Memory Limitations in Human and Animal Signal Detection
Sheila Chase Hunter College, City University of New York
Eric G. Heinemann Brooklyn College, City University of New York
In previous volumes of this series, we described a general theory of memory and decision making that accounts for phenomena as diverse as probability learning, categorization, concept formation, and pattern recognition ( Chase, 1983; Heinemann, 1983a; Heinemann & Chase, 1989). In this chapter we shall discuss how this theory is related to signal-detection theory (SDT), and attempt to account for some of the differences between the behavior predicted by SDT and that actually observed. (The problems discussed in this chapter concern "recognition" rather than "detection" but to facilitate communication we shall use the well-known terminology of SDT.) We shall limit our discussion here to decisions involving stimuli that induce sensations which vary along a single intensitive dimension, such as loudness or brightness.
It is now clear that neither humans nor pigeons behave quite as described by SDT in the form presented by Swets, Tanner, and Birdsall ( 1961). For example, the basic measure of sensitivity, d', is affected by factors not treated by the theory. Four such factors will be considered here: (1) the occurrence of responses that are independent of stimulus value, guessing, (2) improvements in sensitivity that occur during discrimination training ( Chase, Bugnacki, Braida, & Durlach, 1983; Heinemann & Avin, 1973; Swets & Sewall, 1963), (3) the dependence of d' for any particular pair of stimuli upon the range of stimulus values presented during the experiment ( Braida & Durlach, 1972; Chase, 1983), (4) the dependence of d' on the position of the stimulus pair within the range ( Braida & Durlach, 1972; Weber, Green, & Luce, 1977).
According to our theory, a relatively small amount of information is