Abstract
Over the past decades, distributed leadership has been extensively studied in educational settings, owing to its empirical association with school improvement, and enhanced student achievement. Consequently, researchers, policymakers, and those working on educational reform are increasingly focusing on adopting a distributed approach to school leadership. This review aimed to analyse the intellectual structure of the concept of distributed leadership within the field of school leadership over 35 years (1988 to 2023) by employing bibliometric methods to examine the meta-data of 572 selected documents. This review summarised the key features of the distributed leadership knowledge base by analysing its volume and citations trends, the geographical distribution of knowledge production, key scholars and their patterns of collaboration, and its underlying intellectual and conceptual structures. This review offers a resource for scholars by providing insights into the foundations and dynamic evolution of the knowledge base on distributed leadership. The review offers a roadmap for future research directions on distributed leadership, empowering scholars to explore new research possibilities, by building upon the existing knowledge base. Furthermore, this review makes a contribution to the ongoing development of distributed leadership within the field of school leadership. It offers an overview of the evolution of the scholarship on this specific leadership topic, and it offers greater understanding of the development of the research base on distributed leadership.
Introduction
Within the field of educational leadership the discourse has shifted from a consideration of conventional, hierarchical patterns of leadership, in which the school principal enacts a heroic one-person leadership (Adams and Velarde, 2021; Bolden, 2011; Harris et al., 2022; Tian et al., 2016), towards a distributed form of leadership, in which formal and informal leaders, such as principals, assistant principals, and teacher leaders actively participate in school decision-making, and innovative practices (Liu et al., 2022). The emergence of a distributed model of leadership has been heavily influenced by the early work of Spillane et al. (2001), which focused on the nature of interdependencies, and the co-performance of leadership practices in schools. Distributed leadership underscores the idea that multiple sources of leadership operate within any organisation, emphasising the critical aspect of the “leader plus” within organisational change (Spillane, 2006).
Over the past two decades since the work of Spillane et al. (2001) and Gronn (2000, 2002), distributed leadership has been extensively studied in educational settings (Bush, 2023a; Bush and Ng, 2019; Gümüş et al., 2018; Harris et al., 2022; Tian et al., 2016), particularly in Western contexts, such as the United Kingdom, and the United States, owing to its significance in supporting school improvement strategies, and enhancing student outcomes (Harris et al., 2022; Harris, 2008; Walker and Hallinger, 2015). Additionally, it has been shown that, under the right conditions, distributed leadership can positively impact teachers’ attitudes and performance, including their organisational commitment (Liu, 2020; Muthiah et al., 2019; Thien and Adams, 2021), self-efficacy and job satisfaction (Liu et al., 2021), motivation and empowerment (Bektaş et al., 2022), professional learning (Polatcan, 2024), well-being (Liu et al., 2022), and networking (Azorín et al., 2020).
Following the work of Harris (2008), distributed leadership has been conceptualised as a shared, collective, and extended model of leadership practice, where interdependent interactions are more prominent rather than independent or individual actions. In line with Spillane's (2006) seminal theoretical work, this model of distributed leadership also highlights the importance of multiple sources of leadership within an organisation, which is a vital aspect of organisational change. It is important to distinguish distributed leadership from shared leadership, as these terms are often used interchangeably but have distinct theoretical foundations.
Distributed leadership focuses on how leadership is spread across multiple individuals and contexts within an organization, emphasizing interactions, interdependencies, and collective agency (Spillane et al., 2001; Gronn, 2002). It is concerned with how leadership is constructed through the collaboration of formal and informal leaders, emphasizing leadership as a practice rather than a role (Harris and DeFlaminis, 2016). Distributed leadership fosters a fluid and dynamic form of leadership, where various stakeholders, including teachers, administrators, and external partners, contribute to school leadership based on their expertise and situational needs (Harris and DeFlaminis, 2016).
In contrast, shared leadership refers to the distribution of leadership roles and responsibilities within a team, highlighting the collaborative nature of leadership within a team (Pearce and Conger, 2003). This model emphasizes team collaboration, where members share decision-making power and influence. Shared leadership is commonly seen in smaller, project-based teams, where leadership roles are distributed among team members to achieve a specific goal (Carson et al., 2007). The key distinction is that shared leadership often applies to team dynamics, whereas distributed leadership takes a broader organizational perspective, involving multiple stakeholders at different levels of the organization (Spillane, 2006; Bolden, 2020).
