The size of the sample, in non-statistical sampling, depends upon the risk of a significant exception[fn]. Provided that items associated with a high risk of significant exception are selected as key items then the following table may be used as a guide to determining the sample size of the remaining population, or each of the remaining stratum, whichever is appropriate.
| Risk of significant exception |
Stratified population [fn] |
Unstratified population | ||
|---|---|---|---|---|
| Sample size (n) |
Max. sample size (n) |
Sample size (n) |
Max. sample size (n) | |
| Mod to high | 20% of N | 30 | 40% of N | 60 |
| Moderate | 10% of N | 25 | 20% of N | 50 |
| Mod to low | 5% of N | 20 | 10% of N | 40 |
| Low | 1% of N | 10 | 5% of N | 20 |
| N = number of items in the stratum or population (excluding key items). | ||||
Example. Assume there are 1000 items in an unstratified population and that the auditor considers there are 25 items that have a high risk of a significant exception with the other items having a moderate risk of significant exception. The auditor will select the 25 high risk items as key items and a sample of 50 items from the remaining 975 items. (Sample size = 0.20 x [1000 - 25] = 195, with a maximum of 50 items).
On occasions when the auditor wishes that the sample have increased reliability [fn] (for example, where the effectiveness of a control procedure is critical to the prevention of a material misstatement), or where the tolerable level of exceptions is close to the expected level, then a 50% increase in the above sample sizes (including the maximum size) is appropriate.
The smaller the sample size, the greater is the sampling risk, and vice versa.
For an alternative view, see The CPA Journal article Customized Tables for Tests of Controls: A Spreadsheet.
Audit testing procedures, index
page
Copyright, Australian Educational Research Pty Ltd. Any person accessing this site agrees to the Terms of Use.