VERBAL AND NONVERBAL CLUES - T0 TRUTH AND DECEPTION DURING POLYGRAPH EXAMINATIONS Thesis for the Degree of M. S. MICHIGAN STATE UNIVERSITY FRANK S. HORVATH 1972 i III III III I I III III III I b LIBRARY Michigan Staic I University . VERBAL AND NONVERBAL CLUES TO TRUTH AND DECEPTION DURING POLYGRAPH EXAMINATIONS By Frank S. Horvath AN ABSTRACT OF A THESIS Submitted to The College of Social Science Michigan State University In Partial Fulfillment of the Requirements for the Degree MASTER OF SCIENCE School of Criminal Justice 1972 e I, I ..? u Approved: ;_,\x.¢\/\ Lk.j12..('v x... Dr. Leon weaver, Chairman fl/{L/ (7) 7fly gin/magi PrdfessorTRobert‘C—_T?ojanowieb RAF-(’4' 0 p and.“ Professor Ralph E. Turner ABSTRACT VERBAL AND NONVERBAL CLUES TO TRUTH AND DECEPTION DURING POLYGRAPH EXAMINATIONS by Frank S. Horvath The behavior of one hundred polygraph subjects, fifty known truthful and fifty known lying, was analyzed to de- termine whether their behavior provided clues to the subject's actual truthfulness and deception. The behavior studied was broken down into three categories: elicited verbal behavior, spontaneous verbal behavior, and non- verbal behavior. Elicited verbal behavior was defined as the subject's verbal answers to each of fourteen structured pre—test interview questions. A significant number of the truthful subjects answered ten of these questions with answers, which, according to accepted theory were answers typically given by truth-tellers; a significant number of the lying subjects gave answers typical of liars to six of the questions. Only three of the interview questions were found to be significant discriminators between the truthful and lying subjects. Spontaneous verbal behavior was defined as those com- ments, requests, and complaints which a subject makes during a polygraph examination without direct prompting from the Frank S. Horvath examiner. The hypothesis that lying subjects exhibit spontaneous verbal clues more often than truthful subjects was found to be correct. Nonverbal behavior was defined as the subject's de- meanor and appearance during the examination. It was hypothesized that the subjects would exhibit nonverbal be- havior, which according to prior research was more typical of their actual classification (truthful and lying) more often than nonverbal behavior which was, according to prior research, not typical of their actual classification. The hypothesis was found to be correct; a significant number of the truthful subjects more often appeared as "truth- tellers" than "liars"; whereas, a significant number of the lying subjects more often appeared as "liars" than as "truth-tellers." VERBAL AND NONVERBAL CLUES TO TRUTH AND DECEPTION DURING POLYGRAPH EXAMINATIONS A Thesis Presented to the Faculty of the School of Criminal Justice Michigan State University In Partial Fulfillment of the Requirements for the Degree Raster of Science by (:) Frank S. Horvath 1972 DEDICATION To My Parents ii ACKNOWLEDGEMENTS The writer is indebted to John E. Reid, Director, John E. Reid and Associates for his cooperation in this research project. Also, the writer gratefully acknowledges the assistance of the following persons: James and Sally Hennessy for their advice regarding the statistical test- ing of the data; Dr. John H. McNamara for his advice and assistance concerning the methodology used in the study; Mrs. Sheryl Ten Broeke for her painstaking efforts in typing the many manuscripts; my wife, Jan, for her patience, advice, and assistance throughout the many months of the preparation of this study; David Leflet for providing in- formation concerning non-polygraph interrogation; and, finally, Dr. Leon Weaver, Dr. Robert Trojanowicz, and Professor Ralph F. Turner for their critical review and advice concerning the final report of the study. iii TABLE OF CONTENTS PAGE LIST OF TABLES ................................. IX CHAPTER 1 INTRODUCTION .......................... The Problem ......................... Statement of Purpose ................ Need for the Study .................. Definitions .00....OOOOOOOOOOOOOOO... -p ¢> xx n) e4 I4 Examiner 00......OOOOOOOOOOOOOOOOOO Interrogation, Interrogational DeVice 00.0.0000...OOOOOOOOOOO... Interview ......................... Interview Sheet ................... Lying Subject ..................... Polygraph, Polygraph Instrument ... Polygraph Examination ............. Polygraph Records ................. Polygraph Subject Behavior ........ Structured Pre-Test Interview ..... Subject ........................... Truthful Subject .................. Coqmoxoxmmmm-b-b-P. 2 REVIEW OF THE LITERATURE .............. Behavioral Characteristics of Polygraph Subjects ................ 8 iv CHAPTER PAGE Behavioral Characteristics of Subjects in Non-Polygraph Interrogation ...... 12 Sweating ............................ 15 Dry Mouth ........................... l5 Clenching of Hands .................. 13 Respiration ......................... 15 General Nervousness ................. 14 Color Change ........................ 14 Heart beat, pulse ................... l4 Elbow Position ...................... l5 Fidgeting ................ ...... ..... 15 Facial Tics ......................... l5 Adam's Apple Activity ............... l6 Behavioral Characteristics Used for Detecting Deception ................. 17 Association and Reaction Time ....... 17 Muscular Movements .................. 19 Voice Quality ....................... 19 Eye Movements ....................... 21 Facial Expression and Body movement ...OOOOOOOOOOOOOOOOOOOO... 22 Nonverbal Behavior As a Communication System ................ 25 Detecting Body Movement and Facial Expression Significance ........... 26 CHAPTER PAGE GSR Responsiveness Related to Emotional Expression .............. 28 Classifying Behavioral Communication 31 Paralanguage...................... 31 Proxemics ......................... 32 Kinesics .......................... 32 Summary ............................... 33 3 THEORETICAL BASE OF DATA STUDIED ........ 35 Elicited Verbal Behavior: The Structured Pre-Test Interview ....... 35 You Question ........................ 38 Suspicion Question .................. 39 Who Not Suspect Question ............ 39 Think Stolen Question ............... 40 Best Opportunity Question ........... 41 Fingerprint Question ................ 41 Borrow Question ..................... 43 Ever Happen Before Question ......... 44 Think Question ...................... 45 Approach Question ................... 45 How Will L.D. Come Out on You queStion 0.000.000.0000...00.0.0... 46 Truth Serum Question ................ 47 What Should Happen to Doer Question .................. ..... ... 48 Attitude Question ................... 49 vi CHAPTER PAGE Spontaneous Verbal Behavior .......... 51 Nonverbal Behavior ................... 52 4 DESIGN OF‘THE INVESTIGATION ............ 54 Background Information of Data Studied ............................ 54 Source of Data ..................... 54 Subjects Studied ................... 57 The Original Investigations ........ 59 The Polygraph Examiners ............ 60 Method of Study ...................... 61 Elicited Verbal Clues .............. 61 Spontaneous Verbal Clues ........... 64 Nonverbal Clues .................... 66 Limitations of the Study ............. 7O 5 RESULTS OF THE STUDY ................... ‘ 74 Elicited Verbal Clue Results ......... 74 Spontaneous Verbal Clue Results ...... 78 Nonverbal Clue Results ............... 80 6 DISCUSSION ............................. 82 7 SUMMARY ................................ 9O BIBLIOGRAPHY .0............OOOOOOOOOOOOOOO0...... 102 vii CHAPTER PAGE APqu-DICES .................O.................... 11'1" Appendix A (Section I) Master Data Sheets: Elicited Verbal Clues .. 115 Appendix A (Section II) Selected Actual Answers to the Structured Pre-Test Interview Questions ................ 118 Appendix B Master Data Sheets: Spontaneous Verbal Clues 0............COOOCOOOOO......OOOOOOOOOO 1:; Appendix C Master Data Sheets: Nonverbal Clues ........ 126 Appendix D Discussion of Additional Spontaneous Verbal Clues .....OOOOOOOOOOOOOOOOOOIOOOOOOOOOOOO... 129 viii TABLE 10 LIST OF TABLES Comparison of Background Information of Polygraph Subjects ............... Comparison of the Original Investi- gation Issues ....................... Elicited Verbal Glue Results: Truthful SUbjeCtS 00000000000000.0000 Elicited Verbal Clue Results: Lying SubjeCts ...OOOOOOOOOOCOOOOOOOO A Comparison of Truthful and Lying Subjects on Three Types of Spontaneous Verbal Clues ............ Grouped Spontaneous Verbal Clue Results Comparison of Truthful and Lying Subjects On "Typical Truth-Teller" Nonverbal Clue Descriptors OOOOOOOOOOOOOOOOOOOOO Comparison of Truthful and Lying Subjects On "Typical Liar" Nonverbal Clue Descriptors 00.000000000000000...0.... Results of Classification of Subjects on Nonverbal Clues ................... Number of Subjects Voicing Concern for Either Control or Relevant Test QUEStiOnS 000000.000. ooooo 000000000000 ix PAGE 59 6O 75 78 81 83 85 86 88 131 Chapter 1 INTRODUCTION THE PROBLEM Polygraph examiners frequently disagree about whether or not judgments concerning a polygraph subject's behavior should be used in making truth or deception decisions. On the one hand, some examiners insist that behavior judgments should not be used at all. They insist that behavior is too evanescent and impalpable to be objectively studied. These examiners wish to replace behavior judgments with a total reliance on physiological concepts, which they feel to be more substantial. They maintain that polygraph re- cord analysis, unsupported by behavior judgments, is "pure lie detection." Thus, the proponents of this method liken the polygraph examination to a mathematical procedure where- in any subjectivity is eliminated.1 ISee for example: U.S., Congress, House, Committee on Government Operations, Foreign Operations and Government In- formation Subcommittee, Use of Pol aphs as "Lie Detectors" B the Federal Government, Hearing, 89th Cengress, 2n5 Sess., AprII 7, 8, and 9, I963, Part 1, Panel Discussion with Pri- vate Polygraph Practitioners (Washington: Government Print- ing Office, 1964), esp. pp. 33 and 91, testimony of Cleve Backster; and, see: Cleve Backster, "Lie Detection Comes of Age," Law and Order, date and page unknown, reprint supplied to the erTer by Mr. Backster, January, 1970. On the other hand, some examiners maintain that while judgments of a polygraph subject's behavior should not be used to make truth and deception decisions, they should be used to reinforce polygraph record analysis.2 These exam— iners have intuitively developed techniques to elicit and observe behavior and to relate what is observed to the record analysis. The combined use of record analysis and behavior judgments is said to make the polygraph examination a diagnostic technique in which all factors of the subject and inquiry at hand are studied in terms of truth and decep- tion. This technique is likened to medical or psychiatric examinations, wherein both subjective and objective means of evaluating physical and mental well being are used. STATEMENT OF PURPOSE The purpose of this study is to determine whether the verbal and nonverbal behavior of polygraph subjects pro- vide valid clues to the subjects' truthfulness and decep- tion. In this study the behavior of known truthful sub- jects and known lying subjects will be analyzed; and, while verbal and nonverbal behavior will be considered separately, 2John E. Reid and Fred E. Inbau, Truth and Deception: The Polygraph ("Lie-Detector?) Technique (Baltimore: The WiIliams and Wilkins Co., 1966), pp. 11-16, 152. \Jd verbal behavior will be further broken down into that which is either elicited or spontaneous. NEED FOR THE STUDY The only research on polygraph subject behavior was published in 1953.3 Consequently, there is a need to update that research and to relate it to new developments in polygraph technique. Furthermore, the past research did not specifically distinguish between verbal and nonverbal behavior or between elicited and spontaneous verbal be— havior. These distinctions are necessary for the further refinement of polygraph technique. The primary source of behavior data in polygraph examinations occurs during the structured pre-test inter- ”I- I) 0 view. This interview has en develOped and refined large- ly on an intuitive basis and, while it is claimed to be effective, it has never been statistically evaluated. The need for such an evaluation was pointed out by Jesse Orlan— ski in his study on "lie detection" for the Institute of. A [I- "1- n c u . L. _ l '_‘. ' J- . eiense Analyses. ihus, it is important that an evaluation Ci of behavior data be accompanied by a statistical evaluation of the polygraph interview itself. 7 )This research will be cited and discussed in the next chapter of this paper. 4Jesse Orlansky, An Assessment of Lie Detection Capa- bilit , Declassified Version, Technical Report 62-15’(Arling— ton, 6a.: Institute for Defense Analyses, Research and Engineering Support Division, July, 904), pp. 29-50. III I I’l‘l[‘t\[.[]’l’l‘ n'll DEFINITIONS Examiner The term examiner is used synonomously with polygraph examiner, the person administering a polygraph examination. Interrogation, Interrogational Device The term interrogation refers to any type of question- ing wherein the purpose of the questioning is to influence the questionee into answering truthfully. Such influence is exerted through the use of interrogational devices such as sympathy, empathy, flattery, etc. Interrogation is to be distinguished from interviewing. Interview The term interview refers to a period of questioning during which the interviewer maintains a neutral, objec- tive attitude toward the interviewee. Interrogational devices are not used; the interviewer neither attempts, nor desires, to influence the interviewee to answer his questions truthfully. Interview Sheet An interview sheet is the paper on which a polygraph examiner records data pertaining to a subject's verbal and nonverbal behavior, along with other information pertaining to the subject's examination. Lying Subject As used in this study the term lying subject refers to a person who admits subsequent to a polygraph examina- tion committing the offense investigated by that examina— tion. A lying subject did not tell the truth to those questions concerning the area of inquiry of the examina- tion. Polygraph, Polygraph Instrument A polygraph is a multi-channel diagnostic instrument which records in ink on a chart paper certain of a person's physiological activity. Usually the polygraph used in "lie detection" records, at least, a person's blood pressure— pulse rate, respiration rate and amplitude, and Galvanic Skin Response (GSR). Polygraph Examination A polygraph examination is the procedure employed by a polygraph examiner to determine the veracity of a person‘s statements to a specified area of inquiry. All polygraph examinations consist of at least a pre-test interview, a series of polygraph tests, and an interpretation and analysis of polygraph record data. Polygraph Records Polygraph records are the chart paper recordings made by the polygraph instrument of a subject's physiological activity. Polyggaph Subject Behavior Polygraph subject behavior is defined as the words, actions, and appearance of a person taking a polygraph examination. The person's words are either elicited verbal behavior (clues) or spontaneous verbal behavior. Elicited verbal behavior is that which a subject says in response to a direct question from the examiner; spontaneous verbal behavior is that which a subject says without prompting from the examiner. The person's actions and appearance, or demeanor, are defined as nonverbal behavior. Structured Pre-Test Interview A structured pre-test interview is a period prior to polygraph testing in which a subject is asked questions by the examiner from a standardized format. The purpose of these questions is primarily to elicit from a polygraph subject verbal and nonverbal behavior which the examiner observes and records. The terms interview and pre-test interview are used synonomously with structured pre-test interview in this study. Subject The term subject refers to the person taking a poly- graph examination. Truthful Subject As used in this study the term truthful subject refers to a person who neither committed nor aided in the commis- sion of the offense investigated by a polygraph examination. The truthful subject told the truth to those questions con- cerning the area of inquiry of the examination. Chapter 2 REVIEW OF THE LITERATURE Reviewed in this chapter will be the available literature pertaining to verbal and nonverbal behavior of polygraph subjects. However, because this literature is very limited, and in order to provide a complete as- sessment of the scope of this study, data pertaining to the use of behavior judgments in situations similar to the polygraph examination will also be presented. These latter data will be broken down into three categories: 1) observational comments concerning the use and value of behavior judgments in non-polygraph interrogation; 2) research which has been conducted in an attempt to detect deception by evaluating behavior in the absence of poly- graphic data; and 3), research which has been conducted in an attempt to assess the use of nonverbal behavior as a communication system, separate from a spoken language. BEHAVIORAL CHARACTERISTICS OF POLYGRAPH SUBJECTS The foremost advocate of the viewpoint that a judgment of a subject's behavior should be a standard part of all 8 polygraph examinations is John E. Reid. Through his ex- perience and expertise in the art of non-polygraph police interrogation Reid hypothesized that certain behavior pat- terns were common to lying persons while certain others were common to truthful persons. While this is an assump- tion made by almost all criminal interrogators and is fre- quently commented upon in interrogation manuals, Reid voiced the belief that the ability to accurately interpret behavior patterns was not, as many felt, a "sixth sense," but rather, a good use of the five senses. This led him to conclude that an objective study of the behavior of persons taking polygraph examinations might be useful in diagnosingiruth and deception. In 1953, Reid and Richard Arther reported the results of a five year study of 809 polygraph subjects.5 Although they cautioned that in the polygraph examination "no speci- fic type of behavior - even though it is highly typical of one [Iying7 or the other [truthful7 group - should ever be considered proof of guilt or innocence,"6 they did find that the conduct and statements of the lying and the truth- ful differed widely in some respects; in others they were quite similar. 5John E. Reid and Richard 0. Arther, "Behavior Symp- toms of Lie-Detector Subjects," The Journalflgfmppimipalmggy, Criminolo and Police Science, 44, 1 (May-June, 1953), pp. IOZ-IU8; hereafter this will be referred to as Reid in the text. 6Reid and Arther, "Behavior Symptoms," p. 104. 