1
New Perspectives on an Old Idea
A Short History of the Old Idea
KENNETH LEITHWOOD, BLAIR MASCALL, AND TIIU STRAUSS
Current conversations about educational leadership have increasingly included questions about its sources. Distributed leadership is much in vogue with researchers, policy makers, educational reformers, and leadership practitioners alike (Hammersley-Fletcher & Brundrett, 2005; Storey, 2004). Not surprisingly, however, there are competing and sometimes conflicting interpretations of what distributed leadership actually means. As Harris (2004) has noted, the definition and understanding of distributed leadership varies from the normative to the descriptive. Some, for example, are attracted to increasing, or otherwise manipulating, the distribution of leadership as a possible strategy for school improvement. This perspective, reflected in most of the chapters in this text, encompasses âsharedâ (Pearce & Conger, 2003), âdemocraticâ and âdispersedâ (e.g. Ray et al., 2004) conceptions of leadership, as well. Others (e.g. Spillane et al., Chapter 5 in this volume) employ the concept as a means of âsimplyâ better understanding the meaning and nature of leadership in schools. Not surprisingly, then, the literature about distributed leadership remains diverse and broad based (Bennett et al., 2003).
Although current interests in leadership distribution represent a shift in both educational leadership research and policy from a preoccupation with those in formal roles, the study of distributed sources of leadership can be traced back at least to the work of Gibbs in 1954 (Gronn, Chapter 2) and possibly as far back as the mid-1920s. As MacBeathâs example of the counsel to Moses reminds us (Chapter 3), however, the actual practice of distributed leadership is as old as human efforts to organize. Pearce and Conger (2003) provide a very useful synopsis of the roots of research on leadership distribution and trace their evolution to the present. The significant shift in interest toward sources of leadership reflects, they suggest, disillusionment with âgreat manâ conceptions of leadership and bureaucratic organizational structures. This shift also reflects growing appreciation for the contributions to productivity of the informal dimensions of organizations (Tschannen-Moran, 2004), the untapped and often unrecognized leadership capacities found among those not in positions of formal authority (Gronn, 2003), and the extent to which the capacities of those at the organizational apex alone have been overtaken by the complexities of the challenges they now face (Wheatley, 2005). Through a normative lens, leadership increasingly is conceptualized as an organization-wide phenomenon (Ogawa & Bossert, 1995) in which flatter organizational structures and leadership distributed over multiple people and roles are being advocated as solutions to these dilemmas (Manz & Sims, 1993).
Indeed, the overwhelming disposition of the contemporary, normatively-oriented literature on distributed leadership is enthusiastic optimism about its anticipated benefits. As compared with exclusively hierarchical or âfocusedâ forms of leadership, distributed leadership is thought to more accurately reflect the division of labor which is experienced in organizations from day to day and to reduce the chances of error arising from decisions based on the limited information available to a single leader. Distributed leadership, it is argued, also enhances opportunities for the organization to benefit from the capacities of more of its members, permits members to capitalize on the range of their individual strengths, and develops among organizational members a fuller appreciation of interdependence and how oneâs behavior affects the organization as a whole. Through increased participation in decision making, greater commitment to organizational goals and strategies may develop. Distributed leadership, some claim, has the potential to increase on-the-job leadership development experiences and reduce the workload for those in formal administrative roles (presumably by increasing the workload of others). The increased self-determination believed to arise from distributed leadership may improve membersâ experience of work. Such leadership might allow members to better anticipate and respond to the demands of the organizationâs environment. Solutions to organizational challenges may develop through distributed leadership which would be unlikely to emerge from individual sources. Overlapping actions that occur in some distributed leadership contexts provide further reinforcement of leadership influence (e.g. Burke et al., 2003; Cox et al., 2003; Gronn, 2002; Grubb & Flessa, Chapter 7). Englandâs National College of School Leadership is advocating more distributed leadership as a strategy for increasing both administrator and teacher retention, as well as minimizing the significant negative consequences typically associated with leadership succession (NCSL, 2004).
