On 2 September 1999, I sent the following message to Secretary of Energy Bill Richardson and the directors of Lawrence Livermore National Laboratory, Sandia National Laboratories, and Los Alamos National Laboratory. They did not reply -- George Maschke,
DEPARTMENT OF ENERGY 1000 INDEPENDENCE AVE SW WASHINGTON DC 20585 Dear Secretary Richardson: I urge you to rescind your proposed polygraph rules and to use your discretionary authority to halt all polygraph screening of Department of Energy and contractor employees. In your notice of proposed rule-making and public hearings regarding polygraph policy dated Aug. 18, 1999, you invite the public to "comment on the balance [you have] struck in [your] proposal between legitimate national security interests and regulatory limitations to protect employees from inappropriate or imprudent use of polygraph examinations and the results of such examinations." Your proposed rules fail to protect both national security interests and employees. While you may indeed have struck some sort of "balance" by failing on both scores, I hope to convince you that your proposed rules are unwise and that you should not implement them. Your decision to rely on Counterintelligence Scope Polygraph (CSP) exams fails to protect our national laboratories from espionage because such polygraph screening is without validity. As Dr. Drew C. Richardson of the FBI Laboratory testified regarding polygraph screening before the Senate Committee on the Judiciary in 1997 (1): [Polygraph screening] is completely without any theoretical foundation and has absolutely no validity. Although there is disagreement amongst scientists about the use of polygraph testing in criminal matters, there is almost universal agreement that polygraph screening is completely invalid and should be stopped. As one of my colleagues frequently says, the diagnostic value of this type of testing is no more than that of astrology or tea-leaf reading. Dr. Richardson further testified: If this test had any validity (which it does not), both my own experience, and published scientific research has proven, that anyone can be taught to beat this type of polygraph exam in a few minutes. Polygraph screening has absolutely no validity and anyone can be taught to beat this type of polygraph exam in a few minutes. Don't you suppose that any spies at our national laboratories will take those few minutes and more to learn to beat the polygraph? In para. II of your notification of proposed rule, you state, "DOE is aware of no scientific studies that establish that polygraph examination results are unreliable for use as an investigative tool, as DOE today has proposed to use them." You disregard one of the most fundamental principles of rational discourse: it is incumbent on the one making an affirmative assertion to prove it (not on others to disprove it). Is the DOE aware of any scientific studies that establish that polygraph examination results *are* reliable for use as you propose?! As Dr. Richardson testified, there is almost universal agreement among scientists that polygraph screening is completely invalid and should be stopped. Your alleged unawareness of scientific studies establishing that polygraph examination results are not reliable for counterintelligence security screening leads to the inescapable conclusion that you have not seriously studied the matter. I refer you first to a comprehensive review of the available scientific literature on polygraph validity conducted by the U.S. Congress' Office of Technology Assessment (OTA) in 1983 (4). It states: OTA recognizes that the administration as well as NSA, CIA, and DOD believe that the polygraph is a useful screening tool. However, OTA concluded that the available research evidence does not establish the scientific validity of the polygraph for this purpose. (p. 100) The OTA report goes on to state: One area of special concern in personnel security screening is the incorrect identification of innocent persons as deceptive. All other factors being equal, the low base rates of guilt in screening situations would lead to high false positive rates, even assuming very high polygraph validity. For example, a typical polygraph screening situation might involve a base rate of guilt of one guilty person (e.g., one person engaging in unauthorized disclosure) out of 1,000 employees. Assuming that the polygraph is 95 percent valid, then the one guilty person would be identified as deceptive but so would 50 innocent persons. The predictive validity would be about 2 percent. Even if 99 percent polygraph validity is assumed, there would still be 10 false positives for every correct detection. (p. 100) The OTA review assumes that a polygraph screening validity rate of 95% entails that 95% of guilty subjects will be detected. But with an extremely