Measuring research-related academic performance in a comprehensive university: individual heterogeneity, institutional context, and multidimensional output profiles

Abstract Assessing research-related academic performance remains a central challenge in scientometrics because scholarly output is multidimensional, unevenly distributed across individuals, and embedded in disciplinary and institutional contexts. This study examines the relative contribution of between-individual heterogeneity and institutional context to research-related academic performance in a Romanian comprehensive public university, using a five-year panel dataset comprising 444 academic staff members and 2,220 person-year observations from 2017 to 2021. Performance was measured through two complementary approaches: a relative performance indicator expressing each individual’s annual research score in relation to the benchmark for their current academic position, and a multidimensional activity profile based on 20 research-output indicators. Generalized linear mixed models, exploratory generalized additive models (GAM), and multivariate constrained ordination were used. The most parsimonious mixed model retained year, age, academic title, and study field and yielded a marginal R 2 of 23% and a conditional R 2 of 78%. Constrained ordination showed a similar pattern, with respondent identity explaining 49.79% of adjusted variation in activity profiles. Over time, output profiles shifted toward Q1/Q2 publications and independent citations. In the GAM analysis, work seniority was not associated with relative performance, whereas time since last promotion showed a nonlinear relationship, with performance increasing during the first years after promotion and declining thereafter. Overall, persistent between-individual differences accounted for the largest share of structured variation, while institutional, field-specific, and career-stage factors contributed to differences in the level and composition of academic output.