Working in Virtual Teams: A Systematic Literature Review and a Bibliometric Analysis

Modern developments in technology have changed the way we socialize, communicate and work. Globalization, Information and Communication Technologies, digital culture and the increase in the amount of technology available for online communication mean that more organizations are implementing virtual teams. The growth in the use of virtual teams in organizations has incited researchers to investigate the different aspects, factors and challenges of these teams. This article uses a systematic literature review and a bibliometric analysis of virtual teams to identify the most relevant articles on the subject. These articles are then thoroughly reviewed and finally, a summary is made of all the research published over a five-year period. The systematic review of literature proposed by Ramey and Rao [1] and enhanced by Pulsiri and Thesenvitz [2] was used to examine the Scopus and Web of Science databases to identify the theories, research problems, research methodologies and results of 2354 studies on virtual teams published between 2015 and 2019. The main topics of the existing research in the field are reviewed, and the main limitations, problems and existing gaps in research are presented.

about virtual teams and their performance in the Scopus and Web of Science (WoS) databases. These articles were then reviewed and summarized, by means of a systematic literature review and a bibliometric analysis.
The article is organized in the following way. Firstly, the fundamental concepts which the research uses are defined and some related research papers are quoted. Then, the inherent methodology of a systematic literature review process and bibliometric analysis are discussed. Following this there is an analysis and discussion of the results, and finally conclusions are drawn.

II. CONCEPTUAL BACKGROUND
This section presents the concepts used in this study and gives a brief explanation of each one.

A. VIRTUAL TEAM
A virtual team is a geographically dispersed group of individuals who work together to achieve a common goal. ICTs allow each team member to communicate and coordinate from different locations in different time-zones outside the boundaries of the organization [7], [8]. These working groups have been given different names, such as, virtual teams, distributed teams, remote teams, computer-based teams, online teams and cross-site teams, and are used interchangeably in the literature on the subject [9].

B. CLOUD COMPUTING (CC)
This technology is usually explained in one of the two following ways. The first does not make any difference between the cloud and the Internet and includes them both in the same concept. Therefore, a reference to cloud computing includes any service available on the Internet. The second explanation takes a wider view and uses the term ''cloud computing'' to refer to a range of technologies that provide a number of advantages for both the customer and the service provider. This type of technology provides ''economies of scale'' for Internet services by reducing costs and increasing scalability [10]. The cloud is understood to be all the tools used to store, and possibly process, the information of an organization, but they are not physically positioned at the premises and can only be accessed by connecting to the Internet [11].
Cloud technology is a breakthrough, but it should be adopted in a modular way which allows for the constant reconfiguration of resources and services so that the changing needs of the market can be met [12]. The cloud as a new technology and business model represents progress, but it must be adopted in a modular way which provides a wide range of reconfiguration possibilities, so that resources or services can be accumulated in a flexible manner, in order to meet changing market demands [13].

C. BIG DATA (BD)
Big Data (BD) is the result of technological advances that have made the consumer a data source. This means that there is a large number of high-speed and diverse ''data sources'' available. This has led to the definition of Big Data as the three ''V''s of Volume, Velocity and Variety by [14].

D. BUSINESS INTELLIGENCE (BI)
BI derives from the management of knowledge by companies and is made up of a set of strategies, actions and tools for the creation and administration of knowledge by analyzing the data held by an organization or company [15].

E. COMPUTER MEDIATED COMMUNICATION (CMC)
One of the main changes being seen in organizations, which has been researched on various occasions, is the use of virtual teams in the different departments of an organization [16]. Virtual teams are essential for the success of a knowledge-based organization in a volatile, dynamic and globalized market. In addition, a very important factor is the reliance on collaborative technologies using Computer Mediated Communication (CMC), which provides rapid and effective collaboration and communication channels for team members [17].
Current research is dedicated to finding relationships between the increasing use of virtual teams in organizations, and advances in CMC technologies. These technologies have strongly influenced organizations, not only in the way members of the organization collect, communicate, share, and distribute data, but also in the dynamics of the relationships between team members [18].
Information and Communication Technologies (ICTs) have become an important tool for business globalization when designing the working strategies of any type of organizational structure. Modern design methods use CC tools and systems with access to Big Data and BI [19].
Literature on the matter shows that using virtual teams means more efficient work and this creates value for organizations. Other advantages include increased scope, improved response times and the dynamic adaptation of organizations to new situations [8], [20], [21].
Virtual teams also allow companies to make better use of their resources by hiring staff with the necessary experience, regardless of their geographic location. This can improve the overall performance of the organization [22]. A virtual team also helps improve creativity, facilitates the acquisition and exchange of knowledge, and enables organizations to respond faster to market changes [23], [24].
Another line of research, which is similar to the previous one, investigates the importance of CMC technologies by analyzing the factors that influence the efficiency of teams using these CMC technologies compared with traditional face-to-face teams [25], [26].
CMC technologies allow the interaction of virtual teams in several ways, at the same time and place (e.g., email in an office), at the same time and in different locations (e.g. instant messaging), at the same place at different times (e.g., video conference), and at different times in different locations (e.g., online classrooms) [18].

