This sample demonstrates report architecture only. It does not certify an AEDT, employer, vendor or dataset and must not be reused as evidence of an actual bias audit.
1. Audit scope
Illustrative AEDT: ExampleRank v3.2. Use: candidate ranking for a fictional NYC hiring workflow. Output family: continuous score. Audit period: fictional 12-month historical dataset.
2. Data basis
The sample assumes historical data with documented outcome semantics, sex and race/ethnicity categories, controlled unknown states and a defined analysis population. No demographic values are inferred.
3. Methodology
The sample follows the score-output structure in 6 RCNY §5-301: full-sample median, category scoring rates and impact ratios for sex, race/ethnicity and intersectional categories, with documented treatment of unknowns.
| Category | N | Scoring rate | Impact ratio |
|---|---|---|---|
| Example Category A | 420 | 0.58 | 1.000 |
| Example Category B | 315 | 0.49 | 0.845 |
| Example Category C | 205 | 0.46 | 0.793 |
4. Limitations and auditor notes
A professional report should distinguish regulatory calculations from professional judgment, identify exclusions and unknown states, document data provenance and avoid converting a single ratio into an automatic discrimination or LL144 pass/fail conclusion.
5. Deliverables
A production engagement may include a technical audit report and a publication-ready Summary of Results. Internal evidence and integrity controls should remain separate from public disclosure.