Master’s Thesis
Master’s Programme in Communication Science
The interplay between virtual teamwork, workplace conflicts, the
solution-oriented communication strategy and team performance
Author: Konstantina Mermela (11391464) Supervisor: Dr. Pernill van der Rijt
2
Abstract
Contemporary organizations are increasingly relying on virtual teamwork to pursue their objectives. The spatial and temporal distribution and the asynchronous communication, which characterize virtual teams, provoke the rise of three types of workplace conflicts: task, relationship, and process conflict. The aim of this thesis is to examine the relationships between virtual teamwork and types of conflicts, as well as the effects of them on team performance. The role of the solution-oriented communication strategy for conflict management is investigated as a variable that mitigates the potential negative effects of workplace conflicts on team performance. Through the execution of an online survey, we collected operationalizable data from 174 individuals who participate in work teams. The statistical results were proved to be insignificant for our hypotheses; nonetheless, team cohesion was revealed as a salient variable, which intervenes between teamwork and workplace conflicts, and is positively related to team performance. The current thesis offers to communication practitioners and organizations an intelligible view of the communication challenges taking place among virtual teams, and an accurate guide for dealing with workplace conflicts.
3
Introduction
In the ever-changing business world, corporations invest in the integration of virtual work environments and mobile work systems to their organizational structures (Shah, 2014; Chudoba, Wynn, Lu, & Watson-Manheim, 2005). Virtual teams are crucial for the survival of contemporary companies due to globalization, increased market competition, mergers and acquisitions and corporate layoffs (Lurey & Raisinghani, 2000). Virtual teamwork is a notion which stems from the context of new ways of working. It signifies the accomplishment of specific organizational objectives by geographically dispersed or remote individuals, who work interdependently and simultaneously without sharing physical co-presence (Shah, 2014; Tjosvold, West, & Smith, 2003; Mihhailova, Oun, & Turk, 2009). The rise of virtual teams generated not only from the geographical distribution of modern organizations, but also from the development of communication and information technologies (ICTs) (Ter Hoeven & Van Zoonen, 2015; Shah, 2014). Employees who work in virtual teams are enabled to access information related to their work, to communicate in real-time with their colleagues and to coordinate their activities across the boundaries of time and space (Ter Hoeven & Van Zoonen, 2015; Lurey & Raisinghani, 2000; Pauleen, 2004).
The formation of virtual work teams brings about a large number of advantages. Virtual teamwork is related to moderated demands for office spaces; reduced travel time and expenses; productivity benefits; better resource utilization; improved performance; greater visibility of working progress; and utilization of employees’ talents and competencies in a more efficient way (Shah, 2014; Tjosvold, West, & Smith, 2003). From the employees’ side, the spatial distribution and the asynchronous nature of virtual teamwork afford them with greater control over their work, as they can optimize their schedule, which in turn enhances their work and life balance (Ter Hoeven & Van Zoonen; Putnam, Myers, & Gaillard, 2014; Tjosvold, West, & Smith, 2003).
4
However, except for the abovementioned benefits and tremendous potential of virtual teams, the space-time dispersion and the use of ICTs might also pose challenges to
communication and effective teamwork (Kankanhalli, Tan, & Wei, 2006; Griffith, Mannix, & Neale, 2002). The most prominent challenge which is linked to virtual teamwork and which will be the focus of research in the current thesis is the conflicts that arise within virtual teams. Conflict is considered to be a “normal and natural part” of a workplace. It takes place when two or more interacting individuals, who have divergent beliefs, skills, values and attitudes, have to perform an activity in a mutually dependent manner (Zafar, Ashfaq, Ali, & Imran, 2014). Conflicts and disputes in the workplace have several origins, namely lack of consensus, antagonism, espousal of different communication behaviors and tactics, as well as a general heterogeneity among team members (Jehn, 1995; Putnam &Wilson, 1982).
Especially in virtual teamwork contexts, the above-mentioned origins of conflicts are
complemented by lack of face-to-face contact, communication delays, time zone differences and cultural diversity (Kankanhalli, Tan, & Wei, 2006; Chudoba et al, 2005). These
disturbances might hinder the restoration of the relationships among colleagues and, subsequently, team performance (Kankanhalli, Tan, & Wei, 2006; Nataatmadja & Dyson, 2006).
Task, relationship and process conflict are the prevalent conflict types, which are typically found in every workplace and constitute an inevitable part of the organizational life (Jehn, 1997; Putnam & Wilson, 1982). Each of these workplace conflict types might reduce the assessment of information provided by team members and waste the time and energy of employees. Moreover, the group tasks or projects themselves take second place (Jehn, 1995; 1997). Nevertheless, workplace conflicts could also lead to positive outcomes for team performance, if they are managed effectively. Therefore, team members should develop and
5
integrate some communicative strategies in their work routine, in order to deal with emerging conflicts and disputes (Jehn, 1995; Putnam & Wilson, 1982).
Three principal communicative conflict management strategies are proposed by organizational communication theorists (Putnam & Wilson, 1982; Wilson & Waltman, 1988): the non-confrontational, the control, and the solution-oriented strategy. Non-confrontation involves indirect communicative behaviors, such as withholding of negative feelings,
avoidance or withdrawal from disputes, and silence. It is considered a “lose-lose” approach to conflict management. The control strategy pertains to direct communication tactics for
handling conflicts. In particular, controlling entails persistent advocacy of one’s beliefs or position, non-verbal forcing and taking control of the interaction between team members. This strategy is considered to be a “win-lose” conflict approach (Putnam & Wilson, 1982; p. 647). Herein, the solution-oriented communication strategy will be the focus of research, as it constitutes a “win-win”, widely acceptable, and efficient strategy in dealing with different types of conflict (Putnam & Wilson, 1982; Gross, Guerrero, & Alberts, 2004; De Dreu et al., 2001;Kankanhalli, Tan, & Wei, 2006).
In the light of the abovementioned notions and facts that take place in organizational contexts, the current thesis aims to examine the relationships between virtual teamwork, workplace conflicts, communication-based conflict management strategies and team performance. Therefore, the principal research question is formulated as follows:
“How do workplace conflicts and communicative conflict management strategies intervene between virtual teamwork and team performance?”
The intention of this research study is to shed light to the different types of conflict which arise in the context of virtual teamwork and establish causal relationships between these variables. The rise of workplace conflicts has its origins in the lack of efficient strategic
6
communication tactics between team members in organizations and businesses of any size (Zafar et al., 2014); therefore, the significance of internal corporate communicative strategies for conflict resolution is emphasized. There are quite a few theories and empirical research regarding conflicts in virtual teams (Gross, Guerrero, & Alberts, 2004; Kankanhalli, Tan, & Wei, 2006; Holahan et al., 2014; De Jong, Schalk, & Curseu, 2008; Mortensen & Hinds, 2001); however, to our knowledge, there is not sufficient cross-sectional quantitative research done about the types of conflict that stem in organizational virtual teams, as well as their impact on work performance on a team-level scope (Kankanhalli, Tan, & Wei, 2006; De Jong, Schalk, & Curseu, 2008; Mortensen & Hinds, 2001). Furthermore, a supplementary innovation of the current thesis is the attempt to establish a causal relationship between each type of workplace conflicts with perceived team performance, while including in the model the role of the solution-oriented communication strategy. To conclude, communication practitioners and organizations will be offered a clearer and more intelligible view of the communication challenges taking place among virtual teams, as well as a more accurate guide for dealing with conflicts, aiming at the empowerment and the enhancement of team
performance.
