Data envelopment analysis (DEA) was originally conceived as a tool for examining the relative efficiency of production units on the basis of ex post data on outputs produced and inputs consumed. Latterly, DEA-inspired approaches have assumed status within the toolkit of investigators concerned with multi-attribute decision making. In its original environment, DEA essentially classifies the production units into two groups: the relatively efficient and the relatively inefficient. However, in the multi-attribute decision making environment it is often desired to rank a set of alternatives or select a shortlist from the set for more detailed scrutiny. With an emphasis on ranking or selection, discrimination in DEA, or its lack, becomes an issue.
