Deception Detection: The Truth About Lie Detectors

I am proud to welcome Dr. Katherine Ramsland as a guest on my blog. Dr. Ramsland has published 33 books and teaches forensic psychology and criminal justice at DeSales University, where she chairs the Social Sciences Department.  Among her books are Inside the Minds of Serial Killers, The Human Predator (a history of serial murder), The Criminal Mind, and The Unknown Darkness, with former FBI profiler Gregg McCrary.  In April, she will published The Devil’s Dozen: How Cutting Edge Forensics Took Down 12 Notorious Serial Killers.  She has also written a series of books to clarify facts about investigations, notably The Forensic Science of CSI, the Science of Cold Case Files, The CSI Effect, and True Stories of CSI.  Later this year, she offers The Life of a Forensic Scientist, with Dr. Henry Lee and The Forensic Psychology of Criminal Minds. Dr. Ramsland writes:

The New York Times ran a commentary recently that noted the use of psychological evidence in serial procedurals.  “The Mentalist” is one of the most popular shows on TV now, and “Lie to Me” has an intriguing premise about rare people who are “naturals” at spotting liars.  Yet research indicates that there’s no simple formula for catching a liar.  Even many people with repeated exposure to deception perform no better than chance when judging deception, but they can slightly improve their skills with solid observation and sophisticated techniques.             

A popular notion is that lying requires more effort than truth-telling, so it produces such physiological signals as a heightened pulse rate, dilated pupils, twitches, and certain facial expressions – especially when the stakes are high.  However, truthful but anxious people may also display such symptoms, while lying psychopaths may not.

Accuracy lies in questioning persons of interest long enough to observe their default behaviors.  People who feel anxious usually either freeze or defend themselves, thus displaying behaviors of discomfort.  While there are no hard-and-fast rules, the types of behaviors that can signal discomfort, and thus potential deception, include:

overgeneralizations, deflections, and increased vocal pitch

speech hesitations and pauses, a lack of spontaneity

an increase in number of shrugs, blinking, and nervous habits

changes in the eye pupil

venting the body, like pulling a shirt or collar away

feet pointed toward an exit

blanching, flushing, sighing

reduced use of hand gestures

These behaviors occur more often in those with motivation to deceive–possibly because they are trying to plan and control what they say.

Statement analysis is a common tool for interrogations.  An investigator asks an open-ended question, “What happened?” and leaves the person to fill in all the blanks.  The subject picks the starting and ending point.  Statement analysis focuses on several things: what’s said about events leading up to a crime, the crime itself, and what’s said about the aftermath.  Investigators watch for the distribution of detail in each area, and note whether subjects provided more information than requested or skipped something crucial. Also, a change in tone or speed of delivery can reveal their comfort (or not) with what they’re saying. 

A similar method called Criteria-based Content Analysis closely examines how an incident is retold, comparing it against the typical method of recall in a truthful session vs. fabricating a supposed recollection. 

Computer software known as Linguistic Inquiry and Word Count (LIWC) analyzes written content, derived from the statement analysis approach, and looks for three markers: fewer first-person pronouns, more words that convey negative emotion, and fewer exclusionary words (except, but).  The software has been more effective than human judges, but the accuracy rate is still only about 67 percent. 

The polygraph in use today is a compact portable device that measures three or four key involuntary physiological responses to questioning: skin conductivity, abdominal and chest respiration, blood pressure, and heart rate.  Some questions are designed to establish baseline responses, some are neutral, and others attempt to register “guilty knowledge,” or at least a sense that the person knows something that confirms him or her as a suspect.  However, despite claims by examiners, the accuracy rate by disinterested evaluators is not high enough for admissibility.

Even less accurate is the Psychological Stress Evaluator (PSE), sometimes referred to as a Voice Stress Evaluator.   Supposedly, during a lie, the voice reaches a higher pitch than when someone is telling the truth.  While the PSE does measure variations in emotional stress, that’s not necessarily indicative of deception. 

