Polygraphs have major flaws. Are there better options?

Research proceeds on alternatives, but some doubt whether true lie detection is possible.

See full article...

 

[RZetopan]

I don't think the polygraph in theory is bad, but the testing is bad

When they do a polygraph test, thy first measure your baseline responses by asking you simple questions like stating your name by the test giver and a simple yes or no from you

The rest of the test is to measure the variations in responses from your baseline.

The problem comes from getting the baseline correct. If you can throw off the baseline, then everything will appear true This can be done by inflicting physical pain on yourself or even psychological pain. Conversely if your baseline is too mild, any response to any question will generate a false positive.

That reminds me of the dowser who told me that he had an 85% accuracy rate. But then he went on to “explain” that he didn't count the times that he failed, since that just meant that he wasn't trying hard enough. One can make an endless number of excuses for failure, while quite carefully avoiding the real reason for failing.

 

[nosh]

You're trying to find a reason for it not to work, which is a different process than trying to find a way to get it to work. I'm sure that with enough data someone will get it to work.

That assumes that lie and truth are actually different enough from a neurological view that it could be done. That assumes a very naive idea about what remembering is.

Every act of remembering something is an act of recreation. Telling a story often enough (even with no intent of embellishing it) will shift it (a big problem for criminal investigations). Without enough information our brain with confabulate the missing pieces (that usually only becomes apparent with hypnosis or Alzheimer's disease). So any difference in this regard between true and lie is a gradual one.

The biggest difference between them is intend. But you can equally well tell a truth with ill intent as a lie, so that does not help.

 

[Idiotzoo]

There are already much better and considerably more reliable options such as goat entrails or picking through some dust touched by individual, looking for patterns to confirm their guilt.

 

[dzid]

That assumes that lie and truth are actually different enough from a neurological view that it could be done. That assumes a very naive idea about what remembering is.

Every act of remembering something is an act of recreation. Telling a story often enough (even with no intent of embellishing it) will shift it (a big problem for criminal investigations). Without enough information our brain with confabulate the missing pieces (that usually only becomes apparent with hypnosis or Alzheimer's disease). So any difference in this regard between true and lie is a gradual one.

The biggest difference between them is intend. But you can equally well tell a truth with ill intent as a lie, so that does not help.

Yep. That's why I've been writing down significant events that happen in my life for years (not really something I do often, but occasionally I do because I want to remember them as close to how they occurred as possible).

 

[Madestjohn]

That assumes that lie and truth are actually different enough from a neurological view that it could be done. That assumes a very naive idea about what remembering is.

Every act of remembering something is an act of recreation. Telling a story often enough (even with no intent of embellishing it) will shift it (a big problem for criminal investigations). Without enough information our brain with confabulate the missing pieces (that usually only becomes apparent with hypnosis or Alzheimer's disease). So any difference in this regard between true and lie is a gradual one.

The biggest difference between them is intend. But you can equally well tell a truth with ill intent as a lie, so that does not help.

Also (if you will allow the trivial diversion) we might be barking up the wrong tree looking primarily at neurological activity
.. or at least exclusively the neurons

With more and more information coming out recently about how important glial cells, astrocytes and microglia, are in attention, learning, and memory (both formation and retrieval).
With indications recently that they might be more in control of the neurons than the other way around.

We like to simplify classify and model systems to better understand them, but we shouldn’t mistake those simplifications for what they are attempting to model

 

[bregalad]

Back in the 1990s I assisted a polygraph examiner by measuring the variations on the chart paper and recording those values. At no time did I know the questions that were asked or the identities of the people he tested. Discretion was of utmost importance so all interviews took place evenings and weekends when there were no employees in the building.

This was a private business for clients involved in out-of-court negotiations or civil litigation who wanted a highly respected investigator on their side.

I understand that he asked questions in a flat unemotional manner and clients were limited to answering "yes" or "no". I was surprised by how short the charts were. I seem to recall that most polygraph sessions consisted of 12 questions.

I was told that the key to good data is focusing on a single event or accusation. An experienced examiner knows how to ask good control questions and phrase everything just right. Analysis involves playing back the recordings and matching the charts to the exact moment a particular word was spoken or a particular eye movement was observed.

