Audit sampling involves selecting and evaluating less than 100% of the items in an account balance or class of transactions (PCAOB, n.d.). There are two general approaches to audit sampling: statistical and non-statistical. Statistical sampling is an objective method, as it involves a random selection of the sample size and the items selected for examination. The sample is then evaluated using mathematical methods to draw conclusions that are projected to the population as a whole. Further, this approach to sampling is most appropriate when examining a large group of homogeneous transactions. Such transactions may take place in the ordinary course of business and should have minimal impact on the financial statements. Contrarily, non-statistical sampling, is more subjective--the sample size and the items selected are determined based off the auditor's judgment (Gleim & Hillison, 2021). This approach is best applied to a small number or group of transactions, often times those determined to be significant (material) or of higher risk. Thus, these transactions may carry a greater risk of misstatement and have the potential to cause the financial statements to be materially misstated. Also, though the non-statistical approach may be more cost-effective and less time-consuming, it does not offer the ability to quantitatively assess sampling risk or the confidence level of how representative the sample is of the population. Being subject to auditor judgment, there may also be room for bias.

Overall, each approach has its advantages and disadvantages; the desired outcome, auditor training, and the control environment are just a couple of factors to consider when deciding which method to apply. Nevertheless, when applied appropriately, both approaches can be used to effectively and efficiently gather sufficient, appropriate evidence to form conclusions about an account balance or class of transactions.