Distributed leadership has become widely promoted and endorsed in educational policy circles globally, largely because it acknowledges the diverse expertise and capacities of formal and informal leaders within schools, fostering collaborative decision-making, and ultimately, focusing on the practice of leadership rather than the role of responsibilities of the leader (Harris et al., 2022).
In order to have a more comprehensive overview of distributed leadership, this review analysed the intellectual structure of the literature on the topic over 35 years (1988 to 2023) using a bibliometric analysis approach. To our knowledge, this is one of the first comprehensive review that employs bibliometric analysis to systematically examine the evolution, trends, and key contributions within this particular knowledge base. The specific research questions that guided this review were:
What is the volume, citation level, and geographic distribution of publications on distributed leadership between 1988 and 2023? What are the patterns of collaboration among authors, institutions, and countries in publications on distributed leadership between 1988 and 2023? What is the intellectual structure and scholarly influences of the distributed leadership knowledge base between 1988 and 2023? What is the conceptual structure (including topics, foci, and frequently used concepts) within the knowledge base on distributed leadership between 1988 and 2023? What are the latest developments in distributed leadership research?
The bibliometric review provides insights into the broad development of the evidence base concerning distributed leadership in schools, ultimately enabling researchers to comprehend both the current state, and future direction of the scholarship on this topic.
Method
This article utilised a bibliometric review approach, which is a variant of the systematic review process. While similar to systematic reviews in selecting papers, it serves a distinct purpose. Bibliometric reviews map how various papers or studies on a specific concept were published and cited through a bibliometric lens. This method provides a review that is impartial, unambiguous, and empirically based, examining the trends of knowledge production in a particular field of study (Aung and Hallinger, 2022). Unlike other review methods, such as scoping, integrative, or systematic reviews, which aim to evaluate and synthesise findings through critical appraisal to highlight gaps in the literature (Hallinger and Wang, 2020), bibliometric reviews focus on bibliographic data to identify influential works, key scholars, and emerging trends (Donthu et al., 2021).
The search process was directed by PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) (Moher et al., 2009). PRISMA is a set of guidelines that provide a framework for researchers to report on systematic reviews and meta-analyses in a transparent, complete, and standardised manner. Adhering to the PRISMA guidelines increases the transparency, reproducibility, and accuracy of the reviews and meta-analyses, and helps readers to assess the quality and validity of the findings. As such, PRISMA was employed to screen and select the records for analysis (see Figure 1), which was generated using the tool developed by Haddaway et al. (2022). The review process comprised three distinct stages, namely, identification of sources, inclusion and exclusion criteria, and extracting and analysing the data.
PRISMA flow diagram detailing steps in the identification and screening of source documents for the bibliometric analysis of distributed leadership research (Moher et al., 2009) (colour online).
Identification of sources
Scopus was selected as the only data set for this bibliometric review as it is among the most reputable databases, with publications in a wide range of fields, including social sciences, education, and management. Moreover, its bibliographic data can be exported to a bibliometric software for analysis (Pan et al., 2023). Researchers have also pointed out that Scopus provides a more comprehensive and superior coverage of available high-quality work on a particular scholarly subject (Gümüş et al., 2021). It is noted, however, that Scopus documents tend to be primarily written in English. Consequently, the results from the bibliometric analysis in this article must be interpreted with this limitation in mind.
The search string on the Scopus database was developed after all relevant keywords were determined. A total of 2334 articles were initially retrieved using the search string shown in Table 1. Although a specific date range for the search was not set, it included any publications up to the current year that contained the specific search terms in the title, abstract, or keywords, with the first publication appearing in 1988.
Search strings.
Inclusion and exclusion criteria
Various inclusion criteria were used at the initial stage of this review (see Table 2). First, in terms of the topic relevance, documents needed to centre on, or at least include distributed leadership from the perspective of principals, or others within the school. Second, the nature of the distributed leadership research was contextualised to that within schools and preschools only. Thirdly, the selection of source types comprised Scopus journal articles. Fourth, the timescale between 1988 to 2023 was selected. Fifth, the selected articles had to be peer reviewed for assured quality, and finally, all documents selected for this review needed to be written in English.