10 Of the 486 lying subjects Reid studied, none of them was anxious to take a polygraph examination; none had made the initial request. Many of the lying subjects either postponed their examination date, were late for their appointment, or failed to appear for their original appoint- ment. Usually prior to their polygraph examinations these subjects tried to explain that they had a nervous, or other mental or physical defect which could cause "false" lying reactions on the polygraph. Many of them complained of physical pain or discomfort from the polygraph apparatus; generally, lying subjects would want to leave the examining room as soon as possible and would complain that the ex- amination was taking too long or that they were late for an appointment elsewhere. In contrast to the lying subjects Reid studied, many of the 323 truthful subjects were glad to take a polygraph examination to prove their innocence. They often requested the examination and voiced their confidence in both the polygraph instrument and the procedure. If they expressed a fear that their nervousness or physical abnormality might affect their results, they did not persist after the examiner assured them that their peculiar difficulty made ' little difference. Aside from studying the general statements of poly- graph subjects, Reid also analyzed the conduct of his subjects. 11 He found that lying subjects often look worried and high- ly nervous. Sometimes they act aggressively, express a bitter attitude, appear to be in a shocked condition, ex~ perience mental blocks, are evasive, have a "dry mouth," sigh and yawn frequently, fail to look the examiner in the eye, and move about. Sometimes, the lying subject is overly friendly and polite to the examiner. The general conduct of truthful subjects often dif- fers from that of liars. They are usually "at ease, light hearted, and talkative."7 They appear to be sincere and straightforward when talking to the examiner. If truth- ful subjects exhibit anger or impertinence, they can usually be relaxed by the examiner and even though they are cooperative in this respect, they do not appear to be overly friendly or polite. Reid concluded his study with the finding that be— havior judgments provide a definite advantage. Primarily, they allow the examiner to assess the subject's attitude and to treat him accordingly. The highly nervous subject must be relaxed; the quiet subject must be reassured, etc. Also, behavior judgments did seem to provide some basis for discriminating between lying and truthful persons, although as Reid pointed out, the advantage in this respect could 7Reid and Arther, "Behavior Symptoms," p. 106. 12 not be definitely stated, "since practically all behavior symptoms are subject to general rule exceptions."8 BEHAVIORAL CHARACTERISTICS OF SUBJECTS - oi'idefi-jij‘jfir‘:iighbfe‘mfofi_' ‘ Reid's finding that some behavior patterns differ between the lying and the truthful in polygraph examina— tions was precipitated by his observations of persons under- going non-polygraph police interrogation. Like Reid, other non-polygraph interrogators have also observed and commented upon these same and other behavioral characteristics. Un— like Reid, however, none of these other interrogators have offered any objective proof of the reliability of their observations. Even so, it is believed that these observa- tions, at least those supplementing what has previously been reviewed, still warrant comment here. They are con- sistently reported on in interrogation manuals written by different authors, and, in that sense at least, have some reliability and some bearing upon the general topic of this stuy. William Dienstein has suggested that the following behavioral characteristics provide helpful, but not com- pletely accurate indications of a person's deception: 8Reid and Arther, "Behavior Symptoms," p. 107. 13 Sweating "Sweating with ruddy or flushed face indicates anger, embarrassment, or extreme nervousness. Sweating with pallid face, 'cold sweat' may indicate shock or fear. Sweating 0 hands may indicate tension."’ Dry Mouth "Continual swallowing by the subject, licking of the lips, and drinking excessive amounts of water are indica- tive of tension."10 Clenching of Hands "The clenching of the hands indicates tension and may be coupled with anger. The wringing of the hands, rub- bing against the clothing, twisting and knotting handker- chiefs, or manipulating objects are indicators of tension.‘11 Respiration "Controlled breathing indicates a critical question. It may be betrayed by a slight gasp, a sudden intake of breath, a holding of the breath or a sudden expulsion of the breath."12 9William Dienstein, Egghpigues for thg Crime Investi- gator (Springfield: Charles C. Thomas, 1952), p. 109. loDienstein, Techniques, p. 109. llDienstein, Techniques, p. 109. 12Dienstein, Techniques, p. 109. 14 General Nervousness "Constantly moving about in the chair, pulling of the ears, rubbing of the face, picking or tweaking the nose, crossing and uncrossing the knees or legs, shifting the position of the feet (there is a tendency to brace the feet on critical questions), rubbing the hair, eyes, or eyebrows, biting or snapping of fingernails, all may indi- cate tension."13 In addition to those characteristics cited by Dien- stein, other interrogators - Reid, Inbau, Aubry, Arther, Caputo, and O'Hara - have reported that the following characteristics may be indicative of lying: Color Change "A flushed face indicates anger, shame or embarrass- ment but not necessarily guilt. A pale face is a more reliable sign of guilt."14 Heart beat, pulse "An increase in the rate of heart beat is indicative of deception. The pulse beat is observable at times in the veins of the neck."15 lZDienstein, Techniques, p. 110. 14Charles C. O'Hara, Fundamentals of Criminal Investi- gation (Springfield: Charles C. Thomas, 1956), p. 111. 15O'Hara, Fundamentals, pp. 111-112. l5 Elbow Position “Due to the large nervous area under the upper arms, some interrogators attach considerable importance to the position of the elbows, claiming that when the elbows are hanging in a relaxed position, the suspect is under no pressure, but when the elbows are brought in close to the sides, it is a strong indication that the questioning has become critical."l6 Fidgeting "Involuntary movements consist of jerking and twitching of the hands, face, legs, arms, feet and in rare instances, the general involvement of the whole body."17 "Shifting of movements and crossing of legs most often occur right after the suspect is asked an important question about the Facial tics, (grossly exaggerated muscle spasms of various parts of the face) are said to be positive indica- jzh. R. Kidd, Police Interrogation (New York: The Police Journal, 19407, pp. 94-95. 7—— 17Arthur Aubry and Rudolph Caputo, Criminal Interroga- tion (Springfield: Charles C. Thomas, 1965), p.‘130. Richard Arther and Rudolph Caputo, Interrogation 18 for Investi ators (New York: William C. Copp and Asso- ciates, I959), p. 92. 16 tions of lying. "These tics take the form of strange grimaces of all the facial muscles in general; sometimes these tics may extend to and involve the arms, legs, and even the whole body will occasionally participate in charac- teristic muscular incoordination."19 Facial tics and bodily twitching, tremors, and spasms may also be of an hysterical nature, eSpecially those consisting of sudden, abrupt and rapid involuntary contractions of a muscle or a group of muscles - rapid blinking of the eyes, shrugging of the shoulders, twitching face, protrusion of the tongue, and vague motions of the hand and arms, legs and feet."20 Adam's Apple Activity "The fact that an acceleration of its (Adam's Apple or Epiglottis) up and down movement is experienced by many offenders when questioned - and particularly when first accused - is well recognized among experienced interroga- tors."21 lgAubry and Caputo, Criminal Interrogation, p. 131. 20Aubry and Caputo, Criminal Interrogation, p. 131. 21Fred Inbau and John Reid, Criminal Interrogation and Confessions (Baltimore: The Williams and Wilkins 50., I962), p. 30. l7 BEHAVIORAL CHARACTERISTICS USED ""IERIDETEUTINGTEEEERERMT“" Reviewed next will be some of the research which has been conducted in an attempt to correlate deception with specific behavioral characteristics. Although much of this consists of work which may be considered outdated,it is nevertheless, relevant and should be noted here. In fact, some of this research provided the impetus for the de- velopment of the polygraph instrument and its application to the detection of deception. Association and Reaction Time Carl Jung was perhaps the first to attempt the detec- tion of deception with the use of associative reaction time.22 The basis of this attempt rests upon the theory that if a person B asked to associate freely to a number of words, some of which are critical to a given crime and some of which are non-critical, he will, if lying, give longer reaction times when answering to the critical words. A shortcoming of this technique is that to truthful per- sons critical words may be unfamiliar and may, therefore, automatically elicit a longer reaction time. 220. G. Jung, "On Psychophysical Relations of the Associative Experiment," Journal of Abnormal Psvcholo v, 1, (1907), pp. 247-255; also see: C. G. Jung, "The Association Method, American Journal of ngchology, 21, (1910), pp. 219-269. 18 Subsequent to Jung's studies, other investigators have researched the correlation between deception and reaction times to various stimuli. harston, in 1920, concluded that subjects with short reaction times and subjects with long reaction times were displaying deception responses when told to obey or disobey requests to perform arithmetic opera- 23 tions. More recently, Ellson, et al., conducted research on reaction time in answering questions about an artifi- cial crime. He concluded that reaction time significantly differentiated between truthful and deceptive responses.24 Although others have investigated reaction time as an indi- cator of deception, and some have quarreled with its value, generally researchers have found that reaction times will probably increase with deception involving a significant stimulus.25 29William M. harston, "Reaction Time Symptoms of Deception," Journal of Experimental Psychology, 5, (1920), pp. 72-77. 24D. G. Ellson, R. C. Davis, I. J. Saltzman, and C. J. Burke, A Report of Research on Detection of Decep- tion, (Prepared for Office of Naval Research, Contract hgonr-lBOll), Dept. of Psychol., Univ. of Ind., (1952). 2DSee also; H. B. English, "Reaction Time Symptoms of Deception," American Journal of Psychology, 57, (1926), pp. 428-429; Eva R. Goldstein, ifReaction Times and the Consciousness of Deception," American Journal of Psychology, 34, (1923), pp. 562-581; w. h. harston, "Negative Type Reaction-Time Symptoms of Deception," Psychological Review, 52, (1925), pp. 241-247. 19 Muscular Movements The correlation between deception and voluntary and involuntary movements of a person's hands when responding to word associations was investigated by Luria, a Russian Psychologist.26 He presumed that a change in the pattern of hand movements would occur as a person responded to a critical word association. Unfortunately, even though Luria reported considerable success with his technique, there does not seem to be any recent research to corro- borate his findings. John Reid does advocate the use of a muscular movement recorder in polygraph examinations; data are not available, however, which either proves or disproves the value of such a device although Reid does present some evidence that voluntary muscular movements may, in some cases, be directly related to deception.2] Voice Quality There are several studies which have explored the relationship between voice quality or modulation and de- ception. Two of these have attempted a subjective analysis 26A. R. Luria, The Nature of Human Conflicts: 0n Emotion, Conflict and"Will, trans. and ed. w.‘H. Gantt (New York: Liveright, 1932). 27John E. Reid and Fred E. Inbau, Truth and Deception The Polygraph ("Lie Detectgr")Technique (Baltimore: The WiIliam and Wilkifl Co., 1966), see esp. pp. 208, 213, 264, and 271. I). C) of deception. Fay and Middleton28 found that about 55 per- cent of the time, persons hearing other persons lie or tell the truth over a public address system could discriminate between the two types of statements. In a more recent study, 01echowski found that three separate groups of per- sons could not discriminate between liars and non—liars from hearing a tape-recorded interview. he presented the recordings of 20 liars and 20 non-liars to 21 experienced detectives, 145 psychology students, and 200 high school students. None of these groups was more successful in detecting the liars; also, he concluded it was not possible to establish a "lie symptomatic" 29 from the tape recordings An analysis of the correlation between voice modu- lation and deception was conducted by Alpert, Aurt/ 3berg and Friedhoff.30 They measured the amplitude of aperson's verbal response with a voltmeter during a deception situa- tion. Amplitude increases in a low frequency band to critical stimuli were found to be significantly different 28P. J. Fay and w. c. Middleton, "The Ability to Judge Truth-Telling or Lying from the Voice as Transmitted Over a Public Address System," Journal of General Psychology, 24, (1941), pp. 211- 215. 29Richard Olechowski "H‘Eperiments Ober den Stim_mund Sprechausdruck bein Lugen," AEXperiment on Voice hodulation While Telling Lies7 Zeitschrift fur Erperimentelle und Angewandte Psycholosie, 14, (1967), pg). 474- 482; from PSYEhological Abstracts, 422, (Aalil,1968), p. 484. 7 ’On. Alpert, R. L. Kurtzbclg and A. J. Friedhoff, "Transient Voice Changes Associated with Emotional Stimuli,” Archives of General Psychiatry, 8, (1963), pp. 262-565. 21 from increases to non-critical stimuli. Eye Movements Polygraph examiners who advocate using judgments of behavioral characteristics believe that observing a person's eye contact reveals much about whether or not a lie is be- ing told. This idea is apparently based upon the common assumption that a liar cannot look another person in the eye. For example, it is not at all uncommon for a parent who suspects his child of wrongdoing to command the child to "look him in the eye" and deny the wrongdoing. The child who cannot do this may be presumed to be lying. Some research data suggest that a correlation exists between eye movements and deception. Berrien conducted a study in which eye movements were evaluated during a mock crime situation.51 Although he detected no difference in the amount of eye movement during critical and non- critical questions, he did find that during a pre-question period, the eye tremors of lying persons decreased; eye tremors of truthful persons did not. Furthermore, he found that evaluators of the eye tremors could discriminate between the lying and truthful persons about 70 percent of the time. 5IF. K. Berrien, "Ocular Stability in Deception," Journal of Applied Psychology, 26, (1942), pp. 55-65. 22 Ellson, et al., tried to determine if lying produces characteristic patterns of eye movements which can be used as indicators of deception..:52 For his study he assigned students to one of five groups, each of which was to view a quadrant box. Two of these groups were to take a coin from a quadrant of the box and later to deny they took it; two were to take a coin and admit they took it; one was simply to view the box when no coin was present. After- wards, all groups were tested to determine if eye move- ments of the liars could be differentiated from the truth- tellers. In all cases, eye movements to the quadrant from which the coin had been taken were greater than movements to other quadrants. Also, the difference in this respect was greater for those persons who were instructed to lie about taking the coin than for those who told the truth. Facial Expression and Body Movement Paul Ekman and Wallace Friesen have conducted exten- sive research on the correlation between facial expression and body movement and deception.99 Their work is primarily oriented toward improving psychotherapy and psychotherapuetic ‘7 ,,,. ~ v. - )2Ellson, Report of Research, Chapter A, pp. 25-96. 7 ’BPaul Ekman and Wallace V. Friesen, "Lonverbal, Leakage and Clues to Deception," Psychiatry, 52, (1969), 23 techniques. However, because they believe that "actions speak louder than words" and that during deception, body movements and facial expression emerge as clues to the truth, their work is of interest here. Ekman and Friesen considered two forms of deception: alter deception, where ego, the deceiver, conceals informa- tion from the other interactant, alter; and self deception, where ego is the object of his own deception, concealing information from himself.34 Since during a polygraph examin- ation only alter deception seemsto be relevant, the follow- ing discussion will not consider Ekman's data on self de- ception. During alter deception a person (ego) is aware of what he is attempting to conceal and plans his behavior accordingly. He realizes that the safest way to avoid detection is to cut off all communication; he also realizes, however, that by doing this, he may be giving himself away, so he chooses a second alternative. He offers informa- tion (to alter) contrary to, or at least selectively misrepresentative of, the concealed information. In doing this, Ekman reports, a person engages in "nonverbal simula- tion" which is imperfect and as a result offers a major JEZEkman and Friesen, "Nonverbal Leakage," p. 89. 24 form of deception clues to an observer (alter), if he is capable of reading the clues. Ekman has found that the hands and feet of a person are more likely to offer deception clues to an observer than is the face, or facial expression alone. He points out, with regard to the hands, that even though a person may be smiling and pleasant, he may, at the same time, be "digging into his cheek, protectively holding his knees, and so forth."'55 In the same way the feet can also offer clues to deception. These would include "aggressive foot kicks, flirtatious leg displays, autocratic or soothing leg squeezing, abortive restless flight movements,...tense leg positions, frequent shift of leg posture, and rest- less or repetitive leg and foot acts."36 It is interesting to note that even though Ekman reports the hands, legs/feet to be the best source of de- ception clues, he feels that this is true only because these are the least within ego's awareness, and thus, are less subject to control. "The best sender, the face, is most closely watched by alter, most carefully monitored by ego, most subject to inhibition and dissimulation, and 35Ekman and Friesen, "Nonverbal Leakage," p. 99. 36Ekman and Friesen, "Nonverbal Leakage," p. 99. 25 thus the most confusing source of information during de- ception,"37 The face reportedly becomes a very good sender of deception clues if one considers "micro-expressions." These, Ekman points out, are more often than not over- looked by an observer; because of their short duration and the observer's limited experience in evaluating them, they are confusing, contradictory, and misleading. Thus, even though the face may actually leak the most it is at the same time equipped to lie the most. It is a deceptive source of deception clues.38 NOI‘TVERBAL BEHAVIOR AS A "TRflMMHIRfiEERflTTEEnflflr‘ Research on nonverbal behavior as communication will be discussed in this section. Much of the information considered first will have to do with the relationship between the emotional eXpression of stress (or other affec- tive states) and the ability to detect it. Then, general comments will be made regarding nonverbal behavioral characteristics as a communication system which is "in ac- cordance with an elaborate and secret code that is written nowhere, known by none, and understood by all."59 EiEkm 38 an and Friesen, "Nonverbal Leakage," p. 99. Ekman and Friesen, "Nonverbal Leakage," p. 99-100. 