Abstract
Digital disruption has changed organizations in an unprecedented way. The thriving field of digital leadership is expanding fast and few retrospective studies on this evolution have been made so far. This study presents a bibliometric and network analysis combining both Scopus and Web of Science databases to provide fresh insights into the evolution of the digital leadership research field. This study is based on a review of 79 publications from 57 journals, published between 2000 and 2020. The newness of the topic and the range of journals in which it is published confirms that digital leadership has gained interest from several different areas. Bibliometric analysis provides a description of the research field identifying the leading publishing journals, affiliation statistics, and most influential authors and expressive publications in the research field. Network analyses identify keyword evolution over time, co-citation relationships, and research clusters. Content analysis is used to identify key topics in the field with attention paid to interrelations among them. A brief description of each paper in the dataset and its methodological approach is provided. The results suggest that the topic will continue to attract more research, as it has not yet entered its maturity stage. This paper contributes to the literature by analyzing the relationship between digital leadership and e-leadership. This study also identifies the most leading digital leadership capabilities for a fast-changing world. Limitations and future avenues are also discussed.
Introduction
The COVID-19 pandemic has affected workers in a variety of ways. The lockdown has forced an abrupt change from face-to-face and analog interactions to the digital realm (Faraj et al., 2021). Teleworking, dispersed teams, and remote management were new experiences for many organizations (Kirchner et al., 2021). Some types of work that previously was uncommon or unaccepted in certain organizations (i.e., virtual work) became typical to workers around the globe (Wang et al., 2020). COVID-19 has fostered experimentation as digital technology has made the world more decentralized (Fenwick et al., 2021). However, due to digital technology, people could maintain connections as they were forced to maintain social distance (Peng, 2021). Digital transformation leads to a work transformation (Nagel, 2020), shaping how people work, think, and interact (Hai et al., 2021). Although no one can foresee how the rapid change to digital work will affect organizations in the future, effective digital transformation programs may not be simple to achieve and sustain (Wade & Shan, 2020). Therefore, leadership practices must adapt to a new setting to attain an effective, enduring performance (Contreras et al., 2020). The pandemic appears to be a major test for leaders worldwide, as it is not possible to predict what the new world will resemble (Dirani et al., 2020). While digital transformation represents challenges to leaders (Bartsch et al., 2021), digital leadership (hereafter, DL) helps organizations to deal with risks and constant uncertainties (Fenwick et al., 2021). Against this backdrop, DL has become a key element in the effort to modernize organizations, and it is worth considering in detail (Peng, 2021) as leaders experience new challenges due to digital disruption (Kane et al., 2019).
Some of the works on DL focus on e-leadership (Avolio et al., 2000; Avolio & Kahai, 2003; Kissler, 2001), leading virtual teams (Mitchell, 2012; Perizade et al., 2017), performance (Fernandez & Jawadi, 2015; Wakefield et al., 2008; Wolor et al., 2020), communication (Braun et al., 2019; Darics, 2020; Narbona, 2016), the public sector (Banerjee & Chau, 2004; Rubino-Hallman & Hanna, 2007), technology (Larson & DeChurch, 2020; Leduc et al., 2015), and competencies (Maduka et al. 2018; Roman et al., 2019). Also, encompassing reviews have been done (Avolio et al., 2014; Cortellazzo et al., 2019). Although in the literature there are some definitions of DL, such as it is conceptualized as the human aspect of a leadership operating with digital tools in the virtual world (Narbona, 2016), or it is a combination of leadership capabilities and digital technology to support the decision-making process (Sasmoko et al., 2019), the pattern of research in this field and the link with other related constructs (e.g., e-leadership, virtual leadership), and digital transformation remains unclear. Avolio et al. (2014) reviewing the e-leadership approach, suggest it may be time to rethink what constitutes leadership in an Advanced Information Technology (AIT) scenario (e.g., internet, email, video conferencing, virtual teams). Cortellazzo et al. (2019) stated in the DL review that there is a need for more theoretical contributions between leadership and digital transformation. Therefore, we address the following research questions: What is the structure of the DL domain, and what has been its evolution? What can we learn from that? Is a new leadership paradigm emerging following the digital disruption we are facing? This study aims to close the gap by examining the content, characteristics, and relationships of all the publications from the Scopus and Web of Science databases from 2000 to 2020 to present an extensive and holistic perspective on the domain of DL. Thus, it can complement and go further than a literature review, which is not necessarily extensive enough, or meta-analyses that do not adequately emphasize the relationships between the publications within the area (Ertz & Leblanc-Proulx, 2018; Zupic & Čater, 2015).
A bibliometric study is a quantitative approach for studying the bibliographic material and mapping the research field without subjective bias (Zupic & Čater, 2015). It brings into light the most influential works and authors, the evolution of the most used keywords, the most associated topics, and the dominant outcomes, among other pieces of evidence. The visualization technique of the network analysis completes the analysis of the research field (Cobo et al., 2011a). This study uses VOSviewer for the network analysis that uses the VOS (visualization of similarities) algorithm. VOS algorithm visualizes similarities between objects (e.g., citations) based on their location, and the distance reflects their similarities (van Eck & Waltman, 2010). Network analysis allows for visualizations of the scientific fields in which network nodes represent units of analysis (e.g., documents, authors, journals, words), and network ties represent similarity connections, with the strongest drawn closer together (Zupic & Čater, 2015). Bibliometric tools and network analysis foster a unique understanding of the DL field's progress and the connected topics within this domain by making sense of unstructured data in rigorous ways and advancing the domain in meaningful ways (Donthu et al., 2021).
