Hacker News new | past | comments | ask | show | jobs | submit login
The Lie Generator: Inside the Black Mirror World of Polygraph Job Screenings (wired.com)
73 points by rms 11 months ago | hide | past | web | favorite | 94 comments

I once was gassing up at a station, and for whatever reason the card reader on the pump wasn't charging, but would let me pump gas anyways. So me, and everyone else who'd used that pump that day, received a phone call from the police telling me to go pay for the gas. I figured it was just an honest mistake and didn't think anything of it.

Fast forward a bit, to where I am undergoing a polygraph examination for the NSA. The exam made me uncomfortable and nervous, but I thought everything was going well. Except for when my interviewer came back and told me I was showing sensitivity towards the hiding crimes question. WTF? And when they do this, they're just giving you enough rope to see if you hang yourself with it. But I had no idea why (or even if) I was showing sensitivity to this question.

They called me in for a 2nd polygraph, this time I didn't show sensitivity to hiding crimes, and I figured I was good to go.

No. I get called into a 3rd exam (each exam was separated by a couple months, mind you). This time the interviewer told me "You did better at the hiding crimes question than I thought you would" W.T.F.?!?! The interviewer then left the room and came back with a manilla folder, from which he procures a piece paper which he reads that I had a suspected larceny charge back at home. I honestly had no idea what he was talking about until I remembered the gas station incident. But after I try telling him about it, he tells me that he doesn't believe me and that he thinks I stole that gas. This leaves me extremely flustered and the rest of the polygraph was a train wreck.

3 strikes and I'm out, my conditional employment with them was terminated.

What irks me the most though, is that when I got back home I retrieved the larceny report from the court house, and in that document the whole story was laid out and my account of the situation was corroborated. So what the hell? Why throw me through such a ringer?

Fuck the polygraph.

...received a phone call from the police telling me to go pay for the gas. I figured it was just an honest mistake and didn't think anything of it

So, given that you paid-up for the gas when requested, you didn't intend to permanently deprive the owner of it which suggests that technically it wasn't larceny and could have been sloughed-off as a misunderstanding. Given you were dealing with an NSA interview, sounds like they were playing a psychological setup game to determine how calmly you would respond under pressure in the context of an intimidating polygraph test. Seems like they weren't too thrilled with the responses they elicited. Would you really want to work for an organization that plays those sort of mind games?

> Would you really want to work for an organization that plays those sort of mind games?

With each other, on a regular basis, for little more than shits, giggles and office-politics? Probably not.

Which is why it's more likely that this mind game isn't just some office antic, but instead means to mimic some class of situations that some kinds of three-letter-agency personnel could be required to overcome.

>Would you really want to work for an organization that plays those sort of mind games? reply

No, I don't anymore. But at the time I was excited for the opportunity

You dodged a bullet.

That hiring practice sounds terrible when you frame it like that. It seems like a great way to build a terribly aggressive, pressure-driven monoculture.

"Would you really want to work for an organization that plays those sort of mind games?"

I would like to work for an organization that pays me so I can buy food and pay rent.

I'd definitely want that type of ruthless organization defending my country.

It’s a subjective interrogation with a prop. You tingled his spidey sense.

And was he wrong?

You pumped gas without paying. Did you acknowledge it as a criminal act or as good fortune or not at all? Look at it from the interrogators perspective.

The OP said the card reader didn't charge. The implication is that they swiped their card and pumped gas thinking they paid.

How is that a criminal act?

I think OP's point is that he didn't even remember the event, and the NSA was using the polygraph as a prop to try and get him to fess up to something he believed was trivial, so it was erased from his memory.

Yes; the background investigation produced a police report which I hadn't mentioned in my application. At the time of the gas station incident, I didn't know that I had just created a paper trail that would damn my background check.

On one hand I understand how a discrepancy in the background check would raise concern, but on the other hand I feel that the police report easily shows no wrong doing on my part, and that this "blemish" could have been cleared up at the 1st polygraph interview, rather than stringing me along for 3.

