtual team will be positively related to Feedback Inquiry through mediation of Perceived Proximity. This effect will prevail under low, rather than high levels of team interdependence.

H7b: Global Identity of an employee in a multicultural virtual team will be positively related to Feedback Seeking Frequency through mediation of Perceived Proximity. This effect will prevail under low, rather than high levels of team interdependence.

These hypotheses are visualized in figure 1.

(15)

15 Figure 1: Conceptual Model

3. Method

3.1 Sample and data collection

Data has been collected at organization X, an online tech organization. At that time there were 707 employees working from home in virtual teams using digital channels for communication. Among the 707 employees, 47 different nationalities were counted from all over the globe. The majority of the employees work in Western Europe and a few work in South East Asia. The working language is English and employees were hired with the requirement of a sufficient level of English, assessed during the hiring process.

Data has been collected by means of an anonymous online Survey through the software Qualtrics. To perform the analyses, Statistical software Package for Social Sciences (SPSS) was used. Survey administration started on May 28^th, 2021 and was closed one week later on June 4^th, 2021. Through pilot testing it has measured that completing the survey took 8.58 minutes average.

All 707 employees received a request to fill in the online survey through their work e- mail, see appendix 1 Invitation Email and appendix 2 Research Survey. To reduce social desirability bias the survey has been conducted anonymously (Joinson, 1999), detection of IP Addresses through Qualtrics was de-activated and the Survey Introduction mentioned the following statement; “Please keep in mind that - there are no right or wrong answers - only your OWN opinion is important”. From the 707 employees 395 employees opened the survey.

To ensure high research quality and prevent careless responses, all responses below three minutes were removed with the assumption that it takes at least more than three minutes to fill in the survey consciously (Meade & Craig, 2012). Among the removed responses the majority only opened the survey or completed the first few questions. This left a sample of 333 respondents (response rate 47.1%). The sample consists of employees working on a partime or

(16)

16 fulltime base (M^age = 38.11, SD^age= 10.32, age-range = 22-74 years). From the 35.4% are female. The sample covered 41 different nationalities (Dutch = 28.9%, Italian = 15.9%, French

= 10.7%, German = 8.4%, Spanish = 6.1%, British = 5.2%, South American = 4.2%, Asian = 3.6%, African = 2.6%, other European = 13.2%, North American = 0.6%, Australian = 0.3%, Canadian 0.3%). The average amount of years that respondents worked at organization X is 2.71 years (MEmployment duration at x = 2.71, SD Employment duration at x= 2.19, organization X Employment duration-range = 0-9, N = 333). The job roles among this sample includes all roles that one can expect in an online auction platform (Customer Experience = 19.4%, Tech = 17.1%, Core Business (Sales, Marketing & Auctioneers) = 51.9%, Staff & Other = 11.6%).

3.2 Measures

To ensure validity and reliability previously researched survey questions have been adopted consistent with the original versions of the conceptual model (Saunders, Lewis, 2018).

All original studies showed a Cronbach’s alpha higher than 0.79. The measurements were conducted at the individual level and all scales were measured either on a seven-item or on a five-item Likert-type scale, consistent with the original versions. All items used in the survey were derived from English studies. As the working language of the sample is English, no translations were required. All items in the sample were tested on reliability to ensure internal consistency of the scale of the variables. Cronbach’s alpha was consistently higher than 0.7 with an exception on Feedback Inquiry which led to removal of Feedback Inquiry’s fifth item.

By computing the items into variables based on the sum of scores, the survey was tested on validity based on the Critical Values of Pearson Correlation Coefficient table (Ramsey, 1989).

Except for the item of Feedback Inquiry that was excluded earlier, all items were significantly higher than 0.113 (327 DF (.05) = .113). This reconfirmed the decision to remove this item from further analyses.