The current thesis consists of five sections. After the introduction, a review of the theoretical foundation is presented, on which the principal variables and notions of this paper are based. In the third section the research methodology is analyzed, while section four contains the results, as they were revealed by the statistical analysis. Finally, the results are interpreted, and theoretical and practical implications of the thesis are discussed in the last section.
7
Literature Review
2.1 Properties of virtual teamwork in the contemporary workplace
Organizations are increasingly relying on teamwork in order to overcome complex problems and to pursue their goals (Van Roosmalen, 2012). When teamwork takes place in a virtual context, it provides organizations of any size with tremendous flexibility (Shah, 2014). This flexibility of working arrangements encompasses three prevalent dimensions of
virtuality: spatial flexibility, temporal flexibility, and the utilization of information and communication technologies (Weber & Kim, 2015; Chudoba et al., 2005; Ter Hoeven & Van Zoonen, 2015).
Spatial flexibility refers to the extent to which employees work on their tasks and projects, while they are distributed across different geographic locations (Weber & Kim, 2015). It implies that physical presence is no longer an integral and fundamental part of communication and of the interactions between employees; team members can effectively cooperate without being in the same physical location (Shah, 2014). According to Chudoba et al. (2005), spatial flexibility forms “the nexus of all conceptualizations of virtual teamwork”, meaning that geographic dispersion is the main assumption of virtual work groups. Therefore, location becomes less important to modern organizations, where the perception that work is “something you do, not someplace you go” is gaining ground (Helms & Raiszadeh, 2002, p. 241). In this way, the most qualified people can be brought together and form a work or project team, even if they are located in different cities, countries or continents (Pauleen, 2004; Chudoba et al., 2005).
Temporal dispersion refers to the variation and the degree of overlap in work hours between members of collaborating work teams (Colazo, 2014; O’Leary & Cummings, 2007). The variation pertains to difference in time zones or difference in work hours, due to the
8
unique employment conditions of each individual employee (e.g. part-time employment, fully mobile workers). Consequently, members of virtual teams might have to work and collaborate across different time zones (Chudoba et al., 2005; Colazo, 2014). This postulation induces the notion of asynchronous communication, which enables the cooperation between temporal dispersed team members (O’Leary & Cummings, 2007). In addition, people might hold different perceptions of time (clock-time perception in Western societies versus event-time perception of Japanese culture). These divergent perceptions imply a cultural and work process diversity among individuals who come from different countries, and have adopted divergent national or professional cultures (Saunders, Van Slyke, & Vogel, 2004; Chudoba et al., 2005). All things considered, temporal dispersion of virtual teams increases the reliance on asynchronous communication, and it extends employees’ work hours, due to elongation of response time and delays in coordination of tasks, roles and processes (Weber & Kim, 2015; O’Leary & Cummings, 2007). In addition, multicultural collaboration poses challenges to the communication and collaboration among members of virtual teams, as they might hold discrepant work attitudes and worldviews (Leung, Ang, & Tan, 2014).
The third dimension of workplace virtuality, namely the use of information and communication technologies, is considered the “enabler” of virtuality (Chudoba et al., 2005, p. 281). It supports the collaboration, the coordination and the communication about work-related activities among team members in different time zones and places (Weber & Kim, 2015). Without ICTs, the implementation of virtual teaming environments would be impossible (Chudoba et al., 2005). Distributed co-workers rely on online communicative technologies for enhancing the “breadth and depth of their communication” and subsequently, for counterbalancing the lack of face-to-face communication (Weber & Kim, 2015, p. 389). According to Rice (2017), ICTs in the workplace refer to the computer systems and to
9
information and communication content within and across all the relevant people and affiliated organizations.
Distributed work arrangements became popular in the 1990s; nowadays employees can be fully or partially involved in distributed virtual teams during their workweek. Partially virtual teaming environments might include both distributed work and face-to-face
collaboration (Chudoba et al., 2005; Helms & Raiszadeh, 2002). Therefore, the level of virtual experience (or else virtuality) might vary in each work team (Mihhailova, Oun, & Turk, 2009). Nevertheless, co-working across distance and time with people who hold different cultural and professional values can negatively affect communication in the team, and can escalate the different types of workplace conflict (Chudoba et al., 2005).
2.2 Virtual teamwork and workplace conflict
Workplace conflicts are broadly defined as both the manifest and latent disagreements occurring between two or more interdependent individuals who attempt to collaborate and cope with discrepant views and incompatible work-related goals and ideas (Gross, Guerrero, & Alberts, 2004; Kankanhalli, Tan, & Wei, 2006; Jehn, 1995; Oni-Ojo, Iyiola, & Osibanjo, 2014). In particular, each individual employee may have different work methods and approaches in achieving project goals. Moreover, there are individual attributes which can prove challenging for teams; for example, distinct personalities, dissimilar experiences, interpersonal skills and communication capabilities, as well as inconsistent goals among colleagues (Oni-Ojo, Iyiola, & Osibanjo, 2014). These distinct characteristics of employees often prompt three types of workplace conflict: task, relationship and process conflicts (Jehn, 1995; Jehn, 1997; Behfar et al., 2011; Jehn & Mannix, 2001). This differentiation of
workplace conflicts is deployed in the next sections in order to investigate whether
10
The notion of task conflict refers to the awareness of differences in viewpoints and ideas by the members of group tasks. It results in the rise of tensions, unhappiness and antagonism among team members, and an unwillingness to work together in the future (Jehn, 1995; Behfar et al., 2011; Griffith, Mannix, & Neale, 2002). Precisely, task conflict stems from the disagreements regarding the assigned work or project(s) being performed (Jehn, 1995; Jehn & Chatman, 2000). Individuals who participate in work teams might understand and interpret facts and issues concerning their work, as well as the content and the objective of their assigned tasks, in incompatible ways (De Dreu & Weingart, 2003). For instance, team members often disagree about determining which data should be included in a report or about whether they should implement a new inbound marketing campaign or not (Jehn, Chadwick, & Thatcher, 1997).
In regards to virtual teamwork, geographic dispersion and spanning in multiple time zones can cause confusion and irritation to team members, as they might not be available for discussion and elucidation of task-related issues regularly (Mortensen & Hinds, 2001). In addition, another characteristic of virtual work teams as was discussed in previous sections, is the element of cultural diversity, which can be combined with tenure and informational diversity. According to this, team members with different educational and cultural background, work experience, general or specialized knowledge and functional expertise might perceive the context of each task or project in incompatible ways (Jehn, Chadwick, & Thatcher, 1997; Mortensen & Hinds, 2001). Thus, frequent and substantial communication is a necessity for virtual team members, in order to develop a common understanding about their assigned tasks. Nevertheless, they rarely or never see each other in person and as a result, they have to use ICTs for the successful accomplishment of their task-orientation (Gross, Guerrero, & Alberts, 2004;Kankanhalli, Tan, & Wei, 2006). The empirical study of Kankanhalli, Tan and Wei (2006) proved that ICTs can also pose challenges to team-level communication
11
regarding tasks, since the amount of task-related information might be either incomplete or higher than what members can manage. Therefore, we hypothesize that as the involvement in virtual teamwork increases, there is greater argumentation, debates and disagreements about the assigned work and tasks. Hence:
H1: Higher involvement in virtual teamwork leads to higher levels of task conflict.
