Distributed leadership: taking a retrospective and contemporary view of the evidence base

ABSTRACT

Distributed leadership is one of the most influential and well-discussed ideas to emerge in the field of educational leadership. Prompted by the foundational and seminal work of Spillane et al. (2001) the idea of shared or ‘stretched’ leadership that incorporates both formal and informal leaders, has been of interest to researchers, policy-makers practitioners and educational reformers around the globe. Distributed leadership has captured the attention of many international researchers and as this article will show, has been the focus of a great deal of empirical enquiry. This article looks at the two decades of research that followed the pivotal Spillane et al. (2001) article on distributed leadership. Firstly, it takes a retrospective view by drawing upon selected literature from 2001 to 2011 mapping out the main findings based on this empirical terrain. Secondly, it offers a contemporary view by exploring recently selected literature on distributed leadership from 2011 to 2021. The article does not claim to be a systematic review of the literature but rather, offers some insights into selected evidence over two decades. The article considers how far distributed leadership remains a relevant concept for those working within the field of educational leadership.

KEYWORDS:

• Leadership
• distributed leadership
• literature review
• school improvement
• school effectiveness

Introduction

Few ideas, it seems, have provoked so much attention, debate, and controversy, in the field of educational leadership, as distributed leadership. Unlike many other popular leadership ideas, labels, and theories, distributed leadership has stood the test of time and it could be argued has experienced a renaissance during COVID times, where all leadership practice was virtual, remote, and inevitably distributed (Harris and Jones, 2020).

While the notion of shared, collaborative, or participative leadership is far from new, within the field of educational leadership, distributed leadership theory has provided a new lens on a familiar theme. The seminal work of Spillane, Halverson, and Diamond (2001b) sparked renewed interest in leadership as practice focusing particularly on the interactions between leaders, followers, and their situations. In essence, distributed leadership theory implied a fundamental re-conceptualisation of leadership as practice and challenged much of the conventional wisdom about leadership defined as discreet leadership roles or functions. Distributed leadership implies a model of shared, collective, and extended leadership practice where the emphasis is upon interdependent interactions rather than individual and independent actions (Harris, 2013). Distributed leadership underlines how multiple sources of leadership operate within any organisation and it suggests that this ‘leader plus’ aspect is a critical feature within organisational change (Spillane 2006, 3).

From the outset, however, distributed leadership has been a leadership idea accompanied by controversy and critique. Writes like Fitzgerald and Gunter (2006) and Hargreaves and Fink (2009) called into question the motivation of those espousing distributed leadership. In their view, distributed leadership was little more than a palatable way of encouraging teachers to do more work, a way of reinforcing standardisation practices, simply old managerialism in a contemporary guise. Instead of being a more democratic form of leadership. Hargreaves and Fink (2009) proposed that distributed leadership could in fact be another, more attractive, mechanism for delivering government policy. This point has been accepted by many of those writing about distributed leadership, insofar that it is acknowledged that there can be a ‘dark side of distributed leadership (Harris, Jones, and Baba 2013) if it is used to coerce or impose ideas through the guise of a more democratic or collaborative way of working.

In their critique, Fitzgerald and Gunter (2006, 335) go a little further suggesting that teacher leadership, which is closely associated with distributed leadership ‘merely cements authority and hierarchy whereby leaders monitor teachers and their work to ensure set of predetermined standards are met’. They argue ‘that teacher leadership is deeply rooted in neo-liberal versions of the performing school and that it is a management strategy and not a radical alternative’ (Fitzgerald and Gunter 2006, 335). Again, the dangers of romanticising the notion of distributed leadership in ways that make it susceptible to manipulation and false representation has been addressed by many of those writing about distributed leadership.

It is important, however, not to lose sight of these critical perspectives as they afford the opportunity to consider whether distributed leadership could just be a more palatable way of engineering greater organisational control. Hence, those researching in the leadership field, have ensured that the shortcomings and limitations of distributed leadership, as well as the possibilities and opportunities, are represented in their writing (e.g. Harris 2007; Hallinger and Heck 2009).

