Polygraphs: The evidence

[qwert]

We can't get around the scare factor, when we test polygraphs. How do we get that in a lab?

I don't quite understand what you are saying. I don't care what the mechanism is, whether it's fear, intimidation, placebo effects, whatever. I'm asking whether you believe the process of polygraphy produces result better than chance, regardless of the reason. I don't really care about the why.

We find out whether lumps are cancers or not, because we can do biopsies to check. What we can't use polygraphs for, is to detect lies, unless we know what are lies and what are not.

But in the lab, we CAN know whether the subject is lying or not.

I do agree that lab tests are of limited use in real life due to the fact that:
(a) in real life the subject typically has a vested interest in "getting away" with the lie
(b) in real life polygraphs are typically used in "high-stress" situations.

But this doesn't invalidate the fact that the polygraph process apparently works better than chance at detecting deception in the lab.

 

[NoZed Avenger]

Okie doke.

Please specify the degree of disagreement it takes.

The line-drawing fallacy has been explained, linked to, clarified, summarized, illustrated, annotated, construed, elucidated, pointed out, rendered, set out, set forth, spelled out, and shown to you more than once, and yet you cannot help yourself from engaging in it once more.

Apart from equivocation, it seems to be your favorite fallacy -- as it is here in this "argument."

 

[Lucky]

The polygraph is based on the assumption that you can detect if people lie or not based on their physical reactions: Sweating, heartbeat, etc. We know that lying isn't the only reason why people would feel this way, especially during a polygraph test.

Which, of course, completely blows the claim out of the water.

That is a very, very serious misunderstanding of how science works, at the most basic level.

The fact that the same effect can occur for reasons other than the one we're interested in would 'blow the claim out of the water' only if the claim was for a perfect technology that worked with 100% accuracy every time. (Of course, that could not be a scientific claim.)

Have you any idea how commonly we use tests that are known to be far from perfect, in many everyday areas – medicine, for example? In all scientific fields (though, no doubt, to a lesser extent in particle physics than in medicine) we expect that, no matter how sound the underlying theory, the accuracy of our tests will be degraded by confounding factors.

Actually, in the case of polygraphy, the original rationale requires that deceptive and fearful responses give essentially the same results (and therefore can be distinguished only imperfectly, and only by expert questioners). This is mistaken - although it is not my area, I am somewhat persuaded by the NA report and digithead's posts on the issue of GKT vs CQT that polygraphy is likely to work better with questioning strategies that target cognition rather than emotion.

I wonder whether you realise how harmful your bulldog, non-scientific 'skeptic' approach is to the causes you're trying to promote? A casual reader of this thread might notice how much better informed the scientist posters are than you on this subject, and that we are concentrating on the point that polygraphy is a valid concept with some support from laboratory studies (because you've forced us to do so if we don't want your multitude of incorrect statements to go unchallenged). And might well conclude that real-world use of the polygraph is therefore justified.

As I have said, I agree with you that there is no case for any use of routine polygraphy. In real-world applications it cannot be shown to be either safe or effective, because the test's discrimination (though well above chance levels) is far too low to allow confidence for practical purposes. It has become established because of uncritical acceptance of outrageous claims of efficacy. But we have to counter this with evidence and scientific analysis, not with political-style campaigns to label it 'pdeudoscience'.

 

[CFLarsen]

I don't quite understand what you are saying. I don't care what the mechanism is, whether it's fear, intimidation, placebo effects, whatever. I'm asking whether you believe the process of polygraphy produces result better than chance, regardless of the reason. I don't really care about the why.

But you can't ignore the why.

We know that polygraphs "work" because they intimidate people. They do extract confessions, but not because the apparatus can determine if you tell the truth or not. You confess, because you are bamboozled into confessing what the interrogators already know.

But in the lab, we CAN know whether the subject is lying or not.

I can do the same, proving that astrology works.

Now what? You're f'ed.

