Discussions
buergoy
Chapter2-LieDetection.pptx
Lie Detection Chapter 2
Chapter 3: Lie Detection
In this chapter
The Complexity and Pervasiveness of Deception
Can We Tell When Others Are Lying?
The Polygraph
From Physiological Arousal to Cognitive Load
Some Emerging Technologies for Lie Detection
Lying
Multi-faceted and complex
Lie, mislead, misinform, trick, fake, betray, fool, cheat, con, delude, deceive, distort, exaggerate, pretend, feign, falsify, presenting “alternative facts”
Irony, jokes, sarcasm, mistakes – not lies…at least according to adults
All are violations of the truth
Complexity and Pervasiveness of Deception
Common in social interactions
An adaptive behavior
Allows human beings to have harmonious group interactions
Many types of lying and for many different reasons
What Do You Think?
Uncovering the truth is a central goal of the legal system.
Can We Tell When Others Are Lying?
Lie detection accuracy is almost equal to chance
Detection is poor even among those with training
Mistaken liar’s stereotype
Flawed interpretation of verbal and nonverbal behavior often fuels confirmation bias
Despite decades of research effort to maximize the accuracy of deception judgments, detection rates barely change.
There is no “Pinocchio response,” no dead giveaway, no obvious visible signal that someone is telling a lie.
6
Can We Tell When Others Are Lying?
Are some people able to detect lying better than others?
College students slightly better at distinguishing true from false confessions; police detectives were not
Detectives slightly more confident about accuracy of their judgments and showed a bias toward judging false confessions as true
Detectives’ tendency to infer guilt increased with job experience and interrogation training
Types of Lies
Commission
Saying things that are not true
Omission
Leaving out things that are true
Antisocial
To protect or benefit the self
Prosocial
To protect or benefit others
Do polygraphs detect lies?
The Polygraph
Theory: Lying causes physiological arousal
1960s–1980s: Polygraph was lucrative business
Polygraph Act of 1988
9/11/01: Resurgence of polygraph use and surrounding controversy
Under what circumstances should the legal system trust the results of a polygraph exam?
The Polygraph
Lie detection device
Assumption:
Lying will cause physiological changes/arousal
Controversial technique with a controversial history
Still routinely used for employment screening
police departments, FBI, CIA, DEA, Secret Service
Polygraph Protection Act (1988)
Do polygraphs detect lies?
During questioning, polygraphs measure:
Breathing
Perspiration
Heart rate
Blood pressure
Polygraph Process
Depends on the measurement of relative arousal
Measures differences in states of arousal
Polygraph must be used with questioning
Comparison Question Test (CQT)
Questioning Procedure: CQT
Comparison Question Test (CQT)
Question Types:
› Comparison Questions – “known lie” questions
Behaviors that are uncomfortable for suspects to discuss but not directly related to the crime
Deliberately broad so that anyone answering “no” would be lying
Might be instructed to lie to these questions
Examples: Have you ever done anything dishonest?; Have you ever tried to get someone else in trouble?
› Relevant Questions (about the crime in question)
− Did you kill Joe?; Did you rape Lisa?; Did you steal the money from the safe?
Questioning Procedure: CQT
Not deceptive if:
Individual reacts more strongly to the comparison than the relevant questions
Deceptive if:
Individual reacts more strongly to the relevant than comparison questions
Better at classifying liars (~83-89%) than truth tellers (~53-76%)
Heavy burden on skill of the examiner
Must come up with good comparison/control questions!
Questioning Procedure: Positive Comparison Test
Sub-category of CQT
Suspects are asked the same (relevant) question twice and are instructed to…
Lie once
Tell the truth once
Questioning Procedure: Positive Comparison Test
Interrogator: “Did you kill John Doe?” Suspect: “No.”
Interrogator: “Did you kill John Doe?” Suspect: “Yes.”
Uses the relevant question as its own control
– Allows for direct comparison of responses to the same question
Polygrapher
Convince the suspect that polygraph is effective
Design good comparison questions
Interpret results
Weaknesses of the Polygraph
Emotionally non-reactive people!
