Accounting technology / Glossary / Audit sampling

What is audit sampling?

Definition

Audit sampling is applying audit procedures to fewer than all the items in a population, chosen so that every item has a chance of selection, in order to reach a conclusion about the whole population. Auditors use it where testing every item is impractical, such as checking a sample of sales invoices to supporting delivery records.

Also calledstatistical samplingsample testing

Compare the softwareBest audit software →All 8 audit vendors →

Standards: ISA 530; PCAOB AS 2315
Methods: Statistical and non-statistical
Category: Audit

How audit sampling works

The auditor defines the population and what counts as an error, then sets the sample size from the level of assurance needed, the error the auditor can tolerate and the error expected. Statistical sampling selects items at random or by monetary unit, so larger balances are more likely to be picked, and measures sampling risk mathematically. Non-statistical sampling uses judgement to set sizes and select items but must still be free of bias. Errors found in the sample are projected to the population and compared with tolerable misstatement. Selecting only the largest items, or all items over a threshold, is not sampling, because the rest have no chance of selection. International Standard on Auditing 530 sets the rules under ISAs; PCAOB AS 2315 covers US public company audits. Data analytics now tests some populations in full, leaving sampling for evidence that has to be inspected item by item.

Common questions

What is monetary unit sampling?+

A statistical method that treats each unit of currency in a population as a sampling unit, so items with larger values are more likely to be selected. It suits testing for overstatement in balances such as receivables.

What is sampling risk?+

The risk that the auditor's conclusion from a sample differs from the conclusion that would be reached by testing the whole population. Larger samples reduce it.

How large should an audit sample be?+

It depends on the assurance needed from the test, the tolerable misstatement, the expected misstatement and the size and variability of the population. Firms set sample sizes in their methodology, often through tables or software.