Detecting Liars by Content

Interesting:

Kevin Colwell, a psychologist at Southern Connecticut State University, has advised police departments, Pentagon officials and child protection workers, who need to check the veracity of conflicting accounts from parents and children. He says that people concocting a story prepare a script that is tight and lacking in detail.

“It’s like when your mom busted you as a kid, and you made really obvious mistakes,” Dr. Colwell said. “Well, now you’re working to avoid those.”

By contrast, people telling the truth have no script, and tend to recall more extraneous details and may even make mistakes. They are sloppier.

[…]

In several studies, Dr. Colwell and Dr. Hiscock-Anisman have reported one consistent difference: People telling the truth tend to add 20 to 30 percent more external detail than do those who are lying. “This is how memory works, by association,” Dr. Hiscock-Anisman said. “If you’re telling the truth, this mental reinstatement of contexts triggers more and more external details.”

Not so if you’ve got a concocted story and you’re sticking to it. “It’s the difference between a tree in full flower in the summer and a barren stick in winter,” said Dr. Charles Morgan, a psychiatrist at the National Center for Post-Traumatic Stress Disorder, who has tested it for trauma claims and among special-operations soldiers.

This is new research, and there are limitations to the approach, but it’s interesting.

Tags: ,

Posted on May 14, 2009 at 1:30 PM β€’ 52 Comments

Comments

Martin β€’ May 14, 2009 1:51 PM

The Department of Redundancy Department enthusiastically approves of the quotes in this blog post. πŸ™‚

PopeSidious β€’ May 14, 2009 1:53 PM

So what about someone who distrusts the police, and is telling the truth, but not an iota more than required?

PiP β€’ May 14, 2009 2:04 PM

This reminds me of the post on “don’t talk to the police” where police will interrogate and look for subtle details in the story to be inconsistent, which they use as indication that a person is lying. According to this research, the exact opposite is true.

the other Alan β€’ May 14, 2009 2:54 PM

“. . . don’t talk to the police.” Quite correct.

When (in the U.S.) they alter the Miranda warning from “anything you say, can and will be used against you in a court of law” to “. . . may be used properly for or against you in a court of law”, it’s good advice.

Seth Breidbart β€’ May 14, 2009 3:01 PM

Interestingly, I’ve found that people who feel guilty (insecure) are more likely to keep talking to provide more facts and justifications. People who feel secure just answer the question.

AMcguinn β€’ May 14, 2009 3:03 PM

I find it all very plausible and not very useful. It will detect lies, except for those told by good liars, and produce too many false positives to be relied on.

It doesn’t mean it isn’t good and interesting, but I’m wary of it being oversold

HJohn β€’ May 14, 2009 3:12 PM

AMcquinn: “I find it all very plausible and not very useful. It will detect lies, except for those told by good liars, and produce too many false positives to be relied on. It doesn’t mean it isn’t good and interesting, but I’m wary of it being oversold.

I agree. I’m guessing there is a high enough accuracy rate to make it somewhat useful, but a high enough error rate to be dangerous if depended on too greatly. Not to mention, the ones you need to catch the most are the ones most likely to be false negatives. False positives can also be costly in terms of misdirected resources, as well as the consequences to the wrongly accused.

Angel One β€’ May 14, 2009 3:34 PM

Seriously? This is new? Every parent in the world who has had a teenager lie to them knows this. Every teacher who has caught a student knows this. Every person who has ever confronted a coworker or colleague with a case of suspected duplicity knows this.

Didn’t you ever hear the old joke about the students who decide to play hookey to skip a test. When they do arrive in school they claim they had a flat tire, so the teacher gives them a 1 question test. The question is “which tire”. It’s very simple – catch the liars with lack of detail. This may be the first time someone has formalized and quantified it (20% is new), but the concept is as old as time.

Steve β€’ May 14, 2009 3:36 PM

@PopeSidious

at least in America, you aren’t required to tell the police ANYTHING.

o.s. β€’ May 14, 2009 3:36 PM

Interesting…I wonder how well this technique would work on car salesman?

o.s. β€’ May 14, 2009 3:42 PM

Interesting…I wonder how well this technique would work on car salesman?