All of that would likely be true of a regular employer. Probably even would be good enough for the military.

But the people you were dealing with, as a rule, don't believe in happenstance, or coincidence. And they absolutely live by attention to detail. I don't know much about your situation, but based on what you've shared, you tripped a lot of alarms.

Only if the agents do not read the folder.

Forgetting to pay at the pump once in your life can hardly be called criminal behavior.


Well then congrats to the NSA; they successfully weeded out a spazzy dork

To be honest, the most amazing part of your story is everything that happened at the gas station. Not so much that you were caught on camera pumping the gas, but that they went after you for it, and that it wasn't pursued in a civil context. Not even a mailing you a bill, or sending a collections agency after you.

Okay, sure, there was no reason not to go inside and pay cash, and ask what's up with the pump, and maybe tell them to put a sign up, when it's all of maybe $20, maybe $40 at stake for a single tank. But I really want to know the background story from the perspective of the gas station. Like, how bad was the hit on that malfunction for them?

I really have to wonder, like was it a Mom & Pop station, or was it some huge multinational chain? And how long did the incident last? Did they lose an entire underground tank's worth of gas in less than a week, or even one day, with no transactions posted? I see "out of order" signs all the time. How did station employees not notice, and disable the pump in time? How many people got scooped up in the drag net?

They must have lost at least six figures worth of gas, and if it wasn't an independent station, I can only imagine that multiple station attendants lost their job for not disabling the pump. For a gas station to open a criminal investigation, to the tune of police reviewing possibly more than 24 hours of footage, running plates for every car that skipped out on a pump's error message, and making phone calls to track down individuals to dispense a warning under penalty of criminal charges, it must have been a real mess, and a total fiasco for the station owner.

Considering the insurance required for handling hazardous materials in a motor vehicle context, where a sleepy trucker could send an entire 18 wheeler plowing into the pumps, destroying an entire station, or a rusty leaking tank having the same effect, it's surprising that they wanted heads to roll over a malfunctioning credit terminal.

>And how long did the incident last?

Not long, I left the gas station and received a call from the police within an hour or two, and I kind of remember them telling me it was only happening that morning.

I only remember that it was a Sinclair station

That's the thing, they want to know if you would bring up things even if you didn't think they knew about them. They want people who would do that, due to believing the polygraph works.

So they questioned them about this thing they knew about and he didn't reveal it, so they thought he was lying even though he honestly didn't remember it.

It doesn't work out well for people who know how polygraphs work in general or for people who do not reveal things they know about (intentionally or otherwise).

How did the police get your number?

Vehicle registration?

It was my understanding that polygraph tests were mostly useless and just a way to claim someone is lying through clever use of pseudoscience

They're not useless, but yes, they are a form of trickery. It's a psychological technique that tries to make people who are lying or hiding things more likely to be nervous and more likely to confess (or be led down a line of questioning which eventually reveals new information).

The actual results from the machine are pretty much just a red herring, but that doesn't necessarily mean the machines have no valid uses. If use of a polygraph has helped tricked even one criminal into (truthfully) confessing or scared away even one person with malicious intent, then they're useful.

I wouldn't want a polygraph to be used for evidence in court, but I would want them to be used if I were interviewing for FBI/CIA/NSA. (But I definitely wouldn't want them to be used for firefighter and parademic positions.) I know it's a bit paradoxical that they can be both useful and a sham, but I think that's how law enforcement views them, too. This article presents a lot of issues with how they're used, but I wouldn't expect them to be phased out for a very long time (unless someone makes a version that actually has more scientific validity).

Polygraphs should nevern be used. An administrator can interpret the results any way they see fit. It's a perfect avenue for bigotry and other unfair hiring practices. It has no place in society.

Again, it's not about the results. But I agree it shouldn't be used for hiring, except for extremely sensitive positions.

