Hacker News new | past | comments | ask | show | jobs | submit login
My NSA polygraph experiences (antipolygraph.org)
98 points by peter123 on Jan 10, 2009 | hide | past | web | favorite | 46 comments

Polygraph results have been inadmissible in state court cases in my state for a long time, because of a late professor of psychology who did much research on polygraphs and debunked them.




The worst risk in federal government use of polygraphs is believing their results, whether it's a result that clears the person who took the polygraph examination or one that implicates the person who took the polygraph examination. Pseudoscience is not a fact-finding tool.

Every time I read one of these accounts, or watch a police interrogation on TV, part of me feels a bit curious. How I would perform in that sort of situation? We (i.e., even hackers) have ways of testing our intelligence and athletic stamina, but what about various forms of psychological fortitude?

It sounds odd, but maybe a mutual interrogation club could be a rewarding way to try this out. Like, person A tells you a secret and then locks you in a room with person B, who knows part but not all of the story and must try to extract an additional detail from you. One either side of the interrogation, you could learn a lot about high-stress psychological situations, and their impact on you and other people.

but what about various forms of psychological fortitude?

Gahhh, there's so much misinformation here. Most people have this idea of what a poly is thanks to TV and movies.

All I'm going to say is that polys are not fun for people who are insecure, have emotional/anxiety issues, or go in with the mindset that they have to "beat the test" or "outsmart the examiner". It's not a measure of psychological fortitude at all, and anyone who views the test as such will be extremely stressed out by it all.

The only legitimate reasons polys are stressful is because they waste half your day, coop you up in a room and tell you to stay as still as possible for most of the 4 hours. That's not really anybody's idea of fun. For the majority of people who don't let all the silly worries get to their head, you go in, chill out, tell the truth, and leave.

I'm not interested in "learning how to beat the test". I'm interested in understanding the experience of an interrogation, from both the perspective of the interrogator and the person being questioned. And I'm wondering if a form of role play would be a feasible way of exploring that experience, without going to the trouble or risk of committing a crime or becoming a police detective.

There are clearly lots of aspects of interrogation that are difficult in uninteresting ways, like sitting still for many hours. But the reason TV/movie interrogations are so intriguing to watch (and, I imagine, to act) is that they represent an unusual and intense interaction between people. People who are trying to manipulate each other in real time in pursuit of some valuable information, and willing to break a lot of normal social boundaries in that pursuit.

That certainly sounds like it entails some interesting psychological challenges. If I think I'm secure, relatively free of anxiety, and willing to engage in those challenges, maybe it would be a revealing and worthwhile endeavor?

I received a government polygraph test over a year ago.

A mock interrogation would not be very effective. The most important aspect of a polygraph is leverage. There has to be an incentive for the person being questioned (e.g. they want an out for committing a crime) so that they want to tell the examiner the truth. There should also be a real fear of failing the test. Your brother questioning you about who stole the cookie from the cookie jar wouldn't cut it.

As far as I can tell, the goal of the test is to make you spill your guts. You might consider yourself confident or be a pathological liar, but when you're afraid of lying, your body will react to it. It's a very unique experience and I don't think practicing lying or rehearsing your responses would be useful. They'll dwell on questions that they think you're trying to deceive them on and ask you over and over. Each time, you'll give them a little more ("well, one time, but I didn't inhale!") to get them off your back. They'll do the good cop/bad cop routine and occasionally storm out of the room. Eventually, hours later, you'll crack and think, "fuck it, I've already told them this much, might as well give them the rest." At that point, you're probably crying or look pretty distraught so they know they've got everything.

You could always apply for a job that requires security clearance. In my case, I once applied for a job at the CSE (Communications Security Establishment -- Canada's equivalent of the NSA), which required top secret clearance. I would have had to submit to a polygraph had I not backed out to accept another job.

Having a psychology background, I was curious about the test and somewhat disappointed that I didn't have to take it.

I have the same reaction whenever I read about pyschological experiments like the milgram's experiment. I think that I would have stopped administering schocks long before the end, but given the statistics of people who didn't I'm probably overestimating myself.

That's why I sometimes think it would be beneficial to do such experiments to better understand your character, your limitations and how you react under pressure...

Maybe a group of people who would pick one such experiment a week and have some administering the test and some receiving it would be fun....

I wonder if there are some trends for the people that were the minority in this kind of experiment.

Something like this? http://yudkowsky.net/singularity/aibox Not an interrogation as such, but definitely a test of psychological fortitude, and of the consequences of "being locked in a room" (even if voluntarily) with someone who is trying to get inside your mind (oddly, to write to it instead of to read from it.)

