Why Lie Detector Tests Can’t Be Trusted • Hacker News Board

Re: Why Lie Detector Tests Can’t Be Trusted

#61

Post by Smithalicious » Mon, Jul 29, 2019, 6:36 PM UTC

nocsi wrote: post #38

Polygraphs aren’t really used for lie detection. Think of it as a tool for an enhanced interview. It allows for the interviewer to have many more dimensions of straws to grasp at in assessing you. “Lies” are assessed as anomalous readings measured against control questions. This is something polygraphers are open about and inform you of the control stage during the start. Afterwards, it’s all really an internal battl…

It's the classic "I already know the truth and lying will only make your punishment worse" trick that most moms seem to pull at least sometimes

Re: Why Lie Detector Tests Can’t Be Trusted

#62

Post by throwawy123 » Mon, Jul 29, 2019, 6:36 PM UTC

JackFr wrote: post #58

I passed my NSA polygraph, though it was uncomfortable. However my psych interviewer was able to get disqualifying info out of me using just technique. It was only years later while reading a book about the Shin Bet (Israeli military intelligence), I recognized the way I was manipulated. No hard feelings, after all I was telling the truth, and I probably would have hated the job anyway.

What sort of disqualifying info, if you don't mind my asking?

There were a lot of interesting questions in the psych interview, from porn use to romantic relationships and digital pirating...

And, just spitballing, I wonder if it's a crime to lie in the psych interview. I know it is during the polygraph, since the examiners are federal investigators.

Re: Why Lie Detector Tests Can’t Be Trusted

#63

Post by taftster » Mon, Jul 29, 2019, 6:40 PM UTC

the-pigeon wrote: post #50

taftster wrote: post #23

Earlier quoted context omitted.

Yes, exactly. Polygraphs are just a form of interrogation technique, disguised as "science". If the science was debunked, then the credibility of the machines would be questioned and the interrogation process would not yield as positive (if you can call it that) results. So the whole mystique of polygraphs must be guarded to keep up the facade of the lie detector's true capabilities.

I mean they have been debunked and anyone that Google's it knows it. Doesn't mean they don't work as an intimidation tactic. Though most of the agencies that use them are probably filtering out a lot of good employees by putting them through a test designed to mess with them. Instead of focusing on whether the person has the skills needed for the job.

It's got to be pretty intimidating. Getting locked up in a room with the investigator, being told that you're not actually required to be there, but it will seriously jeopardize your chances of employment if you decide to leave or decline, etc. I'm sure it works well as an intimidation tactic, as you say.

People need jobs and are willing to put up with lots. Going through with a polygraph is likely just one of a number of the things that people in these industries have to put up with.

Re: Why Lie Detector Tests Can’t Be Trusted

#64

Post by nxpnsv » Mon, Jul 29, 2019, 6:41 PM UTC

jpmattia wrote: post #46

I have no idea why there is not an accurate calibration of polygraphs published. Here's a straightforward calibration: Have the subject write down a random number from 1 to 1024. Perhaps this random number is assigned via a phone app. Have the subject put the paper in his pocket. An maximally accurate polygraph will require no more than 10 questions to determine the number in the subject's pocket. [However, I think w…

Trying to keep a random number secret likely is less stressful than an actual lie. Even if lie detectors worked this wouldn’t be a great approach. My way is better, scrap them for parts and buy beer for whatever is left over.

Re: Why Lie Detector Tests Can’t Be Trusted

#65

Post by wycy » Mon, Jul 29, 2019, 6:41 PM UTC

JackFr wrote: post #58

I passed my NSA polygraph, though it was uncomfortable. However my psych interviewer was able to get disqualifying info out of me using just technique. It was only years later while reading a book about the Shin Bet (Israeli military intelligence), I recognized the way I was manipulated. No hard feelings, after all I was telling the truth, and I probably would have hated the job anyway.

Can you elaborate on the way you were manipulated?

