Leniency bias gets most of the attention in performance management literature. Its mirror image, severity bias or strictness bias, is equally damaging and often harder to detect because it feels, to the manager practising it, like rigour rather than error.

A manager who rates most of their team as below average, who applies a standard so demanding that almost nobody meets it, or who uses low ratings as a motivational tool, is producing review data that is just as distorted as the leniency inflater. The difference is that the employees receiving the ratings are also being demotivated, which has an immediate and visible cost.

This article explains how to identify severity bias, why it happens, and the structural and conversational interventions that correct it.

How severity bias manifests

• A manager rates 60-70% of their team as below the midpoint rating while comparable managers in similar functions rate 20-30% of their teams below the midpoint
• The manager's stated rationale for low ratings is often aspirational: "I rate harshly to push people to improve" or "I set high standards"
• Turnover is higher in the team than in comparable teams, but the manager interprets this as departure of low performers rather than as flight from a punitive rating environment
• The manager's high performers receive the same ratings as average performers, because even above-average work is not good enough to reach the manager's threshold

Why managers develop severity bias

• They were managed harshly themselves and believe it was the reason they developed. They are applying the management style that they experienced as formative
• They genuinely do have a high-performing team that they are measuring against a world-class standard rather than the organisational standard. This is rare but real
• They are uncomfortable with performance management in general and use low ratings as a distance mechanism: if everyone is rated low, nobody can argue that specific people were singled out
• In some African organisations where senior leadership publicly values only extraordinary performance, managers learn that average or even good ratings are seen as insufficient by their bosses. The strictness cascades down from above

The cost of severity bias

Severity bias is not just a measurement problem. It is an engagement crisis in a specific team:

• High performers who work hard and receive below-expectations ratings conclude the system is rigged. They leave. The manager loses the people they most need to retain
• Average performers become demoralised. There is no point in improving if improvement produces the same rating as staying where they are
• The team develops a collective resentment toward the manager that affects collaboration, psychological safety, and willingness to bring problems forward

How HR addresses severity bias

Step 1: Surface the distribution in calibration

The same calibration challenge question that addresses leniency addresses severity: "This employee is rated below expectations. What specific evidence from this cycle supports that rating?"

For a manager with severity bias, the evidence may be real but the standard applied is different from what the organisation intends. "Below expectations" in this manager's system means "not quite as good as they could theoretically be." In the organisation's system it means "not consistently meeting core role requirements."

Calibration makes that standard divergence visible and correctable. It is not a personal attack on the manager's judgment. It is alignment on what the rating words mean.

Step 2: Share distribution data with context

Show the manager their distribution alongside the distributions of peer managers in comparable functions. "Your team's distribution shows 65% below the midpoint. Comparable teams in operations and finance show 18-25% below the midpoint. Either your team is significantly underperforming relative to similar teams, or there may be a standard difference. Let us explore this together."

The comparison data is not accusatory. It is the starting point for a calibration conversation that the manager needs.

Step 3: Clarify what "meets expectations" means at the organisational level

Severity bias is often a definition problem: the manager has a higher personal standard for what constitutes meeting expectations than the organisation has set. The conversation should return explicitly to the rating anchors: "Our definition of meets expectations is [specific anchor language]. Does your rating of this employee match that definition?"

Step 4: Address the motivational theory directly

If the manager believes they are using low ratings as motivation, address this explicitly: "Low ratings can motivate the people who have the most options to leave. They rarely motivate people to stay and improve, particularly for sustained periods. The research on this is consistent: accurate ratings tied to specific development plans produce better performance outcomes than systematically low ratings."

Frequently asked questions

What if the manager is right and the team is genuinely underperforming?

This is possible, and it requires a different conversation. If the evidence genuinely supports that the team's output is substantially below the organisational norm, the question is why: is it a management problem, a skills problem, a resource problem, or a design problem? Low ratings without a structured response to the underlying cause are not performance management. They are documentation without development.

How do you address severity bias from a very senior manager?

Senior manager severity bias requires a conversation at the executive or board level, not just an HR calibration session. When a senior leader's team is being systematically underrated, the damage is compounded by the manager's influence over career decisions. The CEO or relevant C-suite leader needs to be brought into the calibration outcome and the standard alignment conversation.

The bottom line

Rating severity is not rigour. It is a misapplication of a high personal standard to an organisational measurement system. The employees affected experience it not as a high bar that inspires improvement but as an unfair environment that inspires exit.

HR's role is to ensure the rating standard is consistent across managers, which means addressing both inflation and severity with the same calibration rigour. A distribution that is uniformly low is just as problematic as one that is uniformly high.

‍