Audit and Assurance · Audit procedures

Audit Sampling and Selecting Items for Testing (ISA 530)

Updated 11 October 2026 · Fact-checked

Audit sampling means applying audit procedures to fewer than 100% of items in a population so the auditor can conclude on the whole population. Under ISA 530 every item must have a chance of selection. You choose a method, size the sample, test it, project errors and evaluate the result.

Understand Audit Sampling and Selecting Items for Testing

An auditor cannot check every transaction. Time and cost make it impractical, and it is not needed to gain reasonable assurance. So the auditor tests a sample and draws a conclusion about the population, which is the full set of data the sample comes from.

ISA 530 says sampling must give all items in the population a chance of being selected. That is what lets you conclude on the whole population. If you hand-pick items, that is not sampling. It is selective testing, and you cannot project the result to the rest of the population.

Sampling can be statistical or non-statistical. A sample is statistical only if it combines random selection with probability theory to evaluate results and measure sampling risk. Random selection uses tools such as random number generators or random number tables. Systematic and monetary unit selection can be used in a statistical sample only if the start is random and the results are evaluated using probability theory, but they are not random selection themselves. Selection methods alone do not make a sample statistical, because non-statistical sampling can use them too. Non-statistical sampling relies on auditor judgement for size, selection and evaluation. Both are acceptable if they give sufficient appropriate evidence.

Haphazard selection is not statistical. It avoids conscious bias, but the auditor must still guard against bias or predictability. Block selection, where you take a run of consecutive items, is generally not appropriate. Items outside the block have no chance of selection, so it is not true sampling.

Sampling risk is the risk that the auditor's conclusion from the sample differs from the conclusion if the whole population were tested. Non-sampling risk is any other risk that leads to a wrong conclusion, such as using the wrong procedure, misreading evidence or failing to recognise an error. Increasing sample size reduces sampling risk. It does not cure non-sampling risk. Better planning, supervision and review do.

An anomaly is a misstatement clearly not representative of the population, such as a one-off error from a known cause. The auditor must be highly certain it is an anomaly before excluding it from projection. Otherwise, errors found are projected to the population and compared with tolerable misstatement (or tolerable rate of deviation for tests of controls).

Key rules to remember

Sampling interval (systematic selection): Sampling interval = Population size ÷ Sample size; Choose a random start within the first interval, then select every nth item. Beware hidden patterns in the population.
Monetary unit sampling interval: Sampling interval = Population value ÷ Sample size; Each ₹/$ unit is a sampling unit, so larger items are more likely to be selected. Good for overstatement testing.
Projected misstatement: Projected misstatement = (Error found ÷ Value of sample) × Population value; Used when errors are found in a sample. Compare the result with tolerable misstatement.
Projection by error rate: Projected errors = (Errors in sample ÷ Sample size) × Population size; Used for tests of controls and item counts.
Factors affecting sample size: Higher tolerable misstatement → smaller sample; higher assessed risk → larger sample; Sample size rises with risk, expected misstatement and population variability. It falls with higher tolerable misstatement. Population size has little effect on sample size unless the population is very small.

How to solve Audit Sampling and Selecting Items for Testing questions

Use this order for any sampling question, whether it asks you to explain, select, size or evaluate.

1Define the objective and the population. Say what assertion you are testing, for example existence of receivables, and what the population is.
2Decide whether sampling is suitable. If only a few key items matter, consider testing 100% or selecting specific items instead.
3Consider the factors that set sample size: tolerable misstatement, assessed risk, expected errors and population variability. Population size matters little unless the population is very small.
4Choose a selection method: random, systematic, monetary unit or haphazard. Match it to the population and justify it. Haphazard is not statistical, and you must avoid bias or predictability. Avoid block selection, because it is generally not appropriate when not every item has a chance of selection.
5Perform the procedure on each selected item. If an item cannot be tested, use alternative procedures. Do not just ignore it.
6Investigate errors. Decide if each is an anomaly or a real error, and find its cause and effect on other audit areas.
7Project errors to the population and compare with tolerable misstatement. Add any other known misstatements.
8Conclude. If projected error is near or above tolerable, extend testing, ask management to adjust, or consider the audit opinion.

Quickest way: Sampling question in four lines

When to use it: Use in Section A or B objective tests, or when you are short of time in Section C.

Check the wording: if every item must have a chance of selection, it is sampling. If items are chosen by size or risk, it is selective testing.
For selection, match the method: random suits any population where items can be numbered, systematic suits a population with no pattern matching the interval, monetary unit suits populations where larger values carry more risk, and haphazard is non-statistical and acceptable only if the auditor avoids bias. Block selection is generally not appropriate.
For calculations, divide population by sample size for the interval. For projection, scale sample error up to the population.
For risk, remember sampling risk falls with a bigger sample while non-sampling risk does not.

