Defining the Objective of the Sample

Auditors resort to sampling because they have a question they want to answer. This question can take many forms and its precise nature helps to determine what sampling methodology would be most appropriate to obtain an answer.

A central question to consider when defining the objective of a sample and selecting an appropriate approach is whether the audit team will need to extrapolate the results of its analysis to the entire population to be able to conclude on the audit objective. If yes, this necessarily implies a generalizable sampling approach (either attribute or variables sampling).

For example, if auditors want to estimate the frequency of a certain characteristic in a population, such as non-compliance with a policy, then a generalizable sampling approach would be most appropriate because there is a need to extrapolate the result of the sample to the entire population. In this case, the best approach would be attribute sampling because of the binary nature of each sampling unit (each unit is either compliant or non-compliant). Alternatively, auditors may want to assess a program’s impact in a population using a continuous variable such as dollars or size. In that case, a variables sampling approach would be appropriate.

A related question is whether the audit team wants to provide a quantitative measure of the sample’s accuracy; that is, the confidence interval. If so, this also implies using a generalizable sampling approach. Similarly, if the audit team does not think it will be able to support its findings from the sample with other sources of information, then being able to provide a quantitative measure of sampling risks by using a generalizable sample would help the team to increase the credibility of its findings.

Another approach applies in cases where auditors seek to identify the causes and repercussions of specific problems without having to quantify, extrapolate, or measure their extent. In such cases, auditors can use their judgment to select sample units that exhibit characteristics of interest and there is no need to select the sample randomly or to extrapolate the analysis results to the entire population. A purposeful sampling approach is the most appropriate choice in these circumstances. A purposeful sample is also recommended if the auditors are aiming to demonstrate extreme variations or exemplary occurrences.

• PreviousStep 2 – Selecting a Sampling Approach
• NextOther Factors to Consider When Selecting a Sampling Approach