I don't get it; how in character were they supposed to be when deciding? Sure, there could be convincing arguments for release made by an actual AI, but for ME the real person sitting in IRC it would be easy to just say no.

The classification levels below "top secret" involve a laughable laxity. But spying is a very real problem, and the government needs to screen people who handle actual state secrets.

I'm guessing that they are (in 4 hours) trying to simulate some of the pressures that go on when people are recruited and act as spies, and especially when they are blackmailed into spying. All the pressure tactics probably give testers a fair idea of how quickly people crack.

Huh, in that case it is better to learn to lie very well, then tell them after the successful test what you lied about. That seems like a different kind of training.

This sounds very similar to my wife's experience with taking a polygraph at the NSA for an internship (and this was before 9/11). They hounded her over the drug use question over and over. She was so upset by the whole experience she nearly flew home right after and I can't say I blame her.

Is there any evidence ( even anecdotally ) that drug use and spying are related?

The government security clearance process seems to be very focused on recreational drug use history.

It may not have anything to do with fear of spying. Drug use is something that a) most people have done, and b) are least willing to admit to. While some of us may shy away from telling people of our sock fetish or whatever, admitting to drug use has real and far-reaching consequences.

So in other words, getting you to admit to drug use would likely be a reasonable representation that they have ascertained "the whole truth" from you.

Also if the operator can't see an alteration in the graphs, he can think that the subject has been "inmune" to the machine and try to break him "manually" like in the linked article or like arockwell's wife. The drugs would be used because, as the interviewer himself said in the linked article: "everybody have done drugs some time", so it's a good guess, even if it's "most people" instead of "everybody".

Perhaps I'm the only person here who doesn't fit (a), but I have never done any drugs().

() Excluding alcohol and caffeine

I haven't either - but being in college right now I would say that a very large number of people have :)

The issues that government agencies have with drugs are (1) blackmail potential and (2) risk behavior.

To maintain a drug habit, one usually needs a relationship with a dealer, and drug dealers (in addition to being criminals) are generally pretty awful people. So, if you're a regular illegal drug user, there's someone out there who is (a) committing a serious crime, (b) probably of bad moral character, and (c) knows that you've done something illegal. Although it's extremely unlikely that a drug dealer is going to blackmail his clients into revealing state secrets, putting oneself into such a position is taken as a sign of bad judgment.

The second issue with drug use (as pertains to the psychedelics, most of which are not illegal, e.g. salvia) is the risk attitude it shows. Drug use is taken as an indicator for how a person handles low-frequency risks with very high loss potential. Since no one who hasn't used psychedelics can understand their subjective effects and risks, a person who takes a mind-altering drug is undertaking at least an epistemic risk of blowing out his mind. There's obviously a valid philosophical debate regarding how the risks of drug use compare to those of other pleasurable activities (sex, skydiving) and, frankly, I don't think psychedelic drug use is worse, from a cost/benefit perspective, than a host of "normal", socially acceptable human behaviors. However, the psychological risk inherent in drug use is taken by government agencies to be fairly singular: people who die in skydiving accidents don't present a risk of revealing state secrets or creating image problems for government agencies.

Is there the possibility that everyone who answers "no" truthfully to the drug question is intentionally hounded like this? Intelligence work is incredibly stressful. This seems like it would be an opportune time to see how someone handles stress conditions without water boarding them.

The Lie Behind the Lie Detector pdf (referenced in the write-up) is also a good read.


I wonder how much better people would take it high or drunk. Maybe that's why they make you wait 4 hours, to let anything you may have taken to relax you wear off.

Cannabis edibles can easily exceed an effect duration of 4 hours.

Really though, the long interrogation time is probably so because people eventually become uncomfortable, being in the same state for such a long time. They want to see how you react in that situation.

Especially things like relaxants, maybe Valium or Vicodin, last about that time frame... and granted, these might as well be prescribed and perfectly legal!

Folks that want to apply these techniques have jobs and management and organizations that I simply don't want to have to deal with; fixed, inflexible, distrustful, political, bureaucratic.

There might well be good and valid reasons for any or all of these organizational choices, but I too have a choice in the matter. And I choose not.

I disagree, while stressful depending on the situations, certainly there a few agencies that do have flexible work schedules... one that I refuse to name, that you can find with a little digging, adheres to the employee's work schedule.. so if you like working at night, they'll provide that schedule for you, granted that you've gone through some reasonable hoops to get there.

