The next two chapters in the book, Social Network Theory, edited by Alan Daly, are examples of social network analysis studies in elementary schools. Chapter 3 examines the dynamics of and changes in tie formation in a districtwide mathematics reform. Chapter 4 examines network characteristics before and after school-based change efforts aimed at implementing the Literacy Collaborative model of instruction. Going back to the typology described in chapter 2, the first study is an example of Type 2 and 4 because the researchers examine at dyad and network levels and make conclusions about the consequences for ties as a result of the change efforts. The second, on the other hand, is both Type 4 and 6, comparing engagement in coaching (an indicator of implementation of the LC model) with network density (network level variable), coach centrality and isolation (node level variable), and school climate. Both drew on data from larger, longitudinal studies.
One brief observation as to my general understanding of social networks is that these methods of analysis gives us a different way to understand our reality and experiences. It’s not about searching for the “perfect” school-based social network, because these are dynamic systems.
Network Formation in the Context of a District-Based Mathematics Reform, by Coburn, Choi, and Mata.
Question: Coburn, Choi, and Mata attempt to characterize the dynamics of tie formation and change among elementary school teachers. Methods: Four elementary schools were selected based on recommendations from district directors for their variable professional communities and teachers’ expertise. In each school, six focal teachers (one from each grade) were selected. Interviews and classroom observations of focal teachers were conducted plus interviews with mathematics coaches, school principals, and six “nonfocal” teachers and observations of teacher interaction on matters of mathematics instruction (ex. PD, grade-level meetings). As an egocentric approach, a subset of interview questions was gathered from nonfocal teachers specifically about who teachers talked to regarding mathematics instruction, the frequency and content, and why those people. Finally, they also conducted interviews with district leaders, observed professional development sessions, and collected relevant district documents. Analysis: The focal teachers’ networks were mapped each year. A hybrid coding approach was used for the interviews to draw out expected themes (ex. proximity, homophily, expertise) but allowed for relevant themes to emerge. To examine the networks more holistically, size and diversity of ties were also calculated and reported. Finally, “expertise” was defined and compared for perceived expertise, tie formation, and actual expertise. Findings were reported by year and then compared. This was intentional to delineate the phases of implementation of the change effort. Conclusions: 1) Existing organizational norms, structures, and practices affect tie formation. For example, the school-level professional development corresponded with new tie formation around reform activities and an increase in forming ties based on expertise. 2) Networks became larger, more diverse, and more expert. Most notably, the accuracy of teachers locating expertise in a network increased, but even in year three, teacher accuracy in locating expertise was only 52%. 3) Changes to tie formation continued even after the district withdrew support due to policy and leadership change. This was evident in a continued increase in tie formation based on expertise, but the teachers also showed a return to forming the majority of ties based on proximity.
My comments mirror quite a bit of what the researchers bring out in the discussion. First, the importance of social networks in organization seems obvious and yet education has been slow to examine this (Little, Forward, xi), but these results support the importance of “organizational embeddedness to social network research.” I think this means that the organization has such a high influence on social networks that it is an appropriate level of network analysis. The results of this study, measured changes in tie formation, and the understanding that “relationships and collegial support are central for the retention, increased professionalism, and depth of engagement of educators” (Daly, p.1), provide strong rationale for me to pursue research in this area, or at least to have some component of social network analysis in any study. Another intriguing result around expertise in the network is the “degree to which teachers know where expertise is located in their environment” (p.47). With only limited information about what teachers do in their classrooms, it makes sense that teachers might not actually know who would be the best person to ask for advice, thus relying on those closest (proximity) or most like themselves (homophily). Social network analysis of this type might thus be used to make leaders aware of the need to create norms and structures that focus on developing strength-based connections. Knowledge of others’ expertise is related to network learning, which can also be called network capacity or transactive memory, for collective problem solving, coordination, and group performance” (p. 48). For me, this begins to tap into how we might find proximate measures of emergent characteristics, such as network learning, capacity for change, or internal/external credibility.
