The Moment Bound with Unrestricted Random, Cell and Sieve Sampling of Monetary Units

Abstract

Abstract Since its introduction by Dworkin and Grimlund (1984), the moment bound has emerged as a less conservative method of evaluating monetary-unit samples than existing methods for estimating the total error amount in substantive auditing. Arthur Andersen has recently adopted it in place of the Stringer bound (Felix et al., 1990). In this paper we examine its performance with unrestricted random, cell and sieve sampling; the behaviour of the sampling methods is compared by means of an empirical investigation using two new accounting populations. Our results show that the method of sample selection has no effect on the tightness of the moment bound, but does appear to affect the reliability and precision. While the moment bound is reliable in most cases, some reliability problems arise and these are most frequent with unrestricted random sampling. We also observe substantial gains in the precision of the moment bound with cell and sieve sampling in populations with large line items. The reduction in sa...