Psychiatrist Lawrence Farwell developed the Brain Fingerprinting process, based on the notion that all experiences, including a crime, are stored in the brain. The electrical activity of a suspect’s brain is monitored with sensors on a headband attached to a computer, while the subject is exposed to words or images that are both relevant (“probes”) and irrelevant to the crime. Certain information would be meaningful only to the actual perpetrator and would include such items as what was done to a victim, where the victim was taken, items that were removed from the victim, and items that might have been left at a scene. The subject would not see this list until the test itself was performed.  Irrelevant stimuli might include a different type of weapon, the wrong landscape, a different MO, or acts not performed during the commission of the crime.  

Probes are known only to the investigators, the test-maker, and the perpetrator.  If the brain activity shows recognition of relevant stimuli-a distinct spike called a MERMER (memory and encoding related multifaceted electroencephalographic response) – then the subject has a record of the crime stored in his or her brain.  Innocent people will display no such response to crime-relevant stimuli.  To strengthen the results, Farwell might test the suspect’s alibi for the time of the crime, by devising a scenario to test to see if the brain has a record.

At the University of Texas Southwestern Medical Center, fMRI scans are used to detect differences in neural activity between lying and truth-telling.  In the experiments, subjects were paid to perform a “theft” of one of two items (either a ring or a watch) and conceal information from the researchers.  First, each was asked neutral questions while being scanned, as well as questions about minor wrongful deeds commonly committed.  This way, the researchers could identify typical neurological patterns during truthful responses.  Then each subject responded to questions in a way that was truthful about an object he or she did not steal, but deceptive about the stolen object.  The rate of accuracy for the fMRI was around ninety percent.  Apparently, the trick lies in scanning brain regions that activate to suppress information and resolve internal conflicts; these regions are quiet when the person is telling the truth.   

There is as yet no “one size fits all” signal in the neurocircuitry that a person is lying, but it does appear that brain scans are better at revealing “tells” than is watching someone fidget and sweat under questioning.  Identifying the right combination of brain signals for a high rate of accuracy when a person lies or hides the truth is still in the future, but possibly not far away.  

Some researchers believe that certain people with high levels of emotional intelligence have a knack for spotting a liar; in fact, they can see certain signals that others cannot.  Dr. Paul Ekman and Dr. Maureen O’Sullivan float the notion that a few rare people are “naturals,” i.e., are highly accurate at knowing when someone is trying to deceive them.  (In fact, these researchers consult for Lie to Me.)  Often, these lie-detectors have jobs where it’s an important skill, such as law enforcement or psychotherapy.  When the stakes are high, such as with a violent crime or a threat assessment, they’re even better at it, because they’re more vigilant.  Ekman believes the best cues are found in the voice and face for deception about feelings, and find the best “hot spots” in gestures and words when a person lies about beliefs and actions.  Extremely slight gestures can “leak” emotional states that a person is trying to hide, providing a “tell” to a skilled and observant detector. 

However, other research contradicts the notion that certain select people are human diving rods.  Psychologists Charles Bond, Jr. and Bella DePaulo ran a large-scale study and found that lie detection is not about the observer but the observee.  A person’s perceived credibility plays a strong role in whether someone judges him or her to be deceptive.  That’s not necessarily because a person is honest; it’s because they comport themselves in a credible manner.  Participants in the study more often believed liars with high credibility ratings than truth-tellers who were perceived as low in credibility.  When Bond and DePaul evaluated numerous other studies about deception, they realized that individual differences among judges of deception hovered near the same rate as chance (50%).  No one appeared to have an innate advantage.  No “naturals” stood out.

In the real world rather than a lab, lies are often identified in context, when compared over a period of time to other behaviors or narratives.  The judgment generally involves a number of factors taken together, not just a person’s response to some questions at the time a lie is told, or their pupil contractions or fidgeting.

See also: Serial Killers and the Writers Who Love Them: Facts about Popular Myths by Dr. Katherine Ramsland

add to del.icio.us : Add to Blinkslist : add to furl : Digg it : add to ma.gnolia : Stumble It! : add to simpy : seed the vine : : : TailRank : post to facebook

Follow

Get every new post delivered to your Inbox.

Join 5,675 other followers