F. REMOTE WORK
Remote Work (RW) is defined as work done at a location far from the company headquarters or factory, where the worker has no personal contact with other co-workers, but is able to communicate with them using modern technology [27].

G. RELATED WORKS
As shown in Table 1 SLR and bibliometric analysis have been used in ''several previous research papers'' [28], [29]. Some of the research papers which use SLR to investigate virtual teams and remote work have been relevant and interesting for researchers and this method of literature review has proven useful and, above all, can be reproduced. Clark et al. [29] analyzed 135 articles about engineering, concluding that the relevant factors were empathy, trust and technology. Santos Rocha and Fantinato [28] investigated virtual teams with SLR in order to show that Business Process Management (BPM) is a potential area for the successful use of lineal software products. Nguyen-Duc et al. [30] investigated geographical separation as a relevant factor in the performance of virtual teams, while Yu [31] studied the relationship of global business with virtual teams, and especially the aspects of communication, leadership and trust.
Acharya [32] studied remote workers in virtual teams in the area of administration and economy in three cities in India. Charalampous et al. [33] and Donnelly and Johns [34] researched the areas of psychology, business and administration. The first analyzed 63 quantitative, qualitative and mixed studies, finding social isolation to be a negative factor. The second focused on empirical work and concluded that the location of the remote worker appears to be an important factor. Finally, Vayre [35] used the SLR method in a detailed and complete analysis of virtual teams, identifying motivation as a positive factor and family conflict as a negative factor.

III. RESEARCH METHODOLOGY
The aim of this research was to use a systematic literature review (SLR) and a bibliometric analysis to investigate virtual teams working on programming projects. To achieve this, the following factors were considered in the methodology.
There were no previous articles that used both bibliometric analysis and SLR methods when investigating VT. However, research by Vatananan-Thesenvitz [2] used the CIMO (Context, intervention, mechanism and outcome) model to show the usefulness of literature reviews using BA and SLR. According to these authors, such revisions can be automatic and give an overview of the subject being investigated (bibliometric analysis) and then select the most relevant articles (SLR). Therefore, the benefits of both literature review systems are coupled and the resulting research is more comprehensive. Therefore, this methodology could be used in this research to analyze the factors that affect the performance of VTs, since it analyses a large amount of literature and then selects the most relevant articles with a systematic review. Other research in other areas used a similar methodology, such as Caiado et al. [36] who analyzed the literature on eco-efficiency, Pulsiri, Pulsiri and Vatananan-Thesenvitz [2] and Vallaster et al. [37] who analyzed articles about entrepreneurship.
However, few articles in the literature provide a comprehensive analysis of cutting-edge research to show where new virtual teams are being used, and where new collaboration and communication technologies have been incorporated into virtual teams. The importance of using the systematic SLR review methodology proposed by Ramey and Rao [1], [31], [50], [51] is based on the following four points: 1) It reduces bias and/or subjective judgment when choosing the search items. 2) It summarizes the results of existing studies and evaluates any inconsistencies with previous studies. VOLUME 8, 2020 3) It identifies, as far as possible, gaps in knowledge or incongruous or weak findings that can be areas for future research. 4) It provides criteria for new research activities and for maintaining periodic updates on the subject. The prevalence of virtual teams in organizations, the existence of different team structures, the emergence of different CMCs to facilitate communication in virtual teams, and the growing amount of literature investigating a variety of virtual team features requires a careful literature review to understand the current state of research on virtual teams and provide areas for future research.

B. RESEARCH QUESTIONS
The main aim of this research is to analyze the most important articles about Virtual Teams in the existing literature with bibliometric tools. Two large databases, Scopus and WoS (Web of Science), were used to collect the information. These articles were then systematically reviewed and summarized. The result is a systematic framework with categorized search terms that sorts, structures and gives important information about Virtual Teams for researchers, knowledge engineers and professionals working with virtual teams. This study has the following research questions: • Q1. What are the main areas and topics of current research in virtual teams?
• Q2. What are the main constructs that have been investigated in current research on virtual teams?
• Q3. What are the gaps in existing research and possible areas for future research?

