What data do I need for a bias audit?

To conduct a meaningful bias audit, you need candidate pipeline data that includes:

1. anonymous ID (no names or PII required)
2. which stage each candidate reached (applied, screened, interviewed, offered, hired)
3. self-reported race/ethnicity and sex/gender categories (collected via EEO voluntary disclosure)
4. (optional but recommended) — any scores generated by your AI hiring tools
5. to ensure the audit covers the required time period

You need at least one year of historical data, or a statistically significant sample. The minimum recommended sample is 30 candidates per demographic group for reliable statistical analysis.

OnHirely's auto column mapping feature automatically maps your CSV columns to the required fields, making data preparation simple.