J Educ Change (2010) 11:1–21 DOI 10.1007/s10833-008-9099-1

Thinking more deeply about networks in education ´ vila de Lima Jorge A

Published online: 22 November 2008
Springer Science+Business Media B.V. 2008

Abstract In education, initiatives to restructure and reculture schools through their involvement in intra- and inter-institutional networks have grown in number in recent years. Networks of teachers and schools (often linked to institutions outside education) are becoming a key focus of change efforts promoted by professionals and policymakers. However, in current research, writing and policymaking, the use of the network concept has been generally vague, normative and mostly instrumental. This paper develops a discussion of network concepts and proposes more systematic, less normative ways of addressing and researching network issues in education. It outlines a set of key dimensions of intra- and inter-organizational networks and makes a case for more a comprehensive research approach to these phenomena. Keywords

Networks
Educational change
Teacher collaboration

Introduction Contemporary societies are increasingly characterized by highly varied and complex forms of organization. Moves to new organizational moulds date back to at least the midtwentieth century, when workplace programs started to emphasize group processes and teamwork as critical factors for increased organizational productivity and innovation. Later, these changes extended to relations among organizations and lead to the extensive formation of inter-organizational networks. Networks have been classified as the organizational form of the information age (Castells 2001) and the inter-organizational network, in particular, has been described as ‘‘a fundamental unit of analysis in the study of advanced industrial societies’’ (Benson 1975, p. 229).

´ . de Lima (&) J. A Department of Education, University of the Azores, R. Ma˜e de Deus, 9501-801 Ponta Delgada, Portugal e-mail: [email protected]; [email protected]

123

2

J Educ Change (2010) 11:1–21

In the corporate world, acquisitions, joint ventures, research and development partnerships and other types of alliances are radically changing conventional ways of doing business (Perrow 1992). Between 1995 and 1998 there was a 48% increase in the number of fastest growing American companies that used strategic alliances, and from 1987 on the number of strategic alliances worldwide grew at an annual rate of 25% (Barringer and Harrison 2000, pp. 367–368). In education, moves to restructure schools and a growing emphasis on intra- and interinstitutional networks of professionals and organizations have become part of the policy landscape for at least two decades. Especially since the mid 1980s, many educational organizations have rushed to form or to join networks, or have been vigorously encouraged to do so. Today, these organizational forms are regarded as ‘‘a new construct for conceiving of educational provision and a new strategy for achieving reform’’ (Chapman and Aspin 2003, p. 653). Networks in education vary widely in size. Teacher networks can range from small groups of three or four colleagues who plan their lessons together in a given subject department (Lima 2003), to much wider groups of teachers, such as the National Writing Project in the USA (Lieberman and Wood 2003). Networks of schools can also range from small aggregates of 4 to 6 schools located in particular geographical areas to nation-wide systems such as the Coalition of Essential Schools (USA), or the Specialist Schools and Academies Trust (England), whose members count by the thousands. Advocates of the creation and development of teacher and school networks believe that they can be powerful instruments for breaking teachers’ classroom and institutional isolation (Pennell and Firestone 1996) and that they have ‘‘enormous potential to genuinely affect classroom practice’’ (Andrews and Rothman 2002, p. 509). Despite their growing prevalence, networks have become popular mainly because of faith and fads, rather than solid evidence on their benefits or rigorous analyses of their characteristics, substance and form. With notable exceptions (especially, Lieberman and Grolnick 1996, 1998; Lieberman and McLaughlin 1992), scholarly writing on educational networks has tended to focus almost exclusively on how good networks are and how much more we need to invest in them. However, there is nothing inherently positive or negative about a network: it can be flexible and organic, or rigid and bureaucratic; it can be liberating and empowering, or stifling and inhibiting; it can be democratic, but it may also be dominated by particular interests. What actually occurs in concrete educational networks is something for researchers to determine. This will only be achieved if we approach networks in a more systematic, more complex and less normative manner than has been the case in most studies conducted so far (there are, of course, notable exceptions, which will be referred to throughout this paper). For the field of educational research to overcome the currently prevailing normative and instrumental view of networks, these systems need to be regarded as conceptually, pragmatically and politically problematic. The main reasons for approaching educational networks in such a way are threefold: (1) theoretically, the complexity of these formations needs to be matched up with conceptual tools that do not oversimplify their nature and operation, and, despite the progress that has been made so far, this has not been achieved satisfactorily until now; (2) practically, an adequate theory of effective network design needs to take into account what forms, types and network dynamics produce which results and under what circumstances; and (3) from a policy point of view, interventions that aim at creating or modifying existing networks need to be informed by empirical research that shows what goes right and what can go wrong in distinct types of networks.

123

J Educ Change (2010) 11:1–21

3

The main purpose of this paper is to illustrate how complex educational networks are and to suggest several conceptual and analytical tools that can be used to approach them in innovative ways. In the paper, I outline a set of key dimensions of educational networks and propose some new lines of research that might allow us to learn more about these intricate systems. The paper reviews and discusses key research findings on networks in education and in organizations more generally, and proposes a provisional outline for a future research agenda in the field. The paper is organized into three main sections. Firstly, I clarify some key terminology and conceptual issues regarding networks in general. Then, I sketch out a list of key dimensions for the analysis of networks in education. Finally, I propose a few more specific research themes that would allow us to learn more about the interactional side of these complex systems.

The network notion: conceptual clarification Due to the growing prevalence of networks in many domains of social life and to their roots in sometimes very distinct intellectual, policy and practical traditions, they have been referred to by many different terms and expressions. With respect to inter-organizational networks in education, some of the most widely used terms include alliances, coalitions, collaborations/collaboratives, clusters, consortia, development groups, families, partnerships, federations, groupings, territories, trusts, and zones. In popular parlance, teachers and others tend to use many of these terms interchangeably. But terminology options are not innocent choices; each term conveys a particular set of meanings that may or may not be verified in the facts. Most of the times, the term ‘‘network’’ is used normatively to advocate what organizations must become, rather than to describe what they are. Also, today, the notion is totally trivialized and applied indistinctively to very dissimilar phenomena. This ‘‘threatens to relegate [the concept] to the status of an evocative metaphor, applied so loosely that it ceases to mean anything’’ (Nohria 1992, p. 3). Two narrow conceptions of the term ‘network’ have become prominent in the general literature on the issue (Nohria and Eccles 1992). The first is that of the network organization. Kindred labels include the ‘‘post-industrial,’’ ‘‘post-bureaucratic,’’ ‘‘heterarchical,’’ and ‘‘cluster’’ organization (see, for example, Bell 1973; Castells 2001). In this conception, the network is taken as a broad organizational template for contemporary societal formations. It is regarded as a new organizational form characterized by ‘‘a fluid, flexible, and dense pattern of working relationships that cut across various intra- and inter-organizational boundaries’’ (Nohria and Eccles 1992, p. 289). In the second conception, a network is understood mainly as a technological system that enables fluid and flexible interconnections among actors located in distinct intra- and inter-organizational levels and sectors. Regrettably, the two uses of the term ‘‘network’’ have tended to converge, so that the network organization is increasingly regarded by many as an electronic network and/or a set of spontaneous, unconditioned ties that are the opposite of hierarchy or bureaucracy. However, as Nohria and Eccles (1992) remind us, ‘‘the structure of any social organization can be thought of as a network’’ (p. 288). For instance, a hierarchy is simply ‘‘a centralized network in which the vast majority of ties flow to or from one particular node’’ (Podolny and Page 1998, p. 59). Networks need to be regarded as social phenomena that may take many forms. Even though they may eventually have an electronic infrastructure, their