C. SLR PROTOCOL
To answer the questions above, a systematic literature review and a bibliometric analysis were used to find the most important research papers on virtual teams. A systematic review and analysis were carried out which classified and categorized the research papers identified. The dependent, independent and moderator constructs and terms were also analyzed and classified, as shown in table 2.
The SLR methodology used in this research was proposed by Ramey and Rao [1], [50]. This methodology was chosen because it is simple to apply, uses well-structured steps and has also been used in several previous research papers about SLR. This methodology is complemented by the ideas in the research by Pulsiri and Vatananan [2], that incorporates automation and bibliometry and consists of the following stages, planning, realization, analysis and synthesis and finally, reporting. The research methodology used is shown in Figure 1 below: The SLR in this study was planned with the topic and research questions that were given in the Research Questions section above, Q1-topics and areas, Q2-constructs, and Q3-gaps in research. The published literature on Virtual Teams was reviewed in order to identify the main research topics in this area, prepare the research terms and define a framework of the constructs to be used in the research. In the review stage, a thorough search of the literature was carried out using numerous rules and search parameters to identify the most relevant set of articles for the following study. These rules and parameters included the year of publication, the databases in which the document is included, the search keywords and the inclusion and exclusion terms and criteria.
The process started with setting the time period covered by the study. The main purpose of the study is to show the current state of research on the subject and so only research papers published between 2015 and 2019 were included. Two of the most highly acknowledged databases, WoS and Scopus, were used to find the initial group of research papers. The SCOPUS database provides the largest database of summarized and indexed articles and the largest source of citations and search terms for finding relevant literature [53]- [55]. The next step was to define rules with logical connectors to fine-tune the literature search for the terms and keywords used in the research. The search terms and keywords were initially chosen from the results of an exploratory search which included ''virtual team(s)'' and ''virtual collaboration''. The keywords and terms for the search were selected from the range of keywords included in the literature found in the exploratory search.
During this initial search, the main concept term (i.e. virtual team[s]) was found to have several synonyms. For example, some researchers used the term ''online group'' and others ''computer-mediated team'' or ''distributed team''. Different databases were also found to have different term selection and algorithmic search behavior and these differences affect the search results.
To ensure that an exhaustive list of documents was retrieved from the different databases, and to ensure that biases were minimized, the search terms and keywords were organized into several search strings that would work with all of the databases (Table 3). Advanced search features were also applied to maintain the consistency of the results from all the databases. Synonyms of all the words were also included in the terms searched for. In addition, an iterative process was used for the search terms and keywords in the databases, as shown in Table 3 below.
The Publish or Perish (PoP) software [56] (www.harzing. com/pop.htm) was used to search for article topics with the keyword ''Virtual Team''. This software is produced by Research in International Management, and is used to retrieve and analyze academic research articles on the SCOPUS and Web of Science databases, among others.
The data was collected on January, 2020. The results of an initial search were collected and put into a file with a bibliographic format ''.BIB'' using the title and summary references. After this process there were an initial set of 2354 relevant published works that were then used for further analysis.
The term ''Virtual Team'' and its variants were used as the key phrase in the search of the titles, summaries, and keywords of the articles in the databases. The next step was to formulate the inclusion and exclusion criteria. Researchers held collaborative meetings to prepare the criteria. First, exclusion rules were drawn up and subsequently applied to the articles found in the searches. Four limiting rules were applied to the content of documents and related fields: 1) Articles and documents which describe and deal with any aspect of the communication, collaboration and coordination of virtual teams are considered, since the field is multidisciplinary (e.g. technical aspects, behavioral aspects, management, tasks, processes, leadership, knowledge management, etc.). 2) Articles and documents should consider a virtual team as a relationship for communication, collaboration, and coordination between its members. 3) Articles and documents must provide a clear link or contribution to the different aspects (e.g. problem, research question, method, results, or findings) of communication, collaboration, and coordination in virtual teams. 4) The main objectives and research questions of the articles or documents should be clearly and sufficiently described and explained, and the methodology should propose appropriate steps to address the research problem and answer the research questions.
The following criteria ( Figure 2) were obtained after using the above rules: 1) The results of the research should only include studies published in English, as these are currently the most accessible and disseminated publications worldwide. Studies published in languages other than English were excluded, resulting in 2330 remaining articles.
2) The results of the research should only include research articles and conference documents. 432 documents were therefore excluded, leaving 1898 remaining for analysis. 3) Because the goal was to only consider high-quality publications, only those which had undergone peer review and, for which the complete text was available were included in the results. This criterion was applied to the previous 1898 articles and 342 articles and publications that were not reviewed by peers or for which the full text was not available were excluded. 4) The titles and summaries of all the remaining articles and documents were then analyzed using content-related selection rules. This stage of the process resulted in 724 valid articles (832 were excluded). 5) The full texts of the remaining articles were analyzed using the same content-related selection rules as in step 4, above. This resulted in a total of 585 articles being excluded, which gave a final result of 139 acceptable articles for the analysis.
Once the final set of articles and documents for analysis had been obtained, the following stage of the SLR process was to extract and summarize the data. This is a complex and iterative process in which the articles are reviewed in depth allowing similarities and differences to be found, and, limitations and research gaps to be identified for future research in this field of study. The next section describes the data extraction and summary stages in detail.

D. PRELIMINARY DESCRIPTIVE ANALYSIS PRIOR TO THE SLR
A preliminary descriptive analysis was carried out prior to the SLR analysis. The reviewed articles and documents were grouped according to citations, type of publication, language used, and country of publication. This was done to identify the general profile of the different documents found in the bibliographic search. Table 4 above shows the queries used to obtain the articles and documents for this study from the databases. The software with the queries was applied to the data retrieved from the SCOPUS database and the articles and documents were grouped by year, document type, and author. The initial 5-year search range was extended to see how the groups of publications had changed over the years. The difference between the total number of articles found in this search (1782) and those obtained in the search carried out in the complete process (2354) is due to the fact that in this case only the Scopus database was used, since the objective was an initial approximation and the Scopus database contains the largest number of documents about this research topic.

