What is and is not audit sampling, and sampling versus nonsampling risk

In this lesson you will

Apply the AU-C 530.05 definition of audit sampling.
Distinguish sampling from 100% examination and selecting specific items.
Explain the two types of sampling risk and nonsampling risk.

The definition

AU-C 530.05 defines audit sampling as the selection and evaluation of less than 100% of the population of audit relevance such that you expect the items selected to be representative of the population and likely to provide a reasonable basis for conclusions about it. Representative means that evaluating the sample gives conclusions similar, subject to sampling risk, to those you would reach testing the whole population.

The key is intent: you plan to conclude on the whole population from the sample.

Three ways to select items

AU-C 330.A66 lists the means of selecting items for testing:

Selecting all items (100% examination). AU-C 330.A68 says it may be appropriate when the population is a small number of large items, when a significant risk exists and other means are not sufficient, or when an automated, repetitive calculation makes it cost effective. Not sampling.
Selecting specific items. High-value or key items, unusual or risky items, or all items over an amount (AU-C 330.A69). Not sampling. AU-C 330.A70 says results cannot be projected to the population, and selecting specific items does not by itself give sufficient evidence about the remainder.
Audit sampling. Testing a representative sample to conclude on the population.

Common practice combines them: test all items over a threshold, then sample the remainder.

AU-C 530.A3 adds another exclusion: examining a few transactions to understand operations or clarify your understanding of internal control, as in a walkthrough, is not audit sampling.

Statistical and nonstatistical

AU-C 530.05 defines statistical sampling as having both random selection and the use of an appropriate statistical technique to evaluate results, including measurement of sampling risk. Anything lacking either is nonstatistical. AU-C 530.A14 makes the choice a matter of judgment and states that sample size is not a valid criterion for choosing.

Sampling risk

AU-C 530.05 defines sampling risk as the risk that your conclusion based on a sample differs from the conclusion if the whole population were tested. Two types:

Type	Test of controls	Test of details	Effect
Effectiveness risk	Controls more effective than they are (risk of overreliance)	No material misstatement when one exists (risk of incorrect acceptance)	Wrong opinion. The auditor is primarily concerned with this type.
Efficiency risk	Controls less effective than they are	Material misstatement when none exists	Extra work to correct the initial conclusion

You control sampling risk through sample size and selection. AU-C 530.A12: the lower the risk you accept, the larger the sample.

Nonsampling risk

AU-C 530.05 defines nonsampling risk as the risk of an erroneous conclusion for any reason not related to sampling risk. AU-C 530.A4 gives examples: inappropriate procedures, misinterpreting evidence, and failing to recognize a misstatement or deviation. Testing every item does not remove it. It is reduced through planning, supervision and the firm's quality management.

Selecting specific items is subject to nonsampling risk only (AU-C 330.A69), which is why judgmental key-item selection has no sampling risk to measure but still needs a well-designed procedure.

Common findings

Key items tested and the results "projected" to the population.
Remainder populations above tolerable misstatement left untested.
Walkthrough selections described as samples.
Samples drawn from listings not tested for completeness (AU-C 530.A8 refers to AU-C 500.08).

The audit sampling methods guide compares approaches, and the next lesson begins with attribute sampling.

Worked example

Splitting a population

Ridgeline Medical Equipment, Inc. reports fixed asset additions of $2,960,000 across 214 items. Performance materiality is $70,000.

Step 1, specific items: 9 additions over $70,000 total $1,820,000. All 9 are vouched. This is not sampling, and no projection is made.

Step 2, remainder: 205 items totaling $2,960,000 - $1,820,000 = $1,140,000. The remainder exceeds tolerable misstatement of $70,000, so testing only the key items would not provide sufficient evidence about it (AU-C 330.A70). The team applies audit sampling to the remainder.

Step 3, completeness of the population: the team agrees the additions listing to the fixed asset rollforward and the general ledger, both $2,960,000.

Step 4, nonsampling risk: the program defines a misstatement as a cost not supported by an invoice or an item that should have been expensed under the capitalization policy. The senior briefs staff on the policy before testing.

Results: key items had one misstatement of $12,000 (factual). The sample is evaluated separately and projected only to the $1,140,000 remainder.

Self-check

Answer each question before you open the answer.

Is testing all items over $50,000 audit sampling?
Show answer
No. It is selecting specific items. Results cannot be projected to the rest of the population (AU-C 330.A70).
What two characteristics make a sample statistical under AU-C 530.05?
Show answer
Random selection of sample items and use of an appropriate statistical technique to evaluate results, including measuring sampling risk.
Which type of sampling risk affects audit effectiveness?
Show answer
Concluding controls are more effective than they are, or that no material misstatement exists when it does. It can lead to an inappropriate opinion, so the auditor is primarily concerned with it.
Does testing 100% of a population eliminate nonsampling risk?
Show answer
No. Nonsampling risk comes from using inappropriate procedures or misinterpreting evidence, which can happen on any item tested.

Sources

Back to the syllabus

Standards are paraphrased from the primary sources listed above; read the source before relying on a requirement. This lesson is training material, not professional advice. Report an error