Department of Defense
ANNUAL POLYGRAPH REPORT TO CONGRESS
Fiscal Year 2002 Office of the Assistant Secretary of Defense
(Command, Control, Communications, and Intelligence)
TABLE OF CONTENTS
I DoD Polygraph Statistics II Fiscal Year 2002 Counterintelligence-Scope Polygraph (CSP) Examinations CSP Refusals Specific CSP Examination Results Significant Information Developed III Utility of the Investigative Polygraph IV Training and Qualification Standards for Department of Defense Forensic Psychophysiologists (Polygraph Examiners) V Polygraph Research
EXECUTIVE SUMMARYDuring Fiscal Year 2002 (FY02) nine Department of Defense (DoD) agencies maintained operational polygraph programs. DoD employs the polygraph as one of many investigative tools to detect deception and assess the credibility of individuals involved in criminal investigations, counterintelligence cases, foreign intelligence and counterintelligence operations, as a condition for access to certain positions or classified information, and for requests for exculpation.
Approximately 74 percent of DoD polygraph tests are conducted as a condition of access to classified information or positions under the DoD Counterintelligence-Scope Polygraph (CSP) Program. The DoD CSP Program is authorized by Public Law 100-180. The purpose of the CSP Program is to deter and detect activity involving espionage, sabotage, and terrorism.
The DoD Polygraph Institute (DoDPI) provides basic and continuing education for all federal polygraph examiners. In FY02, the Accrediting Council of Independent Colleges and Schools accredited DoDPI. The institute currently offers a program in conjunction with Argosy University, Washington, DC to transfer graduate credit from the DoDPI Program towards a Masters Degree in Forensic Psychology, the Forensic Psychophysiology Track. During FY02, a total of 18 federal polygraph examiners were enrolled in this program and 14 were awarded Masters Degrees.
DoDPI also administers the Quality Assurance Program (QAP), which requires that all federal polygraph programs undergo a biennial inspection to ensure that their program meets or exceeds the standards outlined in the Federal Examiner's Handbook. During FY02, QAP inspected 13 federal polygraph programs, including eight non-DoD programs, all of which are now deemed to be in compliance.
The Research Division at DoDPI is staffed by a team of doctoral level scientists. Significant research in and development of alternative tools to detect deception and assess credibility has been limited because the Congressionally appropriated funds for the DoDPI Research Division have not been increased for over ten years. However, the Research Division has made significant progress through collaborative research projects to develop new, and less intrusive, technologies to detect deception and assess credibility.
Just after the close of FY02, the National Research Council (NRC) released a report on the existing scientific evidence of the validity of the polygraph technique. The report recommended that a significant investment in the research and development of new technologies to detect deception and be used for credibility assessment was required to produce tools for these requirements based on sound scientific principles. The Department concurs with this recommendation, and will pursue funding sources to be dedicated to this research.
Finally, it is important to note that the NRC Report also concluded that the polygraph technique is the best tool currently available to detect deception and assess credibility. The Department will continue to use the polygraph technique as it has in the past, until improved technologies or methodologies are developed as a result of scientific research.
IThe Department of Defense has used the polygraph technique for almost half a century. It is used as one of the tools presently available to detect deception and assess the credibility of individuals involved in criminal investigations, counterintelligence cases, foreign intelligence and counterintelligence operations, exculpation requests, and as a condition for access to certain positions or information. The polygraph technique is often successful in developing essential information used to resolve national security issues and criminal investigations. A statutory quota of 5,000 for CSP examinations limits DoD use of the polygraph.
DoD Polygraph Statistics
The following table reflects DoD Polygraph Program statistics for FY02.
Type of Exam Number Percent of Total Criminal 2,283 19.7% Exculpatory 354 3.1% CSP* 8,512 73.6% All Others** 417 3.6% Total*** 11,566 100%
* See page 3 for further breakdown of CSP examinations.
* Includes examinations conducted in support of personnel security investigations, counterintelligence and intelligence operations, and polygraph assistance to non-DoD federal agencies.
** Does not include polygraph examinations conducted by the National Security Agency (NSA). A breakout of polygraph examinations conducted by the NSA is contained in a classified table submitted with this report. Polygraph examinations conducted by the National Reconnaissance Office are conducted under the authority of the Director of Central Intelligence and are not reported in this report.
IISection 1121 of the National Defense Authorization Act for Fiscal Years 1988 and 1989 (Public Law 100-180, December 4, 1987; 101 Stat. at 1147) authorizes the Department of Defense to conduct CSP examinations as a condition for access to certain information.
