Using social network analysis to better understand student support structures
and facilitate success in biomedical education and training
Social network analysis
- • The methods and theory first were studied quantitatively by social and behavioral scientists.
- • The canonical text is “Social Network Analysis: Methods and Applications” by Stanley Wasserman and Katherine Faust, Cambridge University Press, 1994.
- • The tradition extends to the 1920s.
- • Network science, as a discipline, now extends well beyond the social context, seeing contributions from applied mathematical and statistical sciences and applications to the natural sciences, social and organizational sciences, economics and political science, and information science.
Biomedical science trainees
- • Students arrive at biomedical training programs with a set of existing social ties and social connections that offer support and buffer them from stress.
- • We invite and implore students to expand their existing social networks to include fellow students, faculty mentors, fellow lab members and others.
- • The success of our students depends on their ability to navigate toward the larger social network of biomedical students and biomedical scientists.
Many students lack confidence or feel anxious when they are first thrown into big science — with the attendant needs to navigate complex research protocols, fit in and function with teams of seasoned students and postdocs, and locate and establish working relationships with new research mentors.
For students unacquainted with this traditional academic-training model, these tasks may seem burdensome on top of the expectations of earning good grades and completing demanding coursework.
The pressure to fit in can be especially challenging for first-generation and underrepresented minority students, who may have limited exposure to the social norms and expectations of a laboratory or other research environment. Additionally, first-generation and URM students may arrive at universities with different sets of familial and cultural experiences.
For more than a generation, formal, federally supported programs have been in place to identify and recruit underrepresented students to biomedical-research training programs. Students in these programs typically are offered support addressing their economic needs and educational preparation.
However, interventions designed to help these students have shown mixed success, and we argue that this is largely because efforts to date have failed to take a holistic view, including both the academic and social needs of students and trainees.
The standard model
Considerable research and intervention-related attention has focused on the dyadic mentor-protégé relationship. That body of work assumed students seek and obtain all necessary support from individuals who serve in defined mentoring roles.
Solid mentoring relationships certainly can facilitate training experiences in the biomedical sciences, offering students access to tangible, informational and emotional support. Mentored research experiences have been shown to enhance undergraduate students’ motivation, interest and readiness for careers in biomedical research (1, 2, 3, 4).
As scientists and educators, some of us may have leaned hard on our mentors for professional and perhaps also personal support. Yet not all students find suitable or supportive mentors.
Underrepresented and minority students in the sciences are at risk for inadequate mentoring relationships (5). Students can and do persist and succeed in the absence of strong mentoring relationships, but these students are at higher risk for attrition. Additionally, years of effort to improve mentoring in science have yielded only limited results with respect to retention and success of underrepresented students. Some scientist-educators simply may lack interest or skill in providing substantive emotional support, especially in a competitive funding environment where research must come first.
A growing body of evidence suggests that it is critical to look beyond dyadic mentoring relationships to support students effectively, especially URMs, in the biomedical sciences pipeline, and here we briefly present an innovative way for doing so.
Looking beyond mentors
Social network analysis (SNA), as a theory and data analytic method, allows for investigation of larger and more comprehensive social networks, going well beyond examination of dyadic relationships.
SNA emerged in the 1970s, and network theory, or the related study of the processes and mechanisms interacting with network structures that yield individual- or group-based outcomes, is developing rapidly as a field of study (6).
This cross-disciplinary research area is now of interest not only to social scientists but also to epidemiologists, biologists and physicists. Network theory, commonly associated with big data, has been applied to such disparate topics as the prediction of flu pandemics, the assessment of individuals’ voting behavior and understanding marriages between Florentine families in the 15th century.
The methodology also can be adapted to help us understand and intervene with networks of students in the biomedical sciences pipeline. By better understanding the structure and characteristics of students’ social networks and associating network structure with student characteristics (e.g., gender, ethnicity, grade point average, publications and presentations), it may be possible to help the students to support themselves.
In brief, a social network is defined as a set of actors or nodes (for example, a group of students) that are interconnected by a set of defined ties (for example, amount of lab time spent together or interpersonal support). Ties interconnect these actors to form paths, and within a network, the pattern of ties represents structure. These structures can be represented mathematically. Hence, the position of actors has meaning, as do the characteristics of the network structure.
Examples of network structure characteristics include actor centrality — how prominent an actor is within the network. Actors who are prominent are extensively involved in network relationships. SNA researchers have derived numerous measures of actor centrality. In contrast, actors may be isolates and remain relatively unconnected to other actors within the network. Typically this would be seen as a liability.
