Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Psychology. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 05 December 2023

Virtual Teams and Digital Collaborationfree

Virtual Teams and Digital Collaborationfree

  • Conny H. AntoniConny H. AntoniTrier University


Collaborating in teams by using various digital information and communication technologies (ICTs) to perform interdependent tasks and achieve common goals relevant for one’s organization is increasingly the new normal. Such more or less virtual teams—which can be all human or human-agent teams (HATs) (i.e., including autonomous software agents with artificial intelligence)—are complex dynamic open socio-digital systems embedded in an organizational, economical, and societal context. How and to what degree team members use ICTs to perform their tasks and to manage situational demands influence team processes and emergent states, such as transactive memory systems and team mental models, and thus team effectiveness. Research on input-mediator-output-input models of teamwork has shown that these processes are reciprocal, influencing team development over time. Research on virtual team effectiveness shows negative effects of virtual teams on team functioning and effectiveness primarily when short-term laboratory teams are studied, whereas no or lower effects were found for long-term organizational teams. These results have practical and theoretical implications, such as to support the launch of virtual teams by team-building interventions and trainings and to prefer longitudinal and field studies to examine processes and outcomes of virtual human teams as well as HATs.


  • Organizational and Institutional Psychology
  • Social Psychology

Relevance of Virtual Teams and Digital Collaboration

Modern information and communication technologies (ICTs), such as email, chat, video conferencing, augmented and virtual reality, and collaboration software with shared networked databases, enable synchronous and asynchronous communication and information access in real time and digital collaboration within and across locations, countries, and company boundaries. Members of virtual teams can be geographically dispersed but still coordinate with other team members and quickly access the same information. Especially for organizations operating in a global market, it has been crucial to take advantage of the opportunities of virtual teams. Specialists for complex and customer-specific project tasks can be recruited across the world more easily and contribute their knowledge to different teams working in parallel as needed. Thus, digital collaboration promises companies organizational flexibility and acceleration of work processes. Furthermore, employees may benefit from more flexible work arrangements granting more autonomy to decide where and when to work, reducing commuting time and costs when working from home. Already in the 1990ies, virtual teams were announced as the new way to work and the peopleware of the 21st century (Lipnack & Stamps, 1997). Although many knowledge workers had already been collaborating digitally (at least to some extent) for a long time, the beginning of the covid pandemic in 2020 was a game changer. Almost from one day to the next, offices were shut down and people worked from home. Collaborating digitally became the new normal. Companies and employees that had no experience with virtual teams realized the chances and risks of virtual teamwork and digital collaboration. After the forced experience, companies and employees seem more willing to continue virtual teamwork than before. In the future, digital collaboration will increasingly encompass software agents, which can act and decide autonomously based on artificial intelligence (O’Neill et al., 2022). Only time will tell if the vision of human team members and software agents meeting in the metaverse via avatars will come true.

The focus of this article is on virtual teams and digital collaboration in teams. Virtual team research deals with phenomena and questions on the team level, such as how working with collaborative ICT affects the emergence of trust between members or influences the relationship between team processes and outcomes. In contrast, telework and telecommuting research focuses on the individual level of analysis, such as on the relationship between individual characteristics and individual adjustment, well-being, and performance comparing office work with non-office contexts using individual and mobile technologies (Raghuram et al., 2019). The structure of the article is as follows. First, I define the concept of virtual teams and team virtuality. Then, I describe the factors influencing the processes and the development of emergent states and outcomes of virtual teams using an input-mediator-output-input (IMOI) model of team effectiveness before deriving practical implications and future research perspectives.

Defining and Measuring Team Virtuality

Lipnack and Stamps (1997) define a virtual team, like any team, as “a group of people who interact through interdependent tasks guided by common purpose” but which “works across space, time, and organizational boundaries with links strengthened by webs of communication technologies” (p. 7). The authors combine aspects of ICT-mediated communication and distributed work with core aspects of a definition of teams in general. However, the mentioned aspects of teams in general might also apply for people working interdependently along a process chain in different departments or even companies. Hence, this definition might fit for digital collaborative work better than for virtual teams. To differentiate teamwork from collaborative work in general core aspects are missing, such that teams have specific common goals linked to organizational goals and members have different roles and responsibilities for attaining them (Kozlowski & Ilgen, 2006). Teams have boundaries and linkages to the broader system context and task environment and are—apart from cross-organizational project teams—embedded in an encompassing organizational system.

With respect to team virtuality, Schulze and Krumm (2017) point out that at least four different types of definitions of virtual teams exist in the literature: (a) definitions treating virtual and face-to-face teams as a dichotomy; (b) definitions based on a single dimension of team virtuality, ranging from low (merely face-to-face) to high (entirely virtual) virtuality; (c) definitions with multiple dimensions of virtuality; (d) definitions which emphasize how individuals perceive and react to discontinuities in which routine behaviors do not produce expected effects (e.g., Watson-Manheim et al., 2012).

