NSA Polygraph Screening Format The 1983 report of the Congressional Office of Technology Assessment (OTA),
Scientific Validity of Polygraph Testing: A Research Review and Evalation, indicates at p. 24 that the National Security Agency (NSA) uses a relevant/irrelevant polygraph screening format: "...NSA reports that it uses primarily the R/I technique."
For those who have read Professor David T. Lykken's
A Tremor in the Blood: Uses and Abuses of the Lie Detector or the 1st edition of
The Lie Behind the Lie Detector (the next edition will be updated), it may come as some surprise that the NSA would be using a technique that is largely discredited, even within the polygraph community. But the OTA report explains the mystery at p. 17:
Quote:...the R/I technique as used by the Federal Government involves somewhat different types of questions than the traditional R/I, and it must be explained separately. As currently used by Federal examiners, the R/I relies on a type of control question, and is claimed to be a version of the control question technique....
...
...The Federal Government version of the technique is called the general question test (GQT)....
In a separate post, you'll find a description of the General Question Test (GQT) and its "disguised control questions" derived from a DoDPI document dated 1995:
http://www.antipolygraph.org/cgi-bin/forums/YaBB.pl/YaBB.pl?board=Proc&action=di... However, the 1983 OTA report's explanation of the GQT, which is found at pp. 17-18, is inconsistent with DoDPI's 1995 description:
Quote:The Federal Government version of the [relevant/irrelevant] technique is called the general question test (GQT). Like the Reid CQT (discussed below), it uses inclusive control questions, which pertain to the subject's entire life, such that a complete answer would also include the specific incident being investigated. Thus, with a question like, "Did you ever steal anything from a place where you worked?" the theft being investigated would in actuality be part of the answer, Technically these are seen as "relevant" questions, because they are pertinent to the incident in question. Yet they are claimed to function as control questions, because they are intended to provoke a greater response in innocent subjects than questions about the misdeed provoke.
Some plausible explanations for this discrepancy include:
- The GQT format has been modified since the writing of the OTA report;
- The OTA was not informed about the true nature of the "disguised control" questions employed in the GQT;
- The OTA simply got it wrong.
The 1983 OTA report also notes at p. 41 that "NSA conducts the largest number of polygraph examinations of any Federal agency" and, at p. 107, that NSA conducted 9,672 such "tests" in fiscal year 1982. (The number of polygraph "tests" annually conducted by NSA now seems to be classified. The
DoD Polygraph Program Report to Congress for FY 2000 hides polygraph figures for NSA in a classfied table.)
It is sadly ironic that the U.S. intelligence agency best known for its technological sophistication also seems to be the most heavily reliant on pseudoscientific polygraph screening, a procedure, which, as polygraph expert Dr. Drew C. Richardson of the FBI laboratory
testified before the U.S. Senate, "is completely without any theoretical foundation and has absolutely no validity."
Last modification: George Maschke - 02/11/01 at 06:46:29