The next section of this article offers a retrospective view of a selected evidence base¹ (Harris 2008b), published following Spillane, Halverson, and Diamond (2001a). As noted already, this is not a systematic review of the evidence base nor does it claim to be. Rather, evidence about distributed leadership has been selected, from two decades, to outline general trends and shifts in the type of empirical studies that have focused on distributed leadership.

Distributed leadership: taking a retrospective view

Looking back at the selected literature on distributed leadership from 2001 to 2011, it was consistently argued, by those writing about distributed leadership, that this leadership practice was pivotally important to securing organisational improvement. Consequently, knowing if, how and in what way distributed leadership practice influenced organisational outcomes was at the heart of a great deal of contemporary scholarly enquiry during this period (Leithwood, Mascall, and Strauss 2009). A great deal of the research evidence, at this time, also signalled a positive relationship between distributed leadership, organisational improvement, and student achievement (Harris 2008a, 2009; Hallinger and Heck 2009; Leithwood and Mascall 2008). Such studies reinforced the importance of distributed leadership as a potential contributor, under the right conditions, to positive influence organisational change and improvement.

Even though the evidence base about distributed leadership was still emerging and growing, in the decade following Spillane, Halverson, and Diamond (2001b), distributed leadership, as a concept, had already been enthusiastically adopted within educational policy circles in several countries including the UK, USA, Australia, parts of Europe, and New Zealand. Most of these reforms took a normative stance on leadership distribution and emphasised extended forms of leadership at school, district, and system levels. For example, in Wales, distributed leadership was a key part of the national school effectiveness framework in 2012. In England, distributed leadership featured heavily in workforce remodelling and reform along with the introduction of new models of schooling such as federations, partnerships, networks, and multi-agency working. In the USA, distributed leadership was a feature of many of the comprehensive school reform approaches (Mayrowetz 2008).

Despite words of caution and critique from those in the research community, distributed leadership became an idea that was widely advocated and endorsed in policy circles around the world. This proved to be both good news and bad news. On the positive side, there was evidence to suggest that certain forms of collective leadership or forms of distributed influence have a modest but significant indirect effect on student achievement (Leithwood and Mascallm 2008, 546). However, it was also clear, from the available evidence during this period, that it was how leadership was distributed that mattered most of all.

Many studies, in this decade, underlined that distributed leadership was not intrinsically a positive or negative thing, as it depended upon the purpose, form, and nature of the distribution. The empirical work highlighted that it was how leadership was distributed (i.e. the patterns of distribution) that explained its subsequent influence and impact. As Leithwood, Mascall, and Strauss (2009) clearly highlighted in their empirical analysis, purposeful or planned leadership distribution was more likely to impact positively on school development and change. Other researchers similarly reinforced that successful leadership distribution occurred by careful design rather than from some random default position (Day et al. 2009).

Definition and design

In the decade that followed Spillane, Halverson, and Diamond (2001a), distributed leadership became a concept that was interpreted in different ways. The idea of distributed leadership as delegation prevailed and its popular positioning as the antithesis of top-down, hierarchical leadership was prevalent. This proved to be deeply unhelpful as both these positions misrepresented distributed leadership. As researchers at the time underlined, distributed leadership encompassed both formal and informal forms of leadership practice within its framing, analysis, and interpretation. Distributed leadership was primarily about the ‘co-performance of leadership and the reciprocal interdependencies that shaped leadership practice’ (Spillane 2006, 58).