39Edward Sapir, "The Unconscious Patterning of Behavior In Society," Reprinted in: R. G. Mandebaum, Ed., Selected writin s of Edward Sa ir in Lan a e Culture and Personalit (Berkeley: University of California Press, I965); cited in: Starkey Duncan, "Nonverbal Communication," Psychological Bulletin, 72, (1969), p. 118. 26 Detectinngodnyovement and Facial Expression Eignificance In a study reported by Krout4O in 1955 an attempt was made to discover the significance of various bodily movements. He presented a list of 160 of these to 120 persons and asked them: (1) to give the first word which comes to mind and, (2) to write down the meaning the movement would have for them if they observed another person making it. Although Krout's detailed association results are too lengthy to report here, he concluded that responses are made up largely of popular preconceptions of what a movement is supposed to mean and that movement and appearance can be interpreted only in terms of their con- text. Krout's findings, and the findings of others doing similar research,led to the belief that emotion could not be judged from facial expression or bodily movement alone; it was suspected that it was all a matter of inference from know- ledge of the situation. This, however, has seemed to be proven untrue. 40M. H. Krout, "Autistic Gestures: An Experimental Study In Symbolic Movement," Psychological Monographs, 46, (1935); reported in H. Zimmer, PsychOphysiologic Variables as Indications of Emotional Stress, Technical report no. RADG-TR-65-296 (New York: Rome Kir Development Center, 1966), pp. 41-49. 27 There are many studies which have found a significant reliability in the ability of observers to judge certain emotions from the facial and bodily expressions of photo— graphs, etc., of people. Feleky41 found that 74 of 100 per- sons viewing a certain photograph of a person agreed that the emotion of surprise, wonder, or astonishment was depicted. On the basis of viewing another photograph they agreed 50 percent of the time that the emotion of disgust or repugnance was being shown. Munn42 found that college students could judge emotions from the facial expressions on persons in photographs with an accuracy range from 65 to 95 percent; further, his results showed that performance was not improved when the context in which the person was experiencing the emotion was added. Other studies have shown that the "accuracy with which emotions are recognized from facial expression alone increased when still photographs were replaced by short movies and that adding a sound track to silent movies of faceszmcreased accuracy in judging fear and surprise (though not other emotions), even though the voices were 41A. M. Feleky, "The Expression of the Emotions," Psychological Review, 21, (1914), pp. 55-44; in Irving Janis, et al; Personality, Dynamics, Development and As- sessment (New York: Harcourt, Brace and World, Inc., T959), p. 601. 42H. L. Munn, "The Effect of Knowledge of the Situa- tion Upon Judgment of Emotion from Facial Expressions " Journal of Abnormal and Social Psychology, 35, (1940) pp. 324-38; in Irving Janis, et a1; Personality, Dynamics, Develo ment and Assessment (New York: Harcourt, Brace and World, Inc., 1969), p. 603. 28 merely reciting irrelevant, neutral materials."45 GSR Responsiveness Related to Emotional Expression Harold Jones conducted two studies in an attempt to assess the relationship between galvanic skin response (GSR - a physiological phenomenon often used in both ap- plied and experimental "lie detection") and overt emotional expression. In his first study,44 1955, he used 85 pre- school children as subjects and simultaneously measured the intensity of GSR and the amount of overt emotional expression. He concluded that the average galvanic deflec- tion recorded for a given subject gave little or no pre- diction of his average overt response but that the average galvanic deflection recorded for a given stimulus agreed closely with the average overt response for that stimulus. In his second study,45 1950, Jones used adolescents as subjects. Their overt behavioral characteristics were compared with their GSR recordings in eleven test situa- tions. Jones found that the "'high reactives' were less 421rving Janis, G. Mahl, J. Kagan, and R. Holt, Personality, Dynamics, Development and Assessment (New York: Harcourt, Brace, and World, Inc., 1969), p. 605. 44Harold Jones, "The Galvanic Skin Reflex as Related to Overt Emotional Ex ression," American Journal of Psycho- logy, 47, (April, 1955 , pp. 241-251. 45Harold Jones, "The Study of Patterns of Emotional Expression." In M.L. Reynert (ed.), Feelings and Emotions: The Mooseheart Symposium (New York: McGraw Hill, 1950); cited in: Jack Block, "A Study of Affective Responsiveness in a Lie Detection Situation " Journal of Abornormal and Social Psychology, 55, (1957), pp. 11-15. 29 assertive, less animated, less talkative, and less atten- tion seeking than the 'low reactives,”46 and that the high reactive group was "more calm, more deliberative, more good-natured, more cooperative, and more reSponsible than the low reactives, who were judged as irritable, excitable, impulsive, and immature."47 In 1957 Jack Block48 reported a study similar to those of Jones' but one in which the "more stressful cir— cumstance of lying was employed as the stimulus situation"49 and the subjects were older and probably were a more homogeneous group. Also, objective test measures and rat- ings were used for comparison to GSR reactivity. Generally, the results Block reported were: (1) that GSR reactivity could not be ascribed to "intellectual factors as these are ordinarily understood and measured."50 (2) That re- actors seemed to be withdrawn, worrying individuals who turn their "anxieties toward internal routes of expres- sion" and non-reactors seem independent, aggressively direct, and relatively nonconforming.51 (5) "Reactors appeared to be more dependent, dreamy, idealistic and 46Reported in: Jack Block, "A Study of Affective Re- sponsiveness in a Lie Detection Situation," Journal of Abnormal and Social Psychology, 55, (1957), p. 11. 47Block, "A Study," p. 11. 48Block, "A Study," pp. 11-15. 49Block,."A Study," p. 11. DO 51 Block, "A Study," p. 12. Block, "A Study," p. 15. 50 suggestible; non-reactors were evaluated as relatively cool, evasive, opportunistic, and independent."52 The results of the Jones and Block studies seem to suggest that there is a relationship between GSR reactivity and personality structure and behavior. However, the most significant result of their studies, at least for purposes of the present study, is that they did not find a direct reciprocal relationship between behavioral characteristics (overt motor response) and autonomic activity. As Block points out, internal activity (that which is recorded via the polygraph for "lie detection") seems to be an expres— sion of an affective state; it does not itself reduce ten— sion; this can only be brought about by "appropriate "55 Thus, behavioral cognition, fantasy, or motor behavior. characteristics may be tension reducing signals which may or may not be consonant with autonomic activity. There- fore, for "lie detection" purposes the interpretation of behavioral characteristics would seem to be important in- formation which should be considered along with the analy- sis of polygraph record reSponses. When the two are redundant, i.e., when both show truthfulness or deception, for example, the examiner's task would seem to be easier. 552Block, "A Study," p. 15. 53Block, "A Study," p. 14. 51 Obviously, the same could hold true for other diagnostic procedures. Classifying Behavioral Communication Besides being a product of deception and other stress situations, behavioral characteristics also seem to be, in and of themselves, a communication system separate from, and, at times, contradictory to, the spoken word. In fact, since the early 1950's there have been many studies reported which describe systems for recording and classifying various nonverbal behaviors. A brief review of some of these studies follows.54 Paralanguage. Attempts at classifying nonverbal be- havior can be divided into at least three distinct areas, each with its own pioneering investigator. First, the study of voice quality and sounds such as laughing, yawn- ing, and grunting was developed by George Trager. He termed his work the study of "paralanguage" and divided this into two components: (1) vocalizations, and (2) voice qualities. "Trager's work constitutes the foundation for further work in the field; and, in fact, his system is quite applicable in a practical sense, enabling worthwhile transcription of paralinquistic behaviors in empirical studies."55 54The review of the literature which follows was pri- marily obtained from: Starkey Duncan Jr., "Nonverbal Com- munication," Psychological Bulletin, 72, (1969), pp. 118-157. 55Duncan, "Nonverbal," p. 119. Proxemics. A second area of study of nonverbal behavior was developed by E. T. Hall. This area is termed "proxemics" and consists of the study of "social and personal Space and man‘s perception of it."56 Hall primarily is concerned with the concept of culture as a communication system and describes distances or zones of human interaction as vari- ables between cultures. An extension of Hall's work is the concept of "territoriality" which has to do with animal behavior in setting up and defending fixed territories; this is discussed by Ardrey in his book The Territorial Impera— tive.57 Kinesics. Ray Birdwhistell is noted for the develop- ment of "kinesics," or the study of body movements and ges- tures.58 Of the three areas mentioned, this one seems to be the most developed in the study of nonverbal behavior. In kinesics, body movements or gestures are treated as a means of communication;and a system of symbolizing virtually w 56Duncan, "Nonverbal," p. 118; and, E. T. Hall, The Silent Lan» age (Garden City, N.Y.: Doubleday and Co. Inc., 1959). 57Robert Ardrey, The Territorial Imperative (New York: Atheneum, 197C). 58See: R. L. Birdwhistell, "Background to Kinesics, etc.," A Review of General_§emantics, 13, (Autumn, 1955); and, Introduction tO”KEn§sigs_(Louisville: University of Louisville Press, 1952) cited in: Julius Fast, Body Language (New York: M. Evans and Co., Inc., 1970). fly“. ‘2') every possible human movement has been developed which leads to, what could be termed, a non-speech language. The relationship which nonverbal behavior as a com- munication system has to the detection of deception would seem to lie primarily in its value as a means of codify- ing the behavior of polygraph subjects. If for example, symbols were used to describe certain actions (if neces- sary, through a study of video tape replay) such as the system advocated by Birdwhistell, it would seem that the evaluation of verbal and nonverbal behavior of subjects taking polygraph examinations could lend much to the de- termination of truth and deception. SUMMARY This chapter has reviewed some of the literature pertaining to the use of verbal and nonverbal behavioral characteristics in the detection of deception. In an ef— fort to clarify this literature it was divided into four categories: (1) research on behavioral characteristics of polygraph subjects; (2) observational comments of the use and value of behavioral characteristics in non- polygraph interrogation; (5) research directed at detect- ing deception by analyzing specific behavioral characteris- tics; and (4) research pertaining to the use of behavioral characteristics as a communication system. Most of the literature presented in this chapter lends credence to the belief that "lie detection" can and should be accomplished through a study of both autonomic activity and overt verbal and nonverbal behavior. This is not, however, a new idea, but it is one which deserves considera- bly more attention than it has received. For example, some of the first known efforts to detect liars, about 900 B.C., involved the study of behavior. Trovillo reports that in one of the papyrus Vedas the following information was offered as a means of detecting poisoners: A person who gives poison may be recognized. He does not answer questions, or they are evasive answers; he speaks nonsense, rubs the great toe along the ground, and shivers; his face is discolored; he rubs the roots of the hair with his fingers; and he tries by every means to leave the house... Many of the characteristics of poisoners in 9OO B.C. are similar to those Reid found in lying polygraph subjects in 1955 A.D. Certainly, something which seems so basic to human nature to endure for over 2000 years warrants in- vestigation. 59Paul V. Trovillo, "A History of Lie Detection," The Journal of Criminal Law and Criminology, 29, (March-April, 19397. pp. 848-849. Chapter 5 THEORETICAL BASE OF DATA STUDIED Presented first in this chapter will be an analysis of the theory and application of the structured pre-test interview technique. Subsequent to this analysis, a summary of the previous research findings on both spontan- eous verbal behavior and nonverbal behavior will be dis- cussed. A thorough understanding of this chapter is es- sential to an understanding of the methodology employed in the present research. tr “LICITED VERBAL BEHAVIOR: 1 UR THE 0 C ED PRE-TEST INT VIEV6O Prior to detailing the analysis of the structured pre- test interview a prefatory comment is necessary in order to place the interview in the proper context. At no time during the interview does the examiner-through sympathy, accusation, or any other "interrogational device" - attempt to influence the subject into answering the interview questions truthfully. The examiner does not point out to b0A complete description of the structured pre-test interview has not been published. Therefore, the description which follows is based upon the writer's experience in the use of this interview and on the partial description which appears in: John E. Reid and Fred E. Inbau, Truth and Dece - tion, The Polygraph ("Lie Detector")_Technique (Baltimore: The Williams and'Wilkins Co., 1966): pp. 11-16. 5 \fl 56 the subject any inconsistencies in his alibi or statements. Rather, during the interview the examiner's attitude is objective and non-committal; he is cordial, but firm. His purpose during the interview is not to obtain an admis- sion of lying but rather to elicit, observe, and record the subject's verbal and nonverbal behavior. The interview starts when the examiner first greets the subject, who has usually been ushered into the examining room by a secretary. After his cordial, but reserved, greet- ing the examiner immediately proceeds to place the polygraph 61 attachments on the subject. While doing this he asks the subject: "Have you ever had a polygraph ('lie detector') test before?"62 No further comment is made by the examiner during the time when the attachments are placed on the subject. However, the examiner listens carefully to what- ever the subject may say in reaction to the instrument or the test itself. 61The attachments are placed on the subject at this time only for the purpose of allowing him to become accus- tomed to them. At no time during the interview are any polygraphic recordings made. 62Reid and Inbau explain in Truth and Deception, p. 11, Note 15, "If the subject has not been told the purpose of his appearance in the testing laboratory, the examiner should, of course, explain that the test is desired of him as part of the investigation regarding the particular loss, offense, or incident, and as much time should be spent in this preliminary explanation as the circumstances reason- ably warrant. Only thereafter should there be any attach- ment of the instrument to the subject." 57 When the examiner has completed placing the polygraph attachmentS' he sits in a chair in front of the subject. At this time he immediately records whatever comments the subject made up to this point. Then, the examiner proceeds with the interview, asking first, questions about the subject‘s background. For example, he asks the subject his name, age, home address, and marital status and, in most cases, asks basic information regarding the subject's place of employment. All of this information is recorded on the examiner's interview sheet along with the subject's sex and race, and the date. After background information is acquired the examiner proceeds with the asking of the structured pre-test in- terview questions. In asking these questions he uses as a guide abbreviations of each question which appear on the back of the interview sheet. Each of these questions ap- pears on the sheet in a fixed order; however, the exa- miner may, or may not, ask them in that order. Further, depending on the case at hand, he may not ask certain of the questions because they are inapplicable to the facts of the case. As the examiner asks each interview question he re- cords the subject's verbal answer on the interview sheet alongside a number corresponding to the question asked. It is these recorded verbal answers which comprise the 58 verbal behavior studied in the present research. As the subject answers each question the examiner notes the subject's nonverbal behavior; he observes the subject's physical appear- ance, eye contact, demeanor, etc. This nonverbal behavior, while as important as verbal behavior will be discussed separately and does not appear in the description of the interview which follows. You Question63 The first structured interview question is abbreviated on the examiner's interview sheet as "You." The "You" is a guide to the examiner to ask the subject a direct ques- tion in the form of: "Did you do it (the murder, theft, etc.)?“ If, for example, the purpose of the examination is to investigate a certain murder the examiner may say to the subject: "You know, of course, you're going to be tested about the killing of Joe Jones. What I want you to understand is that ii'you didn't commit this killing, the polygraph testing will show that. If you did do this thing, the polygraph will show that. (pubject's name)did you kill Joe Jones?" When confronted with such a direct question concern- ing whether or not he committed the offense, a lying 65'The short phrase which precedes some of the follow- ing paragraphs corresponds to a similar phrase which is the abbreviation of each interview question as it appears on the examiner's interview sheet. 