Go back to the SPSE Home Page http://www.spse.org/
Comments from the hearing in Albuquerque,
New Mexico.
September 16, 1999
DOUG ADKINS
First, I would like to say that I'm here on my own time. I just couldn't justify charging this to any Sandia case.
Second, after hearing about the phenomenal accuracy of polygraph testing, I decided to voice my support for the tests, and would like to further suggest that polygraphs might have much broader application in addressing societal problems.
The main problem that should be addressed is the cynicism that the American public has towards its political leaders. The corrosive effect of this cynicism could be eliminated if every candidate for public office is expected to take a polygraph test as a normal part of getting elected.
Now I know that engineers and scientists may be skeptical of anything as scientifically suspect as polygraph tests, but fortunately our political leaders have no such reservations.
As Rep. Wilson so eloquently stated, "polygraph tests are just another tool that should be available to investigators." Certainly, the voters deserve the same tool in selecting leaders.
To be fair, candidates should not be asked lifestyle questions, but they should be asked a few policy-related questions. For instance:
By having politicians require us to take polygraph tests, we are in a unique position to request that politicians do the same. Let us work together and have NM become the first state where political candidates are routinely asked to take polygraph tests.
DAVID B. BALDWIN
As we struggle to understand the motivation for this effort to change the Regulations to allow for polygraphy at the labs, there are too many unanswered questions. I am getting frustrated because no one seems to be able to answer a number of simple questions.
A few such questions are as follows. What is the intended outcome of this effort? Is polygraphy being introduced as part of a routine screening process? Or is a sample population going to be selected for a "counterintelligence-scope polygraph examination"? Or is it both? Our contact person has advised me that she too would like the answers to these questions, and she expressed her own frustration at DOE's lack of communication. I have read and re-read the proposed rule and am unable to find any definitive answers. The "technical briefing" presented by the "men in black" team last week raised more questions than it answered. Ladies and gentlemen, this whole situation is beginning to stink and draw flies, but it doesn't have to.
We all want strong national security. I served my country with the 5th Special Forces Group in the Republic of Vietnam. I came through the battle of Loc Ninh in '67 and the TET offensive in '68. I was awarded the Bronze Star Medal with the V device for "heroism in ground combat", and I still carry in my side a piece of shrapnel from a 122mm rocket. I value a strong national defense. I especially value our constitution because I have seen first hand what can happen to people who don't have its protection. I fought for our country when Bill Clinton was at Oxford dodging the draft and bashing our country. Consequently, I think I have a bigger investment in national security than most. I also have no interest in seeing a bunch of lawyers get rich because a poorly considered process was implemented in haste. If this must be done then there is a right way to do it.
It seems clear that we are all going to have to learn to live with polygraphy. From my own experience with polygraph testing I believe that if one has nothing to hide one has nothing to fear. However, most of the people I've talked with are either apprehensive or angry about it. Some have even said they will refuse the test. What happens if we all refuse? It is one thing to introduce polygraph testing as part of a routine screening process, and it is quite another to compel a sample of a population to "prove" their loyalty and innocence without probable cause. Therefore, the DOE must be extremely careful avoid even a hint of a witch-hunt. Here are my suggestions on how to do it right.
This will lay out the ground rules up front and answer many of the questions that people have. I believe that a random selection will at least partially mitigate the apprehension and possibly some of the anger that is present. If the DOE continues on its present course a lot of folks are going to become alienated. Don't forget that alienation was the reason that Albert Einstein, Niels Bohr, Enrico Fermi and others chose to work for us rather than Germany. If you come storming in here with a McCarthy style witch-hunt, it will aggravate rather than relieve the anger and resentment that already exists.
On the other hand, we at Sandia National Laboratories have proven time and time again that we can do anything we set our minds to. So if you want real solutions to real problems, then let's pull together, roll up our sleeves and get to work.
David B. Baldwin
PAULINE DOBRANICH
My name is Pauline Dobranich. Although I am employed by Sandia National Laboratories, my comments reflect my personal concerns. I have six comments on the proposed rule on polygraph examinations.
Comment #1: Part 709.4 describes who will be required to take a polygraph. Item 6 states that polygraphs will be required for those positions which involve access to information on the design and operation of nuclear weapons and associated use and control features. Because of the vagueness of this statement, it is not clear who will be eligible: all people with Q clearances or perhaps a subset. Because of this ambiguity, the DOE does not know how many people will be impacted by polygraph examinations. Thus, the DOE cannot properly prepare for conducting polygraphs nor predict the adverse effects.
Comment #2: Parts 709.13 and 709.14 describe the polygraph examination process as voluntary. Yet if the individual is an incumbent in a position where polygraphs will now be required, the DOE may deny the individual access to that information or involvement in those activities. Thus, an individual could lose their position for refusing to take the polygraph examination. Obviously the polygraph examination is not voluntary, and it is ludicrous for the DOE to suggest otherwise.
Comment #3: Part 709.15 briefly describes the process for polygraph examinations and the follow-on evaluation process. The proposed rule does not specify how long the process will take or whether the individual retains their clearance during the process. The eligibility evaluation panel and their qualifications are not defined. The individual has neither legal protection nor is there a process for the individual to appeal the decision from the eligibility evaluation. The individual does not even receive a copy of their records. This process is unsatisfactory because it does not protect the legitimate interests of existing employees.
Comment #4: As described in Section II (entitled "Background"), the President has instructed DOE to develop and implement specific measures to protect highly sensitive and classified information at its facilities. Can the DOE demonstrate that polygraph examinations will provide better protection of highly sensitive and classified information? In 1998, the Supreme Court reaffirmed that polygraph results are inadmissible in court due to reliability concerns. Rather than depend on unreliable polygraph examinations, the DOE should improve the quality of background investigations.
Comment #5: In Section IV, Subsection B (entitled "Regulatory Flexibility Act"), the DOE certified that the proposed rule will not have a substantial impact on a significant number of small businesses. Because the DOE does not know who or how many people will be affected by polygraph examinations, they also do not know how many of these people are employed by small businesses. Therefore, the DOE cannot certify that the proposed rule will not have a substantial impact on small businesses.
Comment #6: In Section IV, Subsection E (entitled "Treasury and General Government Appropriations Act, 1999"), the DOE states that the proposed rule would not have any impact on the autonomy or integrity of the family as an institution. What is the basis of this statement? Because the DOE has not specified the duration of the eligibility evaluation, the individual could be in "limbo" for an extended period of time. This causes concerns about whether the individual will continue to have a job, concerns about making mortgage payments, and morale problems associated with a tarnished reputation. I think that the DOE is extremely naïve to believe that this will not impact the family.
In summary, although Section IV, Subsection I (entitled "Executive Order 12988") states that the DOE has a duty to provide a proposed rule with clarity and without ambiguity, my comments have identified several areas where the DOE has failed to be clear and has not considered the ramifications of the proposed rule. Therefore until the ambiguities have been addressed and the impacts have been evaluated, the DOE should abandon the proposed rule on polygraph examinations.
STANLEY FRALEY
[Webmaster's Note: Dr. Fraley also submitted written comments which may be seen below.]
I am Dr. Stanley Fraley.
The essence of my comment is this:
"The proposed polygraph screening program, if implemented, will result in damage to individuals and to the United States. The proposed use of the polygraph as a means of screening employees to detect 'spies' would not increase national security. On the contrary, it threatens national security."