A DL bibliometric study is important to map the DL domain with its characteristics and evolution, to present how the conceptual items are organized, interrelated, and displayed, and to show the emerging trends of a research field (Donthu et al., 2021; Marsilio et al., 2011). Furthermore, DL is receiving increased attention (Zeike et al., 2019). This study can lead to a broader understanding of the DL domain by focusing on leadership related to digital transformation (Williams et al., 2021). Therefore, this study has several benefits to researchers and practitioners: revealing the evolutionary nuances of the DL field while shedding light on the emerging topics within that field (Donthu et al., 2021); assisting others in absorbing the growing body of DL studies through the application of transparent and replicable protocols; providing a comprehensive, holistic analysis at DL work; and facilitating potential research paths for future studies (Williams et al., 2021).
Finally, DL is worth studying as, in today's digital trends, the leadership role is paramount in propelling a fast decision-making process and promoting change (Jäckli & Meier, 2020). Studying the phenomena of DL is relevant because of its importance to organizations, as digital transformation requires leaders to discover new ways to thrive in times of uncertainty (Matzler et al., 2018). Therefore, the study presents a comprehensive review of the field using bibliometric tools and network analysis. It begins by showing the academic evolution of the field, the more influential papers and researchers, the most used keywords, and affiliation statistics. Then, a network analysis informs the evolution of the author's keywords over the years and the relationship between authors and academic publications. Also, it algorithmically identified three major clusters used to examine research topics through content analysis. Finally, additional insights and potential directions for future avenues are also presented.
Literature Review
Digital transformation is a key theme for businesses in this century (İnel, 2019) and is without precedent in relation to speed, range, and impact (Matzler et al., 2018). Due to technological disruption, organizations currently face enormous changes in the way work is done and how to lead (Schwarzmüller et al., 2018). The rise of digital technologies has accentuated the need to study the role of leadership, as incorporating digital technology into leadership study makes the topic even more challenging (İnel, 2019). Vial (2019, p. 118) defines digital transformation as “a process that aims to improve an entity by triggering significant changes to its properties through combinations of information, computing, communication, and connectivity technologies”. However, skilled human resources are at the heart of the digital revolution (Hanna, 2018). As leadership is context-specific, in times of uncertainty and complexity, prepared leadership is crucial (Bolden & Regan, 2016). As Cortellazzo et al. (2019) stated in the review about leadership and the digitalized world, leaders are crucial to develop a digital culture enabling collaborative processes in complex scenarios with ethical concerns.
In the late 1990s, the rise in information technology (e.g., internet, email, and videoconference) started to significantly transform work and leadership and established a new form to lead: e-leadership (Avolio et al., 2014). In response to those changes, many authors began to investigate the implications of information technology on leadership (Avolio & Kahai, 2003; Kissler, 2001; Pulley & Sessa, 2001). The term “e-leadership” started to be used to refer to leaders who carry out many of the leadership processes through electronic channels and their global reach (Zaccaro & Bader, 2003). One of the widely employed definitions of e-leadership states that it is a social influence process mediated by information technology to produce a change in attitudes, feelings, thinking, behavior, and/or performance with individuals, groups, and/or organizations (Avolio et al., 2000; Hambley et al., 2007; Jawadi et al., 2013; Jiang et al., 2017). Many different types of studies about e-leadership have taken place, such as in the governmental area (Banerjee & Chau, 2004; Rubino-Hallman & Hanna, 2007), SMEs (small-to-medium enterprises) (Belitski & Liversage, 2019), online game players (Lu et al., 2014), and even in the study of the migration decision-making process (Amit et al., 2016). Beyond that, there are several studies about information technology and the impact on organizations, focusing specifically on leadership challenges in a virtual environment (Anoye & Kouamé, 2018). In an e-leadership review, Avolio et al. (2014), inform that AIT technologies have not deeply modified the way leadership is studied or theorized, and it may be time to reevaluate what represents leadership in this scenario with ethical mindset. The ethical aspects remain a concern in the use of AIT, as ethical leadership in the virtual teams differs from traditional face-to-face interactions, and the e-leader need to discern moral dilemmas, evaluate risks, and protect privacy (Lee, 2009). The virtual work settings provide a high level of versatility, allowing organizations to pool staff worldwide, but it also challenges how to lead (Pradhan, 2019). E-leaders can empower followers to participate more in decision-making and becoming self-reliant working in virtual teams as they communicate effectively (Walvoord et al., 2008). Although communication aids in overcoming the uncertainties of the digital environment, it is also another challenge, as most of the exchanged information between leaders and followers is non-verbal, which is reduced in a virtual team (Maduka et al., 2018). Also, trust is a key aspect of any relationship and also a challenge for leaders due to the establishment of trust within the virtual team (Cascio & Shurygailo, 2003).
