The symptom
Consider a department where annual ratings cluster tightly around the midpoint of a five-point scale. Almost everyone lands on “meets expectations.” A handful of genuinely outstanding contributors and a handful of genuinely struggling ones both get pulled toward the same middle rating as everyone else, and the scale’s other four options sit almost entirely unused.
On paper, this looks like a well-calibrated, consistent team. In practice, a manager pressed for specifics can usually name, privately and immediately, who the real top and bottom performers actually are — information the ratings themselves don’t reflect at all. Ask that same manager to justify the formal rating for any individual person, though, and the answer tends to be considerably vaguer: solid, reliable, no particular concerns. The private knowledge and the formal record are, quietly, two different documents.
The consequence compounds over several review cycles. A genuinely strong performer, rated identically to an average one for two or three years running, starts to notice that visible effort produces no visible recognition. A genuinely struggling performer, rated “meets expectations” rather than told directly where the gap actually is, has no formal signal that anything needs to change — only whatever informal, undocumented feedback happens to reach them, if any does.
What leaders usually conclude
The common explanation is that the team is, in fact, genuinely uniform — a group of similarly capable people producing similarly solid work, fairly reflected by similar ratings. This explanation has the advantage of requiring no further action and casting no doubt on the review process itself. It also, conveniently, requires nobody to revisit a rating they’ve already delivered, or to have the harder conversation that a more differentiated set of ratings would have required in the first place.
What the evidence suggests
Industrial-organisational psychology has studied this pattern extensively under the name central tendency bias, alongside its close relative, leniency bias. The consistent finding, summarised in Landy and Conte’s widely used academic treatment of performance appraisal, is that raters cluster scores toward the middle of a scale specifically to avoid the discomfort of justifying an extreme judgement — not because the people being rated are actually that uniform.
The mechanism is straightforward once named. An extreme rating, in either direction, typically requires more from the rater: specific evidence for an outstanding score, and an uncomfortable conversation for a poor one. The middle rating requires neither. It doesn’t need to be defended with detail, and it doesn’t require telling someone directly that their performance has fallen short. For a rater managing a full team under real time pressure, the middle option is simply the path that costs the least effort and the least discomfort — regardless of how accurately it reflects the actual spread of performance underneath it.
This connects to a further, specific pattern worth naming directly: in most rating systems, the middle score is also the one that triggers no administrative consequence. A top rating may unlock a bonus or a promotion conversation, each requiring justification. A bottom rating may trigger a formal improvement process, requiring documentation and difficult conversations. The middle rating triggers neither. Raters who default to it aren’t merely being imprecise — they’re often, consciously or not, selecting the one outcome that closes the file with the least additional work.
Research on this bias also finds it worsens under specific, identifiable conditions: when a rater has limited direct contact with the person being assessed, when the evaluation criteria themselves are vague or open to interpretation, and when the rater is managing a larger number of people than they can genuinely differentiate between from memory. Each of these conditions makes the effortless middle option more attractive relative to the harder work of an accurate, defensible extreme rating — which means the bias tends to get worse precisely in the large, fast-growing teams where accurate differentiation matters most.
Competing explanations
A genuinely high-performing, well-matched team could, in principle, produce ratings that cluster naturally without any bias operating at all. The distinction worth checking for is whether the manager’s own private, informal assessment of the team — asked directly, outside the formal process — actually matches the formal ratings. Where a manager can readily name real differences in performance that the ratings themselves don’t show, central tendency is the more likely explanation than genuine uniformity.
The same pattern appears across many kinds of organisation; see Other Sectors for how it is read elsewhere.
Questions worth considering
- If you asked your team’s manager, privately, to rank everyone from strongest to weakest, would that ranking match the formal ratings on file?
- Does your organisation’s middle rating carry any administrative consequence at all — or is it, in practice, the option that requires the least from the rater who chooses it?
- How many ratings in your last review cycle actually required the rater to write specific, defensible evidence — versus how many simply stated a score?
- Would removing the middle option from your rating scale change the results, or would raters simply find a new middle to cluster around?
More on this topic: Leadership.
Your team is facing similar challenges?
Subscribe to The Riddle Ledger
Perspectives are sent as they are published, with no promotional email, and you can unsubscribe at any time.