Sunday, December 13, 2009

Confessions of a Detected Liar

I wrote this ten or fifteen years ago as an article that I hoped to sell. Either I sent it to Newsweek, then forgot it, or I didn't get around to researching markets. Peter Watts' border adventures reminded me of it:

Confessions of a Detected Liar

by Will Shetterly

Don't trust me; I can't pass a lie detector test.

Like most people, I believed polygraphs were usually accurate. They call them lie detectors, right? The polygraph industry thrives on the promise behind that popular tag for their machines: they detect lies and liars.

I lost my faith in polygraphs in 1981. My fiancee and I had visited my family in Canada and were returning to the US. A customs agent searched the car, as almost always happened when I, then a young guy who looked a bit hip, crossed the border. The agent looked under the passenger seat, frowned, and showed me a small block of hashish in Saran wrap.

He asked if I knew what it was. I suspected I did--I grew up in the '60s and '70s, after all. But I couldn't be sure, so I reached for it to take a better look. When he jerked it away before I touched it, I realized I was in trouble.

I had no idea how the hashish had gotten there. I was driving my parents' car; I knew the hashish couldn't belong to them. I had lost interest in recreational drugs years earlier. I didn't think it could belong to my fiancee, yet there it was.

I soon found out how much trouble I was in. Taking drugs across the border is a federal offense. This was before the RICO laws went into effect, so the police didn't seize my parents' car, but my fiancee and I were both strip-searched. As the driver, I was presumed responsible, even though the hashish was under the passenger seat.

My parents hired a lawyer, who told us that the charges would be dropped if I passed a polygraph. So I went down to the police station a few weeks later, took a polygraph test, and left confident that this was all over. Shortly after, I learned I had failed the test.

I was baffled. I briefly wondered if the hashish was mine and my mind was playing tricks on me. There was the evidence, after all: the lie detector had told me, and the world, that I had lied. Fortunately, my parents still trusted me, though modern science told them not to. They continued to spend money on the lawyer.

Wanting to understand what had gone wrong with the polygraph, I decided to hire a private operator to give me a second test. Somewhere in my files, I still have the letter from him. He said that my test showed evidence of an intent to deceive.

At this point, I began to research polygraphs. I learned that its critics think the test is wrong as often as half of the time—it's no better than guesswork. Its defenders, including the American Polygraph Association, think polygraphs have an 85-95% accuracy rate.

Before I was tested, if you'd asked me how accurate polygraphs are, I probably would've offered a percentage close to that. Ninety-five per cent sounds pretty good, if you don't think about it. But who among us would risk our reputation for honesty on a test that's wrong at least one time out of twenty?

The answer is that many of us are forced to. Look in the yellow pages under "lie detection services" (there's no category for the more accurate "polygraph services") and you'll find a thriving business catering to civil, criminal, business, and domestic investigations. Want to know if someone is embezzling, taking drugs, or sleeping around? Like psychics, phrenologists, and others who claim that their profitable business is a science, polygraph artists promise an easy answer. And, like most easy answers, it's often a wrong one. People unjustly lose jobs, lovers, and reputations every day, thanks to polygraph tests.

For me, there was a happy ending. A friend of the family had lost the hashish several months before my arrest when she was riding in my parents' car. She did not come forward immediately because she thought she had lost it while playing volleyball, not while riding to the game, and she thought I had been arrested for having marijuana, not hashish. When she realized the truth, she told us, we consulted with the lawyers, and the following was worked out:

Our friend gave a full report of the incident to the Ontario Provincial Police. Because the hashish was not found in Ontario, the O.P.P. did not charge her. The O.P.P. sent a copy of their report to the prosecutor in the U.S. That report was accepted, and the charges against me were dropped. My fiancee and I got married, and I suppose it was a happy ending for everyone, if you don't count the four thousand dollars or so spent on my defense. That, I suppose, was part of the happy ending for the lawyer.

But I was one of the lucky ones. Without our friend's statement, I would have been convicted. The hashish was in my car, after all. My only defense was my word. And the judge, prosecutor, and defender all knew the polygraph test showed I lie.

Will Shetterly is a novelist. But since he can't pass a lie detector test, you might want to go to the bookstore and verify that for yourself.