Nonetheless, the concept of e-leadership is changing, moving from an electronically mediated form of intercommunication to a broader view of digital technologies (Belitski & Liversage, 2019; Cortellazzo, Bruni, & Zampieri, 2019). According to this perspective, e-leadership is no longer about only electronic tools and communication, but mainly involves leading effectively in a digital environment (Roman et al., 2019). Digital technology is changing the context in which people work, increasing ambiguity, and the need for change (Pulley, Sessa, & Malloy, 2002). Therefore, many of the assumptions about leadership must evolve. Many challenges may appear to arise for the digital leaders, affecting their leadership effectiveness (Amit et al., 2016). The digital scenario is changing some skills, as the ability to lead networks instead of via hierarchy seems to become more critical because it facilitates fast and collaborative work, removing barriers and increasing agility (Kane et al., 2019). The digital scenario is also implicated in some leadership paradoxes. As technology may change how leadership is performed, some dilemmas appear. DL needs to provide autonomy to its teams without their feeling isolated; to preserve focus and purpose in an environment of continuous change that requires strong adaptability; and, also, to balance efficient well-known responses with new ideas and innovation (Pulley et al., 2002). This scenario leads us to question the structure of the DL domain and its evolution and what we can learn from it. Another question is if a new leadership paradigm is emerging from digital disruption. Using bibliometric tools and network analysis provides a comprehensive review of the field. It can assist in giving the summarization and connections within the research subject to conclude and learn from the complementary pieces of evidence.
Research Methodology
In this study, we use bibliometric analysis, which is all about summarizing past research to advance the lines of research knowledge (Zupic & Čater, 2015). Bibliometric analysis is a quantitative study of written publications (Broadus, 1987; Pritchard, 1969) that analysis the history of scientific works to measure impact (Lee et al., 2005). Following Fahimnia et al. (2015), we use a four-step methodology (determining the search keywords, data cleaning and formatting, initial analysis, and data analysis) to identify the higher impact studies and authors and to analyze the topics of research that contribute to insights for future research in the domain. In line with Ertz and Leblanc-Proulx (2018), the use of the two databases (Scopus and Web of Science) instead of one is an important contribution aiming to obtain more depth and variety in our results.
Determining the Search Keywords
First, we performed a search on the terms “e-leadership” or “virtual leadership” or “digital leadership” or “leading online communities” using the Scopus database, focusing on three main subject areas (i.e., social sciences; business, management, and accounting; and psychology) using article title, abstract, and keywords for the search. We decided to limit our search to articles published after 2000 (included), when Avolio et al. (2000) started to use the term “e-leadership” connecting leadership and technology. Then, we restricted our search by selecting only peer-reviewed journal articles in the English language, since these works are a body of certified knowledge and mostly suggest reliable results (Bhatt et al., 2020; Haleem et al., 2020). The Scopus database was composed of 173 articles. Afterward, the same search was done on the Web of Science database core collection, aiming to avoid missing articles and resulting in a sample of 86 articles, with the vast majority overlapped with Scopus.
Data Cleaning and Formatting
Due to the use of two different databases instead of only one, a higher amount of work related to the standardizing of the data is needed to obtain the consolidated information. Because many publications appeared in both databases at the same time, the duplicates were suppressed manually from the Web of Science data because their citation format was less complete than that of Scopus (Zhao & Strotmann, 2015). Also, the data retrieved from the bibliographic databases normally contain errors, such as misspelled elements, among others. Therefore, an analysis of the retrieved data is necessary (Cobo et al., 2011b). We conducted a review of the bibliometric information and corrected (e.g., in one article the authoŕs name was misspelled) or excluded articles with bibliometric problems. Then we read every abstract of the remaining documents. Our exclusion criteria follow Gümüş et al. (2019) and Soriano et al. (2018), as studies unrelated to the scope of our research, i.e., articles not related to organizational studies, were excluded. In this way, we excluded thirty-nine articles from non-organizational domain (e.g., educational, health areas), three articles that were non-English studies, and sixty-eight that the search terms were missing (e.g., the search results in the databases selected articles with the word “leadership” instead of “e-leadership”). We also confirmed if the article were within the years defined for the search (see the Appendix 1 for all excluded articles). Finally, 79 records remained in our database to be more deeply examined (Table 1). The decrease in the number of articles from the initial dataset were significant, although it is not unusual in bibliometric analysis, as it occurred in a variety of other studies (e.g., Figueroa-Rodríguez et al., 2019; Galvagno & Giaccone, 2019; Gümüş et al., 2019; Keathley-Herring et al., 2016).
Articles Belonging to the Research Dataset.
*Approach: T/E: theoretical/empirical; Quali: qualitative; Quanti: quantitative; MM: mixed method.
Initial Analysis
Figure 1 shows the publishing trend in the number of publications on the DL topic from 2000 to 2020. We can notice that from 2009 to 2011, the number of publications was practically stable. However, mainly from 2018 onwards, the increase rate was more pronounced. The volume of articles published significantly increased by 243% in the last decade (2011–2020); even in the recent 5 years (2016–2020), the growth rate was still more than 100%. This growth rate is in line with other studies such as Wei et al. (2021), in which the number of articles published increased by 211% in the last decade and by over 90% in the recent 5 years. Therefore, concerning not only the emerging subject (Zeike et al., 2019) but also the increased rate of the number of articles during this period, this preliminary result suggests that the topic will continue to attract significantly more research. The topic is still in the growth period and has not entered its maturity stage yet, in line with the conclusion of Ertz & Leblanc-Proulx (2018) and Davarzani et al. (2016) regarding the cumulative growth presented over the years.
Cumulative growth in the number of source documents.