Fiscal Year 2002 Counterintelligence-Scope (CSP)
The purpose of the CSP Program is to deter and detect espionage, sabotage, and terrorism. The following topics are covered during the CSP examination:
Public Law 100-180 (P.L. 100-180) authorizes DoD to administer CSP examinations to persons whose duties involve access to information that has been classified at the level of top secret or designated as being within a special access program under section 4.4 of Executive Order 12958. This includes military and civilian personnel of the Department and personnel of defense contractors and consultants. Since 1991, the provisions of Public Law 100-180 have limited the Department to 5,000 CSP examinations per year. However, DoD personnel assigned, detailed, or under contract to the following positions are exempted from this ceiling and may be subject to a CSP examination:
- Involvement with a foreign intelligence/security service.
- Involvement in terrorism.
- Unauthorized foreign contacts.
- Deliberate failure to protect classified information.
- Damage or sabotage to government information, clandestine collection, or defense systems.
The following table provides a breakdown of CSP examinations conducted by the DoD in accordance with Public Law 100-180.
- DoD personnel assigned, detailed or under contract with the Central Intelligence Agency.
- DoD personnel assigned, detailed, under contract, or applying for a position in the National Security Agency.
- DoD personnel assigned to a space where sensitive cryptographic information is produced, processed, or stored.
- DoD personnel employed by, assigned, detailed, or under contract to an office within the DoD for the collection of specialized national foreign intelligence through reconnaissance programs.
(1) Special Access Programs 2,819 (2) DIA Critical Intelligence Positions 1,345 (3) TOP SECRET 0 (4) Exams for Interim Access to Sensitive Compartmented Information 55 Total CSP Exams Conducted Under the Congressional Ceiling 4,219 Exempted Examination* 4,293 DoD CSP Program Totals** 8,512
* Includes detailees to CIA and NSA, assignees to cryptographic information processing spaces, persons in non-NRO reconnaissance programs.
**NOTE: Does not include polygraph examinations conducted by NSA. A table of polygraph examinations conducted by NSA is contained in a classified annex to this report. Nor does it include examinations conducted by the National Reconnaissance Office (NRO), which are conducted under the authority of the Director of Central Intelligence.
During FY02, two military members declined to submit to CSP testing required as a condition of access to classified information. DoD policy states that those persons who decline to take the CSP examination are denied access to the classified materials in question, but are retained in their position or transferred to other positions in the organization of equal pay and responsibility. Of the two individuals who refused to submit to the CSP examination, one subsequently retired from military service and the other voluntarily left at the end of his term of enlistment.
Specific CSP Examination ResultsThe polygraph examination results for the 8,512 individuals tested under the DoD CSP Polygraph Program are as follows:
Three hundred seventy-two individuals required more than two series (a series is defined as the collection of at least two polygraph charts on an examinee). A total of 87 examinations required more than one day to complete.
There were 8,245 individuals who were evaluated as no significant response (non-deceptive) to the relevant questions and provided no substantive information. The remaining 267 individuals were evaluated as displaying significant responses (deceptive) and/or provided substantive information.
Two hundred forty-seven individuals made admissions relevant to the issues on the CSP topic list, and through further testing, all relevant issues were resolved favorably to the individual.
Seventeen individuals made admissions relevant to the issues on the CSP topic list, and continued to be evaluated as deceptive during further testing.
Of the 267 individuals who were evaluated as displaying significant responses or provided substantive information, 247 received a favorable adjudication, 11 are still pending adjudication, nine are pending investigation, and no one received adverse action denying or withholding access.
Expansion of the CSP Program
Public Law 100-180 instituted an annual ceiling on the number of CSP examinations conducted by DoD. This annual ceiling has remained at the same level since 1991. Since that time, the Department has identified additional vulnerabilities and threats to classified information that did not exist over a decade ago. The broad based use of information technology systems, coupled with the development of information sharing capabilities over the internet and through other electronic media, require the updating of DoD information assurance policies and practices to keep pace with this emerging threat. DoD is considering the development and implementation of enhanced security requirements for information technology professionals with root access to DoD information systems. These enhanced security requirements may require a CSP polygraph examination for access to DoD information systems. Based on the number of information technology professionals assigned to DoD, an increase in the CSP ceiling to prior 1991 levels may be requested from Congress.
Significant Information Developed
The following cases reflect significant information developed during DoD CSP examinations covered by this report. Most of the information that was developed relates to the removal of classified materials, including computer media, from secure environments. Unauthorized contacts with foreign nationals is another common category of CSP developed information. After information is developed from a subject, further CSP polygraph testing is conducted to determine the extent of the violation. Final CSP polygraph examination results, along with a summary of developed information, are provided to appropriate security officials for further investigation and final adjudication.
During a CSP examination, a DoD civilian employee admitted that he removed confidential documents as a souvenir when he left one of his previous assignments. Investigators subsequently recovered the documents from the employee's residence. The employee favorably completed additional polygraph testing.
During a CSP examination, an examinee admitted to the inadvertent unauthorized removal and destruction of a classified document. The examinee successfully completed the polygraph examination after this admission.
