Audit Sampling

Audit sampling is the application of an audit procedure to less than 100 percent of the items in a population, in a way that provides a reasonable basis for drawing a conclusion about the population as a whole. It exists because testing entire populations is rarely proportionate to the assurance gained.

Sampling approaches divide into statistical and non-statistical. Statistical sampling uses random selection and probability theory to quantify sampling risk, allowing the auditor to project results to the population with a measurable confidence level. Non-statistical, or judgmental, sampling relies on the auditor's judgment to select items and size the sample; it is permitted and widely used, but conclusions cannot be mathematically extrapolated.

Two purposes drive design. Tests of controls use attribute sampling to estimate a deviation rate, with sample size driven by the tolerable deviation rate, the expected deviation rate, and the desired confidence level — which is why control testing sample sizes cluster around conventional figures tied to control frequency: annual controls tested once, monthly controls perhaps two to five instances, daily controls twenty to forty. Tests of details use variables sampling - often monetary unit sampling, which weights selection by dollar value - to estimate misstatement in an account balance.

Design also involves stratification, separating high-value or high-risk items for individual examination, and identifying items to be tested 100 percent because they exceed a scope threshold or possess specific risk characteristics. When misstatements or deviations are found, the auditor evaluates both their nature and cause and projects the error to the population. A projected misstatement approaching materiality requires extending procedures or requesting an adjustment.