The New Science of Lie Detection

lie detection quantified impressions e1384443355850

“I’m on my way”, “Sorry, my phone died”, “I’m fine”. Everyone lies daily. According to Pamela Meyer, author of LieSpotting, we are lied to anywhere between 10-200 times every day. Strangers lie up to 3 times within the first ten minutes of meeting each other, and even babies lie by faking a cry just to see what kind of attention they might receive. But, even though we all lie, we’re not very good at identifying the lies being told to us. The average person can detect a lie only about 54% of the time. Our instincts tell us to look for non-verbal cues such as shifty eyes, unnatural body movements, or a shaky voice, but clearly instinct is not enough. New technologies have evolved that are able to more accurately detect if someone is lying (such as polygraph tests or MRI scans) – Google has even created a “Lie-Detecting Neck Tattoo”, a small sensor that measures changes in galvanic skin responses to detect signs that someone might be lying.

Now, lie detection is starting to evolve to include an analysis of thoughts and actions that subjects may not even be aware of. According to Leonard Mlodinow, author of the book Subliminal: How Your Unconscious Mind Rules Your Behavior, our unconscious mind controls the majority of what we say and do. In fact he states, “Some scientists estimate that we are conscious of only about 5 percent of our cognitive function. The other 95 percent goes on beyond our awareness and exerts a huge influence on our lives.” What this means is that if we can analyze our subconscious communication we can detect deception with greater accuracy than traditional methods. The subtle nuances in the words we use and the nonverbal communication we deploy indicate so much about our personalities, our thoughts, our intentions, and can give away the lies we tell.

After performing quantitative communication analyses on thousands of pieces of communication, from financial earnings calls to a corpus of both fake and genuine online reviews, we have identified the subconscious indicators associated with honest and deceptive language. The primary markers of deception are:

  • The number [or type] of pronouns used
  • The overall sentiment behind the words
  • The complexity of how words and sentences are placed together
  • The length of a statement

For example, using our proprietary language processing technology to analyze trustworthy and fake reviews, we found that fake reviews offered 43% more insightful language, were 18% less positive, and 9% longer. We believe data to be the next evolution in lie detection, and we are excited to continue our research.