Common mistakes in Audit Sampling and Selecting Items for Testing

Calling any test of selected items 'sampling'.
Students forget that sampling needs every item to have a chance of selection.
Fix: If items are picked by value or risk, call it selective testing and say the result cannot be projected.
Saying a bigger sample removes all risk.
Students mix up sampling risk and non-sampling risk.
Fix: A bigger sample reduces only sampling risk. Non-sampling risk comes from poor procedures or judgement.
Ignoring errors as 'one-offs' without proof.
Students want to avoid projecting a large error.
Fix: Treat an error as an anomaly only if you are highly certain it is not representative. Otherwise project it.
Forgetting to project errors to the whole population.
Students compare the sample error directly with tolerable misstatement.
Fix: Scale up the error found, then compare with tolerable misstatement.
Saying statistical sampling is always better.
It sounds more scientific.
Fix: Both methods are valid under ISA 530. Statistical gives a measured sampling risk, but non-statistical is often cheaper and simpler.
Not testing an item that cannot be located, such as a missing invoice.
Students assume the item can simply be replaced.
Fix: Use alternative procedures. If none work, treat the item as a deviation (in tests of controls) or a misstatement (in substantive tests).

Worked examples

Example 1

A population of 4,000 sales invoices is to be tested using systematic selection with a sample of 80 invoices. Calculate the sampling interval. The auditor picks a random start of 17. State the first three invoice numbers selected, assuming invoices are numbered 1 to 4,000. Name one drawback of the method.

Show the solution

Sampling interval = 4,000 ÷ 80 = 50.
The first item is the random start: 17.
The second is 17 + 50 = 67.
The third is 67 + 50 = 117.
Drawback: if the population has a pattern that matches the interval, such as errors on every 50th invoice, the sample may be biased.

Answer: Interval is 50. The first three invoices are 17, 67 and 117. A drawback is that a hidden pattern in the population can bias the sample.

Example 2

An auditor tests a sample of receivables with a total book value of ₹20,00,000, taken from a population of ₹2,00,00,000. Overstatement errors of ₹40,000 are found. Tolerable misstatement is ₹6,00,000. Assume the errors are not anomalies and no other misstatements are known. Project the error and conclude.

Show the solution

Error rate in the sample = ₹40,000 ÷ ₹20,00,000 = 2%.
Projected misstatement = 2% × ₹2,00,00,000 = ₹4,00,000.
Add any other known misstatements to the projection. None are given here, so the total stays at ₹4,00,000.
Compare the total with tolerable misstatement of ₹6,00,000. It is below it, at two-thirds of tolerable misstatement.
The auditor must still judge whether sampling risk is acceptable, because the unexamined population may contain further errors. A projection that is only moderately below tolerable misstatement may not give enough comfort.
Investigate the cause of the errors and whether other balances are affected.

Answer: Projected misstatement is ₹4,00,000. With no other known misstatements, this is below tolerable misstatement of ₹6,00,000. The auditor must judge whether sampling risk is acceptable. If it is not, the auditor should consider extending testing or asking management to correct the errors, and should investigate their cause.

Exam tips

In Section C, link method to population. Say why the method fits, such as monetary unit sampling for overstated receivables.
In OT questions, read for the trap: 'every item has a chance of selection' means sampling, 'selected because of size' means selective testing.
Always project errors before comparing with tolerable misstatement, and say what you would do next.
For sampling and non-sampling risk, define both and give how each is reduced. Two contrasting points earn the marks.
Show working in calculations. Even if the arithmetic slips, method marks are available in Section C.

Audit Sampling and Selecting Items for Testing in other exams

The same ground in other exams, if you are preparing for more than one or want another angle on it.

Audit Sampling and Selecting Items for Testing: frequently asked questions

What is the difference between statistical and non-statistical sampling?

A sample is statistical only if it uses random selection and probability theory to evaluate results and measure sampling risk. Random selection uses tools such as random number generators or tables. Systematic or monetary unit selection can be used in a statistical sample only if the start is random and the results are evaluated using probability theory, but they are not random selection themselves. Non-statistical sampling relies on the auditor's judgement for size, selection and evaluation, and haphazard selection is non-statistical. Both approaches can give sufficient appropriate evidence under ISA 530.

What is sampling risk in audit?

Sampling risk is the risk that the auditor's conclusion based on a sample differs from the conclusion if the whole population were tested. You reduce it by increasing the sample size. It exists because only part of the population is tested.

What is non-sampling risk?

Non-sampling risk is the risk of a wrong conclusion for reasons unrelated to sampling, such as using an inappropriate procedure or misinterpreting evidence. A bigger sample does not reduce it. Good planning, training, supervision and review do.

Which sampling method should I choose in the exam?

Match the method to the scenario. Use random selection for any population where items can be numbered, systematic selection where no pattern matches the interval, and monetary unit sampling when larger values carry more risk. Always explain why your choice fits.