The 79 articles in the research dataset published by 57 different journals are related to leadership, human resources, business, innovation, government, psychology, and strategy. The newness of the topic and the range of journals in which it is published confirms that DL has gained interest from several different areas. Such a dispersal of the literature, and the different perspectives it has allowed, confirm the need for structure and discussion about avenues of future research. Most articles were written by two (30.4%) or three authors (27.8%), with a maximum of seven authors writing a single paper. From a methodological perspective, empirical research represents 64.6% of the total articles in the dataset, being closely distributed between qualitative and quantitative studies and with mixed method representing 13.7% of the total empirical studies. These qualitative studies use methods such as general qualitative methods (interviews, content analysis) in 12 articles, case study in 6 articles, and observational in 1 article. The quantitative studies implement methods such as survey (17 articles), econometric model (4 articles), experimental designs (3 articles), and secondary data analysis (1 article).
Data Analysis
Data analysis is performed in two parts, regarding bibliometric analysis (section 4) and network analysis (section 5). Bibliometric analysis uses BibExcel because it allows for the analysis of diversified datasets and is the software most used for executing bibliometric analysis in management and organizations (Zupic & Čater, 2015). VOSviewer works efficiently with different databases (e.g., Scopus and Web of Science), providing visualization and analysis options (van Eck & Waltman, 2010). The visualization of similarities (VOS) algorithm presents similarities between objects (e.g., co-occurrence, co-citations) providing accurate visual information about the distance between any pair of objects (van Eck & Waltman, 2010). Also, content analysis followed the publication co-citation analysis to fully comprehend the research areas of each cluster.
Bibliometric Analysis
Leading Publishing Journals
The main journal repositories that published in the field of DL, five main journals respond for 19 studies, corresponding to 24% of all publications under study. These journals are Organizational Dynamics, Leadership Quarterly, Management Science Letters, Frontiers in Psychology, and Leadership and Organization Development Journal. As expected, among the most published studies, there are journals concerning organization and leadership topics as well as psychology.
Influential Authors and Affiliation Statistics
In terms of the number of authors’ local citations, i.e., the number of citations within the 79 publications inside our dataset, most authors cited work in the United States, as shown in Table 2. When considering the most cited authors, names such as Bass and Yulk, which are particularly impactful in leadership studies, appear in the list as expected.
The Most Cited Authors Based on the Number of Publications.
*Local publication: publication within the 79 publications of the dataset.
**Country: country of the author's affiliation.
Analyses of the author's affiliation show the international dimension of the research field studied (Cobo et al., 2011b). The research of organizations contributing to the DL literature was produced in 28 countries around the world. While most publications on DL are from the United States (39%), Europe accounts for 31.4% of the topical research, with the United Kingdom and France leading the European continent in this domain. The research is spread among 13 European countries. Following Europe, Asia accounts for 21.9% of the total organizations of the research, with Indonesia leading this continent, followed by China.
The California State University is the leading university regarding the DL topic. In second place, following the American university, are three organizations from Asia, namely, Bina Nusantara University (Indonesia), City University of Hong Kong (China), and KDI School of Public Policy and Management (South Korea), with four publications each. Although most of the DL literature comes from organizations in the United States, the most productive universities are spread through the United States (California State University and University of Southern California), Europe (Aix Marseille University, National University of Ireland, and University of Reading), and Asia (Bina Nusantara University, City University of Hong Kong, KDI School of Public Policy and Management, and Singapore Management University).
Keyword Statistics and Most Cited Articles
Co-word analysis is a technique that considers the text in the publications and presents joint occurrences among the main concepts related to a field (Cobo et al., 2011b). Usually, the words for this analysis are from author keywords, but they can come from document titles or abstracts (Donthu et al., 2021). A pool of over 336 author keywords was drawn from the 79 publications based on the number of occurrences, and data refinement was used (e.g., singular/plural) to obtain the final keywords (Davarzani et al., 2016). From Table 3, we find that most of the keywords correspond to the search criteria used (Soriano et al., 2018). The most frequent keywords are those related to e-leadership, digital aspects (e.g., digital leadership, digital transformation, digital technology), and virtual leadership and virtual teams.
The Most Popular Keywords.
Table 4 shows that the most cited articles are about e-leadership and virtual teams, including the seminal article about e-leadership from Avolio et al. (2000). The term “digital” is more recent and, although it appears in the list of most popular keywords, it is not yet seen in the titles of the most cited articles. Although citation is mainly a measure of impact, usually the highly cited papers have enough time to establish citations (Zupic & Čater, 2015). We acknowledge that these most cited papers are at least seven years old and that more recent papers are not contained in the set of articles most cited in our citation analysis. Moreover, the main subject areas concerning the global citations are business, management, and accounting; computer science; social sciences, and psychology. In conclusion, although the most popular keywords are related to e-leadership, virtual leadership and teams, and digital aspects, the term “digital” does not appear in the title of the most cited articles yet. Typically, the highly cited papers need some time until other studies use them as references.
The Set of Cited Articles Ordered by Number of Local Citations.
*Local citation: citation within the 79 publications of the dataset.
Network Analysis of Publications
Network analysis allows for visualizations of the scientific fields in which network nodes represent units of analysis (e.g., documents, authors, journals, words), and network ties represent similarity connections, with the strongest drawn closer together (Zupic & Čater, 2015). Mapping and clustering are complementary to each other. Mapping provides a picture of the structure of a bibliometric network restricted to two dimensions, and clustering does not have dimensional restrictions, but works with binary rather than continuous dimensions (Waltman et al., 2010). VOSviewer construct maps of authors or publications based on co-citation information and generate maps of keywords based on co-occurrence data (van Eck & Waltman, 2010). A co-occurrence network is the interconnection of terms based on their relatedness (van Eck & Waltman, 2011). After the relatedness of the unit of analysis has been determined, VOSviewer clustering technique assigns each topic to exactly one cluster, not having overlap of clusters, and there are no unit of analysis without a cluster assignment (van Eck & Waltman, 2017).