Global Identity

Global Identity was measured using five items developed and validated by Shokef and Erez (2006, 2008) and Harush, et al., (2018) (Cronbach’s α = 0.89). The measurements consisted of items that assessed the extent to which respondents identified themselves as part of a global environment. Example items are “I see myself as part of the global international community” and “I relate to people from other parts of the world as if they were close acquaintances/associates”. The items were measured on a seven-point likert scale ranging from

(17)

17 1 (“not at all”) to 7 (“to a great extent”). Reliability, validity and descriptive tests have been applied (α = 0.88, MGlobalIdentity = 5.54, SDGlobalIdentity= 1.11, r ≤.77, p<.05).

Perceived Proximity

Perceived proximity was measured using seven items of the perceived proximity scale (O’Leary, Wilson & Metiu, 2012). The original conceptual model study from Harush, et al.

(2018) created one item based on the Perceived Proximity scale of O’Leary, Wilson and Metiu (2012) to measure Perceived Proximity. However using one item to measure a variable decreases validity and limits the interpretation of measurements (Saunders & Lewis, 2018).

For this reason it was decided to differentiate from the original conceptual model study and adopt seven items of the original source instead. For this scale reliability coefficients ranged from 0.47 to 0.70 (O’Leary, Wilson & Metiu, 2012). The measurements consisted of items that assessed to what extent the respondents felt a psychological proximity towards team members.

In this context the survey referred to “team members” as the group of colleagues that the respondent regularly had work related interactions with. Example items are “Even when we are not working in the same place, I still feel close to my team members?” and “I have a warm feeling about my team members”. The items were measured on a five-point likert scale ranging from 1 (“not at all”) to 7 (“to a great extent”). Reliability, validity and descriptive tests have been applied (α = 0.87, MPerceivedProximity = 5.09, SDPerceivedProximity = 1.06, r ≤.65, p<.05).

Team Interdependence

Team interdependence was measured using the four items used by Harush, et al., (2018) (Cronbach’s α = 0.79) which are based on the validated team interdependence scale by Van Der Vegt, Emans, Van de Vliert, (2001). The measurements consisted items that assessed the extent to which respondents were interdependent on team members in completion of their daily tasks. In this context the survey referred to “team members” as the group of colleagues that the respondent regularly has work related interactions with. Example items are “I depend on my team members for the completion of my work” and “In order to complete their work, my team members have to obtain information and advice from me”. The items were measured on a seven-point likert scale ranging from 1 (“not at all”) to 7 (“to a great extent”). Reliability, validity and descriptive tests have been applied (α = 0.87, MTeamInterdependence = 4.74, SDTeamInterdependence= 1.41, r ≤.79, p<.05).

(18)

18 Relational Conflict

Relational conflict was measured using the three items used by Harush, et al., (2018) (Cronbach’s α = 0.89) which are based on the Relational Conflict scale of Jehn and Mannix’s (2001). The measurements consisted items that assessed the extent that respondents experienced relational tension, emotional tension and anger in the team. Example items are

“How much relationship tension is thee in your team?” and “How often do people get angry while working in your team?” In this context the survey referred to “team” as the group of colleagues that the respondent regularly has work related interactions with. The items were measured on a seven-point likert scale ranging from 1 (“not at all”) to 7 (“a lot”). Reliability, validity and descriptive tests have been applied (α = 0.87, MRelationalConflict = 2.20, SDRelationalConflict

= 1.10, r ≤.87, p<.05).

Feedback Inquiry and Feedback Seeking Frequency

Feedback was measured through Feedback Seeking Frequency and Feedback Inquiry, using nine items of the Feedback Orientation scale developed and validated by Ashford and Cummings (1983). Only items related to interaction with colleagues within the context of feedback seeking behavior were selected to ensure relevance for the conceptual model. Three out of the nine items were related to the frequency a respondent applies feedback seeking behavior towards colleagues either through observations or through inquiry. An example question is “In order to find out how well you are performing in your job, how frequently do you seek information from your colleagues about your work performance?”. These three items were measured on a five point likert scale ranging from 1 (“very infrequently”) to 5 (“very frequently”). Reliability, validity and descriptive tests have been applied on the three items (α

= 0.87, MFeedbackFrequency = 2.97, SDFeedbackFrequency= 0.86, r ≤.74, p<.05).

