This paper suggests a new DEA-based approach to the identification of the efficient category value for controllable categorical factors when controllable categorical factors are involved in the DMU's production process. Though the main objective of this approach is to identify the efficient category value for nonhierarchical categorical factors, it can also be employed to identify the efficient category value for hierarchical categorical factors without any large adjustment. This new approach is applied to an empirical data from an Agricultural Resource Management Survey for soybean production in Pennsylvania. The result of the empirical application illustrates how our approach leads us to the efficient category value of categorical factors by using both radial measure and nonradial measure. Moreover, it shows us the existence of substantial heterogeneity in the use of categorical factors and the opportunity of improvement in DMUs' selection of category value for categorical factors.
Read full abstract