Evaluation of university libraries using a multi-stage non-parametric model

Abstract

The present work aims to evaluate the efficiency of a set of university libraries belonging to a Brazilian federal higher education institution using Data Envelopment Analysis (DEA). To prevent the internal stages of resource processing in services from being neglected, we chose to use a multi-stage method known as Network DEA (NDEA). Thus, to obtain efficiency indices in the two internal stages and consequently the verification of the overall efficiency index, seven variables were used: three exogenous inputs, one intermediate variable, and three exogenous outputs. The results showed that the libraries located in inland cities obtained greater efficiencies in the first stage because they are smaller and offer their services in the number of hours near the larger libraries. It was also possible to identify the best performing libraries in the second stage and those that performed well in both evaluation stages. For better visualization of the results, we developed a bi-dimensional representation of the results. It was noticed that in future works the evaluation can be expanded, being analyzed libraries belonging to different institutions and different countries, in addition to using a model that allows the identification of benchmarks for inefficient units.