摘要：

Signal detection theory, as developed in electrical engineering and based on statistical decision theory, was first applied to human sensory discrimination 40 years ago. The theoretical intent was to provide a valid model of the discrimination process and the methodological intent was to provide reliable measures of discrimination acuity in specific sensory tasks. In a detection task, the observer decides how likely the presence of a signal must be before he or she will report that a signal is present rather than just noise. In a recognition task, in which a signal is known to be present, the observer decides how likely the presence of Signal A must be relative to Signal B in order to report A. In other words, the observer sets a decision criterion, or a response threshold, along a probabilistic decision variable. The immediate import of this finding was to undermine the venerable concept of a physiologically determined sensory threshold. The first studies showed also that an analytical method of detection theory, called the relative operating characteristic (ROC), can isolate the effect of the placement of the decision criterion, which may be variable and idiosyncratic, so that a pure measure of intrinsic discrimination acuity is obtained. The model and methods were then used in other areas of psychology in which discrimination is central, including recognition memory, conceptual judgment, and animal learning. This book attends to both themes, ROC analysis in the psychology laboratory and in practical diagnostic settings, and to their essential unity, by presenting 12 selected articles. The value of assembling them was suggested when one article in preparation in the mid 1980s turned into four separate ones. Together, those four articles give a unified, current statement of the concepts just mentioned: one shows that existing data across the range of discrimination tasks produce ROCs of a certain form; the second develops the implications of that form for theory and measures; the third treats issues in applying the methods and measures that arise in diagnostic settings; and the fourth discusses the science of setting an appropriate decision criterion. These four articles appear, however, in three different journals. Moreover, the seven papers chosen here to illustrate diagnostic applications were published in six diverse journals and a technical report. (From the preface) Contents: Part I: Theory, Data, and Measures. 1. The relative operating characteristic in psychology; 2. Form of empirical ROCs in discrimination and diagnostic tasks; 3. Indices of discrimination or diagnostic accurracy. Part II: Accuracy and efficacy of diagnoses. 4. Measuring the accuracy of diagnostic systems; 5. Choosing the right decision threshold in high-stakes diagnostics. Part III: Applications in various diagnostic fields. 6. Medical imaging techniques: A review 7. Medical imaging techniques: An illustrative study; 8. Enhancing and evaluating diagnostic accuracy; 9. Information retrieval methods; 10. Predictive validities of amplitude tests; 11. Accuracy and response bias in survey research; 12. System operator response to warnings of danger. Appendix: Computer programs for fitting ROCs. Author index. Subject index.

展开