123

4

J Educ Change (2010) 11:1–21

nature and impact usually extend beyond this dimension and give way to much broader, more profound and more complex patterns than the mere technological side would suggest. We would be better equipped to understand these social systems if we used the term ‘‘network’’ in a strictly descriptive and analytical sense. Thus, in this paper, I will use the concept to refer simply to a set of actors (technically called ‘‘nodes’’ in the field of social network analysis) and the direct and indirect ties that exist among them (Wasserman and Faust 1994). As we will see later in the paper, the form taken by the set of these ties is one of the factors that more clearly distinguish networks from one another and that may account for important variations in their performance.

Networks in education: some key dimensions For systematically studying networks in education in a more descriptive and analytical way, both at the intra- and inter-organizational level, we need to focus our research efforts on several key dimensions, some of which have received little attention until now from the educational research community.1 In this section, I briefly outline these dimensions and refer to some of the main research questions that we need to address with respect to each of them. Genesis Issues of network genesis refer to the reasons and motivations behind the creation of networks and the factors that lead actors to join them. Who is behind the formation of a network? Does it emerge as a sort of grass-roots movement, perhaps with founding fathers2 from within its field, or is it stimulated by external actors? In other words, is the network an emergent social system or an externally sponsored one? The question of who initiates a network is important because it relates to the critical issue of network ownership. Often, teacher networks and especially inter-school networks are externally sponsored; that is, they are encouraged by central government initiatives and are key elements of states’ efforts to support the implementation of legislated reforms or changes (see, for example, Pennell and Firestone 1996). Contrary to what several recent contributions on the issue suggest, in many cases, organizations are simply left with no option and forced to join. But, besides coercion, there may be other reasons why actors join networks. They may spontaneously and voluntarily organize themselves so as to seek access to particular resources, to profit from economics of scale, to share risks and costs, to gain access to difficult-to-reach domains, to develop products and/or services, to learn from others, to capture first-mover advantages, to lobby collectively, or simply to neutralize or block competitors (Barringer and Harrison 2000). The full understanding of the genesis of intraand inter-organizational networks requires determining which of these (or other) motives apply to a given network.

1

This list of dimensions is by no means intended to be exhaustive. It simply points to a set of key areas where more research is needed.

2

The issue of the importance of founding fathers and its consequences for networks was first raised by Lieberman and Grolnick (1996).

123

J Educ Change (2010) 11:1–21

5

Composition For purposes of classification, network composition involves a basic polarity between individual-actor networks and collective-actor ones. The key defining characteristic of the latter form is that it aggregates whole organizations or institutions, while the former simply comprises individuals. The network concept can thus be applied to both teacher networks (composed of individual actors) and networks of schools (comprised by collective actors and eventually, other institutions outside education). Additionally, networks can be often viewed as organizational actors per se. This involves a further level of complexity which I call supra- or meta-networks; that is, networks of networks, where each network is a distinct actor per se. The meta-network is thus a special case of a collective-actor network. In reality, there can be multiple combinations of these conceptually distinct levels, so that we will often find mixed-level networks, i.e., networks that are comprised simultaneously by individual and collective actors. Structure A key message of the field of social network analysis is that structure matters, both as an antecedent and as a product of social phenomena (Borgatti and Foster 2003). Whereas network structure has been the object of extensive attention in this field, progress in this respect has been much slower within the educational research community (for exceptions, see, for example, Lima 1997, 1998, 2007, 2008; Bakkenes 1996). Relevant network structure issues may be classified into two broad categories: wholesystem structural properties and ones related to individual members’ structural locations (for applications, see Lima 2004, 2008). Since the latter are more relevant for analyzes focusing on specific actors, rather than on whole-networks, in this section I will focus mainly on the former. Issues of whole-system structural properties refer to features of the whole set of relations and positions that comprise a network, abstracted from its members’ individual characteristics and specific network locations. These issues relate to an entire network and can be used to distinguish it from other networks. They refer to aspects such as network density, centralization and connectedness, which are key conceptual tools in the field of social network analysis (Scott 1991; Wasserman and Faust 1994). Density The density of a network refers the extent to which all theoretically possible relations among actors within it are actually activated. High degrees of density suggest strong levels of network integration and of actor cohesiveness (dense networks), while low ones tend to express sparser forms of relating (sparse networks). The denser the relations among a network’s members across a variety of roles and relations, the more integrated it will be as a social system. Networks vary in their degree of integration and it is important to determine this empirically. Baker (1992) distinguished ‘‘thick’’ network organizations (integrated across formal boundaries between organizational units and over many types of socially important relations) from ‘‘thin’’ ones (with little practice of cross-sector integration over multiple types of relations). Current promoters of networks in education advocate that individual and collective actors should maintain a significant amount of interaction with one another,

123

6

J Educ Change (2010) 11:1–21

through extensive, reciprocal exchanges of information and resources—in short, a dense and ‘‘thick’’ pattern of interaction. This view assumes that networks are characterized by high degrees of relational density, but this should be demonstrated empirically rather than merely claimed. When a network’s density is low, isolates emerge—actors who don’t have any ties to other network members (e.g., teachers in a school who never share professional materials or ideas with their school colleagues, or schools that are formally members of a network, but never contribute to the ongoing exchanges that unfold within it). In less dense networks, there may also be a high relative number of pendants (e.g., teachers who are only connected to a single colleague to whom they always ask for professional advice while remaining isolated from the rest of the staff). The proportion of isolated actors and pendants can be informative of the nature of a network (Lima 1997), and the identification of their characteristics and the reasons behind their isolation can reveal the true network that often lies behind the public image that its members, managers or sponsors like to sell. Centralization A network’s degree of centralization describes the extent to which relations and communication patterns within it are centered around one or only a few particularly prominent actors or subgroups. Network structures can range from centralized to decentralized patterns of interaction among their members. This structural feature is particularly useful for understanding power location and distribution in networks, a very much neglected topic in educational research. A common finding in social network analysis is that central positions in a network tend to be associated with power and influence, either within or across organizations (Brass and Burkhardt 1992). A network’s degree of centralization stems from the centralities of its participant members and from the way these centralities are distributed within the network. Conventionally, social network analysts distinguish four main ways in which an actor can be central in his or her network (Freeman 1979; Bonacich 1987): by developing connections with many others (degree centrality), by becoming important as a liaison between otherwise directly unconnected members (betweenness centrality), by maximising the number of direct contacts with others in the network (closeness centrality), and by becoming connected to participants who are themselves connected to many other actors in the network (Bonacich’s modification of the original degree centrality concept). Due to their advocacy of teacher and school resistance to top-down, mandated initiatives, many educational researchers have been reluctant to address power issues among teachers and educational organizations, and this plays out in the way they have analyzed and interpreted interaction in networks. However, as Hargreaves and Fink (2006) have recently recognized, many educational networks are more like clubs or groups than the decentralized structures suggested by the network metaphor. As the authors note, these networks ‘‘tend to have controlling hubs with spokes of communication that extend out to the rest of the network in simple and centralized pathways of influence instead of connecting through multiple nodes of interconnected influence that follow less predictable and geometrically precise patterns’’ (p. 179; see also Busher and Hodgkinson 1995, 1996). A clear empirical example of this is provided in Wohlstetter et al.’s (2003) research on four ‘‘school families’’: the Livingston and Patterson School Families used team meetings for information dissemination rather than for decision making. In the Patterson School Family, the

