Editorial II: Solid as a ROC

Abstract

Receiver operating characteristic (ROC) curves were developed to assess the performance of radar operators during the second world war. The radar operators had to be able not only to distinguish between friend or foe blips on their radar screens, but also between signal and noise. The ability of the radar operators to make these life-or-death distinctions was termed the receiver operating characteristic. Plots of right vs wrong answers were called ROC curves, and these curves were taken up in the 1970s by the medical profession to determine the relationship between sensitivity (true positive rate) and specificity (true negative rate) of diagnostic tests. ROC curves are used in medical imaging, materials testing, weather forecasting, information retrieval, polygraph lie detection, and aptitude testing. Though the ROC itself is sound, the values obtained from diagnostic tests often require qualification because the test data on which they are based are of dubious quality.1