## SA-530: Audit Sampling
### Introduction
- Traditional 100% checking is economically wasteful and still cannot give absolute satisfaction
- Statistical theory: a randomly drawn sample reveals the characteristics of the population
- Sampling is not mandatory; extent of checking is the auditor's judgment (no statute specifies it)
- Auditor should prefer statistical sampling as it is more scientific
---
### Sampling Risk
| Risk Type | In Test of Controls | In Test of Details | Effect on Audit |
|---|---|---|---|
| Over-reliance | Controls appear more effective than they are | Misstatement appears absent but actually exists | Inappropriate audit opinion (affects effectiveness) |
| Under-reliance | Controls appear less effective than they are | Misstatement appears present but actually does not exist | Inefficient audit (affects efficiency) |
---
### Key Definitions
| Term | Meaning |
|---|---|
| Sampling | Applying audit procedures to less than 100% of items; all units have equal chance of selection |
| Sampling Unit | Individual items making up the population; conclusions drawn from the sample are projected to the entire population |
| Anomaly | A misstatement/deviation NOT representative of the rest of the population |
| Tolerable Rate of Misstatement | Monetary amount set by auditor — actual misstatement in the population should not exceed this |
| Tolerable Rate of Deviation | Rate of deviation from a prescribed control — actual deviation should not exceed this |
| Stratification | Dividing a population into sub-populations with similar characteristics |
| Non-Sampling Risk | Risk of erroneous conclusion for reasons other than sampling (e.g., inappropriate procedure, misinterpretation of evidence) |
| Population | Entire set of data from which the sample is selected and about which the auditor wishes to draw conclusions |
---
### Characteristics of Population
| Characteristic | Requirement |
|---|---|
| Completeness | Must contain all items of the financial reporting period |
| Appropriateness | Must be relevant to the specific audit objective (e.g., credit sales → examine debtors) |
| Reliability | Must contain supporting documents; unreliable population = irrelevant population |
---
### Statistical vs. Non-Statistical Sampling
| Feature | Non-Statistical | Statistical |
|---|---|---|
| Basis | Knowledge and experience of auditor | Random selection + probability theory + mathematical methods |
| Bias | Risk of personal bias | No personal bias |
| Objectivity | Neither objective nor scientific | More objective and defensible |
| Best suited for | Smaller organizations; year-end transaction checks | Large organizations with huge transaction volumes |
| Advantage | Simple to operate | Minimum sample size; calculated risk; better description of large data |
| Disadvantage | Sample quality cannot be measured | Not always appropriate (e.g., if staff lacks knowledge) |
Statistical sampling may NOT be appropriate when:
- Another approach provides satisfactory information with less effort
- Audit staff lacks knowledge of the sampling technique
---
### Methods to Collect Sample
| Method | Key Features |
|---|---|
| Simple Random | Random Number Table/Generator; every item has equal chance; suitable for homogeneous populations |
| Stratified Random | Extension of simple random; heterogeneous population divided into homogeneous strata; reduces sample size without increasing sampling risk; results projected per stratum |
| Interval (Systematic) | Population ÷ sample size = sampling interval; e.g., 500 vouchers ÷ 50 = every 10th item; multiple starting points recommended to reduce variability |
| Monetary Unit (MUS) | Value-weighted selection; auditor's efforts directed toward high-value items; conclusion expressed in monetary amounts |
| Block | Continuous range of transactions selected randomly; risk — if management knows the selection pattern, misstatements in other blocks may be overlooked |
| Haphazard | No structured technique; auditor's bias cannot be fully avoided; NOT appropriate for statistical sampling |
---
### Factors Affecting Sample Size
| Factor | Effect | Applies in TOC | Applies in TOD |
|---|---|---|---|
| Higher RMM | Increases sample size | ✓ | ✓ |
| Higher tolerable rate | Decreases sample size | ✓ | ✓ |
| Higher expected rate (deviation/misstatement) | Increases sample size | ✓ | ✓ |
| Higher desired assurance | Increases sample size | ✓ | ✓ |
| Greater reliance on other substantive procedures | Decreases sample size | — | ✓ |
| Appropriate stratification | Decreases sample size | — | ✓ |
| Change in population size | Negligible effect | ✓ | ✓ |
---
### Audit Procedures and Projecting Misstatements
Steps when a selected item cannot be tested:
1. Apply alternative procedures on a replacement item (e.g., if sale document is lost, obtain DC from debtor under SA-505; if no reply, examine subsequent cash receipts)
2. If neither procedure nor alternative can be applied → treat as: Misstatement (TOD) or Deviation (TOC)
Projecting Misstatements:
- Exclude anomalies from the main projection
- However, the effect of anomalous misstatements must be added back to the projection of non-anomalous misstatements
- TOD: Auditor SHALL project misstatements to the population
- TOC: No explicit projection of deviations is necessary
Nature and Cause of Misstatements/Deviations:
- If misstatements share a common feature, extend procedures to all transactions with that feature
- Intentional misstatement/deviation = fraud
- For anomaly: obtain high degree of certainty + perform additional procedures to confirm it does not affect the remainder of the population