Co-Occurrence Analysis
Figure 2 shows the authoŕs keyword co-occurrence network to understand the keyword evolution over the years based on the frequency of words used in the articles belonging to the dataset. Thus, corresponding terms extracted from authoŕs keywords from Scopus and Web of Science dataset are used in this analysis (Soriano et al., 2018). The keywords co-occurrence networks examinations combine the unified approach to clustering and mapping bibliometric networks (Waltman et al., 2010). Recent studies employed this method as effective to perform clustering and grouping literature data graphically to explore trends (Prashar & Sunder, 2020). VOSviewer software was used to mining the Scopus and Web of Science datasets containing the bibliographic information of the articles without modifications. A threshold of keyword frequency was established to be 3 (Dai et al., 2020), which we found to properly cover the main keywords of most articles. Density based clustering employing full counting method and association-based normalization algorithm was used (Kriegel et al., 2011; Prashar & Sunder, 2020). The keywords that meet the established threshold were then mapped, and cluster and co-occurrence networks analysis was performed.
Author's keyword co-occurrence network from dataset publications.
From the chronological perspective, which was aimed at understanding the temporal distribution of the keywords, e-leadership, virtual teams, and virtual leadership are some of the keywords most used until 2016. The term “virtual”, focusing on the virtual aspect of leadership brought by the Internet and the new sorts of communication technology, seems to change to “digital”. Digital transformation and DL are newer keywords, reinforcing the digital aspects of market orientation from today's publications. These findings are in line with the co-word analysis from Table 3.
Regarding cluster analysis, four clusters were created regarding the authoŕs keywords. The first cluster have the terms “virtual teams” and “virtual leadership”, while the second cluster have the terms “e-leadership”, “ict”, “communication” and “leadership”. Digital aspects, such as “digital technology” and “digital transformation” appear in a third cluster, highlighting these terms together. Finally, the fourth cluster relates to market change, containing the keywords “digital leadership”, “dynamic capabilities”, and “market orientation”, suggesting a new market demand towards digital leadership and its capabilities in a changing business scenario. Therefore, we conclude that there is a trend in the use of the keywords. The keyword “ICT” (information and communication technology) appears together with “e-leadership” and “communication” studies, reinforcing the aspect that e-leadership is mediated by information technology. Nonetheless, the evolution of the keywords is transitioning from the e-leadership and virtual aspects to the digital term (e.g., digital technology, digital transformation, digital leadership), reinforcing todaýs market orientation due to the digital-oriented era.
Co-Citation Analysis
Figure 3 graphically shows the networks of co-citation relationship for authors produced by VOSviewer, and Figure 4 displays the publicationś co-citation relationship by VOSviewer as well. Nonetheless, the clustering was run independently for each map. Co-citation analysis is about the relatedness of items and is determined by the number of times they are cited together. It is a reliable measure of the impact of publications of a field in the scholarly community through the identification of its core works (Batistič et al., 2017).
Visualization of authors’ co-citation relationship.
Visualization of publications’ co-citation relationship.
The publications that are more co-cited are more likely to present similar subject areas having a strong co-citation relationship within each cluster, being the base of the different subfields (Cobo et al., 2011a). Clustering techniques are used to identify groups of related publications, authors, or journals (van Eck & Waltman, 2017). The VOSviewer mapping technique generated three clusters, each containing a group of articles with a similar co-citation profile. Therefore, each cluster has a probability of sharing the same theme or knowledge base. The size of the clusters ranged from 5 publications in cluster 3 to 11 publications in cluster 1. Afterward, we conducted a content analysis to determine the research focus in each of the three clusters. The articles in each cluster were scrutinized to identify the specific topics related to each cluster. Attentive analysis of the articles of each cluster can delineate the cluster's area of research focus (Fahimnia et al., 2015). Figure 4 indicates each cluster generated by VOSviewer. Cluster 1 is shown in red, cluster 2 in green, and cluster 3 in blue. Table 5 shows the publications in each cluster.
The Publications of Each Cluster.
Both clusters 1 and 2 seem to approach a virtuality component and aspects of trust in organizations (Jarvenpaa & Leidner, 1999; Thomas & Bostrom, 2010). It seems that the question of how to build and maintain trust is a concern with both virtual teams and leadership effectiveness. The papers in cluster 1 address the characterization of virtual teams and the challenges related to their dynamics and functioning in the organizations. Topics covered include the main aspects of virtual teams (Kirkman & Mathieu, 2005; Maznevski & Chudoba, 2000), different elements that drive groups to higher levels of team virtuality (Kirkman & Mathieu, 2005), and differences between virtual teams and face-to-face teams (Zaccaro & Bader, 2003). This cluster has the two oldest publications within the three clusters, considering electronic mail as a tool to speed up the exchange of information within teams (Sproull & Kiesler, 1986) and some aspects of information processing to reduce equivocality and uncertainty in interdepartmental relations (Daft & Lengel, 1986). Also, team virtuality (e.g., geographic dispersion, electronic dependence) can hinder innovation through different mechanisms, but a psychologically safe environment can help to overcome it (Gibson & Gibbs, 2006). Communication is important to the operation of effective virtual teams (Maznevski & Chudoba, 2000), and empathy for other team members is another aspect of highly effective virtual teams (Kayworth & Leidner, 2002). “E-teams” is labeled when interactions among team members occur through electronic communication (Zaccaro & Bader, 2003).