123

J Educ Change (2010) 11:1–21

7

teams were not conceived as decision-making bodies. Their purpose was to serve as a sounding board for reform ideas and to distribute information to schools in the network. Consequently, meetings consisted mainly of reports by central administrators. Similarly, in the Livingston Family, teams were largely advisory to the School Family Facilitator. The few times we observed decisions being made at team meetings, the decisions concerned matters of minor importance, such as how to distribute tickets at a Familywide event or how to arrange substitute teachers for a professional development day. In this Family, the real decision making was conducted behind the scenes and then brought to the teams for ratification. The School Family Facilitator made all budget decisions, had the power to overrule any team decisions, and also served as the team leader for a majority of the teams. (p. 414) The use of centrality and centralization measures can largely improve our understanding of educational networks. In this respect, difficult but critical questions need to be addressed: What is the amount of direct influence exerted by particular actors on the direction taken by their network as a whole? How far do certain teachers in an intraorganizational context conduct collegial matters according to their own individual agendas? And how far are inter-school networks dominated by particularly powerful institutions, either schools or other types of organizations? Many networks don’t display such centralized patterns as those illustrated above. In fact, in Wohlstetter et al.’s (2003) own research there is ample evidence that other networks display very different patterns of operation. However, to be sure, we need to conduct extensive and systematic analyses of the relational patterns that actually exist within and across these kinds of systems, rather than merely rely on speculation and wishful thinking. Connectedness Connectedness refers to the overall unity of a network. There is fragmentation when the network (either formally or informally) is partitioned into sub-sets of actors where internal cohesion is much greater than their members’ links to the wider network. This partitioning can sometimes compromise a network’s global unity as a relational system. In this respect, network structures can vary between strongly integrated/connected and fragmented forms. In education, the possibility of network fragmentation is real. Indeed, as Busher and Hodgkinson (1996) point out, interschool networks face continuous pressures toward fragmentation3: some of the participant organizations resist moves toward collective action and are keen to preserve their autonomy; others are seduced by other networks, or see additional advantages in joining them; still others believe that the cost of membership and participation is too high to afford. Over time, these centrifugal forces threaten to prevail over centripetal ones. When this occurs, a network is shattered into distinct sectors and individualistic pursuits and its overall effectiveness in accomplishing common gals becomes compromised—it becomes a network only in name, as professional isolation persists as the norm. The field of social network analysis offers several useful concepts and techniques for identifying network fragments, such as components, cliques, clans, social circles, factions, fuzzy groups, etc. (Davis and Carley 2008; Freeman 1992; Scott 1991). These measures

3

For examples of intra-organizational network fragmentation, see Lima (1997, 2007).

123

8

J Educ Change (2010) 11:1–21

can be usefully applied to network contexts, resulting in the identification of relevant subgroups or key components within these systems (Lima 1997). Substance Issues of network substance refer to what network members interact about: missions, purposes, values, social norms, conceptions and perspectives, among many other things. Some key questions in this regard are: Do network actors share core beliefs? What views of teacher professionalism do they endorse? What conceptions of pupil learning do they share? What perspectives about the distribution of power in the educational system and within schools do they endorse? Do variations in these beliefs explain, for example, differences in structure, stability and effectiveness across networks? In this respect, there is a basic difference between networks that operate in relatively circumscribed domains of activity (single-purpose networks) and ones that intervene in multiple domains (multiple-purpose networks). Each of these types of network carries advantages and disadvantages. As Borys and Jemison (1989) note in a more general context, while a broad and evolving purpose (open, continuously updated) may hold the network sufficiently together in the face of disagreement over narrow issues, it may not provide enough detail to adjudicate among distinct interests. Due to the complexity of dealing with several domains in parallel, multi-purpose networks (especially, collectiveactor ones) may take on particular characteristics—for example, they may be less effective in achieving their intended ends (Bell et al. 2006). On the other hand, a narrow and focused purpose may allow partners to be clearer about what they expect from each other, but it may leave many ‘‘important fringe issues’’ unaddressed, preventing the network from moving into promising but not-agreed-upon areas. Research on the substantive side of network life implies performing a series of critical research tasks. The first is the mapping of the number and nature of the domains in which networks intervene. A second task is identifying and classifying the ways networks operate (or not) on the basis of particular value systems. For example, some will operate as affiliation networks (Carmichael et al. 2006); that is, as networks whose participants share a commitment to concepts, principles or communities. Others may operate in a way similar to social movement or ‘‘professional-area movement’’ organizations whose programs are ‘‘aligned with other organizations in a larger change effort sharing a general philosophy and seeking similar results’’ (Pennell and Firestone 1996, p. 48). A third research task involves determining the extent to which there is substantive consensus or dissension among network members with regard to the fundamental meaning of their collective action. Although it may be convenient for researchers (and even participants) to assume that educational networks embody a purpose that is shared by all members, this may actually be far from reality. This third research task underscores the importance of network identity. As Berry et al. (2004) put it, in this respect, there is the disturbing possibility that perhaps networks are merely abstractions in the minds of scholars, and we are the ones who attach meaning to them, irrespective of what their own participants believe their activities mean. Perhaps to the extent that such structures exist, they are merely temporary constructions in the minds of their participants rather than persistent phenomena for which participants share a common understanding. Or perhaps on close examination we might find that (…) most network participants hold varied and

