# Polygraph reliability **URL:** \ **Category:** Factual Questions\ **Created:** [March 31, 2017, 2:05pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411 "2017-03-31T14:05:38Z")\ **Posts on this page:** 20\ **Page:** 4
**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\ **Post date:** [April 3, 2017, 8:40pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/61 "2017-04-03T20:40:33Z")
> [@Czarcasm](#): > > Where did you get the “60%” figure from? > **[The Truth About Lie Detectors (aka Polygraph Tests)](https://www.apa.org/topics/cognitive-neuroscience/polygraph)** > > Most psychologists agree that there is little evidence that polygraph tests can accurately detect lies. \*The cumulative research evidence suggests that CQTs detect deception better than chance, but with significant error rates, both of misclassifying innocent subjects (false positives) and failing to detect guilty individuals (false negatives). \* > **[The Truth About Lie Detectors (aka Polygraph Tests)](https://www.apa.org/topics/cognitive-neuroscience/polygraph)** > > Most psychologists agree that there is little evidence that polygraph tests can accurately detect lies. _“Proponents will say the test is about 90 percent accurate. Critics will say it’s about 70 percent accurate,” said Frank Horvath of the American Polygraph Association. "Many people refer to polygraph tests as lie detector tests, and that’s a bit of a misnomer._ ---
**Author:** ![Czarcasm](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/czarcasm/32/4050_2.png) [@Czarcasm](https://boards.straightdope.com/u/Czarcasm)\ **Post date:** [April 3, 2017, 9:25pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/62 "2017-04-03T21:25:21Z")
> [@DrDeth](#): > > [The Truth About Lie Detectors (aka Polygraph Tests)](http://www.apa.org/research/action/polygraph.aspx) > \*The cumulative research evidence suggests that CQTs detect deception better than chance, but with significant error rates, both of misclassifying innocent subjects (false positives) and failing to detect guilty individuals (false negatives). > \* > > [The Truth About Lie Detectors (aka Polygraph Tests)](http://www.apa.org/research/action/polygraph.aspx) > _“Proponents will say the test is about 90 percent accurate. Critics will say it’s about 70 percent accurate,” said Frank Horvath of the American Polygraph Association. "Many people refer to polygraph tests as lie detector tests, and that’s a bit of a misnomer._ Two quotes from the same source, neither of which say “60%”. Perhaps you are a little unclear on how cites are supposed to work? ---
**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\ **Post date:** [April 3, 2017, 9:56pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/63 "2017-04-03T21:56:38Z")
> [@Czarcasm](#): > > Two quotes from the same source, neither of which say “60%”. > Perhaps you are a little unclear on how cites are supposed to work? One says \* better than chance\* the other _about 70 percent accurate_; perhaps you are unclear on how math works? ---
**Author:** ![Musicat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/musicat/32/20189_2.png) [@Musicat](https://boards.straightdope.com/u/Musicat)\ **Post date:** [April 3, 2017, 9:57pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/64 "2017-04-03T21:57:03Z")
> [@TriPolar](#): > > Polygraph machine manufacturers don’t like too cooperate with scientific studies. Traditionally they have insisted that the results are dependent on the skill of the operator, which is a clue to how little they are based on scientific principles. Noting the post above, many people certainly will crack under the pressure of a polygraph exam, even if a fake polygraph machine is used. But a good interrogator can get the same results. > [@DrDeth](#): > > Cold reading is quite accurate, far more than a “blind guess”. … Not according to [Ian Rowland,](https://www.amazon.com/Full-Facts-Book-Cold-Reading/dp/0955847605) who wrote the book on the topic, literally. Cold reading, or the acceptance of its accuracy, is subject to confirmation biases. So is the interpretation of a polygraph chart. Lines drawn on the chart don’t go from “yes” to “no”, but usually end up somewhere in between. It’s up to the examiner to decide if the line is closer to one or the other, made harder by the lack of a firm line to cross. What does 60% between yes & no mean? 40% maybe? ---
**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\ **Post date:** [April 3, 2017, 10:03pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/65 "2017-04-03T22:03:41Z")
> [@Musicat](#): > > Not according to [Ian Rowland,](https://www.amazon.com/Full-Facts-Book-Cold-Reading/dp/0955847605) who wrote the book on the topic, literally. That very expensive book? Do you own a copy? A book with horrible reviews? ---
**Author:** ![Musicat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/musicat/32/20189_2.png) [@Musicat](https://boards.straightdope.com/u/Musicat)\ **Post date:** [April 3, 2017, 11:09pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/66 "2017-04-03T23:09:18Z")
> [@DrDeth](#): > > That very expensive book? Do you own a copy? A book with horrible reviews? I do not own the book, but I have met Ian in person on more than one occasion and watched him work as a magician and mentalist professionally. He was a poster on SDMB a while ago under the name [ianzin,](http://boards.straightdope.com/sdmb/member.php) last posting in November 2016. He knows his shit, no matter what the book critics may say. And he’s not alone in his assessment of cold reading. This subject was a frequent topic in skeptic meetings/conventions I used to attend. The procedure is pretty well defined by the experts now, whether white hat or black hat. ---
