8 Unconscious Bias Examples in Hiring—and What to Check

Unconscious bias can appear when a recruiter favors a familiar background, an interviewer mistakes confidence for competence, or a panel assumes an older candidate will struggle with new software. The common problem is an assumption taking the place of job-related evidence.

The Royal Society’s selection-panel briefing describes unconscious bias as judgments shaped by experience, assumptions or interpretations without awareness that this is happening. For more on the distinction, see our unconscious bias definition.

The examples below are hypothetical warning signs—not proof of someone’s motives. Several biases can overlap, and the same behavior could be conscious rather than unconscious.

Eight examples you can recognize in hiring decisions

1. Affinity bias: shared interests become a hiring advantage

Example: “We went to the same university and clicked immediately. I can see them fitting in.” The interviewer gives the candidate a higher teamwork score without evidence of how they collaborate.

What to check: Remove the shared background from the explanation. What collaboration example still supports the score?

The U.S. Office of Personnel Management (OPM) identifies this as a “similar to me” rating error. Our affinity bias guide explores the mechanism in more detail.

2. Halo effect: one strength lifts unrelated scores

Example: A candidate explains their career clearly and confidently. The interviewer also rates their analytical ability highly, although their problem-solving answer lacks a workable solution.

What to check: Score communication and problem-solving separately. OPM specifically warns against letting a strong oral-communication rating influence a problem-solving rating, irrespective of the evidence for that competency.

3. Horns effect: one weakness becomes an overall verdict

Example: A candidate stumbles over one introductory answer. The panel subsequently describes them as “not capable,” despite a strong work sample and detailed technical answers.

What to check: Identify the actual weakness and its relevance to the role. A weak answer may justify a lower score for that answer; it does not automatically establish weaknesses across unrelated skills.

4. First-impression bias: the decision arrives before the evidence

Example: An interviewer decides in the first minute that a candidate “doesn’t seem senior enough.” Later answers are evaluated against that early impression rather than the role’s requirements.

What to check: Delay the overall recommendation until the required questions are complete. Record what the candidate demonstrated before deciding whether they meet the level.

5. Contrast effect: the previous candidate becomes the benchmark

Example: A capable applicant follows an exceptional speaker and receives a low communication score. The same answer might have seemed strong after a less fluent interviewee.

What to check: Compare the answer with the scoring standard, not the person interviewed immediately before them. OPM lists both first impressions and contrast effects among common interviewing mistakes.

6. Gender stereotyping: identical behavior gets different labels

Example: One candidate’s direct disagreement is called “leadership”; another’s comparable disagreement is called “abrasive” because it conflicts with the evaluator’s expectations about gender.

What to check: Describe the behavior: Did they explain the disagreement, listen and propose a resolution? Apply the same standard to everyone rather than scoring an adjective.

7. Age stereotyping: age stands in for learning ability

Example: “They may not adapt to our new systems” is written about an older applicant without asking about recent learning or assessing the relevant software skills.

What to check: Use evidence of learning and task performance. Neither age nor graduation year answers whether someone can use the tools required for the job.

8. Presentation bias: interview manner becomes a skill score

Example: Limited eye contact or a flat speaking style is interpreted as low engagement, even though the candidate gives relevant answers and completes the task successfully.

What to check: Separate interview manner from the competency being assessed. If communication matters, define the required behavior—such as explaining an issue clearly—rather than assuming a preferred presentation style proves competence. OPM warns against judging candidates merely by how they act during an interview, rather than by evidence relevant to the competency.

What changes when AI scores the candidate?

Software does not have unconscious attitudes, but automated scoring can create similar unfair barriers. For example, a system that rewards similarity to current successful employees may reproduce exclusions in that workforce. The U.S. Department of Justice’s ADA guidance identifies this disability-discrimination risk and warns that facial or voice analysis may screen out qualified applicants with autism or speech impairments.

For HR, ask what the score measures and whether it evaluates relevant job skills. “AI-generated” is not evidence of fairness. Use our AI hiring vendor validation checklist for a fuller review.

Replace the impression with a checkable note

A useful interview note follows this pattern:

Criterion → observed answer or behavior → scoring standard → remaining uncertainty

For a role whose conflict-resolution rubric requires evidence of attempting direct resolution and knowing when to escalate, replace “not a culture fit” with a specific note:

“On the conflict-resolution question, the candidate described escalating the disagreement but did not explain an attempt to resolve it directly. The rubric requires evidence of both; the direct-resolution component remains unestablished.”

That creates a specific assessment another reviewer can examine. For a wider process redesign, see how to reduce hiring bias.

Candidates can ask which competencies will be assessed and offer concrete examples tied to those requirements. A vague rejection alone cannot establish unconscious bias, and candidates are not responsible for correcting an employer’s process.

This article is informational, not legal, HR or employment advice. Legal obligations depend on jurisdiction and circumstances; confirm requirements with qualified counsel.