What is the difference between calibration and normalization in performance ratings?

What is the difference between calibration and normalization in performance ratings?

• A. Normalization adjusts ratings to reduce manager bias; calibration ensures comparable rating scales across managers or departments.
• B. Normalization adjusts ratings to reduce manager bias; calibration ensures comparable rating scales across managers or departments.
• C. Calibration is about salary decisions.
• D. Normalization changes employee performance data.

Answer: (D)

Calibration and normalization address different fairness concerns in performance ratings. Calibration is about aligning how managers apply the rating standards so a given level of performance means the same thing across departments or teams; it creates a shared standard so ratings are comparable in interpretation. Normalization, by contrast, is a post-rating adjustment that shapes the overall distribution of scores to reduce individual rater tendencies (like leniency or severity) and make ratings comparable across the organization. In short, calibration ensures consistent standards among raters, while normalization adjusts the ratings themselves to fit a common distribution. The idea that normalization simply “changes” data is too vague—normalization targets bias through distribution control, whereas calibration targets rater consistency.

Calibration and normalization address different fairness concerns in performance ratings. Calibration is about aligning how managers apply the rating standards so a given level of performance means the same thing across departments or teams; it creates a shared standard so ratings are comparable in interpretation. Normalization, by contrast, is a post-rating adjustment that shapes the overall distribution of scores to reduce individual rater tendencies (like leniency or severity) and make ratings comparable across the organization. In short, calibration ensures consistent standards among raters, while normalization adjusts the ratings themselves to fit a common distribution. The idea that normalization simply “changes” data is too vague—normalization targets bias through distribution control, whereas calibration targets rater consistency.