DEFENSE SECURITY SERVICE
1340 BRADDOCK PLACE
ALEXANDRIA, VA 22314-1651
MAY 15 2001
Honorable Orrin G. Hatch
Senate Judiciary Committee
Room 224, Dirksen Building
Washington, DC 20510
Dear Mr. Chairman:
The following responses (Attachment) are to questions of Mr. Michael H. Capps based upon the Senate Judiciary Committee Hearing on "Issues Surrounding the Use of Polygraphs," April 25, 2001.
CHARLES J. CUNNINGHAM JR. Attachment
Response to Senator Patrick Leahy's Questions Regarding the Hearing before the Senate Judiciary Committee on "Issues Surrounding the Use of Polygraphs," April 25, 2001
1. In Mr. Keifer's testimony, he refers to "prior studies" indicating that the polygraph has "an accuracy rate" of between 90 percent and 99 percent. Is there any report in the peer-reviewed scientific literature establishing that polygraph screening has a higher accuracy rate than 90 percent? If so, could you please identify that study?
In the testimony of Mr. Richard W. Kiefer [sic] before your Committee, Mr. Kiefer indicated that polygraph testing had an accuracy rate of between 90 percent and 99 percent. There are no research studies involving counterintelligence polygraph screening examinations that support an overall accuracy of 90-99 percent. When addressing specific issue examinations a significant body of literature demonstrates that polygraph decisions for criminal investigations have an error rate of ten percent or less. In the U.S. Supreme Court case U.S. v. Sheffer [sic] [523 U.S. 303 (1998)], numerous laboratory and field studies were identified by the Committee of Concerned Social Scientists as Amicus Curiae in Support of the Respondent as "high quality laboratory studies of the control question test." The Committee of Concerned Social Scientists reported the following:
In nine laboratory studies, 91 percent of the subjects were correctly identified when inconclusive opinions were excluded.
In five field studies, the accuracy reported for correctly identifying the guilty was 95 percent; the innocent were correctly identified in 75 percent of the examinations.
In five field studies using the original examiner's opinions, the accuracy rate for correctly identifying the guilty was 97 percent; the innocent were correctly identified in 98 percent of the examinations.
2. Mr. Keifer opines that, if Robert Hansen [sic] had been given a polygraph examination, he would have "reacted with greater than 99% certainty." Yet we know that Aldrich Ames was not caught even though he was given two polygraph examinations while he was at the CIA and that other guilty people passed polygraph tests. Is there any reliable basis to estimate the probability that a particular person would or would not pass a polygraph test?
There is no body of research that allows one to predict with certainty the outcome of a given polygraph examination. Estimating the probability that a particular person would or would not pass a polygraph test, a priori is not currently possible. Research indicates (Raskin, 1988) that with the levels of oversight imposed on government examiners which require standardized polygraph procedures and standardized numerical evaluations an accurate polygraph decision is usually the result.
While it is widely publicized that Aldrich Ames was not identified as a spy as a result of his polygraph testing, it is seldom stated that during the conduct of his two polygraph
examinations, significant responses to the relevant questions did occur. House Intelligence Committee Chairman Representative Dan Glickman related that Ames had been detected as deceptive in two of his answers (New York Times, August 10, 1994). Similarly, then-CIA Director James Woolsey (New York Times, March 8, 1994) acknowledged that the CIA had failed to follow-up on Ames' failure on two polygraph questions.
As in other forensic disciplines, false negatives and false positives do occur in polygraph testing. However, based upon government research and statistics routinely collected by DoD, the occurrence of false positives in counterintelligence scope polygraph examinations is not believed to occur in significant numbers. Research has shown that using a two-step process can mitigate these results. In this process, the examination proceeds from the multiple issue questions asked during the initial phase of testing to a single-issue test format during the subsequent phase. These additional testing phases are administered on an as needed basis to clear up issues to which the examinee showed significant responses. This methodology is consistent with Meehl and Rosen's "successive hurdles" approach (1955), which serves to lessen the adverse effects of imperfect validity and unbalanced base rates. This approach is similar to medical diagnostic screening strategies wherein tests with high sensitivity are given to the population of interest followed by testing those with positive results using procedures that have better specificity.
Using DoD statistics for fiscal year (FY) 2000 as an example, 7,890 examinations were conducted involving counterintelligence scope polygraph examinations.* Of the 7,890 examinations, 7,688 were evaluated as no significant response (non-deceptive) and 202 were evaluated as significant response (deceptive). Of the 202 evaluated significant response, 191 individuals made admissions to the relevant issues tested. Through additional polygraph testing, all relevant issues were resolved favorably for the examinee, i.e., they were able to maintain their clearance.