Approaches using multiple dimensions differ in terms of facets and measurement. For example, Chudoba and colleagues (2005) consider the aspects of geographic locations, time zones, cultures, work practices, organizations, and ICTs as potential discontinuities. They asked team members to rate these aspects (e.g., how often team members experienced working with people at different sites). They used a factor analysis to develop a three-dimensional measure of team virtuality encompassing team distribution, workplace mobility, and variety of work practices. Others, such as Gibson and Gibbs (2006), use objective indicators to measure team distribution and combine them with external or team members’ ratings on electronic dependence, dynamic structure, national diversity, and psychological safety. Criteria such as dynamic structure, national diversity, and psychological safety apply also to merely face-to-face teams. They can be used to differentiate between teams in general but appear less suited to define team virtuality. Therefore, most definitions of team virtuality focus on geographic dispersion and technology use (Gilson et al., 2015).

Geographic dispersion implies that team members cannot interact face-to-face besides when meeting in person or using videoconferencing tools. However, digital collaboration can also take place to varying degrees when team members are co-located. Focusing on ICT use and characteristics might therefore tap the core of team virtuality. For example, Kirkman and Mathieu (2005) define team virtuality as the extent to which team members use virtual tools to coordinate and execute team processes, the amount of informational value provided by such tools, and the synchronicity of team members’ virtual interaction. By integrating informational value and media synchronicity as criteria for team virtuality, they build on media richness theory (Daft & Lengel, 1984, 1986 and media synchronicity theory (Dennis et al., 2008), which explain how ICTs influence information and communication processes in organizations.

Handke and colleagues (2021) introduce the concept of team perceived virtuality. They describe team perceived virtuality as an emergent state influenced by structural team virtuality and other team characteristics and team processes, which in turn influence team outcomes. Team perceived virtuality consists of two dimensions: collectively experienced distance and collectively experienced information deficits. However, by focusing on these two dimensions, they adopt only a deficit-oriented view, implying that increasing team virtuality means more distance and information deficits, leaving out any positive effects, such as increased flexibility.

Furthermore, the above-described definitions of virtual teams focus implicitly on human teams. However, the rapid development of artificial intelligence allows researchers and developers to design software agents with higher levels of agent autonomy and to integrate them in digital collaboration and virtual teams. O’Neill and colleagues (2022, p. 911) define human-autonomy teams or human-agent teams (HATs):

as interdependence in activity and outcomes involving one or more humans and one or more autonomous agents, wherein each human and autonomous agent is recognized as a unique team member occupying a distinct role on the team, and in which the members strive to achieve a common goal as a collective.

To be considered autonomous, the agents must meet at least partial levels of self-directed behavior (agency) and agent autonomy. The concept of levels of automation was suggested by Parasuraman and colleagues (2000) and ranges from low (level 1: agent offers no assistance) to high (level 10: agent decides everything). Partial agent autonomy means that the computer can at least suggest and execute actions and decisions if they are approved by the human (level 5) or do not veto actions and decisions before they are automatically executed (level 6). Whether one perceives an autonomous software agent as a virtual teammate or collaborator or as a machine with artificial intelligence performing collaborative tasks might also depend on the similarity of perceived agent personality characteristics to the individual (Dryer, 1999). Boundaries might get more fluid in the future when humans and software agents interact via avatars in the metaverse and when software agents can recognize emotions and react accordingly (Lee et al., 2021).

Integrating research on HATs and following Kozlowski and Ilgen’s (2006) definition of teams, I define virtual teams as two or more human or autonomous software agents who (a) collaborate to perform organizationally relevant tasks; (b) interact at least to a certain degree virtually to achieve one or more common goals; (c) are interdependent with respect to workflow, goals, and outcomes; (d) have different roles and responsibilities; and (e) are embedded together in an encompassing organizational system or cross-organizational agreement with boundaries and linkages to the broader system context and task environment. This definition implies that virtuality can have varying degrees and that different types of virtual interaction are used. Virtual teams with human members can vary the degree and type of virtual interaction over time depending on task and situational demands as well as human member needs. They can interact virtually exclusively or alternate virtual and face-to-face interaction or combine the two interactions. Most of the existing virtual teams have some face-to-face contact (Hertel et al., 2005). Partially virtual teams, which alternate virtual and face-to-face interaction or combine them, with some team members interacting face-to-face and others virtually, are also called hybrid teams (Hosseini et al., 2017). Virtual, digital, or e-collaboration is collaboration among individuals or autonomous software agents to accomplish common tasks using electronic technologies, which could happen even without interacting or communicating (Kock, 2005), for example, by telework.

The facets used to define virtual teams are also used to define virtual organizations, where virtual teams are embedded and members collaborate digitally. For example, Ahuja and Carley (1999, p. 742) define virtual organizations “as a geographically distributed organization whose members are bound by a long-term common interest or goal, and who communicate and coordinate their work through information technology.”