During this decade, researchers who investigated distributed leadership tended to focus on two pressing questions. Firstly, what influence if any, did different patterns or configurations of distributed leadership make to classrooms, schools, and school systems? Secondly, how could we know? On the second of these issues, it was clear that the methodological challenges associated with research into distributed leadership were complex and considerable. The proliferation of case-study accounts of the single, often ‘heroic’ leaders in the leadership literature, that persist even now, can be largely explained by the methodological simplicity and ease of data gathering. If leadership is understood primarily as role, then it is relatively straightforward to collect evidence about the actions and responsibilities of those occupying such positions. Many leadership case studies have been compiled based upon the actions, usually self reported, of individual leaders. It remains a moot point, how useful these narratives have been to the field and how far they have moved the knowledge base on leadership forward.

Without question, it is more difficult to empirically investigate distributed leadership, because of the multiple sources of influence, but it is not impossible. As Spillane (2006) noted, data collection methods needed to be more sophisticated and nuanced to capture distributed leadership practice. His research work, with colleagues over this decade, showed that it was perfectly possible to research distributed leadership through the development of new research techniques and tools.

Early work by Spillane, Halverson, and Diamond (2001a) focused on the nature of interdependencies and the co-performance of leadership practices in schools. Implicit in the notion of ‘co-performance’ was the possibility that those performing the practices might be pursuing different or even contradictory goals. ‘From a distributed perspective, leaders can interact in the co-performance of leadership routines even when they seek different or conflicting outcomes’ (Spillane 2006, 84). Thus, taking a distributed perspective recognises the possibility that people may be working with different goals or outcomes in mind.

Empirical evidence

But what exactly do we know from the early empirical evidence about distributed leadership? To answer this question it is important to note that the empirical base about distributed leadership could be found in different research fields and traditions. Early studies that offered research-informed insights into distributed leadership were contained in literatures pertaining to social psychology; organisational theory; teacher leadership; and school leadership. This research terrain was diverse and was reflected in different methodological positions.

Yet despite the miscellaneous nature of the early evidential base, consistent messages about distributed leadership and organisational change were forthcoming. For example, there was evidence that suggested a positive relationship between school level change and wider leadership distribution, albeit focused on changes in teaching and learning (Harris 2008a; Leithwood and Mascallm 2008). Indeed, there were a growing number of early studies that focused attention explicitly on the impact of distributed leadership upon teaching and learning (Leithwood, Mascall, and Strauss 2009). Such studies provided evidence about the nature, form and impact of distributed leadership practices in schools. Studies by Camburn and Han (2009); Hallinger and Heck (2009) and Leithwood, Mascall, and Strauss (2009) highlighted a positive relationship between distributed leadership and certain student learning outcomes. These studies also tended to reinforce a positive impact on teachers’ level of self efficacy and motivation. Leithwood, Mascall, and Strauss (2009) concluded that the ‘field as now much closer to developing impact studies of distributed leadership because of the theoretical, conceptual and empirical work undertaken’. They noted that the major challenge to the field was to start designing such research studies.

Early research studies also tended to conceptualise distributed leadership as a form of work-redesign and focused on distributed leadership as a form of job re-design or work restructuring (Harris 2008a). Other researchers paid close attention to the different patterns of leadership distribution in schools (Leithwood, Mascall, and Strauss 2009; Spillane and Camburn 2006). As noted earlier, the central question pursued, at this time, concerned different patterns of distributed leadership that existed in schools and whether, and in what way, if any, they made a difference in organisational outcomes. Generally, the empirical findings suggested that the configuration of leadership distribution was important and that certain patterns of distribution had a more positive effect upon organisational development and change (Leithwood, Harris, and Hopkins 2008).

The early evidence proposed that there were different formations of leadership distribution in schools, some were random, some emergent, and some were carefully orchestrated. Leithwood, Mascall, and Strauss (2009) suggested that there were different patterns of distribution reflecting differing degrees or active alignment and co-ordination. This work reinforced the importance of planned, aligned distributed leadership practice that is purposeful and focused. Other evidence pointed towards the importance of ‘principal directed’ approaches to distributed leadership (Harris 2008a; Day et al. 2009).

