DoD Relying on Polygraphy in Iraq WMD Search

[George W. Maschke]

A Fox News report titled, "Sources: 'Chemical Ali' Alive and Well" indicates that the Department of Defense is relying on the pseudoscience of polygraphy in its thus-far fruitless search for "weapons of mass destruction" in Iraq.

Senior defense officials told Fox News that it is obvious that top Iraqi leaders in custody are lying to interrogators about Iraq's weapons of mass destruction programs.

The infamous Huda Mahdi Ammash, dubbed "Mrs. Anthrax" for her work on Iraq's bio-weapons program, has failed multiple polygraph tests, senior defense officials said.

[George W. Maschke]

Now it seems that all Iraqi scientists questioned by the Defense Intelligence Agency about biological, chemical, and nuclear weapons have passed DIA polygraph "tests." Associated Press writer Dafna Linzer reports in an article titled, "U.S. Working on Plan for Iraqi Scientists":

Two members of the Pentagon's Defense Intelligence Agency involved in questioning scientists in custody told AP the Iraqis continue to deny the existence of illicit weapons programs in Iraq. Dozens of Iraqi scientists have been questioned and less than 30 remain in custody. All of them, including senior members of Saddam's regime, have been subjected to lie-detector tests, which have come up clean on weapons questioning, the DIA officers said.

I don't mean to suggest in any way that any of the scientists questioned were lying, but given what is known about the polygraph's inaccuracy, the only way that any population of "dozens" are likely to all "come up clean" on any particular question (or general area of questioning) is if some criteria other than the polygraph charts are used for the purposes of making a determination.

Note also that the information in this Associated Press report seems to be in conflict with that provided in the earlier Fox News report.

[Marty]

Now it seems that all Iraqi scientists questioned by the Defense Intelligence Agency about biological, chemical, and nuclear weapons have passed DIA polygraph "tests." Associated Press writer Dafna Linzer reports in an article titled, "U.S. Working on Plan for Iraqi Scientists":


I don't mean to suggest in any way that any of the scientists questioned were lying, but given what is known about the polygraph's inaccuracy, the only way that any population of "dozens" are likely to all "come up clean" on any particular question (or general area of questioning) is if some criteria other than the polygraph charts are used for the purposes of making a determination.

This is true in a general "screening" poly but this is closer to a specific incident one. Doesn't the DOE's DL-CQT poly (unlike LE screening ones) produce a very low false (or true for that matter)  positive rate? There may be similar dynamics at work.

-Marty

[George W. Maschke]

Marty,

Interrogating Iraqi scientists about their possible knowledge or involvement in the production of weapons of mass destruction not known to exist is very much a screening situation. But even with regard to a specific incident situation, one would not expect "dozens" of subjects all to pass, even if all were truthful. (Consider, for example, the Molly Bish and Douglas Gene Kirk cases.)

As for the very low fail rates (much lower than the expected false positive rate) associated with the DOE and DoD polygraph programs, these can only be achieved by making ultimate pass/fail determinations based on extra-polygraphic criteria.

[Marty]

Marty,

Interrogating Iraqi scientists about their possible knowledge or involvement in the production of weapons of mass destruction not known to exist is very much a screening situation. But even with regard to a specific incident situation, one would not expect "dozens" of subjects all to pass, even if all were truthful. (Consider, for example, the Molly Bish and Douglas Gene Kirk cases.)

As for the very low fail rates (much lower than the expected false positive rate) associated with the DOE and DoD polygraph programs, these can only be achieved by making ultimate pass/fail determinations based on extra-polygraphic criteria.

George,

One can quibble over whether it is a screening or specific incident but it is closer to the narrow DOE screens than FBI broad screens.

It stands to reason that narrow screens would produce fewer false positives since innocent people are not equally reactive to all accusatory subjects.

That said, the very low incidence of positives on DOE screens may well be due to extra-polygraph scoring techniques. Perhaps they put a "fudge factor" in the numerical calculations in favor of the examinee. Do you have any more detail on why this anomoly is?

-Marty

[George W. Maschke]

George,

One can quibble over whether it is a screening or specific incident but it is closer to the narrow DOE screens than FBI broad screens.

It stands to reason that narrow screens would produce fewer false positives since innocent people are not equally reactive to all accusatory subjects.

Marty,

First, I don't mean to split hairs, but the kinds of polygraph interrogations described in the AP article are quite clearly screening examanations (about behaviors not known to have occurred).

I am not confident that your second point stands to reason. In either case, the procedure is still without validity. Moreover, significant reactions to relevant questions might be anticipated due to the adversarial, involuntary nature of the interrogations and the fact that the subjects' liberty was at stake. (Indeed, some are still being held prisoner.)

In any event, my point remains that an invalid procedure such as CQT polygraphy would not be expected to result in all persons passing (based on chart readings) when "dozens" of truthful persons are interrogated. There can be little doubt that some other criteria are being relied upon to declare a subject as having "passed" or "failed."

That said, the very low incidence of positives on DOE screens may well be due to extra-polygraph scoring techniques. Perhaps they put a "fudge factor" in the numerical calculations in favor of the examinee. Do you have any more detail on why this anomoly is?

-Marty

In the counterintelligence polygraph screening programs, it seems that if a subject reacts strongly to a relevant question, he/she gets interrogated, or, as polygraph polygraphers sometimes say, given an "opportunity to explain" their reactions. If the polygrapher(s) involved (and/or their bosses) are ultimately satisfied with the explanation, the subject is deemed to to be "cleared" and "passes." This is discussed in Chapter 2 of The Lie Behind the Lie Detector (pp. 49-50 of the 3rd ed.) in discussing how 20% of 800 DOE employees polygraphed were initially accused of deception, although all ultimately "passed."

The point I implicitly hoped to make with these posts, and perhaps should now make explicit, is that the DoD's reliance on the polygraph in its hunt for putative weapons of mass destruction seems to be an example of precisely the kind of "belief in its accuracy not justified by the evidence" against which the National Academy of Sciences has warned (The Polygraph and Lie Detection, p. 219).