Stratified Models of Education Production Using Modified DEA and Regression Analysis

In their agenda for research into the determinants of organizational effectiveness, Lewin and Minton (1986, p. 531) called for studies illustrating the “feasibility of using DEA (perhaps in combination with other analytical methods) as a mathematic for relating effectiveness outcomes to features of organization design.” In this chapter, we report the initial results of just such a study. The purpose of the study is to investigate the performance of secondary education in the U.S. Performance is defined as the ability of secondary schools to convert human, physical, and financial resources into educational opportunities, and then to combine these opportunities with student input to produce both intermediate and long-term educational outcomes. At each stage, performance is measured using a modified form of data envelopment analysis (MDEA). The three sets of calculated MDEA scores are then regressed against groups of environmental variables that reflect features of the organizational design and characteristics of the student body of each secondary school in the sample. The modification to DEA and the second- stage regression analysis are structured to take into account the special distributional features of unmodified DEA scores. The empirical results suggest that a two-stage approach that augments MDEA with regression analysis provides a fruitful way of determining effectiveness outcomes and relating them to features of organizational design and the operating environment. Furthermore, regression analysis provides useful information for policymakers interested in how changes in organizational characteristics can improve school performance.