Affinity bias: what it is, examples and how to avoid it in hiring

Affinity bias is the unconscious pull toward people who remind you of yourself. Same university, same former employer, same sense of humour, and a candidate starts to read as "a good fit" before anyone has checked whether they can do the job. In hiring, that pull quietly narrows the shortlist to people who already look like the team.

TL;DR / Key Takeaways

• Affinity bias is the unconscious tendency to favor people who seem similar to you in background, interests, appearance or manner. It is also called similarity bias.
• An Express Employment Professionals-Harris Poll survey found 61% of US hiring managers judge fit mostly through small talk and email tone.
• Textio found candidates who got offers were described as having a "great personality" 12x more often than rejected candidates.
• Measure it with a four-fifths check per funnel stage. A group pass rate below 0.8 of the highest group's rate is a flag.
• Change the process, not intentions. Write the scorecard first, collect skills evidence before interviews, use structured interviews and score independently before the debrief.
• Affinity bias is not illegal in itself, but selection rates it pushes below the four-fifths line can lead to discrimination claims.
• AI tools help only with the same criteria for everyone, explainable scores and a bias audit. Models trained on past hires can copy the bias.

What is affinity bias?

Affinity bias is the unconscious tendency to favor people who seem similar to you in background, interests, appearance or manner. In hiring, affinity bias pushes interviewers toward candidates who feel familiar and away from equally qualified candidates who do not. Affinity bias is also called similarity bias or the "similar to me" effect.

The word "unconscious" matters. Nobody plans to hire their own reflection. The preference arrives as a feeling, usually a warm one, and the reasons get assembled afterwards. SHRM puts it simply: because affinity bias operates before conscious intention, it often goes unnoticed. That is why it survives in teams who genuinely care about fair hiring, and why it is usually defended as instinct: "she just clicked with us".

Affinity bias is one branch of unconscious bias, sometimes called implicit bias. Our guide to implicit vs explicit bias explains the wider category.

What causes affinity bias?

Affinity bias comes from two ordinary things happening at once: people form impressions fast, and sameness feels safe. Similarity gives the brain a shortcut. The conversation flows, and an easy conversation gets mistaken for a strong candidate.

Four mechanisms do most of the work:

• In-group preference: People give more trust and more benefit of the doubt to those they read as "one of us". A shared employer, school or accent is enough.
• Familiarity and fluency: An easy conversation feels like a good interview. Often it just means the two of you communicate the same way.
• The halo effect: One shared positive, such as the same alma mater, spreads a glow over unrelated qualities like judgment or reliability.
• Confirmation bias: Once the interviewer likes someone, the remaining questions look for reasons to keep liking them, and the harder probe goes to somebody else.

Hiring managers lean on exactly the moments where these shortcuts run free. In an August 2026 Express Employment Professionals-Harris Poll survey, 61% of US hiring managers said informal interactions, such as small talk before or after the interview and email tone, are their most common way to judge whether a candidate fits. Another 87% said they can tell from the first conversation whether a candidate will succeed. Small talk and first conversations are where shared hometowns and shared jokes do their work.

What are examples of affinity bias in hiring?

Affinity bias examples in hiring are easier to catch when you stop looking for prejudice and start looking for comfort. The table below shows where affinity bias tends to sit in a normal hiring process, what it sounds like out loud, and what it costs.

A concrete example. Two candidates apply for a senior account manager role at a 900-person SaaS company. One spent six years at companies the hiring manager knows by name. The other ran the same function at a retailer, managing a bigger book with a smaller team. The first candidate gets forty minutes of shop talk about customers they both remember. The second gets a tidy set of competency questions and finishes eight minutes early. Both are scored "good". Only one was given room to be impressive.

That is affinity bias with nobody behaving badly. The interviewer was warmer because the conversation was easier, and the conversation was easier because the backgrounds matched.

