Exploring Determinants of University Employees’ Work Engagement: Insights from Random Forest and SHAP

This study examines the determinants of university employees’ work engagement using explainable machine learning techniques. Data were obtained from the 2022 University Innovation Capacity Assessment, involving 3,649 respondents. Random forest and Shapley additive explanations (SHAP) algorithms were employed to assess feature importance and partial dependencies among factors. The findings reveal that, for full-time employees, community sense, organizational positivity, organizational pride, collaborative culture, organizational learning, and communication were the primary determinants of work engagement. For part-time employees, organizational pride, communication, collaborative culture, community sense, organizational innovativeness, and procedural justice were decisive factors. Additionally, the relationships between the factors and work engagement were predominantly positive. Based on the findings, we provide implications for policy and practice to enhance university employees’ work engagement, contributing to improved effectiveness and performance in higher education institutions.