Audit Evidence, Techniques and Reporting
Domain 4 tests whether your findings would survive challenge. The exam focuses on evidence quality, appropriate techniques and report structure.
Evidence quality
Evidence must be sufficient, relevant and reliable. Direct observation, measurements, original records and system logs are stronger than verbal assurances, undated screenshots or summaries. Triangulate: compare a drawing with a physical walk, and a maintenance record with a test log. When sampling, define the population, method and size; record why the sample is defensible.
Techniques
Use document review, interviews, observation, walkthroughs, re-performance, measurement and sampling. For TIA-942, typical evidence includes as-built drawings, cable labelling, redundancy-path separation, maintenance records, UPS and generator test reports, environmental monitoring and change records. Interviews are useful for context and for finding gaps, but they are rarely sufficient alone for a conformity conclusion.
Reporting findings
A defensible finding states the criteria, condition, cause and effect, and cites evidence. Classify it consistently: nonconformity, observation or improvement opportunity; use major/minor only if your scheme defines them. Avoid opinion. Recommendations should address root cause, not only the symptom.
Worked example
A Rated-3 site claims diverse fibre backbone paths. The contractor says “yes, fully diverse.” One as-built drawing shows both paths sharing a tray for 30 metres, and a physical walk confirms the shared tray. The finding is based on the drawing plus observation, not the interview. Criteria: TIA-942 Rated-3 diverse-path requirement. Condition: shared tray over 30 metres. Cause: design review missed physical separation. Effect: one incident can cut both paths. Recommendation: re-route one path and update drawings. The interview alone would not justify a conclusion.
Traps that cost marks
- Accepting verbal assurance as sufficient evidence.
- Citing “best practice” where a TIA-942 criterion exists.
- Reporting a finding without condition, cause, effect or evidence reference.
- Overgeneralising from one sample to the whole site.
- Confusing an observation with a nonconformity.
- Writing a recommendation that only fixes the symptom.