low base rate of guilt, as is the case with espionage, such an assumption is not warranted. If we allow that not more than one in a thousand persons examined are actually spies, then a validity rate of at least 99.9% can be achieved simply by ignoring the polygraph charts altogether and peremptorily declaring all examinees innocent. Of course, the usefulness of such a "test" for catching spies would be zero. Yet this is essentially how the CSP achieves its claimed validity of 99.9%! The interpretation of polygraph charts is manipulated so that almost everyone "passes." In 1991, Professor Charles R. Honts (3) reviewed Department of Defense data on the CSP examination and concluded: Research and analyses conducted on the Department of Defense's Counterintelligence Scope Polygraph (CSP) Screening Program indicate that the polygraph tests used in that program are unable to discriminate truthtellers from deceivers. It appears that the CSP polygraph examinations correctly classify only about 2% of the guilty subjects. Effective countermeasures exacerbate this problem and may render the CSP Screening Program completely ineffective at detecting deception. (p. 91) A laboratory experiment conducted by Professor Honts also showed the CSP exam to be a poor discriminator (2). The OTA report also acknowledges that deceptive persons can beat the polygraph by employing physical countermeasures, concluding: ...[T]he evidence, while limited, is that deceptive subjects who use physical countermeasures and who can distinguish non-relevant from relevant questions (in a CQT or R/I test) can increase their chances of avoiding detection. The CSP exam is a form of the Control Question Test (CQT) referred to above. David T. Lykken, an emeritus professor at the University of Minnesota, explains how to defeat the polygraph in chapter 19 of the 2nd edition of his often-cited study, A Tremor in the Blood: Uses and Abuses of the Lie Detector (5). Polygraph screening also entails a tremendous ethical problem. You, as Secretary, should be aware that in conducting the Counterintelligence Scope Polygraph, the polygraph examiner will formulate "control" questions (they are not a control in any scientific sense) to ask of each emloyee for purposes of comparison. These "control" questions do not involve matters of espionage, but are more general, often along the lines of, "Have you ever betrayed a friend's trust?" The polygraph examiner tells the examinee that he must answer all questions truthfully, but he actually assumes that the examinee will be deceptive when answering these "control" questions. He deceives the examinee about that expectation. As the OTA report (4) explains at p. 20: The polygraph examiner does not tell the subject that there is a distinction between the two types of questions (control and relevant). Control questions are described as intending to determine if the subject is the "type of person" who would commit a crime such as the one being investigated.... The examiner stresses that the subject must be able to answer the questions completely with a simple "yes" or "no" answer, that the polygraph will record any confusion, misgivings, or doubts, and that the subject should discuss any troublesome questions with the examiner.... Thus, the situation is set up such that the subject is persuaded that the examiner wants the truth. In reality, however, the examiner wants the subject to experience considerable doubt about his or her truthfulness or even to be intentionally deceptive.... The polygraph examiner, assuming first that the examinee is being deceptive when answering the "control" questions, assumes next that if the examinee's physiological responses as measured by the polygraph are stronger when answering a relevant question (e.g. "Have you made classified information available to a foreign government?") than when answering the "control" questions (e.g. "Have you ever lied to a supervisor?"), then the examinee must have been deceptive in answering the relevant question. The conscientious examinee who, at the polygraph examiner's urging, "discuss[es] any troublesome questions with the examiner" and then truthfully answers the "control" questions (and thus exhibits weaker physiological responses to the "control" questions than to accusatory relevant questions like,"Are you a spy?") is the most likely to be found deceptive! As Honts (3) writes: ...Lykken [(5)] has persuasively argued that the individual who tries to be truthful during a pre- employment polygraph examination and who, at the examiner's urging, bares all of his or her past wrongdoing, is the very individual who is most likely to be rejected by the preemployment screen- ing process, whereas the individual who makes minor admissions and then dishonestly maintains his or her innocence is more likely to be given