The Skeptical Psychologist

Investigating questionable, controversial, and novel claims in psychology.

The Polygraph Test Strikes – and Strikes Out – Again

The polygraph is an arousal detector, not a lie detector.

About two weeks ago, I received word that a close friend of mine from elementary school (I'll be altering a few details to protect this friend's identity) - let's call her Ellen - failed a required polygraph test she had taken as part of her application for a high-level government position. I can say without qualification that Ellen, a former social worker, is among the two or three most honest people I've ever known. She is a person of exceptional integrity, and is truthful almost to a fault. In fact, she is in possession of what Freudians would call an "overdeveloped superego" - meaning, in common parlance, that she is highly guilt-prone.

My suspicion is that Ellen is a prime example of what some polygraph examiners informally term a "guilt grabber" - an innocent person who flunks the polygraph test not because he or she is actually guilty of having done anything wrong, but because he or she feels guilty at the mere thought of having done something wrong. I've often fantasized about stealing a cheesecake (not merely a slice, but the whole shebang) from one of my favorite New York City delicatessens, but I don't feel guilty because I've never actually done so and never would. In contrast, a guilt grabber might feel pangs of conscience at the mere thought of having imagined walking off with a cheesecake. Indeed, in Ellen's case, the polygraph examiners left fairly convinced that she had committed a major crime in her youth - which in fact Ellen had not.

The polygraph or so-called "lie detector" test is, as most scientists acknowledge, badly misnamed. It's a detector of autonomic arousal, not of lies. As a consequence, people who become highly aroused in response to the relevant (or "Did you do it?") questions, but not in response to the other questions, will tend to fail the test. In some cases, this arousal almost surely reflects actual guilt stemming from commission of a crime - and for this reason there's general consensus that the polygraph test probably does somewhat better than chance at detecting lies. But in many cases, this arousal just as surely reflects emotions other than guilt stemming from malfeasance, such as understandable anxiety at the prospect of failing the test, indignation at being accused of a crime one didn't commit, and - perhaps in the case of my friend Ellen - guilt at the mere thought of having once fantasized about having committed a crime. My late Ph.D. mentor, David Lykken, who was the world's foremost critic of the polygraph test, put the overall accuracy of the test at about 70%, but not surprisingly, fervent proponents of the test place it higher.

In any case, there is no disagreement among reasonable scholars that, Meet the Parents and other Hollywood portrayals aside, the polygraph is not an infallible detector of lies. Yet survey data suggest that many people believe otherwise; the results of one study revealed that 45% of undergraduates enrolled in introductory psychology courses believed that the polygraph "can accurately identify attempts to deceive."

Many scientists contend that a true lie detector is not possible, at least given the present state of psychophysiological technology. That's because no-one has yet found a "Pinocchio response" - a physiological or behavioral reaction that's specific to lying. Like Pinocchio's protruding nose, this response would be emitted only during lying and never at other times. Some scientists doubt that such a response will ever be found; I remain skeptical myself given the enormous diversity of motivations underlying lies. There are white lies, lies intended to protect oneself, lies intended to protect others, malicious lies, lies told for the mere fun of fooling others (that is, lies told for what psychologist Paul Ekman calls "duping delight"), and many others; the prospect of finding any deep-seated biological or psychological commonalities underpinning of all of them seems doubtful. But perhaps my skepticism will one day prove unwarranted.

Ironically, the developer of the polygraph test, William Moulton Marston, was also the developer of the comic book character Wonder Woman, who lured criminals into telling the truth by encircling their waists with a magical lasso.  For Marston, the polygraph was the real-world equivalent of Wonder Woman's lasso: An infallible truth detector.

image

Given that it's really an arousal detector, the polygraph test suffers from a high rate of what psychologists call "false positives" - innocent people whom the test deems guilty. Ironically, some of these false positives, like Ellen, may be among the guilt-prone and honest among us, although this disturbing possibility has received surprisingly little attention from researchers.

There's at least one other great irony in all of this. In 1988, the federal government passed the Employee Polygraph Protection Act, which banned most uses of the polygraph in the private sector - for example, for pre-employment screening or for trying to ferret out pilferers in the workplace. Yet the government refused to ban the test for screening its own potential employees. In essence, the government told private employees, "We won't let you use this test, as it doesn't work very well, but we'll keep on using it ourselves." More than two decades later, my friend Ellen and hundreds of others are continuing to feel the unfortunate effects of this paradoxical policy.

Scott Lilienfeld is a psychology professor at Emory University in Atlanta and co-author of several books on pseudoscience in psychology.

more...

Subscribe to The Skeptical Psychologist

Current Issue

Just Say It

When and how should we open up to loved ones?