Early evidence also reinforced that distributed leadership had a greater impact upon organisational development where certain structural and cultural conditions were in place (Louis et al, 2009). Within this early literature there was the strong implication that distributed leadership had the potential to positively influence organisational change and student learning outcomes, but only under the right conditions.

Impact on student outcomes

Without question, the relationship between distributed leadership and student learning outcomes, in the early evidence base, was controversial and contested. Positions on the relationship between distributed leadership and student learning outcomes varied. Some writers argued that seeking to explore this relationship was simply a futile exercise. They suggested that the search for normative links between specific leadership distribution patterns and student achievement results is unlikely to yield clear guidelines for practice. Conversely, others argued that distributing leadership is only desirable if the quality of leadership activities contributes to ‘assisting teachers to provide more effective instruction to their students’ (Timperley 2005, 220).

In essence, there were a limited number of studies that explicitly explored the relationship between distributed leadership and learning outcomes. Two studies offered a useful starting point in the exploration of the relationship between distributed leadership and student learning outcomes. The first study by Leithwood and Jantzi (2005) suggested that distributing a larger proportion of leadership activity to teachers had a positive influence on teacher effectiveness and student engagement. They also noted that teacher leadership had a significant effect on student engagement that far outweighs principal leadership effects after considering home family background.

The second study by Mulford and Silins’s (2003) also provided empirical confirmation of the key processes through which more distributed kinds of leadership influence student learning outcomes. Their work concluded that student outcomes are more likely to improve when leadership sources are distributed throughout the school community and when teachers are empowered in areas of importance to them.

There was other early, empirical work that pointed towards a positive relationship between distributed leadership and student learning outcomes. For example, a study of teacher leadership conducted in England found positive relationships between the degree of teachers’ involvement in decision making and student motivation and self efficacy (Harris and Muijs 2004). This study explored the relationship between teacher involvement in decision making within the school and a range of student outcomes. These findings showed a positive relationship between distributed leadership and student engagement.

Camburn and Han (2009) explored the outcomes of distributed leadership by drawing upon extensive evidence from an investigation into the America’s Choice CSR program. A core design feature of this programme was the requirement to distribute leadership responsibilities to teacher leaders in schools to act as a key lever for instructional change. Their research work outlined the impact of this programme in 30 urban elementary schools and investigated the association between the distribution of leadership to teachers and instructional change. This study concluded that distributing leadership to teachers can support positive instructional change under the right conditions.

Hallinger and Heck (2009) explored the impact of system policies on the development of distributed school leadership and school improvement. Their quantitative analysis and results supported a relationship between distributed leadership and the school’s capacity to improve. They concluded that distributed leadership was an important co-effect of school improvement processes. Research by Day et al. found that substantial leadership distribution was very important to a school’ success in improving pupil outcomes. The findings from this study showed that distributed leadership was positively correlated with the conditions within the organisation, including staff morale, which in turn impacted positively upon student behaviour and student learning outcomes.

In their work, Leithwood et al. (2009), highlighted the potentially positive influence of distributed leadership practices on student learning outcomes. In summary, the early empirical evidence about distributed leadership and student learning was encouraging but it was clear that more work was needed to consolidate and confirm this relationship.

The next section of this article looks at scholarly work about distributed leadership, produced between 2011 and 2021. It comments on a selected literature that has been generated by a specific methodology which is explained next.

Distributed leadership- taking a contemporary view

This section of the article looks at a selected, international, contemporary literature about distributed leadership. As noted already, this is not a systematic review of the literature but rather, it offers an overview of evidence based on a selected literature. Only peer reviewed articles in scholarly journals were selected to compile this narrative. It is fully accepted that there are potentially many books, book chapters and reports on distributed leadership, which are missing from this overview. Similarly, opinion pieces, blogs, and other media did not feature in the evidential base selected.

The prime aim of this selected overview is to highlight trends in the more contemporary, empirical evidence base about distributed leadership. It is fully acknowledged that this process may have missed key work published outside academic journals and accepts that it has omitted scholarly work written other languages.

