Challenging The 'Unbeatable' Polygraph 101
George Maschke writes "Dr. Louis Rovner, a prominent California polygraph operator, has (through PR Newswire) issued a press release titled, 'Polygraph Unbeatable, Says California Psychologist.' All too often, such publicly-made claims by those with vested interests in the perpetuation of polygraphy (a make-believe science that offers make-believe security) go unchallenged. So, I've publicly challenged Dr. Rovner to support his claim and pointed out scientific research that contradicts it, as well as the examples of several notorious spies and a serial killer who have beaten the polygraph. See, A Public Challenge to Dr. Louis I. Rovner."
96% accurate? (Score:4, Informative)
"Overall," says Dr. Rovner, "we are confident that polygraph tests have a 96% accuracy rate when done properly."
If that is true, then if you have 1 spy and 49 honest people, this polygraph will likely falsely accuse two honest people as being spies.
Re:So... (Score:5, Informative)
Re:Polygraphs and plants. (Score:4, Informative)
http://antipolygraph.org/cgi-bin/forums/YaBB.pl?b
Don't forget human polygraphs (Score:5, Informative)
Polygraphs are bunk (Score:5, Informative)
Afterward, the guy puts his arm around me and tells me I passed and that one lie that I told about the pot wouldn't be held against me. He patted me on the back and sent me on my may.
One anomalous response was interpreted as a lie. A faulty technology had convinced a total stranger that I smoked pot when I never had. The report of that session went to my new employer who didn't fire me but did make the report available to another employee who happened to be my sister. To this day, she thinks I've experimented with drugs when I haven't. After all, what's my word balanced against a neat-o cool technology with all those scribbling pens and sensors and stuff, right?
Polygraphs are bunk. People who make their living in that industry are, by my definition, liars and should be shunned.
Yes, I know I'm only one data point. But sometimes it only takes one data point to know when a technology has failed and is not trustworthy in broad application.
Re:So... (Score:4, Informative)
Re:false positives (Score:4, Informative)
Other agencies that administer lifestlye polygraph examinations, such as the CIA and NSA, do not make their polygraph failure rates public, though I suspect that they are somewhat lower than the FBI's.
In the Department of Defense (which uses a counterintelligence-scope polygraph), virtually everyone passes: the only ones who "fail" seem to be those who make what DoD terms "substantive admissions."
For information on what you might expect during your polygraph examination, and tips on how you might protect yourself against the risk of a false positive outcome, see Chapters 3 & 4 of The Lie Behind the Lie Detector [antipolygraph.org] (1 mb PDF).
Re:remorse (Score:3, Informative)
A polygraph works based on relative stress. Usually it goes as follows:
First they ask some "control" questions. These are questions they already know the answers to. From that they determine your baseline responses. Then they start asking the real questions. They will usually introduce more control questions during this process so they can make sure the baseline is still working (you won't know which are the control questions during this phase).
Do beat the machine all you have to do is fuck up the baseline. There are different ways of doing this. If you can cause physical pain to yourself that will stress your system. With a little practice many people can learn to control the stress reaction without any physical stimulus needed. By manipulating your stress at appropriate times you can distort the baseline so the polygraph interpreter can't make any solid decisions.
And of course there are people who can just completely eliminate the stress response in their system. Those are often the people who can beat the machine without any training at all.
Re:remorse (Score:3, Informative)
Sort of. This is how they work:
There are null questions, 'control' questions, and pertinant questions which are asked by a tester who also attempts to spook/convince the victim that their 'high tech' equipment actually gives them the ability to tell if someone is lying. No equipment can tell if you are lying short of a PET scan which they can't afford to give you. Even a PET scan may not be able to distinguish from someone making up a story from scratch and someone making up details to fill in gaps in their memory. Making things up to fill in gaps in memory is something we all do.
The tester tells the victim that the null questions are actually 'control' questions. In fact they are merely a ruse. They will ask questions like: what is your name, or what is your date of birth, but the answers and the polygraph data for those questions is discarded.