The written record shows the same pattern. Textio's State of Interview Feedback 2025 found that candidates who received an offer were described as having a "great personality" 12x more often than candidates who were rejected. Personality notes are not wrong in themselves, but when they outweigh notes about job skills, likability is doing the deciding.

What does affinity bias look like in the workplace?

Affinity bias in the workplace looks like opportunity flowing to the people a manager feels closest to. The same pull that shapes a shortlist keeps working after the hire, and its effects compound over years.

• Promotions: A manager backs the person they have lunch with over a stronger performer who is quieter or works remotely.
• Performance reviews: Ratings drift upward for people who share the manager's working style and downward for those who do not.
• Stretch projects: High-visibility work goes to people who "get it", which usually means people who think like the leader.
• Recognition: Praise in team meetings concentrates on a familiar circle while steady contributors go unnoticed.
• Mentoring: Senior people mentor juniors who remind them of themselves, so the next generation of leaders looks like the current one.

Across the teams we talk to, the tell is the same: when the people promoted over two years all share a background with the person who promoted them, affinity bias is the first thing to rule out.

Why is affinity bias a problem for hiring teams?

Affinity bias is a problem because it lowers the quality of hiring decisions and creates legal exposure at the same time. Affinity bias narrows the range of experience on a team, makes outcomes harder to defend, and, because it feels like good judgment, goes uncorrected for years.

Start with the legal edge. Under the Uniform Guidelines on Employee Selection Procedures, US agencies use a rule of thumb: a selection rate for any race, sex or ethnic group that is less than four-fifths, or 80 percent, of the rate for the highest group is generally treated as evidence of adverse impact.

Affinity bias does not need intent to push a selection rate under that line. A few interviewers hiring people who feel familiar, repeated across a year of requisitions, is enough. Affinity bias itself is not a legal category, but the outcomes it produces can be.

The gaps are measurable at scale. In 2024, researchers sent 83,000 fictitious applications to 11,000 entry-level jobs at 97 Fortune 500 companies. NPR reported that the typical employer called back applicants with white-sounding names about 9% more often than applicants with Black-sounding names, and the worst employers about 24% more. The same study found less bias at firms with more centralized HR and standard hiring workflows. Process, not good intentions, was the difference.

Then the quality edge. Every time a familiar candidate beats a stronger unfamiliar one, the team loses skill it needed. Teams that hire fastest feel it most, because each new hire becomes the model for the next.

How do you measure affinity bias in your hiring funnel?

Affinity bias is measured by comparing outcomes, not by asking interviewers whether they are biased. Pull pass rates for each stage of your funnel, split by the groups you can lawfully track, and look for the stage where one group's rate drops well below another's. That stage is where to look first.

A simple four-step check works for most TA teams:

1. Pick one high-volume role and export the last six to twelve months of applicants from your ATS with their stage outcomes.
2. Calculate a pass rate per stage: resume screen, phone screen, interview, offer. Do it for each group you track.
3. Apply the four-fifths check per stage: Divide each group's pass rate by the highest group's rate. Anything under 0.8 is a flag. For example, if 50% of one group passes the interview and 35% of another does, the ratio is 0.7.
4. Compare interviewers: For the flagged stage, compare each interviewer's average score and pass rate. One interviewer who passes far more candidates from their own background than peers do is a coaching conversation, not a scandal.

Two softer signals help as well. Read a sample of debrief notes and count how often "fit", "energy" or "personality" appear next to a pass decision without a skill named. And check where your last ten hires came from: the same three employers or schools is a pattern worth a second look.

Pro tip: Run the check on closed requisitions before live ones. We have seen teams discover that one stage, usually the hiring manager interview, carries almost all of the gap, which tells you exactly where structure will pay off first.

How do you avoid affinity bias in hiring?

Affinity bias in hiring is avoided by changing the process, not by trying harder to be fair. Decide what good looks like before you meet anyone, collect skills evidence before the conversation, ask every candidate the same questions, and score independently before anyone talks.

