Understanding Performance Evaluation Bias for HR Managers

Last reviewed August 2026

Bias in performance appraisal is a systematic error in judgement that skews how an employee's work is rated, so the score reflects the rater more than the work. The most common types are recency bias, the halo and horns effects, leniency and central tendency, similarity bias and confirmation bias. Each can be reduced with standardised criteria, evaluator training, calibration, 360-degree feedback and audits of rating patterns.

What bias in performance appraisal is

Performance evaluation bias happens when personal prejudice and mental shortcuts distort a manager's judgement during a review. The rating stops reflecting the work and starts reflecting the rater. Evaluator bias is the same thing viewed from the rater's side: the predictable ways an individual assessor's judgement drifts.

The cost is real. Research cited widely in performance management shows around 51% of workers believe their reviews are biased or inaccurate. When people stop trusting the process, engagement drops, promotions get misaligned, disciplinary decisions become hard to defend and good people leave. For HR managers, that combination is a retention problem and a legal risk rolled into one.

Most evaluation bias is unconscious. That matters, because it means good intentions are not a fix. Fair reviews come from process design rather than from asking managers to try harder.

Errors and biases in performance appraisal: the full list

HR textbooks tend to split these into rater errors (patterns in how someone uses the rating scale) and cognitive biases (patterns in how someone weighs evidence). In practice they overlap, and a manager can show several at once. This table covers the ones that turn up most in real review data.

The five biases that do most of the damage

These five account for most of the distortion in performance reviews. Each one has a distinct pattern, which makes it easier to spot once you know what to look for.

1. Recency bias

Recency bias is the tendency to weight recent events far more heavily than the full review period. A strong final month erases an average year, or one recent mistake erases eleven good months. It is the most common appraisal bias because it is built into how memory works: the last few weeks are simply easier to recall than last September.

Example: an employee delivers consistently from July to April, has a rough May, and their annual review reads like the whole year went badly.

It also cuts the other way, and employees know it. The pre-review sprint, where effort spikes in the six weeks before appraisals, is a rational response to a process that rewards recency. If your team's output has an annual bump right before review season, recency bias is probably driving ratings.

2. Halo effect

One positive trait inflates every other rating. A manager who sees an employee as likeable and helpful rounds up their scores on technical delivery too, even when the work has gaps.

Example: a friendly team member misses two project deadlines, but their review still reads "exceeds expectations" across the board because the manager enjoys working with them.

3. Horns effect

The reverse of the halo effect. A single negative trait drags down ratings in unrelated areas.

Example: an employee who missed one visible deadline gets marked down on collaboration and quality as well, even though the rest of their output was strong.

4. Confirmation bias

The evaluator has already formed a view and only registers evidence that supports it. Contradictory evidence gets discounted or forgotten.

Example: a manager who decided early that someone is "not leadership material" remembers the stumbles from the year and overlooks the successful project that person led in March.

5. Similarity bias

Evaluators rate people more favourably when they share a background, education, interests or personality style. It quietly rewards sameness and penalises difference.

Example: a manager gives higher ratings to the team member who went to the same university and barracks for the same footy team, without noticing the pattern.

Leniency, strictness and central tendency deserve a mention alongside these, because they are the easiest to find in the data. A manager whose whole team sits at "meets expectations" is either leading a remarkably even team or avoiding hard conversations, and a rating distribution will show which within one cycle.

How to avoid bias in performance appraisal

You cannot train bias out of people entirely, but you can design a process that gives it far less room to operate. These tactics work together.

• Standardise the criteria. Rate everyone against the same defined, behaviour-based criteria agreed before the review cycle starts. Structured evaluation replaces "how do I feel about this person" with "did this specific behaviour happen".
• Keep notes across the whole period. Recency bias thrives on memory. Managers who log examples throughout the year walk into review conversations with a record of the full period, not just the last six weeks. Quarterly check-ins with a written summary do the same job with less discipline required.
• Train evaluators to recognise bias. Awareness training only works when it is paired with follow-up. Teach managers the patterns above, then check calibration across raters each cycle rather than assuming a single workshop fixed it.
• Run calibration sessions. Before ratings are final, managers of comparable teams compare distributions and talk through the outliers. Leniency, strictness and central tendency rarely survive a room where the numbers sit side by side.
• Use 360-degree feedback carefully. Input from more than one rater dilutes any single manager's blind spots. It needs structure to work: poorly run 360 programs can lower engagement rather than lift it, so keep questions behavioural and feedback specific.
• Audit rating patterns. Look at review data across teams and cycles. If one manager's ratings cluster oddly, or scores differ consistently by gender, age or background, that pattern is evidence worth investigating. Bias shows up in aggregate long before anyone spots it in a single review.

Understanding how different people prefer to work also helps here. A manager who knows their team's work types and team dynamics is less likely to mistake a different working style for a performance problem, which is where a lot of similarity bias starts. A Doer and an Evaluator can produce equally good work in ways that look nothing alike, and a rater who only recognises their own style will score one of them down.

The free work personality assessment gives a manager that map in 4 questions per person, so a different working style gets read as a style rather than a performance gap.

Why fixing evaluation bias pays off

Fair reviews are not just an ethics exercise. When employees trust the evaluation process, they act on feedback instead of disputing it, and high performers stay because progression feels earned. Managers get something out of it too: every promotion and pay decision can be defended with evidence.

Biased appraisals are also an early attrition signal that most HR teams miss. People rarely resign over one unfair review, but they do stop believing the process, and disengagement follows. If your attrition rate is climbing among people who were rated well two cycles ago, look at what changed in who was rating them. Tools like Compono Engage make the pattern visible by measuring engagement and culture alongside performance data, so you can see whether your review process is building trust or burning it. The HR glossary has short definitions of the individual biases if you need them for a policy document.

Frequently asked questions

What is bias in performance appraisal?

Bias in performance appraisal is a systematic error in judgement during employee reviews, driven by personal prejudice and cognitive shortcuts. It causes ratings to reflect the rater's perceptions rather than the employee's actual performance.

What is recency bias in performance reviews?

Recency bias is when the last few weeks of a review period carry far more weight than the whole period, so one recent success or mistake outweighs months of consistent work. Keeping notes throughout the year and running quarterly check-ins are the usual fixes.

What is the halo effect in performance reviews?

The halo effect occurs when one positive trait, such as being likeable, inflates a manager's ratings of unrelated areas like technical skill or reliability. Its opposite, the horns effect, allows a single negative trait to drag down every other rating.

How common is bias in performance reviews?

Very common. Around 51% of workers believe their performance reviews are biased or inaccurate, and most evaluation bias is unconscious, so it persists even under well-intentioned managers unless the process is designed to counter it.

How can HR avoid bias in performance appraisal?

Use standardised, behaviour-based criteria, have managers keep notes across the full review period, train evaluators to recognise common biases, run calibration sessions across raters, add structured 360-degree feedback, and audit rating patterns for demographic or rater-level skew.