The inclusion and exclusion criteria.
Exclusion criteria were established to ensure only specific and relevant documents were selected for this bibliometric review. First, books, book chapters, theses and dissertations, conference papers, data, media, and editorials were excluded. At this step, 356 documents were excluded. Second, documents were screened to examine their language (e.g., whether the main text was in English, not just the titles, abstracts, keywords, and references), and quality (e.g., whether they were peer-reviewed, and centred on, or at least partially focused on distributed leadership). Third, documents of other types, such as conference proceedings, conference reviews, theses, dissertations, books, book chapters and editorials were excluded. At this stage, 1406 more documents were ruled out. This screening and assessment process yielded a total of 572 journal articles for review.
Data extraction and analysis
An original dataset was extracted, comprising information that included author names, document titles, publication years, journal sources, citation counts, author affiliations, abstracts, and keywords. The data files were screened by two of the authors independently, based on the inclusion and exclusion criteria. The final dataset (572 articles) was subsequently organised and exported to an Excel file (csv format). Descriptive and bibliometric analyses were conducted. Scopus analytical tools were utilised to provide information on the volume of publications, and citations relating to distributed leadership articles published between 1988 and 2023.
Additionally, the science mapping analyses in this review were performed using the VOSviewer software. The use of VOSviewer software enabled an analysis of a large database of documents, providing insights such as geographical location, authorship, patterns of collaboration, intellectual structure, topics, foci, and frequently used concepts in distributed leadership. This approach offers a broader understanding of a knowledge base and concepts that may be challenging through systematic reviews of literature (Hallinger and Kovačević, 2021). ‘Science maps’ that visually represent the networks of the various dimensions of research carried out by researchers and scholars who share similar themes within the distributed leadership literature were produced. A minimum threshold of 15 citations was set so that the VOSviewer software could generate a relationship map with sufficient authors, and citation numbers. While setting such a threshold may result in the exclusion of less-cited authors, it would not significantly alter the overall results, as key authors in each cluster would still be determined by the highest citations. Particularly, co-authorship (including country of origin, and authors), co-citation (author, and source), and keyword co-occurrence analyses were conducted on the elicited data (Hallinger, 2019; Van Eck and Waltman, 2017).
The co-authorship analysis showed the collaboration among authors from various countries alongside their level of productivity (Bunjak et al., 2022). Co-citation analysis was performed to identify the frequency of two documents, authors, or journals that were often cited together in the distributed leadership literature (Small, 1973), providing a useful ‘visual illustration’ of the intellectual structure of this leadership topic (White and McCain, 1998). This analysis helped identify patterns of collaboration and influence within the field, providing insights into the relationships of ideas and the interconnectedness of researchers. For instance, when two referenced articles appeared together in a citation list, they are considered to be co-cited (Batistič et al., 2017). Additionally, citation analysis was utilised to determine the citation count of the documents in the dataset to identify the most influential publications. Lastly, co-occurrence analysis for keywords identified the general topical foci by highlighting commonly used keywords in the reviewed documents, thereby providing insights on the conceptual structure of the literature.
Findings
The findings from this bibliometric review are presented according to the research questions.
Volume of publications and citations
The first research question was to identify the volume, citations, and geographic distribution of the selected publications on distributed leadership between 1988 and 2023. Figure 2 depicts the results of the analysis. The findings indicate that the literature concerning distributed leadership or the influence of leadership when shared, first appeared in 1988. However, the topic only received modest attention in the subsequent 17 years that followed. From 2005 onwards, however, there was a gradual increase in the number of publications, steadily rising from one to three publications per year in the previous period (1988–2004) to around 10 publications per year. In total, publications on distributed leadership in the last 18 years (2005–2023) accounted for about 97% of the total volume, more than 17 times the amount in the first 15 years (1988–2004). Since 2020, 200 articles on distributed leadership have been published, representing 35% of the total literature over the past 35 years.
Number of publications and citations on distributed leadership between 1988 and 2023 (colour online).