123

J Educ Change (2010) 11:1–21

9

unique understandings of why a particular network exists, who is involved in it, and why they, as individuals, are involved (pp. 546–547). Effectiveness With respect to network effectiveness, the key issue is: How do we know that a given network is making a difference? Or, to put in more adequately, how do we know that this difference is larger than the one that it would make if its members were not organized as a network? This leads to several more specific questions: How does the network affect the actions and performance of its members and of those with whom they interact? How strong is the evidence, for example, of tangible changes in classroom practice and in student learning as a result of the activities undertaken at the network level? To what extent do networks have systems in place that check if work done at the network level cascades back to schools and classrooms, as intended? In this regard, some networks will be effective, while others may be ineffective. As Kahne et al.’s (2001) research shows, being a member of a network is not, by itself, a sufficient condition to generate changes in how a teacher or a school approaches and deals with teaching and learning: a network can often operate in an uncoordinated and incoherent fashion and its activity may be weakly related to the core of classroom life. Currently, in education, there is too much emphasis on network success and too little credible research to support this claim. Whereas a recent systematic research review commissioned by England’s National College for School Leadership (Bell et al. 2006) suggests that ‘‘networks can be a highly effective vehicle for improving teaching, learning and attainment’’ (p. 6), the evidence presented is not overwhelming—of the eleven studies included in the review that investigated and reported on networks’ impact on pupil attainment, achievement or engagement, only six reported a high level of impact. Besides, the authors themselves recognize that ‘‘the studies are rarely explicit about what the networks they examine have contributed to what schools and others do to improve pupil learning’’ (p. 18). Dynamics Network dynamics is the least well-know aspect of networks and probably the most difficult one to do research on. Methods for the characterization and visualization of system dynamics in networks are still evolving.4 However difficult the task, it is crucial that we move from images of networks as fixed sets of actors and relations to more dynamic views of them as moving systems that may be constantly rebuilt and reshaped by the actions and interactions of their members. Here, the key research question is: Do networks develop and change over time? If so, how does this occur? The dynamic nature of network life is recognized by prominent network participants and promoters. For example, Veugelers and Zijlstra (1996) warn that ‘‘networks must develop gradually. It is not sufficient to put schools together and call them a network’’ (p. 78). It is crucial that we pin down and document how these development processes unfold. This calls for a longitudinal approach to the study of networks.

4

Tom Snijders (in Groningen and Oxford) and Matthias Finger (in Berlin), among others, are currently doing cutting-edge work and developing analytical software in this area.

123

10

J Educ Change (2010) 11:1–21

Over time, networks follow specific types of trajectories. In this respect, Kilduff and Tsai (2003) distinguish two change processes that differ fundamentally from one another in terms of structural dynamics: • Goal-directed network trajectories. Network change is driven primarily by goal directedness. Network participants see themselves as part of a whole and are committed to clear and specific network-level goals. All relations between network members are structured in order to achieve these goals. Usually, an administrative entity (a steering committee, a coordinating council, a leadership team, or the like) is formed and endowed with the function of planning and coordinating the activities of the network as a whole. • Serendipitous network trajectories. In this case, ‘‘network trajectories develop haphazardly from the interactions of individual actors’’ (p. 89); no network levels drive the actions of network members. At times, groups of actors may have an interest in sharing some goals, but at others they may not. Additionally, ‘‘actors make choices about who to contact with, what to transact, and so on, without guidance from any central network agent concerning goals or strategy’’ (p. 90). In practice, networks will often exhibit some combination of both kinds of dynamic processes. Over time, it may even be possible that networks alternate between one mode of evolution and the other. However, one type may prevail over the other in a particular network and this may be important for characterizing it and understanding its operation and outcomes. An excellent illustration of the dynamics of networks in education and of the problems associated with it is provided by Lieberman and Grolnick (1996), who note a tendency for networks to move from informality to formality, and from structural looseness and flexibility to rigidity, as they become more successful and expand. With current trends toward the constitution of networks at the local, regional, state, national, and international levels, sometimes gathering thousands of institutions, there is reason to believe that these growth processes may become even more problematic for networks than they were a decade ago. To summarize, the network issues highlighted so far suggest a provisional list of key dimensions for the analysis of networks in education (Table 1). It is worth emphasizing that, in each dimension or sub-dimension, the network types that are listed should not be regarded as polar constructs; the polarities simply express theoretical extremes between which any network can vary and this may be different at different points in time. Also, the proposed analytical framework allows for the application of multiple combinations of types in a given network. For example, an individual-actor network may be emergent, dense, decentralized, strongly connected, multi-purpose, ineffective and serendipitous; or a collective-actor network may be externally sponsored, sparse, centralized, fragmented, single-purpose, effective and goal-directed. The point of the framework is to provide a specific language and precise criteria for the classification of networks, so that this ceases to be a normative judgement and becomes a more systematic description of concrete network patterns and characteristics, identified on the basis of empirical data collected in actual research work.

The interactional side of network life Currently, in education, our knowledge of the abovementioned dimensions of networks is still relatively poor. But even if we had enough information on these general issues, we

123

J Educ Change (2010) 11:1–21

11

Table 1 Key dimensions for the analysis of networks in education Dimension

Sub-dimension

Genesis

Composition

Structure

Density Centralization

Connectedness

Substance

Effectiveness Dynamics

Type

Description

Emergent

Spontaneous, voluntary, generated through ‘‘grassroots’’ processes

Externally sponsored

Conceived and promoted by outside agents as an instrument for their own agendas

Individual-actor

Comprised by individuals

Collective-actor

Comprised by groups, organizations, institutions, etc.

Dense

Most possible relations are activated

Sparse

Few of the possible relations are activated

Decentralized

No particular actor or set of actors is relatively prominent

Centralized

One actor or a small set of actors are the focus of most of the network’s relational activity

Strongly connected

Most actors are strongly connected to one another

Fragmented

Subgroups of actors within the network are weakly connected or unconnected to one another

Single-purpose

Devotes most of its activity to a single purpose

Multi-purpose

Devotes its activity to several purposes

Effective

Achieves its declared aims

Ineffective

Fails to achieve its declared aims

Goal-directed

All relations between network members are structured in order to achieve network-level goals; an administrative entity plans and coordinates the activities of the network as a whole

Serendipitous

Evolves haphazardly from the interactions of individual actors, without guidance from any central network agent

would still need to go beyond them. The general framework that I proposed previously is intended to be a useful tool for the general classification of networks, but, by itself, it runs the risk of ignoring the actual interactional aspects of network life, both within and across networks. This is also a key issue for educational researchers to focus on. We need to know much more about how actors interact in different kinds of networks and what their actual experiences are. In the following sections, I illustrate this argument with a discussion of networks’ internal processes, their relations with outside entities, their dark side and their dissolution. Internal network processes How do different types of networks operate? How are they led? How do network actors participate and communicate with one another and how does this affect their learning and

