Discussant's response to toward standards for statistical sampling;

              Discussant's Response to
Toward Standards for Statistical Sampling
James W. Kelley
University of Georgia
As was expected, Ken Stringer expertly reviewed the brief history of the move toward establishing standards for the use of statistical sampling in auditing. He has also analyzed the situation and referred to specific proposed standards.
His analysis and the recommendations have far reaching implications and are deserving of our most serious attention and discussion. The purpose of this paper is to direct further attention to some of what I hope will prove to be the more serious implications of establishing standards.
Are Standards Needed?
Several years ago over a hundred of us attended a Professional Development Workshop on the use of computers. We spent two full days working on controls over commission checks sent to some independent salesmen. Near the end of the program one of the most inexperienced men in attendance said with absolute innocence, "I don't see why we can't have the salesmen deduct his commission as he sends in the order. Then we wouldn't have to send these checks." You know, he was right. Over 1500 man hours were spent working on a problem that really didn't exist.
This is not to say we have no problem, but we do hear a lot of conversation on what standards are needed and very little on why they are needed. We also hear some recommendations about what the standards should say but almost nothing can be found about what they will mean.
Let us remember first, that all statistical theory can do for the auditor is state in mathematical terms the risk he is taking by not examining all items. Statistical sampling can not tell us what tests are to be performed or what evidence need be collected to satisfy the test. It deals only with the sample size.
It should be understood that the courts have never ruled on the use of statistical sampling in auditing. A lawyer friend of mine recently searched the reported cases involving accountants and found no references to statistics. There was a by-product to this research, however, that surprised me. He found no cases where the auditor's sample size was questioned. I understand the attorneys of a national accounting firm conducted the same type of study with the same results. This suggests that, to date, the courts have not questioned auditors' judg­ments
on sample sizes. While we may properly decide that standards are neces­sary,
there appears to be no overpowering legal reason to do so just now.
It is argued by some that we need to set standards before someone else does. This is a powerful argument, with both rational and, in the light of recent de-
50