DSS Withholds R/I Screening Documentation
Who Do They Think They're Fooling? And What Are They Hiding? The Defense Security Service (DSS) has denied AntiPolygraph.org's 5 Sep. 2001 Freedom of Information Act
request for "[a]ll Department of Defense Polygraph Institute (DoDPI) materials describing the R/I (Relevant/Irrelevant) Screening Test, whether on paper, videotape, computer media, or in any other format."
In its
letter of denial dated 9 Nov. 2001, DSS (DoDPI's parent agency) claims that the requested material is exempt from the Freedom of Information Act because it "relates solely to the internal rules and practices of the Department of Defense, Defense Security Service" and constitutes "investigative techniques and procedures, the release of which could reasonably be expected to risk circumvention of the law."
Before discussing the merits of DSS's stated reasons for withholding the requested material, which is unclassified, some discussion of just what the Relevant/Irrelevant (R/I) technique is is in order. In this polygraph technique, the polygrapher asks the subject a series of relevant questions (e.g., "Did you use an illegal drug?) interspersed with irrelevant questions (e.g., "Are the lights on in this room?"). The questions are asked multiple times in different order over multiple polygraph charts, and may be differently phrased. As with the "Control" Question "Test," the irrelevant questions are not scored. Instead, the polygrapher examines the polygraph charts looking for "consistent, specific, and significant" reactions to a particular relevant question. (Polygraphers have a mnemonic expression for such reactions: "con-spec-nificant.") For example, if a subject consistently shows strong physiological questions when asked, "Did you use an illegal drug?" no matter what the order of the questions or how this question is phrased, deception will be inferred, and a post-test interrogation will follow. In addition, a "breakdown test" (a type of peak of tension test) may follow. For example, the subject who shows "con-spec-nificant" reactions to the question "Did you use an illegal drug?" may be asked questions like the following in a "breakdown test": "In connection with the question on illegal drugs, does anything disturb you about the following things?:
Marijuana?
Cocaine?
Heroin?
LSD?
Methamphetamine?"
Again, the question order is mixed and repeated. If the subject shows "con-spec-nificant" reactions to any one particular question in the list, then the polygrapher infers that this is an area of concern to the subject, and follows up with an interrogation along those lines.
Apart from examination of the polygraph charts, the polygrapher may also use his subjective impressions in making a determination of truth or deception.
The R/I "test" is one of the oldest polygraph techniques. Like the "Control" Question "Test," it is also thoroughly discredited, and there is absolutely no peer-reviewed research supporting its validity. Professor David T. Lykken devotes Chapter 7 of
A Tremor in the Blood: Uses and Abuses of the Lie Detector to the R/I technique. He notes two assumptions on which the R/I "test" depends, the second of which is, as he terms it, "wildly implausible":
ASSUMPTION 1. A guilty subject whose relevant answers are lies will be more aroused by the relevant than by the irrelevant questions and this difference will be revealed by his responses on the polygraph.
ASSUMPTION 2. An innocent subject who is answering truthfully will not be disturbed by the relevant questions and will show no more reaction to them than to the irrelevant questions. It is appropriate to cite here in full Professor Lykken's discussion of the validity of this technique:
Validity of the R/I Test
So much for theory and common sense; what is the evidence? It is astonishing to discover that, in 70 years of use prior to 1997, the only published studies assessing R/I test accuracy using "blind" evaluations of charts obtained from criminal suspects were one described by Larson[reference deleted] in 1938 and another by Horvath,[reference deleted] in 1968. Larson asked nine judges to read the charts obtained from 62 suspects. Only 1 of the 62 suspects had actually lied and yet the number scored as deceptive by the nine judges ranged from 5 to 30. This amount of disagreement among the nine judges indicates poor reliability. The average judge scored about one-third of the innocent suspects as deceptive, which means that two-thirds of these innocents failed to give large reactions to the relevant questions and were scored as truthful, just as Assumption 2 demands. One might have thought that Assumption 2 would nearly always be wrong and that most subjects would fail the R/I test whether innocent or guilty. That is in fact what Horvath reported; all of his innocent suspects were erroneously classified as deceptive by the R/I test, whereas Larson's earlier study reported only 33% false positives. We should not put too much faith in the exact percentage of errors found but we can say that, just as common sense would predict, a high proportion of innocent subjects will "fail" the R/I test. Quite recently the Raskin group of lie detector advocates published the results of a mock crime laboratory study[reference deleted] in which the R/I method classified all 15 of the "guilty" suspects as deceptive but at the expense of identifying only 3 of the 15 "innocent" subjects as truthful. Nonetheless, the U.S. Government appears to be actively relying on this most dubious of polygraph techniques for national security purposes. For example, in October 2001, DoDPI taught its two-week course on the Relevant/Irrelevant screening "test" in Chantilly, Virginia, home to the
National Reconnaissance Office.
Now, as mentioned in AntiPolygraph.org's Freedom of Information Act request, on 25 January 2001, Dr. Andrew Ryan, chief of the DoDPI Research Division, speaking at the first public meeting of the National Academy of Sciences/National Research Council Study to Review the Scientific Evidence on the Polygraph, stated:
"The curriculum at DoDPI, the lesson plans as they are being reviewed by the Middle States and others for the accreditation process, it's -- they're finding out that the curriculum there is based on research.... The curriculum changes that occur there, when we -- when and if we modify the curriculum, the training that the examiners receive, it must be backed up by research findings. So, unlike the early days in the 1950s when the school was located at Fort Jackson [sic] and I guess we would have to say the school, the training, was basically governed and controlled by the folklore of the polygraph industry, and now it's driven by research..."