Of the 202 individuals who exhibited significant responses and/or provided substantive information, 194 received a favorable adjudication, three were still pending adjudication and five were pending investigation at the time of the DoD report, and none received adverse action denying or withholding access.
Using these statistics, the greatest possible number of false positive outcomes for the entire FY is eight. This is a possible false positive rate of less than one percent.
*Does not include NSA or NRO
Federal polygraph programs are designed to assist the investigator in identifying persons who merit further investigation prior to granting or denying access to sensitive information. In this context, the polygraph technique has shown itself to be the most efficient method for providing investigative leads for the adjudicator. However, it is important to remember that the polygraph technique is only one of the steps used in personnel screening.
Polygraph examinations during which significant responses occur to the counterintelligence questions are provided to adjudicators. Based upon the information gathered through all sources of the personnel screening process, agency adjudicators may decide to take no action, conduct a more thorough background investigation, or to forward the information for investigation.
In FY 2000 based on information gathered during DoD polygraph examinations, adjudicators made determinations of the security worthiness of 7,890 persons. As demonstrated by DoD statistics, 202 individuals were identified as significantly responsive to relevant questions relating to security issues, and 191 of these persons provided substantive information that allowed adjudicators to make decisions about referral for investigation or granting the clearance. Only eight persons out of 7,890 were possible subjects of investigations; five were referred for investigation. These statistics also indicate that persons who show significant responses during a polygraph examination usually are not involved in espionage but committed security violations of some kind.
4. Do you believe it is appropriate to exclude someone from government employment, without any independent corroborating evidence of deception or other information indicating that the applicant is unqualified for the position, solely because that person failed a polygraph? If not, what specific steps should be taken to insure that this does not occur?
Federal agencies should not exclude an applicant for employment solely because he/she reacts to relevant questions during a polygraph examination. Polygraph examinations are investigative tools, and sole reliance on them, or any other single tool, may not result in the level of decision accuracy equal to that of an adjudicative process that considers multiple sources of information. The polygraph examination consists of a set of standardized procedures designed to resolve issues during an applicant's screening examination. If at the end of the polygraph testing an applicant continues to demonstrate significant responses to counterintelligence issues and provides no information that would disqualify him/her from consideration, the polygraph decision should be provided to the adjudicators. The adjudicators should seek to verify information gained during the polygraph examination through the background investigative process. Once all information is gathered and provided
to the adjudicator, an employment decision should be made for the best-qualified applicant. The criteria to determine the best qualified should not overly rely on any single tool, including polygraph testing.
5. If someone is told that they have failed a polygraph test, is it more likely that that person will have an adverse physiological reaction if the same questions are asked in a subsequent polygraph test?
The effect of telling a person they failed a polygraph examination cannot be stated with certainty. In the absence of definitive research, field practices have safeguards that consider this factor. Subsequent to initial polygraph testing, if a person reacts significantly to a relevant question, the person is advised of this outcome in a positive, professional manner. The individual is correctly informed that a given question has not been resolved and the examiner solicits an explanation. A relatively short-lived, positive confrontation seldom has deleterious affects [sic] upon subsequent polygraph testing. The concern of most persons is that once an individual has been told that the individual had problems with the espionage question he/she will become sensitized to that question and will consistently and significantly respond to that question regardless of his/her veracity. Government research, DoD statistics, and the daily practical experience of federal polygraph examiners involving counterintelligence scope polygraph testing do not support this intuitive position.
The FY 2000 DoD statistics demonstrate that 199 persons required more than two series of questions to complete their examinations. During this same time a total of 66 examinations required more than one day to complete. In both of these instances, a positive confrontation would have occurred between the examiner and the subject of the examination. As indicated in the DoD statistics, even though persons are confronted about issues arising as part of the screening examinations, the vast majority (all but eight of 191) successfully completed the polygraph process. Research also supports the position that after a person has been found deceptive during initial polygraph testing, subsequent testing results in non-deceptive opinions being appropriately rendered. Research indicates that the error rates for examinees that demonstrate significant responses during initial testing can be mitigated if subsequent examinations are more focused. This occurs because the initial testing of a screening examination involves broad and general questions while subsequent series are able to focus the examinee on more direct issues.