During a CSP examination, the examinee admitted to the removal of classified materials from secure environments on several occasions, so that he could complete office assignments at home. These violations were substantiated through investigation; however, the investigation did not develop any additional information on security violations. The examinee successfully completed a polygraph examination after the close of the investigation.
During a CSP examination, the examinee admitted that he had improperly stored classified materials outside of government control, so that he could complete work assignments at home. He denied that any of the materials had been compromised. The examinee surrendered the classified information to investigators. The examinee then favorably completed his polygraph examination.
During a CSP examination, a DoD civilian employee disclosed a previously unreported romantic relationship with a foreign national. The employee denied disclosing classified information to his foreign national girlfriend. He successfully completed the polygraph.
During a CSP examination, a DoD employee assigned overseas admitted that he had become romantically involved, and eventually co-habitated with a foreign national. The employee also stated that he had discussed classified information with his foreign national girlfriend. After these admissions, he successfully completed the polygraph examination.
During a CSP, a DoD employee admitted that he had developed a close relationship with a foreign national while assigned outside the United States. The foreign national served as the employee's clergyman. On one occasion, the clergyman had asked the employee for personal information about another DoD employee assigned to a different overseas site. After making these admissions, the employee successfully completed the polygraph examination.
During a CSP examination, a DoD employee admitted to using a restricted government computer system to determine the classified location of the U.S. Navy ship to which his son was assigned. The employee then shared this restricted information with other family members. After providing this information, the employee successfully completed the polygraph examination.
Utility of the Investigative Polygraph
During FY02, investigations conducted by DoD obtained significant information from interviews conducted with the aid of the polygraph. DoD policy mandates that the polygraph technique can only be employed to supplement traditional methods of investigation. In the examples detailed below, traditional methods of investigation had been unsuccessful in resolving the matter under investigation. In these cases, information derived from the polygraph examination proved to be invaluable in successful resolution of those matters.
A military member was suspected of involvement in a check-kiting scheme after a worthless check was deposited in his checking account, and the money was withdrawn resulting in a loss to the financial institution. When interviewed by investigators, the suspect stated that he believed the worthless deposit was his reenlistment bonus, which had been electronically deposited into his account. After a deceptive polygraph examination, the suspect admitted knowingly depositing the worthless check, and implicated three additional individuals in the check-kiting scheme.
A military member was accused of sexual assault by three female trainees under his command. The subject was interviewed and denied any inappropriate behavior with the victims. After a deceptive polygraph examination, the subject admitted to kissing and inappropriately touching all three women.
A polygraph examination was administered to a military member implicated in the death of a co-worker. After a deceptive examination, the suspect confessed to killing the victim, and slicing the victim's wrist in an attempt to make the death look like a suicide.
Six military members were suspected of stealing two classified laptop computers from a military installation. Five individuals were eliminated as suspects after non-deceptive polygraph examinations. The sixth individual admitted to stealing the computers after a deceptive polygraph examination. The computers were subsequently recovered from the suspect's residence. Additional polygraph examinations were conducted and determined that the classified information on the computers had not been compromised.
A retired military member working as a DoD contractor was administered a polygraph examination to determine the veracity of critical information he provided relating to the terrorist attacks on September 11, 2001. The examinee was evaluated as deceptive, and admitted that he did not know if the information he provided was factual, nor did he know if the event he reported had actually occurred.
A supplier to the DoD shipped counterfeit and gray market copier cartridges instead of new cartridges that had been purchased directly from the brand-name manufacturer. The supplier provided a letter of authenticity to the DoD, certifying that the cartridges were purchased directly from the brand-name manufacturer. The employee of the supplier who provided the letter of authenticity was interviewed by investigators, and agreed to submit to a polygraph examination to confirm that he did not alter or counterfeit the letter. After a deceptive polygraph examination, the employee admitted that he had altered a genuine letter of authenticity using his office computer, and submitted it to the DoD. The employee also implicated the owner of the supply company in the scheme.
The skeletal remains of a military member missing since 1990 were found concealed in a chimney, located in Germany. An autopsy of the victim could not determine the cause of death, but foul play was suspected. During the course of the investigation, a former military member who had been stationed in Germany at the time of the incident was identified as a suspect. The suspect was interviewed and denied any involvement in the death. A polygraph examination administered to the suspect was evaluated as deceptive. The suspect then admitted that he and four other soldiers had beaten the victim to death and concealed his body in the chimney.
An investigation was initiated after five containers of high explosives were stolen from an off-post construction site and recovered several days later on a military installation. A military member was developed as a suspect in the theft. The suspect denied any involvement in the crime and agreed to undergo a polygraph examination. After a deceptive result, the suspect admitted that he and another service member had stolen the explosives and discussed bombing two separate buildings on the installation.
