May 7, 2001
VIA FACSIMILE AND MAIL
Honorable Orrin Hatch
Senate Judiciary Committee
224 Dirksen Building
Washington, D.C. 20510
Re: Senate Judiciary Committee Hearing - April 25, 2001
“Issues Surrounding the Use of Polygraphs” - Supplemental Written Questions
Dear Senator Hatch:
As a follow-up to my oral testimony before the Committee with respect to the above topic, I am pleased to submit my supplemental responses to written questions submitted by Senators Patrick Leahy and Charles Grassley. I have only responded to those questions where I feel I am most qualified based on my expertise and research.
Questions Submitted By Senator Leahy
1. In Mr. Kiefer'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.
Almost every available polygraph study conducted pertains to specific incident criminal investigations (i.e., identifying the thief who embezzled funds). This question properly addresses the most significant aspect affecting current federal polygraph policies. The Congress needs to be most concerned about the reliability/validity of polygraph screening tests. It is these types of tests that are administered every year to thousands of applicants for federal employment, as well as tens of thousands of current federal employees who undergo routine security investigations. The primary purpose of
the applicant screening test is to determine suitability while the security screening test is designed to expose espionage. However, there is absolutely no scientific evidence that either type of screening test is reliable or valid. The few studies that exist prove that screening tests should be stopped immediately.
The largest study of polygraph tests used for national security screening ever conducted - “Studies of the Accuracy of Security Screening Polygraph Examinations” - was published in 1989 for the Department of Defense’s Polygraph Institute (“DoDPI”) by Gordon H. Barland, Charles R. Honts and Steven Barger. Although the report was never classified, the government declined to publish it in the open literature. Indeed, when the results were first made known to the respective agencies involved there was tremendous pressure to classify the entire report. One of the authors, in fact, was forbidden by his parent agency from publishing or presenting the results. As a concession to the agencies involved, the association of the agency names with their performance data was classified. A copy of the report is at http://truth.boisestate.edu/raredocuments/bhb.html.
The study reports on three mock espionage experiments using different polygraph screening techniques. In Experiment One, 94% of the innocent subjects were cleared, but only 34% of the guilty subjects were identified as deceptive. Thus, the false negative rate (i.e., guilty individuals being declared innocent) was a staggering 66%. Experiment Two correctly classified only 79% of those who were innocent and 93% of those who were guilty. Finally, Experiment Three identified 90% of the innocent subjects and 81% of the guilty subjects. It is important to note that the examiners used in these experiments were trained federal polygraphers who regularly conducted periodic national security tests for their agencies. Following this primary study, four follow-up studies were conducted by the Department of Defense. The results of each supported and strengthened the findings of the primary study.
Professor Honts, one of the primary authors of the DoDPI study and a strong advocate of the polygraph, has harshly criticized the federal government’s use of polygraph testing for screening purposes. I strongly recommend that the Committee review two of his articles on the topic: “The Emperor’s New Clothes: Application of Polygraph Tests in the American Workplace”, Forensic Reports, 4:91-116 (1991)(available at http://truth. boisestate.edu/raredocuments/ENC.html), and “Counterintelligence Scope Polygraph (CSP) Test Found To Be Poor Discriminator”, Forensic Reports, 5: 215-218 (1992)(available at http://truth.boisestate.edu/raredocuments/CSP.html).
With respect to specific incident polygraph studies, from which Mr. Kiefer derives his statistics from, there have been many studies regarding the reliability of the polygraph
when used in this manner. The resulting figures have varied widely. Though somewhat dated, let me recommend one report in particular for review. In November 1983, the Office of Technology Assessment (“OTA”) issued a report entitled “Scientific Validity of Polygraph Testing: A Research Review and Evaluation”. The OTA compiled the results of six prior reviews of polygraph research, ten field studies, and fourteen analog studies that it determined met the minimum scientific standards. The results were as follows:
1) Six prior reviews of field studies:
- average accuracy ranged from 64% to 98%.