In addition to being one of the country's most experienced polygraph examiners, the man I worked for was an expert at reading people and spotting inconsistencies. Like a great poker player he could quickly identify someone's "tell" if they had one. It was probably easier to get an inconclusive result from the polygraph machine than it was to fool the man operating it.

 

[dzid]

Also (if you will allow the trivial diversion) we might be barking up the wrong tree looking primarily at neurological activity
.. or at least exclusively the neurons

With more and more information coming out recently about how important glial cells, astrocytes and microglia, are in attention, learning, and memory (both formation and retrieval).
With indications recently that they might be more in control of the neurons than the other way around.

We like to simplify classify and model systems to better understand them, but we shouldn’t mistake those simplifications for what they are attempting to model

Every time I try to get out, you keep pulling me back in... (something like that).

I think the diversion is apropos, and that's the impression I had after reading this recent paper on ibogaine and the development of derivative / structurally related drugs: dismayed because we're not as far along that road as I had thought, and still hopeful because I believe it is the right direction to continue pressing ahead with the science.

Another way to frame it more in the way I do when thinking about addictions, including OG narcotics and the potential for something similar to be formally defined in relation to social media and the big tech business model, which becomes more plausible by the day: The confluence of a number of important mechanisms in the brain being out of equilibrium is gaining traction as more research is conducted.

There is only one known drug that acts on all of the implicated systems (and that's yet to be agreed among the leading researchers in the field), has real potential in treating PTSD, addictions including alcoholism, opioids, cocaine and maybe others such as cognitive issues related to chronic stress, but is fiendishly complex in everything from functional design of organic molecules, synthetic or semisynthetic production methods, identification of which therapeutic goals are satisfied by a given molecule, and development of a proper testing framework that correctly identifies thresholds for neuronal damage, etc.

When I consider all of that, and then consider the state of our society's justice and health care systems along with my (admittedly anecdotal but extensive) personal history related to addiction and its effect on so many people close to me, and now the available research on (potential) internet addiction:

Would I risk my own children's learning potential? Absolutely not. I don't have children of my own, but feel strongly that I do not want other children to be put at risk either.

 

[SixDegrees]

Back in the 1990s I assisted a polygraph examiner by measuring the variations on the chart paper and recording those values. At no time did I know the questions that were asked or the identities of the people he tested. Discretion was of utmost importance so all interviews took place evenings and weekends when there were no employees in the building.

This was a private business for clients involved in out-of-court negotiations or civil litigation who wanted a highly respected investigator on their side.

I understand that he asked questions in a flat unemotional manner and clients were limited to answering "yes" or "no". I was surprised by how short the charts were. I seem to recall that most polygraph sessions consisted of 12 questions.

I was told that the key to good data is focusing on a single event or accusation. An experienced examiner knows how to ask good control questions and phrase everything just right. Analysis involves playing back the recordings and matching the charts to the exact moment a particular word was spoken or a particular eye movement was observed.

In addition to being one of the country's most experienced polygraph examiners, the man I worked for was an expert at reading people and spotting inconsistencies. Like a great poker player he could quickly identify someone's "tell" if they had one. It was probably easier to get an inconclusive result from the polygraph machine than it was to fool the man operating it.

A "good" polygraph examiner is one who can intimidate the subject into a confession, full stop. There's nothing even remotely valid about the machine itself. That's just a security theater prop.

Consider that with modern technology there is no need at all for the examiner to be present in the first place - if the machines actually worked. Video and physiological data could simply be recorded in response to pre-recorded or even synthesized questions. The reason this isn't done is that the examiner is needed to apply pressure to the subject and subtly harass them into admissions, in loads of cases even false admissions.

A "good" polygraph examiner is an expert liar.

 

[AverageDutchGuy]

It should be noted that virtually all of the top positions in our current government are being held by people who's security clearances should've been contingent upon polygraph examination and yet no one here would believe that a single one of them would be able to pass one (be they flawed or not).

For instance, if Trump was ever asked "have you ever stolen anything?"
Or if Hegseth was asked "have you ever been intoxicated while working for an employer?"