59 subject will usually hesitate before giving a verbal answer; his answer will usually be evasive and will not really answer the examiner's question. For example, a typical liar's answer would be as follows: "I can't understand this whole thing. I hardly knew Joe Jones. I don't know why anyone would want to kill him." On the other hand, a truth- ful subject will usually answer the "you" question in a very direct manner. He typically will say something like this: "I didn't kill Joe Jones. I had nothing to do with it but I sure hope you catch the guy who did." Suspicionpguestion The examiner then asks the subject if he has any sus- picions as to who committed the offense. The subject is requested to reveal his suspicions even if he does not know for sure. .A truthful subject will usually do so. He will reveal names and the reasons for his suspicions. A lying subject will typically refuse to blame anyone and, in fact, may try to shift the blame for the offense to persons who probably could not have committed it. For example, in an investigation involving the theft of money from a bank vault, a lying subject, when asked of his suspicions, may say something like: "It could have been anybody. Even the customers could have gotten that money." Who Not Suspect Question The subject is asked if there is anyone who he would personally vouch for and who he feels did not commit the 4o offense. A truthful subject will typically respond to such a question by naming one or more persons who he feels would not have committed the offense. He may, at times, name himself. For example, the truthful subject may say, "Well. I know I wouldn't do it. And, I know the two fellows I work with, Joe Smith and John Jones, are too honest to do itJ'The lying subject will usually fail to name anyone. He may say that he does not know anyone well enough to vouch for them or that he suspects everyone. He may answer something such as: "I don't trust anyone. I don't pay attention to other people." He does not wantany- one to be relieved of suspicion; he desires to keep the investigation as broad as possible. Think Stolen Question The examiner will ask the subject if he thinks the offense really was deliberate. For instance, in a theft case the examiner will ask the subject if he thinks that the missing items were actually stolen. A lying subject typically reSponds to such a question,even though it is obvious the items were stolen, by answering: "No, I don't think they (the missing items) were stolen. They were prob- ably lost or something. It happens all the time." In contrast to this, the truthful subject tends to readily ad- mit that the items were stolen and may say: "Of course they were stolen. They had to be. They didn't just walk away by themselves."z 41 The lying subject, of course, has good reason for wanting the examiner to believe that the missing items were lost or misplaced. If he can convince the examiner and everyone else of this, he himself would be relieved of suspicion and would be "successful" in his theft. Best Opportunity Qpestion The examiner as:s the subject to assume that, in a theft case, for example, the missing items were actually stolen. He then asks the subject which of the groups of persons, out of those which could have been involved, prob- ably committed the theft. Typically, a truthful SUDjCCt will restrict the theft to one particular group of people and may even name the group to which he belongs as the one most likely to have committed the theft. For example, a truthful bank teller may say: "Well, it had to be one of the tellers. It's impossible for anyone else to have done it.“ A lying sybject will usually be reluctant to restrict the theft to any one group. A lying bank teller may say: "It could be anyone in the bank. Even a customer could have stolen it." The lying subject attempts to broaden the investiga- tion as much as possible and to direct it away from himself. The truthful subject will narrow the investigation to the group or groups most logically involved. 42 Fingerprint Question If the case is one in which fingerprints or footprints might have been left at the scene of the offense, the examiner will say to the subject: "'18 there any reason why your fingerprints should be on [3.g., a beer bottle or a glass in John Jones' hom§7 or your footprints [3.g., on the ground at First and Main Stree_t_7?' Except when the circumstances indicate the possibility of the innocent [truthfug7 subject's prints being there, he will usually say: 'No, there's no reason why my fingerprints [3r foot- print§7 should be there.‘ A lying subject, on the other hand, is inclined to offer an explanation of their pre- sence (e.g., 'Well, I may have handled a beer bottle or a glass in his house, or I may have walked by First and Main Street recently'). In other words, a lying subject may seek to forestall any incriminating inference from such a fingerprint or footprint. He may feel impelled to offer an explanation. A truth-telling person, however, will ex- perience no such concern."64 The "fingerprint" question may be asked in a different form than that stated above. For example, in cases where there are witnesses to the offense, a subject may be told: "INow, if you were at First and hain Streets [or other loca- tion of the murder, rape, etcA7 that night at any time, say so. You could have been there and yet that would not 6 . . . u ., . J. ..- 4Reid and Inbau, Trutn and Deception, pp. lJ-l4. 45 necessarily mean you shot John Jones. Maybe you were there just before or soon after and had nothing to do with it. Someone could have seen you before or after, but later on he may assume or believe he saw you there when Jones was shot. 80 it's important that we get this straightened out before the test. Tell me, then, were you there at any . E‘ time that night?'“6) The lying subject may answer such a question by saying: "Well, I was there that day, and may— be someone saw me; but I didn't shoot anyone." The truth- ful subject, on the other hand, tends to deny that he was at the scene of the offense, barring, of course, situa- tion where his presence is legitimately explainable. Borrow Question In those instances where a theft of money (from an employer) is being investigated the examiner may ask the subject: "Did you ever borrow any money (or merchandise) from your employer even though you paid it back?" A lying subject will usually answer such a question with an admis- sion of borrowing such as: "Well, I did take some money for lunch from petty cash one time, but, I paid it back as soon as I got my paycheck." A truthful subject will typically respond: "I never borrowed anything. The only money I take out of there is my pay." 65Reid and Inbau, Truth and Deception, p. 14. 44 The reason the "borrow" question is said to be helpful is that borrowing is reputedly the first step toward steal- ing, especially in cases of embezzlement. A lying subject, recognizing this, readily admits his "borrowing" as this conduct is minimized in his mind by his more serious steal- ing activity. The truthful subject, however, even if he did borrow on occasion, is reluctant to admit it as its seriousness is maximized in his mind. Ever Happen Before Question A subject may be asked by the examiner if he had ever been involved in an investigation similar to the present one. A subject who states that he had been involved be- fore, particularly if that involvement had occurred on numerous other occasions, is more typically a lying sub- ject. If investigations of a certain type of offense (e.g., theft, murder, rape, etc.) seem to "follow the subject" it is presumed that they may do so because of his involve— ment in them. A truthful subject usually has not been in- volved in investigations of serious offenses on numerous occasions. He may tell the examiner: "No, this is the first time I've ever been questioned about any crime." Whereas, a lying subject may answer: "I've been questioned lots of times about this type of thing, but I've never been convicted or anything." 45 ‘Thipk Question The "think" question is one which an examiner can adapt to most any type of investigation. It is based upon whether the subject ever thought about doing anything similar to, or the same as, the act under investigation. For instance, in a murder case in which the victim was shot to death, the subject would be asked: "'Did you ever think about shooting or killing anyone, 21?“ though ypu didn't? If so, I want you to get this off your mind before the test.‘ The characteristic answer of a person who did not do the killing is: 'I never thought of shooting or kill- ing anyone!‘ His answer is emphatic and unequivocal, for even though he may have had a fleeting thought along that line, he interprets the examiner's question to mean a serious, deliberate thought of killing someone. On the other hand, a lying subjectis likely to respond by saying: tSure, I've thoughtabout doing things like that; everyone does. But I didn't do it,:no6 Approach Question The examiner may ask a polygraph subject, in a theft case, for example: "Have you ever been approached by any- one to help them steal something like this (the stolen item bCHeid and Inbau, Truth and Deception, p. 14. 46 in the investigation) even though you didn't go along with them?" The subject's answer to such a question provides the examiner with some insight into how the subject is viewed by his acquaintances. A lying subject, for instance, may tell the examiner: "Sure, I've been asked a lot of times. But, I've never gone along with anyone in something like that." A truthful subject, typically answers the "approach" question in a very direct manner. He may say something such as: "No one ever asked me to do anything illegal. They know better than that. I‘d turn them in." How Will L.D. Come Out on You Question The examiner asks the subject how he thinks his test will come out; for instance, the examiner may say: "How do you think you will come out on the polygraph (lie de- tector) test today?" The truthful subject will usually give a very direct, unequivocal answer. For instance, he may say: "I'll come out perfect. At least I better come out perfect because I didn't do it." On the other hand, a lying subject will typically offer excuses for the polygraph test indicating that he is lying. He will tell the examiner that he has some im- mediate physical or emotional problem, or, will offer some other reason for the polygraph test to give misleading in- dications. He may say, for instance, something like: "Well, I don't know what the test will show. I'm a very nervous person and I've heard that the 'lie detector' doesn't work well on nervous peeple. Anyway, I don't 47 believe in 'lie detectors.'" Truth Serum Question In some investigations the examiner, for observa- tional purposes, may ask a subject an interview question regarding the subject's willingness to take "truth serum" in the immediate investigation.67 For example, the examiner may say: "We find that the polygraph does not work on everyone. Sometimes we cannot tell if a person is, or is not, telling the truth. In other words, sometimes our tests are inconclusive; and, while this only occurs in about 5 percent of our cases, if it happens in your case, would you agree to take truth serum to resolve this (mur- der, rape, theft, etc.)?" A truthful subject will usually not hesitate in af- firming his willingness to take "truth serum" if the need arises. He may say: "Sure, I'll take that or anything else you want to give me. I want to prove I didn't do this thing." A lying subject, believing in the infallibility of truth serum, will usually decline to take it. Or, if he 67A polygraph examiner, of course is usually neither equipped, nor qualified, to administer a truth serum test. For an interesting discussion of the use of truth serum see: Fred B. Inbau, "Scientific Evidence in Criminal Cases," Journal of Criminal Law and Criminology; 24, (March-April, Mn, pp. 1153-1157. 4s agrees to take it, he will qualify his agreement with con- ditions which would, in effect, preclude, or at least delay, the truth serum test. For example, a lying subject may say: "No, I won't take truth serum. I don't like needles. This ‘lie detector' test is as far as I will go;" or, "Well, I will have to talk to my wife, attorney, and doctor before I take the stuff." When asked the names of his doctor and attorney the lying subject will usually be reluctant to give them or will tell the examiner that he doesn't have any new but will have to get them. What Should Happen to Doer Question When asked by the examiner what should happen to the actual perpetrator of the offense being investigated a truthful subject will characteristically want him dealt with in a very harsh manner. He may express a desire to see the perpetrator "locked up," even for a very minor offense. He may also indicate to the examiner that he himself would like to deal with the perpetrator and inflict his own form of punishment. A truthful subject's typical answer to the "what should happen to the deer" question would be: "He should be put in jail. He caused me a lot of trouble and I'd like to get a hold of him myself." In contrast to the truthful subject, a lying one will usually want the perpetrator treated "softly." He may 49 offer excuses for the perpetrator's conduct and may indi- cate that the perpetrator needs help, not punishment. For instance, a lying subject may respond to the "what should happen to the deer" question by saying: "I don't know what should be done with him. Maybe he just couldn't help himself and the crime wasn't his fault. Maybe he should go to a psychiatrist instead of jail." Or, in a theft case, the lying subject may say merely that the perpe- trator should pay the money (or merchandise) back and be given a second chance. Attitude Question This question is usually the last of the structured interview questions asked.68 The examiner asks the sub- ject how he feels about taking the polygraph examination. While the subject's answer provides the examiner with some insight to the subject's actual attitude, it also reportedly provides the examiner with a verbal clue to the subject's truthfulness or deception. For example, a truth- ful subject will typically respond to such a question in a manner that indicates his willingness, and sometimes eagerness, to take the polygraph examination. He may, for 68As previously explained the order in which the interview questions are asked may vary. However, the "attitude" question is usually asked last in order to allow the examiner to determine if the subject's expressed attitude is consistent with the examiner's impressions of his attitude up to that point. 50 instance, answer with: "I'm glad to take it (the examina- tion). I want to prove my innocence; and, besides, it's a new and interesting experience for me." A lying subject will usually answer the "attitude" question with words that express his contempt for the examination or his disbelief in its effectiveness. He may, for example, answer as follows: "I don't like tak- ing this test. I think it's a silly thing to do and a waste of money; and besides, I don't believe in 'lie detectors.'" After the examiner has asked the structured interview questions of the subject, he then reviews with the subject the exact test questions. He will also, at this time, make any necessary explanations or clarifications to the subject. However, the examiner does one other thing before he starts the actual polygraph testing. He records, on the front of his interview sheet, his appraisal of the subject's behavior during the pre-test interview. He notes, for example, if the subject looked like a typical lying subject or like a typical truthful one, if the subject talked like a typical lying subject or a truthful one, and, if the subject had good eye contact or poor eye contact, etc. These notes provide the information for the nonverbal behavior studied in the present research. 51 SPONTANEOUS VERBAL BEHAVIOR In his 1955 study of polygraph subjects' behavior, John Reid found that lying subjects frequently make certain types of spontaneous statements to the examiner throughout the examination. In some cases they try to explain why their polygraph responses "might mislead the examiner into believing that they are lying. Hence, they complain of being nervous, and if that does not seem to impress the examiner, they further emphasize their 'ner- vous condition' or mention a physical defect which they may or may not actually have. Also, they frequently feel it necessary to assure the examiner that they are very religious, hoping that the examiner will dismiss them as innocent because of their alleged righteousness."6S Reid also found that lying subjects sometimes com- plain of physical pain caused by the polygraph apparatus. On occasion they complain that the examination is taking too long and they frequently inquire of the examiner, throughout the testing, as to how they are doing. WK 09John E. Reid and Richard 0. Arther, "Behavior Symp— toms of Lie-Detector Subjects," The Journal of Criminal Law, Criminology and Police Science, 44, l (Hay47une, 1953), p. 105. 52 Truthful polygraph subjects, Reid found, do not usually complain. While they may refer to their nervous- ness, they do not persist in this after the examiner assures them that nervousness makes no difference. Since they are usually anxious to take the polygraph examination, truthful subjects do not usually ask to hurry the examination. A questioning attitude was found by Reid to be common to both truthful and lying polygraph subjects. They in- quire as to the use of the polygraph apparatus and what the examiner can tell them about themselves. For example, they may ask the examiner if he can tell them what their blood pressure is, etc. NONVERBAL BEHAVIOR Nonverbal behavior is defined as the demeanor and appearance of the polygraph subject during the examination. Reid found that in many cases the nonverbal behavior of lying subjects differed from that of truthful subjects. Lying subjects look worried and highly nervous; they some- times have a "dry mouth," continually sigh or yawn, fail to maintain good eye contact with the examiner, and are usually quite fidgety. At times, lying subjects act aggres- sively, appear in a shocked condition, or may be overly friendly and polite to the examiner. Truthful subjects, Reid reported, are usually "at ease, lighthearted, and talkative."7O They appear to be sincere and direct in talking to the examiner, but they are not overly friendly or solicitous. 70Reid and Arther, "Behavior Symptoms," p. 106. 53 Chapter 4 DESIGN OF THE INVESTIGATIOH There will be three main sections of discussion in this chapter. F'rst, background information of the data studied in the present research will be presented. This will include the source of the data, a brief description of the original procedure used in collecting the data, and an analysis of the sample studied. Second, the hypo- theses formulated for each category of data studied and the methodology employed will be discussed. Finall,, in order to put the study into perspective, some of the limitations of the data and the methodology will be presented. BACKGROUND INFORMATION OF DATA STUDIED Source of Data This study is based upon an analysis of verbal and nonverbal behavior clues recorded during actual poly- graph examinations. While these examinations were conducted during the years from 1964 to 1971, all were conducted according to standard Reid Control Question Technique (RCQT) . U7 4'}. K)”. \J‘. Essentially, the RCQT consists of a structured pre- test interview and polygraph testing. During the interview, the examiner explains to the subject the purpose of the test and the nature of the instrument; he also formulates and reviews with the subject the actual test questions. Moreover, during the interview, the examiner asks certain questions of the subject from a structured format.71 The purpose of these structured questions is to elicit be- havior clues from the subject. As the subject answers the structured questions, the examiner records those answers on an interview sheet. These answers are said to provide elicited verbal clues to the subject's truthfulness or deception. In addition to these clues, however, the ex- aminer also notes nonverbal clues. He observes and re- cords on the interview sheet, the subject's physical re- action to each questiOn and any changes in either the sub- ject's demeanor or attitude. At the conclusion of the interview, which lasts about twenty minutes, the examiner proceeds with the polygraph testing. The polygraph testing consists of the asking of rele- vant, irrelevant and control questions during a number of separate tests. The questions in the 3,5,C,9,and lO posi— tions are relevant and relate to the matter under investi- rl. -. H Z See Chapter 9, pp.95-51. 56 gation, such as, in a murder case, "Did you kill John Jones?" and "Did you shoot John Jones with a .58 caliber revolver?" The questions in the 1,2,4, and 7 positions are irrelevant to the issue being investigated; they deal with such matters as, "Do they call you Joe?", "Are you over 21.years of age?", etc. These irrelevant questions are asked for the purpose of establishing the subject's normal pattern of responsiveness. The remaining two questions are control questions. They are placed in the 6 and 11 positions. A control question is one which is unrelated to the matter under investigation, but is of a similar, though less serious nature and one to which the subject will, in all probability, lie; or at least his answer will give him some concern with respect to either its truthfulness or its accuracy. For instance, in a burglary investigation the control question might be, "Did you ever steal anything?" or "Except for what you have already told me, did you ever steal anything else?" The response or lack of response to the control question (in respiration, blood pressure-pulse rate, or GSR) is then compared with what appears in the tracings when the subject is asked the questions relevant to the issue under investigation. This comparison is the primary basis for 57 the examiner's ultimate truth and deception decision./2 Throughout the testing period the examiner observes the subject. He notes on the interview sheet any comments or complaintsnmde by the subject;he also notes any questions the subject asked during the testing period. These notes are said to provide spontaneous verbal clues to the sub- jects truthfulness or deception. At the conclusion of the polygraph testing the exa- miner interprets the subject's polygraph records and his behavior in terms of truth and deception. If both the records and the behavior are congruent, i.e., if both indicate deception, for example, the subject is reported as being deceptive. However, if the records and behavior are not congruent, the records always take precedence. In other words, behavior clues are used only to reinforce and aid polygraph record interpretation; they are never used to overrule it.73 Subjects Studied The interview sheets of one hundred different poly- graph subjects were randomly selected from the verified files of John E. Reid and Associates, a security consulting 72The previous paragraph is an excerpt from Fred E. Inbau and John E. Reid, "The Lie Detector Technique: A Reliable and Valuable Investigative Aid," American Bar Association Journal, 50, 5 (May 1964), p. 170. 73See: John Reid and Fred Inbau, Truth and Deception, The Polygraph ("Lie Detector“) Technique (Baltimore: The Williams anTWiIkinsTfo” 9667, p. I52. agency specializing in the administration of polygraph examinations, 600 S. Michigan Avenue, Chicago, Illinois. These files consist only of data pertaining to polygraph subjects whose truthfulness or deception is certain. For example, if in a certain murder case, ten subjects are given polygraph examinations and one of those subjects offers a complete and fully corroborated confession, he would be considered a verified lying subject and his case data would be filed accordingly. The remaining nine sub- jects would be considered verified truthful. Fifty of the subjects selected for this study were of the verified truthful category; fifty were verified liars. Table 1 summarizes the background data for each category of subject. Of the fifty truthful subjects studied, 24 were males and 26 females; 47 were Caucasian, and 3 were Negro. The age range for these subjects was from 13 to 68 while the mean age was 34 and the median age, 28. Of the fifty lying subjects, 43 were males and 7 were females; 40 were Caucasian, lO Negro. The age range for these subjects was from 13 to 65 with a mean age of SO and a median age of 25.5. 59 Table 1 Comparison of Background Information of Polygraph Subjects INFORMATION TRUTHFUL SUBJECTS LYING SUBJECTS (N=50) (sto) Sex: Male 24 43 Female 26 7 Race: Caucasian 47 4O Negro 5 10 Age: Range 13-68 15-65 Mean 34 50 Median 28 25 The Original Investigations Of the SO verified truthful subjects, 48 of them were involved in theft investigations, e.g., employee theft, burglary, robbery, etc.; 1 each of the remaining two sub- jects were involved in the investigation of a murder and of sexual offense, respectively. A total of 6 different examiners conducted the examinations on the truthful sub- jects. Of the 50 verified lying subjects, 41 of them were involved in theft investigations, 5 in sexual offenses, 2 in bribery, and 1 each in a child beating investigation 60 and the illegal use of alcohol, respectively. A total of 4 different examiners conducted the examinations on the lying subjects. Table 2 shows the distribution of the types of ori- ginal investigations for each subject category. Table 2 Comparison of the Original Investigation Issues INVESTIGATION ISSUE TRUTHFUL SUBJECTS LYING SUBJECTS (N=50) (N=50) Theft 48 41 Murder 1 Sexual Offense 1 5 Drinking l Bribery 2 Child Beating 1 The Polygraph Examiners No precise data are available as to the background information of the ten examiners who had conducted the examinations on the subjects studied. However, it is known that all the examiners were licensed as Detection of Deception Examiners in the State of Illinois; all had 61 undergone a six month internship training program con— ducted at John E. Reid and Associates, Chicago; all had been taught to conduct examinations according to the RCQT and to observe and record polygraph subjects' be- havior in a similar manner; and, all had a minimum of one year of full time experience in conducting poly- graph examinations. METHOD OF STUDY This section contains a discussion of the procedure used in categorizing and evaluating information from the subjects' polygraph interview sheets. The discussion will involve first, elicited verbal clues, then spontan- eous verbal clues, and finally,nonverbal clues. Elicited Verbal Clues Elicited verbal clues are defined as the polygraph subject's verbal answers to each of the examiner's structured interview questions. In other words, the answers which the subject gives in response to the structured interview questions are supposed to provide a clue to the subject's truthfulness or deception (about the offense being investigated), according to the theory of the interview, as previously discussed. These answers are recorded by the examiner, as the subject 62 says them, on the interview sheet alongside a number cor- responding to the particular question answered. For ex- ample, if a subject answers the "think" question with: "No, I never thought of killing anyone." and, the "truth serum" question with, "Sure, I'll take truth serum.", the examiner records these answers on his interview sheet alongside the numbers 9 and 12, respectively, as these are the ninth and twelfth structured interview questions. The examiner follows this same procedure for recording all of the subject's answers to the interview questions. Three hypotheses were formulated as regards the elicited verbal clue data. These were:(l) The truthful subjects will give significantly more “typical truth-teller" than "typical liar" answers to each of the structured inter- view questions than would be expected on a chance basis for either answer. (2) The lying subjects will give significant- ly more "typical liar“ than "typical truth-teller" answers to each of the structured interview questions than would be expected on a chance basis for either answer. (5) Each of the structured interview questions will provide a basis for discriminating between the truthful and the lying sub- jects. That is, the truthful subjects will give signifi- cantly more "typical truth-teller" and the lying subjects significantly more "typical liar" answers (than would be expected on a chance besis) to each of the structured in- terview questions. . r1 ()9 In order to test the hypotheses it was necessary to classify the recorded verbal answers of all the verified truthful and lying subjects. This classification was done by reviewing each of their answers to all questions asked of them independent of any knowledge of their actual truth- fulness or deception. As each recorded answer was reviewed a judgment was made as to whether or not that particular answer was one which, according to the theory of the struc- tured pre—test interview, would be said by a truthful sub- ject or by a lying subject in response to that particular interview question answered. Each answer for each question was then classified as being either the answer of a "typi- cal truth—teller" or a "typical liar." For xample, if a subject's recorded verbal answer to the "truth serum" ques- tion was similar to: "Sure, I'll take truth serum, or anything else you want to give me.", it was classified as a "typical truth-teller's," answer. On the other hand, if the recorded answer was something like: "I won't take truth serum until I check with my doctor.", it was classified as a "typical liar's" answer. All recorded verbal answers to the structured interview questions were classified independently of other verbal answers. Also, the classification was made only on the recorded verbal answer; any nonverbal behavior which may have been noted at an answer was disregarded. In the case where a verbal answer's actual classification was 64 doubtful, i.e., where a classification judgment was ex- tremely difficult, that answer was classified as a "typical truth-teller's" answer. The reason for this is that for verbal answers to be useful in discriminating between the truthful and the deceptive, they should do so without ex- traordinary interpretation; the presumption of truthfulness should take precedence over ambiguity. The recorded interview answers for all subjects were classified. Then the number of "typical truth-teller" and "typical liar" answers to each interview question was tabulated for both categories of subjects. For example, the number of "typical truth-teller" answers to the "you" question for the fifty verified truthful subjects was determined; then, the number of "typical liar" answers to this same question for the same subjects was deter- mined. This procedure was followed for all interview questions for all verified truthful subjects; and, was repeated for the verified lying subjects. A master data sheet for each category of subject was used to tabulate the elicited verbal clue results.74 Spontaneous Verbal Clues Spontaneous verbal clues are those comments which a polygraph subject makes during the examination without 74See Appendix A, Section I for master data sheets for both categories of subjects; and, Section II for samples of actual verbal answers and their classification. 65 direct prompting from the examiner. When a subject makes such a comment the examiner records it on the front of the interview sheet. Reid's prior research suggests that Spontaneous verbal clues are usually made by lying subjects although he offered no hard data to support this position. Therefore, for the present research, the following hypothesis was formulated: The lying subjects will exhibit spontaneous verbal clues significantly more often than the truthful subjects. It was decided that it may be of some practical value to classify Spontaneous statements, even though this was not essential to the testing of the hypothesis. Based upon Reid's research this classification resulted in three groupings of spontaneous statements: complaints, questions, and requests to hurry the examination. All of the subjects' interview sheets were reviewed. The recorded spontaneous statements were placed into one of the three groups and the frequency of occurrence of spontaneous clues in each group, for both categories of subjects, was tabulated.75 From this the spontaneous clues recorded for the truthful subjects were compared to those of the lying subjects. 75See Appendix B. 66 Nonverbal Clues Nonverbal clues include such things as the general appearance and demeanor of the polygraph subject, the way in which he answers questions, and the examiner's impres— sion of whether or not the subject looks and talks as a truthful subject or as a lying subject. These clues, and any others which the examiner may observe during the in- terview and examination, are recorded by the examiner on the front of his interview sheet. Reid's prior research suggests that the nonverbal behavior of truthful and lying subjects often differs. It also suggests, however, that at times truth-telling sub- jects exhibit some behavior more typical of "liars" and vice versa. Therefore, two hypotheses were formulated as regards nonverbal clue data. These were: (1) A significantly greater number of the truthful subjects will exhibit "typical truth-teller" nonverbal clues more often than "typical liar" nonverbal clues than would be expected on a chance basis. (2) A significantly greater number of the lying subjects will exhibit "typical liar" nonverbal clues more often than "typical truth-teller" nonverbal clues than would be ex- pected on a chance basis. The testing of the hypotheses required that the nonverbal clue data recorded on the interview sheets be 67 classified according to what is supposed to be "typical truth-teller" and "typical liar" behavior. Reid's prior research was used as a basis for this classification. From his data two equal groups of descriptors were developed. One group consisted of nine "typical truth-teller" descriptors. These were: "genuinely friendly," "direct answers," "good eye contact," "cooperative," "light hearted," "composed," "relaxed," "talkative,"76 and "truth- teller‘s appearance." On the other hand, the "typical liar" grouping consisted of: "overfriendly," "evasive answers," "poor eye contact," "uncooperative," "scared," "nervous facial appearance," "nervous bodily movements," “untalka- e"77 tiv and "liar‘s appearance." 76The word "talkative" is used here as meaning responsive to the examiner‘s questions. Reid explains that truthful subjects "usually become more responsive and begin to talk quite freely" after the interview begins. See John Reid and Richard Arther, "Behavior Symptoms of Lie-Detector Subjects," The Journal of Criminal Law, Criminology and Police Science, 44, 1 (May-June, 1953), p. 107. 77The‘word. "untalkative" is used here as meaning only minimal responsiveness to the examiner's interview questions, such as, only a "yes" or "no" when a more lengthy answer would seem to be in order. leid explains that lying subjects are often "afraid to speak for fear of trapping themselves" and rarely, if ever, change from this non-talkative state. See Reid and Arther, "Behavior Symptoms," p. 107. 68 The descriptors "truth-teller's appearance" and "liar's appearance"require some elaboration. By themselves, of course, they are not specific nonverbal clues. Rather, they are an examiner's'shorthand" for his overall appraisal of a subject'esnonverbal clues. However,because this re- searcher was aware that the original examiners; had been trained to use such shorthand and because these descrip- tors do give a good indication of an examiner's nonverbal clue appraisal, they were included. The interview sheets were reviewed for the presence of words and phases which the examiners had recorded to describe the subjects' nonverbal behavior. All words and phases were then converted into the descriptor which came closest in meaning to the examiner's words and phases. For example, if an examiner had recorded the word "fidgety," it was converted into the descriptor "nervous bodily move- ments;" if the word "calm" was recorded it was converted into the descriptor "composed;"if the phrase "dry mouth" was recorded it was converted to "nervous facial appearance" etc. Subsequent to the conversion of all nonverbal clue data to the apprOpriate descriptors, all subjects were classified as "typical truth-tellers" or "typical liars." This was,for the most part, a quantitative classification and was accomplished by assigning one point for each 69 descriptor appearing in either of the "typical"groups and then totaling the points for each group for each subject. The "typical" group with the most points was used as the basis for classifying subjects. For example, if the non- verbal descriptors for a subject were: "composed," "re- laxed," "good eye contact," "evasive answers," and "over- friendly" that subject was classified as a "typical truth- teller." This is so because the first three of these descriptors are those of a "typical truth-teller" while the latter two are those of a "typical liar." Thus, the subject received three truth-teller points and only two liar points. In no case was any single descriptor counted more than one time for any subject regardless of the number of words or phrases which were converted into that descriptor. The reason for this was to lessen the possibility of a subject being classified on the basis of only one outstanding behavior trait. In four instances the total "truth-teller" points equalled the total “liar" points; all of these occurred on the verified lying subjects. In these cases, a qualitative decision had to be made. This was done by giving pre- ference to descriptors which indicated the examiner's over— all appraisal, e.g., "truth-tellerk;appearance" etc. Thus, if the following descriptors were recorded for a 70 subject, "truth-teller's appearance," "good eye contact," "nervous facial appearance" and "uncooperative" that subject was classified as a "typical truth-teller." There were four interview sheets on which nonverbal clue data were not recorded; three of these were sheets of verified truthful subjects and one was of a verfiied lying subject. Therefore, of necessity, the nonverbal clue results are based upon forty-seven and forty-nine sub- jects for the truthful and lying categories, respectively.78 LIMITATIONS OF THE STUDY This research was based on "after-the-faet" infor- mation derived from real-life polygraph examinations. As a consequence, it was not possible for the researcher to allow for all the variables involved in the subjects' exam- inations. For instance, the sample subjects were selected only on the basis of their known truthfulness and decep- tion. No attempt was made to select subjects according to their personal characteristics or to the case type in which they were involved; nor were the truthful sub- jects matched with the lying subjects. Moreover, it was not possible to evaluate the personal characteristics of the original examiners. Although an attempt was made to compensate for these shortcomings by describing the subjects and examiners in as great detail as possible it is readily admitted that any of these variables could 78See Appendix C for the nonverbal clue master data sheet. 71 have influenced the original data on which the research is based.79 However, since the purpose of this research was to determine if a difference exists between the be- havior of truthful and lying subjects, truthfulness and deception were considered irrespective of the possible in— fluence of other factors; the results must be viewed accordingly. There is no assurance that the sample subjects are representative of_§;l polygraph subjects. The reason for this is that only about 30 percent of all polygraph subjects are found to be lying; of these, only a small percentage are ever verified. About 70 percent of all polygraph subjects are reported as truthful?O only a small percentage of these are ever verified. Thus, selecting subjects from the verified files does not necessarily 79See for example: Goldine C. Gleser, Louis A. Gottschalk, and Watkins John, "The Relationship of Sex and Intelligence to Choice of Words: A Formative Study of Verbal Behavior," Journal of Clinical Psyphologv, 15, (April 1959), pp. 182—191; Robert Halmo, Thomas Boag, and A. Arther Smith, "Physiological Study of Personal Inter- action," Psychosomatic Medicine, 19, (1957), pp. 105-119; and, J.S.—Bruner and R. Taguiri, "The Perception of People" in G. Lindzey (ed.), Handbook of Social_§sychology, 2, (Reading, Mass.: Addison-Wesley, 1954), pp. 634-54. 80See: Fred B. Inbau and John 3. Reid, Lie Detec- tion and Criminal Interrogation (Baltimore: The Williams and *Wilkins Co., 1953), pp.‘IIO-113; and, R. A. Sternbach, L. A. Gustafson, and R. L. Collier, "Don't Trust the Lie Detector," Harvard Business Review, 40, (Nov.-Dec., 1962), 72 mean that the subjects are representative of all polygraph subjects. The criterion used as the verification of the sample subjects' truthfulness or deception was a fully corroborated confession. There are some who would quarrel with the validity of such verification.81 Nevertheless, at the present time, a corroborated confession is believed by this researcher to be the only effective index of truth- telling and lying for purposes of real-life polygraph research. It may have been preferable for the coding, or categorization, of the data in the present research to have been done by two, or more, persons. This may have increased the reliability of the coding or, at least, would have made the reliability measurable. However, due to time and expense factors it was not possible to train coders for this purpose. Nevertheless, it is be— lieved that the methodology employed by the researcher, and his experience with and knowledge of the polygraph technique,provided the basis for more accurate coding than could be obtained by persons without such knowledge and experience. 81For discussion of an alleged "false positive" polygraph examination and the problem of using a con- fession as a valid verification of truthfulness and de- ception see: H. B. Dearman and B. M. Smith, "Unconscious Motivation and the Polygra h Test," American Journal of Psychiatry, 119, (May 1963 , p. 1017. 73 Two basic assumptions of the methodology used for studying nonverbal behavior were: that if behavior occurred it was observed and recorded; and, if behavior did not occur it was not recorded. These assumptions were necessary because there is no "check off" system which forces exam- iners to self-code nonverbal behavior during an examina- tion. While these assumptions are open to question, it should be noted that all examiners were trained in the RCQT and all had adequate experience in using it; these factors give some support to the validity of the assumptions. In order to compensate somewhat for the lack of a self coding system, which would have provided for a direct com- parison between the various nonverbal clues, a methodology was used which compared these clues collectively, on a quantitative basis, rather than individually on a quali- tative basis. The results of this methodology may not be as meaningful to polygraph examiners as would be a quali- tative analysis. Chapter 5 RESULTS OF THE STUDY The results of the study will be reported in this chapter. In order to clarify these data the three types of behavior clues studied, elicited verbal, spontaneous verbal, and nonverbal, will be considered separately. ELICI‘ED VERBAL CLUE RESULTS Table 5 displays the number and percentage of "typical truth—teller" and "typical liar" answers given to each structured interview question by the truthful subjects. Also shown are the results of chi-square tests which were run to test the significance of the results of each interview question. In running these tests a chance expectancy was assumed. That is, the truthful subjects who were asked each question were expected to have given as many "typical truth-teller" as "typical liar" answers. This assumption was made because there is no data avail— able on which a different expectancy level could be 74 ‘75 Ho>mH Ho. wsohop Ho.vm pqwoawquamtt Hm>ofl mo. ozomop no .pw psdoamanwant hammmmoo: when: .hpfiszapnoo How coupoonhoo mopmw Apfizd "w mum wHH amp eon OHm **s.oH maul m we on. was *m.m on ad ob am am **s.mm ma e um oe we **m.m~ OH m cm as me **H.HH o o ooa ma ma *th.mm HH m ms 01 me *em.ma ma e pm em um **>.m~ o N ew Hm «m ms. mm s an m H **o.~H AH m no mm am mo. as mm mm em as o.H mm m we ma em *Agv <.m mm mm mm mm se atm.oe a N am as am a .0; mm; .om ¢ and N muesmsd newness? anasaoo muozmn< guoflampugpsna gonna mpoonpsm noozpop aux snags Hausaseg Hwosnsa= Hawspsnm .om quaoa evapavpd noon op nmgmwm ashom manna so» so pzo maoo .n.q AH“: 30: somonmmd Madge omomom smmmmm umbm zonnom mpsfipmnomsam mpfisdpnommo pmom :ofiopm usage poommsm pom 0:: :oaoaqmsm so» coaumoso 3ea> Iaman omndposnpm Mo soapma>mnpn< mpoanpsm Hamspsna "mpflsmom mafia Hangs» ompfiofiam n wands 76 based.82 It should be kept in mind in viewing this table that the results shown across each row are based upon the number of subjects shown in the first column, not neces- sarily the total sample of fifty subjects. An analysis of the data in Table 5 shows that ten of the interview questions gave results in agreement with the hypothesis. In other words, a significant number of the truthful subjects who were asked these ten questions gave more "typical truth-teller" than "typical liar" an- swers. Of these ten questions the "what should happen to deer" question was answered with "typical truth—teller" answers by 70 percent of the subjects asked; this result was significant at the .05 level. The percentage results for the other nine questions ranged from 82 percent to 100 percent; these results were significant at, or beyond, the .01 level. The four interview questions which did not give re- sults in agreement with the hypothesis for the truthful subjects were the "suspicion," "who not suspect," "think stolen," and "fingerprint" questions. The "suSpicion" question gave results in a negative direction; the truth- ful subjects who were asked this question gave a statisti— 82While Reid used a pre-test interview prior to the publication of his article on behavior symptoms of polygraph subjects, he did not mention the idea of a structured inter— view at that time. The first mention of the structure” in— terview, though Reid has never termed it this way, was in: Fred E. Inbau and John E. Reid, Lie Detection and Criminal Interrogatign (Baltimore: The Williams and Wilkins—Co., 1955 , Pp. 10-15. 77 cally significant, greater number of "typical liar" answers than "typical truth-teller" answers. The latter three of the questions not in agreement with the hypothesis, while all giving results in the right direction, i.e., more "typical truth-teller" than "typical liar" answers,did not give results approaching significance. The elicited verbal clue results of the lying subjects are displayed in Table 4. These results, as viewed across each row, are based upon the number of subjects shown in the first column and not necessarily on the total subject sample. The data in ‘able 4 show that a statistically signif- icant number of lying subjects gave more "typical liar" than "typical truth-teller" answers to six of the fourteen interview questions. The "suspicion," "think stolen," "best opportunity," and "what should happen to deer" ques- tions were answered with "typical liar" answers by 91 per- cent, 89 percent, 92 percent, and 85 percent of the lying subjects who were asked these questions, respectively; these results were significant at, or beyond the .01 level. The "who not suspect" and the "truth serum questions were answered with "typical liar" answers by 90 percent and 68 percent of the lying subjects, respectively; these results were significant at the .05 level. It is interesting to note that for the lying subjects two of the interview questions, the "ever happen before" and the "approach" questions, gave results in a negative '78 H¢>~H Ho. season no .pa pnaowmaamam** Hobofl mo. caches no .pd pnwoamanmflmt hnwmmmoes when: .hpfisqapnoo How soapoonnoo mopww spasd Mm use mom mam sma mos mqaaom mo. - mm, mN say oN ma mespappa **m.m mm ma ed a mN noon op :mammm *N.e we AN Nm ma oe esnom spans 50% so on. em sN me oN as see maoo .q.g HHa3.3om Auv H.N em Ha me oN Hm nomonaa< om. mm sN Ne ma me snags Auv OH. we ma em HN mm ouoeom commas amps N.H as m an e ma zonuom s.a as ea mm s HN mpnananomqwm *xo.ea Na MN m N mN spaqspnoamo pmom **n.mH as an HH e mm amaosm Medea *m.¢ om 0 OH H OH somgmsm p02 or: **m.sN am on m e. me cofioaamsm ¢.m es Nm on ma om so» I! m\ Jloz R .02 ¢ one N uuokuqd coavmeso soapmeso Beab mssdaoo mno3ms¢ :HeHHevlnpsue wexm< mpoonnsm Inean cessposupm seezpep x =HMHA Hwoaahe= Hwoaaha: msflhn .om Ho nowpeabonnnd MN mpoonnsm mnahg ”spasmom made Hanno> eopHOfiHm e mapae 79 direction although these results were not statistically significant. In other words, the lying subjects answered both of these questions more often with "typical truth- teller" than with "typical liar" answers. Based upon the data in Tables 3 and 4 it is evident that only three interview questions, by themselves, provide a basis for discriminating between the truthful and lying subjects. The "truth serum,"tne "what should happen to doer,"and the "best opportunity" questions all gave statis— tically significant results in the right direction for both categories of subjects. iowever, the first two of these questions were significant at different levels for the truthful and lying subjects, while the "best opportunity" question was significant at the .01 level for both cate- gories of subjects. The "fingerprint" question is the only one which did not show significant resultszflnreither the truthful or lying subjects. However, since this question did show re- sults in the right direction for both categories of subjects it is possible that significant results would have been found if a larger sample of subjects had been studied. Only twelve of the truthful and twenty-one of the lying sub- jects were asked this question. As explained previously, all polygraph subjects are not asked all structured interview questions depending, of course, upon the case facts at hand. As a consequence of this, it is not possible to analyze the elicited verbal clue results collectively, i.e., to consider the combined 80 effectiveness of the interview questions, as an examiner does during actual testing. All that can be stated in this regard from the data shown in Tables 3 and 4 is that of the total 510 answers given by the truthful subjects, 78 percent of them were "typical truth-teller" answers; where- as, 69 percent of the total 465 answers given by the lying subjects were "typical liar" answers. These percentages, however, have little meaning since they do not reflect that the interview questions, as a group, were 78 percent and 69 percent effective for the truthful and lying sub- jects, reSpectively. SPONTANEOUS VERBAL CLUE RESULTS Table 5 summarizes the occurrences of each of the three classifications of spontaneous verbal clues for both the truthful and lying subjects. As is shown, only 2 of the truthful subjects made any kind of complaining state- ment during their examinations. Both of these subjects complained of the "pins and needles feeling" caused by the polygraph's blood pressure cuff. In contrast to this, 17 of the lying subjects complained during their examina- tions. Of these seventeen, 7 lying subjects complained about the polygraph attachments; lO complained about their own physical or emotional condition. For example, com— plaints about the polvgraph attachments were statements dC.) 81‘ o m OH .m am 5H AQmuzv whomnnsm mQHhH o o N H a m 8 mumv mpoanSm HSMQPSHB is . . on we . 02 m . oz mHzcmoon soapmsflamxm pmme Hosasmmm wmowmaso huhsm op mpmmswmm escapmmsa oa msamagaoo BDMWMDm modao Henno> macoqmpsomm Ho magma wanna go mpomWQSm mswmg was HSMQPSHB we momflnmmfioo < m manta 82 such as: "They're (the attachments) cold.", or, "The chair is uncomfortable." Complaints about the subject's own con- dition included the following: "I've got an awful lot of personal problems. Will that affect the test?"; or, "I've got hemorrhoids and have bad respiration." Table 5 shows that 5 of the lying subjects questioned the polygraph testing procedure; whereas, only 1 of the truthful subjects di‘ so. These questions generally con- cerned the number of individual tests to be given and when the testing results would be known. None of the truth- ful subjects asked the examiner to hurry the examination. However, 5 of the lying subjects told the examiner to hurry because they had an appointment elsewhere. To answer the question of whether the lying subjects exhibited more spontaneous verbal clues than the truthful subjects, all such clues, in each of the three classi- fications, were grouped together for each of the subject categories. The results of this grouping are shown in Table 6; 3 truthful and 25 lying subjects exhibited Spon- taneous verbal clues. The significanceof the difference between the truthful and lying subjects in respect to such clues was tested.1l 2 x 2 contingency table was pre- pared and a chi-square test of independence was run. The 83 Hosea Ho. esohmp pascawfismflm** mpflssflpsoo How soapooamoo mmpmw spasm mm mm omumv .sfimg **bm.am r as m Aomnzv stgpssm meSHo mesao aspam> muomsspsoam maemsmpsomm Na msfipapaaxm poz meflpapfiasm smegmaao m mpomnpsm Hmpoe whommpsm Hepoa Homembm mpafimom mSHo Hennm> muomsmpsomm weQSOHw o manna 84 results of this test, also shown in Table 6, indicate significance at the .01 level. The lying subjects did exhibit a significantly greater number of Spontaneous clues than the truthful subjects. However, as can be seen from an analysis of the table, the absence of Spontaneous clues, while it may give some weight in the direction of the truthful subjects, is not conclusive. The lying subjects were just as likely as not to exhibit such clues. MOHVEHBAL CLUE RESULTS Nonverbal behavior clues are recorded on interview sheets without regard to any standardized "check off" system. Because of this it was considered inadvisable to separately analyze the results of each of the eighteen nonverbal clue descriptors. However, in order to provide some indication of the relative frequency of occurrence of the "typical truth—teller" and "typical liar" descrip- tors for both categories of subjects, Tables 7 and a were prepared. Table 7 shows the number of times each of the nine "typical truth-teller" descriptors occurred for the truth- ful and the lying subjects. Table 8 shows the relative occurrences of the "typical liar" descriptors for both categories of subjects. In viewing each of these tables 85 cm H m N pH N v m e m Amwnuv mpoomnsm msflhA oma e e h ma ma mm on mm mm Aswnrv mpomrnsm adwapsam chasm powpsoo mossnsmaaml Inomooo counsom hacsmanm mumsqu ohm m>Hp m.noaamp Hosea eonuaam passe cauomaoo samnasqac pomuan coca .mxawa unease I $8.430 ecmwmbm mmoamHmUmmn MDHO HdMMszoz :mmnnmalmabma AdonMB: unopmanomon undo aspnobnoz unmaamauapsna HwOflaha= no 333% mass can H3539 no 82.838 b oHAda 86 eNH N m mH u mH eN wN NN Am¢uzv unconnsm mfisg HH 0 a N H o o N H Ap¢uzv upoonpsm Hsmnvsna mumm. muses eprs Indomgd umboz podpnoo Inmmooo thsoHnH HsHomm hHHoom mnm3m:4 ohm chap conundomnd Hspoe Is: condom maobnoz msobhmz mbamdbm Hoom lstmpnb n.9ddn ullullu Hmoanadc HomhMDm mmoemHMUmmn HDAO H4Mdfl>zoz :MdHH HmH H0. .620th He .pm HQmOHstmHm** HpHssHpsoo How SOHPoeHHoo mapsw mpHBs I. 'I’ "i!!! $*¢.®H mm 0% HH 0 Amfiumv mpoempzm mstH **O.