Calibration Presets

Understanding the five preset configurations and choosing the right level of adjustment for your organisation.

Overview

Different organisations and situations require different levels of adjustment. We provide five preset configurations that range from minimal intervention to maximum correction.

Starting Point

Most organisations should start with the Balanced preset and adjust based on specific needs and stakeholder feedback after evaluating initial results.

The Five Presets

Very Conservative

Minimal adjustments, preserving original scores as much as possible

Conservative

Light adjustments for established reviewer-mean differences

Balanced (Default)

Moderate offset shrinkage with bounded spread correction

Aggressive

Strong corrections for variable reviewer patterns

Very Aggressive

Maximum correction for substantial reviewer variability

Very Conservative

When to Use

When you want to preserve original scores as much as possible, making only minimal adjustments for the largest reviewer-mean differences.

Characteristics

Very high shrinkage factor (K=40): Strongly favours organisational patterns over individual reviewer patterns
No spread adjustment: Only applies the offset calculation, not rating compression/expansion

Suitable For

Organisations with highly trained reviewers
Established rating standards
Where preserving original scores is culturally important

Conservative

When to Use

When you trust your reviewers but want to address clear reviewer-mean differences.

Characteristics

High shrinkage factor (K=25): Still favours organisational patterns
No spread adjustment: Only applies the offset calculation

Suitable For

Mature review systems with periodic reviewer calibration
Organisations with established rating cultures

Balanced (Default)

When to Use

When you want moderate offset shrinkage and bounded spread correction as a starting policy.

Characteristics

Moderate shrinkage factor (K=12): Balances individual and organisational patterns
Spread adjustment enabled: Adjusts both reviewer-mean differences and compression/expansion
Moderate spread limits: Scale factor between 0.88 and 1.15

Suitable For

Most organisations
Mixed reviewer experience levels
Standard implementation scenarios

Why This Is Default

The balanced preset provides:

Moderate offset shrinkage
The narrowest enabled spread limits
A predictable starting point that should still be validated against case mix and observed outcomes

Aggressive

When to Use

When observed reviewer-cohort means or spreads vary materially and you want stronger corrections after reviewing assignment context.

Characteristics

Low shrinkage factor (K=5): Gives eligible reviewer-mean differences more formula weight
Spread adjustment enabled: With tighter controls (K=16, minimum 10 reviews)
Wider spread limits: Scale factor between 0.8 and 1.25

Suitable For

Organisations deliberately testing stronger corrections after case-mix review
Minimal reviewer training programmes
Diverse international teams with different cultural norms

Very Aggressive

When to Use

When you deliberately choose maximum correction after reviewing substantial observed cohort differences and assignment context. Use with caution.

Characteristics

No offset shrinkage (K=0): Applies the full eligible reviewer-mean difference once the minimum three total scores exists
Aggressive spread adjustment: K=8, minimum 8 reviews
Very wide spread limits: Scale factor between 0.7 and 1.35

Suitable For

Organisations undergoing major review system changes
Research and analysis purposes
Controlled analysis of maximum permitted correction

Important Considerations

This preset can produce dramatic adjustments. Ensure stakeholders understand the rationale before implementation.

Comparison Table

Preset	Offset shrinkage (K)	Spread adjustment	Three-score offset weight	Best for
Very Conservative	40	None	0.048	Highly trained reviewers
Conservative	25	None	0.074	Mature review systems
Balanced	12	Moderate	0.143	Default starting policy
Aggressive	5	Enabled	0.286	Variable reviewer means
Very Aggressive	0	Strong	1.000	Maximum correction needed

Practical Comparison

For a reviewer mean of 62, organisation mean of 70, and 15 valid reviewer scores:

Balanced Preset

Reviewer offset: +4.31 points (8 × 14 / (14 + 12))
Eligible scale is clamped to 0.88-1.15; the realised scale depends on the observed spreads

Aggressive Preset

Reviewer offset: +5.89 points (8 × 14 / (14 + 5))
Eligible scale is clamped to 0.80-1.25; the realised scale depends on the observed spreads

Effective target adjustments can differ from these reviewer offsets because of co-reviewers, spread correction, or clamping.

Rule of Thumb

The balanced preset is more conservative and suitable for most situations. The aggressive preset produces stronger corrections and should only be used when you have clear evidence of substantial reviewer inconsistency.