Slashdot Mirror
The Truth About the Polygraph, According To the NSA
An anonymous reader writes "The NSA (the secretive intelligence agency that brought you wholesale warrantless wiretapping) has produced a public relations video about its polygraph screening program titled 'The Truth About the Polygraph.' But is the NSA telling the truth? AntiPolygraph.org provides a critique (video)."
Penn & Teller taught a random woman who answered a Craig's List ad how to fake a polygraph response in less than 30 minutes.
Polygraphs aren't lie detectors. They are used to assess truthfulness. Much of the magic is not in the machinery itself but in subjecting the person under assessment to unfamiliar, semi-stressful conditions while asking probing questions. It's basically a game of manipulation for the polygrapher.
I am becoming gerund, destroyer of verbs.
Polygraphs are basically a placebo to make you believe that they can detect your lies. A lot of theater and psychology goes into helping enhance that belief - things like using 'scientific looking' equipment (the more complex the procedure the stronger your belief will be that it 'works'), having the questioner dress in labcoat (it enhances our authority belief), using escalations in authority (switching to a more 'experienced' examiner part way through), pointing to a random squiggle and claiming that it shows you lied on some vague question to convince you to change your answer and admit to something.
So its just like a Scientology body thetan test machine?
It's the same situation down here, too. Police forces and various government offices use polygraphs while hiring. It makes sense, really. They want to make sure that you can lie convincingly. I'm not really sure the purpose of putting sociopaths in power, though.
These men, and others, were all employees of the CIA, NSA, or other intelligence agencies. All of them were subject to taking and passing one or more polygraph tests. They all ended up providing classified information to the Soviets for a relatively minimal amounts of money. The information they disclosed resulted in the compromise of highly useful, and costly, collection systems, data, and human assets, some of whom were killed as a result. In a number of these cases, Aldritch Ames, in particular, the agency they worked had suspicions that something was going on yet these men remained free to continue their spying. Ames was even tested again, passing the test to continue his work.
The polygraph, in these instances, was worthless and, in fact, provided a false sense of security to the detriment of the country's well-being.
Not sure how this got a tagged as an NSA video, it's from the DSS. The DSS is the organization responsible for granting security clearances. The process they're describing is the polygraph you take to receive certain security clearances. Anyone who is taking this polygraph has applied for a Top Secret-level security clearance. This process is pretty much the same for anyone applying for these clearances, doesn't matter if they'll be working at the NSA, another three-letter agency, in the armed forces, or for a private defense contractor.
I like those. I did one of their "test" once. The guy talked to me, and asked lots of questions. I remained calm, and answered every one of them any way I wanted. The needle didn't move. After a few minutes, he began doubting the machine, and then questioned me on if I was operating it right. With the simple instructions "hold these loosely in your hands", there wasn't much for me to mess up. Since he had turned the sensitivity all the way up because he couldn't get a response, when he told me to hold them a little tighter, the needle shot all the way to the right. I suggested he turn the sensitivity down. :)
I held on a little tighter, and he adjusted the machine again, so it was now showing neutral. The questions resumed, and I didn't show any sort of reaction to any of the questions. He got real frustrated with me (Hey dude, reactionary mind. Practice what you preach.), and gave up on it. I guess I wouldn't be a good cult member, if they won't know that I'm lying to them or not. Too bad, I wanted to join up, so I could take over. ;)
If you really don't care about what you're saying, everything will show you're answering truthfully. When you start overthinking the questions, that's where you'll run into trouble. Consider these questions during a polygraph.
(Q = question. T = thought. A = verbal answer. R = Result)
Q: Did you know the victim Bob?
T: Ya, I know bob.
A: yes.
R: Pass
Q: Are you aware that Bob is missing?
T: Everyone knows Bob is missing, that's why I'm here. This is easy.
A: Yes
R: Pass
Q: Do you know where Bob is?
T: Buried in that empty field. Shit, they know I killed Bob. They're going to figure it out!
A: No.
R: FAIL!
Q: Did you have anything to do with Bob disappearing?
T: Oh shit, they know I did it. They know I shot him, and buried him. I'm going to prison forever.
A: No.
R: FAIL!
Serious? Seriousness is well above my pay grade.
