Towards standards for statistical sampling;

              3
Toward Standards for Statistical Sampling
Kenneth W. Stringer
Haskins & Sells
I am always glad to have an opportunity to discuss statistical sampling, which is one of my favorite subjects. I am particularly pleased to be able to do so at the invitation of Howard Stettler because my interest in the subject was first stimulated by reading his article in The Journal of Accountancy in January 1954. At that point I became convinced that statistical sampling is the most rational means for determining the extent of audit tests of details of transactions and account balances. Extensive study and experience in implementation of statistical sampling in our Firm's audit practice in the intervening years has strengthened that conviction.
Although the use of statistical sampling in the profession has not progressed as rapidly as I have considered desirable, I think it is fair to say that interest in the subject is increasing currently. This observation is based on discussions with interested parties in various firms concerning the extent of their current studies and/or applications. The reasons why progress in the meantime has been more evolutionary than revolutionary are understandable, and have involved both statistical and auditing problems. The statistical problems have included the general unfamiliarity of auditors with statistical methods, and technical questions concerning the applicability of certain statistical methods to auditing situations. The auditing problems have related primarily to defining and expressing audit objectives in terms susceptible to statistical measurement, and to the difficulty of combining statistical and subjective evaluations of audit evidence in forming overall conclusions.
Because the auditing problems are equally or more difficult and also are more appropriate for my assigned topic, I will confine my discussion today to them. For this purpose I will review first the evolution of present AICPA literature concerning statistical sampling, and second the current consideration being given to the expansion of that literature.
Present Literature
The first official AICPA literature on statistical sampling was a special report of its Committee on Statistical Sampling, which was published in The Journal of Accountancy in February 1962. Although this report was quite general in its coverage, it was the result of extensive deliberations by the Com­mittee
and established two landmark positions. First, it stated that:
43