If You Want to Catch a Liar, Make Him Draw | Hacker News Reader

If You Want to Catch a Liar, Make Him Draw

neuronarrative.wordpress.com

63 pointsfogus21 comments

I hate stuff like this. The article is about generalizations made from a group of people. If you apply that to a single person (to "catch a liar"), you will only sometimes be correct. Catching a liar is something that you should only do when you know you are correct, and all lie detection schemes cannot do that. Sort of like our system of justice which would rather a guilty person go free that convict an innocent person. Well, in theory at least.

Expecting perfection is unreasonable - everything has a failure rate. The question is what's an acceptable percentage of false positives.

I wonder how effective this will be after it gets published in mass media. I, for one, will now surely remember to use a "shoulder-surfing" perspective if I ever get into such a situation...

Lie detection techniques are a bit like secure software; eventually the best of the best will be good at reducing the reliability of the best expert. And once a technique is covered on enough episodes of CSI/Lie to Me, then the "script kiddies" might be able to fool them sometimes, too.

Later, though, liars will have to eventually beat systems like this:

http://noliemri.com/products/Overview.htm

That is a good point. A trained criminal will probably not fall for this because they will study articles such as this to learn about what they should draw, how they should respond, etc.

Even a lie detector can probably be fooled by someone who knows how to keep themselves calm while lying.

There is no such thing as a "lie detector". It is all nonsense. And it is dangerous and harmful that people think that lie detectors are worth something.

If you want to understand why, you can read what is on http://antipolygraph.org/

I actually agree with you. The polygraph isn't useful. To my way of thinking people can scam it just like they could scam the drawing test. But what is even worse about the polygraph is that it will often misreport someone as lying when they told the truth.

That's why polygraph test aren't usually accepted in court in my country.

For that matter, though, I don't think any type lie detector test, polygraph, drawing, etc is really worthy of being used as court evidence. It may help law officials who suspect that someone is lying, but it isn't going to be conclusive proof no matter how you do it.

I'm glad I read it. I'm pretty sure I would always draw like the "liar". I am a terrible artist and resort to boxes and lines from the overhead view. I better start practicing some truth art.

But then the psychologist determines that you're one of the types that would have normally have used an 'overview' perspective, branding you either 'too smart' or 'a manipulator', neither of which is good.

For another failure mode: they should repeat the experiment such that some people in both groups are told in no uncertain terms that the other party thinks they are lying. How many of the 'truth speakers' would start drawing overview pictures in order to try and convince the other that they must have been there?

Or for another variety: allow everyone to visit the location and have the talk, but tell one half to do everything to lie about it and introduce some stress. Some people crack so soon under stress that they will actually start believing their own lies.

This entire line of investigation is useless.

I, for one, will simply refuse. You don't have to testify against yourself, and not drawing is a lot easier than drawing.

Except, such papers forget the power of reflexivity : after the douchebags read the paper, they will start dissimulating very cleverly, rendering the results effectively useless.

It's like that article back in the 90's which said that "gets-things-done"-type of people often folded their shirts up to their elbows, and all the mouths started imitating this.

It's like that article back in the 90's which said that "gets-things-done"-type of people often folded their shirts up to their elbows, and all the mouths started imitating this.

LOL That's a good point. I never read that article but now that you mention it that is still a relatively good way of telling the doers from the talkers.

But you are right, people are going to purposefully try to make themselves good if they know that there is a way of testing them or observing them.

I guess the key is if you find the ultimate way to tell a douchebag from a professional or a liar from a truthful person then don't tell anybody about it!

The key error inexperienced liars make is usually inventing too much; you ask them about different parts of it and the parts don't add up right. A good lie is short and simple. This is a technique for forcing more detail out of people.

Why will only "douchebags" have a need for these techniques?

This comment is indescribably hilarious.

Using the “sketching the agent” result alone, it was possible to identify 80% of the truth tellers and 87% of the liars – results superior to most traditional interview techniques.

What a great example of the base rate fallacy. See: http://en.wikipedia.org/wiki/Base_rate_fallacy

hmmm... While reading through I visualized drawing the top-down perspective. Guess I'm guilty.

thats the same way I felt, I see myself as a over-view type of person at most times... I always see things from the 50k ft level.

I guess I've been using Google Earth too much. I see things from several miles up. ;)

But seriously, a good criminal would check Google Street View to see what the area looks like before the drawing test. Then they would be able to draw it.

(not directly related to this article) The more I read about the brain, the less I believe in "free will". We have a nice illusion though...

People lie if truth hurts their self-esteem.

Reply on news.ycombinator.com