In this overview, the data retrieval and analysis process followed the review method guidelines outlined by Kitchenham (2004) that also corresponds to the PRISMA (2008) methodology -

1. Setting Keywords

2. Selecting databases

3. Defining search criteria

4. Selected articles

5. Analysis framework

6. Extracting information from the articles into the analysis framework

In terms of searching the academic literature, it was decided to use a range of databases to identify relevant articles and to cross-check these databases to ensure that all the retrieved information was accurate, and no publication was missed. EBSCOHOST was used and the search was subsequently extended to include Web of Science and Google Scholar for cross checking. The databases explored within EBSCOHOST included:

• British Education Index (BEI)

• ERIC

• Education Abstracts

• Education Administration

• Academic Research Complete

The inclusion criteria for this review of journals were:

• Source: refereed journal articles only

• Search Dates 2011–2021

• Language: English

• Search in article: title against the keywords

• Sources: only peer reviewed journals SCOPUS and ISI

The data retrieval and analysis process for all journal articles followed four steps:

• Identification – all studies identified in the initial results of the electronic literature search were recorded.

• Screening – studies that did not fully meet the inclusion criteria were excluded and duplicate results were removed from the list.

• Eligibility – abstract screening for each output was conducted to ensure only relevant and appropriate articles were selected for the next stage.

• Inclusion – full PDFs of the selected articles were read to ensure they remained relevant and appropriate for inclusion in the final selection.

This process generated the following articles (Table 1):

Table 1. Summary for number of papers across PRISMA steps.

Step in PRISMA	Number of papers
Identification	368
Screening	190
Eligibility	121
Inclusion	43

The initial results from the search were significantly higher than the final list of articles selected. Many of the articles that appeared in the identification stage were irrelevant on closer inspection (e.g. many articles contained the word ‘distributed’ but did not focus explicitly on leadership). Also, many articles were duplicates and therefore were removed. Despite a significant number of articles identified initially in the search, the rigorous filtering process for quality, direct relevance and appropriateness resulted in 43 articles being selected finally in this overview.

The findings from this contemporary consideration of a selected literature are summarised under seven key themes that feature heavily and consistently in the articles included. The implications for future research into distributed leadership are then assessed and outlined.

Theme 1 – exploring relationships

Within the selected literature on distributed leadership, large scale, quantitative analyses of relationships feature heavily. There are many studies that have explored the relationship between distributed leadership and some other factor or variable (e.g. academic optimism, trust). In the articles included in this overview, a range of relationships are explored. For example, Amels et al. (2020) examine the relationship between teachers’ perceptions of distributed leadership and inquiry-based work in primary schools and the resulting impact on those teachers’ capacities to contribute to educational change. The results indicate a direct, positive effect of distributed leadership on teachers’ collaboration and collegiality, as well as on their motivation to contribute to educational change.

Another quantitative study by Chang (2011) explored the relationship between distributed leadership, teachers’ academic optimism and student achievement in learning. The study involved 1500 teachers and a structural equation modelling approach was applied to the data. This study established that not only did distributed leadership have a positive influence on academic optimism, but it also indirectly affected student achievement. Similar quantitative studies by Chen (2018); Hulpia, Devos, van Keer (2011), and Aldaihani (2019) focused on the relationship between distributed leadership and instructional leadership, organisational commitment, teacher professionalism and trust, respectively. In all but one of these studies a positive relationship was proven.

Looking more broadly to the larger body of evidence on distributed leadership, since 2011 there has been a burgeoning of studies in many different countries that have explored the relationship between distributed leadership and a wide range of factors. Most of these studies focus on some aspect of organisational change and largely, but not exclusively, highlight a positive relationship. This would be in keeping with the earlier evidence about the potential of distributed leadership to contribute to organisational change. Despite the early evidential base on this relationship and the positive evidence from many contemporary studies, a review by Tian, Risku, and Collin (2015) concluded, that the positive impact of distributed leadership on organisational change and learner outcomes remains questionable.

