Data envelopment analysis and multiple criteria decision making

Stewart has recently presented a survey of theory and practice in multiple criteria decision making (MCDM). While Stewart admits that this is ‘by no means a comprehensive review’, the absence of data envelopment analysis (DEA) from the approaches described is, the authors believe, an important omission. In this brief note they illustrate DEA as an aid to MCDM by reanalysing some data which Stewart used in the article to demonstrate multivariate statistical approaches. One distinction that Stewart draws is between a priori and progressive (interactive) articulation of preference by the decision maker. DEA is flexible since it may be done with preferences articulated a priori, progressively or not at all.