1) ANALYSIS OF CITATIONS
The table 4 shows the period of time, 28 years, which was analyzed in the initial review. Also shown are the total number of citations of ''Virtual Team'', and the number of citations per year, per document and per author. It should be noted that the average number of authors in each publication was 2.64.
The h-index (Hirsch index) and g-index proposed by Egghe were then calculated. These are two of the most relevant measures in a bibliometric analysis [57]. The h-index shows the publications of an author that have had the most citations and the g-index shows the number of articles published by an author and the number of citations they have had. An h-g metaindex also exists, which is calculated as the geometric mean of the two indices, but it is not a relevant measurement for this study.

2) DOCUMENTS AND SOURCES
The analysis of the Scopus database gave information about two different periods of time. One was the Historical Scopus (''Scopus H'') and the other was the 5-year Scopus (''Scopus 2015+''). The information obtained was grouped for the type of scientific document it was, i.e., conference paper, article, book, etc. The source of the document refers to whether it is from a magazine, a document from a congress, a book or a commercial publication. Table 5 also includes more types and sources of document, such as, editorial, note, short survey, errata and indefinite document type. However, these types only have a rating of less than 1% each. It can be seen that the vast majority of the publications were scientific articles, and account for 41,72% of the total number of publications. Documents from conferences were the next largest group (39.93%) and the other types of documents collectively accounted for less than 18% of the total. It is interesting to note the low number of books, that only accounts for about 1% of the publications.
The results of the search were grouped into five different types of media source (Table 6). Journals accounted for the largest group (47,43%). The next biggest group, 356 (36,55%), was conference papers. Books accounted for  almost 10% of the results and there were only 5 commercial publications (0.51%). Table 7 shows the number of documents published in 9 different languages found in the bibliometric study. Most of the documents found were published in English (97.66%). The second most common language was German (1.02%) and the other languages, such as French, Chinese or Spanish, were less than 1% each. Table 8 shows an ordered list of the 20 countries with the most publications. The United States ranks first with a total of 1686 publications, which is more than 40% of the total number of documents. This is followed by the United Kingdom (5.61%) and Germany (5.36%). In the lowest positions are Greece, New Zealand and Norway.

IV. RESULTS
The articles and documents were obtained from the databases so that they could later be analyzed to find the information about the research topic. A summary of the information found in the documents and articles was then used to answer the  research questions defined in the primary and secondary aims.  The h and g indices were then calculated for the results presented in Table 9 and the last two columns show the rankings for the Journals using these indices. The International Journal of E-Collaboration is in the top position for both indices. However, the International Journal of Networking and Virtual Organizations ranks second for the h-index, but third for the g-index. Team Performance Management takes third place for the h-index and joint-third place with International Journal of Networking and Virtual Organizations for the g-index. Human Resource Management Review can be seen to be the Journal in the lowest position for both of the indices. Specialized journals and those concerned with business administration are in the most highly ranking positions. Figure 3 below shows the number of documents and articles published about Virtual Teams over time. The first of these was published by Cohen in 1991 [58].
Unfortunately, research in this field was not very popular until 1995 with only one (1) article in this research area being published. However, from 1996 to 2010, publications about  Table 10 shows the details of the results of the ''Virtual Team'' search including columns for the historical Scopus search with no time limit, and another for the information found on WoS. These columns are included for comparison. It can be seen that there are significantly fewer publications in WoS than in Scopus. Articles are clearly the most common type of document found in all the searches. In Wos, Literature Reviews are the second most common, whilst in the historical search on Scopus they are the third. Table 11 gives detailed information about the results of the database search for the term ''Virtual Team'' and its equivalents. It also includes the results from the Scopus historical search and the search on WoS. Significant differences can again be seen in the results from the two databases, whilst the frequencies of each document type have a similar order to that in Table 9. Table 12 shows the number of documents published about virtual teams for each research area. The largest number of publications are for computer science, business, social sciences and engineering. Virtual teams for medical specialties have the least number of publications.   Table 13 shows the 20 most relevant keywords used in publications on ''Virtual Teams''. The most frequently used keywords in publications after ''virtual teams'' were found to be ''virtual reality'' and ''management''. The terms which were least used were ones related to software and computers. Table 14 shows a coded sample of the 25 most-relevant articles per construct since 2015, in order of number of citations.

V. DISCUSSION OF RESULTS
The bibliographic analysis provided a number of results which can be used to answer the research questions:

Q1. What are the main topics of current research on virtual teams?
This question was answered after the journals that published the largest number of articles had been analyzed and, of course, from the results of the keyword searches for ''Virtual Team'' and related terms.
• The keywords (apart from ''virtual team'') that appeared the most in the bibliographic search were: virtual reality, management and project management, which shows the importance of the work being done in this area. However, the term ''leadership'' which describes the performance of the people involved in the areas was one of the terms which appeared least, along with software design, which is an essential element of remote communication.
• The journals which published the greatest number of articles in the period studied were specialized in technology and, business administration and management. This demonstrates, once again, the interest that exists for this type of remote work in the two areas mentioned above. The years with the highest values of g and h indices were 2007 and 2008. VOLUME 8, 2020  • In addition, the most common format of publication depended on the database used. In the main keyword search, WoS included many more academic articles than the other databases. Scopus, on the other hand, included a variety of formats. When other related keywords were analyzed however, there were more conference papers than articles in this database. The proportion of books in the total number of revised documents was low. While some key concepts such as virtual teams and technologies have been extensively studied, there are still a lot of areas, such as software, leadership or design, which are all essential for high-performance virtual teams, that have still not been studied in depth [48], [49]. One of the most important technologies for improving the performance of virtual teams is virtual reality. Studies, such as that carried out by Lee et al. [50] conclude that it is the perceived enjoyment, and not the usefulness, that influences the use of this technology.
Some videogame companies have used this idea to develop methods of interaction that allow players to perform joint tasks realistically, something that can be extrapolated to the virtual team. There have also been advances, not only in medicine or psychology, but also in other areas such as museum tours, where once again, the user demands a realistic experience [59]- [61].
Therefore, the potential of using this type of application in virtual teams, where interaction is important, seems obvious. Virtual reality can be used to make remote relationships similar to interpersonal ones, thus removing one of the drawbacks of virtual teams and remote communication which are not face-to-face [62]. Therefore, correct software design is critical for an effective and efficient virtual team and this should be studied further.
However, technology is not the most important factor of performance, as [7] showed. Virtual teams are made up of people who have emotions and different needs in all areas of their lives. This has inspired these authors to investigate trust and commitment.
There are also other factors which are not investigated in the literature, such as leadership and, where appropriate, e-leadership. An efficient leader can implement different technological tools effectively so that they are accepted by the whole team. Another essential responsibility of the leader is to monitor projects without having a negative effect on the team [63].

Q2. What are the constructs that have been examined by current research on virtual teams?
A detailed analysis of the most relevant publications was carried out to answer this question. Most of the constructs which were studied were related to the performance of virtual teams. These included: • The technology that is used for remote relationships between team workers. Most researchers consider this to be a very important variable.
• Trust is essential in face-to-face relationships at work and is also essential in virtual ones. It was considered a relevant construct when working remotely in a large number of the documents.
• Communication technology affects personal relationships and even more so in virtual relationships.
• The equipment needed for remote relationships, especially the different types of technology which can be used.
• Although the term ''leadership'' only appeared in 2.60 % of the documents, many authors consider it an important characteristic of a good manager which is useful for gaining the trust of the group.
• Education, including online training, is considered a key factor for performance. While it is true that this term did not appear as such in the bibliographic searches, it should still be considered important. According to Gilson et al. [52] previous research on virtual teams mainly occurred in laboratory conditions using teams of students and comparing them with their face-to-face counterparts. All the articles found in this study shared several key points, which were, firstly, that studies now use a diverse range of disciplinary approaches and study virtual teams in different areas, such as accounting, applied psychology, business management, communication, computer technology, education, engineering, information systems and software design. Second, much of the work uses the same constructs and indicators (e.g., Task Functions, Composition Factors, Technology and Leadership), Mediators (e.g. communication, coordination, conflict and trust),  moderators (e.g., Virtual model and Interdependence), and results (e.g., performance and affective reactions). However, the ways these constructs have been treated are quite different. There has also been a lot of research about the professionals who use virtual teams, including information technology professionals (Baruch and Lin) [64], software developers (Muethel, Gehrlein and Hoegl) [65], research and development professionals (Stark and Bierly) [66], consultants (Suh, Shin, Ahuja and Kim) [67], developers of new product (Stark and Bierly) [68] and engineering teams (Forester, Thoms, y Pinto) [69].
The software used by virtual teams has to be produced by creative people, who use and promote non-traditional development teams that are different to normal business routines (Nash) [70]. These non-traditional teams may consist of individuals working non-traditional hours or in non-traditional locations. Software development usually happens in a globally distributed environment (Agerfalk et al..) [71]. More and more companies use these types of teams, which include members from various backgrounds, as they communicate more clearly and effectively than teams that are homogeneous (Granered) [72]. This has led companies to not only assemble teams with members from different locations, but also to ensure that such teams work efficiently.

A. COMMUNICATION IN VIRTUAL TEAMS
Communication in virtual team has been widely investigated. The action processes, communication, coordination and knowledge sharing are the areas which have been studied the most (Gilson et al.) [52] at they are critical for predicting team efficiency and effectiveness (Kock and Lynn) [73] According to Argaña [74], it is important to consider separately how the content of the communication is interpreted, and also the ease and monitoring of it. Limitations of the interpretation of virtual communication have been identified. One of the challenges of virtual teams is that information must be communicated, shared, and interpreted without the help of gestural language. Some authors have also pointed out that the communication about projects must be very clear and the information must be handled appropriately.
Research on virtual teams usually considers that team members may encounter difficulties when communicating with computers (Gilson et al.) [52]. Despite the barriers and difficulties shown in previous research, current research has shown that as the next generation of employees (i.e. millennials) enter the workplace, many of the previous difficulties lose importance (Gilson et al.) [75]. As Gorman and his colleagues state [76], millennials have the ''ability to effectively use networked digital communication technologies to perform a wide variety of tasks quickly and smoothly''. Indeed, younger generations might see working in virtual teams as commonplace and working in a faceto-face environment as the exception.