Observations on the pattern of average citations throughout revealed low annual rates, with about two citations per paper. In calculating the average citations per paper, the total number of citations received by papers published in a specific year was divided by the total number of papers published in that same year. Subsequently, that result was divided by the number of years after publication. This provided us with the average number of citations per paper per year after publication (Leydesdorff and Bornmann, 2011). The highest normalised citations were by papers published in 2003, 2005, 2008, and 2010 when distributed leadership was at the height of its popularity.
Despite the overall declining pattern in citation levels since 2018, it is notable that the volume of publications on distributed leadership increased steadily in 2016, 2020, and 2023. Also, there was an increase in average citations in 2023 compared to the previous decade (2014–2022). This could be explained by the fact that, during COVID-19, leadership was distributed by default as schools faced unprecedented challenges that required collective and adaptive responses (Harris and Jones, 2020). This shift likely contributed to an increased scholarly focus on distributed leadership during this period. In addition, the increased volume of publications can sometimes dilute the impact of individual articles, resulting in lower citation rates on average (Larivière and Gingras, 2010). Overall, this trend suggests an increasing research interest in the topic of distributed leadership since 2001, which is in keeping with other reviews of distributed leadership over this period (Harris et al., 2022).
Geographic distribution
Next, the geographic distribution of the 572 publications on distributed leadership from the search was investigated. The analysis showed that authors from a total of 59 countries contributed to these publications (see Figure 3). Authors from the United States of America (USA), United Kingdom (UK), Australia, and Canada were the top three contributors to the publications on distributed leadership. Amidst all these Western countries, researchers from South Africa, Turkey, Hong Kong, Malaysia, and Singapore were also noteworthy contributors.
Geographic distribution of publications on distributed leadership (colour online).
Patterns of collaboration between countries
To address the second research question, the patterns of author collaboration within the publications in the bibliometric data file were explored. Co-authorship analysis based on the countries using the VOSviewer software was employed. Figure 4 shows the network of collaborations that exists among the countries in these publications. For the analysis, the minimum number of documents was set at three, and a full counting method was employed. Out of 33 countries, 29 had some degree of author collaboration, and were included in the network visualisation shown in Figure 4. The size of the nodes indicates the strength of the collaboration (Ellegaard and Wallin, 2015).
Collaboration between countries in publications on distributed leadership (colour online).
According to Figure 4, the three biggest collaborators were authors within the UK, USA, Canada, Finland, and Hong Kong. Interestingly, collaborators who worked together most were also among the biggest contributors of publications in total on distributed leadership. Moreover, despite authors from South Africa, Malaysia, and Spain being among the top contributors, they were not identified as having significant author collaboration. In addition, authors from Australia and China, who were portrayed as notable collaborators, were not among the top contributors to publications on distributed leadership.
We acknowledge that setting a minimum requirement of three publications for our analysis might disproportionately represent locations with better access to collaboration opportunities and favor larger nations or institutions with more researchers. Lowering this criterion could reveal newer and emerging patterns of collaboration, although the significant collaborators would likely remain unchanged.
Intellectual structure of distributed leadership knowledge base
Next, three techniques were utilised to investigate the intellectual structure of the knowledge base concerning distributed leadership in both school and preschool settings. The analysis incorporated highly cited documents, the most co-cited sources (journals), and the most co-cited authors. Accordingly, Table 3 shows the 18 most influential documents from the analysis, arranged from highest average citation count per year (CPY) after publication to least. It is observed that the top two most influential publications (Leithwood et al., 2010; Supovitz et al., 2010) were both focused on the influence of shared leadership on student learning and achievement. A review of the rest of the titles also revealed that a significant number of articles among the top 18 focused on the relationship between shared or distributed leadership, and measures of student or school achievement (e.g., Hallinger and Heck, 2010; Heck and Hallinger, 2009).
Eighteen most highly cited articles on distributed leadership.
*citations per year.
It was also noted that the articles in this list were either theoretical (Tian et al., 2016), conceptual (Mayrowetz, 2008), or review papers (Woods et al., 2004). Notably, the articles by Harris (2008) had taken a critical stance on distributed leadership, which contrasts with the normative stance generally adopted by many other authors in the field during the surge of publications on the topic post-2001 (e.g., Harris, 2008; 2013; Harris and Spillane, 2008). Harris (2008) described distributed leadership as leadership shared among many, focusing on teamwork rather than individual efforts. This interpretation highlights the idea that organisations have multiple leaders, which is important in executing organisational changes.