6. Can chemical substances affect the results of a polygraph test? Is there a comprehensive list of prescription drugs and other substances that are known to alter the results of polygraph tests?
Virtually all polygraph examinations require that the examinee demonstrate the ability to respond to at least one question. To effectively alter the results of a polygraph examination, a chemical substance would have to demonstrate a differential effect. That is, the substance would have to suppress responses to some questions but not to others. To the best of our knowledge, no substances with this quality exist. No known drug is capable of selecting only certain questions on which to exert an effect.
The studies listed below indicate that specific drugs (alcohol, propranolol, diazepam, and methylphenidate (meprobamate), and trasicor) do not influence the results of polygraph examinations. Only the Waid, Orne, Cook, and Orne (1981) study suggests a drug effect, and this result is not supported by subsequent research. Moreover this study used a testing format rarely used outside of the laboratory and not at all by federal polygraph programs.
Bradley M.T., and Ainsworth D. (1984). Alcohol and the psychophysiological detection of deception. Psychophysiology, 21(1): 63-71.
Elaad, E., Bonwin, G., Eisenberg, O., and Meytes, I. (1982). Effects of beta blocking drugs on the polygraph detection rate: A pilot study. Polygraph, 11:229-233.
Gatchel R.J., Smith J.E., and Kaplan N.M. (1983). The effect of propranolol on polygraphic detection of deception. Abstract of unpublished manuscript, University of Texas Health Science Center, Dallas, TX 75235.
Iacono, W.G., Boisvenu, G.A., and Fleming, J.A. (1984). Effects of diazepam and methylphenidate on the electrodermal detection of guilty knowledge. Journal of Applied Physiology, 69(2): 289-299.
Iacono W.G., Cerri A.M., Patrick C.J., and Fleming J.A.E. (1987). The effects of antianxiety drugs on the detection of deception. Psychophysiology, 24: 594 (abstract).
Waid W.M., Orne E.C., Cook M.R., and Orne M.T. (1981). Meprobamate reduces accuracy of physiological detection of deception. Science, 212: 71-73.
The Department of Defense Polygraph Institute funded the Cail-Sirota and Lieberman (1995) study. This study resulted in establishing a database relating to drugs and their influence on the outcomes of psychophysiological detection of deception examinations. The Department of Defense Polygraph Institute lesson plan titled, "Pharmacology-Drugs and Psychophysiological Detection of Deception Testing," dated January 2001, provides a list of drugs that examiners could expect to encounter and the drug's effects upon the individual.
7. Is there any research indicating whether certain personality types have an easier time passing polygraph tests?
There is no research to indicate that the effects of personality variables are consistent for the various polygraph techniques. Personality is not an explicit component in any of the theories of polygraph, and it is not viewed as an important factor by field practitioners, which may explain why it has rarely been the focus of investigation in polygraph research. Some researchers have reported on the influence of demographic and psychological variables in their polygraph validity studies. Refer to attached Table for a summary of those effects.
8. Is there any research indicating whether certain ethnic or social groups have an easier time passing polygraph tests?
Since the 1960s, university and government researchers in the U.S. and elsewhere have conducted research on ethnicity and the polygraph. The trend has been that there are no meaningful differences in accuracy, but the research evidence has not eliminated ethnicity entirely as a factor. For example, desert-dwelling Bedouins have shown a dampened responsiveness in one channel but not in another channel. Similar results were found for Icelandic criminals. To date, among the population typically afforded polygraph testing in the U.S., an effect for ethnicity alone has not been shown to be reliable.
How do you insure that routine polygraph tests do not probe into purely private matters? Are there any questions that are off limits? What safeguards exist to prevent the release of private information?
Agencies provide written guidance to examiners that prohibits examiners from probing issues that are not related to the matter under inquiry. Individual agency policy requires that all questions asked during polygraph examinations must be reviewed with the examinee before the examination. Questions asked must be of special relevance to the subject matter under inquiry. Questions probing a person's thoughts or beliefs that are not related directly to the matter under inquiry are prohibited. The probing of a person's beliefs (such as religious beliefs and affiliations, beliefs and opinions on racial matters, and political beliefs and affiliations of a lawful nature) and questions that have no security implication are prohibited.
The federal polygraph standards state that all relevant questions must pertain directly to the matter under investigation or to the issue(s) for which the examinee is being tested. The federal polygraph standards also require that all questions asked during the data collection phase of the examination be reviewed with the examinee prior to the initiation of the examination.
