Updated 2026-09-16
Interviewer bias is systematic distortion in hiring judgments — favouring candidates who resemble the interviewer, overweighting first impressions, or scoring on likability instead of job competencies — that structured interviews, rubrics, and calibration sessions aim to reduce.
| Bias type | What happens | Mitigation |
|---|---|---|
| Similarity / affinity | Preferring candidates with shared background, school, or hobbies | Structured questions; diverse interview panels |
| Halo / horn | One strong trait colours all scores | Score each competency separately on rubric |
| Contrast | Rating against previous candidate, not rubric | Score immediately after each answer; blind order when possible |
| Confirmation | Seeking evidence for early gut feeling | Require written evidence notes per rating |
Structured interviewing with predefined questions and behavioural anchors reduces inter-rater variance and some bias sources compared to unstructured chats. Meta-analyses show higher predictive validity for structured formats — but interviewers can still rubber-stamp rubrics or score charisma when anchors are vague.
Calibration sessions — where interviewers independently score the same recorded answers then discuss discrepancies — improve consistency. Periodic audit sampling of scorecards for disparate impact patterns connects interviewer behaviour to compliance obligations under EEOC guidance and similar frameworks.
AI interview tools introduce their own bias risks — training data, accent handling, question wording — which require separate monitoring. Neither human nor automated screening is bias-free by default; both need governance.
No. They reduce certain bias sources and improve consistency. Ongoing monitoring, training, and diverse panels remain necessary.
It removes some human similarity bias but can introduce model or data bias. Automated tools used for screening should be audited — see adverse impact and NYC Local Law 144 for context, not legal advice.
When bias or unrelated criteria systematically disadvantage a protected group, selection rates may show adverse impact. Structured rubrics and monitoring help detect and address this.
Head back to Hiring glossary or start now.
· 3 free credits · pay per interview · nothing recurring