Individual, Institutional and Systemic Factors Contributing to Academic Research Production: Comparing Traditional and Multilevel Models

Abstract Academic research production is a complex process shaped by factors operating at individual, institutional, and systemic levels. This study examines the relative influence of nine individual-level variables (academic rank, having completed a postdoc, time dedicated to research, interest in research, national collaborations, international collaborations, institutional funding, national funding, and gender), two institutional-level variables (university type and linguistic status), and one systemic-level variable (province). Using survey data from 2,364 professors across 47 universities in five Canadian provinces, we compared multiple linear regression with multilevel modelling to assess research production. Both approaches explain similar proportions of variance in academics' publication index (R² = .35 for multiple regression; R² = .33 for the three-level model). Across models, the strongest predictors of research production are the time dedicated to research and whether academics report collaborating nationally or internationally. Recognising that faculty operate within nested contexts, we highlight the added value of multilevel modelling in disentangling individual and institutional influences, capturing clustering effects, and mitigating aggregation bias. These findings contribute to a deeper understanding of research production across individual, institutional, and systemic levels, while underscoring the importance of employing diverse analytical approaches to both measure research output and assess the effects of influencing factors.