123

12

J Educ Change (2010) 11:1–21

the quality of their relations? These and other process questions raise issues of network management and leadership, participation, learning, and interpersonal communication and trust among network members. Management and leadership The study of management issues in educational networks requires, first of all, information on their operational and coordination mechanisms. For example, many interschool networks are usually coordinated by the leaders of the participant institutions, who usually meet on a regular basis. Are these top leader meetings formally organized? Are minutes or notes taken? How is chairing ensured? Does it occur on a fixed, or on a rotating basis? What are the meeting places? Do they vary? Why? And how are network financing and budgeting handled? What are the consequences of this for the development of the network and for its participants? Leadership processes in networks are more difficult to capture than these more formal aspects, among other reasons, because often formal roles do not correspond to actual leadership practices. In a recent study of leadership in secondary school subject departments that utilized a social network approach (Lima 2008), I showed how leadership patterns can vary among administrative units within the same organization, and how they can differ radically from what is formally predicted. Thus, rather than viewing leadership in networks as a mere activation of the powers with which certain actors are formally endowed, it may be more fruitful to regard it as a quality or property that emerges from processes of social interaction among network members. A key question in this respect is: How far is leadership actually emergent, distributed or shared in a given network? Network advocates claim that distributed patterns are typical of these organizational forms, but, to my knowledge, in the educational literature, no single study has yet supported this widely held belief (see, for example, the research review by Bell et al. 2006). Participation In collective-actor educational networks, it is vital that we understand not only what organizations belong to the network, but also which members of those collective actors actually participate in it. Usually, regular participants in most networks tend to be school managers or teachers with specific coordination responsibilities in their home institutions. The generally held but largely untested assumption is that these individuals serve as bridges between the network and actors affiliated with the participant organizations, but not directly involved in the network’s activities (Day and Hadfield 2004, p. 581; Veugelers and Zijlstra 2002, p. 170). But how well-connected in their organizations are these actors? How much of the benefits of network participation go strictly to them, rather than to their home organization as a whole? Many teachers who are engaged in inter-organizational networks may be actually alienated or distanced (even ostracized?) from their school colleagues. This was recognized by Lieberman and McLaughlin (1992), who spoke of ‘‘the sometimes chilly reception that network teachers encounter in their schools or departments’’ (p. 675). This is also evident in Rusch’s (2005) documentation of ‘‘cultural rifts between network members and local district peers,’’ which led the author to observe that ‘‘network members report hostility, ruptured relationships, and marginalization when they attempt to share their learning in their home school district’’ (p. 87; see Goldsberry 1995, for similar results). It is clear that the impact of educational networks will depend strongly on their members’

123

J Educ Change (2010) 11:1–21

13

connectedness within their local work environments. If they are not at least relatively central actors in their own schools, the so-called ‘‘inter-organizational’’ networks may be little more than mere networks of acquainted individuals who happen to belong to different organizations. For networks deeply committed to educational change, this is one key reason why organizers/network managers should take care in recruiting members whose network properties in their home institution make them adequate for the dissemination of change initiatives and innovative practices. To rely solely on voluntary participation in this respect is to risk assuming a change potential that may not actually exist. Learning In education, there is a common belief that networks generate powerful professional learning. But does learning occur all the time, in every network? And if so, what kind of learning? In some networks, learning can be experience-driven; in others, in can be based merely on the transmission of abstract and de-contextualized ‘‘expert’’ knowledge. There are networks in which learning in mostly collegial and horizontal (Veugelers and Zijlstra 2002, p. 169), and others where it is centered almost exclusively on one-to-one relations with a small group of experts (McDonald and Klein 2003; Wohlstetter et al. 2003). In a recent paper, Stoll (2005) summarized the characteristics of professional learning communities in education under five headings: ‘‘shared values and vision that focus on improving learning and teaching; collective responsibility for the learning of all pupils; reflective professional inquiry to deepen practice; collaboration and teamwork; group and collective learning, as well as individual learning’’ (p. 34). Earl and Katz (2005) also identified seven interdependent ‘‘key features of learning networks’’: explicit purpose and clear focus; ongoing and dynamic relationships between network members; collaboration; enquiry; shared leadership; external accountability and active self-monitoring, and capacity-building and support. These and other relevant key factors could be combined and/or developed into criteria that would then be used to assess whether or not a given network is actually a professional learning community. Again, to assess this, it is not sufficient to determine if key participants in collective-actor networks develop these qualities among themselves; it is also crucial that their home organizations as a whole exhibit the same patterns of relating and learning. Interpersonal relations and trust Formal or prescribed relations do not entirely capture the complete set of ties that make up an organization or a network of collective actors. Informal, emergent relationships that comprise ‘‘hidden’’ networks are just as important and can be of great significance for understanding organizations. Important issues in this respect are the extent to which relations between network members are based on interpersonal ties and the degree to which these constrain or promote the goals of the network. Krackhardt and Stern (1988) suggested that organizations where there are friendship ties that cut across departments may be better prepared to deal with environmental changes and uncertainty than those where these kinds of ties do not exist. Krackhardt (1992) also argues that ‘‘the affective component of the strong tie is important in understanding the dynamics surrounding crises and changes in organizations’’ (p. 221). But ties that are too strong (i.e.,

123

14

J Educ Change (2010) 11:1–21

extensive, exclusive, reciprocal, and intimate) can also be stifling and redundant, and this can have negative implications for the learning potential of a network (Lima 1997, 2001). Some studies also suggest that effective communication among network participants requires face-to-face interaction. The issue is particularly significant when we consider the huge size of some contemporary educational networks. One reason why face-to-face contact may be so crucial in network life is that it can be vital for the generation of trust among network members. As Nohria and Eccles (1992) put it bluntly, ‘‘it is hard to trust someone without—as they say—a hard look in the eye’’ (p. 298). Writing about educational networks, Day and Hadfield (2004) also stress that ‘‘trust is the glue that holds partnerships together’’ (p. 583). It may be the case that networks that do not operate mainly on the basis of face-to-face contact may promote shallower forms of relating and learning, when compared to ones where members meet regularly face-to-face. There is an obvious lack of empirical research in this respect. Furthermore, while we need to know how trust is established in networks, we also need to understand what the limits of trust are in these systems. Are there boundaries that network members never seem willing to cross, no matter how trustful their relationships seem to be? What is then the actual balance between knowledge-sharing and secret keeping and what are the consequences of this for learning? Do actors tend to share primarily what they regard as trivial and keep secret what they regard as really vital? Do they tend to strategically give away only those bits that are less critical for their survival? In a competitive educational world, how far can trust reach? And how do these issues relate to the nature and structure of networks? We need to gain more direct insight into the processes by which interpersonal and trust relations build up in and influence the operation of networks. Network ecology Network ecology refers to the external relations that a network maintains with outside entities, including other networks. This side of network life may play itself out in quite complex patterns. For example, Busher and Hodgkinson’s (1996) work illustrates how complex webs of educational networks can emerge in education: secondary schools can form inter-school networks with similar institutions from the same schooling phase; simultaneously, each can form a network with its main feeder primary schools; moreover, some of the head teachers of these and other institutions can join together in a schoolleader network, while some of its staff can integrate still other professional networks, for example, ones linked to pedagogical or subject-centered renewal movements. Thus, networks of networks can emerge, often cross-cutting one another and forming interlocking structures. Besides the identification and classification of these complex forms, researchers need to pay close attention to the communication and power patterns established between a network and other networks, and between it and more conventional entities, such as school districts, government bodies, political parties, civil society associations, private foundations, higher education educational institutions, the press, and so on. An excellent example of research on educational networks that illuminates the issue of network ecology and its implications for educational change is Rusch’s (2005) study of the Pathfinder network’s relations with schools and administrators not affiliated with the network. A key finding of this study was that ‘‘the fear of sibling rivalry [between schools in a network and schools that were nonmembers of a network] and the fear of competition on the part of central administrators and board members literally froze the [districts’] systems’’ (p. 110). The author also found that the schools that were members of the