I make this statement from a number of different viewpoints. First, as a scientist, I can find no scientific basis supporting the use of polygraphs as an effective screening tool. Second, is my viewpoint as an individual who has taken polygraphs for national security screening purposes. It took me five separate four hour sessions before my responses were deemed to be "well within the norm" for individuals not being deceptive. Prior to that experience I naively and wrongly believed that since I had nothing to hide, and since I didn't feel anxious about answering the relevant questions, then I should sail right through the polygraph.
After the first session I realized, "The polygraph is not a lie detector, and it clearly is not a truth detector." The polygraph is a tool for inquisitors to use to try to elicit confessions. It does not provide a reliable indication when you are telling the truth, and further it cannot detect lies or "deception."
Regarding the word deception, unfortunately, there is deception taking place. One deception is that they will only ask four simple questions. That is not true. A typical "session" lasts one or more hours. This is so the examiner can "discuss" each of the questions with you and ask you to elaborate on any reasons why you might have anxiety or concerns about any of the four questions. This inquisition is the real goal of the process, not the time you spend attached to the machine. After the questions are asked with you attached to the polygraph, the examiner typically will then have a second "informal" discussion with you. This is so you can explain why you seemed to react to the questions under the polygraph. Take the examiners word for it. You did react. There must be something that bothers you that you haven't mentioned. In the end, it will be the examiner's subjective judgement as to whether you are being deceptive.
I am sure that the inquisitors during the Spanish Inquisition believed that their actions were useful and necessary for rooting out heresy. They could point to confessions that "justified" their actions. The practitioners of polygraphy use similar justifications.
However, the major deficiency in the use of polygraphs is not that it is simply an instrument for the inquisition of overwhelmingly innocent employees. It is that it is not an effective tool for detecting spies. Individuals can be trained to pass the polygraph test even when they are lying. Many individuals do not even need training to deceive a polygraph examination. Further, this screening is currently used to accelerate the clearance process in lieu of a more lengthy and thorough background investigation. As a result, the use of the polygraph procedures as proposed, and even as now practiced, is a threat to our national security.
The DOE notice in the Federal Register states that DOE is "aware of no scientific studies that establish that polygraph examination results are unreliable for use as an investigative tool," as DOE has proposed. DOE claims to be unaware of such studies! It might be more accurate to characterize this lack of knowledge as "clueless." There is a significant body of scientific evidence that addresses this issue directly. Others here at Sandia, and possibly individuals previously providing comments, have noted a 1983 Office of Technology Assessment report that suggests that there is up to a 50-percent chance that an individual will be falsely accused of lying. I personally have drawn heavily from testimony before the US Senate Committee on the Judiciary on September 29, 1997 given by Dr. Drew Richardson, a scientist who worked in polygraph research.
There is scientific evidence that is being ignored. This evidence is also very easy to find.
It is especially significant that even DOE does not claim that there are scientific studies that establish that polygraph examination results are reliable for use as an investigative tool.
General Habiger has stated that he wishes to restore the special trust of Congress and the American people in the DOE. If DOE proceeds with this program, then I believe that it will have demonstrated that it should not be trusted by its employees, by Congress, or by the American people, either to treat its employees ethically, or to protect national security interests.
In closing, I wish to reiterate, the proposed polygraph program does not and can not tighten up DOE security. It is a facade that represents a clear danger to our national security.
[Webmaster's Note: These are Dr. Fraley's written comments. His oral comment made at the DOE hearing can be seen above.]
9/27/99
Douglas J. Hinckley
Office of Counterintelligence
CN-1
Docket No. CN-RM-99-POLY
U.S. Department of Energy
1000 Independence Avenue SW
Washington, DC 20585
The following are my written comments regarding the proposed polygraph regulation and on the regulatory review process given in the associated notice in the Federal Register.
I am representing myself.
Dr. Stanley K. Fraley
1417 San Rafael Ave. NE
Albuquerque, NM 87122
The regulations should be modified to specify that "In no case shall individuals be required to undergo a polygraph examination under these provisions more often than once every five years, unless the Secretary of Energy makes a written determination that special circumstances exist that pose an unacceptable risk to national security and that warrant a request to a specific individual to consent to a polygraph examination at a date less than five years since a previous examination under this regulation."
The regulations should be modified to reflect the full range of possible results. Any possible results other than "deception indicated" and "no opinion" should be indicated. The revisions in the regulations to accomplish this should be made available for further public comment.
The regulations should be modified to specify the basis for arriving at a result of "deception indicated" or "no opinion." The regulation should also specify that until such time as there is a clearly demonstrated scientific basis for arriving at the result "deception indicated" that has a demonstrated reliability of at least 99% then the determination of "deception indicated" is not to be used. The regulation should stipulate that the scientific basis for such a determination should be accepted by a panel of experts brought together by the National Academy of Science.
If a result such as "no deception indicated" is also to be a possible result of the examination, then the regulation should specify that until such time as there is a clearly demonstrated scientific basis for arriving at the result "no deception indicated" that has a demonstrated reliability of at least 99% then the determination of "no deception indicated" is not to be used. The regulation should stipulate that the scientific basis for such a determination should be accepted by a panel of experts brought together by the National Academy of Science.
The regulation should state that until such time as a scientific basis for a determination of "deception indicated" and "no deception indicated" meeting the above criteria have been made, the only acceptable result shall be "no opinion."
These two proposed exceptions should be deleted. There is no basis for assuming that Presidential appointees should be exempt from the requirements of the regulation. There is also no basis for assuming that the Secretary of Energy could have a valid reason for excluding any specific individuals "in the interest of national security."
During the public hearings, DOE Officials have stated that the scope of the proposed polygraph examination will be limited to four specific questions that are a sub-set of the counterintelligence topics listed in 709.11 (b). The regulations should be modified to specify that limitation.
The regulation in 709.25 should be modified to read:
"DOE or its contractors may not:
(a) Take an adverse personnel action against an individual solely
on the basis of a polygraph examination result; or
(b) Use a polygraph examination as a substitute for any other
required examination."
The regulation should be modified to add additional qualifications for examiners in order to ensure adequate training in the analysis of physiological data. Specifically, all examiners should be licensed physicians. They should be licensed to practice medicine in the District of Columbia and in the State within which the examination will be conducted.
The proposed regulation should be modified to include a requirement for all examiners to have a favorably adjudicated, full-field Federal Bureau of Investigation background investigation completed within the last five years; a yearly financial disclosure statement made public record, and a yearly psychological examination performed by a licensed psychiatrist. This examination should examine the examiners emotional stability and ability to maintain objectivity while making judgments about people that affect national security.
The examiners should also fulfill all of the requirements deemed necessary for individuals under the PSAP and PAP programs. This would include, inter alia, random drug testing, and the requirement to restrain from the use of alcohol for a period of 24 hours prior to conducting a polygraph examination. This latter requirement is necessary in order to ensure that the judgment of the examiner is not impaired. Tests (such as urine tests) to demonstrate the fulfillment of this requirement should be performed before the examiner is allowed to proceed to conduct an examination. A two-person rule should remain in continuous effect after testing for the presence of alcohol consumed in the past 24 hours in order to ensure that alcohol is not consumed after testing but prior to the completion of all polygraph examinations.