B. TRUST IN VIRTUAL TEAMS
The literature about this states that building trust and having members identify with the team are two important sources of concern. In fact, if team members don't know each other, they don't know if they can trust each other, so before becoming an effective virtual team, the members have to know each other.
Trust is defined as a person's willingness to become vulnerable to the actions of others with the expectation that others will continue with their commitments. Two types of trust have been defined in the literature on the subject. These are called cognitive trust and affective trust (Mayer et al.) [77], [78].
Creating trust in virtual teams is often connected with the exchange of knowledge (e.g. Liu and Li) [79]. Pinjani and Palvia [80] investigated the role of functional and deep-level diversity, finding that deep-level diversity is more relevant to trust and knowledge sharing. In addition, interpersonal trust and trust in technology were also found to be relevant to knowledge sharing (Golden and Raghuram) [81], whilst the lack of interpersonal trust and confidence in technology can be obstacles (Breu and Hemingway) [82].
In a longitudinal study of global virtual teams, Goh and Wasko [83] found that when the team member's actions were visible, trust was not a key factor in resource allocation. In globally distributed teams, trust greatly diminished the negative effects of the diversity of members on performance (Garrison, Wakefield, Xu, and Kim) [84] C. CULTURAL DIVERSITY Cultural diversity is the construct which is used for the globalization of virtual teams (Hoch and Kozlowski) [85], and the findings suggest that perceived differences in national culture and language barriers can adversely affect teams (Au and Marks) [86].
Taking a more detailed approach, Mockaitis et al. [87] found that members of a global virtual team who had a collective mindset accepted team processes favorably. Duranti and Almeida [88] found that some cultures (Brazil) prefer advanced computer-mediated communications such as audio and video conferencing, while others (the United States) preferred less advanced computer-mediated communication tools, such as email and chat.
Some ways to alleviate issues associated with cultural differences include cultural training, face-to-face meetings, and team-building activities. However, team interactions should be monitored to identify potential difficulties (House) [89].

D. TEAM DISTRIBUTION
The physical separation of virtual teams was investigated in the work by Allen [90] on geographic distancing and dyadic communication. The study of the dispersion of teams by O'Leary [91] analyzed three dimensions of dispersion, (1) spatial, which is the actual geographic separation of teams, (2) temporal, which is the time difference of normal working hours in different places, and (3) configuration, which is the distribution of team members across the sites.

E. CHARACTERISTICS OF TEAM TASKS
The literature found in this study included research on the interaction between the type of task a team does and the type of communication used and the impact on team performance (Rico et al., Bell et al., Montoya-Weiss et al.) [92]- [94]. As virtual teams rely heavily on communication technology to coordinate their work, the relationship between the nature of the task and the effectiveness of communication and its impact on team performance must be studied. The variety of tasks is the frequency of unexpected and novel events that occur during information processing. A low variety means that team members can be sure about which future events will occur. A high variety means that team members cannot predict future activities. High variety would also involve a lot of unplanned communication between team members.

F. COHESION
Cohesion is a sense of unity in a team. Cohesion is important for virtual teams, and is associated with an improvement in performance (Lurey and Raisinghani, 2001, Maznevski and Chudoba) [95], [96] and satisfaction (Chidambaram) [97]. According to Salisbury et al. [98] classical research suggested that the physical distance between team members can create a psychological distance between them. Salisbury et al. [98] suggested that the physical dispersion of the virtual team may inhibit the development of cohesion compared to on-site teams. In addition, virtual team members may have different perceptions of what cohesion is. In other words, the perception of cohesion, which is based on communication between the members of the team, is affected by the method used to communicate.

G. LEADERSHIP
A classic definition of leadership states that it is when a person makes other people do something (Kort) [99]. Leadership is an influential relationship between leaders and followers who try to make changes that benefit their mutual purposes (Kort) [99].
Research on leadership in virtual teams has increased rapidly, especially in areas about behavior and leadership traits (Gilson et al.) [52]. Several studies examined the interaction between leadership and virtual work, finding that team members are more satisfied with their team and leader and perceive that their leader is more able to decode messages when the leader is geographically distant from the team (Henderson) [100]. Hoch and Kozlowski [85] found that virtual work reduced the relationship between hierarchical leadership and performance while improving the relationship between structural support and performance.

H. EMPOWERMENT
Empowerment means having the favorable backing of the team leader, especially when this allows members to participate in decision-making. Kirkman et al. [101] showed that teams experience empowerment in four ways, power, which is the collective belief of a team that it can be effective, meaning, the extent to which team members take intrinsic care of their tasks, autonomy, in which team members are free to make decisions, and impact, the degree to which team members feel that their tasks make important contributions. Initially, Gondal and Khan [102] studied the impact of team empowerment on virtual team performance at ten telecommunications companies located in Islamabad. In that study, they found that there is a positive relationship between team VOLUME 8, 2020 empowerment and team performance in telecommunications teams. Also, Kirkman, Rosen, Tesluk and Gibson [101] studied 35 sales and service teams at a high-tech firm and investigated the impact of team empowerment on team performance and the intermediary role of face-to-face interaction. They found that team empowerment is positively related to the two constructs of virtual team performance, which are process improvement and customer satisfaction.