Cluster 2 draws on the leadership aspects and their challenges in managing virtual teams. Avolio et al. (2000), a seminal paper about e-leadership, is present in this publication selection. Also, transformational leadership is approached by comparing face-to-face and virtual aspects (Balthazard et al., 2009; Purvanova & Bono, 2009). A practitioner view is established, providing practical guidelines for the virtual team leader to build a strong sense of team identity and minimize uncertainty and ambiguity (Brake, 2006). Virtual teams need guidance but leaders are often not sufficiently prepared to lead effectively, as specific challenges of remote leadership are unclear (Hertel et al., 2005).
Finally, cluster 3 approaches the interrelation between technology and organization. Adaptive structuration theory (AST) is reported as a theoretical perspective of technology and change. AST can be a viable approach for studying AIT (advanced information technology) in organizations, as AST examines the technology process change and how people interact with technology (DeSanctis & Poole, 1994). Nonetheless, diverging perspectives about technology within organizations may have restrained the topic (Orlikowki, 1992). In this cluster, a notion of what constitutes e-leadership is broadened, and AIT is related to shaping the function of the organizations (Avolio et al., 2014).
In summary, the authors’ and the publications’ co-citation relationship analysis was run independently for each VOSviewer map. Co-citation analysis refers to the relatedness of items by the frequency they are cited together. Each different authors’ and publications’ groups are more likely to present similar subject areas. Cluster 1 presents papers about virtual teams and the challenges related to their dynamics and functioning in the organizations; Cluster 2 focuses on the leadership challenges for virtual teams; Cluster 3 addresses the interrelation between technology and organization.
Discussion
The findings show that an important part of the field is formed by empirical works showing the practical relevance of the domain. Such managerial work continues to grow and, together with the theoretical studies, spur additional research interest in the field. It appears that leadership effectiveness through virtuality was a major concern until digital technologies become more present in organizations. and which emergent leadership capabilities most impact virtual team effectiveness (Ziek & Smulowitz, 2014). More recently, the focus seems to migrate toward analyzing the impact of digital transformation in organizations (Peter et al., 2020). A content analysis of the dataset articles helped us better understand the critical DL capabilities needed to thrive in a digital scenario. It is used to get the most from the surveyed articles to frame the study's theoretical aspects. While bibliometric and network analysis reveal the publication patterns and their evolution, content analysis is complementary to them (Takey & Carvalho, 2016). Our findings concerning DL content analysis are in line with Kane et al. (2019), as while many core leadership capabilities remain the same, the unique characteristics of digital transformation requires new capabilities as well.
The fundamental capabilities normally related to leadership (e.g., communication, direction-setting) continue to apply, but are changing (Pulley et al., 2002). Communication seems to be a constant leadership concern over the years, as it is a topic of continuous discussion both at the beginning of e-leadership conceptualization (Purvanova & Bono, 2009) and now in the digital scenario (Darics, 2020). It is a fundamental aspect that appears in the list of the most used keywords. Communication is a foundational component of DL. It is a skill that provides team effectiveness (Ziek & Smulowitz, 2014), aiding leaders in influencing team interaction (Hambley et al., 2007). Transparency in digital times seems to be an important aspect of DL. Transparency helps leaders to communicate clear progress and problems toward establishing goals, thereby enabling team members to perceive the impact of their work on overall performance (Turesky et al. 2020). Leaders need to clearly indicate the strategic goals to their teams so they can experiment with new ways of doing things in a digital scenario (Kane et al., 2019). Also, DL needs to keep its team headed in the same direction, both intellectually and emotionally. The team needs clear information on what they are working on together, and also care that the work is worthwhile (Amit et al., 2016). A clear vision in a digital environment is crucial. The DL needs to anticipate trends and solve complex problems that are emerging due to technology and lead the team in response to those changes (Kane et al., 2019).
Trust is another essential part of leadership and is particularly relevant in a virtual environment. Building trust with team members is fundamental for effective DL (Campion & Campion, 2020). But it is also a challenge for DL. Trust is easier to build through face-to-face interactions rather than in a virtual environment, and it is a challenge for DL to rediscover how to lead teams based on trust rather than control (Maduka et al., 2018). Leading within a distance, leaders have little or no control over the group (Amit et al., 2016). Transparency also has a valuable impact on trust-building for a dispersed team (Liao, 2017). In virtual teams, trust helps each member better communicate with each other and encourages people to have initiative, and even take risks when performing their tasks (Liao, 2017). Also, recent literature indicates that building trust is one of the most important predictors of virtual team performance (Turesky et al., 2020).
