Co-Plot: A Useful Tool to Detect Outliers in DEA

Abstract

The main goal of this research is to compare and contrast two methodologies that exist within the multi-criteria decision making field. After presenting Data Envelopment Analysis (DEA) and the Multi-Dimensional Scaling approach (Co-plot), the two techniques are evaluated and compared against data on Program Follow Through (PFT) reported by Charnes et al. (1981). It would seem that cross-fertilization of these two methodologies sheds additional light on the definition of DEA efficiency and grouping in Co-plot, it may also be useful to apply both approaches in the future. Comparison shows that Co-plot may be useful prior to a DEA in order to identify outliers that require additional scrutiny. Furthermore, there is a clear connection between DEA efficient units and the results of a Co-plot analysis using variable ratios; consequently Co-plot could be useful as a graphical tool helping to interpret the results of a DEA.

Keywords: Data Envelopment Analysis, Multi-Dimensional Scaling, Co-plot, outliers, efficiency

JEL Classification: C02, C61