The question every reviewer asks first
"Why 25?" That's the comment you'll find on a review note when someone picks a sample size without showing how they got there. It's a PIOOMA number (pulled it out of my audit experience) and inspectors treat it accordingly. The fix isn't complicated, but it does require connecting four inputs into one documented calculation.
Sample size determination is the process of calculating how many items an auditor must test from a population to draw a reliable conclusion about it. ISA 530 governs the process, and ISA 530 .A10 lists the factors that drive sample size for tests of details: the acceptable level of sampling risk, the tolerable misstatement, the expected misstatement in the population, and the characteristics of the population itself. Reduce the acceptable risk and the sample grows. Narrow the tolerable misstatement and it grows again.
The standard doesn't prescribe a formula. ISA 530 .A11 notes that the auditor may use statistical or non-statistical methods, provided the sample is sufficient to reduce sampling risk to an acceptably low level. We've seen most mid-tier firms use either a statistical formula (the AICPA-style or MUS approach) or a table-based method from their methodology provider. What matters is that the inputs are documented and the relationship between risk assessment, performance materiality (PM), expected error, and sample size is visible in the file.
One point that gets missed: ISA 530 .A10(e) requires you to consider the stratification of the population. If you stratify well, the required sample size for each stratum drops because variability within each stratum is lower.
Key Points
- A larger sample does not automatically mean a better test. The design of the sample matters more than the count.
- Reducing the acceptable level of sampling risk forces the sample size up, not down.
- If you can't justify your sample size with reference to the population and your risk assessment, an inspector will challenge it.
- In our experience, most sample size disputes trace back to tolerable misstatement being set without reference to PM.
Why it matters in practice
Getting sample size wrong hits in both directions. Set it too low and the auditor accepts too much sampling risk, potentially missing material misstatements. Set it too high and the team over-audits, burning hours on immaterial items without improving the conclusion.
Worked example: Lindberg Precision AB
Client: Swedish engineering group, FY2024, revenue SEK 380M (approx. EUR 33M), IFRS reporter. The population is 1,247 trade receivable balances totalling SEK 89M.
Step 1: set tolerable misstatement
PM is SEK 2.7M. The engagement team sets tolerable misstatement for this test at SEK 2.7M (equal to PM, since no other substantive procedures address the same assertion).
Step 2: estimate expected misstatement
Prior year testing found SEK 180K of projected misstatement. The team estimates expected misstatement at SEK 200K to allow for growth.
Step 3: determine confidence factor
The team uses a 95% confidence level (risk factor 3.0) because the risk of material misstatement for receivables existence is assessed as significant.
Step 4: calculate sample size
Using MUS: sample size = (population book value x confidence factor) / tolerable misstatement = (89,000,000 x 3.0) / 2,700,000 = 99 items. Rounded to 100.
A sample of 100 items from 1,247 balances is defensible. The four inputs (risk assessment, confidence factor, tolerable misstatement, expected misstatement) are documented and traceable. If the team had set tolerable misstatement at half PM (SEK 1.35M), the sample would have doubled to 198. Nobody enjoys reworking the calculation after a reviewer challenges it, but documenting the inputs up front is how you avoid that conversation entirely.
Key standard references
- ISA 530 .A10 covers the factors affecting sample size for tests of details: acceptable sampling risk, tolerable misstatement, expected misstatement, and population characteristics.
- ISA 530 .A11 permits statistical or non-statistical methods, provided the sample is sufficient to reduce sampling risk to an acceptably low level.
- ISA 530 .A8–A9 addresses sample size for tests of controls. The factors are tolerable deviation rate, expected deviation rate, the assessed risk of material misstatement, and the assurance expected from other procedures.
- ISA 530 .A10(e) requires the auditor to consider stratification of the population when determining sample size.
Related terms
Related tools
Related tools
Related reading
Frequently asked questions
What are the main factors that determine sample size?
ISA 530.A10 lists four key factors: acceptable sampling risk, tolerable misstatement, expected misstatement, and population characteristics.
Does a larger sample always mean a better test?
No. The design of the sample matters more than the count. A large sample using the wrong procedure produces no useful evidence.
Why do inspection findings often challenge sample sizes?
Most disputes trace back to tolerable misstatement being set without reference to performance materiality, or sample sizes documented without the inputs that produced them.