Structure and Evolution of Scientific Collaboration Networks in a Modern Research Collaboratory

Abstract

This dissertation is a study of scientific collaboration at the Center for Embedded Networked Sensing (CENS), a modern, multi-disciplinary, distributed laboratory involved in sensor network research. By use of survey research and network analysis, this dissertation examines the collaborative ecology of CENS in terms of three networks of interaction: co-authorship of scholarly publications, communication activity on mailing lists, and interpersonal acquaintanceship. This study exposes the topology, structure, and evolution of these networks in relation with the disciplinary and institutional arrangements of CENS. Findings indicate that CENS collaboration networks have fluid, non-cliquish, small-world topologies, and are free of prestige-based mechanisms. Further analysis reveals that structural communities in the co-authorship and acquaintanceship networks overlap considerably. They also exhibit little disciplinary and institutional diversity locally, although CENS becomes more inter-disciplinary over time. Overall, results of the structural and evolutionary analyses point to the importance of interpersonal relationships for accomplishing scientific work in distributed environments.

Keywords: scientific collaboration, collaboratories, cyberinfrastructure, social networks, scientific networks, social complex systems, community structure, assortative mixing, homophily