Youdens J Statistic
Youdens J Index (Youden Index) · Also known as: Youden Index, Sensitivity + Specificity - 1
Youdens J statistic, also called the Youden index, measures the maximum difference between the true positive rate and false positive rate across different classification thresholds. It is useful for selecting optimal cutoff points in diagnostic testing.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use Youdens J when selecting the optimal classification threshold for diagnostic tests or when you want to balance sensitivity and specificity equally. It is particularly useful in medical diagnosis where you need to find the threshold that maximizes overall test performance.
Strengths & limitations
- Provides a simple, interpretable metric for threshold selection.
- Balances sensitivity and specificity equally, useful when both matter.
- Directly related to ROC curves; maximizing J often selects the most effective threshold.
- Historically validated in medical diagnosis literature.
- Equally weights sensitivity and specificity, which may not match domain requirements.
- Can be unstable if either sensitivity or specificity is poorly estimated due to small class sizes.
- Does not account for the costs of false positives versus false negatives.
- May select different thresholds than other criteria (e.g., ROC-maximizing threshold).
Frequently asked
How is Youdens J related to balanced accuracy?
Youdens J = sensitivity + specificity - 1, while balanced accuracy = (sensitivity + specificity) / 2. They are directly related: J = 2 * balanced accuracy - 1. Both balance sensitivity and specificity, but J ranges from -1 to 1 while balanced accuracy ranges from 0 to 1.
Should I always use the threshold that maximizes J?
Not necessarily. Maximizing J assumes sensitivity and specificity are equally important. If they have different costs in your application, choose a threshold that reflects those costs. For example, in disease screening, a higher sensitivity (fewer missed cases) might be more important than specificity, suggesting a lower threshold than J-optimal.
How does Youdens J relate to ROC curves?
Youdens J = max(TPR - FPR) over all thresholds. On an ROC curve, this equals the maximum vertical distance from the diagonal (random guessing line). The threshold that maximizes J often corresponds to a point on the ROC curve with good balance between true and false positive rates.
Sources
- Youden, W. J. (1950). Index for rating diagnostic tests. Cancer, 3(1), 32-35. DOI: 10.1002/1097-0142(1950)3:1<32::aid-cncr2820030106>3.0.co;2-3 ↗
- Perkins, N. J., & Schisterman, E. F. (2006). The inconsistency of optimal cutpoints obtained using two criteria based on the receiver operating characteristic curve. American Journal of Epidemiology, 163(7), 670-675. DOI: 10.1093/aje/kwj063 ↗
How to cite this page
ScholarGate. (2026, June 3). Youdens J Index (Youden Index). ScholarGate. https://scholargate.app/en/model-evaluation/youdens-j-statistic
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Balanced AccuracyModel Evaluation↔ compare
- F1-ScoreModel Evaluation↔ compare
- SpecificityModel Evaluation↔ compare