General Prediction Relating Yes‐No and Forced‐Choice Results

One reason for developing detailed models of the detection or discrimination process is to provide a means of predicting performance in a variety of test situations, thereby permitting a psychoacoustic theory that is not bound to any single psychophysical method. Two general methods are often used to study discrimination or detection processes. In the Yes‐No task, a single observation interval occurs; the signal is either presented or not and the subject says that Yes—he hears the signal or No—he does not. In the forced‐choice task, several observation intervals occur; the signal is presented in only one of them and the subject indicates the interval in which he believes the signal has occurred. It can be demonstrated that a number of psychophysical models (e.g., threshold, detection theory) all agree in predicting the same relation between the accuracy of the judgments in the forced‐choice task and the area under the Yes‐No ROC curve. (The ROC curve is a plot of false‐alarm rate versus hit rate for diffe...