Relationship conflict -or else social-emotional conflict- consists of interpersonal animosity, annoyance and dejection among team members (Jehn, 1995; Behfar et al., 2011). The springboard for this conflict type is not related to work and tasks, but to disagreements concerning personal issues, social events, gossip and world news (Jehn & Chatman, 2000; Jehn, Chadwick, & Thatcher, 1997). Hence, discord can be instigated when people who differ from each other in their norms and values, behavior, style, personality or attitude, and they are also aware of their interpersonal incompatibilities, have to cooperate and achieve specific goals (Paul, Seetharaman, Samarah, & Mykytyn, 2005). Relationship conflict enhances anxiety and fear, and inhibits people’s ability to work together efficiently (Jehn, 1995; Behfar et al., 2011). In regards to virtual teamwork, group members have low mutual awareness and familiarity with each other, due to physical distance and the subsequent lack or absence of regular face-to-face communication and interactions (Gross, Guerrero, & Alberts, 2004; Mortensen & Hinds, 2001). According to the social information processing theory (Walther, 1995; 2016), distance and remoteness make the creation of spontaneous and substantial relationships less likely to occur; or they require increased time and increased number of interactions for relational development. This is because team members are accessible mainly by e-mail, phone, videoconferencing or other ICTs (Walther, 1995). Extending the previous thoughts, distant and computer-mediated communication does not allow sufficient social and context information to be conveyed, and thus, virtual teams are likely to experience low level
12
or even absence of relationship conflict (Paul et al., 2005). Therefore, the second hypothesis is formulated as follows:
H2: Higher involvement in virtual teamwork leads to lower levels of relationship conflict.
The third type of conflict, labeled process conflict, refers to the awareness of disagreements between team members about the assignment of duties and resources (Jehn, 1997; Jehn & Mannix, 2001; Jehn, Northcraft, & Neale, 1999; Jehn & Chatman, 2000). More specifically, team members hold discrepant views about how they will delegate the
responsibilities and how they will manage and complete each of the assigned tasks (Jehn, 1997; Jehn & Mannix, 2001; Behfar et al., 2011). In contrast to task conflict, process conflict does not refer to “the content or the substance of the task itself”, but to the means and
strategies for approaching each task or project (Jehn & Bendersky, 2003, p. 201). Arguments about who is responsible for specific duties or parts of a project; how often the team should meet (in person or virtually); through which communication channel they will co-operate and communicate each time; all these are illustrative examples of process conflict experiences within work teams (Jehn & Mannix, 2001; Jehn & Bendersky, 2003).
These disagreements are caused because people hold different insights about how to work on their assigned tasks; more specifically, the procedures, strategies, methods, and policies that should be followed for the successful completion of each task (Jehn, Northcraft, & Neale, 1999; Jehn & Chatman, 2000). Similar to task conflict, it is challenging for virtual groups which consist of members with divergent educational, work, and cultural background to decide how to proceed with a common task (Kankanhalli, Tan, & Wei, 2006). For instance, if a professional with background in finance is asked to complete a project in cooperation with someone with engineering background in virtual contexts, they might identify and choose dissimilar potential courses of action (Jehn, Northcraft, & Neale, 1999). Furthermore,
13
time discontinuities and the lack of regular non-verbal communication of virtual teams can pose challenges to the coordination of the tasks and can also cause misunderstandings
between team members (Mortensen & Hinds, 2001). When complex issues arise regarding the procedures that should be followed for the completion of tasks, or regarding team members’ roles, then there is increased need for explicit and efficient communication between team members. Nonetheless, computer-mediated communication may increase misunderstandings rather than smoothing virtual team members’ interactions (Mortensen & Hinds, 2001; Shah, 2014). Similar to task conflict, the participation in virtual teamwork contexts is expected to be linked to greater coordination and communication challenges and related disputes (Paul et al., 2005). Hence:
H3: Higher involvement in virtual teamwork leads to higher levels of process conflict. 2.3 Workplace conflicts and team performance
In early organizational studies, conflicts were presented as a negative effect of teamwork, which distract team members from their work and tasks (De Dreu & Weingart, 2003). Later on, a general trend was observed in organizational theories, according to which task conflict was considered productive and functional; in contrast, relationship conflict was the counterproductive and dysfunction type of workplace conflicts (Jehn, 1995; De Dreu & Weingart, 2003; Jehn & Chatman, 2000). Nevertheless, empirical evidence provides inconsistent outcomes regarding the direction of each conflict type effects on team
performance (De Dreu & Weingart, 2003; Hanif, Khan, Adeel, & Shah, 2016). Until now, there is an ongoing debate about the relationship between task, relationship, and process conflict types with the overall group performance (Jehn, Chadwick, & Thatcher, 1997; Behfar et al., 2011). In addition, there is a general belief that complete absence of all types of
14
& Chatman, 2000), because team members might not realize their inadequacies and communication problems (De Dreu & Weingart, 2003).
As far as the notion of team performance is concerned, it encompasses two different aspects: objective and perceived performance. The former aspect of performance can be measured by objective criteria, such as sales, customer return or profitability. The latter aspect is related to the perceptions of team members about how well the group is performing (Jehn, Chadwick, & Thatcher, 1997). In the present paper, the focus of interest is the performance-related perceptions of each team member. In general terms, the notion of team performance is widely used in intragroup conflict literature and research, as researchers have always been trying to examine the effects of conflicts on team performance (Hanif et al., 2016; De Jong, Schalk, & Curseu, 2008). As soon as team members gain an understanding of the nature and the effects of workplace conflicts, they will be able to optimize their team performance. The optimization of team performance implies that teams utilize the available resources in more efficient ways, meet their intra-team goals, and contribute to the achievement of the broader organizational objectives (Lurey & Raisinghani, 2000; Van den Vlierte & De Dreu, 1994).
2.4 Task conflict and team performance
Task conflict is the type of conflict that provokes the most controversial effects on team performance (De Dreu & Weingart, 2003; Jehn, Chadwick, & Thatcher, 1997; Hanif et al., 2016). Many literature references support that this type of conflict scrutinizes task-related issues; increases the engagement to information processing; fosters team learning and
knowledge sharing; enhances members’ understanding about the task; and results in the development of creative and effective solutions to problems, ideas and decisions (De Dreu & Weingart, 2003; Jehn & Chatman, 2000; Jehn, 1995). More precisely, when team members offer and evaluate various ideas, solutions and interpretations about the assigned group tasks, then they are able to reach more optimal decisions and outcomes (Jehn, Chadwick, &
15
Thatcher, 1997; Jehn, 1995). In contrast, when team members avoid explicit discussion of task-related disagreements, they inhibit the team from identifying its inefficiencies and from configuring potential news ways to enhance its performance (Jehn, Northcraft, & Neale, 1999; Jehn, 1997).