I. AMOUNT OF VIRTUALITY AS A MODERATING VARIABLE
Martins et al., [103] suggested that the amount of virtuality in a team affects processes (especially in communication) and team results (performance and satisfaction) and that it should be considered as an important variable to understand the performance of the team. Cuevas et al. [104] found that the technological intermediation required for the communication of virtual team members can generate what is known as team opacity. Basically, opacity is described as the experience of ambiguity and artificiality caused by this type of intermediation in geographically dispersed work. It can limit coordination and communication in teams and therefore affect their performance. Martins et al. [103] and Gibson and Gibbs [105] investigated geographical dispersion in the processes and results of the team, considering coordination and communication (both implicit and explicit) as critical processes, as well as the foundations for the performance of a virtual team (Fiore et al.) [106].

J. INTERDEPENDENCE OF TASKS
In traditional environments each sub-team usually has its own goals and objectives, and the tasks of the different sub-teams are not dependent on each other, while virtual teams usually have distributed tasks that are highly dependent on the other members and their tasks (Saldaña) [107].
Several studies state that the interdependence of tasks can improve the performance of virtual teams (Chi, Chang and Tsou) [108].
The interdependence and complexity of tasks has been found to influence the inputs and processes of a team and to also have an effect on the relationship of the process with the result (Bell and Kozlowski, 2002, Bowers, Pharmer, and Salas, Hambrick, Humphrey and Gupta) [93], [109], [110].
In the collaborative environment of software development, the tasks performed by team members are highly interdependent. The interdependence of tasks is the extent to which the team members depend on each other to perform their individual work (Van de Ven) [111] and is considered to be one of the structural variables that most influences team performance (Langfred) [112].
Q3. What are the gaps in existing research and possible areas of future research?
The most important research gaps were identified from the results of the systematic literature review, and are explained below.
• Virtual teams still need to be investigated so that this new type of organization can be fully understood, especially as new technologies for collaboration are constantly being developed. The previous sections include bibliometric and systematic analyses of the existing research in this area.
• The main points and ways in which virtual teams operate are still not fully understood and these areas should be investigated further. One of the fields which has not been widely investigated is that concerning emotions and how they influence performance. The guidelines for future research can be summarized as follows: • Classification of terms: The classification of terms, which means ordering the concepts and terms. This is necessary not only for existing research, but also for future research.
• Model-based study: The variables and constructs found in this study can help researchers understand the type of constructs which have already been investigated and whether these constructs were dependent or independent. In addition, researchers may refer to the documents to find more information about any construct and whether it influenced, or was influenced by, other constructs.
• Study the adoption of new collaborative technologies, as well as the patterns of adoption The adoption and use of virtual teams are driven by new technologies that enable better collaboration of team members. Currently, projects and teams need this new collaborative technology (Matthiesen and Bjorn, Stray et al.) [113], which goes beyond the classic video conferencing software that is often used for distributed meetings. The number of communication tools and platforms has increased substantially in recent years (Cardon and Marshall, Cardon, Huang and Power) [115], [116] An interesting application of technology for virtual and distributed teams is the electronic task board. Some of these are based on Kanban, which was successfully used by Toyota in factories in the 1950s. It started being used for software development in 2004 at Microsoft (Anderson) [117]. Examples of the Kanban method include Trello or Canvanizer. The use of Kanban in software development is still an emerging area and it is currently being promoted (Ahmad, Markkula, and Oivo) [118], not only for software development, but to manage the tasks of teleworkers. The main reported benefits of using the Kanban method are, improved software delivery times, better software quality, better communication and coordination, greater delivery consistency, and a reduction in defects reported by customers.
Other collaborative tools include social software (e.g. instant messaging, tools and wikis) (Giuffrida and Dittrich) [119]. These tools are compatible and integrate with social networks and messaging. The most urgent and informal conversations use instant messaging (IM) tools, such as Whatsapp, Telegram or Webchat, among others. These tools reduce coordination problems in distributed projects (Dittrich and Giuffrida) [120].
Another group of applications are known as social network platforms (SNPs), telecommunication platforms (TCP), or enterprise social software (ESS) (Dittes and Smolnik) [121]. Examples of this type of applications are Slack, Yammer, Microsoft Team, and Facebook Workplace (Leroy et al.) [122]. These applications aim to promote collaborative communication with all levels of an organization (Cardon) [123]. They work on mobile devices and improve transparency the virtual teams (Stray et al.) [124]. These applications improve trust (Moe and Smite, 2008) [125] and transparency in virtual teams, compared to the classic method of communication by email.
The reasons why new collaboration technologies are constantly being developed can be found in recent studies (Stray and Moe) [114]. This study found that collaborative technology has important positive effects, such as increasing the feeling of belonging to a team, increasing informal communication, and reducing the need for e-mails. They also found that essential team members being unavailable, the lack of organizational support for unscheduled meetings, and the varying participation of team members in meetings were all obstacles to effective coordination of virtual teams.