Awareness training names the problem but is weak at changing a decision made in the ninth minute of a friendly interview. Controls work better than good intentions:

1. Write the scorecard before the job goes live: Four or five skills, each with what a weak, adequate and strong answer looks like. Once the bar is written down, "not quite a fit" has to become a rating, and that is usually where the vague objection dies.
2. Put skills evidence before the interview: A work sample or role-based assessment completed before anyone meets the candidate stops a warm conversation from setting the ranking. Our guide on using skills assessments to reduce bias covers how to set this up.
3. Use structured interviews: The same questions, in the same order, rated on the same scale. The US Office of Personnel Management notes that structured interviews raise agreement between interviewers by limiting the discretion an interviewer is allowed, and show less adverse impact. Discretion is the room affinity bias lives in. See the difference between structured and unstructured interviews.
4. Hide names and schools at the screen: SHRM recommends blind evaluations and forms that focus on qualifications rather than personal similarities. Our guide to blind hiring shows what to redact.
5. Score independently, then discuss: Every interviewer submits a rating before the debrief. Otherwise the first confident voice sets the anchor.
6. Replace "culture fit" with named behaviours: Swap "fit" for the two or three behaviours the role needs, and make each one rateable. Read more on culture fit vs skill fit.
7. Mix the panel and rotate the first interviewer: One person's taste should not filter the whole pipeline.

The honest caveat: structure costs time up front and makes interviews less pleasant for the interviewer. A structured interview is a worse conversation than a free-flowing one. You are buying comparability, and comparability is what lets you defend a decision six months later.

Can AI hiring tools remove affinity bias?

AI hiring tools can reduce affinity bias when they score every candidate against the same criteria, but they can also copy it. A screening model trained on who the team hired before can learn the team's past preferences, including shared schools and employers, and repeat them at scale.

Three checks keep AI on the right side:

• Same criteria for everyone: Scores should come from skills the job needs, not from resemblance to past hires.
• Explainable scores: Reviewers should see why a candidate scored as they did, so an odd result can be challenged.
• Human decision, audited tool: In New York City, Local Law 144 lets employers use an automated employment decision tool only if it has had a bias audit within one year of use, the audit results are public, and candidates get notice. Ask any vendor for their audit.

How do you reduce affinity bias in the workplace after the hire?

Affinity bias in the workplace is reduced the same way as in hiring: written criteria, comparable evidence and decisions made by more than one person. The difference is that workplace decisions happen continuously, so the controls need to be part of normal management.

• Promotion criteria in writing before candidates are discussed, with evidence for each.
• Calibration sessions where managers compare ratings across teams before reviews are final.
• Open calls for stretch projects instead of quiet hand-picks.
• A yearly look at who got what: promotions, raises and high-visibility work by team and group.
• Structured mentoring that pairs people across backgrounds, not only by who asks.

How is affinity bias different from other biases?

Affinity bias is one specific bias inside the wider category of unconscious bias, the one driven by similarity. Affinity bias is the same thing as similarity bias, and is often confused with the halo effect and confirmation bias, which behave differently and need different fixes.

The practical reason to keep them apart: affinity bias responds to structure and blind evidence, while confirmation bias responds to independent scoring before discussion. Treating them as one blob is how teams run one training session and call it solved. For the rest of the family, see our overview of common types of hiring bias and our guide to first impression bias.

Hire on evidence, not on familiarity

Testlify is built to put skills evidence in front of the hiring team before the friendly conversation does. Teams turn a job description into a role-specific assessment, pick from 3,500+ tests (including situational judgment and role-based tests), and run structured AI interviews that score every response against the same criteria. Results come with per-skill breakdowns and weighted scoring, and the final decision stays with people.

Testlify's matching model completed a third-party bias audit under NYC Local Law 144 on June 16, 2023; see the NYC bias audit.

Start with one role: write the scorecard, add a skills test from the test library before the first call, and have every interviewer score before anyone talks. If you want help mapping where evidence enters your funnel today, book a demo.

FAQs