Lie Detectors Are Junk Science, but We Keep Using Them

Amit Katwala’s Tremors in the Blood explores how unreliable technologies have been used in our criminal justice system.


You've probably seen a lie detector in a movie or TV show, its stylus scratching an ink line across a scrolling page and jumping when the subject lies. Sometimes the polygraph is presented as infallible; sometimes its scrutiny can be evaded. In a spy thriller, the hero might put a pin in his shoe: Stepping down on a sharp point, the theory goes, will cause sufficient stress to spike his blood pressure, disguising false statements. In other tales, talented operatives can simply meditate their way down to a state of calmness and therefore appear not to be lying.

Even in the early days of the lie detector, the device's advocates seemed dimly aware that simply being questioned by the police might make a suspect flustered or nervous and, thus, perhaps give the impression of being untruthful. Whose blood pressure wouldn't spike when faced with criminal charges? Sure enough, police quickly found that people who were given polygraphs tended to panic. They also tended to confess: to all sorts of  offenses, from card games to illegal alcohol, separate from the crimes under investigation.

Amit Katwala, a reporter at Wired, tackles the lie detector's early history in Tremors in the Blood. He focuses on its origins in Berkeley, California, in the 1920s and on some cases that both brought it to prominence and raised questions about its validity..

The machine attracted controversy right from the start. Katwala covers the case of Henry Wilkens, who was (probably) guilty of killing his wife but managed to get away with it. This was a front-page news story in San Francisco in 1923, and it was an opportunity for the lie detector to prove its value. Unfortunately for the prosecution, when he was subjected to a test in front of a crowd of onlookers, Wilkens passed. In the eyes of the polygraph boosters, the device had failed—or was failed, in being applied incorrectly. Wilkens was acquitted and the San Francisco Police Department swore off lie detectors. 

 The Wilkens case shows more about why police forces (and the public) wanted the lie detector, or something like it. A young woman had been shot to death. It was the sort of violent crime that was on the rise in the early 1920s, panicking suburbanites and leading authorities to grasp at any technique to catch perpetrators. People wanted answers and security.

In the preceding decades, new policing techniques had become standard (from mug shots to fingerprints) and the "science" of law enforcement had been growing apace. The lie detector looked like yet another leap forward. The era had already brought the telephone and the radio; what new magic would be next? It was the same mindset that led people to embrace the eugenic ideas of the Italian criminologist Cesare Lombroso—that there were criminal "types," detectable by their appearance. If some people were just undesirables, the thinking went, wouldn't it be all the better to flush them out with scientific proof?

The only "lie detector" available up to that point was a sharply wielded nightstick, so the mirage of an accurate, efficient, and peaceful truthfinder held an understandable appeal. The same impulse led investigators to try and then abandon various versions of "truth serum" over the decades. 

But the lie detector is just the 20th-century version of witch pricking, revealing a "truth" that isn't there. The National Academy of Sciences has dismissed the polygraph's validity, and the American Psychological Association says there is "little evidence" that it works.

The technology also quickly faced legal challenges. In Frye v. United States (1923), the Court of Appeals of the District of Columbia held that any scientific methods deployed in the courtroom had to be widely accepted by experts, which polygraphs were not. But this inadmissibility didn't stop it from catching on, or from gaining acceptance in the public mind. Millions of tests are given in the United States every year, and it is used on anyone from suspected criminals by the FBI to suspected baby daddies by Maury Povich.

Those who use the test clearly believe it has some probative value—and at one remove, the threat of a polygraph might at least lead suspects to be forthcoming. In his 1991 book Homicide: A Year on the Killing Streets, David Simon detailed how Baltimore police fooled suspects by claiming a Xerox machine was a lie detector: They preloaded it with paper saying TRUTH or LIE and just hit the print button when their target answered a question. The rattled suspect thought the jig was up.

False confessions were an issue even in the earliest days of polygraphs, and nothing its developers tried could remove the risk. Disillusionment both with the technology's limits and with law enforcement's ability to solve crime is a thread through the lives of three major figures in Katwala's book: Berkeley police chief August Vollmer, Berkeley physiologist turned police officer John Larson (who invented the device), and Larson's teenaged assistant Leonarde Keeler (who later developed it further). Keeler became the machine's keenest advocate, eventually hoping to patent and market it to law enforcement and civilian organizations. Nonetheless, even he came to see its flaws and potential for abuse.

That leaves the lie detector in a kind of gray area. Its advocates believe that it has now been sufficiently tweaked to be effective. Critics regard it as junk science. One of those critics is Katwala, who states firmly that the polygraph "does not work." In the final chapter, he explores modern lie-detecting variants, based on eye movement tracking or fMRI scans. None of these can be shown to really work either, but the market for them continues. 

The underlying problems here extend well beyond the polygraph, affecting even legitimate technologies. Attorneys today talk of the "CSI effect," with fictional high-tech detectives shaping how the public expects the police to work with advanced, completely accurate technology. Juries like blue light photos and DNA evidence, the more "sciency" the better. As forms of supposedly scientific proof—such as handwriting analysis—have come under suspicion for being more speculative than we were told, the desire persists for a technique that reveals the truth in the human heart. The polygraph and its modern variants won't be going away any time soon.

Tremors in the Blood: Murder, Obsession and the Birth of the Lie Detector, by Amit Katwala, Crooked Lane Books, 352 pages, $18.99