IVDoD maintains stringent standards for polygraph examiners. The basic curriculum taught at DoDPI is based on forensic psychophysiology, and conceptual, abstract, and applied knowledge that meet the requirements of a master's degree level of study. Candidates selected for assignment as a DoD polygraph examiner must meet the following minimum requirements:
Training and Qualification Standards for DoD Forensic
Psychophysiologists (Polygraph Examiners)
All federal polygraph examiners receive their basic training at DoDPI. In FY02, the Institute trained 84 new polygraph examiners. After completing the basic polygraph training course, DoD examiners must intern under the supervision of a certified polygraph examiner for a period of at least six months. In addition, DoD polygraph examiners are required to complete 80 hours of continuing education every two years. To help meet this requirement, DoDPI offers specialized courses in forensic psychophysiology and related disciplines. In FY02, 780 students attended this specialized training.
- Be a citizen of the United States.
- Be at least 25 years of age.
- Be a graduate of a four-year college.
- Possess two years of investigative or comparable experience with a federal or other law enforcement agency.
- Be of high moral character and sound emotional temperament as confirmed by a background investigation.
- Successfully complete a pre-selection polygraph examination.
- Successfully complete the basic polygraph examiner course at the DoDPI.
During FY02, DoD agencies maintained an average of 155 certified polygraph examiners. Over the preceding eight fiscal years, DoD averaged 160 certified examiners on staff per year.
If DoD institutes a CSP polygraph examination requirement for information technology professionals, a significant increase in the number of certified examiners will be required to handle the increased workload.
VA team of doctoral level scientists staffs the Research Division at DoDPI. Significant research in and development of alternative tools to detect deception and assess credibility has been limited because Congressionally appropriated funds for the DoDPI Research Division have not been increased for over ten years. However, the Research Division has made significant progress toward developing new and less intrusive technologies to detect deception and assess credibility through the use of collaborative research projects.
Forensic Psychophysiology (Polygraph) Research
The National Research Council (NRC) of the National Academies of Science recently released a report, The Polygraph and Lie Detection (2002). In the report, the NRC commended recent efforts by the DoDPI Research Division but clearly stated the need for expanding these efforts in the interest of national security. More specifically, the NRC strongly recommends expansion of collaborative research efforts from both independent and government laboratories. The NRC also concluded that no alternative technique has yet been shown to outperform the polygraph technique. As mandated by Congress, DoDPI continues to increase the number of research partners through collaboration to research alternative technologies.
As the country enters a new era of threat, the need to expand beyond traditional polygraph techniques has never been more mission essential. DoDPI hopes to broaden its focus to include the entire spectrum of credibility assessment. This change in focus will allow for the rapid development of technologies designed to assist the intelligence community, the homeland and national security communities, as well as the military services in the Global War on Terrorism. To this end, DoDPI Research Division strives to unite the efforts of the government and scientific communities to give our polygraph examiner workforce leading edge technology in protecting the Homeland and Americans worldwide.
Through the 2002 Broad Agency Announcement (BAA) and staff solicitation, the Research Division accepted 13 new institutional proposals, and two graduate student dissertation/thesis award proposals. DoDPI granted three institutional awards and one dissertation/thesis award. In addition, two awards were granted from the 2001 BAA funds. A total of five proposals were rejected for lack of scientific merit and five proposals are being considered for future awards if funds become available. Additionally, there are six proposals in the review process for future consideration. There are also ten internal DoDPI projects ongoing.
Research projects during FY02 resulted in 35 scientific papers and reports published and available to the community. This is the fourth consecutive year the DoDPI Research Division increased its production of deliverables to our customers. The progressive increase in activity is a direct result of the recruitment and partnership development with external laboratories.
Collaborative Research Projects
University of New York at Stony Brook
An Examination of Response Parameters of Electrodermal Recording to Standard Stimuli. The objective of this project is to investigate whether equivalent electrodermal responses are obtained to equivalent psychological stimuli presented at different electrodermal tonic levels. The outcome will determine whether resistance or conductance is a more accurate measure during PDD examinations. Reviews of final report completed. Waiting for final revision.
Development of Improved Automated Scoring Algorithms: An Application of Advance Waveform Combining and Classification Technology to the Analysis of Polygraph Data in Psychophysiological Detection of Deception. This project intends to develop a novel waveform application for the scoring of Relevant/Irrelevant test format polygraph data. A number of advanced statistical methods have been developed for the analysis of complex electroencephalographic (EEG) and magnetoencephalographic (MEG) signals, which may also be applied to polygraph physiological data. The goal of this project is to use traditional polygraph data and combine the data into a novel waveform that will enhance the accuracy of the PDD examiner decision. This technique has been shown to be highly accurate and reliable for single-trial analysis of EEG and MEG signals and should be adapted to similar datasets collected from polygraph examinations. Reviews of final report completed. Waiting for final revision.