An IMOI Model of Virtual Team Effectiveness

To summarize virtual team research, I adopt an IMOI model (Ilgen et al., 2005; Kozlowski & Ilgen, 2006). In comparison with an input-process-output model (Hackman, 1987), an IMOI model emphasizes that virtual teams are complex dynamic open socio-technical or socio-digital systems embedded in an organizational, economical, and societal context (Antoni & Hertel, 2009). Employees are organized in teams within or between organizations to contribute to organizational goals by achieving common team tasks and goals with the help of ICTs. Virtual team members (humans and autonomous software agents) learn and develop through the execution of their tasks and the use of ICTs. They have to adapt to changing internal and external situational demands, but their actions and performance in turn also contribute to organizational and situational changes. Important factors influencing team outcomes via team processes and the development of emergent states are ICTs and team tasks, team member characteristics, and situational demands as input factors. The IMOI model (see figure 1) describes the recursive and cyclical processes between inputs, mediators, and outcomes.

Figure 1. An input-mediator-output-input model of virtual team effectiveness.

Information and Communication Technologies

ICTs are important elements of the technical system of virtual teams as team members collaborate using ICTs. The characteristics of ICTs influence task and situational demands, team processes as well as the development and outcomes of virtual teams. But as virtual teams learn and develop, also the characteristics of ICTs can change. Key ICT characteristics or functionalities discussed in the literature are the degree of social presence, media richness, naturalness, and synchronicity.

Social presence means the degree to which one is aware of the other person in an interaction. Social presence theory proposes that media differ in their degree of social presence and that communication effectiveness depends on the task-media-fit. Face-to-face interaction is considered to have the most social presence, and text-based communication the least. For effective communication, the degree of social presence should have the level of interpersonal involvement required for a task (Short et al., 1976).

Similarly, media richness theory states that media can be ranked according to their information richness. In their seminal article, Daft and Lengel (1986, p. 560) defined information richness “as the ability of information to change understanding within a time interval.” They proposed that communication media differ in their capacity to transmit rich information because of the number of channels they use, the number of cues they can transmit in a given time, their capability for immediate feedback, personalization, and language variety. Using these criteria, they presented a hierarchical classification of media in order of decreasing richness: (a) face-to-face, (b) telephone, (c) personal documents such as letters or memos, (d) impersonal written documents, and (e) numeric documents (Daft & Lengel, 1984, 1986). Daft and Lengel argue that when working under bounded rationality and time constraints, organizations should use rich media to reduce equivocality and quickly clarify ambiguous issues to reach a common understanding of a situation. Videoconferences, telephones, and chat software were not available that time and could probably be ranked second, third, and fifth according to this logic. Kirkman and Mathieu (2005) suggest that subjective ratings of the amount of informational value might be better able to assess media richness than objective ranking as proposed by Daft and Lengel (1984). Research supports that people perceive ICT functionalities differently and that intraindividual differences and differences between situational contexts occur over time and can influence performance (e.g., Carlson & Zmud, 1999; Fuller & Dennis, 2009; Hantula et al., 2011).

Why do team members perceive ICT functionalities differently and change their perception? Channel expansion theory suggests that the knowledge base of ICT users and their ability to communicate with an ICT develop with experience and change the perception of media richness (Carlson & Zmud, 1999). Particularly, the experience of using a medium, the experience with communicating partners, and perceived social influence change the perception of media richness not only between and within users but also between situational contexts (Carlson & Zmud, 1999; D’Urso & Rains, 2008; Hollingshead et al., 1993).

Similarly, adaptive structuration theory argues that people actively select how they appropriate and use the features of a technology and that this appropriation may change over time (DeSanctis & Poole, 1994). In line with this reasoning that ICT use is a function less of physical properties than of media appropriation and perception, Handke and colleagues (2018) report findings that individuals change their perception of ICT richness over time and toward communication partners. Fuller and Dennis (2009) showed that while task-media/technology-fit predicted team performance when teams used a technology the first time, over time teams with an initial fit did not perform better than teams with an initial poor fit. In line with adaptive structuration theory, teams with an initial poor media/technology-fit innovated and adapted their ICT use and improved their performance. This indicates that it is not ICTs that determine the processes and performance of virtual teams but rather that it is a recursive relationship in which the two influence each other (DeSanctis & Poole, 1994).

Although there is no technological determinism, ICTs influence media use, team processes, and outcomes. Media naturalness theory argues that people prefer media similar to face-to-face interactions because they require less cognitive effort for knowledge transfer because of the evolution of our brain (Kock, 2004). Kock suggests that the degree of similarity or naturalness of media can be assessed by the co-location of people communicating, the synchronicity to exchange information quickly, the ability to convey and observe facial and body expressions, and, in particular, the ability to convey and listen to speech. Kock (2004) proposes that cognitive adaption to ICT and the degree of schema alignment between team members decrease the cognitive effort required.

