The use of rating‐scale isosensitivity data for testing discrete, two‐state detection models is admitted to be inappropriate, unless additional assumptions are made. It is suggested that simple assumptions may, in fact, be made by which this model may maintain its basic description of sensory processes and yet make predictions about rating results. The predictions are shown to produce the same isosensitivity curves (ROC curves) as did the original, two‐choice version of the model. These functions do not describe the results of rating experiments.