Using Audit Sampling to Gather Evidence Efficiently

Audit sampling lets the auditor test fewer than 100 percent of a population while still forming a conclusion about the population as a whole. It is useful because many account balances and transaction classes contain too many items for full testing. Sampling is also risky because the selected items may not represent the population.

The AUD exam usually tests sampling through design choices: define the population, select an appropriate method, understand what changes sample size, project errors correctly, and respond when results exceed the tolerable level.

    flowchart LR
	    A["Define audit objective"] --> B["Define population and sampling unit"]
	    B --> C["Set tolerable misstatement or deviation"]
	    C --> D["Select sample method and size"]
	    D --> E["Test selected items"]
	    E --> F["Project results and evaluate risk"]
	    F --> G["Conclude or expand procedures"]

What Sampling Risk Means

Sampling risk is the risk that the auditor’s conclusion from a sample differs from the conclusion that would have been reached by testing the entire population. Sampling risk cannot be eliminated unless the auditor tests the full population, but it can be reduced through better sample design and larger sample sizes.

Sampling risk differs from nonsampling risk.

Increasing sample size reduces sampling risk. It does not fix nonsampling risk caused by poor audit procedures or weak execution.

Statistical and Nonstatistical Sampling

Statistical sampling uses random selection and probability theory to measure sampling risk. Nonstatistical sampling uses auditor judgment to design and evaluate the sample but still requires objectivity, representativeness, and documentation.

Nonstatistical does not mean casual. The auditor still defines the objective, population, sampling unit, expected misstatement or deviation, tolerable level, sample size, and evaluation method.

What Changes Sample Size

Sample size is driven by risk, precision, and expected error. The direction matters for exam questions.

For tests of controls, the auditor often thinks in terms of deviation rates. For substantive testing, the auditor often thinks in terms of misstatement amounts.

Projecting Misstatements

When the auditor finds misstatements in a sample, the results must be evaluated against the population. A simple proportional projection is:

\[ \text{Projected misstatement} = \frac{\text{Sample misstatement}}{\text{Sample book value}} \times \text{Population book value} \]

The auditor also considers known misstatements, anomalous errors, qualitative factors, and allowance for sampling risk. If projected misstatement plus an allowance for sampling risk approaches or exceeds tolerable misstatement, the auditor cannot simply ignore the result.

Possible responses include:

• Ask management to investigate and correct identified misstatements.
• Expand the sample.
• Perform alternative procedures.
• Reassess control risk or inherent risk.
• Increase substantive testing in the affected population.
• Evaluate whether uncorrected misstatements are material individually or in aggregate.

Anomalous errors require caution. A misstatement may be treated as anomalous only when the auditor has a high degree of certainty that it is not representative of the population.

Sampling in Controls Versus Substantive Testing

Sampling is used differently depending on the objective.

In a dual-purpose test, the auditor must evaluate both objectives. A document may have proper approval but the amount may still be wrong, or the amount may be correct but the control may not have operated.

Exam Traps

Do not say sampling eliminates audit risk. It only helps control sampling risk.

Do not treat high-dollar targeted testing as a representative sample. Selecting all large items may be useful, but the remaining population still needs an appropriate response.

Do not confuse tolerable misstatement with materiality. Tolerable misstatement is set for a population and is usually below performance materiality or otherwise aligned with the auditor’s planned detection risk.

Do not assume nonstatistical sampling means undocumented judgment. The auditor still documents the rationale and evaluation.

Quick Review

• Sampling tests less than the full population and creates sampling risk.
• Statistical sampling quantifies sampling risk; nonstatistical sampling uses documented judgment.
• Higher risk, higher desired confidence, higher expected error, and higher variability generally increase sample size.
• Sample errors are projected to the population and evaluated against tolerable levels.
• Sampling conclusions must be tied to the audit objective: control deviation or substantive misstatement.

Audit Sampling Knowledge Quiz