The proposed regulation should be modified, first to provide legal counsel of the individual's choosing at the expense of DOE. Second, at the choice of the individual being examined, the counsel or representative should be permitted to be present during the polygraph examination. The methods used for the examination should be "beyond reproach" and therefore the presence of the council or representative should not have an adverse effect on the efficacy of the examination.
The proposed regulation should be modified to require a copy of all records, including video recordings, made during the examination be provided to the individual for use by the individual and the individual's legal counsel or representative. The individual should also be provided with the examiner's report of the examination. This report should include explanatory information regarding the basis for any conclusions reached. If the individual disagrees with the conclusions reached by the examiner, then the individual should be entitled to a hearing with an appropriate independent review board to review these conclusions. This review board should be empowered to decide on the appropriate results of the examination for the purposes of the examinations conducted under this provision. The majority of the members of the review board should be selected from individuals and organizations that are not associated with the DOE polygraph program, including those organizations listed under Part 709.31 of the proposed rule.
1. The provisions of the National Environmental Policy Act were not followed.
The Federal Register notice states that DOE has determined that neither an environmental assessment nor an environmental impact statement is required because the proposed rule is "strictly procedural" and therefore excluded from such requirements. This is incorrect.
The Federal Register notice states that "DOE has determined that this rule is covered under the Categorical Exclusion found in the Department's National Environmental Policy Act regulations at paragraph A.6 of Appendix A to subpart D, 10 CFR part 1021, which applies to rulemakings that are strictly procedural."
The referenced paragraph A.6 defines the category for which exclusions apply by virtue of being "strictly procedural." The paragraph reads:
"Rulemakings that are strictly procedural, such as rulemaking (under 48 CFR part 9) establishing procedures for technical and pricing proposals and establishing contract clauses and contracting practices for the purchase of goods and services, and rulemaking (under 10 CFR part 600) establishing application and review procedures for, and administration, audit, and closeout of, grants and cooperative agreements."
The proposed set of rules will require the establishment of a number of offices to administer the proposed polygraph examinations. This includes adding additional people whose salaries and expenses will be covered by DOE. It also includes new office space requirements for the additional personnel, for the locations at which the examinations will be conducted, and for the records and review processes that will be an important part of the proposed process. The description in paragraph A.6 as to the types of rulemakings that are "strictly procedural" does not fit the proposed polygraph examination program. The proposed rule is not simply "procedural." An environmental assessment or an environmental impact statement is required.
2. The provisions of the Regulatory Flexibility Act were not followed.
Agencies are required by the RFA either to certify that "the rule will not have a significant economic impact on a substantial number of small entities," and provide a factual basis for the determination, or to prepare a regulatory flexibility analysis. The "factual basis" provided by DOE is incorrect.
The subject notice in the Federal Register states that "There may be some affected small businesses that are subcontractors, but the rule would not impose unallowable costs." This statement neglects significant impacts on small entities.
One impact is the ability of the entity to recruit and retain employees who are willing to undergo polygraph testing. Many small entities will likely correctly judge that the adverse impact on employee hiring and retention if they wish to bid on contracts requiring polygraph examinations of employees would be greater than the benefits of obtaining the contract. This will impact these small entities.
A second impact is the time element associated with the necessary clearance process once a determination is made that the entity would like to bid on a contract that would require some employees to have polygraph examinations. Since the numbers of individuals with such clearances will be kept to a minimum, it is unlikely that all small entities will be allowed to request employees be given polygraph examinations just so the entity will be eligible for making a bid. However, because of the time involved in obtaining the necessary clearances, it is also unlikely that the entity will able to effectively compete. This will impact these small entities.
An initial regulatory flexibility analysis (IRFA) must be prepared
and published in the Federal Register. The DOE should at least
publish a summary of this analysis and make the full analysis
available upon request. This initial analysis should address,
at a minimum, the following:
Under Section 603(b) of the RFA, each initial regulatory flexibility analysis is required to address: (1) reasons why the agency is considering the action, (2) the objectives and legal basis for the proposed rule, (3) the kind and number of small entities to which the proposed rule will apply; (4) the projected reporting, recordkeeping and other compliance requirements of the proposed rule, and (5) all federal rules that may duplicate, overlap or conflict with the proposed rule.
While these five factors are necessary elements to an adequate IRFA, they are not the sole factors necessary to perform an adequate analysis. Most important, section 603(c) requires that each initial regulatory flexibility analysis contain a description of any significant alternatives to the proposal that accomplish the statutory objectives and minimize the significant economic impact of the proposal on small entities. These alternatives could include the establishment of differing compliance or reporting requirements or timetables that take into account the resources available to small entities; the clarification, consolidation or simplification of compliance and reporting requirements under the rule for small entities; the use of performance rather than design standards; or an exemption from coverage of the rule or any part of the rule for small entities.
A description of the steps the agency has taken to minimize the significant economic impacts on small entities consistent with the stated objectives of applicable statutes, including a statement of the factual, policy and legal reasons for selecting the alternative adopted in the final rule, and the reasons for rejecting each of the other significant alternatives.
3. The provisions of the Unfunded Mandates Reform Act of 1995 were not followed.
The notice in the Federal Register states that a detailed assessment is required under the referenced act on a proposed rule if the anticipated costs and benefits may result in costs to State, local, or tribal governments, in the aggregate, or to the private sector of $100 million or more. The notice further states that "this rule, as proposed, is not likely to result" in such costs or more in any one year.
The Unfunded Mandates Reform Act applies to:
"any rule that includes any Federal mandate that may result in the expenditure by State, local, and tribal governments, in the aggregate, or by the private sector, of $100,000,000 or more (adjusted annually for inflation) in any 1 year,"
The statement in the notice that "it is not likely" that such costs would occur is an admission that such costs "may result." Therefore the assessment required under the Unfunded Mandates Reform Act must be performed.
4. The provisions of the Treasury and General Government Appropriations Act, 1999 were not followed.
The notice in the Federal Register notes that the referenced act requires a Family Policymaking Assessment for any proposed rule "that may affect family well-being." The notice further states that "today's proposal would not have any impact on the autonomy or integrity of the family as an institution" and that it is not necessary to prepare a Family Policymaking Assessment. This latter statement is incorrect.
Section 654 of the Treasury and General Government Appropriations Act, 1999 requires an assessment with respect to whether
First, the DOE assessment criteria of "would not have any impact on the autonomy or integrity of the family as an institution" is wrong. The assessment needs to be performed with respect to criteria listed in the Act (as listed above).
When the relevant assessment is made, it is clear that the proposed regulations would affect family well-being.
The proposed rule states that the polygraph is used "for the purpose of rendering a diagnostic opinion regarding the honesty or dishonesty of an individual." As a result, the spouse and any other family members of individuals subjected to a polygraph have reason to believe that the results of the examination provide an indication as to whether or not the individual is honest and can be trusted. Trust is an essential element of any family unit. The proposed procedure will unjustifiably call this trust into question and adversely affect the well-being of a number of families. This action would erode the stability of the family. It would also erode the authority of parents.
Additionally, because the family is a cohesive unit, any procedure that calls into question the honesty of one part of the family also calls into question the honesty and integrity of the family as a whole. This will also adversely affect the family well-being. This action would erode the stability of the family. It also adversely affects the relationship and personal responsibility of youth, and the norms of society.