I do agree that lab tests are of limited use in real life due to the fact that:
(a) in real life the subject typically has a vested interest in "getting away" with the lie
(b) in real life polygraphs are typically used in "high-stress" situations.

But this doesn't invalidate the fact that the polygraph process apparently works better than chance at detecting deception in the lab.

Yes, it does, because you cannot ignore the fact that polygraphs "work" by intimidation. In order for the polygraph to "work", you have to be scared out of your skull. For real.

The line-drawing fallacy has been explained, linked to, clarified, summarized, illustrated, annotated, construed, elucidated, pointed out, rendered, set out, set forth, spelled out, and shown to you more than once, and yet you cannot help yourself from engaging in it once more.

Apart from equivocation, it seems to be your favorite fallacy -- as it is here in this "argument."

If Thanz wants to point to a line-drawing, he is the one who has to point out where the line is drawn.

That is a very, very serious misunderstanding of how science works, at the most basic level.

The fact that the same effect can occur for reasons other than the one we're interested in would 'blow the claim out of the water' only if the claim was for a perfect technology that worked with 100% accuracy every time. (Of course, that could not be a scientific claim.)

Have you any idea how commonly we use tests that are known to be far from perfect, in many everyday areas – medicine, for example? In all scientific fields (though, no doubt, to a lesser extent in particle physics than in medicine) we expect that, no matter how sound the underlying theory, the accuracy of our tests will be degraded by confounding factors.

Stop. Right there.

Where did anyone demand that polygraphs work 100%?

No? Hm. B'bye.

 

[Lucky]

The polygraph is based on the assumption that you can detect if people lie or not based on their physical reactions: Sweating, heartbeat, etc. We know that lying isn't the only reason why people would feel this way, especially during a polygraph test.

Which, of course, completely blows the claim out of the water.

That is a very, very serious misunderstanding of how science works, at the most basic level.

The fact that the same effect can occur for reasons other than the one we're interested in would 'blow the claim out of the water' only if the claim was for a perfect technology that worked with 100% accuracy every time. (Of course, that could not be a scientific claim.)

Have you any idea how commonly we use tests that are known to be far from perfect, in many everyday areas – medicine, for example? In all scientific fields (though, no doubt, to a lesser extent in particle physics than in medicine) we expect that, no matter how sound the underlying theory, the accuracy of our tests will be degraded by confounding factors.

Actually, in the case of polygraphy, the original rationale requires that deceptive and fearful responses give essentially the same results (and therefore can be distinguished only imperfectly, and only by expert questioners). This is mistaken - although it is not my area, I am somewhat persuaded by the NA report and digithead's posts on the issue of GKT vs CQT that polygraphy is likely to work better with questioning strategies that target cognition rather than emotion.

I wonder whether you realise how harmful your bulldog, non-scientific 'skeptic' approach is to the causes you're trying to promote? A casual reader of this thread might notice how much better informed the scientist posters are than you on this subject, and that we are concentrating on the point that polygraphy is a valid concept with some support from laboratory studies (because you've forced us to do so if we don't want your multitude of incorrect statements to go unchallenged). And might well conclude that real-world use of the polygraph is therefore justified.

As I have said, I agree with you that there is no case for any use of routine polygraphy. In real-world applications it cannot be shown to be either safe or effective, because the test's discrimination (though well above chance levels) is far too low to allow confidence for practical purposes. It has become established because of uncritical acceptance of outrageous claims of efficacy. But we have to counter this with evidence and scientific analysis, not with political-style campaigns to label it 'pseudoscience'.

Stop. Right there.

Where did anyone demand that polygraphs work 100%?

No? Hm. B'bye.

I repeat one of my many unanswered questions to you:

CFLarsen: Who on earth do you think you're impressing with this foolery?

 

[CFLarsen]

I repeat one of my many unanswered questions to you:

You can repeat as many questions you want.

Where did anyone demand that polygraphs work 100%?

 

[qwert]

But you can't ignore the why.