Trump would "truthfully" state he's never stolen anything. Anything he's ever appropriated rightfully belonged to him. There's no such thing as stealing if your a malignant narcissist. Can't detect someone is lying if they fully an utterly believe they're telling the truth. Which is one of the problems with "ground truth" these polygraph/liedetector people love talking about. Who's ground truth, exactly?

 

[DNA_Doc]

I knew a person who was a polygraph "expert" for a three-letter agency. In the longest conversation I had ever had with him (over beers during a social gathering), he essentially admitted that technological accuracy has nothing to do with why polygraphs are used.

"Lie detectors" work to the degree that people think they work. As others have mentioned, they are props designed to intimidate and coerce. Their presumed neutrality, the presumed accuracy, the physicality (moving needles, flashing lights) all add weight to the investigators' attempts to manipulate one into making a damaging admission.

For interrogators (and one should never forget - if you are being subjected to a polygraph, you are being interrogated), it's just one more tool in the toolkit.

 

[Travis Butler]

Let’s suppose a perfect lie detector could exist, and we made them? Now what? Would they, globally, serve justice or oppression more?

Science fiction has grappled with this, and the perfect lie detector is ALWAYS bad.

Not quite always; H. Beam Piper had a very favorable view of his fancy lie detector (“polyencephalographic veridicators”):

He had, at that, after he’d decided he couldn’t beat the veridicator. Jack found himself sympathizing with Mallin. He’d disliked the man from the first, but he looked different now—sort of cleaned and washed out inside. Maybe everybody ought to be veridicated, now and then, to teach them that honesty begins with honesty to self. [From Little Fuzzy.]

TBF, the veridicator was presented as measuring brain waves instead of physiological responses:

As soon as the veridicator was on, he looked up at the big screen behind the three judges; the globe above his head was a glaring red. There was a titter of laughter. Nobody in the Courtroom knew better than he what was happening. He had screens in his laboratory that broke it all down into individual patterns—the steady pulsing waves from the cortex, the alpha and beta waves; beta-aleph and beta-beth and beta-gimel and beta-daleth. The thalamic waves. He thought of all of them, and of the electromagnetic events which accompanied brain activity. As he did, the red faded and the globe became blue. He was no longer suppressing statements and substituting other statements he knew to be false. If he could keep it that way. But, sooner or later, he knew, he wouldn’t be able to. [From shortly before the prior quote, Mallin’s thoughts at the start of his testimony.]

Of course, all of this was written in the early 60s, and has about as much scientific reality as psi powers; but it sure sounds realistic. (The harmful part is when people read something like this and start thinking it exists in real life.) And also to be fair to Piper, he did recognize that it only detects if the subject believes something is true, not if something actually is true.

---

On the flip side, Harry Harrison was talking about how to beat polygraphs in The Stainless Steel Rat’s Revenge back in 1970 - think anxious thoughts as they ask the neutral questions, then think relaxing ones when they hit the key question.

 

[Dark Jaguar]

Yep. That's why I've been writing down significant events that happen in my life for years (not really something I do often, but occasionally I do because I want to remember them as close to how they occurred as possible).

I've never kept a diary or a journal like that, but maybe I should start. My "mind palace" is calibrated by linking release dates of movies and video games to the events of my life. I've found that at the very least, I can accurately order these events that way. This itself caused a bit of contention with a family member who, for reasons unknown to me, has formed a rather negative view of another relative having scared her when she was just a little kid. But, in relaying the events and saying WHERE it happened, I knew something didn't add up. She'd have been 16 when that halloween prank happened, hardly a "little kid". She swore up and down it happened many years before that, but I knew that in that very household, I had been playing a certain game at the time, looked up the date, and confirmed it was the turn of the millennium fight 2000. I had to confirm, once again, WHERE it took place, and with that in mind, yeah, the memory didn't line up with the math. Suffice to say I got the silent treatment for a bit after that. Hey I wasn't even saying it was wrong to feel a certain way about it, just that the age it happened couldn't have been what was remembered. The facts didn't line up.