The remaining six items measured the respondent’s perception on potential beneficial outcomes versus potential risks of showing feedback seeking behavior towards colleagues.

Example items are “It takes a lot of effort to get useful feedback from my colleagues” and “It is better to try and figure out how you are doing on your own rather than ask your colleagues for feedback”. For these items a five point likert scale was used ranging from 1 (“Strongly disagree”) to 5 (“Strongly agree”). Five out of six items were reverse coded meaning that a relatively low score indicated relatively high levels of Feedback Inquiry. For these five items the five point likert scale was re-coded in such a way that higher numerical scores, represented higher levels of Feedback Inquiry. See appendix 2 Research Survey for more information on the applied items. Testing for reliability of the five items gave a Cronbach alpha lower than the

(19)

19 required 0.7 (Cronbach’s α = 0.68) due to the fifth item with a corrected Item-Total Correlation of less than 0.3 (0.148). Also the validity testing for the fifth item showed correlation values lower than the required 0.113 with the first item (r = -.011, p >.05), the fourth item (r = .11, p

<.05) and the sixth item (R = .01, p >.05) of Feedback Inquiry (Ramsey, 1989). Given these values and since the fifth item was the only non-reversed item of Feedback Inquiry, there is reason to believe that this item confused respondents. After removing the item, the cronbach alpha improved (α = 0.76). Validity and descriptive tests have been applied (MFeedbackInquiry = 4.30, SDFeedbackInquiry= 0.59, R≤.54, p<.05).

Control Variables

Age, Gender, Nationality, Department at organization X, employment duration at organization X, Different Nationalities team and Cultural Intelligence are the control variables of the conceptual model. Like the study from which the conceptual model originates, the measurements were controlled on the amount of different nationalities that respondents worked with on a monthly basis. All respondents worked with at least three different nationalities on a monthly basis (Mnationalities = 8.74, SDnationalities= 4, nationalities-range = 3-20) and are working from home. Cultural Intelligence was measured using eight items of the Cultural Intelligence scale developed and validated by Ang, Van Dyne, Koh, Ng, Templer, Tay & Chandrasekar (2007). (Cronbach’s α = 0.90). The selected items related to interaction with colleagues within the context of cultural intelligence were considered relevant for the conceptual model. The items were measured on a seven-point likert scale ranging from 1 (“Strongly disagree”) to 7 (“Strongly agree”). Reliability, validity and descriptive tests have been applied (α = 0.82, MCulturalIntelligence = 5.36, SDCulturalIntelligence= 0.81, r ≤.59, p<.05).

3.3 Statistical Procedure

The scale means were computed for all variables from the conceptual model and descriptive statistics, skewness, kurtosis and normality tests were applied to ensure research quality. The variables Cultural Intelligence and Feedback Inquiry were negatively skewed with positive kurtosis and Relational Conflict and Different nationalities were positively skewed with positive kurtosis. The non-normality appeared to be caused by outliers which were identified and examined based on Z scores to ensure no response errors were included.

Exclusion of 25 outliers corrected the skewness and kurtosis to values fitting a normal distribution and created the final sample of 308 responses. Despite this, the Kolmogorov-

(20)

20 Smirnoff test shows significance on all variables, which means that the data sample does not have a normal distribution. To solve this one of the approaches considered is transforming the variables. However this is only recommended in case of small sample sizes (N<30) as it changes the scale and therefore complicates the interpretation of the variables and prediction model. Instead we follow the Central Limit Theorem (CLT) which indicates that, regardless of the distribution of the data, when N is at least 30 (preferably at least 50), the sampling distribution has a normal shape, even if the data deviates extremely from normality (Serfling, 1968). As this data sample is larger (N=308) statistical tests will be robust against violations of the univariate normality and outcomes will not be affected severely (Serfling, 1968). To put this to the test, a Pearson correlation which requires the data to be normally distributed, and a Spearman correlation test without this requirement, were executed with the outcomes visible in table 1. In comparison the outcomes on correlation showed minor differences (∆≤.07, see appendix 3 for Pearson correlation outcomes) with two exceptions. The correlation between Age and Team Interdependence showed a weak significant correlation in case of the Pearson test but did not measure a significant correlation in the Spearman test (r = -.14, ρ = -.07). The same applies to the correlation between Perceived Proximity and Relational Conflict (r = -.14 , ρ = -.11). To ensure validity, we will follow the Spearman correlation outcomes that will not require a normal distribution (Field, 2013).