Next, VOSviewer was engaged to conduct a citation analysis based on the sources (cited journals), the results of which are shown in Figure 5. The number of documents published in a given source was set to at least three, where 38 sources met the criteria. The size of the nodes represents the number of citations of a source (Ellegaard and Wallin, 2015). According to Figure 5, the most cited source was Educational Administration Quarterly, which had been cited 2337 times (from 30 documents) while the next most cited source was Educational Management, Administration, and Leadership, which had been cited 1855 times (from 59 documents). With a total of 803 citations (from 35 documents), the Journal of Educational Administration was the third most cited source. School Leadership and Management comes next, with a total of 474 citations (from 29 documents).
Citation network of cited sources (colour online).
As the final technique to understand the intellectual structure of the knowledge base on distributed leadership, co-citation analysis was employed using VOSviewer with a minimum number of 20 citations per author. 87 authors met the threshold (see Figure 6). This meant that even the smallest nodes represented authors who had been cited 20 times, while larger nodes represented those with higher citation counts. The lines connecting each node represent co-citation by linking authors, where links that were more intense in colour suggest increased frequency in co-citations among the authors while the proximity of the nodes indicates similar topics or themes researched by the authors.
Co-citation network of cited authors (colour online).
The intellectual structure of the distributed leadership knowledge base relates to the key themes of the research in this field (Aung and Hallinger, 2022), which were identified when analysing the articles in this review. These key themes were derived from clusters that address specific areas of interest within the field. The analysis used bibliometric data on author co-citation to provide a valuable overview of the intellectual structure of distributed leadership, highlighting areas of interests, limitations, and avenues for future research.
The author co-citation map (see Figure 6) yielded four clusters representing four varying themes, or “schools of thought” (Kovačević and Hallinger, 2020, p. 10), with the red cluster having a total of 34 authors, the green cluster with 23 authors, and the blue and yellow clusters having 18 and 12 authors respectively, amounting to 87 authors. In this co-citation network, the highest and most significant co-cited author was Harris from cluster 2 (in green), who had a total link strength of 35,561 while Leithwood from cluster 1 (in red), who had a total link strength of 30,090, was second. Similarly, the third and the fourth highest co-cited authors were Spillane (total link strength = 29,198), and Gronn (total link strength = 15,571) respectively, both of whom belonged to cluster 3 (in blue).
These results suggest that the work of the major schools of thought on distributed leadership revolves around the work of Harris (2008), Leithwood et al. (2010), and Spillane (2005). The map also indicated the themes for each of the four clusters – distributed leadership and outcomes (red cluster), school improvement and change (green cluster), contextual factors and distributed leadership (yellow cluster), and roles in distributed leadership (blue cluster).
Distributed leadership & outcomes
The largest cluster (red) revolved around the theme of distributed leadership and outcomes, including mostly internationally well-known school leadership researchers, such as Leithwood (756 citations), Hallinger (524 citations), Louis (281 citations), and Murphy (240 citations). These researchers emphasised the importance of shared leadership practices in fostering a positive school climate, improving student achievement, and promoting organisational effectiveness. By empowering teachers, administrators, and other stakeholders to take on leadership roles, and contribute their expertise to school improvement efforts, their work proposed that distributed leadership could create a sense of collective ownership and accountability for student success. This cluster underlined the transformative potential of distributed leadership in driving positive educational outcomes, and fostering continuous improvement in schools. Harris and Jones (2020) found that schools employing a distributed leadership model were better positioned to manage the rapid transition to remote learning during the COVID-19 pandemic. Similarly, Azorin et al. (2020) observed that distributed leadership helped schools continue functioning under unprecedented conditions.