However, task conflict can also interfere with team performance by hindering the implementation of team goals and by pulling a team away from its specified purpose (Jehn & Mannix, 2001). Performance benefits of task conflict are “highly circumscribed” (Bendersky & Hays, 2012, p. 324). More specifically, only when moderate levels of task conflict are observed, there are benefits for team performance (Jehn & Mannix, 2001; Jehn, 1997). De Dreu and Weingart (2003) who conducted a quantitative review on team conflict literature found out that task conflict is disruptive for team performance, whereas Jehn (1995, p. 261), proved that there is an “optimal level of task conflict”, and team performance lessens beyond or below this level (also in Van den Vliert & De Dreu, 1994). These being said, the
hypothesis for the relationship between task conflict and team performance is formulated as follows:
H4: Task conflict type is negatively related to team performance, only when in low or high levels.
2.5 Relationship conflict and team performance
Relationship conflict has been proved to hurt team performance and to be negatively related to member perception of team performance at any point in the life of a work group (De Dreu & Weingart, 2003; Jehn & Mannix, 2001; Jehn, 1997; Jehn, Chadwick, & Thatcher, 1997). According to Jehn (1995), relationship conflict can negatively affect teams in three ways. Firstly, team members’ ability to access new information provided by their colleagues is reduced. Secondly, members are negative towards the ideas and recommendations of their
16
co-members. Finally, team members spend energy, time and effort on avoiding and resolving interpersonal disagreements and incompatibilities, and thus they lose perspective about the task being performed. Under these conditions, not only team processes are delayed, but also team members withdraw themselves from the group; their motivation and engagement to group activities are clearly lessened (De Dreu & Weingart, 2003; Jehn, Chadwick, &
Thatcher, 1997; Jehn & Chatman, 2000; Jehn, 1995). Manifestations of relationship conflict are all the direct and indirect personal attacks, yelling, name-calling or even throwing things to co-members (Jehn, 1997). The organizational climate becomes uncomfortable and members become suspicious, resentful, irritable (Jehn, Chadwick, & Thatcher, 1997;Jehn, 1997). Even though relationship conflict may be displayed differently across teams, its effect on team performance has been proved to be negative (Behfar et al., 2011; Jehn & Chatman, 2000; Jehn, Chadwick, & Thatcher, 1997). Hence, we hypothesize that:
H5: The level of relationship conflict is negatively related to team performance. 2.6 Process conflict and team performance
In regards to process conflict, academic literature oscillates regarding the direction of its effects on team performance. From the one hand, it is supported that open communication and discussion about the best possible ways to perform a task, as well as the arrangements concerning the best fit of skills and duties within the team may enhance team performance
(Jehn, Northcraft, & Neale, 1999; Jehn & Chatman, 2000). Process conflict can be beneficial mainly in early stages of team creation and prior to beginning of teamwork, when
responsibilities and deadlines are understood and agreed upon, and potential problems regarding task procedures are diagnosed (Jehn & Mannix, 2001; Behfar et al., 2011).
Furthermore, changes in assignments and duties during teamwork are occasionally necessary and efficient for group performance (Jehn, 1997).
17
On the other hand, high level of process conflict is a characteristic of low-performing teams. That is because team members are distracted from the effort they should dedicate to the task itself, and their focus is misdirected to irrelevant discussions about role ambiguities and coordination issues (Jehn & Chatman, 2000; Behfar et al., 2011). Typical manifestation of process conflict is disagreements about who is responsible or capable of doing particular parts of the assigned task (Jehn & Mannix, 2001; Jehn, 1995). Process conflict can also lead to extended delays; “heated discussions” (Jehn & Chatman, 2000); inability of meeting deadlines; and feelings of uncertainty and unfairness, which increase members’ intention to quit or switch groups (Jehn, 1997). Empirical research regarding process conflict leans towards the assertion that this type of conflict is harmful for team outcomes (Behfar et al., 2011; Jehn, 1995; 1997; Jehn & Chatman, 2000). In particular, Behfar et al. (2011),
conducted a survey among groups of MBA students; the findings of this quantitative research revealed a negative relationship between process conflict and each MBA group’s grade. In the same way, Jehn and Chatman (2000), examined the relationship between process conflict and group performance in a sample of 545 employees; team performance was measured by departmental records. The research revealed a negative relationship between the variables, in accordance “with Jehn’s (1995) findings and explanation that all types of conflict are viewed negatively” (Jehn & Chatman, 2000, p. 69). Hence, the hypothesis in this case is formulated as follows:
H6: The level of process conflict is negatively related to team performance. 2.7 Conflict types as mediators between virtual teamwork and team performance
As it was mentioned in previous sections, conflicts are interwoven with the nature and the function of workplace teams and constitute “an inevitable and pervasive aspect” of any organization (De Dreu & Weingart, 2003, p. 746; Putnam & Wilson, 1982, p. 629). Notably, conflicts are a normal outcome which results from people’s attempt to communicate,
18
cooperate and coordinate their activities and efforts. In organizational groups, even if team members agree upon their individual and team goals, they still experience a particular level of conflicts, as well as the multidimensionality of these conflicts (Jehn, 1997). As a result, virtual work teams will experience task, relationship, and process conflicts, which in turn have an impact on team performance (De Jong, Schalk, & Curseu, 2008; Gross, Guerrero, & Alberts, 2004). Notably, and according to the theoretical review which was presented until now, virtual teamwork can lead to positive outcomes for team performance, only if virtual team members experience high or low levels of task conflict, and if they encounter low levels of relationship or process conflict during their cooperation. Therefore, the relationship
between virtual teamwork and team performance is expected to be mediated by the inevitable presence of the three different types of workplace conflicts. Accordingly, the seventh
hypothesis is formulated as follows:
H7: The relationship between virtual teamwork with team performance is mediated by H7a) task, H7b) relationship, and H7c) process conflict type, respectively.
2.8 The role of the solution-oriented communication strategy
The communication-based strategic behaviors and tactics team members deploy in order to manage their task, relationship and process conflicts is a salient process of teamwork in organizations. All types of conflict might have a positive aftermath for team performance under specific conflict management strategies (De Dreu & Weingart, 2003; Putnam & Wilson, 1982). According to Putnam and Wilson (1982), there are three predominant communication tactics for conflict resolution in the workplace, which are widely used in teamwork contexts: non-confrontation strategies, control and solution-oriented strategies. The first strategy is based on disagreement avoidance and controversy sidestepping. Control strategies manage conflict by persistent argumentation and by emphasizing team demands. In the current paper, the solution-oriented strategy is examined, as it has been proved by
19
organizational empirical studies to lead to efficient performance and satisfactory outcomes for all team members (Gross, Guerrero, & Alberts, 2004; Putnam & Wilson, 1982; De Dreu et al., 2001). The solution-oriented communication strategy creates a “win-win” situation in the workplace, where all team members adopt a collaborate style to meet their needs; engage in open communication; and accumulate a variety of points of view to reach the most optimal solution to their problems and make compromises (Putnam & Wilson, 1982; Zafar et al., 2014).