2) Ten individual field studies:
- correct guilty detections ranged from 70.6% to 98.6% and averaged 86.3%;
- correct innocent detections ranged from 12.5% to 94.1% and averaged 76%;
- false positive rate (innocent persons found deceptive) ranged from 0% to 75% and averaged 19.1%;
- false negative rate (guilty persons found nondeceptive) ranged from 0% to 29.4% and averaged 10.2%.
3) Fourteen individual analog studies:
- correct guilty detections ranged from 35.4% to 100% and averaged 63.7%;
- correct innocent detections ranged from 32% to 91% and averaged 57.9%;
- false positives ranged from 2% to 50.7% and averaged 14.1%;
- false negatives ranged from 0% to 28.7% and averaged 10.4%.
These statistics led to the enactment of The Employee Polygraph Protection Act of 1988, 29 U.S.C. § 2001 et seq. The Act outlawed the use of polygraph screening tests in the private sector. Prior to enactment, it was estimated that each year at least 400,000 honest workers were wrongfully labeled deceptive and suffered adverse employment consequences. However, the federal government was exempted from the legislation.
Given that there are no studies that support either the need or usefulness of this exemption, the Committee should consider legislation to have it removed.
2. Mr. Kiefer opines that, if Robert Hanssen 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 have passed
polygraph tests. Is there any reliable basis to estimate the probability that a
particular person would or would not pass a polygraph test?
Mr. Kiefer’s statement was worded perfectly for use in live testimony in order to generate shock value, but it has absolutely no basis in fact. It is no more based on reality than the magic of pulling a rabbit from a hat. Indeed, as described above, the only government studies available on screening examinations reveal that guilty individuals are
far more likely to escape detection than even an innocent person will be falsely accused - as high as 66% of the time.
However, more than anything Mr. Kiefer’s statement illustrates the enormous significant dangers that exist with respect to polygraph screening and the negative impact it can have on federal employees. Mr. Kiefer served as a distinguished Special Agent of the Federal Bureau of Investigation for more than two decades, including many years as a polygrapher, and is a former past president of the American Polygraph Association. Based on my experiences, his strong bias is quite typical of government polygraphers in general. With that type of obvious bias revealed publicly, it is not unreasonable to assume that such an attitude during an examination would have negative consequences on many innocent individuals simply because the polygrapher personally believed something was suspect.
In any event, for purposes of my response, let us presume Mr. Kiefer’s statement is accurate and Mr. Hanssen would have registered deceptive in a routine screening examination. What then would have occurred? Based on all publicly available information concerning Mr. Hanssen’s case - and as my legal practice substantially involves national security matters, I am following the investigation very closely - there is little, if any, incriminating evidence that would have been discovered through a follow-up investigation. The overwhelming evidence against Mr. Hanssen was obtained directly from a foreign source or agent. Unlike other spies such as Aldrich Ames, Harold Nicholson, or Edward Howard, there was no suspicious evidence of significant debt, serious employment disputes, drug or alcohol abuse or marital difficulties that would likely have prompted additional investigations and the exposure of espionage activities. Therefore, even if Mr. Hanssen had registered deceptive - and there is no scientific basis to conclude this to be so - the result would have likely been no more indicative of a truthful result as that of a false positive.
While it appears so simple to discuss Mr. Hanssen’s case in retrospect, we cannot use the knowledge we possess now in order to analyze the possible scenarios that could have occurred had a polygraph examination been administered. For all anyone knows, a deceptive reading five, ten or fifteen years ago would have meant Mr. Hanssen was being falsely accused of something he never did, as occurs every year to federal employees and applicants, and his career would have unfairly suffered as a result.
3. Everyone acknowledges that "false positive" polygraph examinations can occur
in which innocent people will show deceptive reactions. In addition, Mr. Kiefer
estimates that "there might be a maximum of 3 spies in a population of 10,000."
Assuming for the sake of argument that Mr. Kiefer's estimate of the frequency of
espionage is correct:
a. Is it not likely that if you give polygraphs to 10,000 people in order to
catch the three spies, you will get hundreds of false positive responses?
b. Assuming that the three spies all fail their polygraph tests, they would be
only three out of perhaps hundreds of employees who failed the test. How are
investigators going to be able to find the three real spies and not unfairly
cast suspicion on all of the innocent employees who have false positive results?