Table 1: Means, Standard Deviations, Correlations

M SD 1 2 3 4 5 6 7 8 9 10

1. Global Identity 5.54 1.11 (.88)

2. Perceived Proximity 5.09 1.06 .35** (.87) 3. Relational Conflict 2.20 1.10 -.07 -.11 (.87)

4. Feedback Inquiry 4.30 .59 .16** .29** -.36** (.76) 5. Feedback Frequency 2.97 .86 .06 .04 .07 .05 (.87)

6. Interdependence 4.74 1.41 .11 .24** .14* .05 .22** (.87)

7. Cultural Intelligence 5.36 .81 .52** .26** -.02 .10 .15** .14* (.82)

8. Age 38.11 10.32 .10 .16** .06 -.05 -.20** -.07 .13* -

9. Gender 1.35 .48 .06 .01 .04 .10 .11 .03 .12* -.27** -

10. Interaction nationalities 8.74 4.00 .02 .04 .08 -.01 .16** .21** -.07 -.14** -.00 - 11. Employment duration 2.71 2.19 -.13* ,17** .28** -.10 -.13* .03 -.05 .46** -.15** .04

**. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.05 level (2-tailed).

(Cronbach Alpha > 0.76)

(21)

21 4. Results

This chapter will mention the remarkable (non) relations between variables. Based on the outcomes shown in table 1, it can be concluded that Global Identity has a moderate positive correlation with Perceived Proximity (ρ =.35, p<.01), consistent with the findings of Harush, et al., (2018). Perceived Proximity however did not show a significant correlation with Relational Conflict (ρ = -.11, p>.05). As mentioned in the previous chapter, if the Pearson correlation would be taken into consideration rather than the Spearman correlation, there would be a weak negative correlation (r = -.14, p<.5). However taking this correlation into account requires a cautious interpretation and implies that the strength and the significance of this correlation is inconsistent with the findings of Harush, et al., (2018), who did find a moderate negative correlation between Perceived Proximity and Relational Conflict on team level (r = - .30, p<.05, N = 83). Feedback Inquiry has a weak positive correlation with Global Identity (ρ

=.16, p<.01) and a moderate positive correlation with Perceived Proximity (ρ =.29, p<.01).

This is consistent with the findings on previous research (Vancouver & Morrison, 1995; Wu, Parker & de Jong, 2014). Feedback Seeking Frequency however does not have a significant correlation with Global Identity (ρ =.06, p>.05), nor with Perceived Proximity (ρ =.04, p>.05).

Interdependence is significantly correlated with Perceived Proximity (ρ =.24, p<.01), Relational Conflict (ρ =.14, p<.05) and Feedback Frequency (ρ =.22, p<.01). Additionally the control variables Cultural Intelligence and Employment Duration at organization X have significant correlations with the majority of the variables in the conceptual model. These correlations are consistent with previous findings (Harush, et al., 2018; Erez, Lisak, Harush, Glikson, Nouri, & Shokef, 2013; Shokef and Erez, 2008; Vancouver & Morrison, 1995; Wu, Parker & de Jong, 2014). It is worth mentioning that Feedback Inquiry seems to have a moderate negative correlation with Relational Conflict (ρ =-.36, p<.01), according to Vancouver & Morrison (1995) this can be explained due to social component within feedback Inquiry and the importance the relationship between the feedback seeker and the provider.