The synchronicity of information exchange is the focus of media synchronicity theory (Dennis & Valacich, 1999; Dennis et al., 2008). It differentiates conveyance and convergence of meaning in communication processes. Conveyance refers to the transmission and exchange of information between team members; convergence refers to the information processes needed for sense making to achieve a common understanding of the information received. Both processes are regarded as necessary for understanding tasks but differ in terms of the media synchronicity required. Media synchronicity can be defined as the extent to which media capabilities enable individuals to work together on the same activity at the same time and to have a shared focus (Dennis & Valacich, 1999; Dennis et al., 2008). Media synchronicity theory proposes that a higher degree of media synchronicity is required for convergence than for conveyance processes. Consequently, it is proposed that the use of higher synchronic media will lead to better communication performance if team members want to achieve common understanding and, vice versa, that lower synchronic media are supposed to lead to better performance if teams want to convey information. Media synchronicity is determined by media capabilities to support information transmission and processing.

Dennis and colleagues (2008) propose that media synchronicity is supported by (a) the speed of media to transmit information (transmission velocity/channel capacity), (b) media with more natural symbol sets, allowing fast encoding and decoding, and (c) symbol sets better suited to the content of a message (e.g., vocal tone to show doubt); they propose that media synchronicity is impaired by (a) a higher extent to which signals from multiple senders can be transmitted simultaneously using a medium (parallelism), (b) messages that can be rehearsed before sending (higher media rehearsability), and (c) messages that can be reexamined during decoding (higher reprocessability) as shared focus is lowered.

Team Tasks, Team Members, and Situational Demands

Key input factors for virtual teams that influence teamwork interacting with ICTs are team tasks, team members, and situational demands. Virtual teams are formed to fulfill tasks relevant for an organization with the help of ICTs that cannot be effectively accomplished by individuals alone. Team tasks influence workflow and coordination demands and thus the structure and design of virtual teams. Team tasks define the minimum requirements that team members must meet in terms of knowledge, skills, attitudes, and other characteristics (KSAOs), such as dispositions or personality characteristics. Tasks are performed under certain situational conditions that pose specific requirements for team members. Moreover, situational conditions might change, and virtual teams have to cope with the resulting situational demands.

Team Tasks

Which types of tasks are suited for virtual teams? Obviously, tasks requiring information processing and specialist knowledge, such as in research and development, are better suited for virtual teams than tasks requiring manual work. Furthermore, if these tasks are separable into subtasks, coordination requirements are reduced. However, subtasks have to be interdependent to some degree; otherwise, only digital collaboration but no virtual teamwork would be needed. Typically, project tasks that are unique and time-limited and require the integration of specialized knowledge for planning and problem solving are well suited for virtual project teams, particularly if the specialists needed are distributed locally (Hertel et al., 2005). Although the time frame for a project task can vary considerably from months to years, all project teams have, by the very definition of a project, a fixed deadline to accomplish the task after which the teams are usually dissolved. Aside from purely project-based organizations, project team members continue to work in their functional job while working part-time in one or more project teams. Hence, project tasks usually imply multiple team membership (Bell & Kozlowski, 2002; Margolis, 2020) in several more or less virtual project and functional teams. Larger projects are often organized as multi-team systems with interdependent project teams working together to achieve a common goal (Shuffler & Carter, 2018; Zaccaro et al., 2020). However, multi-team systems to address larger problem-solving tasks are not restricted to projects. In particular, international companies use virtual multi-team systems to coordinate distributed functional teams and to develop, adapt, and implement global company strategies.

Team Members’ KSAOs

Team tasks determine what KSAOs are required from team members, and they are selected or have to be trained accordingly. As virtual teams perform their tasks using ICTs, team members require, besides task and team, ICT-specific KSAOs. Schulze and Krumm (2017) provide a review of the literature on KSAO requirements of team virtuality. They found that successful performance in virtual teams requires media-specific KSAOs; that is, team members should know about the functionalities of the media and use their potential, how and when to use a certain medium for communication and knowledge transfer, and be able to adapt to channel restrictions of certain media (e.g., when using emails as opposed to videoconferences). Besides having media-specific KSAOs, virtual team members should be able to communicate effectively with distributed team members (communication KSAOs), to act in a way that creates trust (e.g., by being responsive and dependable) and to be willing to trust others (trust-related KSAOs), to be able to work with people from different cultural backgrounds (intercultural KSAOs), to manage oneself effectively (e.g., self-, time-, and project-management KSAOs), and to handle conflicts constructively (conflict management KSAOs). These media-unspecific competences are particularly required when the available media are inappropriate for the team tasks or when team members work in different time zones and cultures.