Attorney General John Ashcroft recently admitted that there exists a 15% false-positive rate. “Spy-Wary FBI Agrees to Polygraphs”, Los Angeles Times, Mar. 2, 2001. Based on this figure, up to 1,500 individuals will be falsely accused of espionage. Even applying the most conservative false-positive figures, say 1%, then 100 individuals will be stigmatized in order to catch three spies. This hypothetical scenario became a reality at the Central Intelligence Agency following the arrest of Aldrich Ames in 1994. Approximately 300 employees had their careers put on hold, some for as long as six years, until they were finally exonerated of any wrongdoing. Some have likely never recovered from the experiences, nor will they.
Given existing policies at the federal agencies, it is virtually impossible to ensure that unfair suspicion will not be conferred on individual employees during a witch hunt for a spy. This is the essence of the public policy balance that this Committee must address. Is it fair and appropriate to knowingly ruin innocent careers while on a fishing expedition for a spy who likely will never be exposed by the polygraph? In my opinion, it is not.
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?
Obviously, I do not. Indeed, this is the very issue that is being litigated in Croddy et al. v. FBI et al., Civil Action No. 00-0651 (Mar. 15, 2000 D.D.C.)(EGS) and John Doe #6 et al. v. FBI et al., Civil Action No. 00-2440 (Oct. 11, 2000 D.D.C.)(EGS). Federal agencies routinely rescind conditional job offers based solely on polygraph results. I would respectfully refer you to the pleadings in these two cases for further discussion of the relevant legal analysis. Copies can be found at the following websites: www.nopolygraph.com, www.stopolygraph.com and www.antipolygraph.org. Based on my experiences, I would recommend that either screening eligibility tests are eliminated or that a requirement be imposed that a background investigation must first be conducted to collaborate any polygraph results before the information can be considered in the employment decision.
9. 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?
Although the American Polygraph Association, the Employee Polygraph Protection Act and many state licensing laws prohibit inquiry into such areas as religious beliefs or affiliations, beliefs or opinions regarding racial matters, political beliefs or affiliations, beliefs, affiliations or lawful activities regarding unions or labor organizations and sexual preferences or activities, there are few prohibitions imposed upon the federal government. For example, the United States Secret Service routinely questions applicants on sexual behavior, both lawful (premarital sex) and unlawful (sexual involvement with animals).
The only means by which to ensure certain areas of inquiry are forbidden is to require the federal government to comply with the Employee Polygraph Protection Act. While some exceptions may be necessary, no agency should be permitted to question individuals on topics that do not reasonably relate to the skills needed to adequately perform the position in question.
With respect to the release of private information, there are essentially no existing safeguards. The extent to which a federal agency can disseminate polygraph results to other federal, state or local agencies is governed by the Privacy Act of 1974, 5 U.S.C.
§ 552a et seq. The sharing of information is explicitly permitted under the Act’s routine use exception. Id. at § 552a(b)(3).
For example, the FBI maintains a system of records - JUSTICE/FBI-002 - within its Central Records System that pertains to applicants for employment with the FBI. The system includes all records and information relevant to an applicant’s investigation, personnel inquiry, or other personnel matters. The FBI may disclose all personal information and records - even if inaccurate - from this system as a routine use to any federal agency where the purpose in making the disclosure is compatible with the law enforcement purpose for which it was collected, e.g., to assist the recipient agency in conducting a lawful criminal or intelligence investigation, to assist the recipient agency in making a determination concerning an individual's suitability for employment and/or trustworthiness for employment and/or trustworthiness for access clearance purposes, or to assist the recipient agency in the performance of any authorized function where access to records in this system is declared by the recipient agency to be relevant to that function.
As a result of this ability to freely share information, individuals who falsely registered deceptive on one agency’s polygraph examination may have that information used against them by another agency, without ever being given an opportunity to challenge the underlying allegation of deception. Unfortunately, the enactment of additional legislation will be required to minimize the extent to which a federal agency
can disseminate information pertaining to polygraph examinations. Current law is clearly inadequate.
Questions Submitted by Senator Grassley
1. Let’s say that an employee polygraph exam ends with a deceptive result but with no admission of guilt. How do agencies deal with this situation? How about with an inconclusive result?