The polygraph examination process is a very stressful process. It does have an adverse effect on a significant number of individuals, regardless of whether or not they successfully complete the process. Such significant stress in one element of a family unit will be reflected on the family as a whole. This will adversely affect family well-being. This action would erode the stability of the family.
And finally, the proposed rule states that individuals have the right to legal counsel, but only at the individual's own expense. This action will have an adverse financial impact on the family and decrease disposable income of families and children.
The formal assessment required by this appropriations act is necessary.
5. The provisions of Executive Order 12866 were not followed.
The notice in the Federal Register states that section 6 of Executive Order 12866 provides for a review by the Office of Information and Regulatory Affairs of regulatory actions that, inter alia, adversely affect public health or safety. The notice further states that DOE has concluded that this proposed rule is not a significant regulatory action. This conclusion is not justified.
Executive Order 12866 defines a significant regulatory action as:
(f) "Significant regulatory action" means any regulatory action that is likely to result in a rule that may:
(1) Have an annual effect on the economy of $ 100 million or more or adversely affect in a material way the economy, a sector of the economy, productivity, competition, jobs, the environment, public health or safety, or State, local, or tribal governments or communities;
(2) Create a serious inconsistency or otherwise interfere with an action taken or planned by another agency;
(3) Materially alter the budgetary impact of entitlements, grants, user fees, or loan programs or the rights and obligations of recipients thereof; or
(4) Raise novel legal or policy issues arising out of legal mandates, the President's priorities, or the principles set forth in this Executive order.
First, the proposed rule will impose significant stress on a large number of individuals. As a result, the proposed rule does affect public health.
Second, the proposed rule may have a significant adverse effect on the qualifications of individuals who will be required to assure the safety of the nuclear weapons stockpile. This could increase the risk of an accident and the consequences of an accident involving nuclear weapons. This potential, if realized, could adversely affect public health and safety in the extreme.
Finally, the proposed polygraph examination program is raising novel legal and policy issues.
A review by the Office of Information and Regulatory Affairs is required.
6. The provisions of Executive Order 12988 were not followed.
The notice in the Federal Register states that the referenced executive order requires, inter alia, that any proposed regulation adequately define key terms. The proposed regulation does not meet that requirement.
The proposed regulation uses the term "polygraph examination results." This term is key in the determination of actions that can be taken following polygraph examinations. This key term is not defined. The proposed regulation indicates that two possible results are "deception indicated" and "no opinion" and implies that other results are possible. The range of possible "polygraph examination results" is clearly not defined.
The proposed regulation uses the term "deception indicated." This is a key term and it is not defined. Given the significance of a potential finding of deception indicated it is very important that this term be defined. Since a polygraph is defined in the proposed regulation as an instrument "used, or the results of which are used, for the purpose of rendering a diagnostic opinion regarding the honesty or dishonesty of an individual," it is also important that the definition of the term makes it clear as to whether or not such a result carries the implication that the individual being tested is diagnosed as being dishonest. The definition should include the expected false positive rate that will be used in rendering such a diagnostic opinion and include the scientific basis that will be used to support this determination.
The proposed regulation uses the term "no opinion" as a possible result of a polygraph examination. This is a key term and it is not defined. Given the significance of a potential finding that does not render a diagnostic opinion that an individual is being honest, it is very important that this term be defined. Again, since the polygraph is defined in the proposed regulation as an instrument "used, or the results of which are used, for the purpose of rendering a diagnostic opinion regarding the honesty or dishonesty of an individual," it is also important that the definition of the term makes it clear as to whether or not such a result carries an implication that the individual being tested cannot be diagnosed as being honest. The definition should include the expected false positive rate on honest individuals that will be used in rendering such a diagnostic opinion and include the scientific basis that will be used to support this determination.
There are clearly possible results of the polygraph examination other than the ones mentioned above. These results will also involve key terms that should be defined. To the extent that any of the terms are related to the rendering of diagnostic opinions as the honesty or dishonesty of an individual being tested, then the definitions should address the false positive and false negative rates that will be used in rendering these "diagnostic opinions" and the scientific basis used to support the determinations of the false positive and false negative rates.
STEVEN HUMPHREYS
Steven Humphreys' comments delivered orally during the public hearing.
Thank you for this opportunity to be here today. There are several sections of the proposed new 10 CFR Part 709 that I wish to comment on.
Section 709.14 What are the consequences of a refusal to take
a polygraph examination?
(a) ... DOE and its contractors may refuse to employ, assign,
or detail the individual to the identified position.
(c) ... DOE may deny that individual access to the information
or involvement in the activities that justified conducting the
examination.
Section 709.23 states that: DOE may not administer a polygraph
examination unless DOE has: (c) Obtained voluntary written consent
from the individual.
If an individual feels that he or she must consent to a polygraph
examination to hold a new position, compete for a promotion, or
even to continue working in the same area where they have built
their professional reputation, can that be called voluntary? Agreeing
to a polygraph examination to prevent such negative repercussions
may also affect polygraph test validity. In Scientific Validity
of Polygraph Testing: A Research Review and Evaluation --A Technical
Memorandum (OTA-TM-H-15, pg. 101, November 1983), the Office
of Technology Assessment wrote:
... conducting polygraphs on this basis could affect test validity. It is generally recognized that, for the polygraph test to be accurate, the voluntary cooperation of the individual is important. For example, NSA has stated that, in conducting screening examinations, "the full cooperation of the individual taking the test is essential or the results will be inconclusive." The polygraph only detects physiological arousal, and under involuntary conditions, the arousal response of the examinee may be very difficult or impossible to interpret.
The provisions of section 709.14 are inconsistent with the requirements for polygraph examinations to be voluntary and possibly even valid. I recommend that section 709.14 be changed so that employee applicants and employees who refuse to take a polygraph examination would have their access authorization eligibility determined using DOE's other investigative tools and techniques. Present employees who become eligible for polygraph examinations and refuse to take them, would continue to keep their present level of access authorization eligibility until a reinvestigation yielded information that warranted their access authorization eligibility to be downgraded or terminated.
Section 709.15 How does DOE use polygraph examination results?
Paragraph (c) DOE will conduct an eligibility evaluation that
considers examination results, the individual's personnel security
file, and other pertinent information. As part of the eligibility
evaluation process, DOE may interview the individual.
As stated in section 709.25 DOE or its contractors may not:
(a) take an adverse personnel action against an individual solely
on the basis of a polygraph result of "deception indicated"
or "no opinion"...
The eligibility evaluation described above would, in many cases,
bring together information that was already known before the polygraph
examination. If action is taken against an individual after the
eligibility evaluation and the only new information is the result
of the polygraph examination, then that process would violate
the requirements in section 709.25. To better meet those requirements,
I recommend that the words "an eligibility evaluation"
be replaced with "a full-field investigation."
Section 709.22 What rights to counsel or other representation
does an individual have?
... The counsel or representative may not be present during the
polygraph examination.
I recommend that an individual be allowed to have up to two representatives in the same room during the polygraph examination. If it is expected that classified questions will need to be answered, then the representatives must have the appropriate access authorization. If unexpectedly, classified questions need to be answered, and the representatives do not have the appropriate access, then the examination will be halted and the examinee will be given at least 48 hours (subject to exemptions of 709.21) to obtain representatives with the appropriate access authorization.
