During the last few years there has been an increasing acknowledgement of the importance of trust in business interactions within the management and organisational literatures . Trust enables cooperation and becomes the means for complexity reduction even in situations where individuals must act with uncertainty because they are in possession of ambiguous and incomplete information. It is not therefore surprising that in the current age of global and digital economy and virtuality  there has been an enormous interest in trust. Handy for example, has put the point quite succinctly: 'Virtuality requires trust to make it work: Technology on its own is not enough' . As Lipnack and Stamps also put it, 'in the networks and virtual teams of the Information Age, trust is a 'need to have' quality in productive relationships' , while according to Platt , trust is essential to any virtual team as these teams do not have everyday interaction, and the possibility of losing trust is much higher.
Despite the overwhelming interest, our understanding of trust in this area has remained limited. To date, there has been no systematic examination of the computer-mediated interactions that take place within a virtual team situation. Reliance on mediated interactions and especially those that are text-based and asynchronous such as email has been seen to inhibit the development of good working and collaborative relationships ; such views often derive from the media richness theory which suggests that face-to-face is a richer information medium . My argument, however, is that even though the value of face-to-face communication in creating and promoting a rich information context needs to be highly appreciated, we also have to acknowledge that significant interactions remain computer-mediated and provide extensive opportunities for the development of trust .
Trust is a state of a positive, confident though subjective  expectation regarding the behaviour of somebody or something in a situation which entails risk to the trusting party. It is a dynamic and emergent social relationship that develops as participants interact with each other over time and depending on the situation.
While trust has been identified as a key feature for the success of virtual interactions, empirical research in this area has remained limited. Jarvenpaa and Leidner  have conducted one of the most detailed research projects into studies on trust and virtual teams thus far. Their eight-week study of seventy-five teams of university students each consisting of four to six members, highlighted significant differences in the behaviours and strategies between high- and low-trust teams and supported the existence of swift trust; this type of trust presumes that roles are clear and that each member has a good understanding of others' roles and responsibilities .
While the study by Jarvenpaa and Leidner  provides useful insights in this area of research, it does have some limitations when attempting to apply its findings in a business context; tasks and projects may not be as well-articulated in this context, and external factors (e.g. clients' specifications) may require changes in the direction of an already assigned project. Panteli and Duncan , in a study of a virtual team project, managed by a virtual organisation and involving a group of geographically dispersed contractual employees, finds that the content, both formal and informal, of communication, as well as its frequency helps in building and maintaining an interactive social situation and can act as the frame for reference for constructing the trust relationship.
Accordingly, within a business environment where conflict and power differentials prevail, building trust is not always a swift process. This argument is further supported by the findings of an empirical study that is presented in the following section.
Tucker and Panteli  pursued a study of eighteen global virtual teams within a global IT organisation. The study involved interviews with individuals who are employed at the specific organisation and who were part of culturally diverse, geographically dispersed and technology-enabled global virtual teams. Furthermore, the interviewees had worked within a global virtual team for more than 2 months - thus allowing some exploration of the changes within the team over time.
Table 1 below details the common features and behaviours observed within the global virtual teams studied in Tucker and Panteli . The teams were categorised as High-Trust teams and Low-Trust teams and are distinguished in terms of the degree of shared goals that they experience, as well as issues of power and communication.
Tables 1 and 2: Differences between High-Trust and Low-Trust Global Virtual Teams:
|Table 1: High-Trust Global Virtual Teams||Table 2: Low-Trust Global Virtual Teams|
|Factors related to Shared Goals:||Factors related to Shared Goals:|
|Awareness of shared goals||Lack of awareness of shared goals|
|Time given to build shared goals||Lack of shared goals|
|Early and open debate of goals||Opinions of others not considered|
|Primacy of team-based goals||Primacy of individual goals|
|Factors related to Power:||Factors related to Power:|
|Availability of facilitators||Power battles|
|Facilitators' focus on win-win||Coercion|
|Recognition of knowledge as power||Misunderstandings and conflicts of interest|
|Recognition that power moves; power in many places||Use of hierarchical power|
|Power differentials minimised||Perception of 'I have power'|
|Face-to-Face where possible||Asynchronous CMC|
|Regular synchronous CMC
|Adverse effects of time difference|
|Social interaction||Little or no social interest|
Power, defined as the capability of one party to exert an influence on another to act in a prescribed manner, is often a function of both dependence and the use of that dependence as leverage . Indeed, power is an important contextual factor that affects trust  in that it suggests the existence of a unilateral dependency or an unbalanced relationship .
