Communities of Transformation: Framing: Social Network and Community of Practice


Section 2: Framing: Social Network and Community of Practice

We drew primarily on two bodies of literature when designing this study: social network analysis and literature on communities of practice (CoPs). Here we review the literature that informed the study design, as well as pushed us forward in our analyses.

Benefits of Networks

Social network analysis focuses more on outcomes, as compared to the literature on communities of practice; thus we used social network analysis to frame our understanding of outcomes (described in section 5). Social networks are defined in the literature as people loosely connected through some form of interdependencies, such as values, preferences, goals, or ideas (Wasserman & Faust, 1994). In general, outcomes are the most commonly studied aspect of networks. Since we had so much rich data in this area, this study focused on a narrower question: How might STEM networks be unique in terms of outcomes that promote change?

Diffusion of innovation or change has emerged as an outcome in many different studies of social networks, which is why social network analysis has been applied to the study of change processes in more recent years (Rogers, 2003; Valente, 1995). Three specific outcomes of social networks have been related to change: learning, social capital, and risk-taking (Borgatti & Foster, 2003; Burt, 2000; Kilduff & Tsai, 2003; Tenkasi & Chesmore, 2003). Many researchers have found a strong linkage between learning and changes in behavior, showing that as people interact with others in networks they are more likely to experience schema changes allowing openness to new approaches (Tenkasi & Chesmore, 2003). Networks also provide social capital that facilitates the change process by providing access to relationships and knowledge about how to overcome barriers (Burt, 2000). While different definitions of social capital exist, most of the theoretical discussions operate under a definition of social capital as the resources embedded in social relations and social structure which can be mobilized by an actor to increase the likelihood of success in purposive action (Daly & Finnigan, 2009). These resources can vary to include knowledge about how organizations work, influence possessed by particular people, or access to financial resources. Finally, achieving long-term change often requires risk-taking that can be less problematic if it is done collectively rather than individually (Valente, 1995). If a person knows that many of her peers are going to join her in an activity or behavior, she is more likely to feel empowered to engage in this behavior (Rogers, 1962; Valente, 1995). While there are other important components in the efficacy of networks, these three—learning, social capital, and risk-taking—are the ones most commonly identified that demonstrate the importance of social networks in achieving long-term transformations. There are of course many other individual and organizational outcomes of networks, but these three are the most often associated with change.

As noted earlier, these outcomes identified in the literature have been drawn from studies of change initiatives in contexts that are quite different from that of STEM reform. Thus, as we examined learning, social capital, and risk-taking within this specific context, it was important to see which components were most salient in STEM professional networks, and whether other outcomes emerged as similarly important for change. As we studied the four STEM reform communities described in section 1, we paid close attention to how learning, social capital, and mediating risk-taking served to facilitate change in this particular environment. We also took into account the broader literature on outcomes4 , and we drew on this literature as we studied those outcomes that emerged as pivotal to change within the networks.

Design of Networks

While we knew much about outcomes of networks and benefits to members, we knew much less about how the design of these networks shaped these various outcomes. This study set out to draw connections between design principles and the achievement of the outcomes described above that foster lasting change. The limited data that exist on this topic come mostly from the social networks literature, and we now turn to a review of that material.

Various studies have identified how the design of social networks impacts outcomes such as achievement of goals and change. This literature, as well as the literature on design in communities of practice (described below) informed our data collection (section 3) and analyses related to design, described more fully in section 6. The most commonly identified design characteristics were: strong and weak ties, heterophily and homophily, subgroups, connectedness, and opinion leaders. We discuss each of these design attributes briefly below.


Strong ties are most useful for the communication of tacit, non-routine, and complex knowledge; in contrast, weak ties—present in networks said to be “less dense”—are better suited for communication of simple and routine information (Nelson, 1989; Tenkasi & Chesmore, 2003). Strong ties are characterized by three defining characteristics: frequent interaction, an extended history, and intimacy or mutual confiding between parties (Kraatz, 1998). Most studies of change find strong ties to be more conducive to deep or complex changes (Balkundi & Harrison, 2006; Tenkasi & Chesmore, 2003). Strong ties are more likely to promote in-depth, two-way communication and exchange of detailed information. Because pedagogical and curricular reform efforts require a deep and complex kind of change, we anticipated that it was likely that strong ties would be important for undergraduate STEM reform networks. While they can be designed to create frequent interaction, networks that are created for the purpose of innovation may be less likely to have extended histories or intimacy among members. Weak ties, on the other hand, are characterized by distance and infrequent relationships that may be casual, less intimate, and non-reciprocal in nature. However, for the dissemination of ideas and public information, weak ties can be extremely helpful. Weak links can also provide exposure to important external ideas that may promote the development of a more robust change idea. Thus, there may be times and circumstances where weak links are important for creating specific types of change, especially in certain phases of the change process. This study examined the degree to which networks can promote strong ties—those that have been found in the literature to be more useful for scaling up change. Also, the study examined when weak ties might be helpful in some aspects of network activity.