It is clear, from the selected evidence in this overview, that the relationship between distributed leadership and many different organisational factors or variables chosen, presents a very fragmented, and at times incoherent picture within the evidence base. These studies do not coalesce in any meaningful way or connect with one another, rather they tend to be singular explorations of variables, without clear theoretical threads or empirical connections, thus delimiting their potential contribution and significance to substantial knowledge generation.

Theme 2 – different contexts

While the early evidence on distributed leadership tended to be Anglo centric with studies largely coming from North America and Europe, the selected contemporary literature on distributed leadership reveals a much broader range of countries and settings. The selected literature included in this overview illustrates that many more researchers from a wide range of countries and contexts are exploring distributed leadership. Of note is the fact that there has been an uptake in research into distributed leadership in middle eastern countries. For example, Aldaihani (2019) examined different applications of distributed leadership in high schools in the State of Kuwait from teachers’ perspectives. As a quantitative study, the study sample consisted of 1210 high schools, and it concluded that the practice of distributed leadership in Kuwaiti schools was low.

Al-Harthi and Al-Mahdy (2017) similarly examined the prevalence of distributed leadership in Egypt and Oman, using a quantitative approach. This study similarly showed that the adoption of distributed leadership practices was low, and training therefore was required. In a study by Hashem (2022) the focus is on Jordan and this is the first of the selected articles to focus on the way culture affects distributed leadership in practice. Hashem (2022) talks about being culturally bound by certain norms and values that make the possibility of distributed leadership practices, within certain contexts, unlikely.

Other studies have looked further afield to encompass countries like Turkey where Coban and Atasoy (2020) found a positive relationship between distributed leadership, organisational innovativeness and teacher collaboration. Crawford (2012) reports upon a study into distributed leadership practices in Spain, while García Torres (2017) and Hairon and Goh (2014) offer some empirical insights into distributed leadership practices in Singapore, again largely focused on examining the collaborative practices that accompany distributed leadership. Kangas, Venninen, and Ojala (2015) explore distributed leadership in Finland noting that it is generally perceived to be an important part of professional development and learning.

Harris and DeFlaminis (2016) report on a major distributed leadership research project in the United States. The article outlines evidence about the impact of distributed leadership in schools and districts engaged in a large-scale research and development project over five years. This is one of the few longitudinal research studies of distributed leadership practice and the findings suggest a positive relationship between distributed leadership and positive individual and organisational change.

It is encouraging that research is being undertaken on distributed leadership in many more countries than was the case in the earlier decade. Yet, the context of these countries is not always given sufficient attention in the analysis of the findings. In countries where more hierarchical structures are in place and where there are high degrees of accountability, it is unsurprising that levels of distributed leadership practice are reported as low. Conversely, where more open, democratic processes are in place, it is concluded that distributed leadership has the possibility of flourishing. By looking at the selected articles that focus on distributed leadership within a specific country, it is concluded that the cultural influences require far more scrutiny and attention to fully explain the findings and to contextualise the conclusions.

Theme 3 – instrument development

One of the pleas from those writing about distributed leadership during 2001–2011 was the need for more sophisticated instruments to capture data about distributed leadership practices. There was also a call for more large-scale mixed methods studies of a longitudinal nature. In the group of articles selected for this contemporary overview, there are a number which highlight the development and implementation of new instruments aimed at capturing specific aspects of distributed leadership. This is both positive and encouraging.

Baloglu (2012) focuses on the relationship between value-based leadership and distributed leadership behaviours of school principals. A causal research design was employed in the study involving 225 primary teachers. Data were gathered by means of a specially developed ‘Values Based Leadership and Distributed Leadership Inventory’. Confirmatory and exploratory factor analyses were used to define the construct validity of the scale. It was concluded that there were four positive dimensions of distributed leadership – teamwork, support, vision creating and control.