¢m m m we vs Abanmv mpooWQSm Hammpsae .... pl . on.” pm $4 . om mmSHo . x Htpho>Som no mdeo Hdphmbsom so am =Haag Hmoaaaaz =Hmafimpngpsne Hausasc= smoaaeao emHMHmmMHo oeHmHmmmHo 90MNMDm meSHo HmQHe>soz so mpemmndm we SOHPQOHmemmHo Mo mPHSwem m mHQmm Chapter 6 DISCUSSION The elicited verbal clue results suggest that the interview questions are more effective on truthful subjects than on lying ones. Of course, these results could have been influenced by many factors, such as, differences in the sample subjects, differences between the examiners, etc. The results also could have been a product of the methodology. All recorded verbal answers which were difficult to classify were intentionally classified as "typical truth-teller" answers. This procedure could have led to the spurious classification of some of the lying subjects' answers; thus, for these subjects the results may have been skewed in favor of a "typical truth—teller" finding on some questions. While three of the interview questions were shown to discriminate between the truthful and the lying sub- jects, this result may actually have no practical signifi— cance. There are two reasons for this. First, such in- dividual question discrimination is not the examiner's basis for deciding the subjects' truthfulness and deception, Thus, examiners may actually use some of the questions only H) ‘ u S 90 to get an indication of truthfulness rather than an indi- cation of deception. In fact, the results would seem to bear out that such use of the questions may be an im- portant factor in the interview. For example, the "sus- picion," "who not suspect" and "think stolen" questions all gave significant results in the right direction on the lying subjects; none of them did so on the truthful subjects. A reason for this finding could be that the "correct" answers to these questions are always known to the lying subject but may not be known to the truthful subject. The lying subject always knows who did, or did not, commit the offense; he always knows if the missing items were actually stolen, etc. This knowledge may cause him to give false or misleading answers to these questions since the true answers will incriminate him. On the other hand, the truthful subject may not have any of this knowledge; if he does, he may be reluctant to reveal it. The examiner, recognizing that both the lying and truthful subjects may deny their "suspicion", etc. possibly only pays attention to the answers to these questions when they are affirmative, i.e., when the subject tells of his suspicions or knowledge of the offender. A second reason that individual question discrimina- tion may be unimportant when applied to actual polygraph \0 H examinations is that the verbal answers occur in combina- tion with nonverbal clues. In other words an examiner is attentive not only to what a subject says, but also, how he says it. For example, even though the results show that both truthful and lying subjects are likely to deny their suspicion of the offender, it is quite possible that they give such a denial in completely different ways. As the truthful subject answers, he may be very direct, and "look truthful," etc. The lying subject, however, recognizing the implication of a truthful answer, may stall, falter, or give other nonverbal clues to his deception. In other words, particular questions may have practical value be- yond their usefulness in eliciting verbal clues. The subject's verbal answers may be used as a cue to the ex— aminer to be more, or less, attentive to nonverbal clues. The spontaneous verbal clue data suggest that Reid's prior finding may be correct: truthful subjects do not seem to exhibit such clues as often as lying subjects. However, spontaneous clues are dependent upon the subject doing some- thing beyond what it would seem most subjects do (only 28 percent of the total sample subjects exhibited spontaneous clues); therefore, it seems likely that future research, using a larger sample of subjects than the present study, could lead to more meaningful data in this area. 92 The results of the present study as regards non- verbal behavior clues seem to substantiate Reid's research. The nonverbal clues which Reid claimed were common to truthful subjects were, in fact, exhibited by the present truth-teller sample more often than were those clues Reid claimed were common to lying subjects. Also, the reverse of this seems to be true; the present lying subject sample exhibited Reid's "liar clues" more often than his "truth- teller" clues. One important qualification of the nonverbal clue results which deserves discussion is that of the effect of examiner bias on the original data. While it was not possible to determine the extent of this effect, and, while it was assumed that it did not occur, it seems likely that such bias could have influenced the results. For example, in actual examinations the ex-tiner may observe that a subject cannot "look him in the eye," but, he may also observe that the subject seems shy - a possible explanation for his "poor eye contact." When the examiner can in his own mind explain a certain behavioral trait (or believes that he can) he may neglect to note that trait on the interview sheet. Thus, the only nonverbal clues which are noted are those which the examiner cannot explain, or per— haps, those which he feels are not in conflict with his 95 judgment of a certain trait's meaning. Rosenthal's work, which deals with the effect of experimenter bias on re- search results, would seem to confirm the possibility of examiner bias on the results of the present study.85 It is possible that some of the descriptors which were used as a basis for classifying subjects on nonverbal behavior were of little actual value to the classification. For example, two of the "typical truth-teller" descriptors "composed" and "relaxed" were observed in almost equal numbers of truthful and lying subjects. This finding would seem to conflict with Reid's data. However, there may be an explanation for this. The present research was based primarily upon observation of subjects prior to actual testing; whereas, Reid's study may have been based upon observation over an entire examination. Lying subjects are said to become increasingly nervous as the examina- tion progresses; truthful subjects, less nervous. Thus, an appraisal over an entire examination might well indicate that truthful subjects are more composed and relaxed than lying subjects. 85Robert Rosenthal, Experimenter_Effects In Behavioral Research (New York: Appleton-Century-Crofts, 1966)- 94 In general, the results of the present study suggest that both the verbal and nonverbal behavior of truthful polygraph subjects differ in significant, observable respects from those of lying subjects. Because of this, it does seem that behavior data can be a source of useful informa- tion in the diagnosis of truth and deception. This does not mean, however, that behavioral data are substantial enough to replace polygraph record interpretation, which, even without such data, can be a highly accurate means of diagnosing truth and deception.84 What it does mean is that the study of polygraph subject behavior has progressed beyond the point of being useful merely in helping an ex- aminer determine how a subject should be "handled," as Reid suggested in 1353. Behavior data would seem to be useful as an aid to the diagnosis of truth and deception, 04See, for example: Frank S. Horvath and John E. Reid, "The Reliability of Polygraph Examiner Diagnosis of Truth and Deception," The Journal of Criminal Law; Crimino- lognyand Police Science,_62, 2 (June, 1971), pp. 276- 281; also see the research on experimental "lie detection," for example: David T. Lykken, "The GSR In the Detection of Guilt," Journal of Applied Psygholqu, 43 (December, 1959) pp. 385-388; Joseph Kubis,5Anal sis of Pol sra hic Data, Technical Documentary Report No. ’ - - 4 lOI (New York: Rome Air Development Center, 1965). 95 as long as they are regarded as a source of hypotheses to be checked against other data within the examiner's know- ledge, such as his polygraph record interpretation. When agreement, or convergence, emerges, the examiner's confidence in the validity of his data, and his diagnosis, increases. When such agreement is not apparent, the examiner can use this as a cue to reassess his data and perhaps, be cautious in his diagnosis. It is left to future research to determine more precisely than did the present study how behavior differs between the truthful and lying subject. Also, future work could emphasize the development and refinement of present techniques of eliciting, observing and recording polygraph subject behavior, especially nonverbal behavior. It is possible that, as Ekman and Friesen found in their studies of psychotheraphy subjects,85 certain areas of the body are more important sources of behavior clues toiruth and deception than others. Moreover, it is not without 05Paul Ekman and Wallace V. Friesen, "Nonverbal Leakage and Clues to Deception," Ps*chia§£y, 52, (1969), pp. 88-106; Paul Ekman and Wallace V. Friesen, "Nonver- bal Behavior in Psychotheraphy Research," in John Shlein (ed.) Research In Psychothepaphy, Vol. 3. (Washington, D.C.: American Psychological Association, 1968). 96 reason to suggest that future work could lead to an accurate, practical, standardized system of recording behavior data for the use of all polygraph examiners. Perhaps, Birdwhis- tell‘s kinesics signs could be a model for this system.36 Regardless of the direction future research takes, further investigation of polygraph subject behavior is warranted. 86An ex mple of these signs appears in: Julius Fast Body Language (New Yo k: M. Evans and Co., Inc., 1970 ,pp. 1664168. Chapter 7 SUMMARY Polygraph examiners frequently disagree about whether or not judgments of polygraph subjects' behavior should be used in making truth and deception decisions. On the one hand, some examiners insist that behavior judgments should not be used. On the other hand, some examiners claim that behavior judgments should be used because they provide important information useful in dia nosing truth and deception. This study was conducted to determine whether poly- graph subjects' verbal and nonverbal behavior provides clues to the subjects' actual truthfulress and deception. Ver- bal behavior was broken down into two categories: elicited and Spontaneous. Elicited verbal behavior was defined as the subjects' verbal answers to a set of standardized structured pro-test interview questions. These questions, asked of polygraph subjects prior to actual testing, are said to elicit verbal clues to the subjects' truthfulness and deception. There are fourteen such questions, altnough not all subjects are asked all questions, depending upon the particular facts of the offense being investigated. k0 Q 98 Spontaneous verbal behavior was defined as those comments, complaints, and requests which a subject makes during a polygraph examination without direct prompting from the examiner. Nonverbal behavior was defined as the actions, appearance, and demeanor of the subject during the pre-test interview and the examination as observed by the examiner. In this study the verbal and nonverbal behavior of fifty verified truthful and fifty verified lying sub- jects was analyzed. This analysis was made by reviewing the interview sheets of these subjects. The interview sheets contain all data pertinent to the behavior and the examination for each subject. In analyzing elicited verbal behavior the actual verbal answer given by each subject to the structured pre- test interview questions was classified as being either a "typical truth-teller" or ”typical liar" answer, according to accepted theory. The number of classifications in either of the two "typical” groups for all interview questions and for both the truthful and lying subjects was compared to the number exoected on a chance basis. The questions to be answered by conducting this classification were: Did a significant number of the actual truthful subjects give more "typical truth-teller" than "typical liar" answers to each of the interview ques- tions; and, Did a significant number of the actual lying 99 subjects give more "typical liar" than "typical truth-teller" answers to the interview questions. It was further hypothe- sized that if these former two questions were affirmed for any of the interview questions then.those interview questions discriminated between the truthful and lying subjects. The results indicated that ten of the interview questions gave results in agreement with the hypothesis for the truthful subjects; only six interview questions gave results in agreement with the hypothesis for the lying subjects. In other words, a significant number of the truthful subjects gave more "typical truth-teller" than "typical liar" answers to ten of the interview questions; a significant number of the lying subjects gave more "typical liar" than "typical truth-teller" answers to six of the questions. Only three of the interview questions were found to give results in agreement with the third elic- ited verbal behavior hypothesis, i.e., where truthful subjects gave "typical truth-teller" and lying subjects gave "typical liar" answers in significant numbers to the same interview question. Based upon prior research on polygraph subject behavior it was hypothesized that lying subjects exhibit spontaneous verbal behavior to a significantly greater degree than truthful subjects. This hypothesis was tested by determining the frequency of occurrence of such behavior 100 for both the truthful and lying subject samples and then conducting a statistical analysis of the data to test for the significance of the difference between the truthful and lying subjects. This analysis revealed that the hypothesis must be accepted. The lying subjects did exhibit spontaneous verbal behavior to a significantly greater degree than did the truthful subjects. Only six percent of the truthful subjects exhibited such behavior; whereas, fifty percent of the lying subjects did so. Prior research results were used as a basis for de— veloping nonverbal behavior descriptors which would provide a means of classifying the present subject samples on their nonverbal behavior. A series of nine descriptors was de- veloped which, according to prior research, were supposed to be typical of truthful subjects. An equal series of descriptors was develOped whichxwmnasupposed to be typical of lying subjects. Then, the nonverbal behavior of the two actual subject samples was converted into the descrip- tors. All the actual subjects were classified as being either "typical truth-tellers" or "typical liars" by allow- ing one point for each descriptor in either of the "typical" groups for each subject. That group receiving the most points was used as a basis for classification. 101 The classification of subjects on nonverbal behavior provided a basis for answering the two following questions: Did a significant number of the truthful subjects exhibit the "typical truth-teller" nonverbal descriptors more often than the "typical liar" descriptors?; Did a significant number of the lying subjects exhibit the "typical liar" descriptors more often than the "typical truth-teller" descriptors? The results indicated that a significant number of the truthful subjects, 94 percent, exhibited more "typical truth-teller" than "typical liar" descriptors. A signifi- cant number of the lying subjects, 82 percent, exhibited more "typical liar" than "typical truth-teller" descriptors. The results suggest that the verbal and nonverbal behavior of truthful polygraph subjects differs in signif- icant, observable respects from those of lying subjects. Further investigation of this phenomenon is warranted. BIBLIOGRAPHY BIBLIOGRAPHY BOOKS Ardrey, Robert. The Territorial Imperative. New York: Atheneum, 1970. Arther, Richard O.,and Rudolph Caputo. Interrogation For Investi ators. New York: William C. capp and Associates, 1959. Aubry, Arthur, and Rudolph Caputo. Criminal Interrogation. Springfield, Ill.: Charles C. Thomas, 1965. Biderman, Albert D., and Herbert Zimmer (eds.). The Manipulation of Human Behavior. New York: Jo Wiley and Sons, 1961. Borg, Walter R. Educational Research An Introduction. New York: David McKay Company,Tnc., I965. Bruner, J. S., and R. Taguiri. "The Perception of People," Handbook of Social Psychology, ed. G. Lindzey. Reading,Mass.: Addison-wesley,‘1954. pp. 634-54. Dienstein, William. Techniques for The Crime Investigator. Springfield: ChafI SOC. Thomas, 1952. Ekman, Paul, and Wallace V. Friesen. "Nonverbal Behavior In Psychotherapy Research," Research In Psychotherapy, ed. John Shlein. Washington, D.C.: American PSyeho- logical Association, 1968. Fast, Julius. Body Language. New York: M. Evans and Company, Inc., 1970. Ferguson, Robert J. Jr. The Pol a h In Private Industr . Springfield, Ill.: Charles %. Thomas, 1955. . The Scientific Informer. Springfield, Ill.: Charles C. Thomas, 1971. Hall, Edward T. The Silent Language. Garden City, N.Y.: Doubleday and Company, Inc., 1959. 103 104 Hamburg, Morris. Statistical Analysis For Decision Making. New York: Harcourt, Brace and World, Inc., 1970. Inbau, Fred E. Lie Detection and Criminal Interro ation. Baltimore: The Williams and Wilkins Company, §942. . Lie Detection and Criminal Interrogation. Baltimore: The Williams and Wilkins Company, 1948. and John E. Reid. Lie Detection and Criminal 9 Interro ation. Baltimore: The Williams and Wilkins Company, 1953. , and John E. Reid. Criminal Interrogation and onfessions. Baltimore: The Williams and Wilkins CEmpany, 1967. Janis, Irving, G. Mahl, J. Kagan, and R. Holt. Personality, Dynamicsl Development and Assessment. New York: Harcourt, Brace and Worid, Inc., 1969. Kidd, W. R. Police Interrogation. New York: The Police Journal, 1940. Koenker, Robert H. Simplified Statistics. Bloomington, Ill.: McKnight and McKnight Publishing Company, 1961. Larson, John A. Lying And Its Detection. Chicago: University of hicago ress, 1932. Lee, Clarence D. The Instrumental Detection of Dece tion. The Lie Test. Springfield, Ill.: Charles C. Thomas, 1953. Leonard, V. A. (ed.). Academ Lectures on Lie Detection. Springfield, Ill.: Charles C. Thomas, 1957. . (ed.). Academy Lectures on Lie Detection, Volume ' II. Springfield, Ill.: Charles‘C. Thomas, 1958. Luria, A. R. The Nature of Human Conflictg: On Emotion, Conflict, andJWill. trans. and ed. W. H. Gantt. New York: Liveright, 1932. O'Hara, Charles C. Fundamentaksof Criminal Investigation. Springfield, Ill.: Charles C. Thomas, 1956. Orne, Martin T., R. I. Thackray, and D. A. Paskewitz. "On the Detection of Deception: A Model of the Study of the Physiological Effects of Psychological Stimuli." To appear in N. Greenfield and R. Sternbach (eds.) Handbook of Psychophysiology, in press. Reid, John E. and Fred E. Inbau. Truth and Deception, The Polygraph ("Lie Detector") Technique. Baltimore: The Williams and Wilkins Company, 1966. Rosenthal, Robert. Experimenter Effects In Behavioral Research. New York: Appleton-Century—Crofts, 1966. Selltiz, Claire, Marie Jahoda, Morton Deutsch, and Stuart Cook. Research Methods In Social Relations. Holt, Rinehart and Winston, 1962. PERIODICALS Adams, J. L. "Laboratory Studies of Behavior Without Awareness," Psychological Bulletin, Vol. 54, No. 5, (1957). pp. 385-405. Alpert, M., R. L. Kurtzberg, and A. J. Friedhoff. "Transient Voice Changes Associated with Emotional Stimuli," Archives of General Psychiatry, Vol. 8, (1963), pp. 36?-3650 Ax, Albert F. "The Physiological Differentiation Between Fear and Anger in Humans," Psychosomatic Medicine, Vol. 15, NO. 5, Sept.-Oct.,91953, pp. 453-442. Backster, Cleve. "Lie Detection Comes of Age," Law and Order, Vol. and date unknown, reprint supplied to the writer by Mr. Backster, January,l97l. Berrien, F. K. "Ocular Stability in Deception," Journal of Applied Psychology, Vol. 26,(l942), pp. 55-65. Bersh, Philip J. "A Validation Study of Polygraph Examiner Judgments," Journal of Applied Psychology, Vol. 53, No. 5, (1969). pp. 399-403. Block, Jack A. "A Study of Affective ReSponsiveness in a Lie-Detection Situation," Journal of Abnormal and Social Psychology, Vol. 55, T195777 pp. 11—I5. 106 Burtt, H. E. "Motor Concomitants of the Association Reaction," Journal of Experimental Psychology, V01. 19’ (19367, PP. 51-630 Crosland, Harold R. "Measurements of Emotion By a Method Which Combines the Word-Association, Reaction-Time Techniques with the Psychogalvanic Technique," Psychological Bulletin, Vol. 28, (1931), p. 575. Cureton, Edward E. "A Consensus As to the Validity of Polygraphic Procedures," Tennessee Law Review, Vol. 22, February, 1953, pp. 728-712} Davidson, P. 0. "Validity of the Guilty-Knowledge Tech- nique: The Effects of Motivation " Journal of Applied Psychology, Vol. 52, No. 1, (19685, pp. 62-65. Davis, R. C., A. M. Buchwald, and R. W. Frankmann. "Autonomic and Muscular Responses and Their Relation to Simple Stimuli," Psychological Monographs, Vol. 69, (1955). pp. 1-71. Davis, R. C., and A. M. Buchwald. "An Exploration of Somatic Response Patterns: Stimulus and Sex Differ- ences," Journal of Com arative and Ph siolo ical Psychology, Vol. 50, (1957), pp. 44-50. Davis, R. C. "Somatic Activity Under Reduced Stimulation," Journal of Comparative and Physiological Psychology, 010 9 on 3, une, 59, Pp. 309-314. Dearman, H. B., M.D. and B. M. Smith, Ph.D. "Unconscious Motivation and the Polygraph Test," American Journal of Psychiatpy, Vol. 119, No. 11, May, 1963, p. 1017. Dienstein, William. "The Importance of Words in Question- ing Techniques," Police, Vol. 1, No. 4, March-April, 1957’ PP. 47-51. DiMascio, A., R. W. Boyd, and M. Greenblatt. "Physiological Correlates of Tension and Antagonism During Psychotheraphy, A Study of 'Interpersonal Psychology,'" Psychosomatic Medicine, Vol. 19, (1957). pp. 99-104. Duncan, Starkey, Jr. "Nonverbal Communication," Psychological Bulletin, Vol. 72, No. 2, (1969), p. 118- 137. 