Exploring team composition using administrative data: evidence from a large UK university

Bibliometric studies examining the relationship between research team size and structure and research outputs typically define teams through co-authorship. In this paper-in-progress, we explore how research teams are structured using universities’ internal administrative data. By integrating datasets on grants, paid staff, contracts, and studentships from one large UK research-intensive university, we reconstruct research teams active between 2010-2023. Data processing and assumptions were iteratively refined through semi-structured interviews with staff across diverse roles and departments. We model team size as a function of discipline, principal investigator (PI) grade as a proxy for seniority, and year using a multilevel Bayesian negative binomial model. Results show that average team size varies across disciplines, tends to increase with seniority, but has overall remained static over the years. This study establishes the methodological basis for future research on research team size, structure, and composition on what we refer to as the “meso” scale.