On the Probability that X Is Greater than Y
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
This chapter looks in detail at the properties, uses, and ways of expressing one very important measure of the difference between two distributions: the probability that a value randomly chosen from one distribution is greater than a value randomly chosen from the other, the Probability Greater Than . This is a very widely used measure, which is also called the Mann-Whitney U statistic, the Outsurvival Statistic, the Area Under the ROC Curve, the c-statistic, the Common Language Effect Size Indicator, the Individual Exceedance Probability, the D-value, and other things. It is closely related by simple transformations to other measures, including the rank biserial correlation coefficient, and in special circumstances to other measures such as the Gini coefficient and the Number Needed to Treat. The assumption of independence of the two values being compared is of critical importance in deciding whether the Probability Greater Than answers the question, and the consequences of this assumption failing are illustrated. The chapter concludes with a detailed discussion of the question of “which of two treatments is better?” and a discussion of the H-measure for comparing classification rules.
Fields of Study & Subjects
Probability and Statistical Research
Statistical Methods in Epidemiology
Statistical Methods in Clinical Trials
Publication details
- Journal / source
- Chapman & Hall/CRC eBooks
- Pages
- 128–168
- Publication date
- 27 March 2026
- Language
- EN
- Document type
- Book chapter