4.1 Regression Hypothesis Testing

In order to identify the relationships in the conceptual model and test the hypothesis, Hierarchical Multiple Regression tests were carried out. For these tests the control variables Employment Duration at organization X and Cultural Intelligence were included, consistent with the research of Harush, et al., (2018). See table 2 for the Hierarchical Multiple Regression Model results. The first test was carried out to investigate whether Global Identity could

(22)

22 significantly predict Perceived Proximity towards team members in virtual teams, after controlling for Employment Duration at organization X and Cultural Intelligence. The results indicated that the overall model explained 17.3% of the variance and that the model was a significant predictor for Perceived Proximity, F(3, 304) = 21.15 p < .01. Therefor H1 is supported as Global Identity contributed significantly to the model (β = .33, p < .01) with a moderate positive effect in terms of strength in relation to Perceived Proximity. Additionally three separate hierarchical multiple regressions were carried out to investigate whether Perceived Proximity could significantly predict Relational Conflict, Feedback Inquiry and Feedback Seeking Frequency after controlling for Employment Duration at organization X and Cultural Intelligence. For the first test, the result indicated that the overall model explained 8%

of the variance and that the model was a significant predictor for Relational Conflict, F (3, 304)

= 8.84, p < .01. In other words H2 is supported as Perceived Proximity contributed significantly to the model (β = -.19, p < .01) with a weak negative effect in terms of strength in relation to Relational Conflict. The second test was applied with Feedback Inquiry as the dependent variable. The results indicated that the overall model explained 10.5% of the variance and that the model was a significant predictor for Feedback Inquiry F (3, 304) = 11.84, p < .01.

Therefore it can be concluded that H5a is supported as Perceived Proximity contributed significantly to the model (β = .32, p<0,01) and shows a moderate positive effect in terms of strength in relation to Feedback Inquiry. The last hierarchical multiple regression was carried out to investigate whether Perceived Proximity could significantly predict Feedback Seeking Frequency. The results indicated that the overall model explained 3.9% of the variance, F (3, 304) = 4.11 p < .01 but Perceived Proximity was not a significant predictor of Feedback Seeking Frequency (β = .08, p > .05). From this it can be concluded that H5b and H6b are not supported. In other words the level of proximity that one experiences towards team members in a virtual team does not affect the frequency someone seeks for feedback through observation and/or inquiry.

(23)

23 Table 2: Hierarchical Multiple Regression Model

Unstandardized Coefficients

Standardized Coefficients

DV: Perceived Proximity B SE Beta p R2 R2 Change F Change p

Step 1 .10** .10** 15.93** .00

Employment Duration at X .09 .03 .19** .00 Cultural Intelligence .35 .07 .26** .00

Step 2 .17** .08** 28.7** .00

Employment Duration at X .11 .03 .22** .00

Cultural Intelligence .12 .08 .09 .13

Global Identity .32 .06 .33** .00

DV: Relational Conflict B SE Beta p R2 R2 Change F Change p

Step 1 .04** .04** 7.28** .00

Step 2 .08** .04** 11.45** .00

Perceived Proximity -.20 .06 -.19** .00

DV: Feedback Inquiry B SE Beta P R2 R2 Change F Change p

Step 1 .01 .01 1.76 .17

Employment Duration at X -.02 .02 -.08 .16

Step 2 .11** .09** 31.66** .00

Employment Duration at X -.04 .02 -.14* .01

Cultural Intelligence -.1 .04 -.02 .75

Perceived Proximity .18 .03 .32** .00

DV: Feedback Frequency B SE Beta p R2 R2 Change F Change p

Step 1 .03** .03** 5.31** .01

Cultural Intelligence .14 .06 .13* .02

Step 2 .04** .01 1.69 .19

Perceived Proximity .06 .05 .08 .19

(24)

24

**. Significance at the 0.01 level (2-tailed).

*. Significance at the 0.05 level (2-tailed).

