Performance Evaluation of University-Based Scientific Research Institutions With a Non-Independent Parallel System.

This article focuses on improving performance evaluation models for university-based scientific research institutions by addressing the interdependence between teaching and research subsystems and accounting for disparities in university development. It introduces two innovative two-stage data envelopment analysis (DEA) models—the group-frontier and meta-frontier evaluation models—and validates them using data from 144 Chinese universities across different regions and development levels. The findings indicate that these models more accurately assess institutional efficiency than traditional methods, which often underestimate performance by ignoring subsystem interactions and heterogeneous development. The study offers practical insights for university administrators and government education departments to optimize resource allocation and enhance both teaching and research outputs.