Washington School of Medicine and Boeing
Noncontact Sensing of Emotion and Stress Using Laser Doppler Vibrometry. This project involves the use of emerging technologies to develop methods for deriving simultaneous information from the laser Doppler signal regarding multiple physiological functions including body tremor, respiration, cardiac function, muscle contraction, and sweating. Laser Doppler Vibrometry recording methods do not require the attachment of physical transducers, and could be adapted to multiple-examination settings. Awaiting final report. Extension to 2003 has been approved.
University of South Carolina
Research Assistant Professor Research/Research Training in Cognitive Psychophysiology and Detection of Deception. The University of South Carolina conducted research on brain activity as it relates to the detection of deception. The project uses high-definition EEG/ERP recordings and correlates these findings with current autonomic nervous system recordings during a PDD examination. Final report in review.
Vericator: Evaluating the Validity of a Voice-based Measure of Detecting Deception. This study assessed the ability of Vericator, a computer-based system that evaluates credibility through speech, to detect smugglers at a mock security checkpoint. A U.S. Customs Inspector questioned participants while Vericator assessed their veracity. For some the Inspector followed a script of questions without follow-up (Scripted); for the remainder, follow-up questions were permitted (Field-like). Smuggling base rates were 34% and 35%, respectively. The proportions of smugglers correctly identified at the checkpoint, i.e., sensitivity, were low (.12 and .18, respectively). Proportions of non-smugglers correctly identified, i.e., specificity, were much higher (.80 and .78, respectively). Post-hoc analyses produced widely disparate sensitivities (.38 to .78) and specificities (.00 to .96). Final report pending.
University of South Carolina
Research Assistant Professor Research/Research Training in Cognitive Psychophysiology and Detection of Deception. The University of South Carolina is conducting a continuation of research on brain activity as it relates to the detection of deception. The project uses high-definition EEG/ERP recordings and correlates these findings with current autonomic nervous system recordings during a PDD examination. The components include research in cognitive psychophysiology to: 1) continue to refine the localization of cortical sources of deception and utilize these underlying neuronal sources to evaluate deceptive responses at an individual level, 2) begin separating the process of deception from the effects of workload and attention, and 3) conduct multiple session research and apply statistical approaches to the data to evaluate possible practice and memory effects. Performance period 1 Jun 02 - 31 May 04.
Washington School of Medicine and Boeing
Noncontact Sensing of Emotion and Stress Using Laser Doppler Vibrometry. This is a two-year continuation of a project investigating laser Doppler vibrometry (LDV) as a method for assessing the physiological signs of stress and emotion. The research is aimed at five goals: 1) Continue to develop scientific basis for the method, and continue to explore the range of physiological phenomena that may be of use for the detection of deception, 2) Complete an ongoing study of the somatic and cardiorespiratory responses associated with exposure to affective pictures, 3) Operationalize forcefully the specific state of fear, using standardized social stress methods, and to assess the associated physiology using LDV measures, 4) Translate to a mock crime scenario the critical LDV measures that most successfully characterize the state of fear, 5) Continue to develop and refine the technical basis of the LDV method by upgrading existing hardware and software capabilities. Performance period 1 Oct 02 - 30 Sep 04.
Exploring Content Coding Procedures for Assessing the Truthfulness of Verbal Statements. The intent of the research is to examine the utility of verbal analytic procedures in assessing honesty and deception in simulated and real employment contexts. Specifically, criteria from various techniques for assessing the truthfulness and deceptiveness of verbal statements based on their content, quality and expression will be tested in a series of experimental studies. Ongoing data collection and analysis. Performance period 1 Jun 02 - 31 May 05.
The University of Utah
Human and Computer Decision-Making in the Psychophysiological Detection of Deception. The objectives of this study are to develop algorithms to measure the 23 types of physiological changes considered to be diagnostic of deception, assess the stability of those measures across test questions and charts (reliability), and assess their diagnostic validity. It will identify the criteria used by DoDPI-trained examiners to evaluate polygraph charts. It will also explore methods for combining the criteria in order to maximize the accuracy of polygraph decisions. Performance period 1 Jun 02 - 31 Aug 03.
The University of Utah
Growth Curve Analysis of Polygraph Data. The polygraph technique is commonly used during criminal investigations to assess the veracity of suspects, witnesses, and defendants. In a typical polygraph examination, a series of test questions is presented from three to five times with a short break between each repetition of the question list. A process known as habituation can produce a progressive decrease in the magnitude of physiological responses over repeated presentations of the test questions. Depending on differences in the rates of habituation for different types of test questions, it may become more difficult or less difficult over the course of a polygraph examination to determine if the examinee is deceptive or non-deceptive. The study will assess if there are diagnostic patterns of change in physiological measures over the course of a polygraph examination. Pending release of funds.