We know that polygraphs "work" because they intimidate people. They do extract confessions, but not because the apparatus can determine if you tell the truth or not. You confess, because you are bamboozled into confessing what the interrogators already know.

But you don't believe that the result of this intimidation can be measured as a physiological response?

I agree that the primary efficacy of the polygraph is due to the intimidation / placebo effect of the subject believing that the polygraph works. But the research does seem to indicate that the subjects reaction to this shows up as a measurable physiological response when deceiving.

How else do you explain the research that indicates that a computer can analyze the results of a polygraph test & determine deception with rates of success better than chance.

Can you perform a similar laboratory test regarding astrology?

On another note, if the polygraph simply does not work better than chance, it seems odd that Michael Shermer bothered to learn polygraph countermeasures prior to being hooked up to one....from the video link you posted in #57

 

[digithead]

On another note, if the polygraph simply does not work better than chance, it seems odd that Michael Shermer bothered to learn polygraph countermeasures prior to being hooked up to one....from the video link you posted in #57

I really should be working on my dissertation but I'll just quickly interject here...

Countermeasures are designed to give the desired physical response to control and relevant questions so that the polygrapher scores the test as NDI (no deception indicated). It does not matter whether the subject is actually telling the truth or not, just that they're showing a response that's indicative of NDI.

Countermeasures can be used both by those that are deceptive to fool the polygrapher and by those that are truthful to ensure that that they pass.

What person would not want to know how the CQT procedure works and what physiological responses indicate NDI? Especially if there are ways to produce such responses which research shows takes minimal training.

And with the ease of finding countermeasure information on the internet, polygraphers, who have no proven ability to detect countermeasures, have now begun to accuse people of using countermeasures so that they will admit they used them...

 

[NoZed Avenger]

If Thanz wants to point to a line-drawing, he is the one who has to point out where the line is drawn.

No one has to point to a line. That's one reason it is called a "fallacy". This has been explained to you more than once, though you either refuse to read about it or simply (and willfully) choose to use the fallacious reasoning, regardless. I lean toward the latter explanation.

 

[CFLarsen]

But you don't believe that the result of this intimidation can be measured as a physiological response?

I agree that the primary efficacy of the polygraph is due to the intimidation / placebo effect of the subject believing that the polygraph works. But the research does seem to indicate that the subjects reaction to this shows up as a measurable physiological response when deceiving.

How else do you explain the research that indicates that a computer can analyze the results of a polygraph test & determine deception with rates of success better than chance.

The problem is that we cannot distinguish between fear of being found out and fear of the polygraph.

Can you perform a similar laboratory test regarding astrology?

If I can dictate what the conditions are, and leave out real-life situations, I can prove anything.

Countermeasures are designed to give the desired physical response to control and relevant questions so that the polygrapher scores the test as NDI (no deception indicated). It does not matter whether the subject is actually telling the truth or not, just that they're showing a response that's indicative of NDI.

Countermeasures can be used both by those that are deceptive to fool the polygrapher and by those that are truthful to ensure that that they pass.

What person would not want to know how the CQT procedure works and what physiological responses indicate NDI? Especially if there are ways to produce such responses which research shows takes minimal training.

And with the ease of finding countermeasure information on the internet, polygraphers, who have no proven ability to detect countermeasures, have now begun to accuse people of using countermeasures so that they will admit they used them...

That's the inherent problem with the efficacy of a technology that hinges on people not knowing how it can be beat: If polygraphs work as claimed, then they would be used in any situation, not just for screening.

But then, people would have a huge incentive to learn how to beat the polygraph. And of course they will.

You build a fence, and expect people not to learn how to cross it.

I have learned to beat the polygraph. Really. What, you don't believe me? I'll take a polygraph test to prove it...

 

[Belz...]

I take comfort in the scientific community sharing my view.

How does this relate, in any way, to what I posted ?