The adoption of this specific strategy is viewed as a socially appropriate tool and is supported by organizations, as it can lighten the negative aftereffects of task-related conflicts on team performance; to wit, team members show appreciation and understanding for their task partners’ values, ideas, and perspectives, and hence they complete their group tasks smoothly and efficiently. In addition, team members who use solution-oriented
communication tactics have more opportunities for sharing information regularly (Gross, Guerrero, & Alberts, 2004; Nordin et al., 2014). As a result of continuous information sharing, the breadth and depth of communication among team members is enhanced (Weber & Kim, 2015). Subsequently, team members have more chances to discuss, understand, and agreed upon their responsibilities and deadlines, and to coordinate their task-related activities (Jehn & Mannix, 2001; Behfar et al., 2010). Therefore, potential negative effects of process conflict on team performance might be downscaled (Gross, Guerrero, & Alberts, 2004). In regards to relationship conflict, open communication and information sharing might lead in aggregation of alternative points of you. Moreover, the solution-oriented conflict management strategy can depersonalize or weaken disagreements which hurt team performance, as it provides team members with opportunities to understand each other in a better way, and it emphasizes on members’ common interests (Gross, Guerrero, & Alberts, 2004; Putnam & Wilson, 1982; Nordin et al., 2014). Taking everything under consideration, it is expected that
20
the solution-oriented communication strategy will moderate the relationship between each task conflict type and team performance, by mitigating the negative effects of conflicts on team performance. Hence:
H8: The solution-oriented communication strategy mitigates the negative effects of H8a) task, H8b) relationship, and H8c) process conflict type on team performance.
The conceptual model provided below depicts the relationships of interest, as they are investigated in the current thesis.
Figure 1. Conceptual model
Method
3.1 Research design
The research question of the current paper is examined through and supported by a quantitative research. A cross-sectional online survey was executed in order to investigate the intervention of workplace conflicts and communicative management strategies on virtual teamwork and team performance. The research instrument was a self-administered
21
online data collection. Five people participated in a pilot study, in order to ensure that the research protocols were followed and the survey items were accurate and comprehensible.
3.2 Sampling strategy and data collection
The restricted available resources (e.g. lack of financial resources, limited time) made it difficult to use a simple random sampling strategy. Therefore, a convenience-snowball sampling procedure was deployed. This method entails the researcher’s approach of a small group of people, who in turn make contact with others; hence, it tends to be fast and
inexpensive as the sampling units are readily available (Bryman, 2016). The researcher posted an anonymous link for the survey on her Facebook networking site and on LinkedIn. Personal messages were also sent to people through these platforms. Both posts and messages included a supportive pithy text, which explained the purpose of the research study and motivated them to participate: the text elucidated that the questions included in the survey pertained to
communication procedures and tensions which arise in the workplace, as well as to new ways of working. In addition, the researcher attempted to motivate people’s participation by stating in the supportive text that the questionnaire would help them to understand more about the dynamics that take place within their work teams. Potential respondents were also asked to forward the questionnaire to their colleagues. Except for the utilization of online social media platforms as distributing channels for the survey, the researcher came in contact with
“Runway East”, a business that provides startup companies in London with coworking spaces and private offices. The employees of the Runway East team completed the questionnaire and distributed it to the companies based in their building. Additionally, the questionnaire was posted to “EF past and present” Facebook group, which is a private group for the startup companies that took part in the accelerator programme “Entrepreneur First” (EF).
Respondents had to be employers or employees who engage in work teams in
22
on Facebook and LinkedIn platforms, in order to receive and complete the online
questionnaire. Moreover, since the questionnaire was in English, respondents needed to be able to read and write in this language.
3.3 The sample
A total of 199 people took part in the survey; however, the operational sample consisted of 174 respondents, who answered all the questions relevant to the variables of interest. Missing items (N=25) were excluded from the sample and the analysis. For the demographic questions we had completed data for 167 respondents. According to the
demographic-related data, 53% of the respondents were male (N= 88) and the mean age was 31 years old (SD= 6.20, ranging from 20 to 51). Concerning the frequency that the
respondents participate in geographically dispersed work teams, 30% (N= 50) reported they never get involved in such teams, whereas 6% (N= 10) indicated they always participate in virtual work teams on a scale ranging from 1= “never” to 7= “always” (M= 3.03, SD= 1.9); In addition, respondents who work in private organizations constituted the majority of the operational sample, 84% (N= 141). The average number of team co-members for the
respondents of the current survey was 58 people (M= 10.13, SD= 8.56, Mdn = 7, ranging from 2 to 60). Finally, the majority of respondents had the Hellenic nationality (85%), whereas 7% were British (N= 12).
3.4 Description of the questionnaire
The distributed questionnaire consisted of seven blocks of questions, the majority of which provided a seven-point answering scale, ranging from 1= “strongly disagree” to 7 = “strongly agree” (see Appendix A). Respondents were asked to indicate their level of (dis)agreement in statements related to a range of questions, which were expected to shed light to the study’s purpose. In particular, the initial block consisted of questions about the kind of tensions arise in respondents’ work teams. Thereafter, they had to indicate the level of
23
cohesion in their work teams, while the third block contained questions concerning workplace tasks. The fourth block exhort respondents to think of disagreements they have encountered with their fellow team members, and to indicate the frequency they engage in specific conflict-management behaviors. The fifth block contained questions which aimed to reveal the level of virtuality in respondents’ teamwork. The sixth block of questions was about respondents’ team performance.
Finally, a few demographic questions followed, where respondents had to indicate their age, gender and nationality; the number of members in their work teams; the time period they had been working in the team they had in mind while completing the questionnaire; how often they participate in geographically dispersed work teams; and whether they work for a private or public organization. Instructions were provided throughout the questionnaire indicating the requisite mode of responding and reminding at certain stages the lack of right or wrong answers.
3.5 Measures
All variables present in the model were formulated as to refer on team level procedures.
Virtual teamwork. A scale designed and validated by Chudoba et al. (2005) was used
in order to define the dependent variable “virtual teamwork” into a measurable item. The scale contained nine items which aimed at capturing the spatial and temporal flexibility of virtual teamwork, the potential different culture of team members, and the use of ICTs. Respondents were asked to indicate their agreement on a seven-point scale ranging from 1 = “strongly disagree” to 7 = “strongly agree” (see Appendix A). The phrase “me and my team members” was included in the beginning of each item, in order to make explicit reference to team level procedures. For instance, the item “Me and my team members work at different geographic locations” was one of the items capturing geographical flexibility; “Me and my team members collaborate across different time zones” refers to the measurement of time
24
flexibility; an illustrative item for the ICTs use was “I work with my team members via internet-based conferencing applications”; finally, the item “Me and my team members speak different native languages” was used to capture cultural differences among virtual team members. The principal component analysis (PCA) revealed that the nine items of virtual teamwork formed one component (eigenvalue 4.27), which explained 47.4% of the total variance. Therefore, a new variable for virtual teamwork was constructed (M= 3, SD= 1.41, Cronbach’s alpha = .85).
