This study evaluates the transformational efficiency of the California State University (CSU) system by examining how financial and human resources are converted into academic and long-term socioeconomic outcomes. It addresses challenges in performance assessment within multi-campus public university systems characterized by institutional heterogeneity and resource constraints.
An output-oriented Data Envelopment Analysis (DEA) model with assurance-region constraints is applied to twenty-three CSU campuses. To enhance reliability in a small-sample setting, the analysis integrates bootstrap resampling, stability-radius diagnostics, and a semi-parametric Corrected Ordinary Least Squares (COLS) frontier for cross-method validation. A second-stage Tobit model examines institutional and socioeconomic determinants.
Efficiency varies substantially across campuses despite shared governance, indicating that performance is shaped by structural and contextual differences. Instructional spending is positively associated with efficiency, while a higher proportion of economically disadvantaged students is negatively related to measured performance. Scale effects contribute to disparities. Robustness diagnostics show that most efficiency classifications are stable, although some frontier positions are sensitive to perturbation.
The results support peer-based benchmarking, targeted resource prioritization, and equity-sensitive performance evaluation. DEA-derived targets provide actionable guidance for campus-level improvement, while robustness diagnostics encourage cautious interpretation of rankings in accountability-driven public systems.
The study advances higher education performance measurement by integrating deterministic and semi-parametric frontier methods with robustness diagnostics in a unified framework. It extends DEA applications by incorporating long-term outcomes and cross-method validation, and reframes efficiency as a context-sensitive construct shaped by equity, scale and institutional heterogeneity.
