Monday, 7 June 2010

Drawing out the truth

Forget expensive fMRI-based lie detection or iffy polygraph tests, give your suspect a pencil and paper and get them to draw what happened - a new study suggests their artistic efforts will betray whether they are telling the truth or not.

Aldert Vrij's new study involved 31 police and military participants going on a mock mission to pick up a package from another agent before delivering it somewhere else. Afterwards the participants answered questions about the mission. Crucially, they were also asked to draw the scene of the package pick-up. Half the participants acted as truth-tellers, the others played the part of liars.

Vrij's team reasoned that clever liars would visualise a location they'd been to, other than where the exchange took place, and draw that. They further reasoned that this would mean the liars would forget to include the agent who participated in the exchange. This thinking proved shrewd: liars indeed tended not to draw the agent, whereas truth-tellers did. In fact, 80 per cent of truth tellers and 87 per cent of liars could be correctly classified on the basis of this factor alone.

'These are high accuracy rates and will be difficult to exceed by any traditional verbal, nonverbal or physiological lie detection tool,' Vrij's team said. 'In fact, we would certainly expect such tools to fare worse.'

Another distinguishing factor was the perspective of the drawing. Fifty-three per cent of truth-tellers penned a drawing from their own first-person perspective at the scene; 47 per cent opted for a birds-eye view. By contrast, 81 per cent of liars went for the birds-eye view and just 19 per cent for the first-person perspective.

The researchers said theirs was the first study of its kind and they acknowledged many questions were left unanswered. 'We believe that our findings justify taking drawing seriously as a lie detection tool, and that it will encourage researchers to carry out drawing research.'

ResearchBlogging.orgVrij, A., Leal, S., Mann, S., Warmelink, L., Granhag, P., & Fisher, R. (2010). Drawings as an innovative and successful lie detection tool. Applied Cognitive Psychology DOI: 10.1002/acp.1627

Post written by Christian Jarrett (@psych_writer) for the BPS Research Digest.


Anonymous said...

Can this be legitimately applied to the courtroom? Forcing a suspect to draw his way out of guilt seems difficult to do, right?

Unknown said...

Yes, I think you're right it's far too early for this to be introduced into the courtroom. But perhaps one day it could be developed into a useful police tool?

Anonymous said...

In the meantime, you simply use the drawing to unravel the defendent's story to get him to betray his lie. No need to submit the drawing in court.

Rob said...

But surely if this finding becomes well-known, liars will be able to adjust their behaviour accordingly? (i.e. they'll remember to include the agent, and will be more likely to draw the scene from a first-person perspective.) So the more this technique is used, the less effective it will be.

Unknown said...

Hi Rob
I think you're right, that must be a risk. However, the authors point out that there are some advantages to drawing as a lie-detection tool that suspects won't be able to get around. For example, imagine you lied about being in a certain place. If you gave a verbal account you could simply omit certain details - for example, you could omit to say where you were standing in relation to other people. But imagine you're asked to draw the scene - you have no choice but to reveal your perspective and the position of others. The other thing to say is just that this is obviously early days - the researchers are merely raising this as a potential avenue for future research. Perhaps there could be ways to develop this as a tool that would be difficult for guilty guilty suspects to counter.

Vaughn Vaughn said...

First, these are fake liars, not real liars.

Second, they may be incompetent liars - are police talented liars?

Third, you're getting 20% of the truth-tellers pegged as liars in one criteria, and 47% in the other. As for the liars, 13% are defined as truthiness candidates in one test, and 19% in the other. No mention is made of cumulative results, but you have a minimum of 33% false results. Better than flipping a coin, but if you had more intelligent & more qualified liars, it could even go the opposite direction.

So, about as effective as fortune tellers.

Forth, there is no pressure here on the truth-tellers nor the liars. Was any reward given to successful liars - like not getting thrown into jail?

Post a Comment

Note: only a member of this blog may post a comment.