Another area of design found to shape outcomes is diversity or homogeneity of ties. Diversity of ties, or heterophily, can lead to more complex thinking about change processes, but homogeneous ties, or homophily, can lead to quicker adoption of change and to stronger relationship development, ultimately encouraging strong ties (Borgatti & Foster, 2003; Moody & White, 2003). Homophily might also lead to greater engagement and participation of network members. We can examine the degree of heterophily or homophily in the structure of social networks and draw conclusions on how those dimensions impact networks as they try to meet their goals.


The development of subgroups (cliques) within networks has also been identified as a strong lever for moving changes forward (Freeman, 1979; Reagans & McEvily, 2003). Within the networks in our study, we examined the types of subgroups that form to facilitate change, and the structural properties that might govern this process.


Another concept, connectedness, provides a measure of how much exposure each individual receives to the innovation (Borgatti & Foster, 2003; Valente, 1995). Individuals can be influenced to alter their behavior when they are surrounded by many people that have adopted a change, even if others throughout the campus or profession as a whole have not done so. This study tests the notion of connectedness by examining individuals who have more exposure to innovations than others within the networks. This approach yields insights on structural considerations for development of leadership and pervasiveness.


The presence of opinion leaders within social networks can help speed up adoption of innovations (Cross & Parker, 2004; Freeman, 1979; Valente, 1995). We examined the way key opinion leaders from STEM are brought into networks and helped to effect change.

These five design features are discussed in the literature, but their treatment is typically constrained to the more organic forms of social networks. While we examined each of these features in our study of STEM networks, we aimed to address a lack in the research in the question of how more constructed networks, like communities of practice, can be designed intentionally to achieve these outcomes. Thus, in addition to testing out some of the existing findings from social network theory about how design can shape outcomes, we explored new areas where little research exists, such as the type of leadership needed within networks, the way that formal organizations can support networks, and the way to create sustainable networks.

Our approach expands upon the organic focus of social network analysis, which limits what we know about the role that can be played by leadership, organizational structures, and intentional support; this is a major gap in the literature, noted by most social network scholars as a critical area for future studies (Mullen & Kochan, 2000; Spillane, Healey, & Kim, 2010). This gap is the reason there is so little literature about fostering and sustaining networks, a progression seen primarily as an organic series of events, rather than as a structured process. It is also important to note that social network analysis is primarily a survey-based, quantitative approach, which does not lend itself to the study of evolution of communities over time. Thus, we turned to the literature on communities of practice (CoPs) in order to partially remedy these limitations. CoPs have been studied through qualitative methods that better capture processes, and these studies have provided insight into the formation and sustainability of networks. There is also an existing body of research, albeit less well defined, about designing CoPs for engagement.

Defining Characteristics of Communities of Practice

A community of practice (CoP) is a group of people who share a concern or a passion for something they do and learn how to do it as they interact regularly (Allee, 2000; Lave, 1988; Wenger, 1998 and 2007). CoPs operate similarly to traditional social networks in that they connect people with a similar interest or value, and they similarly may cross organizational boundaries and be more loosely connected. CoPs are defined by three of characteristics: 1) a domain—a common ground of purpose and value; 2) a community—a set of individuals connected; and 3) a practice—ideas, frameworks, tools, or documents that the community members share. Without these three areas, an entity cannot be defined as a community of practice. CoPs take on many forms: they can be co-located or distributed (i.e., centralized in the same location or not); short-term or long term; homogenous or heterogeneous (focused on the same field or area or more diverse); small or big; housed within a unit or organization or spread across multiple such units; spontaneous or intentional; unrecognized or institutionalized. While their forms can differ, CoPs often emerge within organizations, and they often involve people working day-to-day with one another.