Situational Demands

As virtual teams are complex dynamic open socio-digital systems embedded in an organizational, economical, and societal context, they have to learn to deal with changing situational demands and adversities. Whereas changes can be perceived as being positive, neutral, or negative (such as additional team tasks and responsibilities) and may require adaptation of task strategies and team processes, adversities have a negative connotation and are associated with disruption, stress, and failure (Hartwig et al., 2020). Changes and adversities can originate from within the team or from outside. Internal adversities, such as the failure or the loss of a team member, or external adversities, such as a conflict with another team or a customer, require virtual teams to be resilient to deal with the adversities (Raetze et al., 2021). This may or may not require virtual teams to adapt their task strategy and team processes.

Team Processes

In the IMOI model, team processes are mediators, which mediate the effects of team inputs on team outcomes. Team processes describe how team members interact and exchange information with or without ICTs to coordinate and monitor their taskwork. Taskwork describes what tasks teams are doing, their interaction with tasks, ICTs, and other technical tools and systems.

Team processes can be differentiated in transition, action, and interpersonal processes. Transition processes encompass team mission analysis, formulation and planning, goal specification, and strategy formulation. Team action processes are the monitoring of goal progress, systems, and team monitoring and coordination. Interpersonal processes include conflict and affect management as well as team motivation and confidence building (Marks et al., 2001).

Team processes are dynamic. Over time, team interaction and information exchange, with or without using ICTs, shape team emergent cognitive and affective states, which in turn influence team processes as well as team outcomes, which reciprocally influence team processes and emergent states (Kozlowski & Ilgen, 2006). DeChurch and Mesmer-Magnus (2010) provided support for this conception. Results of their meta-analysis of 65 independent studies show that both emergent cognitive and affective states were strongly associated with performance. Team cognition was significantly related to team affective states, such as team cohesion, and team action and transition processes, as well as to team performance. Team cognition, team cohesion, and team processes predicted team performance, and team cognition explained unique variance in team performance after controlling for team cohesion and team processes.

ICTs do influence team processes, as suggested by social presence (Short et al., 1976) and media richness (Daft & Lengel, 1986) theories. However, there is no technological determinism. As described above, people learn to appropriate new media by using them in virtual team interaction (Fuller & Dennis, 2009), and this experience can change ICT characteristics by, for example, developing and learning to use emoticons in e-mails and instant messaging (Carlson & Zmud, 1999).

Team Learning and Adaptation

Team learning is both a team process and an outcome (Edmondson et al., 2007; Kozlowski & Bell, 2008). As a process, team learning can be defined as the interaction behaviors of team members to acquire, share, refine, or combine knowledge relevant to the team and the task. Team learning behaviors include the reflection of processes and outcomes, seeking feedback and information from outside the team, and storing and retrieving generated knowledge, encompassing transition, action, and interpersonal processes. As an outcome, team learning describes the acquisition of new knowledge, skills, and attitudes by team learning behaviors, which broadens the action repertoire of a team. Team learning is crucial for the appropriation of ICT functionalities in virtual teams and for team adaptation to ICT, team, and situation requirements (Fuller & Dennis, 2009; Handke et al., 2018). Team learning and team adaptation are closely intertwined as team learning can be regarded as both an antecedent and an outcome of team adaptation.

Burke and colleagues (2006, p. 1190) define team adaptation “as a change in team performance, in response to a salient cue or cue stream, that leads to a functional outcome for the entire team.”

Following this definition, team adaptation can be observed as it is manifested in the innovation or modification of actions or existing structures, while team learning can be latent as a change of new knowledge, skills, and attitudes if it is not implemented in actions. Thus, team learning is a necessary but insufficient condition for team adaptation. When teams adapt (e.g., to ICTs and change their ICT use), they may acquire new knowledge, skills, and attitudes because of team adaptation. Oertel and Antoni (2014) showed that interruptive events can trigger team adaptation via reflective team learning. However, depending on the phase of team development, different team learning behaviors may be relevant for changing team knowledge structures. Research indicates that knowledge-based processes (storage and retrieval) play a more important role during early stages of project-based teamwork, followed by a shift to a higher relevance of communication-based processes (reflection and co-construction) in later stages (Oertel & Antoni, 2015).

Team Leadership

Bell and Kozlowski (2002) argue that team virtuality (i.e., spatial distance and ICT-mediated communication) impedes two primary team leader functions (i.e., team performance management and team development). They recommend delegating these functions to the team and implementing team self-regulation processes using team goal and feedback systems (Kozlowski & Ilgen, 2006). Feedback is important not only regarding key performance indicators to coordinate taskwork but also regarding social processes to support team coordination and team cohesion, motivation, and development (Hertel et al., 2005). For example, Ellwart and colleagues (2015) showed that individual online feedback about individual information overload perceptions and task knowledge followed by collective team reflection increased situation awareness and supported virtual team adaptation processes. Self-regulating teams also require shared leadership. Several studies support the positive relation between shared leadership and virtual team performance (Hoch & Kozlowski, 2014). However, studies report inconsistent results regarding the effects of transformational leadership and virtual team performance. Some studies report negative effects (Gilson et al., 2015; Hoch & Kozlowski, 2014), others positive effects (Avolio et al., 2014; Kahai et al., 2013; Purvanova & Bono, 2009).