The Relationship Between Facial Skin Surface Temperature Reactivity and Traditional Polygraph Measures Used in the Psychophysiological Detection of Deception: A Preliminary Investigation. This study was designed to investigate the feasibility of combining traditional polygraph measures including blood volume, respiration, and electrodermal activity with facial skin surface temperature (SST) changes recorded using high definition thermal imaging. Participants were randomly assigned to nondeceptive (n=13) or deceptive (n=12) treatment groups using a mock-crime scenario. The frequencies of accurate determinations made using traditional polygraph measures, SST measures, and a combination of polygraph and SST measures were compared using binary logistic regression. Highest accuracy was obtained using a combination of polygraph and SST measures, suggesting that recordings of facial SST provide information that may be useful when combined with traditional measures during a polygraph examination. These results are discussed in relation to the orienting response (OR) theory proposed by Sokolov, (1963, 1997).
New Decision Rule Development. This study was focused on increasing psychophysiological detection of deception accuracy for the Zone Comparison Test through the modification of the decision rule process. Two two-stage models for producing decisions following conventional physiological data scoring are proposed. The 3T stage uses total score cutoffs to produce decisions, with totals of -6 or less producing decisions of deception indicated, totals of +6 or greater producing decisions of no deception indicated, and totals between these cutoffs producing a no opinion (NO) decision. The 3S stage evaluates scores assigned to individual question pairs in addition to total cutoffs to produce decisions. The assigned scores are totaled for each of three relevant-comparison question pairs, producing a spot score for each question pair. If the total score is -6 or less or if any spot score is -3 or lower, then a decision of deception indicated is produced. If the total score is +6 or higher and if all three spot scores are +1 or greater, a decision of no deception indicated is produced. If neither of these criteria are met, a decision of NO is rendered. The 3T3S model used the total cutoff rule first and the spot score rule only if a NO decision was produced after the 3T stage. The 3S3T model implemented the stages in reversed order. We examined the accuracy of these two-stage models and compared that accuracy with decision models that used only one of the two stages (3T or 3S), using three laboratory and four field data sets. Results showed that across all data sets, the 3T3S and 3S3T models produced 8.3% more correct decisions, 11.5% fewer NO decisions, and 3.1% more incorrect decisions than the 3T and 3S models. Second revision of final report in progress.
Polygraph Question Series Trend Analysis. This project was undertaken to determine whether a different degree of diagnostic value exhibits itself at different points during the question series process. This project will examine this issue using a number of different approaches. First, using hand-scored data from two laboratory studies, the diagnostic contributions from a number of different levels will be considered. These include the relevant-comparison question pair as a function of channel and as a function of question series, all of which to be assessed using a Multiple Regression approach. The contribution of specific question series and data channels will also be assessed, independent of question pair. Second, accuracy rates will be determined using the first three, middle three, and last three question series collected in a five-question series data set. These different, but overlapping data sets will be used to produce decisions using four different decision models. These include one approach using only total cutoffs (3T), one approach using only spot scores (3S), and two approaches using both decision rules in a two-stage process (3T3S and 3S3T). All of this research is exploratory, and conducted to determine empirically if there is an approach equal to or more effective than that produced using the typical three-to-five chart rule in combination with a two-stage decision rule approach. Preliminary data analysis completed.
An Evaluation of TES Decision Rules. In an effort to increase the sensitivity of the TES to deception, decision rules will be applied to field cases to determine how they affect decisions. Score sheet data have been collected from several Federal polygraph programs, and the data are now being converted to electronic form. Ongoing data input.
Comparison of 3 Major Scoring Systems. In the field there are currently three principal ways in which polygraph examiners score relevant questions. One method is to score the relevant question against the stronger of the two adjacent comparison questions. This is the practice being taught at DoDPI. A second method is to always score the relevant question against the comparison question that immediately precedes it. The Utah and the Matte scoring system takes this approach. A third major method uses the Either-Or rule, in which the examiner decides which comparison question to use based on whether there is a strong reaction to the relevant question. The Either-Or rule is taught by Cleve Backster. All three methods are widely used, but the effects of these different systems have not been carefully investigated. In this project we will ask volunteers who are proficient in one of the three systems to score 360 chart segments using the seven-point scoring system. The segments contain one relevant question bracketed by two probable-lie comparison questions. The charts are from single-issue Zone Comparison field cases of criminal suspects. We will assess how total scores are affected by the scoring method, and calculate cutting scores that render similar accuracies for the systems. Scoring complete.
Modified General Question Test Decision Rule Exploration. Previous studies by Senter, Dollins, and Krapohl (2001) and Senter and Dollins (2001) showed that substantial increases in accuracy could be gained within the Zone Comparison Test format by using three to five question series, relative to the standard government practice of using three question series. The current work focused on boosting accuracy through the modification of the decision rule process with the Modified General Question Test (MGQT). A two stage model for producing decisions after score assignment was proposed, in order to compare performance with the pure spot score rule. The first stage uses total score cutoffs to produce decisions, with totals of -6 or lower producing decisions of deception indicated, totals of + 6 or higher producing decisions of no deception indicated, and totals in between these cutoffs producing a no opinion decision. If a no opinion decision is produced using the first stage, the second stage, using a spot score rule, is enacted. In this stage, the assigned scores are totaled for each of the three relevant-comparison question pairs, producing three spot scores. If the total of these three scores is -6 or lower or if any spot is -3 or lower, then a decision of deception indicated is produced. If the total of these three scores is +6 or higher and if all three spots are +1 or greater, a decision of no deception indicated is produced. If neither of these criteria is met, a decision of no is rendered. This project is a retrospective analysis of archived data conducted to determine the effectiveness of this approach with several diverse data sets. This two-stage approach has shown to produce large increases in the number of correct decisions with nondeceptive participants, both with laboratory and field data with the Zone Comparison Test, and the goal is to determine the impact of these rules on MGQT performance. Data Analysis Stage complete.
The Impact of Averaging Assigned Scores on Polygraph Decision Accuracy. The existing body of research on differences between group versus individual performance has produced mixed results. In certain contexts, the group outperforms the individual, and in others, individual performance is superior to that of the group. The present study explored how group performance compared to individual performance in the blind scoring of polygraph question series, using a Zone Comparison Test format. Results provide preliminary evidence that use of a group or collective decision approach in producing polygraph decisions could enhance the validity of the process. Actual implementation of such a process could be achieved using a simultaneous approach where evaluators review polygraph data at the same time as the examiner conducting the exam. Concurrent real time printing or viewing could facilitate such an approach. The blind evaluators could then transmit their assigned totals that could then be combined to produce a group decision, all in time for the examiner to conduct the posttest in the usual fashion. Given the state of high-speed Internet connectivity, such a system that would enable input from evaluators at considerable distance is certainly a possibility. Early report drafting.
Information Processing Associated with Emotional States' Studies. Thermography is a safe, non-invasive technology that measures skin surface temperature on a real time basis. The following studies were designed to determine whether this technology (a special non-contact thermal camera will be used in all four studies) will be useful as a supplement to traditional polygraph measures including blood pressure changes, electrodermal (sweat) activity using metal plates attached to the fingers, and straps attached to the chest and abdomen that records respiratory activity. Cardiovascular activity will be recorded using either electrodes attached to the wrist and ankle (Studies one and two) or a standard blood pressure cuff (Studies three and four). Participants in the first study will listen to sounds (white noise, tones, clicks) of varying pitch and loudness while their physiological responses are recorded. Participants in the second study will listen to positive (e.g. "ecstatic") and negative ("mutilate") affect words taken from a standard set of emotion word stimuli. Some of the participants will also be asked to complete a counting (math) task varying in difficulty and speed of response. Participants who complete the counting task will also receive feedback about their performance as the task is being completed. Participants in the third and fourth studies will be randomly assigned to one of two groups. Participants in the first group will commit a pretend crime by stabbing a mannequin with a screwdriver and stealing money from a purse. Participants assigned to the second group will be informed that a pretend crime was committed, but that they are innocent. Next, all participants in studies three and four will undergo a polygraph examination using traditional polygraph as well as thermography measures. These measures will be examined to determine whether the thermography measures increase the effectiveness of a polygraph examination. Participants in the fourth study will also be asked follow-up questions (about the participant's involvement in the mock-crime) by the polygraph examiner. While asking these questions, the examiner will be receiving on-line data from the thermal camera to help direct the line of questioning. Four phases. Phase one complete. Other phases proceeding.
Research Projects in Review for FY 2003
Penn State University
The Development of a Computerized Question-Response System and the Use of Reaction Time to Detect Deception. During Year 1, we will develop a computerized system to both present questions and record responses. The two-fold purpose of the computerized system will be to standardize the procedure and to assess the diagnostic utility of both behavioral and physiological reaction time data for the detection of deception using the control question technique (CQT). During year two, a large-scale experiment will be conducted with male and female subjects and examiners to assess the efficacy of the computerized system and the contribution of RT data to detection of deception.
Polygraph Decision Support system for Relevant/Irrelevant Format Examinations. There is a need to investigate the validity, reliability, and accuracy of PDD examinations for the Relevant/Irrelevant (R/I) Test Format. There is also a need to better understand and automate the process of determining whether a subject is being truthful or deceptive during a polygraph examination. The development of improved computer algorithms, which accurately process polygraph waveforms to better determine deception or truthfulness, will benefit the field of forensic psychophysiology and terrorist detection.