Its not about the results and that lets the person running the test interpret things anyway they want. We have seen time, and time again that this leads to all sorts of discrimination that is both against our society's morals and not actually useful as filtering people as well the person doing the discriminating thinks.

> but I would want them to be used if I were interviewing for FBI/CIA/NSA


As well as being known to be a form of trickery it's also the case that it is easy to teach someone to manipulate the results at will.

All you achieve requiring it for those organisations is that there will be a) some who are rejected unjustly thanks to a false positive, and b) the people you really don't want in there have another bit of flim-flam to reinforce their apparent legitimacy.

Yes, undoubtedly there are good candidates who are weeded out due to false positives, but there are probably many bad candidates who are weeded out due to true positives (or don't apply at all because they fear the test). It will lead to bad actors who study the test and pass it, but that's why it's critical that law enforcement agencies do not use a passed test as indication of trustworthiness; merely an initial barrier to entry to be passed. The trustworthiness has to come from a thorough background check and normal interviewing.

As long as the examiners, recruiters, interviewers, and higher-ups fully understand polygraphs are a trick and nothing more, I still think they serve a useful purpose for three-letter agency hiring.

You could say the same thing about homeopathy - they are an effective placebo.

Do you support their use too?

I don't support the use of homeopathy, but I don't think that's a great analogy. Polygraphs are used as an interrogation technique, where psychological tricks and deception on the part of law enforcement are "fair game" (we could enter a long discussion as to whether or not deception during interrogations/interviews should be permitted, but regardless, that's the status quo right now).

Trickery and deception when it comes to medicine and personal health are never fair game. I think there's definitely a difference between police telling a suspect "we know you did it" (when they don't actually know) and a doctor telling a patient "take this and you'll feel better" (when the drug actually does nothing).

Also, polygraphs are a little bit more functional than homeopathic medicine: a polygraph can not only induce nervousness, but can also notice it in some cases (even if the nervousness may not be concerned around a particular question and may not necessarily indicate deception and even if there are false positives). This helps create a feedback loop which makes it more powerful than a pure placebo (e.g. a "lie detector machine" you hook someone up to which isn't powered on at all and does literally nothing).

> Trickery and deception when it comes to medicine and personal health are never fair game.

What about when it is known to work?


It's been shown that telling someone you're giving them a placebo can still bring benefit, in which case I'm okay with it. But giving someone a placebo or homeopathy pill (same thing) and failing to disclose it contains no active ingredient is pure deception, unless the patient has given consent that they may receive a placebo, like if they're in an experimental trial. Even if someone does show improvement after receiving the placebo, it's never justifiable to outright lie to them.

It's a great way to discredit someone if you get lucky enough that they perform poorly on the metrics that the test measures.

True story: we hired a guy who failed his polygraph to work in the motor pool for a local municipality. the test was required because he would be working on police cars and fire trucks.

six years later and hes still working for us, we get an order from that same municipality to overhaul the intercoolers on nearly two dozen cop cars. I called up the pool manager and asked about the polygraph, and his response was they use outside contractors to get around the fact they have no certified mechanics.

I was looking for a statement by the "other side", someone in government who could give a justification as to why they're being used.

Is it simply because its the "standard" now, and bureaucrats don't want to stick their neck out by getting rid of it? Is the fact that it is a machine that has been around for awhile, regardless of efficacy, give people that much comfort? Or are enough people really that misinformed?

Its a bit like marijuana legalization coverage - its rare to find arguments for maintaining the status quo as opposed to getting rid of it.

Lots of technology doesn't work, but we're still going to use it. Sometimes because it's better than nothing. Sometimes because no one gets fired for using <insert standard technology here>. And sometimes just because the politicians owe favors to their corporate handlers.

Lie detectors, Sea Wolf submarines, facial recognition software, etc etc etc. Many security technologies lie on the spectrum from "impractical" to out and out "doesn't work". But we just have to get comfortable with them, because they aren't going away.