The Development of Team Emergent States

Closely related to team processes is the development or emergence of cognitive and affective states, such as trust, transactive memory systems (TMSs), and shared mental models (SMMs), as well as team motivation and affective states, such as team cohesion. Team emergent states and team processes interact and mediate the effects of team inputs on team outcomes.

When a virtual team starts from scratch, team members do not know each other or what is expected from them as a team and from each team member. Team members have to get to know each other. They have to agree on their team tasks and goals as well as on their task strategies and roles before they are able to coordinate and perform effectively. Most authors agree that it is more difficult to get started as a team, to learn each other’s expectations and KSAOs, and to develop team trust only by using ICT-mediated communication. Therefore, they recommend kick-off workshops face-to-face to get to know each other, to clarify tasks and goals as well as team member roles and functions, and to develop rules for virtual teamwork and mutual trust (Hertel et al., 2005). Particularly, the development of trust seems to be both a challenge and a requirement for virtual teams.

Team Trust

Team trust can be defined as an emergent shared willingness of team members to be vulnerable to the actions of the other team members, which are important to the team, irrespective of the ability to monitor or control them (Breuer et al., 2016). A meta-analysis of 48 field and six laboratory studies using cross-sectional (34 studies) and longitudinal (24 studies) data shows that the relationship between team trust and team task performance was stronger in virtual than in face-to-face teams (Breuer et al., 2016). Team trust was also significantly related to information processing (knowledge sharing and team learning), contextual performance (e.g., showing extra effort, volunteering, helping and cooperating with others, following organizational rules, and defending organizational goals), and team attitudes, such as team cohesion, satisfaction, commitment, and effort. However, the studies were too few and the sample size was too low to test whether virtual teams differ from face-to-face teams with respect to these variables. Also, cross-sectional studies, same source data, and subjective ratings showed stronger (and significant) relationships compared with studies using longitudinal data, different data sources, and objective data. Nevertheless, these results are in line with the authors' reasoning that when team members trust each other, they are more likely to take the risk to share their knowledge thus supporting team effectiveness via team coordination and cooperation. As practical implications, the authors recommend, besides trust-building activities, documenting team interactions particularly in virtual teams because their results indicate that the need for trust in virtual teams decreases when interactions in teams are documented. They assume that documenting team interactions facilitates peer monitoring and allows reviewing and verifying team agreements and decisions and thus reduces the risk that individual team members think their efforts are exploited by others.

Transactive Memory Systems

Virtual teams are often implemented to solve project tasks, which require the knowledge and collaboration of different specialists. The concept of TMS is highly relevant for explaining team coordination and performance of this type of teams because it explains how teams can use the individual memories and the distributed specialized knowledge of their team members efficiently. The TMS concept combines two components: (a) a transactive memory that connects the knowledge held by each team member to the knowledge held by the others and (b) knowledge-relevant transactive communication processes that occur among group members (Wegner, 1987). In order to encode, store, and retrieve distributed specialist knowledge, teams need a shared transactive memory or metaknowledge of expertise location (knowing who knows what). Metaknowledge is defined as the shared perception of expertise location (i.e., the shared knowledge of the expertise and knowledge domains of the other team members). Research on TMS shows that individual team members can function as distributed knowledge repositories, each specializing in particular areas of knowledge and expertise to extend the knowledge capacity of teams and to improve team performance by the cognitive division of labor (Ren & Argote, 2011). Many studies showed that memory specialization, task credibility, and task coordination in teams improved team coordination and performance (Lewis, 2003, 2004; Liang et al., 1995).

Shared Mental Models

While the TMS concept focuses on how teams can profit from team members with specialized and distributed knowledge, the concept of SMMs explains how shared cognitions about work-relevant aspects allow implicit coordination and enhance task performance (Cannon-Bowers et al., 1993; DeChurch & Mesmer-Magnus, 2010). SMMs emerge in a dynamic process of convergence and divergence of individual mental models in team interaction and communication processes and manifest as emergent states in teams as a higher-level, collective phenomenon (Kozlowski & Klein, 2000). Whether teams collaborate face-to-face or via ICTs influences and impedes the development of SMMs (Andres, 2011, 2013).

SMMs can be described in terms of similarity and accuracy. Similarity measures the extent of match between individual team members’ mental models. Accuracy describes the extent to which the team members’ SMMs correspond to standards, rules, and expert assessments (Mohammed & Dumville, 2001).

