How to Apply Social Network Analysis to Evaluate Professional Development Programs in Biomedical Research

Effective evaluation and assessment of research-focused programs, including centers, core facilities, and institutes, is crucial but inherently challenging. Faculty members are often departmentalized, yet their research necessitates campus-wide access to infrastructure, specialized equipment, students, and interdisciplinary collaborators, thus creating assessment and reporting challenges. Additionally, it can be challenging to demonstrate return on investment for new and emerging programs as well as those at institutions without an extensive history of research. Social network analysis (SNA) tools may be used to address these challenges. SNA provides a data-driven understanding of collaboration dynamics in programs designed to build research community and provides a method to quantify program impact and demonstrate the effectiveness of capacity-building investments in fostering collaborative research and scientific output. Additionally, SNA is useful in pinpointing influential researchers who act as central nodes or bridges, and may guide future efforts in mentorship, grant development, and recruitment, by highlighting areas of productive collaboration, or alternatively, potential fragmentation in the network. The results may be used to guide future resource allocation, inform new initiatives within the program, or alternatively, target interventions if needed to strengthen a specific research area. Initial analysis can help a program assess collaborative health of a research community by providing a baseline for tracking changes in network structure and collaboration patterns over time, setting the stage for longitudinal evaluation. This article outlines the steps required to carry out SNA of bibliometric data from a biomedical research community. The article describes the acquisition of data through PubMed, application of VOSviewer and Gephi, data visualization, statistical analysis, and generation of figures suitable for publication, including examples of results designed for specific audiences and stakeholders. The article also covers creating an author thesaurus for disambiguation of authors who publish under multiple names.