The Fundamental Attribution Error: Why Performance Reviews Blame the Person and Excuse the System

60-Second Summary

Fundamental Attribution Error (FAE): we over-weight disposition and under-weight situation when judging others — the opposite when judging ourselves.
In reviews: missed goals become 'they lack ownership'. Our own missed goals become 'the roadmap changed'. Same evidence, different attribution.
The result: low ratings cluster where managers, tools, or contexts are weak — because the people in those situations look bad. It's mislabelled 'talent'.
Fixes: distinguish situational vs dispositional evidence in writing, force a 'what would a great person do in this situation?' check, and calibrate by looking at outcomes across managers.
You cannot judge performance without judging the system it happened in. Doing so systematically mislabels people.

An engineer misses two deadlines. Their manager writes, 'lacks urgency, struggles with prioritisation.' The manager, one floor up, is behind on their own OKRs and explains it as 'the CRO kept changing the roadmap'. Same evidence, opposite attribution. Multiply that across a review cycle and you have a rating distribution shaped by cognitive bias, not performance.

What Lee Ross actually said

“The tendency for attributers to underestimate the impact of situational factors and to overestimate the role of dispositional factors in controlling behavior.”
— Lee Ross, Advances in Experimental Social Psychology (1977)

Ross's 1977 essay named a pattern that Fritz Heider had already sketched in the 1950s and Jones and Harris had demonstrated in 1967 (subjects attributed a persuasive essay to the writer's real belief even when told the writer was assigned the position). Decades of replication show that attribution is systematically skewed: your behavior is context; theirs is character.

How FAE shows up in performance management

Situational cause vs dispositional label

What actually happened

Roadmap changed 3 times in 6 weeks
Onboarded onto a team with no documentation
Working through a family bereavement
Assigned an unclear scope and a peer who blocks reviews
New tool rolled out without training

How the review sentence reads

'Struggled to focus on priorities'
'Slow ramp; low ownership'
'Inconsistent quarter; energy dropped'
'Not proactive; needs to push harder'
'Below expectations on execution'

None of the right-hand sentences are technically false. All of them describe the person while ignoring the system. And they're what goes into calibration, promotion, and PIP decisions.

The evidence: attribution asymmetry in ratings

62%

of manager rating variance

explained by the rater, not the ratee — Mount, Scullen & Goff (2000)

1.5x

higher 'low performer' rate

under first-time managers vs experienced ones (SHRM benchmarking)

3.7x

increase in negative-trait attribution

when the observer is under time pressure — Gilbert, Pelham & Krull (1988)

~50%

of PIP outcomes overturn

when a second, independent manager reviews the situation (Bock, Work Rules!)

The Mount et al. finding is the killer: the strongest predictor of a manager's rating of an employee is who the rater is, not who the employee is. That is FAE and idiosyncratic rater effects at industrial scale.

Five design fixes for review systems

Building attribution discipline into reviews

1
Separate outcome, behaviour, and context in writing
Force the review template to have three fields: what happened, what the person did, what the situation was. Attribution errors happen when these are merged into one sentence.
2
The 'competent stranger' test
Before writing the final rating, ask: 'if a competent, motivated person had been in this exact situation, would they have delivered materially different results?' If no, the rating is about the situation, not the person.
3
Calibrate across managers, not within
If Manager A rates 40% of the team 'below expectations' and Manager B rates 5%, don't average — investigate. Very often the difference is rater style, team context, or onboarding quality.
4
Require dispositional claims to have >1 example across >1 quarter
'Lacks ownership' needs multiple concrete instances across contexts, not one project that went sideways with three confounding variables.
5
Use skip-levels and peer input as attribution check
If the manager sees 'low ownership' and every peer sees 'blocked by unclear scope', the manager's attribution is the outlier.

The manager telltale

Sentences that start with 'they just…' or 'they always…' are almost always dispositional shortcuts. Sentences that start with 'in the context of…' or 'given the situation of…' are attribution-aware. Read your last five reviews and count the ratio.

FAQ

Frequently asked questions

Isn't situation just an excuse for underperformance?

Sometimes. But treating everything as disposition guarantees you'll misclassify people whose situations were genuinely broken. The point isn't to excuse — it's to correctly locate the cause so the fix works.

How does this differ from making excuses for people?

Attribution-aware reviews still land on a rating. They just require that dispositional claims be supported and that situational factors be surfaced, not hidden.

Doesn't calibration fix this?

Only if calibration examines managers' attribution patterns. Most calibration ratifies bias by forcing a curve; good calibration audits who is rating harshly and why.

Takeaways

The strongest predictor of a rating is the rater — that alone should reshape how you run reviews.
Force writers to separate outcome, behaviour, and context. Merged sentences are where bias hides.
'Competent stranger' test: would a great person have done materially better in this exact situation? If no, don't rate the person low.
Calibration must audit manager patterns, not just outcomes.