The relationship that the digital leaders establish with their teams is crucial for a fast-changing environment. Relationship-building in DL helps to avoid the sense of isolation of followers and support the leaders in coping with diversity (Fernandez & Jawadi, 2015). DL is collaborative and supports a collective action (Ziek & Smulowitz, 2014). Speed is mandatory in a digital scenario, so leaders must have the ability to lead networks of people instead of leading through a linear hierarchy. A network approach generates a fast and collaborative leadership interaction with the team, facilitating the decision-making process, eliminating barriers, and helping the group increase agility and innovation in their work (Kane et al., 2019). Collaboration also helps the team overcome potential anxiety derived from working in a virtual setting and increasing their motivation (Liao, 2017). Nonetheless, a leadership aspect that has remained important over the years is empowering followers, as DL must enable its followers to pursue new initiatives (Kane et al., 2019).
Innovation and adaptability are topics that have strongly appeared more recently, perhaps because, nowadays, organizations may need to change continuously and quickly to thrive in a digital world. DL needs to be change-oriented and, therefore, open-minded, adaptable, and innovative (Kane et al., 2019). DL seems to have a key aspect in driving innovation throughout the organization as a competitive advantage (Doghri et al., 2020; El Sawy et al., 2016). Digital leaders must have an innovative mindset, allowing their followers to try new things, make mistakes, adjust, and scale (Kane et al., 2019). Figure 5 shows different DL capabilities concerning four elements from the content analysis. The DL seems to have an interpersonal orientation (i.e., interacting successfully with others), personal attributes (i.e., managing their inner selves), strategic focus (i.e., helps the organization achieve its future goal), and delivery-related aspects (i.e., the capacity of achieving the desired outcome).
Integrated elements for the DLs main capabilities.
The articles in the dataset revealed a new reality for leaders performing in the digital era. However, none offered a new definition necessary to understand the main aspects of the DL clearly. The articles used many different definitions concerning the topic. One of the most adopted definitions by many authors for e-leadership is from Avolio et al. (2000), that state that it is a social influence process mediated by information technology to produce a change in attitudes, feelings, thinking, behavior, and/or performance with individuals, groups, and/or organizations (Hambley et al., 2007; Jawadi et al., 2013; Jiang et al., 2017; Roman et al., 2019). Nonetheless, there has been little progress in the e-leadership concept from the year 2000 until now (Avolio et al., 2014; Liu et al., 2018). Therefore, more recent articles started to differentiate e-leadership from DL, using the DL concept instead. Some DL definitions that appear in the dataset are DL is doing the right things for the strategic success of digitalization in the organization and its ecosystem (El Sawy et al., 2016); a leadership perspective needed to successfully encounter challenges in the digital age (Jäckli & Meier, 2020); the human aspect of a leadership operating with digital tools in the virtual world (Narbona, 2016); or it is a combination of leadership capabilities and digital technology to support the decision-making process (Sasmoko et al., 2019). However, the existing DL definitions are vague about the main aspects of DL to thrive nowadays.
Although the link between e-leadership and DL is not clearly stated in the literature, maybe the difference is that the latter is not only mediated by technology but has evolved into a way of thinking and behaving in a complex time. The authoŕs keyword co-occurrence network (Figure 2) provides the evolutionary pattern of the most used keywords over the years. The keywords changed the focus on e-leadership, virtual teams, virtual leadership, and communication to a broader view concerning digital transformation and technologies to address the demanding aspects of market orientation and dynamic capabilities through the DL approach. This evidence reinforces the necessity of a new DL definition that addresses essential points for the DL in the current scenario. Therefore, we propose the following definition: DL is an ethical and agile mindset that quickly responds to changes and learns from them, fostering a trust-based culture that values people and its diversity, coaching them to collaborate and thrive in a digital scenario.
Finally, many academics rely on well-established leadership theories to explain the relationship between DL and the digital scenario (Boje & Rhodes, 2005; Cortellazzo et al., 2019). However, topics such as leading in the digital ecosystem, leading within different organizational structures, collaboration for co-generating ideas and strategies for digital change, dealing with a high amount of data, and artificial intelligence are some topics that have not received much attention so far. Perhaps different leadership theories and leadership capabilities may emerge due to the organizational characteristics of digital disruption. Therefore, the potential for an important contribution to the field of DL may lie there. Lastly, it may seem that the digital scenario brings a new leadership paradigm, i.e., DL, which has unique challenges related to digital technologies. As can be observed in the evolution of the keywords over the years, DL must manage the firm's dynamic capabilities while following a market orientation to thrive in the fast-changing and complex digital environment.
Conclusions, Limitations, and Future Work
This paper aims to analyze the DL domain's structure and evolution, learn from it, and verify if a new leadership paradigm is emerging from digital disruption. Although reviews of the literature on DL have been conducted (e.g., Cortellazzo et al., 2019), a bibliometric study coupled with a network analysis remained inexistent. In addition, we used two databases instead of one for maximum coverage. COVID-19 seems to be a major test for leaders worldwide, as it is not possible to predict what the new world will resemble (Dirani et al., 2020). The complex and fast-changing digital transformation scenario represents a challenge to leaders (Bartsch et al., 2021). Therefore, DL has become a key element in the effort to modernize organizations (Peng, 2021) as leaders experience new challenges due to digital disruption (Kane et al., 2019).