**Author:** ![Musicat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/musicat/32/20189_2.png) [@Musicat](https://boards.straightdope.com/u/Musicat)\ **Post date:** [April 3, 2017, 11:19pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/67 "2017-04-03T23:19:36Z")
> [@DrDeth](#): > > That very expensive book? Do you own a copy? A book with horrible reviews? You may have a problem with mathematics. Of the 8 reviews on the Amazon book site, 5 are favorable, even enthusiastically so. Of the 3 who gave negative reviews, one complained about the price, one complained about typos, and one complained about the binding, at the same time complementing the author on his knowledge of the subject. “Horrible”? I do not think that word means what you think it means. ---
**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\ **Post date:** [April 3, 2017, 11:40pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/68 "2017-04-03T23:40:12Z")
> [@Musicat](#): > > I do not own the book, but I have met Ian in person on more than one occasion and watched him work as a magician and mentalist professionally. He was a poster on SDMB a while ago under the name [ianzin,](http://boards.straightdope.com/sdmb/member.php) last posting in November 2016. He knows his shit, no matter what the book critics may say. > > And he’s not alone in his assessment of cold reading. This subject was a frequent topic in skeptic meetings/conventions I used to attend. The procedure is pretty well defined by the experts now, whether white hat or black hat. So, why cite the book then? In any case, does he say a blind guess is more accurate? ---
**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\ **Post date:** [April 3, 2017, 11:43pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/69 "2017-04-03T23:43:31Z")
> [@Musicat](#): > > You may have a problem with mathematics. > > Of the 8 reviews on the Amazon book site, 5 are favorable, even enthusiastically so. Of the 3 who gave negative reviews, one complained about the price, one complained about typos, and one complained about the binding, at the same time complementing the author on his knowledge of the subject. You mean this review? :"\*The absolute WORST of its genre, this book is overpriced mediocrity. As other reviewers have said, Rowland’s book is full of typos, poorly designed, poorly written and desperately in need of an editor. More importantly, there are far better books than this pathetic pamphlet on the subject, which is probably why Rowland couldn’t land a publisher and did an amateurish job of self-publishing it. Rowland’s writing is immature, boring, tiresome and pedantic, and I barely made it halfway through the book before mercifully giving up and just skimming the rest. It’s hard to take what he says seriously as his attempts at whimsy and congeniality fail, like giving each one of his methods cutesy, silly names like “Rainbow ruse” to explain how to use opposites to describe a person, such as “you are generous but also frugal when you want to be.” Just Google some of the Barnum effect stuff for free and you will learn far more than this book offers."\* The \*only \*in depth review? _Horrible_ reviews. ---
**Author:** ![Czarcasm](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/czarcasm/32/4050_2.png) [@Czarcasm](https://boards.straightdope.com/u/Czarcasm)\ **Post date:** [April 3, 2017, 11:48pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/70 "2017-04-03T23:48:13Z")
> [@DrDeth](#): > > So, why cite the book then? > > In any case, does he say a blind guess is more accurate? A blind guess is certainly better, because it isn’t presented as if there is scientific basis or certainty to it, unlike polygraph “experts” who, when presented with evidence that the tool they use isn’t worth shit, make a counter claim that _they_ are the real “polygraph”. Blind guesses are at least devoid of deception when presented as such. ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [April 3, 2017, 11:48pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/71 "2017-04-03T23:48:35Z")
That’s a horrible review, singular. You said “reviews”? ---
**Author:** ![iamthewalrus\_3](https://avatars.discourse-cdn.com/v4/letter/i/258eb7/32.png) [@iamthewalrus\_3](https://boards.straightdope.com/u/iamthewalrus_3)\ **Post date:** [April 3, 2017, 11:59pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/72 "2017-04-03T23:59:11Z")
> [@TriPolar](#): > > Polygraph machine manufacturers don’t like too cooperate with scientific studies. Traditionally they have insisted that the results are dependent on the skill of the operator, which is a clue to how little they are based on scientific principles. Honestly, this should be all we need. It would be easy to design and perform a double-blind study of polygraphs as lie detectors. That no one has come up with any scientific support for that in how many decades is sufficient evidence that they are not useful for that purpose. Same reason we know that cow-tipping isn’t a thing because if it were, it’d be on youtube. ---
**Author:** ![CookingWithGas](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/cookingwithgas/32/485_2.png) [@CookingWithGas](https://boards.straightdope.com/u/CookingWithGas)\ **Post date:** [April 4, 2017, 1:21pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/73 "2017-04-04T13:21:49Z")
Given the evidence against reliability of polygraph examinations (or at least lack of support for their efficacy), it’s surprising they are still a tool of choice when assessing applicants for certain flavors of Top Secret security clearances. ---