They were probably pretty clear about the fact that this guy is NOT supposed to discuss what happens in a government poly especially on a public forum. There are reasons beyond his career to want to keep that information close. People are constantly trying to beat our system and the more they know about it, the easier it is. If this guy is serious about pursuing a career in the intel world he should respect the rules they lay down. They are bureaucratic and often silly, but they have reasons for them beyond what a young hacker can understand. This guy has shown that he is not responsible or mature enough to handle classified material.

Polygraph tests don't depend on secrecy. I, and millions of other people, have received one and it's fairly well documented. From what I can tell, the way that they polygraph criminals and people in intelligence is roughly the same.

It's used as more of a method to trick you into telling the truth than it is to detect lies. I don't think it's very easy to beat unless you teach yourself how to regulate your heartbeat.

It is in fact very easy to beat.

You don't have to control your hartbeat, you have to overreact to control questions (performing heavy mental calculations combined with holding your breath 2 seconds will do the trick!), so your (normal) response to the important questions will be less severe in comparison.

If your response to control question is stronger than your response to relevant questions, you pass the test. It that simple.

It's really not as simple as these anti-polygraph sites would have you believe.

There's a sensor that's strapped around your chest. If your breathing is irregular, they'll detect it. Also, doing anything mentally draining while someone is asking you a question and answering it correctly is not easy. At some point, you're going to have to focus on the question. You'd have to be able to consistently deceive them for every question and keep at it for several hours.

They don't necessarily tell you which questions are the control and they'll slip in questions that you'll have to think about and answer honestly. They'll catch your baseline with those. At times, they'll intentionally try to agitate you by accusing you of lying. It's difficult to keep cool and it will throw you off balance. You'll forget that you were trying to trick them.

The government knows about all of these stupid little tricks. People used to put pins in their shoe or contract their anal sphincter muscle to generate a response. Examiners now have sensors under your feet and ass. These tests aren't designed to simply detect that you're lying. They want to push your buttons and make you crack.

Silence is not a good a working strategy, I guess?

A polygraph is not the same as law enforcement questioning you. You're taking it because you want a security clearance or you want to keep your parole.

You can be stubborn about it and lie and insist that they've got it wrong, but your reading will show that you're holding back on them. If they sense that you're trying to deceive them, they'll fail you.

Thanks. We do not have that polygraphing here in Germany.

Over a 22-year period, Robert Hanssen, one of the most notorious spies in US history, passed several polygraph tests during his tenure at the FBI while he also sold secrets to Russia.

Both hanssen and Ames came up inconclusive in their polys. But they were able to cleverly explain them away. Coming up inconclusive is easy, passing outright is hard. To get a job at an intel shop you must pass. The problem with Ames and hanssen was in the organization, not the poly.

Can you speak from experience?

Excellent question and I'd like to see an answer - as I was thinking this too when I saw that post.


You just made me so glad that this guy posted what he did. Thanks for making an unpleasant read seem a bit more worthwhile.

Ok, let's put it this way, would you want the playbook for your favorite football team to be public? You may be a fan, and you may be paying for the team but you'll loose if their secrets become public. The same goes for the intel world. Part of being on that team is keeping the playbook to yourself.

I was vaguely intrigued by the notion of the OP revealing some unknown insight on polygraphs or whatever, as well... But after reading it, the fact that the story could have come beat for beat out of a John Grisham novel made me dismiss that fairly quickly.

"our system"

Yes, your system. Canadian intelligence is intimately intertwined with American. In fact, it's probably dependent on it.



I think, at least what I found from that site as a pointer from this link is, that people confuse the polygraph with being a lie-detector which in truth(again based on the readings), it is not. What it is, is an agent used in an interrogation, simply put. You are being interrogated and judged on how well they think you can actually be trusted with a security clearance(along with other factors in your application). It will have you believe that they are on to you by showing your reactions to certain questions, when in reality, there is really only one person in that room that truly knows if an answer is true or false, and it is proven that it is not the person administering the examination.

Seems like fun in a sick kind of way. I wonder if there's a way you could turn this into a novelty game and charge for the experience. The biggest problem would be capturing the authenticity I think. Is there a market for doing NSA quality interrogations? :-)

Authenticity could be gained not telling the subject the date of the interrogation.

The effectiveness of the lie detector seems like another version of the belief that God can see your every thought and will judge you accordingly. It is something that is more effective in producing confessions the more people suspect it might be true.

Applications are open for YC Winter 2020

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact