DETECTION THEORY

An explanation of the way signals are perceived against a background of noise.The psychological study of signal detection is an extension of work performed by engineers on the detection of targets by radar in World War II. Investigators (Smith and Wilson, 1953; Tanner and Swets, 1954) found that any signal from the environment is superimposed on both internal and external noise and must be distinguished from this noise. Internal or “neural” noise is produced by the workings of the individual’s body—for example, breathing sounds or sounds produced by changes in posture. It has even been found that some subjects become aware of the sound of their own blood circulating when they are in a soundproofed room. External noise, on the other hand, is any sound from the environment that interferes with the reception of a signal. Even in a soundproof room the subject may hear a high-pitched noise, the so-called “noise of silence.” When a signal is emitted, the subject must detect it against the background of both types of noise.Research has shown that the individual does not merely receive the signal, but makes a decision as to its presence. When the signal is extremely faint or when there is considerable interference, or both, this decision is not based on absolute knowledge but on probabilities (Carterette and Cole,1962),Each individual sets up a criterion value against which he checks the sensory input. However, different subjects have been found to differ in the criteria they use in spite of the fact that they are presumably receiving the same stimuli. A cautious person will only claim to detect a signal when it is strong or when any background noise is faint. Personality factors therefore enter into this decision.Experiments show that the decision is also affected by motivational factors. In one investigation, one group of subjects was given a reward of ten cents for guessing correctly that the signal had been presented and were penalized five cents for guessing incorrectly, while another group was rewarded and penalized only one cent. The first group was found to venture positive guesses more often than the second because they had a chance of winning more money. In addition, it was found that when the subject is told in advance that the signal will be presented nine out of ten times, he will practically always guess “signal presented” in uncertain cases, since the schedule favors this strategy. He also does better when he is told he is right or wrong.These observations have shed new light on the problems of detection. In the early days of psychology it was assumed that a subject is a “passive receptor” who simply reacts to stimuliin the same way that a galvanometer reacts to changes in voltage. The new findings, however, suggest that the process is not nearly so automatic and mechanical. Even in a situation as simple as the detection of signals there is an element of strategy. Detection is a “game” which the individual tries to win. It is therefore more than a sensory experience, since a person uses all relevant information in making his decision and tries to adjust his behavior to maximize his returns. In doing so, he compares signal-plus-noise to noise value alone, and does so within a personality and motivational matrix. Like a computer, he integrates his data and arrives at a decision on this basis. In a word, the process is far more complex than was recognized fifty years ago.