Academic journal article Auditing: A Journal of Practice & Theory

The Effectiveness of Increasing Sample Size to Mitigate the Influence of Population Characteristics in Haphazard Sampling

Academic journal article Auditing: A Journal of Practice & Theory

The Effectiveness of Increasing Sample Size to Mitigate the Influence of Population Characteristics in Haphazard Sampling

Article excerpt

A recent survey of auditors found that the majority of audit samples are selected using haphazard sampling, i.e., selecting items using professional judgment. Yet, over 40 years ago experts in audit sampling expressed concerns that samples selected in this manner may be unrepresentative due to the presence of unintended selection biases. This concern was reinforced when a recent study found that haphazard sample selections of physical items were significantly influenced by their size and location. To mitigate selection bias when using haphazard selection, some experts in the field recommend increasing sample size by up to 100 percent.

This study investigated whether increasing sample size is an effective technique to mitigate selection bias in audit samples. To examine this issue 142 undergraduate accounting seniors selected haphazard samples from two simulated populations: a population of inventory bins and a population of vouchers. The compositions of these samples were then analyzed to determine if certain population items were overrepresented, and if the extent of overrepresentation declined as sample size increased. Consistent with a previously published study, analyses disclosed that haphazard selections were influenced by the size and location of population items. Thus, a presumption that haphazard selection procedures can be expected to yield representative samples as suggested by standard-setting bodies may not be warranted. The data analysis also disclosed that increasing sample size produced no material change in the composition of the samples. This finding indicates that increasing sample size is not an effective technique for mitigating selection bias in audit samples. Although this study found bias in samples selected directly from a population, there are reasons to believe that selection bias may also result when sampling from a listing of a population (e.g., selecting receivables from a trial balance), an important avenue for future research.

Given these findings, audit practitioners should reevaluate the heavy reliance on haphazard sampling. For critical applications random selection techniques should be used since these techniques are not affected by the features of the population. Standard-setting bodies such as the Auditing Standards Board and IFAC may wish to further examine the ability of haphazard selection procedures to produce representative samples. Pending completion of more research, these standard-setting bodies should reconsider the conditions under which haphazard sampling is sanctioned as a reliable audit tool.


Haphazard sampling is a process of sample selection in which the auditor chooses a sample that is expected to be representative of the population without the use of probabilistic randomizing aids (AICPA 1997; IFAC 1998; AICPA 1999).(1) Ideally, haphazard selection is indistinguishable from probability sampling for which each item has the same chance of selection. Guidance provided by the American Institute of Certified Public Accountants (AICPA) and the International Federation of Accountants (IFAC) suggests that when haphazard sampling is used sample elements should be chosen without any specific reason for including or omitting them. In particular, auditors are instructed that sample elements should be chosen without regard to "size, shape, location, or other physical features" (AICPA 1995, [sections] 8220.05) and sample composition should not be distorted by "avoiding difficult to locate items, or always choosing or avoiding the first or last [items]" (IFAC 1998, [sections] 530). Thus, haphazard sampling does not mean that one is careless in the selection process, but rather due care must be exercised to ensure that sample selections are not improperly influenced by the characteristics of population elements (Price Waterhouse 1989; AICPA 1995; Guy et al. 1998; IFAC 1998; AICPA 1999).(2)

Published data suggest that samples obtained through nonstatistical means predominate in the auditing profession (Hall and Walther 1981; Ross et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.