Students in a recent experiment correctly identified 33 out of 38 classmates as liars or truth-tellers based on the story they told. Even more surprising is the fact that all studies using this new false-story-detection technique have shown greater than 70% accuracy. The leading application, of course, is police questioning.

Imagine changing our approach to evaluating witnesses in the courtroom, where eyewitness evidence is notoriously unreliable.

The new procedure would de-emphasize questioning that breaks up the stream of consciousness, and rules of evidence that confuse jurors with legal technicalities. Those devices would be replaced with story-telling. The stories begin with a “base line” measurement of someone’s conversational style such as their description of a wedding or graduation; and follow with the core narrative – first in chronological order, then in reverse chronological order.

The theory is that liars recite their pre-baked script by rote, and look like a denuded tree; whereas truth-tellers present 20-30% more detail and have the appearance of a tree in bloom. Honest reporters stumble and free-associate more than liars.

When we consider that no other assessment technique such as visual observation or the polygraph has proven to be statistically significant in ferreting out prevarication; it may be well-worth experimenting with this story-telling formula on the witness stand.