Performance Review Bias

Performance review bias is any systematic error that makes a rating reflect the person doing the rating, or factors unrelated to the work, rather than the employee's actual performance. Common types include recency bias, the halo and horns effects, central tendency, leniency, strictness and similarity bias. Bias makes reviews unfair, distorts pay and promotion decisions, and can lead to discrimination claims. Most of it is unconscious, so the fix is better process rather than better intentions.

Also known as: rater bias, rating bias, appraisal bias, unconscious bias in performance reviews

Reviewed

Key takeaways

  • Review bias is when ratings reflect the rater, not the work.
  • Research suggests the rater's own tendencies can explain more of a rating than the employee's performance.
  • Most bias is unconscious, so process fixes work better than training alone.
  • Clear criteria, evidence for every rating and calibration are the core defences.
Common review biases: halo and horns, recency, leniency, central tendency and similarity. Naming them is the first defence.

At a glance

Most common types
Recency, halo, horns, central tendency, leniency, similarity
Root cause
Memory, mental shortcuts, unclear criteria
Main fixes
Written criteria, evidence, multiple raters, calibration
Related
Recency bias, horns effect, performance calibration

Why does performance review bias matter?

Scullen, Mount and Goff (2000) found that more of the variance in job performance ratings came from the rater's own tendencies than from the performance being rated. When ratings drive pay, promotion and exits, that error compounds. Bias also tends to fall hardest on people who are less like their manager, which can widen gender and race gaps.

9 common types of performance review bias

BiasWhat happensExample
Recency biasRecent events outweigh the full periodA strong December hides a weak year
Primacy biasFirst impressions stickA slow first month defines the year
Halo effectOne strength inflates all ratingsA great presenter is rated high on analysis too
Horns effectOne weakness lowers all ratingsOne missed deadline drags down every competency
Central tendencyEveryone is rated in the middleA manager avoids "exceeds" and "below" entirely
Leniency biasRatings are too generousA whole team rated "exceeds" to avoid conflict
Strictness biasRatings are too harshA manager believes nobody truly "exceeds"
Similarity (affinity) biasPeople like the rater score higherSame background or interests leads to higher ratings
Contrast effectPeople are rated against each other, not the standardAn average performer looks weak next to a star

How do you reduce performance review bias?

  1. Define each rating level with clear, observable descriptions.
  2. Require evidence for every rating, from across the whole period.
  3. Capture notes all year through 1:1s and feedback, so memory is not the main source.
  4. Use more than one view: self-assessment, peer and 360 feedback.
  5. Calibrate across managers and look at rating patterns by team, gender and other groups.
  6. Separate the conversations about growth and pay where possible.

Worked example

Before calibration, HR shares each manager's rating distribution. One manager has rated 11 of 12 people "meets", a sign of central tendency. Another has rated every engineer from her former company "exceeds", a possible similarity bias. Both are asked to bring one piece of evidence per rating. Four ratings change, and the company adds written rating descriptions for the next cycle.

Frequently asked questions

What is performance review bias?
Any consistent error that makes a performance rating reflect something other than the employee's actual work, such as the rater's habits or impressions.
What is central tendency bias?
Rating most people in the middle of the scale, avoiding high and low ratings, which hides real differences in performance.
What is leniency bias?
Rating people more generously than their performance deserves, often to avoid conflict or protect the team.
What is similarity bias in performance reviews?
Rating people more favourably because they share the rater's background, interests or working style. It is also called affinity bias.
Can AI reduce performance review bias?
It can help by flagging patterns, such as rating gaps between groups or language differences in written reviews, but people still need to make and own the decisions.
How do you prove a performance review is biased?
Look for patterns: ratings that do not match evidence, consistent gaps between groups, or a manager's ratings that differ sharply from peers' views of the same people.

Sources

  • Scullen, Mount and Goff, "Understanding the latent structure of job performance ratings", *Journal of Applied Psychology* (2000)
  • Edward L. Thorndike, "A constant error in psychological ratings", *Journal of Applied Psychology* (1920)

See it in practice

Compare performance review software

Related reading

Related terms

All glossary terms

One culture-first performance system

Generated in minutes. Owned by you.

The Sageo team, together