That's a rather defeatist attitude to take.

One other side is that most applicants don't know that lie detectors are bunk, so their use can result in more honest answers.

The continued use of polygraph "tests", as if they have any validity whatsoever, is honestly a scandal.

Right. The big lie about it is that there's any possibility of the "test" proving a person's statements honest. That's what gets people to agree to them. In reality, once you step into that room, the best you can hope for is to come out without incriminating yourself, just like any other interrogation.

They have been in the news recently in relation to the current Supreme Court fiasco. Why are they they a thing? Who likes them?

Well, they don't have any particular accuracy, but they do allow the operator to generate whatever result they like in a usefully opaque manner.

The linked article discusses a large number of cases of racial bias. Obviously if you want to be biased in your hiring, but you want to hide it, a polygraph is very convenient way to manufacture some cover. (Alternatively of course, some of the police departments may have had no such intention, and merely been the victim of biased polygraph operators.)

As for the recent Supreme Court fiasco, incidents 35 years in the past rarely turn up much hard evidence. A polygraph is, again, a very useful way to manufacture something that looks like evidence. (And that's true regardless of the truth of Ford's claims. Just because polygraph results are fake doesn't mean they aren't accurate sometimes!)

In short:

> Who likes them?

The people who commission them, because they can get the results they want, and the people who operate them, because it's a pretty well paying job.


The lie behind the lie detector.

Ignore the cheesy web design, the PDF book is a fascinating read and goes into the detail of how polygraphs work. They're a bit more complicated than "oh your heart rate goes up", and yes they're complete BS.

Getting boxed is always about who gives the test, which also includes what and how questions are asked. That's why you would always prefer some old salty crusty bastard who can see through bullshit vs some young FNG who thinks he's saving the world one box at a time, which is where stories of fully qualified people getting dropped mostly come from.

In France in the 90's there was an obsession with analysing the handwriting of applicants, the same kind of bunk science. As far as I know it's gone now (I guess nobody applies with a hand-written letter).

Polygraphs are pseudo-scientific nonsense with no actual validity. Why would anybody use that as part of a job screening?

It gives employers a way to "scientifically" justify their hiring practices against discrimination lawsuits.

IANAL but what's to prevent the litigating party from immediately filing a Daubert motion against an employer who tries to say they used the polygraph to 'scientifically' disqualify a prospective job seeker-since you can't just hand a polygraph machine to any on-staff HR rep and go "No go see if Bob here really graduated from MIT like he says", and the wealth of research telling us how problematic results can be?

> IANAL but what's to prevent the litigating party from immediately filing a Daubert motion against an employer who tries to say they used the polygraph to 'scientifically' disqualify a prospective job seeker-since you can't just hand a polygraph machine to any on-staff HR rep and go "No go see if Bob here really graduated from MIT like he says", and the wealth of research telling us how problematic results can be?

Under what circumstances would a case like this even come to pass? From what I gathered from the article, these administrative hearings where applicants may plead their cases are not court proceedings but administrative hearings conducted by law enforcement officials.

Relief would likely come only under another course of action, like a civil rights case brought against the hearing committee (as mentioned in the article).

Unless you are referring to an action brought against the hearing committee on the basis that the test was improper because the person administering the test was untrained. At that point, I would imagine the court would not want to prescribe the hiring practices of local law enforcement and would defer to them (unless those practices could be shown to be discriminatory against some protected class).

Also, I didn't get the sense that these tests are administered by untrained HR reps, but I may have missed something in the article.

> IANAL but what's to prevent the litigating party from immediately filing

Not having the money to do so.

Heh, okay that's a fair and valid counterpoint, but entertain the hypothetical-I'm genuinely curious how that could play out if someone took a prospective employer to court following a polygraph that disqualifies them from further interviews.

Probably DQing yourself from future employment consideration.

Prospective employers don’t prefer litigious candidates. (seemingly regardless of merit).

