122). The term "decision making unit" was coined by Charnes and his associates to describe a set of homogeneous units, each of which uses a common set of inputs to produce a common set of outputs as shown in Figure 2 below. Figure 2. Graphical illustration of a set of DMUs each using a common set of inputs {1, 2,..., I} to produce a common set of outputs {1, 2,..., R}.

Source: Marcoulides, p. 123.

Some examples of decision-making units include the collection of similar firms, departments, or other organizational units and typical DMUs can be comprised of group of schools, hospitals, bank branches and like units (Marcoulides). Based on the mathematical formula developed for this model, business managers can choose those weights that maximize the efficiency of each DMU, subject to the constraint that no other DMU using the same set of weights can achieve an efficiency rating of higher than 1; a rating of 1 indicates that the DMU is efficient compared to the rest of the DMUs in the group (Marcoulides). According to this author, "As a result, management can use DEA not only to identify relatively inefficient units and measure the magnitude of their inefficiencies, but also to provide guidelines for improvement" (Marcoulides, p. 125).

Many business managers also use decision trees to help them in the decision-making process. This technique provides a graphical representation of the decision-making process and are designed to clearly establish the alternatives available to a decision maker, and to provide some indication of the various outcomes that could result (Vernon). According to this author, "Decision trees are in the form of probability diagrams. Each branch of the tree represents an outcome that follows from the premise to which it is attached. For example, two common outcomes are success or failure, and a probability can be assigned to these to provide a quantitative measure of the likelihood of each" (Vernon, p. 45). Conversely, decision trees can also be used to identify associated costs, with each branch representing an amount that that choice represents; using this approach allows managers to take a large decision and break it down into more manageable pieces, summarizing the impact of various options as the decision tree progresses. In this regard, Vernon notes that, "If, for example, the overall success of a decision depends upon the success of several smaller steps, then an overall probability of success can be calculated" (p. 56).

A sample decision tree used to evaluate the economic impact of hiring a consultant for a given application is show in Figure 1 below.

Figure 1. Sample decision tree.

Source: Merrick School of Business, University of Baltimore, 2008, http://home.ubalt.edu/ntsbarsh/opre640a/tree.gif.

Even business research methods that employ decision trees have their advantages and disadvantages, though. While these graphic representations can be highly useful in certain settings (e.g., the decision under consideration is comparable to decisions that have been taken in the past, then the quantitative aspect of their analysis is likely to be accurate, since probabilities can be derived from that experience), for the same reason, decision trees are subject to a much wider range of variables if they involve strategic rather than routine decisions (Vernon). According to this author, "Strategic decisions are by definition steps into the unknown. Decision trees are also subject to prejudices of the decision maker who might, for example, make...

They are also prone to conceal qualitative aspects of decision making, especially human factors, which can be as important to success" (emphasis added) (Vernon, p. 56).
Data-Gathering Method and Sources of Data.

For the literature review, a variety of sources will be consulted including university and public libraries, online resources such as EBSCO and Questia, as well as online organizational resources from companies to identify any salient business research practices. For the survey, a convenience sample of business managers concerning their current business research methods will be used. For this purpose, a custom survey consisting of a series of Likert-scaled questions ranging from "strongly agree" to "strongly disagree" was created (see copy at Appendix a) and will be posted on SurveyMonkey.com, a free online survey service, with invitations emailed to business managers in various industries identified online or personally known to the author with a goal of collecting at least 50 responses. These initial questions may be replaced with others or modified based on the findings of the final literature review and face validity reports from colleagues, co-workers, family members, and others. Statistical analysis of the aggregated survey data will be accomplished using SPSS Version 11.0 (Student Version), with the results reported in tabular and graphic form, as well as being interpolated with the findings of the literature review in the concluding chapter.