The results suggest that the interest in the DL field will continue to attract significantly more research, as it has not entered its maturity stage yet (Ertz & Leblanc-Proulx, 2018; Soriano et al., 2018; Zeike et al., 2019), stating the relevance and actuality of the theme. Moreover, the range of journals within our dataset shows that DL has gained attention from several different areas, revealing the breadth of the field. Although the publications are dominated by academic scholars within the research teams, the topic interests academics and practitioners. Perhaps this explains why empirical studies represent most of the total publications in the dataset. The continued integration of market data into academic studies is likely to benefit both practical and theoretical advancement. The affiliation statistics, which is the international dimension of the research field studied (Cobo et al., 2011b), show that the DL work was produced in 28 countries worldwide. Mainly, DL publications are from the United States, followed by Europe, with Asian publications not far behind. Co-citation analysis enables studying the past influences of a domain (Cobo et al., 2011b). We opted for a co-citation analysis for more coverage, although it is lacking in considering the most recent publications. Old publications collect citations over the years, favoring them over more recent ones. As expected, the results show that the more influential works are older. Nonetheless, in terms of the authors’ citation within the dataset, the authors most cited are from the United States, indicating the strong influence of American authors in the DL field.
Although the most cited articles are about “virtual” and “e-leadership”, speculating about the technology impact on leadership and teams, a keyword analysis detected, in either bibliometric analysis or network analysis, that the term “digital” has become more relevant in addressing the type of challenge facing organizations today. These findings reaffirm the idea stated by Belitski and Liversage (2019) and Cortellazzo et al. (2019) that the concept of e-leadership is changing from an electronically mediated form of intercommunication to a broader view of digital technologies. Therefore, it is becoming mainly about how to lead effectively in a digital environment (Roman et al., 2019). Therefore, it seems that a new leadership paradigm is emerging from the digital disruption we are facing.
The emergent DL paradigm seems to have leadership characteristics needed to thrive in a digital environment, such as communication, direction setting, transparency, trust, agility, collaboration, innovation, empowerment, and adaptability. Also, DL must manage the organizational dynamic capabilities while pursuing a market orientation to thrive in a rapidly changing business scenario.
This paper contributes to the literature by presenting a comprehensive analysis of the rising field of DL, examining the characteristics and relationships of the DL field of publications from a determined timeframe. While some leadership aspects seem to have changed due to digital transformation, not all have. We agree with Kane et al. (2019) that DL must have a combination of leadership skills identified with insights from the past, with the agile mindset to address the needs of a fast-changing environment. The study provides a broad and holistic perspective on the domain that has not been offered before. Also, it differentiates between literature reviews and meta-analyses, as it is extensive and evaluates the relationships between the publications within the area.
While our review is timely and includes the most recent publications, it is not without its limitations. Despite the rigorous procedure of our systematic review, we considered only peer-reviewed articles as these works are a body of certified knowledge and mostly suggest reliable results (Bhatt et al., 2020; Haleem et al., 2020). Nevertheless, a future review can include conference proceedings and other non-peer review manuscripts to expand the boundaries and analyze the nuances of DL's broad and recent phenomenon. Another limitation regards the sample size (79 articles). While we use two databases for higher coverage, and many similar studies in other areas use only one database for this type of analysis (Fahimnia et al., 2015; Galvagno & Giaccone, 2019), we may have missed some relevant articles.
The avenues for future research are numerous. Subsequently, literature reviews can consider expanding the keywords for different searches on the digital phenomena. Our bibliometric study showed that the digital transformation is affecting both leaders and organizations in a variety of ways and, therefore, future research will likely consider different topics and approaches. The field of DL may focus more on digital technologies (e.g., artificial intelligence, machine learning, internet of things) as a major backbone so that new leadership characteristics can emerge as the topic becomes more mature in organizations and the academic literature. With the increase in both AI-based organizational processes and leaders interacting simultaneously with humans and robots, a case study based on technology companies could analyze the impact of AI on the capabilities required by leaders in the context of human and non-human interaction as collaborative units. Moreover, another future study could focus on ethics and removing bias. The datasets used in machine learning systems need to ensure diverse representation to avoid bias amplification and have inclusion in mind (e.g., see Bolukbasi et al. (2016) as an example of gender bias in machine learning datasets). This future study could use a mixed methods approach to offer a holistic view of the phenomenon (Venkatesh et al., 2013). The findings may have implications for HR practice regarding training and hiring policies, as there are more machine-based human interactions. Hence, calling leaders to focus their attention on it. As leadership is context-related, another potential study regards the metaverse ecosystem and its impact on leaders working in this environment. In a recent study, Dozio et al. (2022) discuss design elements in a virtual environment to elicit distinct emotions in people. Suppose the interaction between leader and follower may occur with peopleś avatars in the metaverse. How might this setting affect building trust and connection between them? Maybe grounded theory could provide a methodological way to explain the manifestation of contextualized social interactions (Kempster & Parry, 2011). Further, exploratory and confirmatory studies can be used for these new study scenarios. Besides, future research can examine the evolution of the clusters identified in this paper for cross-validation and advancement purposes. It would be interesting to see whether some clusters disappear while others appear or whether they will fragment into smaller but more focused sub-areas of research. For example, new digital technologies can fragment cluster 3. Finally, bibliographic coupling analysis can be used in future work to study recent publications. Bibliographic coupling employs the number of references shared by two documents to compute their similarity. The more the references of the two articles match, the stronger their connection (Zupic & Čater, 2015). This type of study can increase the visibility of more recent publications in the field.
Footnotes
Acknowledgements
We would like to thank the encouraging words and comments from the Editor, Dr. Daan Stam, and the suggestions from the anonymous reviewers, together they provided valuable contributions to this manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We gratefully acknowledge financial support from FCT- Fundação para a Ciencia e Tecnologia (Portugal), national funding through research grant UIDB/04521/2020.
