Graduate student networks in a convergent research environment for wicked problems
M. Gail Jones, Julianna Nieuwsma, Kathleen Bordewieck, Gina Childers, Steve McDonald, Kimberly Bourne, Maude Cuchiara, Anna-Maria Marshall, Brooke K. Mayer, Christine Ogilvie Hendren, John Classen
Research in Science Education
Network design
Nodes
graduate students in a 10-university, 28-discipline convergent research center and a comparison group in related programs
Ties
undirected roster-based reports that one student knew another at baseline and after one year
Methods
The study compared density, eigenvector centrality, modularity and assortativity by status, affiliation and institution across two waves and groups, alongside survey measures of expectations, impostorism, collaboration and experience.
Source: Research in Science Education
Reviewed summary
Twenty-five graduate students in a convergent research center and 28 comparison students across the same universities and programs reported who they knew before and after one year. The center network became denser, while modularity and assortativity by status, affiliation, and institution declined. The findings show how a deliberately cross-disciplinary environment can be audited for whether formal boundaries remain visible in actual graduate-student relationships.
The center network density rose from 0.081 to 0.165 and mean eigenvector centrality from 0.029 to 0.043; affiliation modularity fell from 0.278 to 0.044 and institutional assortativity from 0.510 to 0.070. The treatment and comparison groups were small and not randomized, baseline composition differed, the tie only measured knowing someone rather than collaboration quality, and many survey outcomes were exploratory. One-year structural change cannot be assigned solely to the center.
SNA
How SNA was used
The study defined graduate students in a 10-university, 28-discipline convergent research center and a comparison group in related programs as nodes and undirected roster-based reports that one student knew another at baseline and after one year as ties. The study compared density, eigenvector centrality, modularity and assortativity by status, affiliation and institution across two waves and groups, alongside survey measures of expectations, impostorism, collaboration and experience.
Source
Research in Science Education, 55(6), 1757-1772
DOI: 10.1007/s11165-025-10249-x