No, I'm not being silly. Your claim that polygraphs don't perform at all, if it were pseudoscience, is simply false. They work - but because they extract confessions because they scare the bejeebus out of people.

Computers extract confessions from people ?

Nonsense. Chance could very well be a factor.

Now you are just being silly. A meta analysis that shows performance of multiple tests performing well above what is expected by chance is the result of... chance?

I see that's why Claus didn't want to answer that particular question of mine...

 

[CFLarsen]

How does this relate, in any way, to what I posted ?

Computers extract confessions from people ?

No wonder you have a hard time understanding. "They" refer to polygraphs, not computers.

I see that's why Claus didn't want to answer that particular question of mine...

You have to learn how to read.

 

[Thanz]

Okie doke.

Please specify the degree of disagreement it takes.

Line drawing is your favourite game, isn't it? Well, it takes just enough disagreement.

It isn't a question of you being wrong or not.

I guess that I am right then. So, what's your problem?

Have you been qualified by Randi to perform preliminary tests?

No? Then stop lecturing those who have on how to go about it.

If you look at what I wrote, I said nothing about how one would go about performing a preliminary test. Rather, I posted about how one would get tot hat point. I think that it is crystal clear that one only gets to that point when the challenge has been accepted. And that is only done by the JREF. Not you. So giving me your opinion on whether a polygraph would qualify for the challenge in no way compromises the JREF's ability to evaluate any such applications themselves.

You misunderstand. It isn't about what qualifies for the challenge, but where the psychic got her notion that she could talk to dead people from.

That comes from real life, and not a lab test. Right?

Why does it matter? How the alleged pschic came to learn of his powers or whether he can find Timmy in the well or whatever is irrelevant. THe challenge would be: can you correctly identify Zener cards in a properly blinded and controlled experiment.

It would approximate random chance if you tested it in a situation where people would actually be intimidated by the polygraph. But you don't get that in lab tests, because there's no real intimidation.

First: any basis for the factual claim in the first sentence?

Second: Your position has been that polygraphs only work because of fear and intimidation. Here you are saying that the fear and intimidation would bring the test dow to random chance level. Which is it?

If you agree that fear and intimidation are not present in lab conditions, and testing done in lab conditions shows an accuracy rate well above chance, then fear and intimidation cannot be the reason the polygraph 'works'.

How can you design a lab test that invokes the same fear reactions as a real life test would?

I wouldn't, because that is not what we are testing. What we are testing is if the theory could be sound - deception causes certain physiological responses, that can be measured, and then interpreted to reveal deception.

No, I'm not being silly. Your claim that polygraphs don't perform at all, if it were pseudoscience, is simply false. They work - but because they extract confessions because they scare the bejeebus out of people.

You are not even logically consistent in this one post! Above, you assert that these factors would bring the hit rate in line with random chance. Now you say that they are the only reason they work!

We can't get around the scare factor, when we test polygraphs. How do we get that in a lab?

We don't want that in the lab. The theory behind the polygraph is not based on fear. Therefore, we would want to eliminate that factor in lab testing.

 

[CFLarsen]

Line drawing is your favourite game, isn't it? Well, it takes just enough disagreement.

If you are going to use a fuzzy term based on your personal appreciation, then we can safely say that you don't think Intelligent Design is considered pseudoscience.

You're a Creationist.

I guess that I am right then. So, what's your problem?

It isn't a question of being right or wrong.

If you look at what I wrote, I said nothing about how one would go about performing a preliminary test. Rather, I posted about how one would get tot hat point. I think that it is crystal clear that one only gets to that point when the challenge has been accepted. And that is only done by the JREF. Not you. So giving me your opinion on whether a polygraph would qualify for the challenge in no way compromises the JREF's ability to evaluate any such applications themselves.

You are in no position to judge that.

Why does it matter? How the alleged pschic came to learn of his powers or whether he can find Timmy in the well or whatever is irrelevant. THe challenge would be: can you correctly identify Zener cards in a properly blinded and controlled experiment.