ETA: Possibly of interest:


  1. The possibility exists that you were tested by an incompetent polygraph examiner.

    Louis Rovner, Ph.D.

  2. It's possible. But there would've been two of them.

  3. I just tweaked the post to make it clearer that I failed two tests by two operators.

  4. The polygraph doesn't make decisions about truthfulness or deception. The polygraph examiner makes those calls, based on his interpretation of his subject's physiological responses to certain questions.

    If both examiners worked for the same agency, they were probably selected for their jobs using the same set of criteria. If those criteria were not sufficient to weed out incompetents, it's not unreasonable to assume that the same mistake was made by two different people.

  5. Not the same agency. They were both in Minnesota, so it's possible that they got their training at the same place, but I chose a second tester at random because I was so surprised by the results of the first test.

    I do think things may've been complicated by the fact that I knew a little about polygraphs, and I had some karate and yoga training, so I was thinking, "Breathe calmly and tell the truth, and you'll be fine." This might've caused something unusual to show up on their tests. On the other hand, if the test only works on a subset of people who know nothing about it, it's flawed. It has to work for everyone. Otherwise, when it's wrong, the consequences are far worse than if there'd been no test at all.

  6. From the 2002 panel from the National Academy of Sciences after a review of the research on polygraph reliability: "Almost a century of research provides little basis for the expectation that a polygraph test could have extremely high acccuracy." Also, speaking of the future of the technology: "The inherent ambiguity of the physiological measures used in the polygraph technique and interpretation will bring only modest improvements in accuracy."

    Even the Supreme Court holds "There is simply no consensus that polygraph evidence is reliable."

    It seems a little reading does give interesting answers.

  7. Stevengould is a living example of the old saying "A little knowledge is a dangerous thing." If he read the NAS report, he would have seen that the panel cited research to show that accuracy for a single-issue exam in extremely high. The quote from Stevengould was in regard to pre-employment and periodic screening exams.

  8. One more thing - there was not one person on the NAS panel who actually knew anything about polygraph testing. No one on that panel had any training in the technique, and not one of them had ever conducted a polygraph exam.

    Not quite the definition of a panel of experts.

  9. It's an interesting idea that one can only be fit to evaluate the mass of scientific data on polygraphs if one is trained to read polygraph results. Do we next only allow self-identified psychics to study ESP? Do we throw out all the biological data we've accumulated because none of the researchers were fruit flies?

  10. Wouldn't trust one as as as I could throw the piece of junk. You see, unlike Will, I can pass one. Have, in fact.

    Had an employer insist on a pre-employment test (asking all kinds of outrageous, and probably illegal--even then--questions). For a barely above minimum wage job, too. I lied. I gave the answers that were patently the "right" answers (political, sexual, economic, etc. Plus the one about never having used drugs. I had. Hardly experimented, but had).


    A family friend in the company's HR dept asked me that evening, having seen the report, about it, because she knew that it was wrong. Well established examiner, as I recall.

    No doubt also incompetent.

  11. Rovner, I did a fair bit of research at the time. Obviously, it's 25 years later now. But at the time, there were two schools of research, one by people who believed in polygraphs, and one by people who didn't. Since they failed on me, I lean toward the folks who are skeptical.

    If the error rate only hurts one out of a hundred people who are tested, that's far too many.

    And since you were a bit abusive to Steven Gould, a science writer who understands research, I'll note your "No one on that panel had any training in the technique, and not one of them had ever conducted a polygraph exam." This is kind of like a Scientologist complaining that their critics have no knowledge of Xenu and have never used an E-meter.

  12. The evidence isn't all in on polygraphs. If you're a really good liar (and/or a stone cold psychopath) you can pass the test: the Green River Killer passed his polygraph with flying colors, while an innocent suspect failed the same test.

    So, whatever else you are, Will, you can take comfort that you're not a psychopath (which I'm sure is a great relief to Emma. ;-)

  13. Wow, fascinating indeed dude. I like it.


  14. @Rovner One more thing - there was not one person on the NAS panel who actually knew anything about polygraph testing.

    This seems sort of like the argument that parapsychologists use when their positions are challenged. If the truth is objective and spelled out in the literature, it shouldn't take smart people that much trouble to understand. Practitioners are not always aware of their biases, of course.