123

J Educ Change (2010) 11:1–21

15

Pathfinder network ‘‘were treated like most innovations in institutionalized environments—that is, as an anomaly that could exist within the system as long as it did not disrupt the system.’’ In this respect, an important task for network researchers is to investigate the actions that networks develop to conquer legitimacy over desired domains. Benson (1975) has written that ‘‘the possession of a domain permits the organization to operate in a certain sphere, claim support for its activities, and define proper practices within its realm’’ (p. 232, emphasis added). This entails viewing relations among networks, and between each of them and other entities, as processes of social struggle aimed at the occupation of spaces of influence and domination. The dark side of networks The dark side of networks is one of the least pursued themes in network research. It involves patterns of behavior and operation related to dysfunctions, destructive conflicts, exploitation and other unforeseen negative effects associated with network constitution and activity. A clear example of this line of research is Rusch’s (2005) study, where the author found that an unintended consequence of the activity of the network was an enhanced culture of competition in the districts into which the schools that comprised the network were formally integrated. The general literature on inter-organizational relationships suggests a few other potential drawbacks that can result from actors’ participation in networks (Barringer and Harrison 2000). Some of these factors may be relevant for networks in education: • Loss of proprietary information (e.g., the risk that teachers do not acknowledge colleagues’ authorship rights to an innovation that is disseminated throughout the whole network); • Management complexities (e.g.., the risk that the need for multiple cross-level meetings and ongoing collective discussions of issues results in costly delays and widespread feelings of frustration); • Financial and organizational risks (e.g., the risk of opportunistic behavior from network partners, such as when members of a partner organization may take advantage of insider information to make sure that they enrol the best students or gain other competitive advantages, such as additional funding from programs that involve competition between individual schools or even between individual teachers); • Power imbalance (e.g., the risk of learned dependence, such as when teachers become complacent because they feel network colleagues will always come up with solutions and help, or when a school that sponsors the improvement of another tends to hand over pre-packaged solutions to problems rather than invest in developing the partner’s capacity for self-directed development). This risk is particularly high in networks that prize the expert rather than the teacher, that favor delivered rather than constructed learning environments and that put a premium on the faithful replication of training models and contents rather than on their adaptation to teachers’ local contexts (McDonald and Klein 2003); • Partial loss of autonomy in decision-making; • Potential clash between partners’ ‘‘cultures’’ (e.g., the risk of conflict when a network brings together schools with a more elitist, cognitive-oriented tradition and others that are more keen on developing the ‘‘whole child’’);

123

16

J Educ Change (2010) 11:1–21

• Loss of organizational flexibility (e.g., the risk that a partnership with a set of organizations forecloses the possibility of establishing relationships with others; or that organizational routines created by alliances make independent action difficult). The list is open to empirical determination, but a few more issues come to mind when thinking about educational organizations. Firstly, networks (especially externally sponsored ones) face the risk of political co-optation, ideological appropriation, and colonization by external interests. This may occur, for example, in officially sponsored network programs in which teachers selected by the state or the district to become network members are trained to operate as key players and advocates for specific reforms in their schools, using a variety of proactive influence strategies to seduce principals and colleagues into their perspectives. Another risk is that of work overload. Participation in networks can be a burden for teachers who are already overloaded with work in their home institution. They may simply not have any more time and energy to devote to additional activities, and releasing teachers from classroom duties during significant periods of time is expensive and thus unaffordable for many institutions. For example, in their research on four ‘‘school families,’’ Wohlstetter et al. (2003) found that teachers in all four School Families were faced with the pressing regular demands of their classroom work, plus the opportunities presented by committee membership and additional training. In some cases, they chose only to do extra work when they were compensated financially, and programs that were not accompanied by financial incentives often died out (p. 419). This leads to a further risk, that of network participants tending to be, as is common, middle- or senior-level school leaders, rather than ordinary classroom teachers. If there are few opportunities for regular classroom teachers to participate in networks, the divide between them and administrators is likely to widen. In addition, networks’ management or steering groups made up largely of principals or their direct representatives from the participating schools may tend to take a managerial perspective on teaching and learning issues, which can lead to a bureaucratization of pedagogy. Another risk, directly related to the previous one, is the potential exclusion of other relevant educational actors (namely, school governors, parents, local community representatives, etc.) from network activities. The estrangement of parents is evident, for example, in Wohlstetter et al.’s (2003) study, where a parent in the Patterson School Family told the interviewers: ‘‘Stakeholders not on the committee only find out what is happening if they are married to someone on the committee … if you are not on the Task Force, you are clueless. You don’t know what is going on (p. 417).’’ Finally, there is the risk of too much integration. In particular, when dense, tightly-knit groups of actors who passionately believe in their pedagogical or political ideas are formed, they may develop strong self-reinforcing feedback mechanisms and thus become particularly insular. For example, Rusch (2005) found that ‘‘the Pathfinder principals who were actively engaged in formal and substantive talk about (…) complex educational problems in their schools and within the network were not engaging with their peers in the same conversation and exhibited the same defensive behaviors as their district peers’’ (pp. 104–105). Communication within a network tends to become easy and self-assuring, but insularity from the outside world may come at the cost of preventing the network from effectively disseminating its potential benefits throughout the system or systems in which it is located. Educational researchers need to explore these dangers of network homophily

123

J Educ Change (2010) 11:1–21

17

(the tendency for similar actors to cluster together in particularly insular ways) in more depth, because these can lead to a convergence toward groupthink (Lima 2001) and to an excessive dependence on effective leaders (Berry et al. 2004), thus giving way to lower levels of network effectiveness. Disengagement and dissolution The current enthusiasm with network forms of organization is difficult to reconcile with the fact that, at least in the business sector, ‘‘an extremely large fraction of network forms of organization do not perform the function for which they were designed’’ (Podolny and Page 1998, p. 71). The simple truth is that, ultimately, the majority of inter-organizational relationships fail (Barringer and Harrison 2000). There are no studies of the failure rate of networks, either in education or in any other sector, but journalistic and managerial sources quoted by Podolny and Page (1998) suggest that at least one common type of network organization—the network alliance—fails in extremely high proportions (for example, in the airline industry, fewer than 40% of regional alliances and fewer than 30% of international alliances were successful). Data on joint ventures also show that less than one third are considered successes. Inter-organizational relationships are complex and difficult to manage. Networks often strive to maintain an unstable, usually fragile balance between competing forces within them. Today, in education, almost every school organization experiences a tension between competition and cooperation. This can lead school leaders into an ambivalent attitude toward networking: some value it, but are eager to resist common initiatives (e.g., the establishment of common records of achievement across schools) that they regard as infringements of their autonomy. To avoid the risk of serious collusion among participants, some inter-school networks adopt the principle of subsidiarity, according to which schools participating in a network may choose to ignore group discussions (Busher and Hodgkinson 1996). However, this solution itself can become a problem, because failing to adopt collective decisions may be a first step toward animosity from other partners and may ultimately lead to reciprocal, circular and mounting processes of relational disengagement. Thus, the stability of a network is in constant danger of dissolution. We need to determine what structural features and process management patterns lead to network failure. As Miles and Snow (1992) warn, network organizations ‘‘can fail because of alterations made by well-intentioned managers’’ (p. 62). This failure may occur when actors’ expertise becomes too narrow and their role is assumed by other entities, or when the network establishes excessive and rigid mechanisms to prevent partner opportunism or exclusive relations between particular sets of partners. Lieberman and McLaughlin (1992) have also warned that ‘‘like a business that is successful and expands too fast, a network can expand too rapidly and risk falling into a kind of intellectual bankruptcy’’ (p. 676). There are many ways a network may lose its vitality over time and eventually collapse. We know very little about these scenarios and need to engage much more deeply in their examination.