It matters because that's where the claims come from: Real life.

First: any basis for the factual claim in the first sentence?

You obviously have forgotten what I responded to.

Second: Your position has been that polygraphs only work because of fear and intimidation. Here you are saying that the fear and intimidation would bring the test dow to random chance level. Which is it?

No, I am not saying that.

If you agree that fear and intimidation are not present in lab conditions, and testing done in lab conditions shows an accuracy rate well above chance, then fear and intimidation cannot be the reason the polygraph 'works'.

Again, I am not saying that.

I wouldn't, because that is not what we are testing. What we are testing is if the theory could be sound - deception causes certain physiological responses, that can be measured, and then interpreted to reveal deception.

Ah, but that's exactly what we are testing: If people lie or not. Lying is inherently linked with anxiety, so if you want to test polygraphs, you have to test people's anxiety levels.

You are not even logically consistent in this one post! Above, you assert that these factors would bring the hit rate in line with random chance. Now you say that they are the only reason they work!

No, I am not.

Are you drunk? Seriously?

We don't want that in the lab. The theory behind the polygraph is not based on fear. Therefore, we would want to eliminate that factor in lab testing.

I think it's time for you to do some studying on polygraphs. Basic studying.

 

[NoZed Avenger]

I think it's time for you to do some studying on polygraphs. Basic studying.

As it is time in the thread to offer unsolicited opinions, I'll take a turn:

I think it's far, far past time to do some studying on logical fallacies and how to avoid them. Very, very basic studying.

 

[Thanz]

If you are going to use a fuzzy term based on your personal appreciation, then we can safely say that you don't think Intelligent Design is considered pseudoscience.

You're a Creationist.

Have you completely given up on logic? Your conclusions in no way follow from your premise. If you really want to know my thoughts on ID, do a forum search.

It isn't a question of being right or wrong.

Of course it is. "Accuracy, always", right? If I am correct in my statements about the challenge, there is no need for you to get huffy about some special status you think you have.

You are in no position to judge that.

Just like the challenge, there is no need to judge. It is self evident. From the challenge rules:

This offer is administered by the JREF, and no one may negotiate or make any changes, except as set forth in writing by James Randi (JR).

Given that, how does your opinion of whether something might qualify, if an application were made, in any way effect the JREF's ability to evaluate challenge applications?

You obviously have forgotten what I responded to.

What a fantastic non-response. I'll take that as a big "no, I can't actually back that up."

No, I am not saying that.

But you did:

It would approximate random chance if you tested it in a situation where people would actually be intimidated by the polygraph.

This can only be read as saying that real intimidation limits the test to the point where it only achieves the same success as random chance.

Again, I am not saying that.

Again, you did. Are you know retracting that stance?

Ah, but that's exactly what we are testing: If people lie or not. Lying is inherently linked with anxiety, so if you want to test polygraphs, you have to test people's anxiety levels.

More specifically, you are testing variations in anxiety levels. That can be done without needing to attempting to simulate the elevated stress levels of say, a murder investigation.

No, I am not.

You have floated both theories in the same post. The first is that real life stress brings the hit rate to random chance - that is, eliminates the effectiveness of the machine. The second is that the machine only works by intimidating the respondent. Both cannot be true.

I think it's time for you to do some studying on polygraphs. Basic studying.

I think it is time for you to learn some logic. Basic logic.

 

[CFLarsen]

Have you completely given up on logic? Your conclusions in no way follow from your premise. If you really want to know my thoughts on ID, do a forum search.

It's not my problem that your logic isn't in the best shape.

Of course it is. "Accuracy, always", right? If I am correct in my statements about the challenge, there is no need for you to get huffy about some special status you think you have.

You need to get it through your head: Skeptica can do preliminary tests for the challenge.

Just like the challenge, there is no need to judge. It is self evident. From the challenge rules:
Given that, how does your opinion of whether something might qualify, if an application were made, in any way effect the JREF's ability to evaluate challenge applications?