P.S. There seems to be a race condition with the comments, two at a time and it drops one of the comments instead of queuing them.

Anonymous β€’ May 14, 2009 3:46 PM

someone who distrusts the police

No one should ever talk to the police ever. There’s a youtube video of a law prof explaining why. I’m sure that a quick google search will identify it. He was adamant that no one should ever talk to the police for any reason. Pretty interesting stuff.

Aliosha β€’ May 14, 2009 4:16 PM

I kind of remember reading once that psychopaths (4% of the population, give and take something) behave in the opposite way, they do not get nervous about getting caught and add a lot of spurious details. Their lies sound a lot like when normal people try to remember the truth, it is just that they are making it up.

(And I remember a legal advice for political protesters: “do talk to the police, but never about what happened. About football, girls, food, weather; be nice to them, but do not tell them anything about the subject.”
Never tried myself, though)

Jason β€’ May 14, 2009 4:19 PM

Sort of old hat, isn’t it. A recent episode of the television show ‘Rules of Engagement’ explained that the secret to effective lying is the inclusion of specific, yet extraneous details; exactly the kind of thing you’d stumble through while tracing your memories back to a past event.

If a television show sit com knows about it, you can be pretty sure it is common knowledge or at least pop knowledge.

tim β€’ May 14, 2009 4:27 PM

Another previous pop culture sighting of the ask-for-details lie detection method: Loveline with Adam Carolla and Dr. Drew Pinsky, at least as far back as 2001.

I’m with Angel One – this has probably been pretty well known, albeit informally, for a very long time, which means that there are probably a lot of liars out there who are savvy to it.

Alex β€’ May 14, 2009 4:53 PM

There is a film called “The Lives of Others,” whose opening sequence lines up well with the excerpt in this posting. The Stasi interrogate a man for many hours, because he is telling the exact same story, word for word, every time he is prompted to start over from the beginning. He eventually cracks, and the man interrogating later recounts this tale with the above psycho analysis.

Rich Wilson β€’ May 14, 2009 4:57 PM

This reminds me of what Bruce has said about Israeli airline security. They’ll have a conversation with you and probe to see just how deep the story goes if they’re driving it. “So this Uncle you’re going to visit, where do his kids go to school?”

Nostromo β€’ May 14, 2009 4:59 PM

This is probably true and verifiable in a statistical sense, but not useful in deciding which of 2 people contradicting each other is lying.

There are good liars, who will pass all the tests and be more likely to be believed than a nervous person who is telling the truth, and there are bad liars. That’s the dominant effect. And it partly explains why there are so many people in jail for crimes they didn’t commit.

David β€’ May 14, 2009 6:13 PM

Well it’s also obvious that lying and inventing stories, almost in real time, will make our brain work harder…
Of course not the right thing to do.

John β€’ May 14, 2009 9:22 PM

Some people are missing the point: it’s not that the stories that liars tell lack detail, or that the details are inconsistent. Drilling down further and further looking for internal contradictions is precisely the wrong way to proceed.

In fact, internal contradictions are probably more likely if the person is telling the truth, as they are drawing from fuzzy memories of real events.

The approach Alex mentions is more in line with what the researchers found. Have the person tell the same story two or three times, and look for variations. Somebody who is lying is likely to have less variation, and to be more internally consistent, than somebody who is drawing from actual memory banks, at least when it comes to peripheral detail.

Garrett β€’ May 14, 2009 9:50 PM

I wonder if Bruce has any thoughts/opinions on micro-expressions to detect liars?

Micro-expressions have recently become popular with the TV series “Lie to Me,” though I first read about them in a Malcolm Gladwell book called “Blink.”

Seems like the ultimate truth-getter, if one has the talent+experience to leverage it.

Gweihir β€’ May 14, 2009 10:59 PM

The real advice here is of course to follow the research and learn to fake it. Even if you cannot really surpresse the signs, you can add them to theother case. And without a really good baseline, it is hard to find out what is what.

Pants on Fire β€’ May 15, 2009 12:59 AM

I also think it’s thousands years old research..

Another nice movie reference is “the usual suspects”, where the main character tells the police a fully invented story, getting details by reading the newspapers posted on a wall in his interrogation room..

