Detection and Discrimination
of Compound Stimuli: Tools for
Multidimensional Detection Theory
In Flatland, Abbott's (1991/1884) classic mathematical fantasy, a two-dimensional world is visited by someone from the third dimension who shows an eager acolyte the splendors of 3D. So far we have described a one-dimensional psychological world that even flatlanders would disdain: Sensation, familiarity, and other such dimensions have been the single subjective variables involved. For the initial applications of detection theory to auditory and visual detection, the idea that a single variable—subjective intensity— characterized the decision process was quite reasonable. We saw in chapter 5 that some apparently more complex problems such as false memory and social judgment can be interpreted unidimensionally as well.
The problems we consider in this chapter are the detection and discrimination of “compound” stimuli, that is, those with two or more perceptually distinct components. The key questions are whether these “cues” are combined by the observer and, if so, in what way. Treisman (1998) offers some nonlaboratory examples: To decide whether there is an aircraft in the sky (a detection task) or whether the aircraft is a plane or a helicopter (discrimination), one may rely on visual appearance or the quality of sound it produces. In assessing the degree of impairment of a particular patient, a clinician's judgment may be based both on a deficiency in movement control and signs of disordered thought. The question may be whether impairment exists (detection) or what type of impairment it is (discrimination). Cues may be in conflict or in agreement, and how they are best combined is a complex problem. Should the nature of combination change, in the plane-spotting example, if clouds limit the view or traffic noise masks the auditory signal?