Conclusion The concept of the ‘‘network organization’’ has been heavily promoted by the popular business press and management consultants in recent years and is also rapidly gaining wide currency among researchers, practitioners, and policymakers in education. Today,

123

18

J Educ Change (2010) 11:1–21

educational literature references based on faith about the positive potential of networks far outnumber research studies on actual patterns of interaction within and across organizations, especially at the collective-actor level. Additionally, most scholarly texts on networks in education tend to emphasise their advantages and positive outcomes over their negative consequences or their failure. As we have seen, the risks that networks face are huge and should not be underestimated. While it is important that we explore the advantages and the added-value that these systems can bring about as effective instruments for professional renewal, school improvement and pupil learning, it is also important that we recognize and investigate their potential disadvantages. A more balanced perspective is clearly called for to guide theory and research in this respect. This paper has proposed a list of key dimensions of networks and ways of analyzing them in a less normative, more descriptive and analytical manner. Although the research issues outlined in the paper were presented in a sequential manner, they should not be regarded as discrete. Indeed, they are interwoven in multiple and complex ways. For example, actors’ reasons for participating in a network (a genesis issue) can be linked to underlying power processes (a structural and dynamics issue), such as when teachers attend network-organized meetings because they feel pressured to learn how to accommodate new state requirements, or simply because their principals demand them to do so. It is also important to keep in mind that actor power in networks should also be viewed as a process variable, for it is achieved via interaction among actors, that is, through processes not only of coercion and imposition, but also of negotiation, interpersonal influence, acceptance and even active compliance. There may also be important relations between network power and other structural issues. For example, governments may prefer some network forms over others. They may tend to favour more centralized (‘‘hub and rail’’) and predictable (single-purpose, goal-directed) networks over more decentralized and unpredictable ones. Just as well, network dynamics may give way to significant changes in network structures, internal processes, or even purposes. These are just a few examples of the many theoretical and empirical inter-relationships that exist among the different research themes that have been highlighted in this paper. Besides helping us reach more precise, systematic classifications of networks, the use of the concepts and dimensions that are proposed in the present paper may also help illuminate practice and decisions about practice in education. A clear example relates to the increasingly common use of networks as reform strategies for teacher learning. It may be the case, for example, that in many contexts, externally sponsored, sparse and highly centralized networks, operating mostly in specific single-purpose domains and designed so as to follow strictly goal-directed paths, stifle rather than stimulate the development and sustainability of professional communities of practice. In other contexts, significantly low levels of teacher competence and collective capacity may call for a higher level of network centralization and a more goal-directed type of network dynamics. In other words, network design for teacher learning may be developed on the basis of the categories proposed in this paper, which may also be used to assess the properties of existing networks at any given point in time. Throughout this paper, I have repeatedly emphasized the need for educational researchers to attend to the complexity of networks in education. This has implications for the research designs that they employ. An extended discussion of this is beyond the scope of the present paper, but my section on the interactional side of network life suggests, at least implicitly, some of the kinds of fundamental data that need to be collected. Still, a few additional, brief and concluding notes on the issue are warranted in this respect. A first necessary step—and one that has not been pursued systematically in the field of

123

J Educ Change (2010) 11:1–21

19

educational research until now—is to bring into the study of educational networks several concepts, techniques and tools that have long been developed and applied in the field of social network analysis, some of which (such as network density, centralization or fragmentation) are basic conceptual building bocks of the present paper. However, caution is needed in this respect: formal social network analysis methods should be regarded as only a set of tools among many other possibilities that are available and adequate for this kind of study, rather than as the ultimate tool for this type of research. Indeed, despite its potential, social network analysis techniques and conventional research strategies also suffer from important limitations, namely, a tendency to develop a static perspective on networks and to overemphasize the primacy of structure over actor agency. As I have emphasized, it is also crucial that we examine network structures and processes micro-analytically. In particular, the study of processes such as learning, trust-building, identity and change in networks calls for phenomenological and ethnographic approaches that will complement more conventional structural modes of inquiry and interpretation in network analysis. It is true that case studies by themselves do not allow for the substantiation of extended structural phenomena that develop across multiple actors, but they are vital for capturing the meaning of actors’ experiences within those structures and for understanding the way they build and rebuild the very structures within which their individual and collective action unfolds.

References Andrews, K., & Rothman, M. (2002). Cultivating innovation: How a charter/district network is turning professional development into professional practice. Phi Delta Kappan, 83(7), 507–512. Baker, W. E. (1992). The network organization in theory and practice. In N. Nohria & R. G. Eccles (Eds.), Networks and organizations: Structures, form, and action (pp. 397–429). Boston, MA: Harvard Business School Press. Bakkenes, I. (1996). Professional isolation of primary school teachers: A task-specific approach. Leiden, The Netherlands: DSWO. Barringer, B. R., & Harrison, J. S. (2000). Walking a tightrope: Creating value through inter-organizational relationships. Journal of Management, 26(3), 367–403. Bell, D. (1973). The coming of post-industrial society: A venture in social forecasting. New York: Basic Books. Bell, M., Jopling, M., Cordingley, P., Firth, A., King, E., & Mitchell, H. (2006). What is the impact on pupils of networks that include at least three schools? What additional benefits are there for practitioners, organizations and the communities they serve? Nottingham: National College for School Leadership. Retrieved from http://www.ncsl.org.uk. Benson, J. K. (1975). The inter-organizational network as a political economy. Administrative Science Quarterly, 20(2), 229–249. Berry, F. S., et al. (2004). Three traditions of network research: What the public management research agenda can learn from other research communities. Public Administration Review, 64(5), 539–552. Bonacich, P. (1987). Power and centrality: A family of measures. American Journal of Sociology, 92(5), 1170–1182. Borgatti, S. P., & Foster, P. C. (2003). The network paradigm in organizational research: A review and typology. Journal of Management, 29(6), 991–1013. Borys, B., & Jemison, D. B. (1989). Hybrid arrangements as strategic alliances: Theoretical issues in organizational combinations. Academy of Management Review, 14(2), 234–249. Brass, D. J., & Burkhardt, M. E. (1992). Centrality and power in organizations. In N. Nohria & R. G. Eccles (Eds.), Networks and organizations: Structures, form, and action (pp. 191–215). Boston, Massachusetts: Harvard Business School Press. Busher, H., & Hodgkinson, K. (1995). Managing interschool networks: Across the primary/secondary divide. School Organization, 15(3), 329–339. Busher, H., & Hodgkinson, K. (1996). Co-operation and tension between autonomous schools: A study of interschool networking. Educational Review, 48(1), 55–64.