Already explained.

What a fantastic non-response. I'll take that as a big "no, I can't actually back that up."

But you did:
This can only be read as saying that real intimidation limits the test to the point where it only achieves the same success as random chance.

Again, you did. Are you know retracting that stance?

More specifically, you are testing variations in anxiety levels. That can be done without needing to attempting to simulate the elevated stress levels of say, a murder investigation.

No, no, no. You are first running a series of questions meant to calm you down, so you get some sort of baseline. Then, when the tough questions start, you get all jittery because being found out is detrimental.

You have floated both theories in the same post. The first is that real life stress brings the hit rate to random chance - that is, eliminates the effectiveness of the machine. The second is that the machine only works by intimidating the respondent. Both cannot be true.

I think it is time for you to learn some logic. Basic logic.

Get back when you have done an inkling of research.

 

[.13.]

And I really don't care. I haven't said a single thin about whether polygraphs are accurate or not. All I did was try to clarify something for .13.. He asked if something being 70% accurate really means that any single output has a 70% chance of being correct. The answer is yes. Your bizzare rants are utterly irrelevant to anything I have said. That said, .13.'s recent posts give the impression that he is also being deliberately obtuse rather than actually being interested in any answer, so there doesn't seem much point in being here at all.

Bolding mine. You have misunderstood me then if you think I was asking that.

Why are you both babbling about irrelevant nonsense? It was stated that if something is shown to be 70% accurate, then when you run it once, there is a 70% chance that it is correct. .13. questioned this. I pointed out that actually, yes, that is entirely true, and is in fact what 70% accurate means. It doesn't matter if you're talking about polygraphs, tossing coins, predicting the weather or anything else. If you are right 70% of the time, then any single outcome has a 70% chance of being right. End of story.

You can argue whether any particular thing, like polygraphs, really are 70% accurate, and I haven't said anything about that. The fact remains that if they are 70% accurate, then they are 70% accurate. The more you argue with this, the more it looks like either you can't understand infant level maths or are just trolling because you can't make a real argument.

Questioning the claim that a polygraph can discriminate between nervous reaction and a lie is not irrelevant nonsense.


Will you please answer wether you are useing a fair coin or not?

 

[Thanz]

It's not my problem that your logic isn't in the best shape.

No, your problem is that you wouldn't know a logical statement if your life depended on it. You can't "safely say" anything about my position on topic "A" based on my statements regarding completely different topic "B".

You need to get it through your head: Skeptica can do preliminary tests for the challenge.

You need to get it through your head: Preliminary tests are only done after a challenge has been accepted by the JREF and a protocol agreed upon. I am asking you if you think a polygraph test would qualify. You have no official say on that (and your initial responses to me acknowledge this) and therefore your opinion has no binding effect on the JREF.

So don't hide behind some special status you think you have to avoid answering a simple question.

Already explained.

No, it hasn't been. How will you effect the JREF's determination on what challenges to accept?

[snip annoying graphic]

Argument by emoticon. Very mature. It simply shows that you have no coherent argument to make.

No, no, no. You are first running a series of questions meant to calm you down, so you get some sort of baseline. Then, when the tough questions start, you get all jittery because being found out is detrimental.

You get a baseline. Then compare answers against the baseline. You are testing variations from the baseline. This can be done (and has been done) without the stress of say, a murder investigation. In controlled lab conditions. Which is the heart of my question: If a polygraph operator can detect deception at a rate well above chance in these controlled, less stressful lab conditions, would that qualify for the MDC in your opinion?

 

[Belz...]

No wonder you have a hard time understanding. "They" refer to polygraphs, not computers.

No, YOU need to follow the conversation. It was explained to you that computers can determine if someone is lying or not and pretty much agree with the people operating the polygraph. How can a computer get those results through extracting a confession ?

You have to learn how to read.

You have to learn to answer people with more than just rhetoric.