Types of workplace conflict. Three different scales were used for the
operationalization of each of the mediator variables: task, relationship, and process conflict on a group level. The scales were initially designed by Jehn (1995), and Shah and Jehn (1993); however, for the current research we used the scales as they were validated, revised and complemented by Behfar et al. (2011). All conflict types were measured on a seven-point scale ranging from 1= “not at all/none” to 7 = “always/totally” (see Appendix A). Task conflict was captured by three items aimed to depict diverging ideas and divergent thinking about the assigned tasks between team members, as well as the discussion of the positive and negative aspects of various alternatives and the consideration of merits pertained to
contradictory opinions about work tasks (Behfar et al., 2011). An illustrative example was “How frequently do members of your team engage in debate about different opinions or ideas?”. The principal component analysis revealed one component (eigenvalue 1.69), which explained 56.27% of the total variance. Based on this, a new variable for task conflict type was constructed (M= 4.5, SD= 1.1, Cronbach’s alpha = .60). Relationship conflict assessment consisted of four items, which indicated the overt and perceived emotional disagreement and interactions among team members, such as “How much friction is there among members of your team?”. The principal component analysis revealed that these four items form one component (eigenvalue 2.8), which explained 69.93% of the total variance; it also appeared to
25
form a reliable scale, Cronbach’s alpha = .86. Subsequently, a new variable for relationship conflict was constructed (M= 3.33, SD= 1.34). Finally, six items encapsulated the notion of process conflict as the extent to which teams debate about tasks’ procedural decisions (Behfar et al., 2011). “How frequently do your team members disagree about the optimal amount of time to spend on different parts of teamwork?” and “To what extent is there tension in your team caused by member(s) not completing their assignment(s) on time?” were two elucidative examples of the items that compose the process conflict type. The six items loaded on one component (eigenvalue 3.26), as revealed by the principal component analysis. This component explained 54.31% of the total variance and formed a reliable scale, Cronbach’s alpha = .82. Based on this, a new variable for process conflict was also constructed (M= 3.2,
SD= 1.17).
Solution-oriented communication strategy. The scale which was used for the
operationalization of the moderator variable, namely the solution-oriented communication strategy, constitutes a part of the Organizational Communication Conflict Instrument. This instrument was developed by Putnam and Wilson (1982), based on a five-category scheme about conflict types which was initially introduced by Blake and Mouton (1964; in Putnam & Wilson, 1982); the current scale was also evaluated by Wilson and Waltman (1988). The original items were formulated as to examine conflicts with the supervisor. In accordance to this study’s scope, the original items were reformulated as to refer to conflicts in work teams. It consisted of eleven items; respondents were asked to think of disagreements they have encountered with their fellow team members and then to indicate how frequently they engage in the behaviors described by the scale items in order to deal with these disagreements, on a seven-point scale ranging from 1= “never” to 7= “always” (see Appendix A). A few examples of this scale’s items are the following: “I integrate arguments into a new solution from the issues raised in a dispute with my team members”, “I will go 50-50 to reach a settlement with
26
my team members”, “I offer creative solutions in discussions of disagreements”. The scale aimed at capturing the extent that people utilize creative and integrative solutions to situations of conflicts and make compromises (Wilson & Waltman, 1988).The eleven items of the solution-oriented communication strategy loaded on two components. The first component (eigenvalue 4.45) explained 40.45% of the total variance and the second one (eigenvalue 1.58) 14.39% of the total variance. It was suggested the components to be combined, as it was used as a whole in other empirical studies (Putnam & Wilson, 1982; Wilson & Waltman, 1988). A new variable was created for the solution-oriented communication strategy (M= 4.9,
SD= 0.84, Cronbach’s alpha = .85).
Team performance. A four-item scale was used to measure the subjective perceptions
of team members about the level of quality regarding their assigned teamwork. The team performance scale was established by Lurey and Raisinghani (2000). Even though team members’ perception of effectiveness may deviate from company’s official standards or industry measures of quality control, team members’ responses about how well the team is performing can help the researcher to draw conclusive results (Lurey & Raisinghani, 2000). In particular, respondents were asked to indicate their (dis)agreement with the statements about team performance, on a seven-point scale ranging from 1= “strongly disagree” to 7= “strongly agree” (see Appendix A). The four items were formulated as follows: “In the past the team has been effective in reaching its goals”, “The team is currently meeting its business objectives”, “When the team completes its work, it is generally on time”, and “When the team completes its work, it is generally within the budget”. The principal component analysis revealed that these four items form one component (eigenvalue 2.6), which explained 65% of the total variance and formed a reliable scale, Cronbach’s alpha = .81. Therefore, a new variable for team performance was constructed (M= 5.41, SD= 1.1, missing cases = 2).
27 Control variable of task interdependence. Task interdependence designates the level of reliance between the members of a group for the successful implementation of their teamwork. This interdependence of different roles, of procedures, and of skills and resources amplifies the demand for smooth and efficient communication, cooperation and coordination among team members (Jehn, Northcraft, & Neale, 1999; Van der Vegt & Janssen, 2003; Jehn, 1995). If these demands are not met, conflict can occur within the group and its members. In highly interdependent work environments, the effects of task, relationship and process
conflict on team performance can be intensified or mitigated (Jehn, 1995). Consequently, it is important to control for task interdependence throughout the model.
A five-item scale sourced from Van der Vegt and Janssen (2003) and validated by the studies of Van der Vegt et al. (in Van der Vegt & Janssen, 2003) was the instrument for the operationalization of task interdependence. Respondents were asked to indicate the extent they (dis)agree with the five statements on a seven-point scale ranging from 1= “strongly disagree” to 7= “strongly agree” (see Appendix A). One indicative example of these items was as follows: “I need information and advice from my team members to perform my job well”. The principal component analysis showed that four of the five items formed one new component (eigenvalue 2.48), which explained 49.58% of the total variance. The item “I have a one-person job; it is not necessary for me to coordinate or cooperate with others” did not load on any component. Subsequently, it was excluded from the scale; the four remaining items formed a new variable for task interdependence (M= 5.35, SD= 1.15, Cronbach’s alpha = .78).
Control variable of team cohesion. Cohesion is perceived as the willingness of team
members to stay together and remain in the group, and it is linked to member satisfaction, innovation and increased performance (Ruga, 2014). Spatial and temporal mobility, in tandem with lack of face-to-face communication of teams might be negatively related to cohesion
28
between team members, meaning that the more the teams are distributed, the lower the level of cohesion among co-members is (Chudoba et al., 2005; Nataatmadja & Dyson, 2006). In addition, team cohesion can be a reason for increased or decreased personal or task-related arguments among team members (Teakleab, Quigley, & Tesluk, 2009; Chudoba et al., 2005; Weber & Kim, 2015). Therefore, the effects of team cohesion were controlled throughout the model.
The team cohesion scale was validated by Teakleab et al. (2009). It consisted of six items, such as “My team members communicate freely about each of our personal
responsibilities in getting a project done” and “We all take responsibility for any loss or poor performance by our team”. Respondents had to indicate their (dis)agreement with the six items on a seven-point scale ranging from 1= “strongly disagree” to 7= “strongly agree” (see Appendix A). The principal component analysis revealed that the six items loaded on one new component (eigenvalue 4.04), which explained 67.25% of the total variance. Based on this, a new variable for team cohesion was created (M= 5.01, SD= 1.26, Cronbach’s alpha = .90).
Control for interrelations between conflict types. Task, relationship, and process
conflict types can interrelate with each other and the presence of one type of conflict can change the effect that a different types of conflict has on team performance (Jehn & Chatman, 2000; Teakleab, Quigley, & Tesluk, 2009). For example, in case of examining the presence of task conflict in virtual teamwork and its effect on team performance without taking under consideration relationship conflict, there is a high chance of drawing “incorrect inferences”. That is because a team which scores high both on task and relationship conflict will be portrayed to perform equally to a team scoring high in task conflict and low in relationship conflict (Jehn & Chatman, 2000). According to Simons & Peterson (2000), teams that report task conflict, also report relationship conflict. Hence, correlations between conflict types were
29
examined, and the statistically significant correlated conflict types were also treated as control variables in the relevant hypotheses.
