An apparatus designed to detect emotional changes that indicate guilt.Attempts to detect guilt on the basis of physical and psychological signs appear to be as old as history. The ancient Hindus noted that guilty persons refuse to answer questions, give evasive answers, speak nonsense, shiver, and turn red. The Chinese forced suspects to fill their mouths with dry rice; if they could not swallow it, they were pronounced guilty. Current procedures follow the principle of measuring physiological changes that occur under emotional stress particularly autonomic changes, since they are not under voluntary control.The most widely used apparatus is the polygraph {Fig. 34), which records changes in respiration (through a chest lie detector).Fig. 34. Portion of a polygraph recording from a lie detector test. The sudden fluctuations in respiration (top line), galvanic skin response (middle line), and pulse and blood pressure (lower line) indicate that the subject lied in answering question 51.In giving a lie detector test, a trained operator attaches the apparatus to the suspect’s body, then asks a long series of questions, some of which are “neutral” (Did you have breakfast this morning ) and some of which are “critical,” since they indicate guilty knowledge (Have you recently had an ax in your hands ). Instead of questions, some operators use a free association test composed of critical words interspersed among neutral words. Sometimes, too, a third type of question or word is also used, involving emotional responses that have nothing to do with the problem (Do you have a girl friend ). Here the object is to establish a base line for the subject’s reactions, since there are wide individual differences even among innocent persons in emotional response to questioning.After the test, a comparison is made of the responses to the critical, neutral, and emotional stimuli, and a judgment of lying or truth-telling is made. Various surveys indicate that experienced examiners make correct judgments in about 80 per cent of cases, though some have claimed an accuracy of over 90 per cent. Because of the nature of the technique, the errors are usually failures to detect guilt, and it is claimed that innocent persons have rarely if ever been judged guilty by the lie detector.Today lie detectors are used in a variety of settings—not only police departments and courts, but governmental agencies, such as the Atomic Energy Commission, the military services, banks, retail stores, hotels, and industrial plants, both in personnel selection and periodic examination of employees. In police work the test is employed chiefly in the preliminary examination and screening of suspects, as well as in persuading criminals to confess. Only a few courts admit lie detection data as evidence, and then only when both parties agree to submit to a test administered by a mutually acceptable examiner. In these cases, the findings are presented by the examiner as expert testimony to be considered along with other evidence by the judge or jury.A professional organization, the Academy for Scientific Interrogation, has been established for improvement of lie detection and other interrogation practices, and training programs for practitioners are now being offered by universities, police departments, and the military services.Many experiments have been conducted on lie detection, typically with simulated crimes. Kubis (1962) set up a situation in which some students stole money from a coin box attached to a pamphlet rack, while others served as lookouts or innocent suspects. Analysis of the polygraph records by various examiners yielded correct identifications ranging from 73 to 92 per cent, far above the chance expectancy of 33 per cent. Other studies indicate that multiple measures, as on the polygraph, are preferable to single measures, since one subject might show his greatest response to stress through increased sweating, another through blood pressure changes, and a third through respiratory reactions. Lykken (1960), however, has performed an impressive experiment in which only the GSR was used. He first allowed his subjects to practice producing GSRs, and then offered them ten dollars if they could avoid detection. The test consisted of multiple choice questions on such facts as the name of the high school they attended and their mother’s first name. The galvanometer tracings indicated with 100 per cent accuracy whether or not they had chosen the correct alternatives.In spite of these positive results, it has been found that the test cannot be counted on with certain types of individuals—particularly hardened criminals and psychopathic (antisocial) personalities who do not feel guilt or react with fear, as well as retarded individuals who do not appreciate the significance of the situation, highly emotional persons who overreact to nearly all questions, and the false confessor who for unconscious reasons confesses to crimes he did not commit. Moreover, although naive subjects find it difficult to impossible to “beat the test,” it may be possible for some persons to train themselves to control their physiological responses well enough to mislead the examiner. In one case, however, a suspect controlled his breathing so completely that the apparatus did not record any of the normal variations that occur under intensive questioning. When the examiner confronted him with this fact and accused him of trying to cover up, he confessed.