Because most research has focused on local CoPs within a single organization or unit, the principles identified in the literature on CoPs must be used with caution. Nonetheless, this research did provide direction for our methodological approaches, and it introduced some key insights—related to design, formation, and sustaining communities—that were valuable for shaping this study. In what follows, we review the defining characteristics of CoPs, as well as a
variant of CoPs known as professional learning communities (PLCs). These definitions helped us frame our understanding of the communities in our study—communities of transformation—which are examined in more detail in section 4. What are these Entities? Communities of Transformation.

In the literature, CoPs tend to be organic, developing naturally from a need shared among people. Most of the early work on CoPs examined the tacit learning that happened while people worked together through apprenticeship practices. Lave (1988) and Wenger (1998), who originated the concept, philosophically disagree about whether CoPs can be intentionally created or structured and whether they are truly organic entities. Even Wenger, McDermott, & Snyder (2002), who went on to examine “non-organic” CoPs, feel that CoPs are defined by a more organic state, and that they can only be nurtured, not created. The community or social aspect of learning is central in this phenomenon; thus, interaction and the relationships that are developed as part of the community are seen as essential to the CoPs existence.

The key activity of a CoP is to develop the domain that is at the center of the community. This becomes the identity of the community, and it serves as a focus for developing the shared repertoire of resources for the practice. The community operates through learning by problem-solving, sharing information, seeking expertise, visiting others, and using other similar approaches (Wenger et al., 2002). The literature describes challenges that emerge over the lifecycle or stages of a CoP. Given the organic nature of a CoP, it tends to go through a natural cycle (potential, coalescing, maturing, stewardship, transformation) in response to challenges that result from growth. In the course of this cycle, a CoP refines its identity and membership, and it incorporates new members and new purposes (Wenger et al., 2002).

Professional learning communities (PLCs) are a particular type of structured and organizationally located CoP that is commonly found in the education sector. PLCs are built on the principle that community is central to learning. PLCs can be considered a type of CoP because each PLC involves a community, has a domain, and involves a set of practices (e.g., the socially defined practices that enable one to become a strong teacher or principal). PLCs are distinctive, however, because they have several facets that are not part of the definition of CoPs. For example, PLCs are always intentionally created and tend to be heavily structured. This is less typical of CoPs. Additionally, the leadership of a PLC usually defines the membership, and a PLC typically includes people based on their roles, rather than on their organic interest in the domain (Stoll et al., 2006). While PLCs entail the exchange of information, expertise, and problem-solving, the character of these interactions in a PLC are often less peer-to-peer than in typical CoPs. While some institutions have set up peer-oriented PLCs with teachers only, these communities are always created and sanctioned by the institution’s administrative leadership (Bond & Lockee, 2014; Stoll et al., 2006). In addition, the nature of the work of PLCs tends to be more narrowly defined around a set of issues, such as student success or teaching broadly understood (DuFour, 2004; DuFour, DuFour, & Eaker, 2008). The model for such cases is more structured and hierarchical than that of CoPs (see, for example Roberts, 1998, in which PLCs are led by school principals). Domain, membership, and community operate differently in PLCs than they do in the other CoPs typically described in the literature. We now turn to the literature on design of CoPs.

Designing Communities of Practice

As with the limited literature on design of networks, the literature on design of CoPs informed our data collection and analyses, especially on the challenge of designing communities for engagement and the achievement of outcomes (see section 6). One of the major findings in the literature on CoPs is that design varies greatly based on the identified goals. There are many different types of CoPs, so there is no single design that guarantees efficacy. While the social network literature emphasizes basic designs and structures that support certain outcomes, CoPs tend to exhibit a less direct connection between design and outcomes. There is not a single best design, but various design principles that can enable a community to meet its specific goals. Several such general principles and practices were identified in Wenger, McDermott, & Snyder (2002) as important for creating learning that leads to change:

  1. Design the community to evolve naturally. Because a CoP is dynamic by nature, in that its interests, goals, and membership are subject to change, it should be designed to support shifts in focus.
  2. Create opportunities for open dialog not only among members, but also with those bringing in outside perspectives. While members and their knowledge are the most valuable resource of a CoP, it is also beneficial to look outside the community to understand other possibilities for achieving learning goals.
  3. Welcome and allow different levels of participation. Wenger identifies three main levels of participation in a CoP. First, there is the core group of members who participate intensely in the community through discussions and projects. This group typically takes on leadership roles in guiding the community. Second, there is the active group of members who attend and participate regularly, but not at the same level as the leaders. Third, there is the peripheral group of members who, while they are passive participants in the community, still learn from their involvement. Wenger notes that the third group typically includes the majority of the community.
  4. Develop both public and private community spaces. While CoPs typically operate in public spaces where all members share, discuss, and explore ideas, a CoP should also offer opportunity for private exchanges. A CoP designed in this way can coordinate relationships among members and access to resources through an individualized approach that is based on specific needs.
  5. Focus on the value of the community. A CoP should create opportunities for participants to explicitly discuss the value and productivity of their participation in the group.
  6. Combine familiarity and excitement. A CoP should offer the expected learning experiences as part of its structure, but there should also be opportunities for members to shape their learning experience together by brainstorming and by examining both the conventional and the radical wisdoms related to their topic.
  7. Find and nurture a regular rhythm for the community. A CoP should coordinate a thriving cycle of activities and events that allows for the members to regularly meet, reflect, and evolve. The rhythm, or pace, should maintain an anticipated level of engagement to sustain the vibrancy of the community, yet not be so rapid that it becomes unwieldy and overwhelming in its intensity.

Lifecycle of Communities of Practice: Formation and Sustaining

The literature on CoPs has developed a framework that examines formation, design, and sustaining of CoPs and is an expansion of the principles above, connecting them to how networks evolve over time. These various concepts were used to inform this study. The best-known framework for the lifecycle of CoPs was offered by Wenger et al. (2002), who created a five-stage community development model based on empirical studies of CoPs. We present this model here as a precursor to our description of the lifecycles of these communities in sections 7, 9, and 10.

The lifecycle model includes the following stages: 1. potential; 2. coalescing; 3. maturing; 4. stewardship; and 5. transformation (Wenger et al., 2002). Wenger and colleagues also outlined specific challenges or tensions for each stage, represented here in Table 2.1. These challenges indicate areas that might impact the viability or growth of CoPs. We briefly review the elements of the model that guide our exploration into higher education CoPs.


Stage Challenge/Tension
Potential: Community starts as a loose network of connections with potential for growing and developing more connections Discover or Imagine: Build on what is present, or explore where potential could lead
Coalescing: More connections are built, coalescing into a community Incubate or Deliver Immediate Value: Allow connections to form and build trust slowly, or immediately try to show the value of the community
Maturing: Membership and depth of knowledge in the community grows Focus or Expand: Direct energy toward internal interests of core members, or expand to meet interests of new members
Stewardship: Actively share and develop knowledge formed through community Ownership or Openness: Balance ownership over community domain with the need to bring in new ideas

The first phase, potential, is the phase where an important topic attracts an informal group of people who are interested in beginning to work together. Wenger et al. (2002) note that at some point the “idea of forming a community is introduced into [a] loose network, and this prospect starts to redirect people’s attention. They start to see their own issues and interests as communal fodder and the relationships in a new light of a potential community” (p. 71). As the sense of this shared domain develops, more systemic planning and activities begin. So the beginning work at the potential stage is to define the scope of the domain that brings people together, to find people who see the value in
increased networking and sharing of ideas, and to identify what common knowledge is needed to further the community. Through this stage the emerging community creates a vision and sense of mission.


During the second stage, coalescing, people come together and launch the community, and they find value in engaging in learning activities together. At this stage leaders in the CoP facilitate dialogue, create informal meetings, develop initial community support and communications, and develop organizational supports for the long run. Within this phase, the focus is on creating enough interest that people continue to participate. Part of creating this interest in continued involvement is achieved by establishing the value of the domain. The community needs to develop trust and strong relationships to get through philosophical challenges and other issues that emerge. The community also needs to develop key avenues for sharing information and creating information-rich resources.


In stage three, maturing, the community begins to take charge of activities, and it grows in size. At this time, the community is involved in many joint activities together. Active learning is taking place, and the growing community develops standards for how its members interact over the long run. In the maturing stage, the community needs to clarify and focus its roles and boundaries. As the community grows, new ideas are brought in that might expand or change the domain of its focus. New members can disrupt the patterns of interaction that the core members of the community have developed. The community needs to find ways to stay focused on its core purpose and mission while it includes greater numbers of individuals. A key issue related to practice focuses on organizing resources and knowledge for the long haul; the community needs to systematize its practices and create a rhythm of activities that community members can count on. Also, the community identifies gaps in knowledge, especially as the community grows. One challenge is to keep creating additional resources to meet the needs of new members. The tension between focus and expansion is palpable in this phase.