**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\ **Post date:** [April 4, 2017, 4:21pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/74 "2017-04-04T16:21:38Z")
> [@Czarcasm](#): > > A blind guess is certainly better, because it isn’t presented as if there is scientific basis or certainty to it, unlike polygraph “experts” who, when presented with evidence that the tool they use isn’t worth shit, make a counter claim that _they_ are the real “polygraph”. Blind guesses are at least devoid of deception when presented as such. Are you talking about cold reading or polygraphs now? The Blind Guess part was about Cold reading, not polygraphs. ---
**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\ **Post date:** [April 4, 2017, 4:22pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/75 "2017-04-04T16:22:43Z")
> [@Chronos](#): > > That’s a horrible review, singular. You said “reviews”? Yes, there are several short bad reviews and several short good reviews. Only one in depth review. Which is a one star. ---
**Author:** ![Czarcasm](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/czarcasm/32/4050_2.png) [@Czarcasm](https://boards.straightdope.com/u/Czarcasm)\ **Post date:** [April 4, 2017, 4:40pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/76 "2017-04-04T16:40:54Z")
> [@DrDeth](#): > > Are you talking about cold reading or polygraphs now? > > The Blind Guess part was about Cold reading, not polygraphs. Since there is no evidence that one-on-one cold reading is better than blind guessing(your friend’s claims notwithstanding), I would say I was talking about both. Neither method, cold reading or polygraphs, have anything solid to back them up, yet both use a mask of authority and science. ---
**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\ **Post date:** [April 4, 2017, 5:06pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/77 "2017-04-04T17:06:39Z")
Obviously you are arguing for the sake of arguing. I have explained two facts: 1. WHY some LEOs use polygraphs, even when they know the limitations 2. HOW some LEOs use polygraphs, even when they know the limitations I agree that polygraphs are unreliable, and pseudoscience. The trickery involved is unethical. I have stated that "_I personally opine that all polygraph machines should be junked and no government agency allowed to spend a nickel on such procedures. But that’s morality. "_ However my two facts stand. LEO & other agencies _do_ use Polygraphs, they do find them useful. I have tried to explain\* why\* they do so. I do not agree with their use. I am not defending their use. I am trying to help those who may be faced with one. Do you disagree that they do use polygraphs? :rolleyes: ---
**Author:** ![Czarcasm](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/czarcasm/32/4050_2.png) [@Czarcasm](https://boards.straightdope.com/u/Czarcasm)\ **Post date:** [April 4, 2017, 5:32pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/78 "2017-04-04T17:32:54Z")
> [@DrDeth](#): > > Obviously you are arguing for the sake of arguing. > > I have explained two facts: > > 1. WHY some LEOs use polygraphs, even when they know the limitations > 2. HOW some LEOs use polygraphs, even when they know the limitations > > I agree that polygraphs are unreliable, and pseudoscience. The trickery involved is unethical. I have stated that "_I personally opine that all polygraph machines should be junked and no government agency allowed to spend a nickel on such procedures. But that’s morality. "_ > > However my two facts stand. LEO & other agencies _do_ use Polygraphs, they do find them useful. I have tried to explain\* why\* they do so. I do not agree with their use. I am not defending their use. I am trying to help those who may be faced with one. > > Do you disagree that they do use polygraphs? :rolleyes: What I am disagreeing with is your continued use of the word “useful” without explaining that it is different from how the OP (and most of the rest of us) define it. It may be “useful”, in getting people to confess, rightly or wrongly, and it may be “useful” as a tool to make your friend’s portfolio look good…but the question is this thread concerns it’s usefulness in telling truth from lie. ---
**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\ **Post date:** [April 4, 2017, 5:45pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/79 "2017-04-04T17:45:45Z")
> [@Czarcasm](#): > > What I am disagreeing with is your continued use of the word “useful” without explaining that it is different from how the OP (and most of the rest of us) define it. It may be “useful”, in getting people to confess, rightly or wrongly, and it may be “useful” as a tool to make your friend’s portfolio look good…but the question is this thread concerns it’s usefulness in telling truth from lie. I have explained at length why and how LEos find it useful. ---
**Author:** ![Czarcasm](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/czarcasm/32/4050_2.png) [@Czarcasm](https://boards.straightdope.com/u/Czarcasm)\ **Post date:** [April 4, 2017, 6:01pm UTC](https://boards.straightdope.com/t/polygraph-reliability/783411/80 "2017-04-04T18:01:00Z")
> [@DrDeth](#): > > I have explained at length why and how LEos find it useful. A Polygraph can also be useful as a nutcracker but, like your examples, this is not the “useful” the OP is asking about. [Previous page](https://boards.straightdope.com/t/polygraph-reliability/783411.md?page=3) [Next page](https://boards.straightdope.com/t/polygraph-reliability/783411.md?page=5)