My apologies, I was hoping from the context here that my "how would it play out" question would have implied "how would it play out in court" since the comment that originally got my curiosity pumping spoke directly to hypothetical discrimination lawsuits.

They still hold up in courts, and it gives an air of authority. A surprising number of people still believe in them as a valid tool.

I'm surprised a 1930's era pseudoscience is still hanging around professional law enforcement circles. The optimist in me hopes it's a clever ruse to screen people who simply aren't team fit. The pessimist tells me they're the caliber of people who also think calling in a psychic to help with murder cases.

Modern fMRI technologies can tell if people are fabricating stories. There's actual science behind them.

> Modern fMRI technologies can tell if people are fabricating stories. There's actual science behind them

I think there was some actual science behind the polygraph too, but just having actual science on poorly informed/motivated participants isn't really enough for tech that will have long term adversaries.

I can't recall if 1/3 or 2/3 of myth busters staff could beat the fMRI once prepared and motivated.

Initial experiments on disinterested subjects given no information about past experiments may have legitimately been about the same for both fMRI and polygraph (at the respective times when participants could have had no information)..

Apparently in 3 tries Mythbusters managed to fool the fMRI lie test once. They didn't experiment much with different techniques, I would fully expect success rates to get much worse than that as people get more time to find good techniques for beating the test.


It wouldn’t be hard to make the scan uninterpretable either, which is a seperate failure mode.

Boy, they sure didn't open with a case I could be sympathetic with. Lied to the state police, lied when applying to a city PD, but this time, oh, this time he's telling the truth!

My guess is, word gets around, and "inconsistencies" is just the excuse they need. I'm not saying it makes it right, because next it's going to mere coincidence that a black woman had "inconsistencies" when applying. But in this case, I might be willing to let it go.

> He had first applied to the Connecticut State Police and was failed for deception about occasional marijuana use as a minor. He then tried again with a police department in New Britain, where a polygraph test showed him lying about his criminal and sexual history.

> This time he had failed the New Haven polygraph for something cryptically called “inconsistencies.” “[But] I’m not hiding anything,” he said at the hearing. “I was being straight and honest and I’ve never been in trouble with the law. I’m not lying about anything.”

His argument seems to be that all of the polygraph tests were consistently wrong and that he didn't do any of those things. This is consistent with later comments by other people in the article:

> While undergoing a polygraph examination for a position at an FBI field office in New Haven in 2010, a black man was told that his recollection of using marijuana only a few times in high school was showing as deceptive, and that he should change his answer. Later, he wrote: “I was convinced that [the examiner] may have made an assumption, based on a stereotype about African Americans and drug use, and used that stereotype to profile me. I also realized that what [he] was asking of me would reflect negatively either way—if I didn’t change my answer I was being deceptive, and if I did change my answer I was lying on my application.”

Yeah, I read it several times, and it still came out to me as addressing only the final application. But now that you point it out, it seems more likely to address all applications.

Regardless, that's about as far as I got because my two personal experiences with polygraphs tells me they're about on the same level as dowsing rods. "Have you ever used marijuana?"

"No", he said, higher than a kite having smoked a bowl an hour before the test. It was asked both times, passed both times.

what kind of job hooks you up to a polygraph to ask if you've used marijuana and doesn't bother to give you a gc/ms screen?

The kind where 30 years ago peeing in a cup wasn’t nearly the industry it is now, and therefore not as affordable or available at all. That’s what kind. But that also kinnnda wasn’t the point of the story.

The kind where the management doesn't actually care if you toke up privately, but wants a good excuse to diqualify anyone they don't like for whatever other reasons.

It's tricky because I interpreted the paragraph differently: the _polygraph_ has deemed him of lying about these things, not that he -actually- lied:

"He had first applied to the Connecticut State Police and was failed for deception about occasional marijuana use as a minor. He then tried again with a police department in New Britain, where a polygraph test showed him lying about his criminal and sexual history."