the benefit of the doubt and passed through.... (p. 98) The same phenomenon of which Lykken speaks regarding preemployment polygraph exams also applies to the CSP exam. Indeed, the CSP is commonly used for preemployment purposes. In every polygraph screening exam, at least one person truly is deceptive: the polygraph examiner. This trickery is unethical, and it boggles the mind that you intend to practice it against men and women who have dedicated their lives to discovering the nature of things. You are demanding of some of America's most loyal, dedicated scientists that they prove their trustworthiness to you, when it is *you* who are not worthy of *their* trust. What do you intend to do with those informed and honest scientists who, in answer to their polygraph examiner's question, "So, what do you know about the polygraph?" reply, "I know that polygraph screening has absolutely no validity, that you are going to formulate "control" questions in reply to which you expect me to lie, that if I answer all your questions truthfully, you will be more likely to conclude that I am being deceptive, that it is many times more likely that you will falsely accuse me or one of my colleagues than that you will catch any real spies, and, oh yes, I am also familiar with and know how to employ polygraph countermeasures"? If you have not already contemplated this situation (and I suspect you have not), you had best do so before implementing your proposed rules. Scientists at the national labs will soon know much more than you want them to know about polygraphy. The use of polygraph examinations for security screening has a validity approaching zero, and even if it had some validity, anyone can still be taught to beat the polygraph in a matter of minutes. By substituting cheaper polygraph exams for more expensive background investigations, you are shirking your duty to protect America's atomic secrets. Indeed, to the extent that you rely on polygraph examinations, you undermine the national security: spies who "pass" by employing countermeasures will escape further scrutiny, while innocent scientists who "fail" be will targeted for rigorous investigation, and their research will be adversely affected. And our atomic weapons programs will lose the services of those principled scientists who refuse to be a party to this modern-day trial by ordeal. Your proposed rules also fail to protect your and your contractors' employees from polygraph abuse (even were we to accept the premise that polygraph screening does not in and of itself constitute abuse): 1) Section 709.4(b)(3) This section allows the Secretary of Energy to exempt anyone he pleases from the polygraph requirement if he certifies that it is in the interest of "national security." One suspects that the key criterion here is being a friend of the Secretary. If the polygraph is truly as reliable an indicator of truth and deception as DOE claims, why should *anyone* be exempted? How could that be in the national interest? 2) Section 709.11(b) and (c) Subsection (b) limits polygraph questions to topics concerning the examinee's involvement in espionage, sabotage, terrorism, unauthorized disclosure of classified information, unauthorized foreign contacts, or deliberate damage to or malicious misuse of a U.S. government information or defense system. And Sec. 709.11(c) provides that DOE may not ask questions that: (1) Probe a person's thoughts or beliefs; (2) Concern conduct that has no security implication; or (3) Concern conduct that has no direct relevance to an investigation. However, the "control" questions that a polygraph examiner must ask in an attempt to elicit a "deceptive" response for comparison purposes will necessarily run afoul of the prohibition on questions concerning "conduct that has no direct relevance to an investigation." 3) Section 709.13(a) DOE claims that polygraph exams will be "voluntary." Sec. 709.13(a) states, "...An individual may refuse to take a polygraph examination, and an individual being examined may terminate the examination at any time." But the consequence of refusing or terminating a polygraph examination is suspension or denial of access to classified information. You lose your job, or you don't get hired. DOE's claim that the polygraph is "voluntary" is an intellectually dishonest exercise in casuistry. 4) Section 709.14.(c) Regarding the consequences of refusal to be polygraphed, this section states in part, "If the individual is a DOE employee, DOE may reassign or realign the individual's duties or take other action, consistent with that denial of access." What is meant by "other action?" Re-assignment to a dead-end job at some remote DOE facility? Reduction in pay? Will DOE look for some other excuse to fire these employees? 