Empirical Evidence About the Consequences of Distributed Leadership
Positive Consequences
The above list of potential positive consequences of distributed leadership is impressive. One might reasonably expect that if even a few such outcomes materialized, the effects on the organizationâs bottom line would be signific-ant. In point of fact, however, the evidence to justify a belief in these consequences is mixed and indirect. On the supportive side of the ledger, for example, the effectiveness of democratic, supportive and shared forms of organizational leadership (defined as control and influence) have received support from research on teacher participation with peers in planning and decision making (Talbert & McLaughlin, 1993) and from tests of shared transformational leadership effects (Leithwood & Jantzi, 2005). A reasonably strong case for the value of distributed leadership can also be found in studies of organizational turnaround processes, as well. While focused leadership seems most useful at the âcrisis stabilizationâ stage, the subsequent ârecoveryâ stage demands widespread sharing of responsibility to be successful (Murphy, in press; Nicolaidou & Ainscow, 2002).
Leadership succession studies also provide evidence in support of the value of widely distributed sources of leadership, as NCSLâs advocacy illustrates. This line of research suggests that the often devastating effects of principalsâ succession on school improvement processes can be significantly mitigated when both the ownership and leadership of such processes is widely dispersed (e.g. Fink & Brayman, 2006). In a recent study of teacher retention, as another positive consequence of shared forms of leadership, Ingersoll (2007) found that teacher leadership or control over some key decisions, in this case school and classroom disciplinary policies, had striking effects on the willingness of teachers to continue in their existing schools. More specifically, âAlmost one in five teachers in schools with a low level of teacher control over student discipline issues were expected to depart, whereas only one in 20 were expected to depart from schools with high levels of teacher control over such issuesâ (p. 24).
Negative Consequences
Empirical evidence about the consequences of distributed leadership is not all positive, however. For example, a recent, very comprehensive, review of teacher leadershipâone approach to the distribution of influence and control in school organizationsâfound only a very small handful of studies which had actually inquired about effects on students and these data were generally not supportive (York-Barr & Duke, 2004). One of the few large-scale empirical studies directly testing the effects of collective school leadership on students (i.e. their engagement in school) also reported non-significant, negative effects. The authors speculated that this might signify a non-linear relationship between the number of sources of leadership and organizational outcomes. Beyond some optimal amount, perhaps âmore leadership actually detracts from clarity of purpose, sense of mission, sufficient certainty about what needs to be done to allow for productive action in the school and the likeâ (Leithwood & Jantzi, 2000, p. 61). Others have offered similar speculations (e.g. Bryk, 1998; Timperley, 2005). Ritchie and Woods (2007) allude to the potential of distributed leadership to increase the burdens and responsibilities of teachers without actually increasing their power. Such leadership may simply be used as a subtle strategy for inculcating among staff the values and goals of more powerful members of the organization.
Additional evidence about the negative consequences of more widely distributing leadership can be found in research carried out in non-school contexts, especially research in which effectiveness is defined as some version of organizational productivity and assessed using objective indicators. For example, Tannenbaum (1961) was able to provide only limited support for his hypotheses about the contributions of âdemocraticâ organizational control structures. And after about 15 years of programmatic research about âorganic management,â Miller and Rowan (2006) reported that âthe main effects are weak and positive effects appear to be contingent on many other conditionsâ (p. 220). It is reasonable to conclude at this point, then, that the positive consequences of more widely distributing leadership in schools cannot simply be assumed; their precise nature remains unpredictable and likely depends on circumstances and conditions that we do not yet understand very well.
Theoretical Explanations for Distributed Leadership Effects
In addition to the empirical evidence exemplified above, a handful of different theoretical perspectives give rise to expectations about mostly positive associations between organizational effectiveness and the distribution of influence and control to more people, especially those not occupying formal leadership roles. We briefly describe four such perspectives, pointing out how each might be used to better understand the nature and consequences of leadership distribution. These perspectives include organizational learning, distributed cognition, complexity science, and high involvement leadership or management.
