Which criterion requires that precision of positive predictions is identical across groups?

Which criterion requires that precision of positive predictions is identical across groups?

• A. Detect Parity
• B. Predictive Rate Parity
• C. Group Fairness
• D. Agentic AI

Answer: (B)

Predicative parity, also called predictive rate parity, is about equal precision of positive predictions across groups. Precision of a positive prediction is the positive predictive value: among those the model flags as positive, the proportion who truly are positive. If this value is identical across groups, the reliability of a positive decision is the same no matter which group you look at. Achieving this often means adjusting thresholds group-by-group when base rates differ, so each group maintains the same PPV. The other terms don’t specifically describe this condition: the idea here is focused on equal PPV across groups, while the remaining concepts are broader or unrelated to this particular metric.

Predicative parity, also called predictive rate parity, is about equal precision of positive predictions across groups. Precision of a positive prediction is the positive predictive value: among those the model flags as positive, the proportion who truly are positive. If this value is identical across groups, the reliability of a positive decision is the same no matter which group you look at. Achieving this often means adjusting thresholds group-by-group when base rates differ, so each group maintains the same PPV. The other terms don’t specifically describe this condition: the idea here is focused on equal PPV across groups, while the remaining concepts are broader or unrelated to this particular metric.