Jakub NarΔ™bski β€’ May 15, 2009 2:04 AM

This would not help, I think, if somebody is telling a lie by changeing a few details in otherwise true story/event.

Marco β€’ May 15, 2009 3:08 AM

In the movie “Reservoir Dogs” by Tarantino, the infiltrated cop is taught a funny story he can use, and he is continuously told how the details are fundamental for the story to be believable.

Karsten β€’ May 15, 2009 4:03 AM

Get people to tell their story or sequence of events backwards. People not telling the truth will have a harder time doing this.

Grey Bird β€’ May 15, 2009 4:08 AM

@ Civil Libertarian:

Showing ID and talking to the police aren’t the same thing at all. Showing ID gives them information, but isn’t something that can be used against you in a court. Talking to them, as discussed the video of the law professor, can get you in a lot of trouble even if you’re innocent.

bill β€’ May 15, 2009 4:40 AM

The best liars are those who believe they are telling the truth.

Anyone promoting superstition (yes I include religion), or anyone who has just witnessed a good magic trick “The card just vanished into thin air! I swear!”

So how nuanced is this detection method I wonder.

bill (again) β€’ May 15, 2009 4:41 AM

It’s like an inverse side-channel attack.

It’s what they’re NOT leaking that provides the information πŸ˜‰

xxx β€’ May 15, 2009 5:48 AM

This is being used by Israeli immigration/border officials for years — they ask you what you’re going to do in Israel and probe for details.

BF Skinner β€’ May 15, 2009 6:34 AM

@Gweihir “The real advice here is of course to follow the research and learn to fake it. ”

No the real advice here is of course to follow the reasearch and teach others how to fake it…for a fee.

confluence β€’ May 15, 2009 7:03 AM

I think this is only effective if you have the element of surprise. If I were falsely accused of something, and then left alone for several hours before someone questioned me, I’m pretty sure I would spend all of the intervening time carefully going over what I remembered, constructing a clear and coherent account with everything presented in logical order, and practicing exactly what I was going to say in front of a mirror. Whoops.

Clive Robinson β€’ May 15, 2009 8:31 AM

Anybody notice about the test subjects being composed of PTSD sufferes?

“Dr. Charles Morgan, a psychiatrist at the National Center for Post-Traumatic Stress Disorder, who has tested it for trauma claims and among special-operations soldiers.”

Well it is not commonly talked about but there is a very high corelation between suffering from PTSD and having been abused as a minor. One charecteristic of those abused as minors is that they fix on irrelevant details during stressfull periods as a learned response to surviving abuse.

Therefor I would treat these findings with a very large grain of salt as it is not likley ro be representative of the majority of the population…

Perhaps Dr Morgan and his ascociates should consider testing a larger % of the population in general.

Anonymous β€’ May 15, 2009 8:46 AM

@ confluence,

“I’m pretty sure I would spend all of the intervening time carefully going over what I remembered, constructing a clear and coherent account with everything presented in logical order, and practicing exactly what I was going to say in front of a mirror. Whoops.”

The easy solution to this is to write it down as nearly immediatly after the event as you can. Then use these notes exclusivly. When somebody tries to draw you simply refere back to your notes very slowly and read it out again.

You destroy their rythem and upset their thinking.

And if they say something you where unaware of make a very big fuss over it demand the persons identity and write it down at the end of your notes.

In general the police hate this as it will conflict with their way of doing things where they use there notes to convince a judge/jury of what they supposadly saw/heard/etc.

Allways allways write it down and that way you will find that you are in a less doutfull position, which the smarter police and cross questioners almost exclusivly use as a method of getting more information.

Anonymous β€’ May 15, 2009 10:52 AM

@ confluence,

“I’m pretty sure I would spend all of the intervening time carefully going over what I remembered, constructing a clear and coherent account with everything presented in logical order, and practicing exactly what I was going to say in front of a mirror. Whoops.”

The easy solution to this is to write it down as nearly immediatly after the event as you can. Then use these notes exclusivly. When somebody tries to draw you simply refere back to your notes very slowly and read it out again.

You destroy their rythem and upset their thinking.

And if they say something you where unaware of make a very big fuss over it demand the persons identity and write it down at the end of your notes.