Results
4.1 Correlations between the control and the main variables of the overall model
The analysis starts by discussing whether there were statistically significant
correlations between the control variables, namely task interdependence and team cohesion, with virtual teamwork, team performance, and each of the tree conflict types. The control variable of team cohesion was proved to form a statistically significant and negative
correlation with relationship conflict, and the strength of the correlation was high, r = -.51, p < .001. The same control variable was also statistical significantly correlated to process conflict and the strength of the correlation was negative and strong, r = -.52, p < .001. Team cohesion was proved to have a statistically significant correlation with task conflict, but the strength of the relationship was weak, r = .23, p < .001. The final statistically significant and positive correlation of team cohesion was observed with team performance; the strength of the relationship was moderate, r = .44, p < .001. Task interdependence was proved to be statistical significantly correlated to team performance, but the strength of the relationship was weak, r = .27, p < .001. Therefore, only the effects of team cohesion were taken under consideration in every hypothesis testing that follows (see Appendix B, table 1).
In addition, as it was mentioned in the method section, the potential interrelations between conflict types might influence the effects of each of them on the dependent variable. The correlation analysis revealed that relationship conflict and process conflict indeed formed a positive and strong correlation, r = .65, p < .001. Hence, every time that relationship conflict consisted one of the main variables in the hypotheses, the effect of process conflict was controlled, and the opposite (see Appendix B, table 1).
30
4.2 Hypothesis Testing
H1: Higher involvement in virtual teamwork leads to higher levels of task conflict. A hierarchical multiple regression analysis was conducted on SPSS Statistics for testing the first hypothesis. The VIF values were below 10 and the tolerance was above 0.1 for both the independent and the control variable, which means that there was no
multicollinearity in the model. The regression model as a whole was found to be statistically significant predictor of task conflict, F (1, 171) = 9.27, p < .05, while the strength of the prediction was very weak (𝑅2= .06). However, the relationship between virtual teamwork and
tasks conflict, while including the effects of team cohesion, was not statistically significant,
b= -0.03, b* = -0.04, t = -0.51, p = .614, 95% CI [-0.14, 0.09]. The control variable of team
cohesion had a significant, but weak association with task conflict, b = 0.20, b* = 0.23, t = 3.04, p < .05, 95% CI [0.07, 0.33]. Subsequently, for every unit increase in team cohesion, task conflict level increases by 0.20. These statistical results reveal that virtual teamwork does not affect the level of task conflict and thus, the first hypothesis is rejected. Nonetheless, team cohesion has a positive effect on task conflict in the team, when the effect of virtual teamwork is held constant (see Appendix B, table 2).
H2: Higher involvement in virtual teamwork leads to lower levels of relationship conflict.
The VIF values were below 10 and the tolerance was above 0.1 for the independent and the control variables, thus, there was no multicollinearity in the model.The hierarchical multiple regression model as a whole was found to be a statistically significant predictor of relationship conflict, F (2, 170) = 71.24, p < .001, whereas the strength of the prediction was weak (𝑅2= .46). Nevertheless, the relationship between virtual teamwork and relationship
conflictwas not statistically significant, when the control variables of team cohesion and process conflict were included in the model, b = -0.01, b* = -0.01, t = -0.24, p = .812, 95% CI
31
[-0.12, 0.09]. Only the control variable of team cohesion had a statistically significant, negative, and moderate association with relationship conflict, b = -0.25, b* = -0.24, t = -3.57,
p <.001, 95% CI [-0.39, -0.11]. Therefore, for every unit increase in team cohesion,
relationship conflict level decreases by 0.25. In addition, process conflict was proved to have a statistically significant and positive association with relationship conflict, b = 0.60, b* = 0.52, t = 7.90, p < .001, 95% CI [0.45, 0.75]. That means for every unit increase in process conflict, relationship conflict increases by 0.60. These findings reveal that virtual teamwork does not affect the level of relationship conflict and thus, the second hypothesis is rejected. However, team cohesion has a negative effect on relationship conflict in the team, whereas process conflict has a positive effect on relationship conflict, when the effect of virtual teamwork is held constant (see Appendix B, table 3).
H3: Higher involvement in virtual teamwork leads to higher levels of process conflict. A hierarchical multiple regression analysis was conducted for testing the third
hypothesis. The VIF values were below 10 and the tolerance was above 0.1 for both virtual teamwork, team cohesion, and process conflict, which means that there was no
multicollinearity in the model. The regression model as a whole was found to be statistically significant predictor of process conflict, F (2, 170) = 73.67, p < .001, while the strength of the prediction was weak (𝑅2= .47). The relationship between virtual teamwork and process
conflict, was proved to be not statistically significantwhen the effects of team cohesion and relationship conflict were included in the model, b = 0.03, b* = 0.03, t = 0.59, p = .559, 95% CI [-0.07, 0.12]. Team cohesion had a significant, negative and weak association with process conflict, b = -0.24, b* = -0.26, t = -4, p < .001, 95% CI [-0.36, -0.12]. Accordingly, for every unit increase in team cohesion, process conflict level decreases by 0.24. Relationship conflict as a control variable was also proved to have a statistically significant effect on process conflict, b = 0.45, b* = 0.51, t = 7.90, p < .001, 95% CI [0.34, 0.56]. Thus, for every unit
32
increase in relationship conflict, process conflict level increases by 0.45. Similar to the previous hypotheses testing, we reject the third hypothesis, asvirtual teamwork does not affect the level of process conflict. However, team cohesion seems to have a negative effect on process conflict in the team, and relationship conflict seems to have a positive effect on process conflict, when the effect of virtual teamwork is held constant (see Appendix B, table 4).
H4: Task conflict type is negatively related to team performance, only when in low or high levels.
The hierarchical multiple regression analysis revealed that the VIF values were below 10 and the tolerance was above 0.1 for standardized task conflict, squared task conflict, and team cohesion, which means that there was no multicollinearity in the model. The regression model as a whole was found to be a statistically significant predictor of team performance, F (1, 168) = 37.29, p < .001, while the strength of the prediction was very weak (𝑅2= .20).
When squared task conflict as a predictor of team performance was examined, it was proved to have no effect on team performance, b = 0, b* = 0, t = 0.28, p = .978, 95% CI [-0.12, 0.13], while the effect of team cohesion was controlled; the (standardized) task conflict variable was also included in the model as an independent variable for the examination of the curvilinear effect. Only team cohesion had a significant, positive and moderate association with team performance b = 0.39, b* = 0.44, t = 6.11, p < .001, 95% CI [0.27, 0.52]. Thus, for every unit increase in team cohesion, team performance increases by 0.39. According to these findings, we reject the fourth hypothesis, astask conflict type does not affect team performance neither when they are in low, nor on high levels. However, team cohesion seems to have a positive effect on team performance, when the effects of task conflict and of the (standardized) task conflict variable are held constant (see Appendix B, table 5).