AntiPolygraph.org is very interested in knowing on what "research" DoDPI's Relevant/Irrelevant polygraph screening "test" is based, and our request specifically included any such research.
Could it be that despite Dr. Ryan's testimony before the National Academy of Sciences, the DoDPI Relevant/Irrelevant "test" is based on nothing more than "the folklore of the polygraph industry?" DSS's
letter of denial suggests that such is precisely the case. While specifically claiming that "the entire R/I Screening Test can be withheld," the letter makes no mention of any research regarding this "test." Indeed, DSS failed to comply with the Freedom of Information Act requirement that "In denying a request for records, in whole or in part, an agency shall make a reasonable effort to estimate the volume of any requested matter the provision of which is denied, and shall provide any such estimate to the person making the request..." Although the Act contemplates exemptions to this requirement, it seems clear that no such exemption applies to the present request. AntiPolygraph.org has
e-mailed DSS requesting an inventory of the responsive materials that DSS identified and reviewed, along with a description of each item.
Let us turn now to the rationale DSS put forward for withholding all documentation of its R/I "test." It invoked exemption (b)(2)(high), which pertains to information which relates solely to the internal rules and practices of the Department of Defense.
This claim is demonstrably false. DoDPI's R/I Screening course is open to
any federally certified or contracted polygrapher, as well as to any polygrapher "employed by a state, county, or local law enforcement or corrections agency who has graduated from a PDD course at a school accredited by the American Polygraph Association." (See the prerequisites listed in DoDPI's R/I screening
course description.)
Clearly, documentation regarding DoDPI's Relevant/Irrelevant "test" does not relate "solely to the internal rules and practices of the Department of Defense."
DSS further claims that the release of the requested material "could reasonably be expected to risk circumvention of the law." This is a tacit admission by DSS that it believes that the Relevant/Irrelevant "test" depends on secret knowledge. But can DSS credibly claim that releasing any information at all about DoDPI's Relevant/Irrelevant "test" could reasonably be expected to lead to circumvention of the law?
The Relevant/Irrelevant technique was not invented by the U.S. Government. The technique was pioneered by Leonarde Keeler (1903-1949), who opened the first polygraph school, the Keeler Polygraph Institute, in Chicago, Illinois. Many of the U.S. Government's first polygraphers received their training there, including the late
Raymond J. Weir, Jr. of the National Security Administration (NSA), who attended the Keeler Polygraph Institute in 1951. There he learned the Relevant/Irrelevant technique, which NSA used for polygraph screening for decades (and may still use). Weir went on to become chief of the NSA polygraph program and president of the American Polygraph Association (APA).
Raymond J. Weir, Jr. apparently did not share DSS's view that to release any information about the Relevant/Irrelevant screening technique "could reasonably be expected to lead to circumvention of the law" because he described it in minute detail in two articles published in the APA quarterly publication,
Polygraph: "In Defense of the Relevant-Irrelevant Polygraph Test."
Polygraph, Vol. 3 (1974), No. 2, pp. 119-166.
"Some Principles of Question Selection and Sequencing for Relevant-Irrelevant Testing."
Polygraph, Vol. 5 (1976), No. 3, pp. 207-222.
In these articles, it is clear that Mr. Weir describes the Relevant/Irrelevant technique as practiced by the NSA. Would the director of the NSA polygraph program have published these articles if such publication "could reasonably be expected to lead to circumvention of the law?"
More recently, DoDPI released to Dr. James Allan Matte, a private citizen who happens to be a polygrapher, a 1995 manual on a different Relevant/Irrelevant technique then used for personnel screening, the
General Question "Test." Dr. Matte cites the 42-page document at pages 92 and 464 of his 1996 book,
Forensic Psychophysiology Using the Polygraph - Scientific Truth Verification - Lie Detection. DoDPI seemed unconcerned that releasing this information to a polygraph insider "could reasonably be expected to lead to circumvention of the law." (When AntiPolygraph.org
requested the same information, DSS
replied that it couldn't find it, claiming that it "most likely has been discarded.")
It seems that it is only when critics of polygraphy request such information that claims of privilege are invoked to withhold it. DoDPI's motto is "Justice and Security Through Truth," but it appears that DoDPI and the DSS don't want you and me to know the truth about their Relevant/Irrelevant screening "test."
AntiPolygraph.org's request for this information is especially timely, because the National Academy of Sciences (NAS) is currently conducting a
Rev iew of the Scientific Evidence on the Polygraph, with a special emphasis on polygraph screening. Is documention of the Relevant/Irrelevant screening "test" to be withheld from the NAS, too? If so, how can the NAS proceed with its work?
On the other hand, if this information
is to be provided to the NAS with the stipulation that it not be publicly disclosed, will the U.S. Government then seek to withhold the final NAS report from the public based on alleged national security concerns?
Watch AntiPolygraph.org for further developments regarding our request for documentation of DoDPI's Relevant/Irrelevant screening "test." Comments welcome here.
Pages: 1
DSS Withholds R/I Screening Documentation (Read 10863 times)