Blitz and Modeste (2015) outline the development and application of a Comprehensive Assessment of Leadership for Learning (CALL) which they advocate is a multi-source assessment of distributed leadership. The authors outline the use of this instrument to examine differences between teacher and leader ratings in assessing distributed leadership practice.

Carbone et al. (2017) describe how they used a ‘Distributed Leadership Benchmarking Tool’ to assess how effectively distributed leadership was enabled at five Australian institutions that were implementing a collaborative teaching quality development scheme premised on bringing teams of academics, co-ordinators, etc. together with the prime goal of enhancing teaching and learning. The conclusion drawn is that the benchmarking tool is useful in assessing the alignment (or otherwise) of teaching and learning.

Mullick, Sharma, and Deppeler (2013) report on a study of teachers’ perceptions about distributed leadership for inclusive education in Bangladesh. ‘A Distributed Leadership Practice for Inclusive Education Scale’ was developed specifically for this study. The results indicate that teachers perceived that distributed leadership practices for IE were present in primary schools in Bangladesh. The findings indicate that teachers’ perceptions about distributed leadership practices for inclusive education have a positive correlation with their satisfaction about the implementation of inclusive education policy.

Of those articles selected as part of this contemporary overview, only Ersozlu and Ulusoy (2016) adapted an existing instrument for their research purposes. The ‘Distributed Leadership Scale’ originally developed by Davis was used to explore aspects of distributed leadership in Turkey.

While the development of new instruments to capture distributed leadership is unquestionably a positive aspect it is noticeable that firstly, many of these instruments capture perceptions, which can be unreliable, and secondly, many of the instruments, with the exception of Erzozlu and Ulusoy (2016) are single use only, in other words they are only reported to be used in one study. This inevitably calls into question the extent to which a cogent knowledge base about distributed leadership can be built without the consistent, repeated use of reliable and valid instruments in different contexts and settings that capture more than just perceptions.

Theme 4 – critique

The early work on distributed leaders, as acknowledged already in this article, was not without critique. The subsequent decade of research on distributed leadership did not fare any better with many writers continuing to find displeasure with the idea. Of the selected contemporary literature within this overview, there are several critical pieces that follow the contours of the arguments made about distributed leadership in the early years of its development. Indeed, many of the critics of the early work on distributed leadership are also vocal in more contemporary accounts and analyses.

For example, Gunter, Hall, and Bragg (2013) present distributed leadership as a study in knowledge production. They outline the development of a mapping framework to present and analyse knowledge production and distributed leadership in schools. They propose that their analysis identified the prevalence and political dominance of functional approaches and provides explanations regarding the interplay between state, public policy and the preferred types of knowledge. Hall (2013) follows a similar line of argument positing that distributed leadership is located within wider structural reform of education in England as part of the New Public management movement in public service delivery. He suggests that there are officially authorised discourses of distributed leadership that are imposed on teachers and schools.

Corrigan (2013) considers the rhetoric and reality of distributed leadership, arguing that there are contrasting oppositional messages in the literature and that this calls into question the resurgence of distributed leadership. Lumby (2017) considers distributed leadership in the context of post bureaucratic organisations and proposes that there is little evidence that distributed leadership offers an authentic re-distribution of power. Woods (2016) suggests that the practice of distributed leadership is characterised by multiple authorities which are constructed in the interactions between people. A key conclusion is that everyone is involved in the ongoing production of authorities by contributing to who is accepted or is excluded from exercising authority and leadership.

Such critiques remain important in challenging assumptions and questioning the purpose of distributed leadership.

Theme 5 – other disciplines

A major difference between the selected evidence on distributed leadership and the early work is the introduction of research that emanates from disciplines other than education. Within the selected articles, a number focused on distributed leadership within other sectors. Currie and Lockett (2011) examine leadership in practice, drawing on Spillane’s (2006) work within the context of health and social care. Floyd and Fung (2015) consider the idea of distributed leadership within universities, noting that there is relatively little written on this topic. The article reports on a funded study which explored how one institution had implemented a newly conceived distributed leadership model. Their findings suggest that the challenge of distributed leadership in universities is complex on several levels that might be antithetical to the strategic direction of the institution.