33
A hierarchical multiple regression analysis was conducted; it showed that the VIF values were below 10 and the tolerance was above 0.1 for relationship conflict, team cohesion and process conflict, which means that there was no multicollinearity in the model.The regression model as a whole was found to be statistically significant predictor of team performance, F (2, 168) = 12.59, p < .001, whereas the strength of the prediction was very weak (𝑅2= .20). However, the dependent variable of relationship conflict was proved to be
statistically insignificant in predicting team performance, b = -0.08, b* = -0.10, t = -1.04, p = .300, 95% CI [-0.23, 0.07], when the control variables of team cohesion and process conflict were present in the model. In regards to the control variable of team cohesion, it was proved to have a statistically significant and positive association with team performance, although moderate, b = 0.36, b* = 0.41, t = 4.89, p < .001, 95% CI [0.22, 0.51]. Thus, for every unit increase in team cohesion, team performance increases by 0.36. Process conflict as a control variable was not statistically significant as well, b = 0.02, b* = 0.02, t = 0.25, p = .806, 95% CI [-0.15, 0.20]. Therefore, this hypothesis is rejected as well, as we assume that relationship conflict does not influence team performance, when controlled for team cohesion and process conflict. Instead, team cohesion positively influences team performance (see Appendix B, table 6).
H6: The level of process conflict is negatively related to team performance.
The hierarchical multiple regression analysis revealed that the VIF values were below 10 and the tolerance was above 0.1 for process conflict, team cohesion and relationship conflict, which means that there was no multicollinearity in the model. The regression model as a whole was found to be statistically significant predictor of team performance, F (2, 168) = 14.94, p < .001, whereas the strength of the prediction was very weak (𝑅2= .20). However,
the independent variable of process conflict was proved to be statistically insignificant in predicting team performance, b = 0.02, b* = 0.02, t = 0.25, p = .806, 95% CI [-0.15, 0.20],
34
when the effects of team cohesion and relationship conflict were also included in the model. Relationship conflict -which was used as a control variable in this hypothesis- was not statistically significant, b = -0.08, b* = -0.10, t = -1.04, p = .300, 95% CI [-0.23, 0.07]. The control variable of team cohesion was found to form a statistically significant, positive, and moderate association with team performance, b = 0.36, b* = 0.41, t = 4.89, p < .001, 95% CI [0.22, 0.51]. Thus, for every unit increase in team cohesion, team performance increases by 0.36. This hypothesis is rejected, because process conflict does not influence team
performance, when controlled for team cohesion and relationship conflict. Instead, team cohesion positively influences team performance (see Appendix B, table 7).
H7: The relationship between virtual teamwork with team performance is mediated by H7a) task, H7b) relationship, and H7c) process conflict type, respectively.
Lambert’s mediation model in combination with Hayes’s PROCESS tool in SPSS were used for testing the mediation model. Three distinct PROCESS analyses were executed for each of the sub-hypotheses. In particular, task conflict1 was expected to mediate the relationship between virtual teamwork and team performance. The relevant results revealed that the overall model was statistically significant, F (3, 168)= 13.81, p < .001, whereas the strength of the prediction was very weak (𝑅2= .20). Team cohesion was included in the model as a covariate. Virtual teamwork was not a statistically significant predictor neither of task conflict, b= -0.03, SE= 0.06, p = .558, nor of team performance, b= -0.03, SE= 0.05, p =.573; in addition, task conflict was not a statistically significant predictor of team performance b= 0.02, SE= 0.07, p =.824. The indirect effect of virtual teamwork on team performance through task conflict was not statistically significant, b = 0.001, SE= 0.01, BCa CI [-0.02, 0.01]; hence, the mediational sub-hypothesis (H7a) is rejected. We can assume that the whole model
1 The fourth hypothesis (H4) revealed that the relationship between squared task conflict and team
performance was not statistically significant. Therefore, only the effect of task conflict is taken under consideration as a mediator in the model.
35
was statistically significant, because team cohesion was proved to play a significant role in it,
b = .39, SE= 0.06, p < .001 (see Appendix, table 8).
In regards to the sub-hypothesis that relationship conflict mediates the effect of virtual teamwork on team performance, the results indicated that the overall model was statistically significant, F (4, 167)= 6.52, p < .001, whereas the strength of the prediction was very weak (𝑅2= .20). The variables of team cohesion and process conflict were included in the model as
covariates. However, virtual teamwork was neither a statistically significant predictor of relationship conflict, b= -0.02, SE= 0.06, p =.762, nor a significant predictor of team performance, b= -0.03, SE= 0.06, p =.609; in addition, relationship conflict was not a statistically significant predictor of team performance b= -0.08, SE= 0.11, p =.444. The indirect effect of virtual teamwork on team performance through relationship conflict was not statistically significant, b = 0.001, SE= 0.01, BCa CI [-0.01, 0.04]; therefore, this mediational sub-hypothesis (H7b) is not supported. We assume that the whole model was statistically significant, because team cohesion was proved to play a significant role in it, b = 0.36, SE= 0.09, p < .001 (see Appendix, table 8).
According to the third mediational sub-hypothesis, process conflict was expected to mediate the effect of virtual teamwork on team performance. The analyses revealed that the overall model was statistically significant, F (4, 167)= 6.52, p < .001, whereas the strength of the prediction was very weak (𝑅2= .20). The variables of team cohesion and relationship
conflict were included in the model as covariates. Nonetheless, virtual teamwork was neither a statistically significant predictor of process conflict, b= 0.03, SE= 0.05, p =.508, nor a significant predictor of team performance, b= -0.03, SE= 0.06, p =.609; in addition, process conflict was not a statistically significant predictor of team performance b= 0.02, SE= 0.09, p =.791. The indirect effect of virtual teamwork on team performance through process conflict was not statistically significant, b = 0.001, SE= 0.01, BCa CI [-0.01, 0.03]; Therefore, this
36
mediational sub-hypothesis (H7c) is also rejected. We assume that the whole model was statistically significant, because team cohesion was proved to play a significant role in it, b = 0.36, SE= 0.09, p < .001 (see Appendix, table 8).
H8: The solution-oriented communication strategy mitigates the negative effects of H8a) task, H8b) relationship, and H8c) process conflict type on team performance.
Hayes’s PROCESS tool was utilized for testing each of the sub-hypotheses. In regards to sub-hypothesis 8a), the variable of task conflict was included in the model, since squared task conflict did not form a statistically significant relationship with team performance (see results of H4). Team cohesion was also included in the model as a covariate. The overall model was found to be statistically significant, F (4, 167)= 9.56, p < .001, but the strength of the prediction was very weak (𝑅2= .26). The interaction effect on team performance between
task conflict and the moderator variable, namely the solution-oriented strategy, was not statistically significant, b = 0.09, SE= 0.14, 95% CI [-0.20, 0.38], t = 0.65, p = .518. Hence, this sub-hypothesis (H8a) is rejected. We assume that the whole model was statistically significant, because team cohesion was proved to play a significant role in it, b = 0.32, SE= 0.10, 95% CI [0.12, 0.52], t = 3.10, p < .01. The solution-oriented communication strategy was also a significant predictor of team performance, b = 0.34, SE= 0.15, 95% CI [0.05, 0.63],
t = 2.27, p < .05, when the effects of team cohesion and task conflict were held constant (see
Appendix B, table 9).
For the sub-hypothesis 8b), the overall model was found to be statistically significant,
F (5, 166)= 11.56, p < .001, but the strength of the prediction was very weak (𝑅2= .27). Team
cohesion and process conflict type were also included in the model as covariates. The interaction effect between relationship conflict and the solution-oriented strategy, controlled for the variables of process conflict and team cohesion, was proved to be not statistically significant, b = -0.08, SE= 0.11, 95% CI [-0.31, 0.14], t = -0.76, p = .451. Therefore, this
