Data Envelopment Analysis Advantages and Problems Demonstrated in a University Comparison Study

Introduction

Everyone must make decisions, and there are numerous techniques to assist when the scenario is complicated. However, one of the biggest challenges for decision-makers is when the data is complex and when you need to know more than which is the best alternative. Only a few techniques are available to rank alternatives and identify improvements across numerous variables and cases. Most techniques allow only a few data types, they accept a limited number of independent or dependent variables, and the routines will either identify something good or bad, rank the metrics, or predict the best value when a model has been developed - but rarely can one decision-making technique perform all these tasks.

Let's consider an example of a complex problem. How would a large multinational corporation leader decide which divisions or locations to terminate (or create) given the high volume of measurement data available? Furthermore, how could that same leader identify what needed to be improved in the lower-performing divisions so that perhaps no one needed to be hired or terminated? Likewise, given the high volume of marketing and benchmarking information available, how do students know which universities are the best choice? Alternatively, how do regulators determine which institutions under their supervision can be deemed safe for the public, or in contrast, which companies need to be improved and how? These are complicated managerial research questions requiring more than statistical correlation or predictive regression to solve!

That is the purpose of this chapter: to illustrate how a complex managerial decision-making approach can be applied to a complex situation. Additionally, the objective is to illustrate the advantages and disadvantages of applying a single best-in-class technique to solve a managerial problem. This chapter will review the literature to identify several critical factors relevant to managerial decision-making technique selection and implementation. The literature review will explain how several relevant techniques function and provide applied examples cited from the literature. The remainder of the chapter will apply the methodology in a study of universities, following the general approach conducted by a U.S.-based higher education association while using secondary data collected by the U.S. Department of Education. The results will be explained, and the implications will be provided to the stakeholders.

Specifically, the chapter aims to explain the data envelopment analysis (DEA) technique, the underlying theories, the advantages, and the disadvantages in a factual higher education comparative university study. There are several variations of how to design and apply DEA. These variations are discussed, and advice is provided about which design to choose based on the managerial research question and the available data types. There are three managerial research questions (RQ) driving this study. The first RQ is how the universities in the sample compare using benchmarks developed by the U.S. Department of Education (e.g., tuition, tenured faculty, etc.). The second RQ is how the lower-performing universities could improve their resource utilization or production compared to the best-in-class institutions within their peer group. A third RQ is also addressed regarding research quality: What are the statistical problems of using the DEA technique, which can be proven through detailed observations of the sample study, and how can the limitations be overcome in future studies? However, one of the challenges with DEA is that it is difficult to explain and understand for managers or nonmathematicians. In this chapter, the author explains the basics of DEA with examples in spreadsheet graphs to give the reader a theoretical understanding. Additionally, the author provides applied examples of DEA, with the estimates interpreted along with implications generalized to stakeholders.