Assessing the Performance of Nonprofit Organizations: A Regression-Based Absolute Weight-Restricted Benefit-of-the-Doubt Data Envelopment Analysis Model

Performance evaluation becomes more complex when a nonprofit organization’s program expense ratio (PER) is represented by alternative measures that capture the same underlying construct. There are differences in PER measurement procedures and therefore it is difficult to determine which single measure may provide the most appropriate basis for performance appraisal. Moreover, conventional aggregation methods may not adequately account for differences among these estimates. This study proposes a benefit-of-the-doubt (BoD) data envelopment analysis (DEA) framework to develop a composite measure of PER incorporating alternative estimates of PER. The proposed approach allows the relative importance assigned to each PER estimate to vary across years while imposing weight restrictions to prevent the DEA model from assigning excessively flexible or zero weights to individual PER estimates. The empirical application considers three alternative estimates of PER over 19 years. The proposed model ensures that all estimates contribute to the evaluation while retaining the BoD principle. The analysis further compares the resulting efficiency scores and rankings with those obtained from the BoD DEA model and examines the robustness of the results under alternative weight restrictions. The findings demonstrate how weight-restricted BoD DEA can provide a more balanced and transparent framework for aggregating alternative estimates of PER while allowing differences in their relative importance across years. The study contributes a structured approach to evaluating alternative measurements of the same performance construct, offering greater transparency in weight determination and a basis for more balanced performance assessment.