School improvement & change
Researchers in the green cluster focused their research on school improvement and change. Research in this school of thought was led by Harris (1012 citations), Fullan (249 citations), and Hargreaves (218 citations). Within this school of thought, researchers emphasised the significance of shared leadership practices in fostering collaborative efforts to address school challenges and drive meaningful change. Their collective work suggests that distributed leadership, which involves various stakeholders participating and leading decision-making processes, is essential for building capacity, promoting innovation, and enhancing organisational effectiveness. It also suggests that under the right conditions, distributed leadership allows schools to adapt to changing needs, respond effectively to challenges, and sustain continuous improvement efforts over time (Harris, 2008). Bush and Glover (2014) found that schools with shared leadership, where teachers are involved in decision-making, were more likely to sustain improvements. Similarly, Hulpia et al. (2011) showed that teacher involvement in decision-making increased organizational commitment were crucial to foster a positive school climate and sustaining reform.
Contextual factors & distributed leadership
The yellow cluster comprised authors who have explored the influence of contextual factors on distributed leadership. Leading this cluster were Hopkins (254 citations), Day (241 citations), and Bush (190 citations). Authors in this cluster researched the contextual factors that exerted a profound influence on the implementation and efficacy of distributed leadership within organisations. They note that the prevailing organisational culture, characterised by norms, values, and beliefs, either fosters or hinders distributed leadership practices. They also highlight how flexible institutional structures can promote decentralised decision-making to facilitate the distribution of leadership roles and responsibilities. Additionally, they note how external factors like government policies and community expectations shape the conditions under which distributed leadership operates. Hallinger and Heck (2009) highlighted how organizational culture, policy frameworks, and the broader educational context shape distributed leadership in schools.
Roles in distributed leadership
Authors leading in the blue cluster included Spillane (948 citations), Gronn (443 citations), Diamond (305 citations), Halverson (298 citations), and Woods (249 citations), with particularly significant contributions to understanding the roles of those within a distributed leadership framework. Their research emphasised the importance of clarifying the various roles that individuals play in distributed leadership processes within educational organisations. They posit that distributed leadership involves the delegation of leadership responsibilities across multiple roles and positions, including formal leaders, teacher leaders, administrative staff, and even students and parents. These researchers underscore the complexity of distributed leadership, illustrating how different roles interact, collaborate, and contribute to organisational goals. Harris and DeFlaminis (2016) highlights the complexity of distributed leadership, where teachers lead within their classrooms and departments, influencing both students and colleagues.
Conceptual structure of distributed leadership knowledge base
The fourth research question explored the conceptual structure of the knowledge base on distributed leadership. This involved exploring the themes embedded in the keywords associated with the reviewed publications. For this keyword analysis, only author keywords that had appeared a minimum of three times were included. To obtain a more representative network, (i) names of countries and places, and (ii) terms associated with specific research methods, and analytical techniques were excluded from the analysis.
Figure 7 shows the network generated by VOSviewer for the keyword analysis. As depicted in the figure, the conceptual structure of the distributed leadership knowledge base in education is composed of several major clusters/themes. Instead of reporting all the clusters, only those with the most frequently used keywords are emphasised here. The most frequently used keyword was ‘distributed leadership’, which occurred 240 times, and belonged to the red cluster. Other keywords in this cluster include ‘professional development’, ‘educational administration’, ‘educational policy’, and ‘teacher education’.
Thematic connections of author keywords (colour online).
The second most repeated keyword was ‘leadership’, which occurred 73 times, and belonged to the blue cluster. Other keywords in this cluster include ‘schools’, ‘school culture’, ‘education policy’, ‘management’, and ‘change’. With a frequency of 42, ‘teacher leadership’ was the third most used keyword, forming a cluster by itself. Although this keyword had been used in association with 39 other keywords, they were all from various clusters.
The fourth and the fifth most repeated keywords were ‘shared leadership’ and ‘school improvement’, occurring 39 and 38 times respectively. Additionally, ‘middle leadership’, an emerging area of research, had 8 occurrences. Similar to the term ‘teacher leadership’, these keywords were presented as single-keyword clusters. These results suggest a substantial conceptual overlap between teacher leadership, shared leadership, and school improvement in the distributed leadership literature.
Latest developments in distributed leadership research
To answer the final research question, VOSviewer was employed to generate a temporal co-occurrence map of the most used keywords, shown in Figure 8. This map provides valuable information in terms of the usage of different keywords in the reviewed documents over time. From the map, it is evident that, in recent years, several concepts have started to receive more attention from distributed leadership scholars while others have become less prominent.
