Efficiency and related issues in higher education: A data envelopment analysis approach

The purpose of this thesis is to examine Data Envelopment Analysis (DEA) as a potential tool for use in managerial accounting in comparison with other approaches for measuring the efficiency of not-for-profit entities. This study will compare DEA to other techniques in measuring the efficiency of Institutions of Higher Learning (IHLs) in the U.S.; The recent studies by the Select Committee on Higher Education in Texas (SCHE) provided an opportunity to examine different ways of evaluating such performances in a context of practical managerial decision making and policy formulation. By comparing DEA results with those secured from the ratio analysis of the SCHE, DEA was found to be superior not only in raising questions but also supplying decision aids with accompanying supporting information.; When DEA was applied to doctoral-granting universities in the U.S., the results were compared with the scholarly literature that is now available on regression studies of educational activities. For purposes of our study IHLs were separated into those with and without medical schools. This distinction proved very important in uncovering substantial differences in behavior between the two groups. Contrary to hypotheses suggested in these regression studies, public universities proved more efficient than private universities "on the average." Using concepts from DEA a distinction was drawn between "managerial" and "program" efficiency. Here it turned out that when medical schools are not present private universities' program is more efficient than public counterpart. In this way new distinctions were made in this study which were not heretofore present in the literature on IHL efficiency.; Another portion of our study dealt with returns-to-scale possibilities which differed markedly (even on the average) between IHLs with and without medical schools. Moreover, analyses by DEA showed marked ranges of variation for returns to scale possibilities for individual IHLs within each group that was concealed by the statistical averaging utilized in these econometric studies.; Recommendations for further research include uses of data for more than a single year and more elaborate tests of the different kinds of models that incorporate the methods and concepts of DEA.