123

20

J Educ Change (2010) 11:1–21

Carmichael, P., et al. (2006). Learning how to learn in classrooms, schools and networks: Teachers’ networks in and out of school. Research Papers in Education, 21(2), 217–234. Castells, M. (2001). The Internet galaxy: Reflections on the Internet, business and society. Oxford, UK: Oxford University Press. Chapman, J., & Aspin, D. (2003). Networks of learning: A new construct for educational provision and a new strategy for reform. In B. Davies & J. West-Burnham (Eds.), Handbook of educational leadership and management (pp. 653–659). London: Pearson/Longman. Davis, G. B., & Carley, K. M. (2008). Clearing the FOG: Fuzzy, overlapping groups for social networks. Social Networks, 30, 201–212. Day, C., & Hadfield, M. (2004). Learning though networks: Trust, partnerships and the power of action research. Educational Action Research, 12(4), 575–586. Earl, L., & Katz, S. (2005). What makes a network a learning network? National College for School Leadership, UK. Retrieved from http://www.ncsl.org.uk/networked/networked-research.cfm. Freeman, L. C. (1979). Centrality in social networks: Conceptual clarification. Social Networks, 1(3), 215–239. Freeman, L. C. (1992). The sociological concept of ‘‘group’’: An empirical test of two models. The American Journal of Sociology, 98(1), 152–166. Goldsberry, L. (1995). The evolution of a restructuring school: The New Suncook case. In A. Lieberman (Ed.), The work of restructuring schools (pp. 136–156). New York: Teachers College Press. Hargreaves, A., & Fink, D. (2006). Sustainable leadership. San Francisco, CA: Jossey-Bass. Kahne, J., O’Brien, J., Brown, A., & Quinn, T. (2001). Leveraging social capital and school improvement: The case of a school network and a comprehensive community initiative. Educational Administration Quarterly, 37(4), 429–461. Kilduff, M., & Tsai, W. (2003). Social networks and organizations. London: Sage. Krackhardt, D. (1992). The strength of strong ties: the importance of philos in organizations. In N. Nohria & R. G. Eccles (Eds.), Networks and organizations: Structures, form, and action (pp. 216–239). Boston, MA: Harvard Business School Press. Krackhardt, D., & Stern, R. (1988). Informal networks and organizational crises: An experimental simulation. Social Psychology Quarterly, 51(2), 123–140. Lieberman, A., & Grolnick, M. (1996). Networks and reform in American education. Teachers College Record, 98(1), 7–45. Lieberman, A., & Grolnick, M. (1998). Educational reform networks: Changes in the forms of reform. In A. Hargreaves (Ed.), International handbook of educational change, Part One (pp. 710–729). The Netherlands: Kluwer. Lieberman, A., & McLaughlin, M. W. (1992). Networks for educational change: Powerful and problematic. Phi Delta Kappan, 73(9), 673–677. Lieberman, A., & Wood, D. (2003). Inside the National Writing Project: Connecting network learning and classroom teaching. New York: Teachers College Press. Lima, J. A. (1997). Colleagues and friends: Teachers’ professional and personal relationships in two Portuguese secondary schools. Unpublished Ph.D. Dissertation. University of the Azores, Portugal. Lima, J. A. (1998, April). Improving the study of teacher collegiality: Methodological issues. Paper presented at the American Educational Research Association Conference, San Diego. Lima, J. A. (2001). Forgetting about friendship: Using conflict in teacher communities as a catalyst for school change. Journal of Educational Change, 2(2), 97–122. Lima, J. A. (2003). Trained for isolation: The impact of departmental cultures on student teachers’ views and practices of collaboration. Journal of Education for Teaching, 29(3), 197–218. Lima, J. A. (2004). Social networks in teaching. In F. Hernandez & I. Goodson (Eds.), The social geographies of educational change (pp. 29–46). Dordrecht/Boston/London: Kluwer. Lima, J. A. (2007). Professional development in departmentalised organisations: Lessons for school improvement from a case study of two curriculum departments. School Effectiveness and School Improvement, 18(3), 273–301. Lima, J. A. (2008). Department networks and distributed leadership in schools. School Leadership & Management, 28(2), 159–187. McDonald, J. P., & Klein, E. J. (2003). Networking for teacher learning: Toward a theory of effective design. Teachers College Record, 105(8), 1606–1621. Miles, R. E., & Snow, C. C. (1992). Causes of failure in network organizations. California Management Review, 34(4), 53–72. Nohria, N. (1992). Is a network perspective a useful way of studying organizations? In N. Nohria & R. G. Eccles (Eds.), Networks and organizations: Structures, form, and action (pp. 1–22). Boston, MA: Harvard Business School Press.

123

J Educ Change (2010) 11:1–21

21

Nohria, N., & Eccles, R. G. (1992). Face-to-face: Making network organizations work. In N. Nohria & R. G. Eccles (Eds.), Networks and organizations: Structures, form, and action (pp. 288–308). Boston, Massachusetts: Harvard Business School Press. Pennell, J. R., & Firestone, W. A. (1996). Changing classroom practices through teacher networks: Matching program features with teacher characteristics and circumstances. Teachers College Record, 98(1), 46–76. Perrow, C. (1992). Small-firm networks. In N. Nohria & R. G. Eccles (Eds.), Networks and organizations: Structures, form, and action (pp. 445–470). Boston, Massachusetts: Harvard Business School Press. Podolny, J. M., & Page, K. L. (1998). Network forms of organization. Annual Review of Sociology, 24, 57–76. Rusch, E. A. (2005). Institutional barriers to organizational learning in school systems: The power of silence. Educational Administration Quarterly, 41, 83–120. Scott, J. (1991). Social network analysis: A handbook. London: Sage. Stoll, L. (2005). Developing professional learning communities: messages for learning networks. National College for School Leadership, International perspectives on networked learning. Retrieved September 4, 2006, from http://www.ncsl.org.uk/networked/networked-research.cfm Veugelers, W., & Zijlstra, H. (1996). Networks for modernizing secondary schools. Educational Leadership, 54(3), 76–79. Veugelers, W., & Zijlstra, H. (2002). What goes on in a network? Some Dutch experiences. International Journal of Leadership in Education, 5(2), 163–174. Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications. Cambridge: Cambridge University Press. Wohlstetter, P., Malloy, C. L., Chau, D., & Polhemus, J. L. (2003). Improving schools through networks: A new approach to urban school reform. Educational Policy, 17, 399–430.

123