Adverse-impact monitoring
The four-fifths rule (EEOC Uniform Guidelines) flags a group whose selection rate falls below 80% of the top group's rate. Meritline computes this from your explicit protected-attribute mapping — never from names, photos, or extracted text.
Why this dashboard exists
- Protected attributes come from a separate voluntary channel (e.g. self-reported EEO forms) — never inferred from the CV or name.
- This monitor measures outcomes. It never influences a score.
- It measures your decisions, not the model's: a candidate counts as selected when you shortlist them.
The scorer is not blind by default. The model reads the CV text in full, including the candidate's name and contact details. Blind screening — an optional toggle — redacts those direct identifiers before scoring (best-effort, not a fairness guarantee); see methodology §4. This dashboard detects disparate outcomes after the fact; it does not prevent them.
Checking your sign-in…
Action audit trail (this run)
No actions recorded.
Meritline recommends — you decide. Rankings are a starting point — every action stays with the recruiter. Nothing here rejects a candidate on its own.