    I see you did your Ph.D. on physiological correlates of deception back in the 70's, so I'll give you the overwhelming benefit of the doubt when it comes to what was known back then. Certainly there are nice physiological correlates of all sorts of human behavior, deception and otherwise. P300, one of my most favorite misused signals. But I'm curious how much academic research you've done on the subject lately? I wasn't able to find anything in PubMed with (Rovner[AU] and polygraph).

    It seems that, as a practitioner, you have a vested interest in a particular interpretation of the findings and the use of polygraph techniques.

  15. It's funny that I came across this article today, after I had just read a chapter in The Traveler about the failings of polygraphs and how the best thing to do was take an image of the person's brain as they spoke using a room built within an MRI machine. The technology was futuristic, of course (no noise, gigantic, etc.) but I believe that that's the direction things will go in. Polygraphs detect emotional signals that could be true for a wide variety of emotions. The brain is more precise.


  16. What is sounds like is that you got caught with hash crossing the border, though you could lie your way through a test, failed twice and then figured out a way to have a friend take the fall in a way that would get you off the hook and not get them in much trouble.

    And you can give me a lie detector test to see if I am telling the truth.

  17. Doug, it is possible to make anything sound like a lie. But the family friend is still alive. Though I can't pass a polygraph, odds are that she could. What's most important is the Ontario Provincial Police trust her.

    For me, here's the best argument for why the hash was not mine: I'd crossed the border many times before. I almost always got searched because I was a young guy who clearly wasn't military. The only reason I didn't search the car myself before crossing that time was it was my parents' car. The idea that there would be illegal drugs in their car never occurred to me.

    The other thing I knew at the time: If I wanted to smuggle drugs across the border, I would fill up my mom's trunk, because they never searched the cars of nice middle-class women.

    One last point, in case you're really suggesting I'm lying now about something that happened 25 years ago: Why would I? If I had done what you're suggesting, I would brag about fooling the cops and the polygraph operator.

    PS. Since that incident, I clean my car before crossing the border, no matter how sure I am that nothing's been in it. The wind could blow the butt of a joint into the car, or something could stick on the bottom of someone's shoe. Border guards don't speculate: if they find something in your car, it's yours. So, whether you believe my story or not, always clean your car before the border.

  18. Oh, I realize I can't pass a polygraph to prove this, but here's a little more about my drug use: I used a lot of illegal drugs in my teens, then had a few bad experiences with them, and effectively stopped using all illegal drugs a few months shy of my seventeenth birthday. I support the legalization and regulation of most drugs today, but that's not personal. It's because I think it's a horrible waste of our tax dollars.

  19. Rovner you're overtly being an obnoxious prick. The fact you're claiming that the cases of polygraph failure are attributed to poorly trained polygraph readers is one more reason to call the entire polygraph business a sham and prohibit such results from entering a court of law, the first being that the polygraph does not measure by any means truthfulness, but changes in physiology that -- aside from wishful thinkers such as yourself -- specialists can find multiple causes for.

  20. Rovner is clearly protecting the house of cards he helped build. If the success rate of these tests are any less than what you expect from a condom, then the results are worthless. I find it easy to believe Will's story while Rovner just comes off as a sanctimonious prick. An incompetent one at that. Rovner should know better than everyone else (provided he ain't lying about his credentials) that there is always the possibility of human error. And he also knows that most people will tend to believe it is the machine that does all the work, not the interpretor. I also wouldn't bet his ego has never played a role. Nor has his prejudices, biases, dispositions, financial rewards, etc. Louis Rovner is not unlike highly skilled purloining scam artists.

  21. I have passed numerous Polygraphs regardless of actual deception versus perceived deception. I even failed at the "Is your name ______" question before.

    I am 100% familiar with Polygraphs, have taken a lot of them (over 15 in the last 7 years).

    Polygraphs are a way for whomever is conducting them to get people to tell the truth. It does not "DETECT" the truth, however; if deception is found, throwing that in the face of the subject is quite an intimidating way to get them to speak.

    Typically, when deception is found after the analysis is complete, the first question is "So, why are you lying about ____ ?" This indicates that the tester is trying to use his interpretation of the results to coerce an answer out of the subject.

    These tests can easily be beaten and can easily show results that are not 'truth'.