In addition to the above sections, to ensure that all polygraph examinations that are given properly, I recommend that Part 709 include a section that describes the minimal set of records that must be taken during the pre-test interview, polygraph test, and post-test interview. I recommend that Part 709 specify video recording of the examiner and examinee during the interviews and polygraph test, and also the polygraph trace during the polygraph test. The required written materials should also be specified. I also recommend that the examinee be allowed to examine all of the records after the polygraph examination.
Comments delivered in writing at the public hearing.
Section 709.4 To whom does the polygraph examination requirement
under this part apply?
Paragraph (b) This part does not apply to:
Paragraph (1) A Presidential appointee, if such an appointee has
received a favorably adjudicated, full-field Federal Bureau of
Investigation background investigation;
Many of the National Laboratory's employees have received a favorably adjudicated, full-field Federal Bureau of Investigation background investigation and will be asked under Part 709 to submit to a polygraph examination. If the Department of Energy believes that it is in the best interest of National Security for these employees to submit to a polygraph examination, then I believe that it is in the best interest of National Security for Presidential appointees to submit to a polygraph examination. For this reason I recommend that the Presidential appointee polygraph examination exemption be removed from Part 709.
GLENN KUSWA
I am Glenn Kuswa, President Elect of the New Mexico Academy of Science. I am also a manager at Sandia National Laboratories, where I have been employed for nearly 30 years, including a total of six years on extended assignments to DOE Headquarters in Washington, DC.
I first wish to present a short statement from the New Mexico Academy of Science.
The New Mexico Academy of Sciences believes that there is inadequate scientific basis supporting the efficacy and reliability of polygraph testing. The incidence of false positive outcomes and the resulting harm to individuals make polygraph testing an unfair and inappropriate tool in a free society.
I next wish to present some personal views that suggest some direct harms that will arise from wide use of polygraph testing, and mention some alternatives that may improve security.
Polygraph exams seemingly answer frustrations of some political leaders and security professionals, because the technique appears to be a scientific means that rapidly detects security risks. The weaknesses in polygraph testing are self-evident because of failures to detect proven spies, for example, in the Ames case, and because there is a substantial failure rate, requiring stressful repeat tests, sometimes without ultimate resolution. The false failure rate is a subject of much discussion and debate, fed by incompletely reported data, often shrouded by secrecy, and sometimes published without the benefit of peer reviews that are common in most of science. I will not enter that debate here, except to note that polygraph error estimates range from a few percent to twenty percent or more.
False polygraph results cause hardship to persons who fail exams. No matter what is said about the way failed individuals will be continued in employment, their trust, earning capacity, and opportunities for service to the nation and promotion are diminished, perhaps for an entire career that may last 30 years or more. The proposed policy appears to require that employees take the stress, inconvenience, and risk of these unfair and largely unscientific exams with little or no benefit in return, other than the hopefully continued honor to serve the country.
The DOE system should continue to attract and to hire the best possible candidates. We need the best because we cannot afford to be mere caretakers of the complex defense technologies developed in the past, but we have to work on advancements and improvements. Merely preserving our secrets is an open invitation for our adversaries to overtake us in a matter of a few years, even if our secrets remain protected. Bright people have their pick of jobs, and will go where they can most readily apply their talents. Weapons programs mandate publication reviews, and restrict some discussions to a defined need-to know community; these reasonable rules are followed by our work force, but they are a necessary disincentive when hiring. Polygraph testing will likely be a much more potent repulsive force in staff hiring and retention.
An ordinary construction project requires an environmental impact study to assure the well-being of plants and creatures as low as insects and worms. Protection of the individual is what this hearing is all about, but it falls far short of studies aimed at lower life forms. Action should await two studies:
1) How will polygraph testing influence hiring and employee retention programs? Such a study must be done in a way that informs participants fairly of risks and the potential benefits and defects associated with polygraph testing, because these facts are not widely known, but will surely emerge after a program is put in place.
2) Trusted and unbiased legal and medical experts who would be given free access to statistics kept secret by other agencies that have long used polygraph testing should examine the usefulness of polygraphy.
I question the quality and vision of the program to be set up. We were informed in an introductory briefing by the DOE last week that the existing limited DOE polygraph program had been assessed as flawless in its procedures during a recent certification audit. But audits are not the only test for quality. The head of the testing program stated that there is no plan for training of examiners in the DOE, except to require periodic refresher courses, and that the program is to be operated "by the book" with 'no room for inventiveness or imagination', using vetted examiners from other agencies. A technique as fraught with uncertainties as polygraph testing should not be treated as so routine.
Many of the testers work on contract to the government. This might indicate a shortage of qualified testers, and gives little comfort to test subjects. Continued practice that seems to restrict development of new talent could escalate future costs. I am also concerned that one government agency should feel justified in hiring examiners away from other agencies, rather than planning together to share resources and to provide for future needs.
Examples of proven ways to improve security include more sting operations conducted within careful ethical practice, random inspections of employees entering and leaving the work place, more thorough background investigations, instructing all employees to be more cognizant of the signs and traits associated with disloyal persons, etc. Such activities yield tangible evidence and proof, and therefore appear to be more fundamentally fair. The only substantial stress on employees would be on those being examined for a specific reason, and the false accusation rate after investigation would be very small, perhaps zero.
STEWART SILLING
Comments for Polygraph Hearing by Stewart Silling
In the late 19th century, it was thought by some that you could tell whether a person was a criminal or not by measuring the shape of his head. This technique had two problems: false positives, and false negatives. But national security was at stake. Society had to be protected from criminals. So, many authorities thought this was a useful test.
In the days of witch trials, a suspected witch was sometimes made to recite the 23rd Psalm or other text. If she mispronounced any of the words, this proved she was a witch. This technique had two problems: false positives, and false negatives. But national security was at stake. Society had to be protected from witches. So, many professional witch hunters thought this was a useful test.
Of course, DOE would never conduct a witch hunt. But why is it so easy to see the absurdity in this witch test and so hard for many to see it in the polygraph test, when the two methods are almost identical? Both methods are based on the premise that if you get nervous when you say something good, then you must be bad.
A leading authority on polygraph testing from DOD, who spoke here at Sandia last week, admitted that polygraph testing has two problems: false positives and false negatives. But national security is at stake. Society must be protected. So, these problems don't matter.
The Security Czar has been quoted as saying the reliability of polygraph testing is 99.9%, a claim that some would say is outlandish. But let's assume the figure is correct. This means that out of 1000 people, perhaps one false positive will result. People who are ignorant of mathematics would then be 99.9% certain this person is a spy.
Chances are, this false positive person is simply predisposed to failing polygraph tests, so he would also fail a second or third or fourth test. By then, the pressure on the investigators to dig up some dirt on him would be overwhelming, because no one wants to go 0-for-1000. They might find out that this guy once ate in a Chinese restaurant. Or that he has a Persian cat.
Meanwhile, a real spy would be well versed in the techniques for passing the test. One of the most damaging spies in U.S. history, Aldrich Ames, repeatedly passed polygraph tests at the CIA. Ames could never have gone undetected for so long without being shielded by his exemplary polygraph test results. What more conclusive proof could you ask for that polygraph testing is not only worthless, but actually damages the national security rather than enhancing it?
Soldiers sometimes have to make the ultimate sacrifice for the country, and perhaps the rest of us should not complain too much about making lesser sacrifices, such as merely losing our jobs and reputations. But it is one thing for a general to order his troops into battle. It is quite another for him to order them to play Russian Roulette.
To address the insider espionage threat, DOE should adopt reasonable methods as surveillance, improved security of computer systems, and possibly even undercover operations.
Let's put polygraph testing where we put cold fusion, pyramid power, and astrology: in the trash can that contains discredited and evil ideas.
ALAN P. ZELICOFF
[Webmaster's Note: Dr. Zelicoff also submitted written comments which may be seen below.]
In government, in academia, and at the National Laboratories it is common practice to reflect on an event and summarize the lessons learned during the course of that event. Usually this exercise is carried out after a period of reflection, sober thought, or perhaps after a few so not-so-sober moments with colleagues. Regrettably, we at the DOE laboratories will not have that luxury, as Gen. Habiger has already stated that there will be no debate, and Ms. Lowe has indicated in off-the-record conversation that even she (the lawyer for the panel, no less) is not certain what rule changes (if any) would mandate a re-hearing to discuss the results of this process. It would be far below the standards of professional scientific conduct if I did not comment on the disingenuousness of this approach. Gen. Habiger has stated that the rules no longer permit debate on the scientific merit of polygraphy; yet, you are ostensibly here to listen to the scientists from the premier laboratories in the US government, scientists who have prepared careful reviews of the scientific literature on polygraphy, and who have raised legitimate and unanswered questions on the arbitrariness of the exam, particularly as relates to people with medical problems, on medications, or who may otherwise be very different from the populations upon which you base uncritically reviewed conclusions about the utility, veracity and robustness of the test. You are perfectly prepared to dismiss all of this work. Thus, in my view, this is not a hearing; it is not even a "listening tour" as you have forestalled reasoned debate, and it is not clear how we will receive answers to our questions. Based on the non-responsiveness of Dr. Barland and Mr. Renzelman during and after the "technical briefing" (your term, not mine), I am skeptical that you will make a good faith effort to address the concerns, nor will you respond to follow-up questions should your answers prove to be as incomplete as the technical briefings provided today.
What have I learned today? I have learned that post-graduate study at the DOD Polygraphy Institute does not include instruction in the principles of the scientific method, nor does it imbue in its students the responsibility to carry out research subject to the review of scientific colleagues and publication in scientific journals. I have learned that even for educated people, and 4-star generals with the rank of "czar" (an odd notion in our pluralistic society) that the need to do something overwhelms the need to do something useful, data be damned. I have learned that for this panel the search for truth appears to be not nearly as important as producing sound-bites for the media. In short, I have learned that the metrics of accuracy, reliability, safety, and security that you demand of us, your employees at the National Laboratories, you do not demand of yourselves. Leaders should set the standard. You have failed. Shame on you.
What is my empirical evidence for these conclusions? First, let me address the technical. Dr. Ryan and Dr. Barland before him addressed this and other audiences of scientists and engineer with the following:
In short, the presentations you have made today, and at the technical briefing are an insult to this scientific community or any group of scientists. Your credibility, which did not begin on a high plane to start with was further undermined by your poor science and your preference to tell anecdotes. Sorry, but you will have to do better than this.
Next, allow me to review empirical evidence of a political nature. On the 12:00 news today, Gen. Habiger stated: "I don't understand why the people who are entrusted with our nuclear weapons would object to answering 4 simple questions". General, I understand that while serving with the "most ethical Administration in history", you labor under the murkiness of Presidential confusion as to the definition of what the word "is" is, and the sudden reclassification of direct lies uttered on national television into mere "misleading statements". It must indeed be challenging to set new ethical standards under such conditions. But General, we are all adults here so let's get real. Isn't it just possible that your statement about 4 simple little questions was a bit incomplete, perhaps misleading? Do you not wonder why people doubt your sincerity of purpose during your listening tour when you fail to note that polygraphs take at least an hour, and that it is not unusual for them to go on for four hours? For 4 simple questions General? I don't expect that the ability to simple statistics is a requirement to get 4 stars -- after all, that would require doing long division with a calculator. But, until today, I had assumed that integrity was one necessary box to check. My colleagues and I will have to re-examine that assumption.
Further General, if the evidence presented by Sandia scientists today does not make the case for a re-examination of the utility of polygraphs then nothing will convince you. You can not make chicken salad out of chicken feathers General. Until today, I had assumed that keen judgement, skepticism, and the ability to question so obvious a group of self-interested technology peddlers was another requirement for getting a star. This is another assumption I will have to re-examine.
In short General, you have used the opportunity of this public hearing to politicize a critically important issue, and while you may have "listened" there is no evidence that you have "heard". I fear that in your zeal to show firm decisiveness, you have capitulated to flimsy demagoguery. Instead of intellect, you have given us only attitude. This is the worst kind of arrogance -- the arrogance of a "know-nothing".
But, let me end on a more positive note. With the vote of "no confidence" today from Senator Bingaman, this matter has suddenly be lifted out of obscurity into grist for inside-the-Beltway debate. The Senator has probably done himself short-term harm with some of his colleagues on the Hill, and maybe even in the White House. I, for one, admire him. As I have had some modest contact with his staff over the past few weeks, I believe he has become energized by the sheer intellectual emptiness of this rule making process, and the panel's arrogant dismissal of all contrary scientific evidence and its willingness to sacrifice national security on a political altar. So, I guess I should be grateful to you for that much. The Senator's insistence on a well-structured review process, sound science, and, if necessary, rejection of a bad-idea is the stuff that makes this country a great place. The debate is not over. The debate has only begun. You have guaranteed it.
Thank you.
[Webmaster's Note: These are Dr. Zelicoff's written comments. His oral comment made at the DOE hearing can be seen above.]
Response to DOD-PI polygraph questions
Alan P. Zelicoff, MD
October 9, 1999
Dr. Andrew Ryan
Chief of Research
DOD Polygraph Institute
By e-mail: ryana@jackson-dpi.army.mil and by fax: (803) 751-9125
RE: DOD-PI responses to technical questions during DOE Polygraph Hearings
Dear Dr. Ryan:
First, I thank you for your efforts in answering the questions I submitted to you on September 15, 1999. I have reviewed each of your answers as well as each of the articles you referenced in response. Unfortunately, I find your answers quite incomplete from a scientific standpoint; in fact, most of the references you provided to me strongly underscore the lack of utility of polygraphy as a screening tool. Further, the references barely begin to address the questions I put to Dr. Barland and to you during the polygraph "technical" meeting or the subsequent formal DOE Polygraph Hearings led by Gen. Habiger. There may have been some misinterpretation on your part, and I will attempt to once again clarify each of my questions in the discussion below.
I will take each of your replies to my questions in turn, and for your convenience and that of anyone else who may wish to read this correspondence, I will repeat the original questions and each of your answers, followed by a critique of the references as well as the answers themselves. Of course, I would be most pleased to receive any follow-on commentary from you.
Question #1: What happened to the 15 people who had a "significant" response on the DOD-PI polygraph study?
Your answer: "I assume you are referring to the table that presented actual data from the DoD polygraph program for the FY '98. This was not a research study, rather the information that was reported to congress in the Department of Defense Polygraph program Annual Report to Congress Fiscal Year 1998. The answer you are seeking can be found in this report beginning on page 4 and continuing on page 5. I would forward a copy to you but we do not have any additional copies at DoDPI."
Critique and Commentary: I would greatly appreciate a brief summary of the report that specifically addresses the 15 people who had a significant response in the DOD-PI FY1998 study that was presented during the "technical" meeting at Sandia and the other DOE National Laboratories. However, I am somewhat surprised to read your characterization of this work as something other than a "research study"; it was certainly presented that way during the briefings given at Sandia on September 7 and September 19. If it is not a research study, and merely constituted part of a report to Congress, can we agree that it did not meet widely accepted scientific methodological standards required for journal publication as critiqued by several Sandia and Los Alamos staff during the briefings?
It is clearly the view of Sandia staff that the work in the DOD-PI report to Congress does not qualify as being scientifically credible, mostly due to the poor experimental design, lack of characterization of the population, and several potentially erroneous measurement techniques detailed to General Habiger on September 19. Thus, I conclude that this work has no validity in countering conclusions on the non-utility of polygraphy in a screening mode.
Question #2: What reference(s) does Dr. Barland have which demonstrates "no effect of commonly used medication" on the utility of polygraphy?
Your answer: "The statement made by Dr. Barland referred to the studies listed below:
(a) Bradley, M. T. & Ainsworth, D. (1984). Alcohol and the psychological detection of deception. Psychophysiology, 21, (1): 63-71.
(b) Iacono, W. G. Boisvenu, G. A. & Fleming, J. A. (1984). Effects of diazepan and methylphenidate on the electrodermal detection of guilty knowledge. Journal of Applied Physiology, 69, (2): 289-299.
(c) Iacono, W. G. Cerri, A. M. Patrick, C. J. & Flemming, J. A. (1992). Use of antianxiety drugs as countermeasures in the detection of guilty knowledge. Journal of Applied Physiology, 77 (1), 60-64.
(d) O'Toole, D. Yuille, J. C. Patrick, C. J. & Iacono, W. G. (1994). Alcohol and the physiological detection of deception: Arousal and memory influences. Psychophysiology, 31, 253-263.
(e) Kamei, K. Imamura, Y. Aoki, T Yashita, S. Yamaoka K, & Suzuki, A. (1965). Effects of tranquilizers on polygraph tests. Polygraph Reports, 80-84. National Institute of Police Science Research Materials No. 35" [I have added lettering and ordering of these references for convenience]
Critique and Commentary: I have reviewed each of these articles in detail (save for Kamei et.al. which is not readily available). First, let me note that references (b) and (c) are not in the Journal of Applied Physiology, but rather are in the Journal of Applied Psychology (emphasis added). I trust this was merely a clerical error.
Allow me to now take each of these references in turn for discussion and critique.
Reference (a) demonstrates that alcohol use shortly before taking a polygraph test does not affect the results of the Control Question Test (CQT). However, according to the authors:
"Those ingesting alcohol prior to the crime were more likely, when classified on the composite scores, to be classified incorrectly as innocent or inconclusive than those who were sober during the crime." And regarding the GKT, "alcohol effects were found with the SRR and heart rate decelerations. The SRR results showed that those who committed the crime sober had detection scores more in the guilty range than those who committed the crime while intoxicated, regardless of the test state. Test state interacted with crime state on the heart rate measure such that, only subjects who were intoxicated during both the crime and the polygraph test had low detection scores." Further, "as a countermeasures alcohol was not effective with the SRR and only effective with HR deceleration if the crime had been committed while intoxicated.
"The results are potentially of great practical importance for field interrogations since suspects who committed a crime while intoxicated would have a better chance of appearing innocent than those who committed the crime sober."
"In summary, alcohol intoxication during enactment of a mock crime affects detection rates of both CQT and GKT. Because these effects were the result of intoxication during the crime, it was speculated that emotional and memory processes were involved at this point" (emphasis added).
Thus, it appears that alcohol use at the time of a crime substantially affects the polygraph when utilized in the CQT mode. Further, while I agree that subjects' use of alcohol at the time of a polygraph did not affect correct classification of guilty subjects, the question that I asked was in the context of the effect of medication(s) on false positive classification of innocent subjects. I will return to this below.
Reference (b) looks at the effects of diazepam (a sedative benzodiazapine) and methlyphenidate (a stimulant) on the misclassification of test subjects by the polygraph in a laboratory setting. Regrettably, the polygraph test administered in this study is a guilty knowledge test (GKT), which is totally different in psychological construct and theoretical basis from the methodology utilized in the setting of a screening polygraph. The test subjects are male undergraduates of age 19-28; further, there is a 75% pre-test probability of subjects being in a guilty group. 3 or 12 "innocent" subjects were classified as inconclusive (Table I, page 293 of article). But most important (as is the case the other studies see below) NONE of the "innocent" subjects were tested while on medication. Hence, this study addresses only the ability of guilty subjects to fool polygraphy with drugs. The question we asked during the polygraph "technical" meetings and polygraph hearings referred to the effect of drugs on proper classification of innocent subjects. Therefore, this study does not address our question. Accuracy of "innocent" classification, if inconclusive results are considered, is 12/15 = 80%. Thus, the false positive rate is 20% in innocent subjects not taking medications. From this study, we simply do not know what the false positive rate would be for innocent subjects taking drugs. Finally, the sample population is hardly representative of staff members at the National Laboratories who are much older, and tend to be on at least some prescription medications.
I can only conclude that this study further underscores the high false-positive rate of polygraphy, even in the setting of the GKT which, a priori, is favorably biased toward polygraph diagnosticity. This reference does not contribute to your contention that false-positive results are of minimal significance when polygraphy is used in the field and under conditions much less favorably biased than the GKT setting, i.e. screening.
Reference (c) is a follow-on study to reference (b). My criticisms are the same as for reference (b). I also note that in this study, 2 of 15 innocent subjects were classified incorrectly as guilty. The false positive rate was thus 13%, hardly reassuring to National Laboratory staff.
Reference (d) reviews the effect of alcohol on the utility (accuracy) of polygraphy utilizing both CQT and GKT methods. The results from this paper are as follows:
Overall accuracy of Polyraph Tests.
For CQT:
- Of 64 guilty subjects, 30 (46.9%) were inconclusive
- Of 16 innocent subjects, 10 (62.5%) were inconclusive
Excluding the inconclusives,
- 25/34 (73.5%) of guilty subjects were classified as deceptive, and
- 5/6 (83%) of the innocent subjects were classified as truthful
With a post hoc manipulation of cutoff scores to maximize the number of accurate decisions while minimizing the number of inconclusive judgements:
- 12/64 (18.8%) of guilty subjects and
- 4/16 (25%) of innocent subjects were classified as inconclusive.
Excluding the inconclusives:
- 37/52 (71.2%) of guilty subjects and
- 8/12 (66.7%) of innocent subjects were correctly classified.
For GKT:
Excluding inconclusives,
- 80% (47/59) of decisions involving guilty subjects were correct
- 93.8% of innocents were classified correctly.
In short, the false positive rates in this study vary from approximately 6% to over 30%. The study suffers, of course, from being rather small, but it does not lend credibility to the notion that false positives are "rare," as stated many times by both Dr. Barland and you.
Further, I would ask you to take note of another article by W.G. Iacono (the same author in references (b) and (c)) in Psychophysiology, 28(6), pgs. 632-938 entitled: "A Comparison of Field and Laboratory Polygraphs in the Detection fo Deception." This study addresses the problems that may occur when polygraphy transitions from the laboratory (as in all of the references you provided) to the field. Specifically, Iacono notes that:
"the CQT is susceptible to false positive errors when subjects are tested under field-like circmstances, a problem that is not likely to be eliminated by refinements in instrumentation and scoring."
False positive rates of over 44% were found among innocent subjects.
Question #3: Has any DoD polygraph study been subjected to scientific review and published in a major scientific journal? If so, please name the studies and dates of publication.
Your answer: The brief history of polygraph research reflects the profession as an emerging science primarily supported by polygraph examiners and a select group of interested researchers. Moreover, funding for scientific research has been minimal and until recently most of the research was conducted by individual agencies searching to develop more efficient and more accurate methods that supported their specific mission. Therefore, examiners wishing to contribute to the knowledge base for other examiners conducted most of the historical research and the results were primarily internal agency reports. Thus, as you have noted there is a dearth of peer-reviewed articles available to the interested researcher. In recent years this situation has improved. Most significantly, the Federal polygraph research programs were consolidated and the DoDPI is now congressionally mandated to conduct research in all areas related to the detection of deception. This central facility is now producing the quality of research needed to validate the science. Additionally, significant changes (see below) are underway to improve the quality of research and to answer the vital research questions concerning the utility, validity, and reliability of polygraph and alternative methods to the detection of deception. A summary of our mission can be found on the DoDPI web page located at www.dodpoly.org."
Critique and Commentary : The operating principles DOE laboratories are founded in the scientific method. You have agreed with our position that there is a dearth of scientific articles on polygraphy published in peer-reviewed journals. I can only conclude that the scientific credibility of polygraphy, particularly as concerns field use of polygraphy in the screening mode, remains in doubt.
QUESTION #4 : If there is no "gold standard" for the true positives and true negatives, how is it possible to calculate the Bayesian "diagnosticity" of a positive polygraph result? (Bayesian diagnosticity = ratio of true positive rate/False Positive rate).
QUESTION# 5: Assuming that the DoD 1998 data is correct, what is the Bayesian diagnosticity of polygraphy? What is the Bayesian diagnosticity if the false positive rate is 10% (the lowest figure published in a scientific journal to our knowledge)?
Your answer: "I will combine questions #4 and #5 as they address the same issue. As stated in the town meeting the difficulty in conducting polygraph research is in the absolute determination of ground truth. For example, in analog studies where ground truth is programmed it is possible to calculate Bayesian diagnosticity using standard statistical procedures outlined in such books as Hayes Statistics for Psychologists and many other graduate level statistics texts for the social sciences. All DoDPI projects are subjected to this standard as well as the following criteria concerning "effect size."
Additionally, it is my opinion and numerous others, that after decades of research based on proving the null or rejecting the null around some sacred criterion is in need of review. This is not my original thought -- it is the consensus of many scientists with far greater stature than I. The inherent problems of significance testing that assumes randomization from the general population is possible or that the event/behavior actually occurs in the general population has led many to suggest the more appropriate method is to focus on effect size as opposed to probability of error in such a hypothetical population." (followed by a discussion of effect size).
Critique and Commentary: I must regard this answer as non-responsive, or in agreement with our position that the postive and negative predictive value of polygraphy is lacking when subjected to standard methodologies ("ground truth"). There is no question that more work needs to be done, and while you are entitled to your opinion, it must be held to the same standards of excellence as we perform in our certification of the stockpile (or other high-consequence ascertainments) before it can become policy of such consequential importance to the careers of individual staff and the very security of the stockpile we agree must be protected. Anything less is not only unfair; it is likely disastrous for stockpile stewardship and maintenance.
Question #6. What is the change in electrodermal potential response to questions (both calibration questions as well as standard polygraph questions) as a function of disease state or progression for the following diseases? [Please provide references from the medical literature].
Your answer: " In terms of the specific diseases listed above I am not aware of any study that addresses the specificity you are requesting for these individual diseases."
Critique and Commentary: Thank you for acknowledging that there is no scientific literature on this subject. I assume that you would agree with me that in the absence of such evidence, it remains an open question of extreme importance, especially in view of the literature that you have provided that demonstrates a high false-positive rate of classification of innocent subjects under ideal laboratory conditions. I would be pleased to review the classified reference1 you have noted, but I am very skeptical that it is either of proper scientific design, or that any results therein would reverse the body of literature that you (and I) have referred to above.
Summary and Conclusions
The literature you have provided is, in general, seriously flawed in the following ways: the study groups are not representative of the DOE population; there is an extremely high false positive rate for misclassification of innocent subjects; there are no studies that rule out the effect of chronic disease states or medication use on the false positive misclassification rates of polygraphy for innocent subjects; and there are documented deficiencies when polygraphy is taken from an ideal laboratory environment utilizing the guilty-knowledge test (GKT) and moved to far more chaotic field applications where screening polygraphy is attempted. By your own admission, there is no published scientific evidence that polygraphy has any utility (and, in all probability has a high dis-utility) when used in the mode proposed by the Secretary of Energy and General Habiger: screening of large numbers of National Laboratory staff.
It is disturbing in the extreme that the senior leadership of the Department of Energy would launch such a sweeping use of polygraphy when the basis of the test is in so poorly documented, even in the very literature that experts in the DOD and DOE quote as justification for its application. I believe I speak for the vast majority of staff members at the DOE National Laboratories when I assert that such shoddy scientific basis and reasoning for the application of screening polygraphy as presented by DOE and DOD technical briefers would never pass muster in a high school science fair, let alone in professional technical circles. While we all agree that there are security concerns at the laboratories that must be addressed, screening polygraphy is clearly of no utility in this regard.
Because the DOE Implementation Rule calls for screening polygraphy (or is at best ambiguous on who is subject to polygraphy and under what circumstances), the scientific references you have provided eviscerate the justification for the policy the Rule seeks to implement. Obviously, the need to complete revamp the Rule is manifestly clear. Additional review and hearings should also be held as part of the Department's due diligence should the Department decide to pursue any use of polygraphy in the National Laboratories.
There are only three possible outcomes from efforts to address our security problems: we can make the situation better, things may remain the same, or we can take a bad situation and make it substantially worse. From the literature that you have provided to me, there is no question that use of screening polygraphy falls into the latter category.
Surely we can do better than this.
Respectfully Yours,
Alan P. Zelicoff
Senior Scientist
Center for National Security and Arms Control
Sandia National Laboratories
Albuquerque, NM 87185-5800
Phone: (505) 844-8020 FAX: (505) 844-0600
e-mail: apzelic@sandia.gov
1 Cail-Sirota, J., & Lieberman,
H. R. (1995). (DoDPI95-R-0005). Fort McClellan, AL: Department
of Defense Polygraph Institute. NOTE: For Official Use Only.
DTIC# ADB211213 [Back to the text.]
|
Go back to the SPSE Home Page http://www.spse.org/ |