Temporal keyword map of author keywords used in distributed leadership literature.
Keywords in yellow are those that have been used more frequently within the distributed leadership literature within the last ten years. These keywords included those denoting particular theories or forms of leadership, for example, ‘shared leadership’, ‘instructional leadership’, ‘transformational leadership’, ‘leadership for learning’, and ‘principal leadership’. Similarly, keywords such as ‘middle leaders’, ‘middle leadership’, ‘teacher autonomy’, and ‘teacher job satisfaction’ seemed to form a group of trending keywords. Other keywords included ‘school culture’, ‘social justice’, ‘inclusive education’, and ‘teacher education’. Meanwhile, keywords such as ‘educational administration’, ‘management’, and ‘governance’ were used less frequently in recent literature within the field. These keywords suggest a shift in the power base of leadership within schools, reflected in distributed leadership practices, but also in the use of terms like shared or teacher leadership. This shift is significant because it signals greater scholarly interest in leadership as practice rather than leadership as a role within the field of educational leadership.
Discussion
The purpose of this review was to analyse the intellectual structure of the distributed leadership literature from over three decades (1988 to 2023) using a bibliometric analysis approach. This section interprets the findings according to the research questions, namely, the total volume, citations, and geographic distribution; patterns of collaboration; the intellectual structure; conceptual structure; and the latest developments in distributed leadership research. Several findings are given emphasis in this section.
First, the growth trajectory of the literature on distributed leadership signals a significant shift in scholarly attention to collective or collaborative leadership practices over the last two decades. Prior to 2001, the literature on leadership as collective or shared influence was emerging but not dominant within the field of educational leadership. This growth in the empirical attention on distributed leadership post-2001 can be attributed to the fact that distributed leadership as a concept was not fully recognised within the leadership field until the significant and seminal publications on distributed leadership by Gronn (2000, 2002) and Spillane et al. (2001; 2004). The substantial increase in the literature post-2001 suggests that Gronn (2000, 2002) and Spillane et al. (2001; 2004) provided a strong theoretical position for the conceptual development, and empirical testing of distributed leadership in practice. Through the work of Spillane and colleagues, greater recognition was given to the pivotal role that formal and informal leaders play in securing school improvement. It also signals greater interest in and exploration of the evolving dynamics of distributed leadership roles and practices (Bush, 2023b; Harris et al., 2022; Liu et al., 2022).
Second, the total number of publications on distributed leadership reached its highest point in 2023, surpassing all previous years. As noted earlier, the COVID-19 pandemic significantly reshaped leadership practices in schools, marking a departure from traditional hierarchical structures in favour of a flatter, more decentralised leadership culture to cope with the demands of working at a distance (Harris and Jones, 2020). During the pandemic, the focus shifted from individual leadership actions to collaborative interactions, promoting joint activity, and collective practice among educational leaders. Distributed leadership at this time, therefore, emerged as a crucial framework for interpreting how leadership is exercised in complex and disruptive situations (Harris and Jones, 2020). The pandemic necessitated a transition towards distributed leadership as an essential primary mode of operation for school leaders, emphasising capacity building over control, and mobilising collective engagement and action among stakeholders (Azorín et al., 2020).
Third, although distributed leadership has drawn some academic attention in Africa and Asia (e.g., Grant, 2017; Kılınç et al., 2024; Sibanda, 2018; Thien and Adams, 2021), the distribution of the literature highlights a concentration within Anglo-American societies, such as the USA, UK, and Canada, possibly because the leading researchers writing about distributed leadership, as shown in this review, reside in those countries. This pattern of distribution aligns with reviews from other related areas of school leadership (e.g., Adams et al., 2023; Gümüş et al., 2021; Hallinger and Kovačević, 2021). The results indicate that distributed leadership is still very much a ‘Western’ concept, and that its practices have not yet received adequate attention in other regions and contexts although this pattern is gradually changing (Harris et al., 2022). This review emphasises the need for more research in different contexts to ensure a comprehensive understanding of distributed leadership practices in schools globally (Hallinger and Kovačević, 2021).