Psychopaths, sociopaths
Assumes that innocent people won’t react strongly to questions about a crime
Innocent suspect may be nervous about being a suspect
Convince suspect the polygraph is flawless
Secret background information
Bogus pipeline – anxious of being caught
Weaknesses of the Polygraph
Lack of standardization – considerable variability
Subjectivity in scoring – differences
Determination of lying is a “judgment call”
Use of countermeasures
People can try to beat polygraph
The Polygraph Controversy
Not falsifiable
You can always explain away results not consistent with the outcome
Labeled deceptive – later confesses, convicted = confirmed
Labeled deceptive – fails to confess, acquitted = they beat the polygraph!
Labeled truthful – fails to confess, examiner will never know!!!
Research on the Polygraph
Claim high accuracy rates (90-100%)
Laboratory studies
Mock crime, ½ commit crime, all deny
Strength = ground truth known
Field studies with actual suspects
Strength = realism and high arousal
Weakness = criterion of accuracy is subjective
Hit and Misses in Forensic Testing
The results of polygraph testing and some other types of testing are designed to provide yes-or-no answers to specific questions (e.g., Is a person’s response to a question the truth or a lie? Do these two fingerprints come from the same person? Did a likelihood-of-violence assessment correctly predict that a person would become violent?).
As shown in the table, the yes-or-no answers produced by such tests can be classified into four categories: true positives, true negatives, false positives, and false negatives.
Correct answers (true positives and true negatives) are called hits, and incorrect answers (false positives and false negatives) are called misses.
For polygraph testing a positive result is an indication that the person is lying in response to a question, and a negative result is an indication that the person is telling the truth.
23
Research on the Polygraph
Review of research
109 studies of CQT
In controlled conditions, polygraphers can:
Catch 77% of guilty suspects
Exonerate 70% of innocent suspects
Still missing 21% of guilty and falsely accusing 16% of innocent people; some unclassifiable
Research on the Polygraph
Inflated accuracy rates in lab studies?
Consequences for lying are very low
Little motivation to tell successful lies
Unrealistic and artificial situation
Participants are not experienced criminals
No experience with countermeasures
Simple lies not realistic
Polygraph administrator blind (most likely not the case in real world)
Research on the Polygraph
Categories of empirical studies:
Laboratory studies use mock juries and crimes
Field studies use actual suspects
Accuracy rates inflation causes:
Lies told in controlled studies tend to be more simple than those told about real crimes
Limited or no subject experience or training in countermeasures
Low consequence for lying
27
Research on the Polygraph
Under carefully controlled conditions, (some would say, “under highly unrealistic conditions”), polygraphers appear to be able to catch nearly three-quarters of deceptive suspects.
Is this rate of accuracy strong enough to have practical utility for the legal system? Why? Why not?
Most polygraph tests given during criminal investigations not done under controlled conditions.
Cannot tell if above-chance accuracy rate points to identification of guilty suspects.
28
Research on the Polygraph
National Academy of Science report conclusions:
Scientific basis of comparison question techniques (CQT) was weak, extant research was of low quality, claim of high accuracy was unfounded, and error rate unknown.
Most polygraph tests given during criminal investigations not done under controlled conditions.
Cannot tell if above-chance accuracy rate points to identification of guilty suspects.
29
Can the polygraph be used to induce confessions?
Polygraph as Coercion
Used to induce confessions
Presented as opportunity to prove innocence
Attempts to avoid test may be viewed as guilt
Not avoiding viewed as innocence
Results are sometimes manipulated by polygrapher to coerce suspect
Liar!
I did it!
Legal Status of the Polygraph
23 states banned polygraph evidence in courts
Many allow in special cases; Only 1 state (New Mexico) routinely permits
Results may be especially important before trial
E.g., establish credibility, elicit confession
Legal system’s concerns about polygraph:
Lack of consensus in scientific community
Diminishing role of jury
Jurors and the Polygraph
Jurors find results persuasive
Results can change outcome of a trial
Effectiveness of results based on:
Persuasiveness of polygraph testimony
Strength of other evidence
Sophistication of jurors
Instructions from judge
Expert and Nonexpert Opinions About the Reliability and Usefulness of Polygraph Test Results as an Indicator of Truthfulness
Alternatives to the Polygraph
Criterion-Based Content Analysis (CBCA)
Microexpressions (like Lie to Me)
Guilty Knowledge Test
Cognitive Load Technique
Strategic Disclosure of Evidence (SUE)
fMRI/EEG
Thermal imaging, eye movement
Guilty Knowledge Test (GKT)
Uses polygraph equipment but does not attempt to detect lies
Detects criminal/guilty knowledge:
– Recognition of scenes/events from the crime will elevate physiological arousal
GKT: Example
You know that Nicole has been found murdered, Mr. Simpson.