Gilles et al. (2021) describe the utility and impact of a distributed leadership model to implement a National Health Service (NHS) national quality improvement programme. They report on the inclusion of different groups in decision making. Heikka and Hujala (2013) investigated the distribution of responsibilities for leadership in early childhood education highlighting that developed or extended forms of distributed leadership were not used. Jonasson, Kjeldsen, and Ovesen (2018) consider the implementation of distributed leadership during a hospital merger suggesting the need to understand the complex dynamics of widened and restricted leadership distribution.

Compared to the field of education, explorations of distributed leadership in other sectors remain relatively limited. The studies that were selected, however, offer new insights into the challenges of enacting distributed leadership in different contexts, and structures. These studies add value to the literature because they provide an important comparative element that can offer researchers in the field of education new insights and perspectives.

Commentary

An assessment of this selected, contemporary evidence base confirms that there is still much scholarly interest in distributed leadership. In fact, the ongoing scale of research activity suggests that empirical interest in distributed leadership has not waned in the last decade. This conclusion is also supported by Gumus et al. (2018) who found that distributed leadership is the most researched model of leadership along with instructional and transformational leadership.

The selected studies within this overview largely tend to be quantitative in orientation and focused on exploring specific relationships. In contrast, the qualitative studies selected in this overview are smaller in scale. Within the contemporary selected literature, there were six articles that drew upon case study data (Ng and Ho 2012; Or and Berkovich 2021; Piot and Kelchtermans 2015; Tahir et al. 2016; Tam 2018; Torrance 2013). The largest sample in these selected articles was six schools. This raises important methodological considerations and highlights the need for more studies that integrate multiple sources of robust evidence from different sources of data (qualitative and quantitative data).

The 64-million-dollar question is does the contemporary evidence about distributed leadership offer any new conclusions? To answer this question, requires a return to the 2008 article, Seven Strong Claims about Successful School Leadership (Leithwood, Harris, and Hopkins 2008). In 2008, the authors made two strong claims about distributed leadership. In 2020, the authors of this work revisited the claims, based on the evidence that had emerged since 2008. They concluded that a considerable body of relevant evidence had been reported since 2008, significantly increasing certainty about the validity of their claims about distributed leadership. Namely, that school leadership can have an especially positive influence on school and student outcomes when it is distributed (Claim 5) and some patterns of distribution are more effective than others (Claim 6). Hence, the most recent evidence provides further validation of these two claims and underlines that distributed leadership remains an important area of empirical enquiry for those working in the educational leadership field.

So, what has the contemporary evidence about distributed leadership yielded? While the introduction of more large-scale quantitative studies is to be welcomed, the black box of distributed leadership practice remains only partially open, there is more focused and collective empirical work to do. There are strong rays of recent light, however, in the form of the work around networking and professional community that show the benefits of collaborative working and distributed leadership practices. These offer strong lines of future empirical enquiry into distributed leadership (Azorin, Harris and Jones 2020).

It is concluded that future research into distributed leadership would benefit greatly from large-scale, longitudinal, multiple methods studies, if the knowledge base is to improve significantly. Such studies could record, capture, and illuminate the way in which distributed leadership is best deployed. An assessment of this overview of selected, contemporary evidence would also suggest that more co-ordinated, collaborative, connected set of scholarly activities could contribute greatly to a better and more grounded understanding of distributed leadership. If undertaken across different contexts or settings, this collaboration would ensure that the future knowledge base on distributed leadership is more reliable offering the real possibility of robust policy advice and sound professional guidance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The evidence base on DL (2001–2011) was considered in an earlier article and some of it is represented and highlighted here.