Fourth, findings from this review of the evidence relating to distributed leadership research (see Table 3) indicate a clustering of papers in the decades directly following Spillane et al. (2001). Another review of the contemporary evidence base confirms sustained scholarly interest in distributed leadership from 2013 to 2018 (Leithwood et al., 2020). This assertion is corroborated by Gümüş et al. (2018, p. 37), who identified distributed leadership as “the most researched model of leadership”. Overall, review findings suggest an interesting juxtaposition in the distributed leadership research landscape. While older papers continue to exert an overly dominant influence within the field, and are highly cited (e.g., Leithwood et al., 2010; Supovitz et al., 2010; Harris, 2008; 2013; Harris and Spillane, 2008), there is still a great deal of contemporary research interest in the topic, but without the same citation power as this review has shown (Harris et al., 2022; Tian et al., 2016).
This review indicates that researchers are still actively exploring and engaging with distributed leadership concepts, generating fresh insights, and contributing to the ongoing evolution of the field. The scholarly interest in distributed leadership highlights its enduring importance and relevance in educational leadership research despite a reduction in citation levels in the past decade.
Limitations
Limitations in this bibliometric review are duly acknowledged. First, the full literature search yielded a database of 2334 English sources. However, this review focused solely on journal articles, thereby reducing the size of the dataset to 572 journal articles in the English language. Thus, this review's findings may not represent the full distributed leadership literature and should be interpreted with caution as it also excludes books, book chapters, unpublished master's theses, and doctoral dissertations. Nonetheless, it is posited that this review presents a reasonable representation of the distributed leadership literature, providing a sound starting point (e.g., topics, authorship, citations, concepts, etc.) for identifying future directions for further developing the knowledge base.
Second, as an international research database such as Scopus relies heavily on English sources, the study does not capture 100% of the scholarly work on distributed leadership literature, particularly those in non-Western and non-English contexts. Therefore, future reviews on the topic might include literature from local databases, and those in their own native languages to better uncover the related hidden literature. Third, a minimum citation threshold of 15 was applied, potentially introducing bias by excluding newer or less-cited works. Similarly, a minimum threshold of three papers was set for co-authorship analysis, which might favour larger nations or institutions and overlook new collaborations.
Additionally, a minimum citation threshold of 20 was applied for co-citation analysis, which may favour well-known authors and exclude newer contributors. However, these thresholds were necessary to maintain the robustness and clarity of the analysis, as lowering them would not significantly change the overall results. In addition, while the co-citation network of authors was analysed to identify key scholars and understand the broader intellectual structure and collaboration patterns within the field of distributed leadership, future research could incorporate paper-level co-citation analysis to complement the current findings and offer a more detailed perspective on thematic evolution within the literature.
Implications and future directions
This current study highlights a serious imbalance in the intellectual structure of the knowledge base on distributed leadership. In summary, research on distributed leadership is disproportionately represented in the USA, UK, and Canada. Studies in the African continent remain limited, and research interest on the topic in Asia is only found in Turkey, Hong Kong, and Malaysia. The concentration of research in such countries highlights a potential cultural bias, and a limitation to the knowledge base on distributed leadership in its entirety. This limitation calls for a more concerted effort in other countries and contexts to explore distributed leadership in a critical and applied way. It is suggested that education researchers in non-Western societies and cultures need to contribute to the field in ways that respect and highlight differences and departures from the existing assumptions made in the distributed leadership knowledge base (Gümüş et al., 2021).
Conclusion
This review sought to present one type of analysis of the distributed leadership literature, and to illuminate key trends in its knowledge production process. This review offers guidance to researchers by providing data maps that can assist scholars in framing directions for future research on the topic of distributed leadership. It also highlights potential avenues for distributed leadership research in future years. The review did not seek to summarise or synthesise the findings reported in the corpus of studies represented here, but merely provides signposts for future directions in this field of scholarship.
It is hoped that these findings will serve as a catalyst for future exploration by researchers on this important topic, creating avenues for continued research on distributed leadership practices in schools. It is anticipated that this review will be useful for researchers in focusing their future research efforts, and by so doing, facilitate the strengthening of and contribution to the development of the distributed leadership knowledge base in educational leadership and management.
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
