Abstract
Data envelopment analysis (DEA) is a data-oriented mathematical programming approach that evaluates a set of peer decision making units (DMUs) dealing directly with the observed inputs and outputs (performance measures). Empirically, in order to have a logical assessment, there should be a balance between the number of performance measures and the number of DMUs. Accordingly, applying an appropriate method so that one can select some performance measures is very crucial for successful applications. In this paper, we suggest the envelopment form of selecting model under constant returns to scale (CRS) from both individual and aggregate points of view. We also show that applying these selecting models leads to the maximum discrimination between efficient units.