Centrality, Connection, and Commitment: The Role of Social Networks in a School-Based Literacy Initiative, by Atteberry and Byrk,
Question: How do network characteristics interact with change efforts aimed at instructional improvement? This change effort was already observed to have had variable engagement by different school sites. The independent variable was degree of adoption of the change effort, but measured engagement in coaching as a proxy. The dependent variables measured were network density, coach centrality, isolation (structural holes), and school climate. Methods: Data collection happened of seventeen schools over four years. Baseline data was collected during the year of nonimplementation while coaches were trained. The grade levels involved were kindergarten through third-grade. Teachers completed surveys during this first year and in the final year. Coaches maintained online logs include what they did and with whom. Literacy data was also collected from student standardized tests, but not used in this analysis. (Interestingly, compared to the first study, the data collection here is much simpler, just surveys and online logs.) Analysis: Network measure analysis included graphic visualizations of school (before and after) and summary variables, such as density, coach centrality, and network fragmentation. Findings: Some schools showed an increase in network density over the course of the study, which is theorized to be an indicator of growing critical dialogue around literacy teaching. Five schools actually showed a decrease in density. About half the schools showed an increased number of staff forming ties around literacy with the coach. Additionally, there was initial variability in the “coach-to-be”s rank in terms of their network centrality to literacy teaching and learning. Among the fifteen schools with similar faculty sizes, the more central the coach-to-be was prior to the intervention, the “more coaching she conducted per teacher over the following three years.” This correlation “suggests that the coach’s position in her network may be a supportive condition for increased coaching” (p.62). Conclusions: Evidence here, as with the first study, shows that the school social network was reshaped during the intervention. Furthermore, “in all schools, the coach became the most central character in [the literacy based] networks by the end of the study” (p.62). The variation in change of the networks, though, supports that pre-existing conditions and the intervention themselves reshape social networks differently.
My comments plus discussion: This variation in how social networks are reshaped further supports the need for social network analysis as a key part in understanding educational change. Both studies point to the central network role of coaches, which seems like an interesting path for further investigate in terms of social networks. Finally, the idea that the “health” of a network might be an indicator for successful school reform goes along with my growing interest, as mentioned above, about network level emergent factors such as credibility. The authors here do not define network health, and yet I think many might have an intuitive sense of what it might look like.
The result from the first study regarding teachers’ ability to locate expertise, however, might suggest that individuals have a poor understanding of the whole network picture. An individual in one network group or isolated from the network might perceive the network as being relatively healthy while an individual in another place in the network might perceive the opposite. This may suggest that outside perspectives are needed for characterizing networks (i.e. individuals cannot accurately describe their own network.) In order to create network level, emergent characteristics, though, studies might need to define something like “coherence” across the network relative to a certain dimension, or maybe a degree of coherence that indicates how uniform it is? For me to define “internal credibility” could thus involve coherence between perceptions those two individuals, which might be measured by ability to detect expertise, or it could involve their accuracy of understanding the structure of the overall network, with some sort of mechanism related to transparency.
Meta-reflection:
Writing these summaries was HUGELY helpful in understanding the studies, particularly the methods. After reading the studies, I thought I understood them, but when I had to pick out and summarize the methods, I realized that I wasn’t actually clear on how they’d gotten their data. As I try to learn how to write a literature review, particularly how to critique it, I think writing out this kind of summary will be instrumental. I struggled to define what types of studies these are even though when I wrote about the rubric in chapter 2, I thought it would be obvious. And, to be honest, I kind of wished I had a teacher who would just give me the answer so I could know for sure. Such a rational, scientist-minded way of looking at the world. Trusting my own conclusion is not nearly as convincing. I also think writing these summaries helped me to see how their conclusions might apply to my future research, rather than just reading the conclusions and treating them as interesting. I didn’t see the connection that I made in the final paragraph until I started writing the paragraph itself and further developed the idea upon revising the entire post.
More to come!
Julie Kallio, PhD 