Using SNA to see ties among those in a science department
A network of students and faculty mentors working within a biochemistry department can be studied with regard to, say, interpersonal support.
The researcher selects students, perhaps first-year students, and faculty members who define the network to be studied. The researcher also identifies the type of tie (support) or relationship to be studied.
At the outset of the academic semester, students and faculty members have little familiarity with one another, and hence no one individual supports any other. The network, in effect, is unconnected with regard to support. However, as individuals begin spending time together and interacting, they begin exchanging support. In other words, at some second measurement point, all individuals in the network can respond to the question “Who here provides you with support?” By the end of the first semester of students’ first year, all individuals in the network (actors or nodes) can be connected, and the network can be examined for structural features.
Such a study would employ a full-network, or whole-network, research design that examines the supportive ties among first-year students and their faculty members. If the study is conducted longitudinally, the investigator can study how supportive relationships among students and faculty members evolve over time.
Alternatively, investigators might elect to study an egocentric, or personal, network, which can be considered a subset of the full network and either extracted from the whole or studied in the main. Egocentric networks can be created in a more streamlined manner — that is, simply by asking the students, “Among people in the biochemistry department, who are the three of four people who provide you with support?” In this case, the full network is not used; hence, the structural features of the full network cannot be analyzed thoroughly.
SNA in action
Some researchers who study students in the biomedical sciences pipeline are beginning to use network analysis. For example, Cecilia Rios-Aguilar of Claremont Graduate University is studying URM community college students’ success as it relates to their virtual communities, the supportive online relationships that students create and use to help one another achieve success.
Rios-Aguilar argues that successful students are more connected both academically and socially. In this sense, students’ virtual communities augment, or perhaps supplant, mentoring relationships.
In community-college settings, students have more limited face-to-face access to mentors and peers. By contrast, biomedical science students at four-year institutions frequently interact with mentors and peers in person, and those interactions have the potential to provide different kinds of support or assistance — tangible, informational and emotional. Those interactions, such as team-oriented lab work, in-class presentations and one-on-one coaching sessions, also make demands upon students’ interpersonal skills.
Teaming up with social scientists
Supportive social networks are critical to all students, especially those just entering competitive, research-oriented fields such as those in the biomedical sciences. For underrepresented minority students, a supportive social network may be a critical adjunctive factor to their academic success.
Social network analysis offers many opportunities to investigate and intervene with the aim of first understanding and then strategically strengthening students’ networks. It also allows investigators to determine whether the impact of the emergent social milieu exceeds the impact of the mentor-protégé relationship. If so, then support for students in the biomedical sciences pipeline should become more holistic and perhaps less focused on mentoring. Students in these programs still may need assistance with their economic or educational readiness, but additional focus could be placed on strengthening students’ abilities to support one another.
Application of methods derived from network theory will allow us to understand and then perhaps bolster students’ confidence, increase their retention and success in the laboratory, and thereby increase their interest in careers in research. Meaningful collaborations between biomedical scientists and educators and social scientists trained in SNA are needed to assess and refine our training paradigms in ways that promote the supportive social networks necessary for student success.
- 1. ↵ Lopatto, D. Survey of Undergraduate Research Experiences (SURE): first findings. Cell Biol. Educ. 3, 270 – 277 (2004).
- 2. ↵ Pfund, C.; Pribbenow, C.M.; Branchaw, J.; Lauffer, S.M.; & Handelsman, J. The merits of training mentors, Science 311, 473 – 474 (2004).
- 3. ↵ Russell, S.H.; Hancock, M.P.; & McCullough, J. Benefits of undergraduate research experiences. Science 316, 548 – 549 (2007).
- 4. ↵ Seymour, E.; Hunter, A.-B.; Laursen, S.L.; & DeAntoni, T. Establishing the benefits of research experiences for undergraduates in the sciences: first findings from a three-year study. Sci. Educ. 88, 493 – 534 (2004).
- 5. ↵ Chew, L.D.; Watanabe, J.M.; Buchwald, D.; & Lessler, D.S. Junior faculty’s perspectives on mentoring. Acad. Med. 78, 652 (2003).
- 6. ↵ Borgatti, S.P. & Halgin, D.S. On network theory. Organization Science 22, 1168 – 1181 (2011).
Suzanne E. Barbour (email@example.com) is professor of biochemistry and molecular biology. Victoria A. Shivy (firstname.lastname@example.org) is an associate professor of psychology.