Cannon-Bowers and colleagues (1993) initially differentiated four types of mental models: the equipment, task, team interaction, and team mental model. They suggested that particularly task (e.g., task procedures), team interaction (e.g., roles and role interdependencies), and team mental model (e.g., teammates’ abilities and preferences) across tasks and situations should be shared in teams.

While the task and interaction mental model were considered as moderately dynamic and the team model as highly dynamic, the equipment mental model was regarded as highly stable across tasks and situations. Cannon-Bowers and colleagues (1993) assumed that equipment mental models (the knowledge of equipment functioning and limitations, operating procedures, and likely failures) focus on individual taskwork and do not need to be shared. Although later research integrated task and technology/equipment aspects in the concept of task mental models and team and team interaction aspects in team mental models (Mathieu et al., 2000), equipment aspects were neglected in research for a long time.

In the face of the multitude of collaborative ICTs that support the interaction and coordination of virtual team members, researchers called for considering ICT SMMs (Schmidtke & Cummings, 2017). Müller and Antoni (2020, 2022) showed that a shared understanding of ICT functionalities, task-specific ICT use, ICT adaptation, and ICT netiquette within a team had an impact on virtual team coordination and performance via team communication. Results also indicated that information about the advantages and disadvantages of ICTs can influence ICT mental models and that explicit planning of ICT use contributes to ICT SMM similarity (Müller & Antoni, 2022).

Furthermore, it has been shown that a temporal SMM, defined as shared knowledge about deadlines, pacing, and sequencing of tasks, contributes to team coordination and performance (Gevers et al., 2006; Mohammed et al., 2015).

Team Motivation and Team Affective States

Several studies showed that team motivation and team affective states, such as team cohesion, can be impaired if team members interact only virtually, because feelings of anonymity and low social control might support social loafing. On the other hand, team motivation and cohesion are considered crucial for virtual team functioning and performance (Hertel et al., 2005). Study findings indicate that positive team affective tone is positively and negative affective tone is negatively related to team cooperation and indirectly to team performance (Lin et al., 2017).

Team Outcomes

Team outcomes are often conceptualized as team effectiveness encompassing multiple dimensions and perspectives, such as team performance as evaluated by different stakeholders, team satisfaction (i.e., the satisfaction of team members with their team), and team viability (i.e., their willingness to continue work together as a team).

Impact of Team Virtuality on Team Effectiveness

In their meta-analysis of 428 samples from 398 studies on the relationship between team design characteristics and team performance, Carter and colleagues (2019) found only a very small negative relationship (r = −0.05) between team performance and dispersed and virtual teams. This finding does not support the results of prior studies and predictions of social presence theory (Short et al., 1976) and media richness theory (Daft & Lengel, 1984) described above. For example, Baltes and colleagues (2002) had reported results of their meta-analysis of 27 studies and 52 effect sizes that computer-mediated communication decreased group effectiveness and member satisfaction and increased time required for task completion compared with face-to-face teams. Lim and colleagues (2007) reported in their meta-analysis of 33 laboratory studies with 62 data points that virtual teams needed more time to reach a decision but achieved higher decision quality, but the authors did not find differences in terms of decision satisfaction between virtual and face-to face teams.

Ortiz de Guinea and colleagues (2012) analyzed 80 data sets from 79 studies and reported negative effects of virtual teams compared with face-to-face teams on team functioning: virtual teams had more task conflicts, lower communication frequency, and less knowledge sharing, and lower performance. The effects for task conflicts, knowledge sharing, and lower performance were stronger for short-term teams than for long-term teams. However, long-term teams had more relationship, process, and other conflicts and a lower communication frequency than short-term teams. Interestingly, in studies with continuous measures of virtuality, they found that the relationship with task conflict was more negative (i.e., lower conflict for more virtual teams) and that the relationships with knowledge sharing and satisfaction were more positive, and they found no association with performance. Other meta-analyses found positive effects of virtual teams compared with face-to-face teams. For example, compared with face-to-face teams, virtual teams generated more ideas, needed less time for task completion, and team members were more satisfied, if there was a fit between the group support system and the task and if the group received appropriate support (Dennis et al., 2001).

Also, Fjermestad (2004) reports an increased number of ideas generated in teams using synchronous group support systems, while no differences were observed between face-to-face teams and teams using synchronous group support systems with respect to satisfaction and usability. However, face-to-face teams showed higher levels of consensus and perceived quality, communicated more, and required less time to complete the tasks. Similarly, Rains (2005) reports that teams using synchronous group support systems generated a larger number of unique ideas and experienced less member dominance than face-to-face teams.

