The halo effect in interviews: definition, examples and how to prevent it

The halo effect in interviews is the tendency to let one strong positive impression, such as a confident manner, a well-known former employer or one excellent answer, raise an interviewer's ratings on other, unrelated competencies. Its mirror image, the horns effect, lets one negative impression drag every rating down. The U.S. Office of Personnel Management defines the halo effect as allowing ratings of performance in one competency to influence ratings for other competencies, and the standard fix is to rate each competency only from the evidence collected for it.

This page explains where the term comes from, how the halo and horns effects show up in real notes and scorecards, worked examples, and the controls that stop them. It sits alongside our wider list of common interviewer mistakes and the stage-by-stage guide to reducing bias in interviews.

Where does the term "halo effect" come from?

The psychologist Edward L. Thorndike named it in "A Constant Error in Psychological Ratings", published in the Journal of Applied Psychology in 1920. He had noticed, in a 1915 study of employees at two large industrial corporations, that ratings of the same person on intelligence, industry, technical skill and reliability were "very highly correlated and very evenly correlated". The raters seemed unable to judge each quality separately; they formed a general view of the person as rather good or rather inferior and let it color everything.

He found the same pattern in U.S. Army officer ratings. One conscientious flight commander, rating 137 aviation cadets under instructions to judge physique, intelligence, leadership and character independently, produced correlations of .51, .58 and .64 between intelligence and the other three, which Thorndike argued were higher than reality could explain. In teacher ratings he cites, the rating of a teacher's voice correlated .63 with the rating of her intelligence. His conclusion is still the best short instruction for interviewers: the observer should report the evidence, and each quality should be rated on its own evidence, without knowledge of the evidence about the others.

Note what Thorndike did not say. Some qualities genuinely go together, so a candidate who scores high everywhere is not proof of halo. The error is when ratings are more uniform than the evidence behind them.

How is the halo effect different from other interview biases?

OPM's Structured Interviews: A Practical Guide (September 2008) lists halo with leniency, strictness, central tendency, rater bias and similar-to-me as the common rating errors, and first impressions, negative emphasis and contrast as common interviewing mistakes. Halo is the one that hides best, because the scores look decisive and consistent.

How does the halo effect show up in interview notes and scorecards?

• A flat profile. Every competency gets the same score, or within one point, even though the questions covered very different skills.
• Adjectives instead of evidence. "Sharp", "polished", "impressive" written once at the top and used to justify every rating.
• One example doing all the work. The same story cited as evidence for problem solving, leadership and communication.
• Scores for things nobody asked about. A 4 for stakeholder management when no question covered it. That score can only have come from the overall impression.
• Brand-name halo. A well-known employer or school on the resume lifts ratings of the answers. OPM's guide warns that interviewers who look at resumes during the interview may rate candidates with strong resumes more favorably.
• Presentation halo. Fluent, confident delivery is scored as if it were the content. Unless the role is about presenting, rate what the candidate did, not how smoothly they described it.
• Horns from one moment. A late arrival, a dropped video call or an awkward answer to the first question lowers scores on competencies covered well later.

What does the halo effect look like in practice?

An invented example: a scorecard for a customer success manager, four competencies, scale of 1 to 4.

WITH HALO
Overall: Very impressive. Led renewals at a well-known SaaS company.
Account planning        4   Clearly strong.
Handling escalations    4   Great presence, very calm.
Data and reporting      4   Seems very analytical.
Cross-team influence    4   Would work well with product.

WITHOUT HALO
Account planning        4   Built quarterly plans for 30 accounts;
                            described cutting churn from a named
                            segment and how she measured it.
Handling escalations    3   One example; resolved it, but escalated
                            to her manager before trying options.
Data and reporting      2   "Our analyst ran the dashboards." Could
                            not describe a metric she defined.
Cross-team influence    -   Not assessed: no question asked.

The first version is not dishonest. The interviewer genuinely formed a strong impression from the renewals story and the candidate's composure. But three of the four scores rest on that impression, and one is for a competency nobody tested. The second version is less flattering and far more useful: the next interviewer knows to probe data skills and cross-team work.

The horns version is just as common. A candidate who stumbles through the first question, then gives a precise, well-evidenced answer on reporting, still gets a 2 on reporting because "they never really recovered". The fix is the same: score the reporting answer against the reporting anchor.

How do you prevent the halo effect in interviews?

1. Map every question to one competency before the interview, so each answer has a home. The interview scorecard template is laid out this way.
2. Take notes of what was said, under the competency it belongs to. Verbatim fragments, numbers and examples, not verdicts. Evidence-based interview feedback explains the method.
3. Rate each competency only from its own evidence, against written anchors. If the evidence is under a different heading, it does not count here. Behavioral anchors for interview scores has wording to start from.
4. Use "not assessed". A competency you did not ask about gets no score, not a guess.
5. Rate straight after the interview, as OPM's guide recommends, while the specific answers are fresh, rather than at the end of the day when only the impression is left.
6. Read the resume before, not during. Pull questions from it in advance, then close it.
7. Score independently before any discussion. One person's halo spreads fast in a debrief. See independent interview feedback before the debrief.
8. Split competencies across a panel, so each interviewer goes deep on two or three and nobody rates everything from one conversation.

Tools can help with the evidence step. Interview Signal, which we make, fills the scorecard with the candidate's own quotes under each score and removes any quote it cannot find in the transcript, which makes a score with no evidence under it easy to spot. The rating remains the interviewer's call.

How does the halo effect spread between interviewers?

Halo does not stay with one person. A recruiter's screen note that says "excellent candidate, must hire" reaches the hiring manager before the candidate does, and the hiring manager walks in expecting excellence. A well-known referrer does the same. So does a debrief where the most senior person speaks first.

• Hand over questions, not verdicts. Pass on what is still unknown ("could not yet describe a metric she owned") rather than an overall rating. Interview handoff notes between rounds has a format.
• Keep referral context out of the scorecard. Who referred the candidate is a sourcing fact, not evidence about any competency.
• Reveal scores together. In the debrief, everyone's ratings go up at once, and the discussion starts with the competencies where the panel disagrees.

How can you check your own scorecards for halo?

• The spread check. Look at your last ten scorecards. If most candidates have every competency within one point, ask whether the questions really failed to separate them or whether you rated the person.
• The evidence check. For each score, point to the line in your notes that supports it. A score you cannot trace to evidence is a halo or horns score.
• The swap check. Imagine the same answer from a candidate with a less impressive resume, or one who started badly. If your score would change, the score is about the impression.
• The panel check. In calibration sessions, compare how often each interviewer gives flat profiles. Patterns across many candidates show habits that one scorecard cannot.