After reading this article you will learn about the signal detection theory.
Experiments in psychophysics were primarily concerned with the establishment of absolute and differential thresholds, very soon it was realised that such attempts were not going to be very successful, because one’s ability to become aware of and differentiate, between stimuli, in larger life situations is far different from the controlled laboratory experimental situations.
In view of this, it was felt that psychophysical experiments, instead of trying to establish universal and constant values of thresholds, should design and carry out experiments, that would help in studying people’s ability at detecting signals and differentiating them, in life-like situations, taking into account the fact that in such situations other neutral and extraneous stimuli also operate. This resulted in a series of experiments on what is known as signal detection and the development of signal detection theory.
In brief, this theory attempts to understand the factors that determine whether an individual experiences and reports, or does not experience and reports so, in case of stimulations which are around the threshold values. In other words the theory is primarily concerned with what happens in the case of “near threshold” stimulations, and what determines the probability of their being received.
Stated in the form of mathematical models such experiments enable psychologists to compare the abilities of different persons to detect sensory stimulation, even when the circumstances they are in, and the manners of presentation of the stimuli differentiating them respond differently.
Experiments in the area of signal detection employ certain special methods or procedures. This involves, some trials on the one hand where the experimenter presents the particular stimuli, and other trials where the particular stimulus is not present.
The latter are known, as “catch” trials, because they are primarily directed at understanding the response biases or the tendency of the individual, to respond even when the actual stimulus is not there. This would give us an idea of how such response biases in the persons affect his responses, on the real trials.
In as far as studying the expectancy of the subject goes, the same thing is assessed by varying the ratio of the occurrence of the stimulus in the set of trials whether the stimulus actually occurs, like every 5th, every 10th, etc.
Employing such designs, the experiments are conducted. Whenever a stimulus is correctly reported, it is scored as a Hit and whenever it occurs and the subject fails to respond it is called a Miss. Whenever no stimulus is given, and the subject reports a stimulus it is called a False Alarm.
Whenever no stimulus is presented, and the subject responds correct, it is called a Correct Rejection. On the basis of a number of experiments, and also the subjects response pattern, a pattern of distribution of the subjects responses has been arrived at known as the ROC or Receiver Operating Curve. The shape of the curves shows a typical pattern and the subjects sensitivity is measured, in terms of a trial.
Some of the general inferences that have been drawn from studies are:
(a) If the percentage of trials with the signal or correct stimulus increases, there is an increase in the Hits. Thus if the percentage of signal trials increases, the individuals ability to discriminate between the real trials and false alarms increases where the person is more sensitive.
(b) On the other hand, if the person is not so sensitive the increase in hits is almost the same as the increase in the False Alarm trials.
(c) The sensitivity of the individual also increases as a function of the intensity of the stimulus.
Receiver Operating Curves (ROC) thus give us an indication of the overall sensitivity of the individual and for each individual one can arrive at an ROC.
Signal detection theory and ROC, in addition to helping psychologists to assess the roles of sensitivity and response bias, can also help in analysing the cause of failure to detect important signals. The role of response biases and circumstantial factors can be analysed and suitable steps can be taken including training of the subject.
It has also been suggested that they may be applied as tests of memory and recall, as for example in a courtroom where witnesses have to identify persons. But while psychologists may suggest many possibilities, still one has to go a very long way in applying these findings to sensitive and serious situations like courtroom identifications.
Steps can be taken to enhance the efficiency in detection of signals on the part of those in high security jobs like Radar observation, aircraft signal observers and many others.
Some of the recommendations of a general nature that have been made are as follows:
(1) For those whose jobs involve detection of signals like those who drive for a long time or command an aircraft, or in charge of security operations, work has to be in short spells, with frequent breaks to avoid the adverse effects of fatigue and boredom.
(2) Those on monotonous jobs like assembly line jobs, are advised to take periodic diversions like physical exercise, listening to music, or some other forms of recreations.
(3) Signal detection jobs have to be paced and spaced appropriately. So also the complexity of the signal detection may be appropriately fixed, to keep people alert and comfortable, and not be burdened.
(4) Feedback about Misses or Hits is used to improve the sensitivity.
(5) Signal detection experiments are finding increasing application, in many spheres.