In general the police hate this as it will conflict with their way of doing things where they use there notes to convince a judge/jury of what they supposadly saw/heard/etc.

Always always write it down and that way you will find that you are in a less doutfull position, which the smarter police and cross questioners almost exclusivly use as a method of getting more information.

Oh on another note, during the 1970’s when various military organisations where trying to train better snipers (or assasins) they found that you could be showing people an increasingly worse series of films de-sensitise them to the brutality and carnage on the film. The trick was to make people answer questions like “the car at the left had what number plate?” or “the shop on the right was selling what sort of product?”. It made the test subjects look all around the act of brutality looking for inane and unrelated items. In the process it speeded up the desensitisation process…

Which further suggests it is a way the brain deals with almost unimaginable horrors and still function effectivly.

crispy β€’ May 15, 2009 5:43 PM

The first example I heard of something like this was the Triangle Shirtwaist Factory fire litigation in the early 1910s, where a too-perfected story told by the chief witness for the plaintiff was used to convince the jury it had been rehearsed – looks like New York lawyers had almost 20 years on the Stasi in this one.

From listening to panhandlers trying to con passersby here in SF, I’d say a key element is that when a liar tells a story every single fact they recite will tie in to the lie; when in the truth of life a lot of things don’t make sense, or happen just because.

Terry Karney β€’ May 16, 2009 9:32 PM

As an interrogator I could have told you this, if anyone had thought to ask. We teach people to look for things which don’t add up.

Not for things which vary, but which aren’t consistent with things which have been said before. If I got a couple of sources who give me exactly the same story, I’ll find it questionable.

When asking, “repeat” questions I am asking for some piece of information the source has already provided. I’m doing it in a different context. If the answers match up, I’ll assume the source is telling the truth as best he knows it.

If every time I come to that subject I get an almost identical answer, I’ll start to suspect the source is hiding something.

Interesting to have someone else explain it to me.

Anonymous β€’ May 17, 2009 5:04 PM

@ Clive Robinson

I agree, people who have been traumatised are not representative of the untraumatised population.

Someone is abused over a long period of time will frequently encounter strong motivation to improve his or her lying skills, so as to avoid a particular instance of abuse. Additionally, the person may start lying to him or her self, trying to pretend that his or her life is better than it really is.

If someone starts lying to him or her self, I don’t think any lie detector will notice.

Mark β€’ May 18, 2009 7:53 AM

@o.s.
Interesting…I wonder how well this technique would work on car salesman?

Or indeed any salesmen…

Another practical application would be for politicans. Assuming that it wouldn’t be easier to have something which detects when they are telling the truth.

king-ink β€’ May 19, 2009 7:48 AM

I guess this result was already known for quite some time…

The movie “Das Leben der Anderen” starts with a lecture given by the leading role, who is an interrogation specialist of the Stasi (the former Eastern-Germany secret police). In this lecture he explains how to tell whether someone is guilty or not.

  • A guilty person (lying) always gives literally the same answers to the same question.

  • A non-guilty person will get outraged over the fact that he is suspected of something and that he is being detained, interrogated, and in general being treated badly. A guilty person accepts this as punishment for his act πŸ˜‰

bob β€’ May 19, 2009 9:59 AM

Theres a TV show called “Lie to me”? How could they sell advertising on it when C-SPAN runs two channels 24×7 where you watch non-stop lying?

@the other Alan: I am puzzled by your discriminator “in the US”. Since the Miranda warning is a US Supreme Court case, it only applies in the US to begin with.

I have frequently been accused of providing too much detail (hence all my parentheticals) when I answer questions (believing that people should be given adequate background to make informed decisions before making decisions; kind of the opposite of a “sound bite” or contemporary journalism where only enough information is given to provide a desired result).
So I would be in deep trouble in a police interview because they are looking to close a case, not solve a crime and the more info you give the more trouble you give in. On the other hand they would probably handcuff me and I dont think I could tell them the day of the week without using my hands, so maybe it would all average out.

Leave a comment

Login

Allowed HTML <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre> Markdown Extra syntax via https://michelf.ca/projects/php-markdown/extra/

Sidebar photo of Bruce Schneier by Joe MacInnis.