Example of haphazard sampling in Auditing
|
 Show
Our study investigated whether increasing sample size was an effective technique to reduce selection bias. We asked 142 undergraduate accounting seniors to use haphazard sampling in selecting from two simulated populations—one of items in inventory bins, another of vouchers. We analyzed the composition of the samples to determine whether certain items were overrepresented and whether the extent of overrepresentation declined as sample size increased. Consistent with previous research, we found size and location influenced haphazard selections. Increasing sample size did not produce any material change in the samples’ composition. This finding showed that more was not necessarily better—at least for haphazard sampling—and bias seemed to prevail. How deep do auditors’ biases run? Although we found bias in samples selected physically, we suspected it also might have played a part in choosing items from a list (for example, picking accounts receivable from a trial balance), an important avenue for future research. Given our findings, auditors should reevaluate their heavy reliance on haphazard sampling. For critical applications, they should use random selection techniques, which skirt the problem of sampling bias. Standard-setting bodies, such as the ASB and IFAC, should reconsider conditions under which they sanction haphazard sampling as a reliable audit tool. For the full text of the research paper, see “The Effectiveness of Sample Size to Mitigate the Influence of Population Characteristics in Haphazard Sampling,” Auditing: A Journal of Practice & Theory, March 2001, vol. 20, no. 1. THOMAS W. HALL, CPA, PhD, is professor and chairman, Department of Accounting, University of Texas at Arlington. His e-mail address is . TERRI L. HERRON, CPA, PhD, is a certified information systems auditor and assistant professor, Department of Accounting and Finance, University of Montana, Missoula. BETHANE JO PIERCE, CPA, PhD, is associate professor and TERRY J. WITT, CPA, is professor emeritus, Department of Accounting, University of Texas at Arlington. This series is based on work published in Auditing: A Journal of Practice & Theory. The intent is to bridge the gap between researchers and practitioners by offering concise practice summaries of cutting-edge research in the field of auditing. This is when the auditor attempts to randomly select transactions from the population to test. Although the auditor thinks the approach is completely random, this approach is still somewhat skewed as a human being can’t select anything at random. Back To All Questions You might also be interested in... What type of sampling method is used for controls testing?Attribute sampling will be used primarily for testing internal controls. Attribute sampling will test for seeking specific characteristics. The other type of sampling is variable sampling, which is used for substantive testing. What is sampling risk?Sampling risk is considered the risk that the sample will not represent the actual population. When conducting tests based on samples, the assumption is that the sample is indicative of the entire population. The auditor should apply professional judgment in assessing sampling risk. In performing substantive tests of details, the auditor is concerned with two... What is risk of incorrect acceptance in audit sampling?Risk of incorrect acceptance is the risk that the sample allows the auditor to conclude that financial statements are not materially misstated, when in fact they actually are. This will affect the effectiveness of the audit detecting existing material misstatements. Paper F8, Audit and Assurance and Paper FAU, Foundations in Audit require students to gain an understanding of audit sampling. While you won’t be expected to pick a sample, you must have an understanding of how the various sampling methods work. This article will consider the various sampling methods in the context of Paper F8 and Paper FAU. This subject is dealt with in ISA 530, Audit Sampling. The definition of audit sampling is: ‘The application of audit procedures to less than 100% of items within a population of audit relevance such that all sampling units have a chance of selection in order to provide the auditor with a reasonable basis on which to draw conclusions about the entire population.’ (1) In other words, the standard recognises that auditors will not ordinarily test all the information available to them because this would be impractical as well as uneconomical. Instead, the auditor will use sampling as an audit technique in order to form their conclusions. It is important at the outset to understand that some procedures that the auditor may adopt do not involve audit sampling, 100% testing of items within a population, for example. Auditors may deem 100% testing appropriate where there are a small number of high value items that make up a population, or when there is a significant risk of material misstatement and other audit procedures will not provide sufficient appropriate audit evidence. However, candidates must appreciate that 100% examination is highly unlikely in the case of tests of controls; such sampling is more common for tests of detail (ie substantive testing). The use of sampling is widely adopted in auditing because it offers the opportunity for the auditor to obtain the minimum amount of audit evidence, which is both sufficient and appropriate, in order to form valid conclusions on the population. Audit sampling is also widely known to reduce the risk of ‘over-auditing’ in certain areas, and enables a much more efficient review of the working papers at the review stage of the audit. In devising their samples, auditors must ensure that the sample selected is representative of the population. If the sample is not representative of the population, the auditor will be unable to form a conclusion on the entire population. For example, if the auditor tests only 20% of trade receivables for existence at the reporting date by confirming after-date cash, this is hardly representative of the population, whereas, say, 75% would be much more representative.
|