5) Section 709.24(b) This section requires that before administering the polygraph examination, the examiner must "[e]xplain to the individual the characteristics and nature of the polygraph instrument and examination." Is the polygraph examiner to explain to the employee that he has crafted "control" questions and assumes that the employee will be "deceptive" when answering them? And that if the examinee truly feels that he is answering all questions honestly, he is likely to be found "deceptive?" If not, and the polygraph examiner is to do his utmost to deceive the examinee into believing that the polygraph is virtually infallible and to deceive him about the true nature of the "control" questions (in conformance with the standard procedures endorsed by such prestigious organizations as the Department of Defense Polygraph Institute and the American Polygraph Association), then I suggest that you strike the words "and examination" from Sec. 709.24(b). 6) Section 709.25(a) This section states that DOE and its contractors may not "[t]ake an adverse personnel action against an individual solely on the basis of a polygraph examination result of "deception indicated" or "no opinion" except when the Secretary or the Secretary's designee makes a written determination that the information to which the individual has access is of such extreme sensitivity that access under the circumstances poses an unacceptable risk to national security or defense..." This means that any employee whose polygraph results are "deception indicated" or "no opinion" may indeed have his or her access to classified information suspended based on the polygraph alone. Career over. In the current political atmosphere, can anyone seriously imagine that the Secretary or his designee will not revoke the clearances of employees whom the FBI says lied about counterintelligence issues during their polygraph exams? 7) Section 709.26(e) This section states, "DOE must afford the full privacy protection provided by law to information regarding an employee's refusal to take a polygraph examination." Such privacy protection is virtually impossible to ensure. Counterintelligence officials will doubtless share information on the polygraph "refuseniks," who, one suspects, will have a hard time ever holding a security clearance again. When such employees are reassigned for their refusal to be polygraphed, how will their personnel and security files explain their sudden departures from their former jobs for new, less prestigious ones? What are they to tell their colleagues? Mr. Secretary, I hope I have convinced you to reconsider your proposed program of polygraph screening. I would remind you of the dénouement of the Hans Christian Anderson fable, "The Emperor's New Clothes," with which Professor Honts opens his review of polygraph screening in the American workplace (3): In a sudden blinding flash the vain Emperor saw how he had been tricked and cheated. But being a mighty Emperor, with purple blood in his veins, he held his head all the higher and walked on and on down that long street. I urge you not to repeat the Emperor's mistake. Don't think that you must continue on this misguided path. Your proposed rules fail to protect national security interests and they fail to protect employees. Please exercise your authority as Secretary of Energy to rescind these proposed rules, and halt polygraph screening within the Department of Energy. Ask Congress for the funding you will need to conduct thorough background investigations with periodic updates for scientists with access to atomic secrets. You will catch more spies (if there indeed are any at the national labs), provide real rather than illusory deterrence against espionage, and avoid falsely accusing the innocent. Sincerely, George W. Maschke gmaschke@ucla.edu ----------- References: 1. Drew C. Richardson. Testimony to the Senate Committee on the Judiciary, Sep. 29, 1997. Available on-line at: http://www.nopolygraph.com/drewtest.htm 2. Honts, Charles R. "Counterintelligence Scope Polygraph (CSP) Test Found to be Poor Discriminator," Forensic Reports, 5 (1992):215-218. [Since the date of my letter, the author has made this article available online]: http://truth.boisestate.edu/raredocuments/CSP.html 3. ---. The Emperor's New Clothes: "Application of Polygraph Tests in the American Workplace," Forensic Reports, 4 (1991):91-116. [Since the date of my letter, the author has made this article available online]: http://truth.boisestate.edu/raredocuments/ENC.html 4. Scientific Validity of Polygraph Testing: A Research Review and Evaluation -- A Technical Memorandum. Washington, DC: U.S. Congress, Office of Technology Assessment, OTA-TM-H-15, November 1983. http://www.wws.princeton.edu/~ota/disk3/1983/8320.html 5. Lykken, David T. A Tremor in the Blood: Uses and Abuses of the Lie Detector. 2nd ed. New York: Plenum, 1998.