Really, all you need is to convince the person you're investigating that it works ... then if they refuse|agree to take a polygraph they're probably guilty|innocent.
Actually, as AntiPolygraph.org pointed out, it convinces people to submit to an interrogation without a lawyer. Standard interrogation techniques can get you to confess to things (sometimes to things you're not guilty of). They can also collect information that they can use against you in combination with other (mis)information.
See the Youtube video of a law school class by law professor James Duane http://video.google.com/videoplay?docid=8167533318153586646#. (Or see http://flexyourrights.com/)
Duane said, don't talk to the police if you're innocent. Don't talk to the police if you're guilty. Don't talk to the police without a lawyer.
You can tell the complete truth, and make a true statement that can be used against you to convict you.
Like: "I never liked the guy."
Or: "I was in the next town." Then they finds a witness who honestly thinks she saw you near the scene of the crime, and they use that to impeach your credibility.
When I was 19 I worked at a pawn shop. After working there for 6-8 months something (I don't know what) happened and everyone was lined up from 3 stores for polygraphs. We were let know in no uncertain terms we would lose our jobs if we failed. I was so nervous that I bombed miserably and got fired. I had done nothing. Polygraphs are simply a way to kick you in the nads and see what responses they get.
Except it's actually so bad for their intended use that it's a hindrance to their job, not a help.
Let's say you have a test for a virus that is 99% accurate, but 1% of the time, it gives a false positive (says someone has the virus, but doesn't). Let's also say that there are 1000 people that might be infected, but nobody is sure. All of them are tested, and on average, 10 of them will show up as false positives. That number can probably be delt with, perhaps with a more expensive (but more accurate) test, or maybe the treatment is no big deal (so they can just get an injection and go on).
OTOH, let's say that 1 million people might be infected. Test all of them and there ends up being 10,000 false positives. Now the costs of the more accurate test start rising. Perhaps the treatment is more dangerous or expensive (rabies shots used to be pretty nasty, for example), so you really don't want to use it on people who aren't really sick.
So your accuracy rates need to be in line with how many people are going to be screened. If its use is highly targeted, then a test that's 90% accurate might be OK. If it's more of a general screening, then it needs to go into the five-9's kind of accuracy, perhaps more.
Polygraphs are nowhere near 90% accurate. It's maybe 70% accurate, and has both false positives and false negatives. For general job screening, like the NSA is using it for here, that's nowhere near good enough. It might be good enough for police investigations as a way of seeing if they're on the right track, but there's a reason it's not considered admissible court evidence in the US.
Of course, all this is only focusing on the basic statistical issues. There's a whole other set of arguments surrounding privacy, which matter even if a future technology is 99.999% accurate.
Not a typewriter
I would refuse to take a polygraph test under any circumstances. Therefore, my cooperation has no correlation with my guilt or innocence.
By the way, why would I refuse to take such a test? Simple. It cannot help you in any way. They can take the things you say in the test and use them against you in court, but no matter how flawlessly you pass the test, as a defendant you cannot call on any of that testimony in the court room. Only the prosecution can call testimony from police interviews. So basically, it can hurt you, but it legally cannot help you at all.
Not really. I'm neither guilty nor innocent, I'm not even connected to a crime and I will refuse to take a polygraph test if ever asked- for any reason.
It's the same when a cop pulls you over and asks to search the car or something. I always default to no you may not. They usually reply with something about something to hide and I reply that if they knew that, they wouldn't need my permission to nibshit through my stuff. I then ask if I'm free to go. Of course they will not find anything if they look, but I'm more worried about them finding something that wasn't there before they looked. If they are honest, then it won't be a problem. If they are not honest, then it's a door to escape the issues at their hand.
It might be a different story is there was some trust surrounding the officers enforcing the laws, but a fe bad apples spoiled that a long time ago and continue to keep it rancid today. The problem is that you cannot tell which are the good cops and which are the bad cops and it's best to just not take chances. Especially when they want to search you or pin something on you that you had nothing to do with.
Yesterday, upon the stair
I saw a man who wasn't there.
He wasn't there again today.
I think he's from the NSA.
Protoplasm. Quiet Protoplasm. I like quiet protoplasm.