Re: Why Lie Detector Tests Can’t Be Trusted

#66

Post by RHSeeger » Mon, Jul 29, 2019, 6:42 PM UTC

khawkins wrote: post #34

There's a lot of hyperbolic claims being made in the article and these comments which seem unsupported by science. The answer to the question, "can polygraphs be utilized to draw out truth?", is a resounding yes: >Notwithstanding the limitations of the quality of the empirical research and the limited ability to generalize to real-world settings, we conclude that in populations of examinees such as those represented…

What is the use case for a test that can indicate that something is probably the truth or probably a lie when focused on specific incidents? _(According to the study, the reliability drops when used for screening)_

Re: Why Lie Detector Tests Can’t Be Trusted

#67

Post by teddyh » Mon, Jul 29, 2019, 6:43 PM UTC

throwawy123 wrote: post #36

I took a polygraph twice for an NSA internship (TS/CSI Clearance). It's an interesting process. In the first phase, you go over your security forms (SF-86 and related) in detail with your polygraph examiner. They ask a lot of detailed questions, sometimes including things you aren't legally required to disclose (drug use and foreign travel outside the listed time limits). In the second phase, they hook you up to the…

> the examiner was convinced I used more drugs than I let on

No they weren’t. They merely pretended they were convinced. It’s all part of the act to make you nervous.

Re: Why Lie Detector Tests Can’t Be Trusted

#68

Post by jpmattia » Mon, Jul 29, 2019, 6:45 PM UTC

nxpnsv wrote: post #64

jpmattia wrote: post #46

I have no idea why there is not an accurate calibration of polygraphs published. Here's a straightforward calibration: Have the subject write down a random number from 1 to 1024. Perhaps this random number is assigned via a phone app. Have the subject put the paper in his pocket. An maximally accurate polygraph will require no more than 10 questions to determine the number in the subject's pocket. [However, I think w…

Trying to keep a random number secret likely is less stressful than an actual lie. Even if lie detectors worked this wouldn’t be a great approach. My way is better, scrap them for parts and buy beer for whatever is left over.

> Trying to keep a random number secret likely is less stressful than an actual lie.

Which begs the issue: If a lie detector is accurate only for answers that are stressful, then it sure looks like we need a detector to figure out if an answer is stressful or not.

Or to put it differently, now you have two problems: Lie detection and stressful-answer detection.

Nevertheless, I concede that your beer detection plan is superior.

Re: Why Lie Detector Tests Can’t Be Trusted

#69

Post by hirundo » Mon, Jul 29, 2019, 6:45 PM UTC

pixelbath wrote: post #22

glitchc wrote: post #2

Lie detector tests rely on psychological intimidation. The testee invariably gives up compromising information over the course of the test, allowing the tester to apply leverage or outright coercion. If no useful information is divulged, the test is marked a fail and requires a repeat.

In my experience, that's not how the session goes. All the questions are reviewed in advance with the interviewee; there are no surprise questions. The interviewer I worked with even mentioned which questions he wanted and expected lies to. Many of these were somewhat personal questions ("Do you live at X; Is your mother's name Y?"), but nothing overly invasive. Granted, I did not participate in any polygraph session…

I had the "intimidation" kind of polygraph session when I was questioned by a couple of police detectives about a theft in a building where I was a security guard. Intimidation and threats. Like if you don't answer right we're headed to your place to search everything. They did not review the questions in advance; they did try to surprise me; there was no attempt at calibration. The lie detector seemed to be a prop. After the test one of them came to me and said "you passed" while looking at me closely like my reaction was just the last test. I laughed. I guess it was the right answer.

Re: Why Lie Detector Tests Can’t Be Trusted

#70

Post by alexhutcheson » Mon, Jul 29, 2019, 6:47 PM UTC

RankingMember wrote: post #27

rndgermandude wrote: post #19

Earlier quoted context omitted.

If you're concerned about this, you really shouldn't read up on the actual "forensics" the FBI did in the past, like "hair analysis", because it will ruin your day: https://www.fbi.gov/news/pressrel/press-releases/fbi-testimo... Or "bite mark analysis" (don't google that, it will ruin your day as well).

Is there a proper response to this other than just banging your head against a desktop until you're blissfully unconscious? I can't think of any other response that would properly articulate my response to skimming that article.

https://www.innocenceproject.org/getinvolved/