Boise State University
Effects of Comparison Question Type and Between Test Stimulation on the Validity of Comparison Question Tests. Comparison Question Tests (CQT) are the most commonly used type of psychophysiological detection of deception test in, law enforcement, forensic practice and in national security screening settings (Honts, Raskin & Kircher, 2002; Raskin & Honts, 2002). Such tests play an important role in the Government's national security and law enforcement programs. However, many aspects of the polygraph testing procedure, as it is used in practice, lack strong empirical validation, and in some cases some aspects of the testing procedures lack any empirical validation. Two aspects of comparison test administration are currently the topic of some controversy in polygraph literature. Those areas of controversy concern the type of comparison question used, and between chart stimulation of questions.
University of Nebraska Kearney
Polygraph personnel security screening with subjects for whom English is their second language: The effects of Examiner Language and Interpreter Presence. Although the use of foreign-born employees is not uncommon in the Department of Defense, cross-cultural research in the use of polygraph has been limited to polygraph examiners' perceptions of differences (Yankee, 1991). Although American and Japanese polygraph examiners suggest that communication issues between examiners and polygraph subjects may lead to inaccuracies, no empirical research has examined whether or not bilingual participants respond differently to the Test of Espionage and Sabotage (TES), of any other polygraph examination format, as a function of language used. With recent increases in pre- and post-employment testing by U.S. Government agencies, such as DOD, CIA, FBI, and DOE and the increased needs and concerns associated with hiring employees not native to the United States, it becomes even more important to makes sure that the current screening procedures used to test employees are accurate with all subjects.
Georgia Institute of Technology
A Radar-Based Non-Contact Physiological Detection of Deception Sensor. GTRI is proposing an emerging technology program (Topic two) to develop a non-contact PDD sensor based on existing Radar Vital Signs Monitor (RVSM) technology. The RVSM is able to detect cardiac function, respiration, muscle contractions, eye blinks, and possibly an indicator correlative with galvanic skin resistance without contact. The frequency of operation of the sensor allows the radar signal to penetrate nearly all types of clothing. The resulting information could be used alone to detect deception, or fused together with the traditional polygraph to increase sensitivity and reduce false positives. The end goals of the proposed program are: Construction of an RVSM prototype for PDD; Development of the necessary foundational science and experiments to examine RVSM information for PDD; Explore fusion of the RVSM with traditional polygraph.
The University of New Brunswick
The Influence of the Threat of a Polygraph Test on Deception Ability. Overall, the argument presented in this paper is that the polygraph interrogation situation is not just a simple measurement occasion. Advocates and critics of the procedure argue that cognitive and psychological schemas involving motivation and emotion are created and modified by the very fact that a polygraph test is included in an interrogation sequence. The contemplation of a polygraph test may well serve a function as an intensifier of stress that acts directly on the interrogation following the polygraph. This intensification may be felt by all suspects but may affect guilty suspects the most. The present study is designed to look at possible effects on behavior indices of deception and truth telling on suspects contemplating a polygraph test.
An Examination of the Impact of Evaluating Additional Question Series Using Field Data. The proposed study is an attempt to validate the findings of Senter, Dollins, and Krapohl (2001), and Senter and Dollins (2002). These previous studies showed statistically significant increases (13.0% and 10.5%, respectively) in the number of correct decisions produced when three or five question series are used to produce decisions relative to when only three question series are used. One limitation of these previous studies is that data were collected using participants in laboratory mock crime scenarios. Thus, the possibility remains that this increase in the percentage of correct decisions may be limited to data collected in laboratory settings, limiting the utility of the technique in the real world. In the present study, we will attempt to replicate the findings of our previous studies using data collected during actual criminal investigations. Roughly 500 sets of question series will be collected from specific issue field cases conducted by examiners employed by the United States Army Criminal Investigation Division Command. Five question series will be collected from each examinee, instead of the usual three. Following the confirmation of deception or truthfulness for each case where possible, blind evaluators who are blind to examinee veracity will assign scores to each relevant-comparison question pair for each data channel and question series for each case. These assigned scores will be used to produce decisions by calculating different combinations of spot versus absolute scores and the three versus three to five question series contingency rule. These decisions will be used to assess accuracy (for confirmed cases) and utility (for unconfirmed cases) for various data collection approaches. To begin early CY '03.
ContractsJohns Hopkins University
New Feature Development and Countermeasures Detection Improvement. This project is designed to develop improved methods of evaluating physiological data known as features in polygraph examinations. This research is also designed to improve the accuracy of detecting polygraph countermeasures.
PolyPlot. Computer program for generating and modifying polygraph charts. Construction of a polygraph simulator suitable for use an instructional and research aid for DoDPI. The system will provide DoDPI instructors with the ability to generate unique, realistic charts with various test formats, including ZCT, MGQT, TES, and R/I. The resulting charts can then be used in a classroom environment to demonstrate characteristics of polygraph tracings that may not be readily accessible using existing field data.