4.2 Mediation Hypothesis testing

PROCESS macro for SPSS of Hayes (2013) has been used tor testing the mediation hypothesis H3 and H6a. Mediation regression tests have been applied based on model 4, applying bias corrected bootstrapped confidence intervals (using 5,000 bootstrap samples), while controlling for Employment Duration at organization X and Cultural Intelligence, consistent with the research of Harush, et al., (2018). Since the variables in the conceptual model use different scales, the standardized coefficients have be used to reflect on the strength of the effect. The direct effect of Global Identity on Relational Conflict (c1 = .10, p >.05) is not significantly different from zero, t = .1.336, p =.782, with a 95% confidence interval from - .047 to .247. The indirect effect of Global Identity on Relational Conflict through Perceived Proximity (a1b1 = -.08), is statistically different from zero, as revealed by a 95% BC bootstrap confidence interval that is entirely below zero (CI = -.143 to -.027). This means that two employees who differ by one unit in their reported Global Identity are estimated to differ by .08 units in Feedback Inquiry. In other words those who identify themselves more with their global surroundings are more likely to feel close to team members in virtual teams which in turn leads to slightly lower experience of relational tension in their multicultural diverse teams.

Based on the outcomes in table 4, it can be concluded that H3 is supported.

Table 4: Mediation test model 4 with dependent variable Relational Conflict Consequent

Perceived Proximity (M) Relational Conflict (Y)

Antecendent Coeff. SE p Coeff. SE p

Global Identity (X) a1 .331 .061 <.001 c1 .100 .0575 .183

Perceived Proximity (M) --- --- --- b1 -.241 .075 <.001

Employment duration at X (C) .097 .024 <.001 .132 .029 <.001

Cultural Intelligence (C) .115 .075 .130 .032 .088 .720

constant i1 -.880 .412 <.05 i2 1.677 .481 <.001

R2 = .173 R2 = .088

F(3, 304) = 21.152, P<.001 F(4, 303) = 7.089, P<.001

The direct effect of Global Identity on Feedback Inquiry (c1 = .04, p >.05) is not significantly different from zero, t = .916, p =.360, with a 95% confidence interval from -.041 to .114. But the indirect effect of Global Identity on Feedback Inquiry through Perceived Proximity (a1b1 =

(25)

25 .06), is statistically different from zero, as revealed by a 95% BC bootstrap confidence interval that is entirely above zero (CI = .027 to .099). This means that two employees who differ by one unit in their reported Global Identity are estimated to differ by .06 units in Feedback Inquiry. In other words those who identify themselves with their global surroundings are more likely to feel close to team members in virtual teams which in turn leads to slightly higher risk tolerant perceptions on seeking feedback from team members in multicultural diverse teams.

Based on the outcomes in table 5, it can be concluded that H6a is supported.

Table 5: Mediation model 4 with dependent variable Feedback Inquiry

Consequent

Perceived Proximity (M) Feedback Inquiry (Y)

Antecendent Coeff. SE p Coeff. SE p

Global Identity (X) a1 .331 .061 <.001 c1 .036 .039 .360

Perceived Proximity (M) --- --- --- b1 .178 .035 <.001

Employment duration at X (C) .097 .024 <.001 -.034 .015 <.05

Cultural Intelligence (C) .115 .075 .130 -.033 .046 .478

constant i1 -.880 .412 <.05 i2 4.566 .253 <.001

R2 = .173 R2 = .107

F(3, 304) = 21.152, P<.001 F(4, 303) = 9.091, P<.001

4.3 Moderation Hypothesis testing

For testing the moderating hypothesis, PROCESS macro for SPSS of Hayes (2013) has been used. Moderation regression tests have been applied while using model 1 and applying

Antecedent Coeff. SE p LLCI ULCI

Direct effect c1 .100 .075 .782 -.047 .247

Total effect c .020 .073 .782 -.123 .163

BOOT SE Boot LLCI BOOT ULCI

Indirect effect a1b1 -.080 .030 -.143 -.027

Antecedent Coeff. SE p LLCI ULCI

Direct effect c1 .036 .039 .360 -.041 .114

Total effect c .095 .039 <.05 .018 .172

BOOT SE Boot LLCI BOOT ULCI

Indirect effect a1b1 .059 .018 .027 .099