Discussant's response to use of decision theory in auditing -- A practitioner's view;

              Discussant's Response to
Use of Decision Theory in Auditing—
A Practitioner's View
William L. Felix, Jr.
University of Washington
Jim Loebbecke is to be complimented on his willingness to provide a prac­titioner's
view of a topic area that is, I am sure, very difficult to deal with.
Researchers interested in using decision theory as a descriptive model of auditing practice need feedback and criticism from practitioners. Otherwise, it is very likely that inaccurate models and inferences from these models will result. I would particularly like to observe the importance of contributions such as the auditor's logic process that Jim provided in his 1974 comment on my paper. These views of the basic audit process, including (1) steps for the collection of particular kinds of evidence and (2) steps that represent major decision points in the audit process, are the kind of basic information and feedback that is necessary to construct useful models.
Significance of Research on Decision Theory
The significance of the research efforts in decision theory may not be entirely clear. The line of argument that is most appealing to me is that a decision theory model requires rather precise specification of relationships and decisions that may have previously been left rather vague and imprecise. The result of greater precision ought to be better understanding of the consistencies and differences in the various opinion formulation processes being used in the profession. In addition, the resulting well structured and understood models would improve the ability of researchers to investigate problems and by implication, for the profession to adapt to changes.
Possible Misconceptions
On the first page of his paper, Jim indicates that approximately 30% of the practitioners he surveyed have some idea as to what decision theory implies with regard to auditing. Being realistic, I suspect that this is a fairly optimistic figure. There are probably only a handful of practitioners in this country that can approach Jim's knowledge of this particular application of statistical inference. Yet I think that his paper indicates that he has some misconceptions about the nature of decision theory, either as a normative model of the auditor's decision processes or as a descriptive model of how auditors behave. Let me illustrate my point with a couple of observations from Jim's paper. In the last paragraph on page 110 of his paper under "point 1" he states that the decision to extend work is
118