Anyway, that's all a detective needs to do. Fit the facts to a story. Whether someone BELIEVES those facts is irrelevant, so I'm really not sure what the point of a lie detector, even one that WORKED, would be. A case only needs to establish facts from fiction, and how reliable a witness is. Whether they were intentionally deceiving or just misremembering is beside the point.

 

[MjchX11R6]

Polygraphs are actually complete bullshit. The man giving the test influences it more than anything else.....

 

[Madestjohn]

Not quite always; H. Beam Piper had a very favorable view of his fancy lie detector (“polyencephalographic veridicators”):



TBF, the veridicator was presented as measuring brain waves instead of physiological responses:



Of course, all of this was written in the early 60s, and has about as much scientific reality as psi powers; but it sure sounds realistic. (The harmful part is when people read something like this and start thinking it exists in real life.) And also to be fair to Piper, he did recognize that it only detects if the subject believes something is true, not if something actually is true.

---

On the flip side, Harry Harrison was talking about how to beat polygraphs in The Stainless Steel Rat’s Revenge back in 1970 - think anxious thoughts as they ask the neutral questions, then think relaxing ones when they hit the key question.

I always like the idea of the tack in the shoe method.. just remember to press down when your not lying

But of course ‘lie detector ltd’ says this is impossible


Many people have heard of things such as putting a tack in their shoes and pressing down on it. Such actions simply do not work as they will be instantly detected.
If someone is engaging in countermeasures during their lie detection testing, they will quickly be asked by the examiner to stop. If they continue, they will end up with a report at the end of the process that states the person was purposefully not co-operating. Innocent people do not engage in this type of behavior and conclusions of dishonesty will only be reinforced.
A properly trained and experienced examiner will observe or detect any such attempts at obstruction.


Which seems like the exact type of statement made with absolute certainty and over stated confidence that a inexperienced liar might attempt used to mask a lie
And that a lie detector could actually detect

​

 

[Dark Jaguar]

Polygraphs are actually complete bullshit. The man giving the test influences it more than anything else.....

Polygraphs are useful in one major way: The public believe they work. This means the police can tell a suspect that they failed the test, and then hopefully get a confession out of them.

Never mind that people are known to confess to things they didn't do if you put them under enough stress.

 

[dzid]

I've never kept a diary or a journal like that, but maybe I should start. My "mind palace" is calibrated by linking release dates of movies and video games to the events of my life. I've found that at the very least, I can accurately order these events that way. This itself caused a bit of contention with a family member who, for reasons unknown to me, has formed a rather negative view of another relative having scared her when she was just a little kid. But, in relaying the events and saying WHERE it happened, I knew something didn't add up. She'd have been 16 when that halloween prank happened, hardly a "little kid". She swore up and down it happened many years before that, but I knew that in that very household, I had been playing a certain game at the time, looked up the date, and confirmed it was the turn of the millennium fight 2000. I had to confirm, once again, WHERE it took place, and with that in mind, yeah, the memory didn't line up with the math. Suffice to say I got the silent treatment for a bit after that. Hey I wasn't even saying it was wrong to feel a certain way about it, just that the age it happened couldn't have been what was remembered. The facts didn't line up.

Anyway, that's all a detective needs to do. Fit the facts to a story. Whether someone BELIEVES those facts is irrelevant, so I'm really not sure what the point of a lie detector, even one that WORKED, would be. A case only needs to establish facts from fiction, and how reliable a witness is. Whether they were intentionally deceiving or just misremembering is beside the point.

There's a research paper I posted over in Observatory (Misc) a while back on the subject of "event segmentation" and the granularity of those events in terms of how the brain stores and processes the continuous stream of sensory input in the course of our day-to-day existence.

A really interesting part of that paper discusses how the degree to which a person moves among the societal hierarchy (rather than "spacial" (place to place) tends to lead to more finely grained event segmentation.

 

[Dark Jaguar]

There's a research paper I posted over in Observatory (Misc) a while back on the subject of "event segmentation" and the granularity of those events in terms of how the brain stores and processes the continuous stream of sensory input in the course of our day-to-day existence.