The real control questions, are questions the tester assumes that the victim will lie about. They want to see what the polygraph data looks like when you do lie. They might ask: "Have you ever told a lie?" Everyone has told a lie. If the victim lies and says: "No, I've never told a lie." then the tester gets his data. If the victim is truthful, and admits to having told a lie, then the tester follows up with something like: "Yes, I know everyone has told a lie about SOMETHING at one time or another, but what I'm interested in is, have you ever told a lie about something important? The tester assumes that the victim has told a lie about something important. If the victim denies lying about something important, then the tester get's their control data, if the victim admits lying about something important, then the tester ups the ante, saying something like: "Yeah, lying is neccessary sometimes, but have you ever told a lie that hurt someone else, or got them in trouble?"
The ante is upped until the the victim denies something. Testers like to pressure victims to lie on the control questions by asking control questions that seem pertinant to the issue at hand. For a job interview, they might ask about stealing from work. Even someone who never filched anything can be coaxed into lying about the five extra minutes they 'stole' on coffee break by a clever questioner.
Then there are the pertinant questions. These are the questions that the tester is trying to determing whether you are lying or not about.
The testers ignore they null questions - they are only part of the ruse. They expect to see the needle move in response to the provocative 'control' questions, and the needle not to move in response to the pertinant questions. This is the ideal "He's telling the truth" result. If the needle moves in response to the controls, and the pertanent questions, then the result is marked as "He's lying."
The third possibility is that the needle never moves for control questions, or for the pertanent questions. If the victim admits to every control question the tester asks, they will have eventually wound up admitting to having sex with their neighbors dog while a busload of fourth graders watched, or some other outrage. If they did not admit to any of the control questions, and the needle didn't move, then the test is inconclusive. They could be a saint, or more likely they have no concience and can lie without making the needle move. Saints and Psychotics are filtered out of the Job Candidate pool in favor of imperfect, morally conflicted people who make mistakes and lie about it, but have at least not lied about any of the 'pertenant questions'. It pays to be what the testers know how to test. As for sainthood, no good deed goes unpunished.
It's possible to get the needle to move by biting your cheek, clenching your anus, stepping on a tack in your shoe, or
Re:Polygraphs are bunk (Score:3, Informative)
1. Reaction when no question is being asked.
2. Reaction to a question where you have no reason to lie.
3. Reaction to a question that where the true answer is embarrasing. For this question, the polygraph is detecting an emotional response rather than a lie (e.g. Have you ever imagined what it's like to have sex with your mother?)
4. Reaction to a question where they explicitly expect you to lie.
If any of these aspects is missing, then the test is incomplete. Until then, you cannot tell whether a general statement is truthful or a lie.
The common patterns for this minitest would be #3 being higher then normal, incidating that you are generally truthful, and #3 and #4 indicating that you are generally lying. Other cominations are possible - if reaction to the four tests are the same, then either you are a nervous wreck or extremely relaxed.
That's definatly true. All polygraphs do is measure body reaction to something, which doesn't always indicate truth or a lie. The only true way to tell if a person is lying is if you take a look at the person's brain to examine the memories precisely - a method that's extremely invasive.
Re:96% accurate? (KNOW YOUR STATISTICS!) (Score:3, Informative)
to have what you suggested, the test would have a 96% "POWER" (or a 4% beta error for n=50).
Alpha error (Type I error) = Probability(X returns false | X is actually true) => false negative
Beta error (Type II error) = Prob(X returns true | X is actually false) => false positive
The actual probability that a 96% accurate polygraph machine would do what you describe (1S + 49 => 2S that are really H) is:
(.96^47)(.04^3) =
This says that it got 47 people right (47H) and 3 wrong (2H + 1S). THAT is what a 96% accuracy means.
Do I believe the doctor's claim that the machine is that accurate? Hell no. I think the doctor that created the claim has a failed understanding of whata statistics actually mean and is giving a stat for something else entirely.
-Ab