In stage four, stewardship, the community is well established and needs to find ways to sustain energy, to renew interest, and to continue to gain new members. At this point community leaders address organizational issues that may hinder their ongoing development, and they often forge linkages with other groups. In this stewardship phase, the community strives to sustain its momentum as continued new members join, as energies decline over time among longtime leaders, and as the original ideas of the community can fade in urgency and become less intellectually interesting. Stewardship is a balance between creating ongoing ways to bring in new ideas and remaining focused. Communities in this phase work to bring in new energy and new people, while supporting long-time leaders. Wenger et. al. (2002) describe the maturing and stewardship phases under the same broad label of maturing, and they see these two phases as hard to separate distinctly.


The last stage is transformation. Wenger et al. (2002) note that a tension exists in this phase, between the community’s sense of ownership and its openness to new ideas—an openness that is never fully resolved and often results in crisis. As the community widens its boundaries, it risks diluting its focus. If the community stays closed, on the other hand, it can suffocate itself. It is a natural feature of the lifecycle of a CoP that these events should occur, and sometimes the influx of new members in the transformation stage creates a new focus for the community; this leads it to transform. Other times the community may cease to exist, because members no longer feel that its purpose is relevant or needed.

Localized and Distributed Communities of Practice

Most of the literature reviewed to this point applies to more localized communities of practice. The STEM groups in this study, however, fall into what is called a “distributed community of practice” that cannot rely on regular face-to-face meetings and interactions as the primary vehicle for connecting members. Many communities of practice entail daily interactions among members, but distributed communities of practice have particular challenges that need to be considered in the design, formation and sustaining aspects: the distance that separates their members, the size of the community, and the need to work together across cultural differences (Wenger, McDermott, & Snyder, 2002). Because of the distance they cover, distributed communities must connect members more intentionally, and their design must think through challenges such as different time zones and the lack of spontaneous interactions among members. Second, distributed communities are often much larger in size than local communities of practice, with sometimes hundreds or thousands of members. Because people are unlikely to know each other well, or to have much face-to-face interaction, the community must wrestle with the question of the right size, and it must recognize when becoming too large impacts its viability. Thus, structures need to be created to adapt to growth. Third, distributed communities often run into issues related to the different cultures represented by their members. When people from across the country and world collaborate, they may not understand each other’s languages, customs, or styles of interaction, and this can create communication barriers that eventually lead to problems within the community. This study acknowledges and takes into consideration these factors connected to distributed CoPs, which have not as yet been studied in depth.


This study was informed by the literature on networks and communities of practice. The network literature informed our original thinking about outcomes of community involvement and design, while the CoP literature informed our thinking about how these communities form and evolve, how they are sustained, and how they can and should be designed to maximize engagement. While both of these literature bases (social networks and communities of practice) were informative, neither was a direct fit for the undergraduate STEM groups that were the focus of this study. These STEM reform groups are not organic, as is typically assumed in social network analysis, nor are they as tightly developed and structured as communities of practice. Instead, most STEM reform groups are semi-structured and fluid, and they best fit the model of a distributed community of practice—a model that has not yet been the object of much research. Therefore, a study of these unique STEM groups was needed to identify the outcomes, design, formation, and sustainability issues that pertain to their work. We drew on the earlier research described above for concepts, theories, and framing, but, because we knew that the STEM groups in our study did not match the literature, we were open to new concepts and principles that emerged as we strove to understand their formation, design, and sustainability. We now turn in section 3 to the methodology we employed for this study.

4. In terms of individual outcomes, the following have been identified: social support, sense of belonging, information sharing, community, more meaningful participation, enjoyment of work, confidence, help with challenges, expansion of skills, enhanced professional reputation, increased employability, and stronger sense of professional identity (Rogers, 2003; Valente, 1995). In terms of organizational outcomes, networks have been shown to help execute strategic plan, increase retention of talent, increase capacity for knowledge, allow for more alliances with external groups, foresee technological developments, improve quality of decisions, improve problem-solving, increase coordination across units, provide additional resources for implementing strategy, and strengthen quality assurance (Wenger, McDermott, & Snyder, 2002).