VI. INTERNAL AND EXTERNAL VALIDITY
This research is an example of the application of the formal systematic review methodology proposed by Ramey and Rao [1] and Rao and Ramey [50], who proposed a formal systematic review process after modifying a method used by medical doctors so that it could be used in scientific research. The methodology proposed by Pulsiri and Vatananan-Thesenvitz [2] was included to incorporate automation and bibliometry in a study in which terms were systematically classified using the methodology given above. This classification summarized the main research topics investigated in the field of virtual teams in companies over a five-year period. The researchers identified the main topics in the field, and the external validity [28]. The main constructs identified in the results provided the researchers with a set of dependent, independent, moderator and mediator variables, and their corresponding relationships. These variables and the relationships which exist between them can be used when designing regression or structural models in studies of quantitative analysis of Big Data. The internal validity or causal relationships can be found in this way [30].

VII. CONCLUSION
Rapid developments in information technologies have resulted in the rapid advance of virtual convergence. The aim of this study was to systematically review the literature about virtual teams. This article is an example of using a systematic literature review with a bibliometric review. The main conclusion was that although there is still a lot of research to be done and that in the last few years there have not been as many publications on the subject as there were in the period 1998-2010, studying this area has become increasingly important since the appearance of the coronavirus pandemic or covid-19. Technological advances permit teleworking and collaboration in virtual teams. Fear of new pandemics and savings in transport costs will definitely promote this technology in coming years.
The number of published documents about virtual teams was higher in the United States than in other developed countries such as the United Kingdom or Germany. The most commonly published format was the article, followed by conference papers. There are still only a limited number of books published on this subject.
In line with previous research [29], [126], the main factors that need to be researched are related to emotions, which is an area of study that is becoming increasingly important in a wide variety of areas such as marketing. Other areas which can be investigated are communications, technology and trust. A considerable amount of knowledge has been gained about virtual teams over the last five years. However, the relationships between the different aspects of virtual teams and how they interact has not yet been researched.
In the future, researchers could collect more bibliometric information from other academic databases, such as Web of Science, Google Scholar, and Dimensions, in order to complement and add to the information in this study. Above all, the relationships between constructs should be investigated using statistical techniques such as regression and structural equations. In addition, research can use a SLR combined with a bibliometric analysis to identify the constructs which are to be investigated.
This work contributes to the existing research about the constructs and relationship between them in research on virtual teams, and by suggesting future research in this area. It also suggests a system which can be used, by combining existing methods, to identify a higher number of more precise results in literature reviews, which is an essential first step in any investigation. VÍCTOR M. GARRO ABARCA was born in San José, Costa Rica, in 1969. He received the bachelor's degree in computer science from the Technological Institute of Costa Rica, in 1994, and the master's degree in project management from the Technological Institute of Costa Rica, in 2010. He is currently pursuing the Ph.D. degree in business administration with the Technological Institute of Costa Rica, in 2020.
Since 1995, he has been an Associate Professor with the Technological Institute of Costa Rica. His experience includes directing various postgraduate courses, international consulting activities such as an Internet expert, home automation consulting for the energy industry, and collaborations with public and private organizations. He has carried out educational and work virtualization research projects, financed by national and international organizations such as Tec-Virtual, Huascaran, Multimedia Portals, and Universal Medical Records.
Mr. Garro Abarca has been a member of the Costa Rican College of Computer Professionals since 1996, and a collaborator in the Technological Research Club of Costa Rica since 2000.
PEDRO R. PALOS-SANCHEZ was born in Badajoz, Extremadura, Spain, in 1968. He received the B.Sc. degree in computer science from Extremadura University, Spain, the master's degree in marketing from UOC, Spain, the MBA from Camilo Jose Cela University, Spain, and the Ph.D. degree in business administration from the University of Sevilla, Spain.
He has served as a Professor with the University of Extremadura, Pablo de Olavide University, and International University of La Rioja, Spain. He is currently an Associate Professor with the University of Seville, Spain. His experience includes the direction of numerous postgraduate courses and collaborations with private and public organizations. He has participated in research projects. He has participated with different papers, courses, and articles in various conferences and programs on the digital economy, entrepreneurship, and management both nationally and internationally, being his lines of research digital marketing, business organization, and information systems. He has been a guest editor of several international scientific publications.
Dr. Palos-Sanchez is a member of the Andalusian College of Computer Professionals and the Economy College of Sevilla.
ENRIQUE RUS-ARIAS was born in Extremadura, Spain, in 1969. He received the bachelor's degree in economics and business administration from the University of Extremadura, Spain, in 1995, and the master's degree in research in social and legal sciences from the University of Extremadura, in 2010. His experience includes managing several of his own companies and collaborating in public and private projects. He is currently a Researcher with the University of Extremadura, where researching human behavior and general and business economics is one of his main tasks. His main lines of research are entrepreneurship and teleworking or remote work. He combines this activity with that of the Editor in Economipedia magazine.