A really interesting part of that paper discusses how the degree to which a person moves among the societal hierarchy (rather than "spacial" (place to place) tends to lead to more finely grained event segmentation.

Funnily enough, I'm in many ways still the shut-in, while my sister's far more socially active.

 

[dzid]

Funnily enough, I'm in many ways still the shut-in, while my sister's far more socially active.

That's cool. That description probably would have applied to me and my partner as well. I experienced a lot more social interaction with her than I likely would have otherwise.

 

[djdaedalus]

Trump has proved that you don't need to have people lie, you just have to have people who don't know what the truth is anyway, and will say whatever you want them to say.

 

[randomuser42]

At least when it comes to security clearance screening the alternative is to research your background and interview your references and family. And that's what they do. I feel like the polygraph is just a way for them to have one final screening step they can manipulate to basically filter out anyone they "don't like" for reasons that wouldn't pass HR muster (so, illegal). It's only done for some Top Secret clearances, not all and not Secret. If it were effective they'd just do it for all of them. But so many people need Secret clearances that it's not worth the trouble to deal with someone having unilateral veto power.

 

[dzid]

Trump has proved that you don't need to have people lie, you just have to have people who don't know what the truth is anyway, and will say whatever you want them to say.

IMO, they're likely paid well and/or encouraged to abandon any notion of loyalty and their own self-respect because they cannot live with the idea that something they've done and are ashamed of might come to light.

If that is the case and the issue is that they cannot live with such a thing: then don't! Easy peasy, and we have one less problem to deal with.

 

[Carrie]

I was polygraphed twice as part of job applications. As a young woman in my twenties, it was a shitty experience.
I got the job both times, though.

 

[Dark Jaguar]

I was polygraphed twice as part of job applications. As a young woman in my twenties, it was a shitty experience.
I got the job both times, though.

I can safely say through the many many job applications I've been through in the past year and a half or so, not once was a polygraph a part of the process. Frankly, it should be illegal for ANY company or government agency to REQUIRE a polygraph for any purpose whatsoever. They should be left as curiosities someone can privately experiment with with friends, like seances and phrenology. I'm so sorry you've been put through such an unnecessary thing.

Oh, while I have heard many times, and believed, that polygraphs are not admissible as court evidence... I want to double-check that with people who know the legal system around here.

 

[mmiller7]

The only thing I would glean from a person who flawlessly passed a polygraph would be to keep an eye on them for their moderately increased chances of being psychopathic

edit: mmiller7, Garak is so appropriate lol

Couldn't find a meme in the first few seconds of search but I also recall something about "never tell the truth when a lie will do" from that episode...I think it was when Bashir went to Cardassia looking for a fix for the implant

 

[kimjc01]

Honestly, I would rather just say "give me the scopolamine", I have nothing to hide and hope they don't ask too many private questions, lol.

 

[Alexstarfire]

Is there anything better? Yea, flipping a coin.

 

[Feanaaro]

The problem is not in the science, it's in the underlying epistemological assumptions. In order for a polygraph, or any analogue system for lie detection, to work, one has to postulate some sort of correspondence between intentional meaning and some sort of physical, material, reality. There is no indication that such would be the case, which is to say that while conceptually we understand what a lie is, and all lies falls under the concept, each lie as it is uttered is a distinct empirical event which can correspond to any number of underlying physical realities.

 

[Madestjohn]

I can safely say through the many many job applications I've been through in the past year and a half or so, not once was a polygraph a part of the process. Frankly, it should be illegal for ANY company or government agency to REQUIRE a polygraph for any purpose whatsoever. They should be left as curiosities someone can privately experiment with with friends, like seances and phrenology.

Oh, while I have heard many times, and believed, that polygraphs are not admissible as court evidence... I want to double-check that with people who know the legal system around here.

There are unfortunately plenty of ‘ forensic science ‘ that have been thoroughly discredited (some that were never credible in the first place like bite marks, hair comparisons, and even the ass cheek impressions left behind on dusty car seat which in one case was matched to a specific pair of jeans and accused murderer ) that still have been used in court to sentence people

And that ignores all the blatant and routine fraud by ‘experts’ who are employed by law enforcement agencies to facilitate convictions

 

[Dark Jaguar]

Honestly, I would rather just say "give me the scopolamine", I have nothing to hide and hope they don't ask too many private questions, lol.

