Audit evidence, sampling, testing and working papers
What the exam tests
This domain tests whether evidence is sufficient, appropriate and properly documented to support conclusions. Expect questions on evidence reliability, sampling methods, exception evaluation and working paper review.
Core rules and concepts
- Evidence must be sufficient in quantity and appropriate in quality. Reliability rises from independent external confirmation, direct inspection and reperformance, and falls with unsupported inquiry and management-generated summaries.
- Common procedures: inspection, observation, inquiry, confirmation, recalculation, reperformance and analytical procedures. Inquiry alone is rarely sufficient.
- Sampling may be statistical or non-statistical. Attribute sampling tests control deviation rates; variables sampling estimates monetary misstatement. Sample size rises with higher confidence, higher expected deviation and lower tolerable deviation.
- The auditor evaluates sample results against tolerable deviation, investigates anomalies and considers whether exceptions can be projected to the population.
- Working papers must be complete, indexed, cross-referenced, dated and reviewed. They should show the test performed, evidence obtained, exception and conclusion.
Worked example
A population of 2,000 loan approvals is tested for evidence of credit approval before disbursement. The auditor sets tolerable deviation at 5% and confidence at 95%, then selects 60 files using random sampling. Four files lack approval before disbursement, a deviation rate of 6.7%, above tolerable. The auditor does not conclude the control is effective. The auditor discusses the exceptions, tests whether approvals were obtained later, expands the sample or performs substantive testing on the affected loans, and documents the revised conclusion.
Traps
- Relying on inquiry and screenshots as sole evidence.
- Treating a sample exception as isolated without investigation.
- Confusing population with sampling unit.
- Working papers that record conclusions but not evidence tested.