> the _polygraph_ has deemed him of lying about these things

I think the article's point is that the polygraph _operator_ decided that (perhaps not consciously), not the machine itself.

> Boy, they sure didn't open with a case I could be sympathetic with. Lied to the state police, lied when applying to a city PD, but this time, oh, this time he's telling the truth!

You misunderstood the article. There's no evidence to suggest he lied ever.

Across lie detectors, forensic testing, and every similar discipline there should be frequent blind auditing by a neutral third party, and the results should be made available to the public.

I don't see how that would change things. The TSA got 97% failure rate in independent testing, did you see any change at the airport?

Make mandatory annual operator tests, all failed ones get fired immediately. Throw in some safeguards against secret handshakes between operators.

Obligatory link to http://antipolygraph.org, which provides detailed research on why the polygraph is ineffective, details on how to defeat the polygraph, manuals used to train polygraph investigators, and many other juicy tidbits.

I wonder what the correlation is between high neuroticism and failing a polygraph.

Might as well use divining rods.

And here I thought having to do FizzBuzz for two different interviewers was onerous.

Onerous? If you think 6-20 lines of code is onerous, you really should reconsider your career choice.

Polygraphs can produce errors that may be used as a wrongful justification for an employer to dismiss your application.

...on the other hand

A fizz buzz application that produces errors _should_ be used as a justification for a potential employer to dismiss you entirely.

It was a joke. Except for the part where I had to do it twice.

I hope this is a joke

The part where I had to do it twice for two different people isn't.

I think that's pretty obvious.

I've had people get visibly angry with me that I'd dare suggest that they do fizzbuzz.

I think every time somebody starts with "And here I thought...." it's something humorous.

Because it is an incredibly stupid test.

That senior devs with decades of experience fail.

I've literally seen someone who supposedly wrote firmware for the space shuttle fail fizzbuzz.

If the candidate was verifiably a NASA engineer the person half assed it probably just didn't care. I'm surprised that person just didnt respond with a very intricate set of functions that resulted in a middle finger. That's honestly insulting to use that as a test for a senior. Clearly there was no vetting process prior to a coding evaluation and the hiring process needs a lot of work.

They worked for one of NASA's contractors, not NASA itself.

Have you been involved in the hiring process? Most of the people actively looking for work don't have a job for a reason. It's not 'insulting' to see if you can even vaguely do the job in an interview.

There's only so much vetting you can do before the 'can you vaguely code your way out of a wet paper bag' step.

Also: you didnt used to work in Boulder recently, did you?

Yes I have been involved in that process at multiple companies. If they were in for an actual onsite interview the only thing I cared about was how they effectively they communicate and their thought process to a problem. Vetting isn't hard, the phone screen should have already done that well before a single line of code is written.

No I have never worked in Boulder.

You literally can't tell until they write code. Having done hundreds of interviews, that's the clearest take away I have.

The candidates you're trying to avoid are really good at passing a phone screen since they have a lot of experience at it, constantly looking for jobs.

You literally can tell by asking a couple questions and letting them run on. You can smell bullshit over the phone pretty quickly.

You sometimes have to be careful and look if people are overthinking the problem. I have been in interviews where I was desperately looking for an efficient solution for a problem and couldn't find one while the interviewer only wanted to get a simple for loop.

Having sat in that interview he wasn't overthinking it. Despite writing firmware for decades he had never heard of the modulus operator, and when we explained it to him, he didn't see the point of it.

Remember those people in college who contributed nothing to the group project, but still somehow squeked by with a C? Those people now have jobs, and some have impressive resumes, particularly from .gov contractors where less than mediocrity reigns unfettered.

I know the modulus operator but I don't recall ever using it in more than 20 years :)

Are you a firmware dev?

Nope. Do firmware devs need modulo? I thought they do everything with XOR.

Yes, it comes up a surprising amount.

Applications are open for YC Winter 2020

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