Los Alamos Monitor
March 4, 2001
Speaker Rates Polygraphs as Unreliable
by Roger Snodgrass
Monitor Assistant EditorConventional "lie-detecting is an outdated and dubious practice that is inherently biased against innocent people. That is the conclusion of William Iacono, a psychology professor from the University of Minnesota-Twin Cities, who spoke Thursday at a Physics and Theoretical Division Colloquium at Los Alamos National Laboratory.
Polygraph testing has been an increasingly prescribed antidote to recent security lapses and espionage fears at the laboratory and in the national defense complex in general.
Iacono, according to the lab's announcement, "is widely regarded as one of the world's foremost clinical psychophysiologists," and "as a national expert on polygraphic interrogation and lie detection."
His talk, he said, would answer the question, "If you've got nothing to hide, you've got nothing to worry about, right?" -- a question from the cover of a vintage magazine from the 1940s with an artist's rendering of a polygraph examination of that era. The fading respectability of the procedure was reinforced by Iacono with the information that the "lie-detector" was invented in 1924 by William Moulton Marston, and that Marston, who wrote under the pen name of "Charles Moulton," was also the creator of the Wonder Woman comic strip. Wonder Woman, recalled Iacono, had a magic lasso that caused those encircled by its coils to tell the truth.
Since 1938, said Iacono, the basic principles of polygraphing have not changed. They measure the same fundamental channels of physiological activity -- blood pressure, respiration, and skin response -- as they did more than 60 years ago. The tests are administered by practitioners who do not need a college degree and have little physiological or psychiatric training.
They are taught during the seven to 12 weeks of their training that the techniques they are learning are infallible.
Iacono described the basic polygraph test as assuming that one would breath more shallowly when one is lying, that one's blood pressure would be elevated, and that one might perspire more from nervousness when telling a lie.
There are two types of scoring for the test, he said. One is "global" -- based on all the information available. Factors to be considered include the way one handles oneself during the test, whether eye contact is made, nervousness, how forthcoming one is during questioning, one's past record, and so forth.
Another scoring system assigns numbers to activity levels on a minus three to a plus three scale. The bottom line is that the subject is judged deceptive if there is a negative total; truthful, if positive.
The problem, said Iacono, is that however complex the system becomes (and there are now sophisticated, proprietary software routines, whose secrets are not even revealed to the government), the polygraph "does not measure lying." Instead, it measures indirect indicators that must be interpreted in order to draw a conclusion.
There are two basic types of questioning routines, explained Iacono. The first and most researched is called the controlled question test.
"These procedures are not standardized," he said. "They vary from exam to exam depending on what's in style. The relationship between the examiner and subject is critical. The nature and wording of the question varies. The grouping and ordering of the questions is thought to be important."
By comparing responses to relevant and irrelevant questions, the person running the test is supposed to be able to determine areas of questionable veracity. By spotting some weakness or anomaly in the physiological response, the polygraph practitioner is trained to bore down on areas presumed to be vulnerable in the subject's story.
"The major purpose is to gain an admission," said Iacono.
The problem with the technique, he said, is that innocent people will have a more emotional response to control questions such as "Have you ever told a lie?" in part because they will be more conflicted about the risk of self-incrimination. The guilty, especially a well-trained spy, will have learned techniques for fooling the test.
Polygraph-spoiling techniques, readily available on the Internet, include ways to enhance one's physiological response, particularly to the control question. Tricks include doing mental arithmetic, curling one's toes, or lightly biting one's fingers. Having established a strong response to the control question, the subject minimizes any reaction to a relevant question, such as "Did you take the hard drive?"
In a brief survey of research results on polygraph testing, Iacono said that research was skimpy and flawed. And even at that, a Department of Defense survey of a particular type of test found that 13 percent of "innocent" people were found "guilty" and 17 percent of "guilty" people were found innocent.
Elsewhere on Thursday, Attorney General John Ashcroft gave a hesitant endorsement to continued polygraph testing during an FBI press conference in Washington when asked about Robert Hanssen and the current Washington spy scandal.
According to a transcript of the conference, made available by the Federation of American Scientists, Ashcroft was asked if he thought that "polygraphing would have done much to have caught Hanssen any earlier, and are you ready to have the FBI do required testing of its agents...?"
Ashcroft responded: "Well, let me just say that -- (pause) -- I have to try and sort out what I'm -- there are certain things regarding intelligence questions that I -- It's my understanding that there have been cases in the past that polygraphing did not work on. I think you could name them. So the polygraph is not a sure way. The polygraph is said to have about 15 percent false positives and has an impact on the way an agency operates.
"Nevertheless, I believe that there are applications for polygraph that are important, and the director and I have agreed that because of the national security involved and the risks involved and the very important consequences of breaches, that we should elevate the use of polygraph in certain cases as it relates to the Bureau."
Polygraphing is also the subject of an ongoing study by the National Research Council, which held its first meeting on Jan. 25. A panel sponsored by the Department of Energy has been directed "to conduct a scientific review of research on polygraph examinations that pertain to their validity and reliability, in particular for personnel security screening." The panel's final report is scheduled to be issued by April 2002.