The rights of the individual examinee are a primary consideration of the Quality Assurance Program, which is a program administered by the Department of Defense Polygraph Institute. This program verifies that each participating agency complies not only with their agency policies but also adheres to the federal polygraph standards. The Quality Assurance Program assures compliance through the use of biennial on-site inspections in which the agency's quality control procedures, policies, and the samplings of the examinations conducted by that agency are reviewed. During the inspection process a random sample of between 50 and 100 polygraph examinations is reviewed. As part of the examination review, the notes produced by the examiner during the examination are scrutinized to ascertain what issues were discussed with the examinee during the conduct of the examination. Any discrepancies are noted and corrective action recommended to the agency.
Table 4. Ethnic Influence on Detection of Deception
|Gudjonsson (1979)||Criminality (high)||X|
|Buckley, & Senese (1991)||Gender||X|
|Cutrow, Parks, Lucas & Thomas (1972)||Gender||X|
|Furedy, Davis & Guervich (1988)||Gender||X|
|Honts & Hodes (1982a,b)||Gender||X|
|Barland & Raskin (1975)||Intelligence||X|
|Bradley & Janisse (1981)||Introversion||X|
|Gudjonsson & Haward (1982)||Introversion||X|
|Stellar, Haenert & Eiseln (1987)||Introversion||X|
|Bradley & Klohn (1987)||Machiavellianism (high)||X|
|Kunzendorf & Bradbury (1983)||Prevalence of Visual Imagery (high)||X|
|Raskin & Hare (1978)||Psychopathy||X|
|Patrick & Iacono (1989)||Psychopathy||X|
|Heckel, Brokaw, Salzberg & Wiggins||Psychotic/Delusional||X|
|Iacono, Boisnevu & Fleming (1984)||Psychoticism||X|
|Buckley & Senese (1991)||Race||X|
|Windel & Hogan (1975)||Race||X|
|Balloun & Holmes (1979)||Socialization||X|
|Honts, Raskin & Kircher (1985)||Socialization||X|
|Gudjonsson & Haward (1982)||Socialization||X|
|Waid, Orne & Wilson (1979a, 1979b)||Socialization (low)||X|
Response to Senator Charles E. Grassley's Questions Regarding the Hearing before the Senate Judiciary Committee on "Issues Surrounding the Use of Polygraphs,"
April 25, 2001
1. Let's say that an employee polygraph exam ends with a deceptive result with no admission of guilt. How do agencies deal with this situation? How about an inconclusive result?
My understanding is that policies concerning the use of polygraph vary across agencies. I cannot state with certainty how individual agencies express and carry out certain policies.
2. My understanding is that most examiners within the ranks of federal law enforcement are non-supervisory or journeyman level personnel. Can we expect these agents to adequately administer polygraph examinations to Senior Level officials within their own agency?
Based upon their training and experience, senior field examiners are capable of conducting valid examinations of senior agency personnel. If a well-trained examiner uses a standardized polygraph procedure, very accurate results can be expected (Raskin, 1988). It is noted that DoDPI is the U.S. government's consolidated training facility for polygraph examiners from all Federal agencies. To qualify for entry into the program, a candidate must be a U.S. citizen, be at least 25 years of age, hold a 4-year degree or demonstrate an ability to master graduate-level courses, have two years of investigative service, have completed a background investigation to confirm a sound temperament and character, and be nominated and supported by his or her home agency. The DoDPI polygraph curriculum is taught at the master's degree level and provides a balance of a challenging academic load and technical skills practica. Those students who satisfactorily complete the DoDPI education program are released to their home agencies where they serve internships and remain subject to quality control and continuing education requirements for their entire professional careers as Federal polygraph examiners.
3. Will there be adverse consequences for employees who refuse to take a polygraph examination?
My understanding is that policies concerning the use of polygraph vary across agencies. I cannot state with certainty how individual agencies express and carry out certain policies.
4. If there are to be adverse consequences for not taking the exam, will this create an uncooperative emotional condition that could affect the results of an exam?
To my knowledge there has never been any research to address this issue.
5. FBI regulations prohibit the use of the polygraph as a "substitute for logical investigation by conventional means" (FBI Poly. Reg: 13-22.299(2)). Does this mean that, if all other factors are in order, the failure of a polygraph examination in the context of a national security update will not necessarily result in an adverse action?
I am not familiar with the policies and practices of the FBI and cannot provide information responsive to this question.
AntiPolygraph.org Home Page > Reading Room