Audit sampling is the use of an audit procedure on a selection of the items within an account balance or class of transactions. The sampling method used should yield an equal probability that each unit in the sample could be selected. The intent behind doing so is to evaluate some aspect of the information. Audit sampling is needed when population sizes are large, since examining the entire population would be highly inefficient. There are multiple ways to engage in audit sampling, including the following:
- Block sampling. A consecutive series of items are selected for review. Though this approach may be efficient, there is a risk that a block of items will not reflect the characteristics of the entire population.
- Haphazard sampling. There is no structured approach to how items are selected. However, the person doing the selections will probably skew the selections (even if inadvertently), so the selections are not truly random.
- Personal judgment. The auditor uses her own judgment to select items, perhaps favoring items that have larger monetary values or which appear to have a higher level of risk associated with them.
- Random sampling. A random number generator is used to make selections. This approach is the most theoretically correct, but can require more time to make selections.
- Stratified sampling. The auditor splits the population into different sections (such as high value and low value) and then selects from each section.
- Systematic sampling. Selections are taken from the population at fixed intervals, such as every 20th item. This tends to be a relatively efficient sampling technique.