107 Ekman, Paul and Wallace V. Friesen. "Nonverbal Leakage and Clues to Deception," Psyghiatry, Vol. 32, No. 1, (1969), pp. 88-106. English, Horace B. "Reaction-time Symptoms of Deception," The American Journal of Psychology, Vol. 57, No. 3, 1113', 192 9 PP. 428- 90 Fay, P. J. and W. C. Middleton. "The Ability to Judge Truth-telling or Lying, From the Voice as Transmitted Over a Public Address System," Journal of General Psychology, Vol. 24, (1941), pp. 211-215. Floch, Maurice. "Limitations of the Lie Detector," Journal of Criminal Law - Criminologl. Vol. 40, No. 5. Jan.-Feb., 1950, pp. 651-653. Gleser, G. C., L. A. Gottschalk, and W. John. "The Rela- tionship of Sex and Intelligence to Choice of Words: A Normative Study of Verbal Behavior," Journal of Clinical Ps cholo , Vol. 15, No. 2, April, 1959. pp. 182-191. Goldstein, Eva R. "Reaction Times and the Consciousness of Deception," The American Journal of Ps cholo , Vol. 34, No. 4, ct., 19 3, pp. 5 2-81. Guertin, Wilson H. and Paul L. Wilhelm. "A Statistical Analysis of the Electrodermal Response Employed in Lie Detection," The Journal of General Psychology, Vol. 51, (1954). pp. 155-160. Gustafson, L. A. and M. T. Orne. "Effects of Heightened Motivation on the Detection of Deception," Journal of Applied Psychology, Vol. 47, (1963), pp. 408-411. and . "Effects of Perceived Role and Role Success on the Detection of Deception," Journal of A lied Ps cholo , Vol. 49, No. 6, December, 1965, pp. 412-417. and . "The Effects of Task and Method of StImulus Pfesentation on the Detection of Deception," Journal of Applied Psychology, Vol. 48, No. 6, Dec., 9 49 PP. 38 “387. and . "The Effects of Verbal Responses on the Laboratory Detection of Deception," Ps cho- physiology, Vol. 2, No. 1, (1965), pp. 10-15. 108 Harman, G. W., and J. E. Reid. "The Selection and Phrasing of Lie-Detector Test Control Questions," Journal of Criminal LawE Criminology and Police Science, 701. 46, O. 4’ NOV.- 60., 195 9 PP. 578-582. Heckel, R. V., J. R. Brokaw, H. C. Salzburg, and S. L. Wiggins. "Polygraphic Variations in Reactivity Between Delusional, Non-Delusional and Control Groups in a 'Crime,'" Journal of Criminal Law, Criminology and Police Science, V01. 53, No. 3, September, 1962, pp. 380-383. Horvath, Frank S. and John E. Reid. "The Reliability of Polygraph Examiner Diagnosis of Truth and Deception," The Journal of Criminal Law, Criminology and Police Clence’ Ol. , NO. , 1.1113, 1 71, pp. 7 - . and . "The Polygraph Silent Answer Test," The Journal of Criminal LawL Criminology, and Police 'Science, 1972, in press. Inbau, Fred E. "The Lie Detector," Boston Universit Law Review. Vol. XXVI, No. 2, April, 1946, pp. 264-27 . and John E. Reid. "The Lie Detector Technique: I Reliable and Valuable Investigative Aid," American Bar Association Journal, Vol. 50, No. 5, May, 1964, p.. 470} "Lie Detector Test Limitations," Reprinted from the Jgurnal of Forensic Sciences, Vol. 2, No. 3, July, 1957. pp. 255-262. . "Scientific Evidence in Criminal Cases: II Methods of Detecting Deception," Journal of Criminal Law and Criminology, Vol. 24, No. 6,‘Mar-Aprii, 1954, Pp. 11.40-1158. Iwahara, Shinkuro, Kyoko Miseki, Noriko Shiokawa and Reiko Yoshida. "Uso Ni Taisuru GSR To Seikaku Tokusei Tono Kankei Ni Tsuite (GSR to Forced Lying and Personality Traits)," Tohoku Journal Experimental Psychology, Vol. 1, No. 3,‘(1955), pp. 82-85. Jaffee, Cabot L., Edward Millman and Bernard German. "An Attempt to Condition an Eyeblink Response to Verbal Deception," Psychological Reports, Vol. 19, No. 2, (1966), pp. 421-422. I 109 Jones, Harold E. "The Galvanic Skin Reflex as Related to Overt Emotional Expression," American Journal of Psychology, Vol. 47, No. 2, April, 1935, pp. 241-251. and D. Wechsler. "Galvanometric Technique in Studies of Association," American Journal of Psychology, Vol. 40, (1928), pp. 607-612. Jung, C. G. "The Association Method," American Journal of Psychology, Vol. 21, (1910), pp. 219-269. "On Psychophysical Relations of the Associative xperiment," Journal of Abnormal Psychology, Vol. 1, (1907), pp. 247-255. Kubis, Joseph F. "Instrumental, Chemical and Psychological Aids in the Interrogation of Witnesses," Journal of Social Issues, Vol. 13, No. 2, (1957), PP. 40-49. . "Experimental and Statistical Factors in the Diagnosis of Consciously Supressed Affective Experience," Journal of Clinical Psychology, Vol. 6, (1950), pp. 12- 16. ‘ Kugelmass, Sol and I. Lieblich. "The Effects of Realistic Stress and Procedural Interference in Experimental Lie Detection," Journal of A lied Ps cholo , V01. 50, No. 3, (1966), pp. 211-216. , , and Zeev Bergman. “The Role of Lying in Psychophysiological Detection," Psychophysiology, Vol. 5, No. 3, (1967), pp. 312-315. and . "Relation Between Ethnic Origin and GSR Reactivity in Psychophysiological Detection," Journal of Applied Psychology, Vol. 52, No. 2, (1968), pp. 156-162. Lacey, J. I. and B. Lacey. "Verification and Extension of the Principle of Response Stereotypy," American Journal of Psycholo , Vol. 71, (1958), pp. 50-73: Landis, 0. and W. A. Hunt. "Word Association-Reaction Time and the Magnitude of the Galvanic Skin Response," American Journal of Psychology, Vol. 47, £1935), PP. 14 3-115 0 Langfeld, H. S. "Psycho-Physical Symptoms of Deception," Journal of Abnormal Ps cholo , Vol. AV, Nos. 5 and 6, Dec., 1920 and March, 1921, pp. 519-28. llO Larson, John A. and George W. Haney. "The Polygraph and Personality Study," American Journal of Psychiatry, Vol. XI, No. 6, May, 1932. Lazarus, R. S. and R. McCleary. "Autonomic Discrimination Without Awareness: A Study of Subception," Psychological Review, Vol. 58, (1951), pp. 113-122. , M. Tomita, E. Opton and M. Kodama. "A Cross- Cultural Study of Stress-Reaction Patterns in Japan," Journal of Personality and Social Psychology, Vol. 4, 00 9 19 9 pp. 2 - 33¢ Lykken, David T. "The GSR in the Detection of Guilt," Journal of Applied Psychology, Vol. 43, Dec., 1959, pp. 385-388. . "The Validity of the Guilty Knowledge Technique: The Effects of Faking," Journal of Applied Psychology, Vol. 44, (1960), pp. 258:262. Malmo, Robert B., Thomas J. Boaz and A. A. Smith. "Physio- logical Study of Personal Interaction," Psychosomatic Medicine, Vol. 19,No. 2, (1957), pp. 105-119. , C. Shagnass and F. H. Davis. "Symptom Specificity and Bodily Reactions During Psychiatric Interview," Psychosomatic Medicine, Vol. 12, (1950), pp. 362-376. Marston, W. M. "Negative Type Reaction-Time Symptoms of Deception," Psychological Review, Vol. 32, No. 3, May, 1925, pp. 241-247. "Reaction-time Symptoms of Deception," Journal of Experimental Psychology, Vol. 3, (1920), pp. 72-87. Martin, Barclay. "Galvanic Skin Conductance as a Function of Successive Interview," Journal of Clinical Psychology, Vol. 12, No. 1, Jan., 1956, pp. 91-94. Morgan, Mildred I. and R. H. Ojemann. "A Study of the Luria Method," Journal of Applied Psychology, Vol. 26, (1942), pp. 168-179. Olechowki, Richard. "Experimente Ober Den Stimmund Sprechausdruck Bein Lugen (Experiments on Voice Modula- tion While Telling Lies)," Zeitschrift fur Experimentelle und An ewandte Ps cholo ie, Vol. 14, No. 3, (1967), pp. 474-485, Psychological Ebstracts, Vol. 42, April, 1968, p. 484. 111 Orne, Martin T. and Richard I. Thackray. "Group GSR Tech- nique in the Detection of Deception," Perceptual and Motor Skills, Vol. 25, (1967), pp. 809-816. Reid, John E. and Richard O. Arther. "Behavior Symptoms of Lie Detector Subjects," Journal of Criminal Law, Criminology and Police Science, VOl. 44, No. 1, EEWCJune, 1953, pp. 104-108. . "A Revised Questioning Technique in Lie-Detection Tests," Journal of Criminal Law Criminolo and American Police Science, Vol. XXXVII, No. 6, March-April, 1947, pp. 542-547. Rose, A. "Psychophysical Tests for Falsehood," Scientific American, Vol. 110, June 20, 1914, p. 502. Ruesch, Jurgen. "Nonverbal Language and Therapy," Ps chiatr , Vol. 18, No. 4, November, 1955, pp. 323- 330. Runkel, J. E. "Luria's Motor Method and Word Association in the Study of Deception," Journal of General Psychology, Vol. 15, (1936), pp. 23-37. Shakhar, Gerson Ben, Israel Lieblich and Sol Kugelmass. "Guilty Knowledge Technique: Application of Signal Detection Measures," Journal of Applied Psychology, Vol. 54, NO. 5’ (1970), pp. 409-113. Sternbach R. A., L. A. Gustafson and R. L. Colier. "Don't Trust the Lie Detector," Harvard Business Revigw, Vol. 40, No. 6, Nov.-Dec., 1962, pp. 127-134. Suzuki, Sadao, Takami Watanabe. and Koichi Shimizu. "Effects of Visual Feedback by the Skin Potential Response," Japanese Journal of Psychology, Vol. 40, No. 2, (1969), pp. 59-67. Trovillo, Paul V. "A History of Lie Detection," The Journal of Criminal LayyCriminology, and Police Science, Vol. ‘29, March-April, 1939, pp. 848-881 and May-June, 1939, PP. 104-119 . van Buskirk, D. and F. L. Marcuse. "The Nature of Errors in Experimental Lie Detection," Journal of Experimental Psychology, Vol. 47, (1954), pp. 187-190. 112 Wenger, M. W., T. L. Clemens, D. R. Coleman, T. D. Cullen, and B. T. Engel. "Autonomic Response Specificity," Ps chosomatic Medicine, Vol. 23, No. 3, (1961), pp. TgS-l93. Winter, J. E. "A Comparison of the Cardiopneumopsychography and Association Methods in the Detection of Lying in Cases of Theft Among College Students," Journal of Applied Psychology, Vol. 20, (1936), pp.fi243+248. Yankee, William J. "An Investigation of Sphygmomanometer Discomfort Thresholds in Polygraph Examinations," Police, Vol. 9, No. 6, July-August, 1965, pp. 12-18. PUBLICATIONS OF THE GOVERNMENT Ellson, D. G.,R. C. Davis, I. J. Saltzman and C. J. Burke. A Report of Research on The Detection of Dece tion. Prepared for Office of Naval Research, Contract N6 onr- 18011, Department of Psychology: University of Indiana, 1952. Kubis, Joseph F. Analysis of Polygra hic Data. Technical Documentary Report No. AD - R-EE-lOl. New York: Rome Air Development Center, Griffis Air Force Base, January, 1965. . Studies In Lie Detection: Com uter Feasibilit Considerations. Technical Report No. RADC-TR-62-2OS. New York: ’Rome Air Development Center, Griffis Air Force Base, June, 1962. Kugelmass, Sol and Israel Lieblich. Some Correlates of Orientinngesponse Indices In Experimental Lie Detec- tion. Contract AF61-(052)-839. Air Force Office of Scientific Research, Nov. 1, l964-April 30, 1965. Orlansky, Jesse. An Assessment of Lie Detection Capability Declassified version). Technical Report 62-16, pre- pare or the Institute of Defense Analyses, Research and Engineering Support Division, July, 1964. U.S. Congress, House, Committee on Government Operations, Foreign Operations and Government Information Subcommittee; Use of Polygraphs As "Lie Detectors" By the Federal Govern- ment, Hearing, 89th Congress, 2nd sess., April 7, B, and 9, 1964, Part 1, Panel Discussion with Private Polygraph Practitioners. Washington, D.C.: U.S. Government Print- ing Office, 1964. 115 Violante, Robert and Sheela Ross. Research on Interroga- tion Procedures. Contract Nonr 4129 (DO), MEnlo Park, Calif.: Stanford Research Institute, May lS-October 31, 1964. Zimmer, Herbert. Psychophysiologic Variables As Indica- tions of Emotional Stress. Technical Report No. -TR- 5-29 , ew ork: Rome Air Development Center, Griffis Air Force Base, September, 1966. APPENDICES APPENDIX A: SECTION 1 MASTER DATA SHEETS: ELICITEI) VERBAL CLUES .tionx 116 VERIPISD mururux. SUNECTS tuna latex-V1“ 8% MASTER DATA SHEET: ELIGITED VERBAL BLUES 78 18 M 82 70 30 '06 87 1 3 10 13 100 69 h') u 31 67 1 3 33 9h 12 67 33 83 17 59 53 ‘07 , 2h 62 ‘07 68 | DEAF OIlu Axf.l"9.Exocf.9.0.9..-axis!-m.1.m...o.m.o.-x-omxoomxma'.9.1.oxmxmx-xo.oao.m.u.0.9.m.m.v.0ooxoonanoax I 2 8 173 5 05 31 8..- 7. 3. 3 .2 296m622§2251w12wmwnn azewwnsnnfinamwawsafimawwmmfiww . III 3 8L Lxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx W .1 xx xxxx x x x xxx xx xxx xx xx x x xx x x "on n x x x x x x xx x x xx xx xxx x xxxxxx ma Coax—L 232233302121152652320222h1h3132h32h012122h23232320 m AB; 1 =25: 816509O121705635667697987679088076k9170o6699991778 .u I. 1 1111 1 1 I. I. la y 99‘ 0.1.4 0&873231h2915151180899n01912110h08392v1602121...9u9d 0 11 111111 1 1 11 1 11 1111111 1 I. Igloo-All]. I. 1 a...) 5 H "H I X XX XX ‘ I a u . c533d£ wade: #H: ".m.e l "SOHpmmSG soaopn #QHQH :.pommm3m p.:6adoa H .H.e mamomg 039 Mo Pmoz .maohnw HommmSm P.fi©H503 H: u.H.e .a.a =.msopm wzw memm .memamm 90% noSob m.H= “.m.e ”SOHPmmsw pommmSm #02 0:2 =.ammp op mHoB mqwme .H.H 90H HngH .woom oop 5mg» gong H.som H .02: “.m.H .e.s :.pflsm> map :H 03» hago map mam: H wqw mama: ".m.a "soapmmSG nOHOHmmSm =.mHowmp ms @800 mng cm: m>.m3 .mnHoo Ho meHH haw **.H.H mmm H.goH@ H mH Ham gmo H mgHzp HHzo map .HHmH: “A.m.Hv HooHHSm mgHHH =.maomhs Hdmao op mH mum: aw H nommma pmHHw mge .mm:oa mqflmmfla wasp *H.H :sz om>Ho>gH Hog aw H Ham HHHngHsHH gmo H .oH: A.m.Hv HomHQSm Hzng5HH "moapmoSO Sow nOHpGOHHHmmeo mhm3mm< HmsPo< QOHpmmSG 3mH>HmpmH mZOHemMDO 3mH>MMEZH ammalmmm QMMDHUDMEm OB mmmzmz< H¢D90¢ QMBOQHMm ”HH ZOHBOMm 120 =.Momn pH P59 mhmsHm H .H.H .HHspm wad .mmxoo How mmnd£o cmzonnon H .smmwz ".moH .s.a =.Axdwn way aoHHV gm0H m pmsh .Hm>oq .02: .m.e "SOHmeSG BoHHom =.%Hsz0Hmwooo #H HQ 0% m2 .Aso on PamHe mquHmHmwQHH as ngpv mamgp Undead .H.H mmHzoma Nonmx m m.mnmge .950 Page madeH p.2Mo H= ".m.H =.#G£P 50% HHmp awe H .mnmnp pom mam hmza A.xop QmMO mnp so .a.e on op manHmHmmnHH he HoHv HH@ #6 nommma oz .02: ".m.e "SOHpmmsw PQHHmHmqum .H.H =.mHmH m>mg HHw m: .UHH HHm m3 mmmsm H .HHm2: ".m.H =.HoHHmp Hmzpogd swap Hmnpo mgo on an mmxomno m.pH dam 9:500 .H.a pHsm> mnp HOH meHmnommmH mam m3 .mHmHHmp was: ".m.e "QOHPmodg HHHQdPHommo wmmm =.mHm£ mxmpmHa m mp ©Hdoo pH .mmopnm>QH mno swap mHoa cm: m3 pmmm map nH .maHp paosm HHm> a HH oHSm HoH 302M HH.mz .msHmmHa m.pH Hgng H.:oo H .mngqsooom .H.H :H onHmHa H .quopm mm: pH anHH p.go© H .02: “.m.H qOHHMOHHHmmMHo mnmzmq< Hm590¢ QOHpmmSO BmeanmM 121 :.omHm manpom wad onH£owa map so mwQan madm pHommH HH.H wad mmanp omonp .H.H 50% oHop H .mxnoz pH Bo: Bong p.:oo H .HHmB: ".m.H :.pH ApHB .e.e ow op mQHmpom om: H Sogm HHHB pH 309% H: ”.m.a "QoHpmoSJ Hdow no pSo maoo .H.H 3oz .H.a :.Hoo£om EoHH mst .Smmh go: ".m.H .H.H :.po: Hoppop hone: u. .H U) “QoHpmoSQ flowchgm< :.oa :pHB 080: Ammnoa qunv maom 0H09 0» 0HHH 0.H .0:Ha 0H.50H moHHm pH .pmmgog .H.H on.:oh HH .onHE m.Hoonhao>o op deooo pH: ".m.H .H.H =.HH00H Hoop Homo: 000: p.:00 H .02: “.m.H ”:0Hpmoso Hnge :.A%mnoe Ho mmOHv mac anopm QH oobHob .H.H uoH :00p 0>.H .mnmg mer 0HHHP 0:H 0H 0H:H= ..m.H :.mHMoH m HoH Hogoa szB omMHoB o>.H .e.e cam MHoz HE pSon meOHpmosw soon Hobos obmn H: ".m.e "QoHpmodg oHoHom nmmmmm Hmbm nowprHHHmmeo mHmzmmd HwSpo¢ qumeSd Boflpnoan $3080 .8: H0033? HoBde HoHHoplnpSleHwommma* :.ma no .H.H :0HH00HH0H 0 m.pH .HHHQ0HH .HH 0HHH p.000 H: ".m.H :.mnohnm>m op mHsp 30am op paws H ogw HMoH .H.H 0p qugpon 0>0g H .HH 0H0p 0H Ham: H H00H H: ".m.H ":0Hpmmso 00SHHHH: .H.H :.HH H0 00Hm00 HH0 p0 H00H 0H 000: 0.H= ".m.H =03$H 039 %0 PQQP .H.e lam pmoHHSH map op oopSOomoHH .BmH op ms m.pH: ".m.s "QoHpmm5@ Hoom op mommmm oHSOAm pom? 0.9.:03 H .0p 0:0: p.:00 H HH 0pH oxwp op omHHsvoH H S< .pon Ho pmsp op momma o.H .H.H HH Song p.noo H wHonMpnoQ adHoom coma 50h om: «.m.H .H.H =.0H omHm 0: 0H003 H .m0H: ”.m.H "SoHpmosa asnom mpzhe noHpGOHHHmmeo mHmBmQ4 depo< QoHpmmdd 3mH>HmeH APPENDIX B MASTER DATA SHEETS: SPONTANEDUS VERBAL CLUES 124 .oopmHH mum mosHo Hmpnob mSoo:Monmm mQHpHpHQNo mpomHQSm omogp tho* o H N .28: H mm H 0m x 5 H55: .. monHmHomfl mzHfimzool: A. Howwmom DE mammDQWm I II." I! II" C. ‘0' ' macflmmbm Hbmmabma HMHmHmm> *mMDHO Rdmmm> mpomz mSoondonmm qupHpHflxm mpoomQSm mmogp hHQO* NNN PH N N NNNNM x xx NXNNNNN Hdeoa a: so NH mm mm Hm Om mm mm Hm om mH NH #H NH HH Mmmpm EFEESga mZOHamMDU mZHdHHSOO mBUMbmbm wZHNH QMHmHmm> *mMDHo H mbomz¢ezomm «Bmmmm ¢e¢m mmeméz APPENDIX 0 MASTER DATA SHEETS: NONVERBAL CLUES 127 '23 J.“ i .fifi. 9.! mm cum W but" fluorine" w W“ hurt". a! g 953.3588: a3§82l8£ :3 if: 3 up; i .ufi w. mnnnnn Banana nunfinnnnnnnnn nunnnn 110000 000001 00151000001000 000000 302833 161150 Ike’s-“533.53... 5"..11 12:.ssvoomnnunuunuuuuanaa:n unfinfinnnnnnnnfifln 9.000200100000000 0661‘EJ‘6’666‘6h flannlnflzfiflflawuunubwfluww “1'! LI. 11 100 128 I £16 mm mm: W flu “3,915 b .. . ...1 128 luv-m. CUJB mun! 3000:“; .23 Jan .3! 0'8 .hé no] "rm“: nu“ Wu .0!!- hdi gho- ul: '9‘.- tag 30.80.— WW" Mm wnnnuuunn: l‘lshssn": 0110002000 nnuannnnnnnnnnnn 1231‘1120'3222001 1.201123200022330 I I I I I I II I II I I I I II III I III I III I I I IIII I III I I I I IIII IIII I I I I ununnwnnnnnnnnnnnnnunun la‘lz‘ooasa‘hlaskzkhia“ 200°13‘8203001000100200 IIII II IIII IIIII I I II I I I II I I I II I I I I I II II I I II IIIII II IIIII I I II I III I I I I I I I III! I I II I III I II II 123.- 56 189mumnn um. Hmwmaana ”an flannnnfinfiflflnmuunubuflwww l 9 "fl - mica ram-um:- 19 125 2 3 222628 TL 0 ”’16“ “if APPENDIX D DISCUSSION OF ADDITIONAL SPONTANEOUS VERBAL CLUES Appendix D DISCUSSION OF ADDITIONAL SPONTANEOUS VERBAL CLUESB7 Spontaneous verbal clues which were not reported in the body of this study but which perhaps warrant some attention involve the test questions for which a subject voices concern. In 1953 Reid did not comment on this aspect of behavior. However, he has since stated that "experience has indicated that a subject who makes some admission or correction with respect to his answer to the control question [as the testing progresseg7...is 871t is of some historical significance, perhaps, to note that Inbau in 1942 reported that there seemed to be a difference between truthful and lying subjects as regards spontaneous verbal clues. For example, he reported at that time that liars more frequently asked if their nervousness would affect the test results and that they "more frequently asked to go to the lavatory to urinate just before the test. They also more frequently comment that the pneumograph and cuff being adjusted reminds them of the electric chair. These remarks and behavior cannot of course, be relied upon as proof of deception, since they are also made (though in fewer instances) by inno- cent [truthful persons. We mention their occurrence more as a matter 0 academic interest than as helpful criteria in dia osing deception." See: Fred E. Inbau, Lie Detec- tion, fgaltimore: The Williams and Wilkins Company, 1942) pp. 25-28 and note 14. 150 Appendix D (continued) 151 usually telling the truth about the relevant issue."88 Table 10 below indicates the frequency with which the two subject samples voiced concern for either the test control questions or relevant questions. As can be seen, the truthful subjects did more often "point to" the control questions than to the relevant questions. The opposite of this was true for the lying subjects. Thus, it appears that a subject who "points to" a test control question may, as Reid stated, be truthful al- though certainly the results are far from conclusive in this respect. Table lO NUMBER OF SUBJECTS VOICING CONCERN FOR EITHER CONTROL OR RELEVANT TEST QUESTIONS Subject Voiced Concern Voiced Concern Category For Control For Relevant Questions Questions % *5 Truthful NO' 1 N04 1 Subjects (N=50) 18 36 2 4 Lying Subjects (N=50) 7 14 11 22 88 The P01 1 liams an note 58. John E. Reid and Fred E. Inbau, "Lie Detector" ns ompany, Truth and Dece tion, (DaItimore: The Techni ue, 9 p0 299 and p0 309 "7'1? 111111'1111 1117 ’1'“