In considering power within virtual teams there is an increasing recognition in the literature that knowledge is indeed power and that teams are often formed to create knowledge through combination and exchange. Within these teams, the team member with power at any given time is the one with the most relevant knowledge at that time. The study found that in high-trust teams, power differentials do not disappear; rather, power shifts from one member to another throughout the life cycle of a project depending on the stage and requirement of each stage.
Several interviewees described the power within their team as originating from knowledge and noted that at any given point in time the most powerful was the individual with the most relevant information. In these situations coercive power was rarely used, and significant emphasis was placed upon collaboration and the use of persuasive power:
"Power tended to move based on whatever activities were going on at that time. I guess it followed those that were most knowledgeable at any point in time. This is not surprising as the reason we selected the external design company was because of their knowledge."
Shared goals are and should be a key characteristic of virtual teams. They could provide a means to developing a common sense of identity for team members which can be of particular benefit to those global virtual teams which meet infrequently or perhaps not at all. These benefits include the establishment of a foundation upon which to build trust and minimise the use of coercive power in pursuit of a collaborative and productive relationship. However, the study finds that even though shared goals are important for the success of virtual teams, these should not be taken for granted. Indeed, goals may not be shared either because they do not exist at all, or team members have not become aware of them, or conversely have their own priorities or share different interpretations of the team's role. Furthermore, this study has also shown that the construction of shared goals is often by no means a one-off activity, but rather a process that requires the participation of all parties involved. Though this could be a time-consuming, iterative and difficult process, these findings allow us to argue that it is far better to invest in it (and as early in the project as possible) than deal with the vicious, destructive downward spiral that result from team members with conflicting goals and poor levels of trust.
As shown in Table 1, shared goals were evident in all of the 'high-trust' teams and, not surprisingly, these teams were also considered to be working well. Conversely in all of the scenarios where trust was described as low, shared goals were lacking. In the situations where team members were of the opinion that trust had been broken, the level of emotion was high.
Moreover, facilitators are found to have an enabling role in constructing shared goals and minimising destructive power differentials. The role of a facilitator is to help in team-building techniques at the early stage of the virtual work project. Similar to the power examples given earlier, situations were described in which facilitators were attempting both to rebuild and improve trust. In these examples the use of shared goals features prominently:
"We had a very definite vision of how we wanted the relationships to work. We were keen to engage and excite the other companies. We gave them an overview of our business and worked hard to try and give them the full picture to create a vision if you like."
"At the very start of the project the project managers from each company got together and put together a comprehensive contract... It was developed jointly and was very comprehensive. We went through a lot of iterative discussions to make sure that the document was extremely well thought out."
Further to the issues of shared goals and power, Tucker and Panteli  found support for the need for face-to-face interaction. However, the opportunities to meet face-to-face have been severely limited by economic pressures and more recently terrorist attacks. Under these circumstances, those virtual teams that work well tend to undertake regular communications via synchronous, 'live'
such as the telephone and videoconferencing systems. Participants confirmed that synchronous media offered more feedback and therefore facilitated understanding more effectively than asynchronous technologies such as voicemail and email. The use of asynchronous technologies was, however, regularly used for documenting and recording agreements and providing brief, simple updates to work progress. The teams that worked well also included a social and fun element in their interactions that appeared to help in creating a stronger shared social context.
Overall, the study has found that the process of jointly constructing team goals holds significant value as it may provide the 'glue' to hold team members together long enough to make possible the development of mutual trust. It has illustrated the significance of shared goals and power in influencing trust development; these factors were not identified in the context of university settings as the tasks are often well-articulated in advance while power differentials, which could influence the degree of inter-dependence among members, are not significant in the case of university students.
The issues presented in this article have been receiving a lot of attention from both academics and practitioners. Context is critical to understanding trust and the fragile sphere of virtual relationships requires a much higher level of trust than do conventional hierarchically controlled settings.
The study presented in this article reinforces arguments in the existing literature on the significance and complexity of trust dynamics in building effective virtual teams. It goes further than the existing research however to identify and illustrate the significance of shared goals and power in influencing the development of trust. In particular, it was found that a focus on jointly agreed goals can help to provide limits or boundaries within which trust can be nurtured. However, the study has also clearly indicated that the construction of shared goals is not often a one-off activity and frequently requires the involvement of all parties involved. Though this could be a time-consuming, iterative and difficult process, this research concludes that it is far better to invest in building trust as early in the project as possible in preference to handling the destructive vicious circle generated by teams with conflicting goals and poor levels of trust. Furthermore, the findings suggest that shared goals may be used to minimise power imbalances, and the resultant use of coercive power, by driving progress through the pursuit of mutually beneficial objectives based on a perceived equity of return.