I unfortunately DO have things to hide that could cost me a lot of jobs. Nothing critical to job performance, no hidden murders or anything either, but I do live in a very red area and there are things about me potential bosses don't need to know. I'm not in an area where I have the luxury of picking my boss' politics.

 

[smallscreensw]

But critically, they only accurately judge truth-tellers around 57 percent of the time.

Put another way: would you step on a commercial jetliner if you knew it had a 43 percent chance of fatally crashing? That is the level of scrutiny these devices should be put to since, indeed, they can also destroy lives.

 

[Danellicus]

When discussing the efficacy of polygraph usage in the USA please remember that the spies Aldrich Ames, Robert Hanssen and Ana Montes all passed multiple polygraph exams over many years. Each was caught only through standard counterintelligence methods.

But you can thank Ana Montes for the configuration that now has you sitting on a twitch detector.

 

[smallscreensw]

Some years ago I read about a case where a couple of cops administered a ‘lie detector’ test. They sat the guy down next to a big machine (a copier) connected him to it with a helmet and wires (all fake) and asked him questions. Each time he answered the machine spit out a photocopy of a sheet that said “You’re Lying”. Finally the guy broke down and confessed. As I remember the confession was thrown out because even the court thought that was going too far. Official ‘lie detectors’ aren’t any better.

That's funny and tragic all at the same time. It sounds like something out of a "Three Stooges" feature.

 

[Danellicus]

[...] I fell asleep during one (didn't pass that time, for some reason).

I was in the chair for 2.5 hours as the examiner asked a plethora of boring questions. When I jerked awake - you know that thing when you have a second or two of microsleep and then you 'jerk' back to full consciousness? - the pens must have jumped. The examiner got really excited and went back to re-ask a few of the questions. But I went back to normal the pens went back to whatever was normal for them.

 

[Madestjohn]

That's funny and tragic all at the same time. It sounds like something out of a "Three Stooges" feature.

Or an episode of the wire (one co-written by ex-police officer)

 

[Derecho Imminent]

Seems like human nature to get anxious at a question like "would you leak national secrets". Even an innocent person would feel uncomfortable being asked sensitive questions.

 

[Waco]

It should be noted that virtually all of the top positions in our current government are being held by people who's security clearances should've been contingent upon polygraph examination and yet no one here would believe that a single one of them would be able to pass one (be they flawed or not).

For instance, if Trump was ever asked "have you ever stolen anything?"
Or if Hegseth was asked "have you ever been intoxicated while working for an employer?"

Polygraphs are not part of the clearance-granting process.

Not that I think any of them should have a clearance (or their jobs) but

 

[J Carr]

Polygraph tests are complete and unadulterated BS.

Never agree to take one.

If you are compelled to take one (to get a national security job, for example), tell the polygrapher that the blood pressure cuff is too tight and uncomfortable. Ask that it be fitted to your calf. You can thank me later.

 

[J Carr]

Polygraphs are not part of the clearance-granting process.

Not that I think any of them should have a clearance (or their jobs) but

Yes they are required for TS/SCI clearances.

The President is not required to take one because the President is elected by the people, not hired for the job. Furthermore, the President is the head of national security and is the final authority on security procedures, such as polygraphs.

 

[TekaroBB]

Polygraphs are not part of the clearance-granting process.

Not that I think any of them should have a clearance (or their jobs) but

It's being phased out, but up until recently it very much was assumed anyone with clearance was either grandfathered or had taken a poly.

 

[dzid]

Yes they are required for TS/SCI clearances.

The President is not required to take one because the President is elected by the people, not hired for the job. Furthermore, the President is the head of national security and is the final authority on security procedures, such as polygraphs.

What was it that Vance's economic(?) advisor was doing? Passing tips on how to take drugs on an airplane flight was one, and a big spender on research chemicals that I'd be like "thanks man, but that's going to need to go to an analytical lab with real credentials."