Organizational Learning Theory
According to this perspective, learning can take place outside individual brains (Weick & Roberts, 1996). An organization can be more intelligent than any one of its individual members, reflecting Gronnâs (2002) concept of âconcertive actionâ as a type of distributed leadership. This collective learning depends on the nature of a small set of key organizational conditions such as a culture of collaboration (Ritchie & Woods, 2007; Starbuck, 1996) in support of such learning. Hutchinsâ (1995) description of how the navigation team, on a disabled aircraft carrier, managed to bring the ship under control without being able to use their established routines and without any central direction, is often cited as the type of evidence capable of supporting these claims. Applied to the concept of distributed leadership, this line of theory raises the expectation that distributed leadership will lead to improved organizational capacity; it also suggests the need for implicit coordination, if not intentional planning, of leadership distribution if the superior capacity development assumed by collective learning is to materialize in practice.
Distributed Cognition
Theories of distributed cognition offer a set of ideas closely related to those found in theories of organizational learning as explanations for the potentially positive effects of distributed leadership (e.g. Salomon, 1993). This line of theory points to the different sets of capacities that exist not only within individual members of an organization but also in the technical and physical artefacts that constitute the setting in which people find themselves. From this perspective, capacities are distributed throughout both the social and material conditions which constitute the organization. When applied to the concept of distributed leadership, this line of thinking encompasses Jermier and Kerrâs (1997) concept of âsubstitutes for leadershipââalthough it would be more accurate to describe what they have in mind as substitutes for leadership by people. These substitutes include the direction and influence on organizational members exercised by the technical and material artefacts found in organizations as, for example, its policies, procedures, culture, and shared mental models.
If our concept of leadership includes at least influence in pursuit of the organizationâs directions, such non-social features of the organization must count as forms of distributed leadership. In this text, the work of Spillane et al. (Chapter 5) makes explicit reference to distributed cognition theory, even borrowing and adapting Perkins (1993) âperson plusâ concept to label a pattern of distributed leadership as leader plus. This pattern captures instances in which membersâ individual contributions add up to more than the sum of their parts through the interdependent nature of relationships among them. Spillane and his colleagues, however, limit their purview, as do the writers of all other chapters, to the social sources of leadership and cognition within organizations.
Much like organizational learning theory, distributed cognition gives rise to the expectation that the existing capacities of individual members of the organization, along with the sources of influence to be found in the organizationâs technology and other artefacts, are radically underutilized in contexts of control firmly exercised by formal leaders at the organizationâs apex. Unlike organizational learning theory, however, distributed cognition is mostly about using existing capacities more fully, as distinct from learning new capacities; it leads to the assumption that considerable value could be added to the organizationâs effectiveness by simply making better use of the existing capacities without the additional investment required to learn new capacities. The most powerful forms of leadership distribution, from this perspective, would be founded on close knowledge of where in the organization was located the expertise needed to respond productively to a perceived challenge. Given such knowledge, the job of formal leaders is to bring those perhaps disparate sources of expertise together and insert a coordination function into their collective problem solving processes.
Complexity Science
Although its natural environment is the physical sciences, concepts from complexity science offer evocative metaphors for trying to better understand social organization. Complexity science appears to be a largely unused tool of some promise for unpacking the nature and consequences of distributed leadership. Indeed, this theoretical perspective has significant implications for focused leadership or those in formal leadership roles, as well. According to Uhl-Bien et al. (2007), for example, the rapidly changing context in which knowledge organizations (that would include schools) find themselves means that âorganizations must increase their complexity to the level of the environment rather than trying to simplify and rationalize their structuresâ (p. 301). Citing the Law of Requisite Complexity, this means that:
It takes complexity to defeat complexityâŚ. Knowledge Era leadership requires a change in thinking away from individual, controlling views, and toward views of organizations as complex adaptive systems [CAS] that enable continuous creation and capture knowledge. In short, knowledge development, adaptability, and innovation are optimally enabled in organizations that are complexly adaptive (possessing requisite complexity).
(p. 301)
Since bureaucracy and hierarchy are simplifying strategies, complexity science suggests that optimal adaptability is most likely when those closest to the action are empowered to shape the organizati...