Purvanova and Kenda (2022) criticized studies that reported negative relationships between virtuality and team effectiveness because they analyzed primarily short-lived student teams and because organizational virtual teams were severely underrepresented in these studies. Therefore, Purvanova and Kenda (2022) compared the impact of virtuality on team effectiveness of 73 independent samples of organizational virtual teams and 109 independent samples of non-organizational virtual teams. They found that, in organizational teams, virtuality did not show a direct positive or negative relationship with any of the team effectiveness outcomes they examined. They had analyzed the following outcomes: productive outcomes (earnings, accuracy, and process improvements), performance outcomes (including both externally rated and team member–rated team performance), social outcomes (including cohesion and team trust), and individual team member outcomes (including project/task satisfaction and relational quality). Supplemental analyses showed that these neutral relationships between team virtuality and team effectiveness in organizational teams were not moderated by virtuality operationalization (technology dependence versus geographic dispersion), industry type (information technology [IT]/telecommunication, service, and production), company type (multinational and domestic), occupation of team members (IT/engineering, research and development [R&D], and consulting/management/sales), national diversity within teams (homogeneous vs. heterogeneous), and gender diversity within teams (percent males). They also found that results from non-organizational teams were significantly more negative than results from organizational teams. However, this was not the case when studies used graduate student participants, long-term teams, continuous virtuality measures, and classroom projects instead of laboratory tasks stimulating greater participant investment.

Outcomes of Human-Autonomy Teaming

A special type of virtual teams are HATs. In their review of 76 studies on human-autonomy teaming, O’Neill and colleagues (2022) report that communication among autonomous agents and humans tended to be different than communication among humans. Performance of HATs was typically lower than performance of human–human teams because of their lower-quality communication. High reliability of autonomous agents showed consistently positive effects on outcomes such as trust, workload, and performance. However, reliability interacted with transparency. Lower agent reliability could be partially compensated by higher transparency. When humans were aware of lower agent reliability, they were more trusting and showed higher performance. However, studies on transparency on autonomous agents showed mixed effects. On the one hand, it seems that transparency can clarify the reasoning and decision making of autonomous agents. On the other hand, it may lead to less vigilance in overseeing or questioning the autonomous agents’ work.

With respect to HATs, they found that higher levels of agent autonomy and interdependence of autonomous agents and humans led to better team outcomes. Although one could assume that autonomous agents are designed to perform tasks and roles that are difficult or too difficult for humans, they did not find evidence that autonomous agents were particularly useful for teams working under conditions of high task difficulty. However, the studies they reviewed were laboratory-based and did not include field settings. Most of them used only a single human–agent dyad performing an action or execution task with a moderate level of difficulty and very limited levels of communication capabilities and partial autonomy of software agents. As studies focused on performance, workload, trust, situational awareness, team coordination, and shared mental models as dependent variables, research on team viability, development, and learning in HATs is lacking. Owing to these restrictions, caution is advised in generalizing these findings.

Practical Implications

Research findings that stronger negative effects with respect to virtual team functioning were found in short-term compared with long-term teams indicate that teams would benefit from team-building interventions to support team learning and adaptation to the specific challenges of virtual team work, such as developing task, team, temporal, and ICT shared mental models. Meta-analytic studies show that team-building interventions have stronger effects if teams are large (Klein et al., 2009). This might indicate that team coordination and developing a common understanding are more challenging with increasing team size but that small teams can more easily regulate themselves.

Since the development of trust seems to take more time in virtual compared with face-to-face interactions (Breuer et al., 2016), face-to-face kick-off workshops or team development interventions are recommended (Hertel et al., 2005). Besides trust-building activities, documenting team interactions particularly in virtual teams is recommended, as the need for trust decreases when team interactions are documented. Study findings on HATs suggest that it is important to provide transparency regarding the capabilities and roles of software agents, particularly if systems lack reliability, but also stress human responsibility to prevent complacency (O’Neill et al., 2022). Findings of communication theories suggest that this implication holds for ICTs in general.

Future Research

Study results indicate a serious method bias, as more negative effects of team virtuality are reported when short-term laboratory teams are compared with long-term organizational teams (Purvanova & Kenda, 2022), and additional longitudinal studies, particularly with organizational teams, are needed. Furthermore, as team virtuality seems to be multi-dimensional, continuous measures or considering the different dimensions of team virtuality seem to be promising. It might be also worthwhile to examine non-linear effects of team virtuality and to study different forms of hybrid teams. Many studies have focused on the effects of team virtuality; therefore, more studies on mediating and moderating variables are needed to learn more about the causal mechanisms. As virtual teams are a multi-level phenomenon, consisting of individual team members and being embedded in organizations and societies, multi-level studies would be promising. Besides virtual team effectiveness, cross-level effects of virtual working conditions and team processes on individual outcomes such as perceived life-domain balance and stress could be interesting. Particularly, virtual leadership research is still needed (Avolio et al., 2014). Last but not least, research on HATs is still in its beginning stages. Studies analyzing larger teams, more complex tasks, more autonomous agents, and team interaction processes and outcomes both in the laboratory and in the field are needed.

Further Reading