Unfortunately, it is difficult to provide a precise answer to this question as procedures differ from agency to agency. Typically, however, should either of the situations occur above, the agency will initiate further investigation into the individual’s background and activities. Oftentimes, the employee may be transferred to a non-sensitive or less sensitive position and may even have promotions withheld. On paper, the employee may very well not suffer an adverse personnel action. By this I mean, they will continue to hold employment and remain at the same pay grade.
The most recent example describing this type of circumstance is that of the FBI. By Memorandum dated March 16, 2001, the FBI announced it would institute counterintelligence-focused polygraph examinations to employees who occupy certain assignments or occupations. With respect to those employees who experience trouble with the polygraph, the Memorandum noted:
Experience has shown that most FBI employees taking the counterintelligence-focused polygraph examination successfully complete the test. However, there may be a very small number of employees whose tests are either inconclusive or are indicative of deception. Polygraph examiners will attempt to fully resolve all unexplained responses through the effective use of thorough pre-and post-test interviews. If, upon completion of a thorough examination, there is still an inconclusive or deceptive response, it will be considered “unexplained”. Consistent with existing policy, no adverse action will be taken based upon the polygraph results alone. However, more extensive investigation will be initiated to resolve the unexplained test results.
However, realistically, an employee in this situation will unequivocally suffer the equivalent of an adverse personnel decision. Some agencies, such as the CIA and FBI, have taken years to finally resolve a false-positive or inconclusive polygraph result. Some employees may be suspended with pay, which is not always considered an “adverse action”. Employees at the CIA who found themselves in such a position were not permitted to attain overseas assignments. This is often the end of a career for individuals
employed within the Directorate of Operations. Scientists under contract at the Department of Energy who experience polygraph problems will find themselves transferred to other positions, which often would negatively impact upon their careers. In my written testimony, I described the situation of FBI Special Agent Mark Mallah. In his case, it took approximately two years of intensive and intrusive investigation before he was finally exonerated. He was so disgusted by how he was treated, he resigned in protest. Unfortunately, Special Agent Mallah’s reaction is not unusual, and the U.S. government has lost many fine employees strictly because of false polygraph results.
3. Will there be adverse consequences for employees who refuse to take a polygraph examination?
Again, this can differ from agency to agency. However, most agencies will react in a similar manner. For example, the FBI Memorandum referred to above states that those employees who refuse to take the test will be subjected to administrative actions which may include transfer, a finding of insubordination and disciplinary action or a reevaluation of the employee’s security clearance.
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?
Again, by viewing this question solely by the legal definition of “adverse action” (such as those actions that can be appealed to the Merit Systems Protection Board,
5 U.S.C. § 1201.3), the conclusion would be accurate. However, as I described above, reality dictates otherwise. For all intents and purposes, the employee does suffer “adverse consequences”, though it might not legally be in the form of an “adverse action”.
This question, however, does raise a larger issue. If such a prohibition exists with respect to employees, why should applicants receive any less consideration? How “logical” is that? There is no question that FBI applicants who have received a conditional offer of employment, but who then fail their polygraph examination (or register inconclusive) are not afforded the opportunity of a background investigation. Their job offer is immediately rescinded. More than that, the polygraph result is maintained in that individual’s personnel file, and will be freely disseminated as permitted by law. One polygraph examination may stigmatize an individual throughout the federal government thereby precluding their future employment and contribution to the United States.
There is something inherently wrong and unfair with the current federal polygraph policies that are implemented throughout the different law enforcement and intelligence
agencies of our government. Without intervention by this Committee, there is little chance these policies will ever change.
I trust this additional information proves to be useful. I would be happy to elaborate further upon any question, or respond to additional inquiries.
Mark S. Zaid
cc: Senator Patrick Leahy
Ranking Minority Member
Senator Charles Grassley
Senator Arlen Specter
Senator Richard Durbin
The agencies have since been identified as the Army INSCOM, the Air Force Office of Special Investigations, the National Security Agency and the Central Intelligence Agency.
AntiPolygraph.org Home Page > Reading Room