How was she killed? Was she drowned? Was she hit on the head with something? Was she shot? Was she beaten to death? Was she stabbed? Was she strangled?
Research on the GKT
Seems promising!
Higher rate of correct classification than polygraph
97% of innocents and 88% of the guilty
8 laboratory studies
96.7% innocents correctly classified
88.2% guilty correctly classified
Field studies
97% innocents correctly classified
76% of guilty correctly classified
Limitations of GKT
Must be enough factual evidence from crime scene
Facts must be known only to the perpetrator
Perpetrator must remember the crime scene
Limited applicability
Human Beings as Lie Detectors
“I’d know a false confession if I saw one” (Kassin et al., 2005)
Prison inmates gave true and false confessions on video
Police detectives and college students judged the truthfulness of the confessions
Who does better?
College student accuracy = 59%
Police accuracy = 48%
– But were significantly more confident in their judgments
Police demonstrated a bias toward lies
Are experts better than laypeople?
Lie Bias – the belief that people are usually lying (police investigators)
− Bias increased with interrogation training and job experience
Truth Bias – the belief that most people are usually telling the truth (lay people)
Does training improve our abilities?
Training doesn’t improve deception detection
Increases confidence and justifications for judgments
Are experts better than laypeople?
Why aren’t police better than laypeople at detecting deception?
Believe that people feel guilt/shame when lying
Look for signs of shame/guilt (e.g., averting gaze, nervousness)
Liar’s Stereotype: There is no Pinocchio's Nose
Behavioral indicators are flawed indicators of deception
» 70% of people believe liars (Bond & DePaulo, 2008):
Avert their gaze
Squirm
Touch their bodies more
Stutter
Lie Detection: The Science
DePaulo et al. (2003): 116 studies; 158 total cues
Words, not behaviors, distinguish liars from truth- tellers
Liars speak with higher-pitched voice
Liars stories are less plausible
Liars stories follow a structured production
Truth tellers give more sensory and perceptual details
Truth tellers give more spatial and temporal details
Truth tellers speak for longer period of time
Liars are less vocally expressive
Confirmation Bias and
Deception Detection
Police overconfident in their deception detection abilities
Develop a belief that suspect is lying during the interview
Then interrogate suspect using high pressure interrogation techniques
Suspect’s behavior and the confession confirms officer’s initial presumptions of guilt
SPOT = Screening Passengers by Observation Techniques
Step 1
Behavior
observation
BOO
Step 2
SPOT referral
BOO
TSO
,If
Step 3
LEO referral
Avoids Eye Contact
T
Excessive Head- urning
Repetitive Face Touching
Y . Exaggerated crying
awnmg, Laughter or Chatt;r
Jumpy "Ada m's Apple"
Excessive Fidgeting
or Leg Shaking
Excessive Shuffling of Feet
Who gets pulled aside for ‘selective
screening’?
The Way TSA Agents Profile You At The Airport Is Not Based On Legit Science, Report Says
A new report finds no "valid evidence" for the TSA's $186 million "Behavioral Detection Assessment" program. But the agency says it will keep doing it anyway.
"TSA does not have valid evidence that most of the indicators in its revised list of behavioral indicators can be used to identify individuals who may pose a threat to aviation security," the report concludes.
The GAO analysis "found that 98 percent (175 of 178) of the sources do not provide valid evidence" and only one behavior (left unidentified in the report) has two research articles that offer support for it being a suspicious thing to do at the airport.
Despite the report 's damning findings , it appears unlikely to change the TSA's operations.
"While we respect the opinions of our GAO colleagues , TSA remains steadfast in the effectiveness of behavior detection and committed to training our officers in these widely used techniques ," TSA spokesperson James Gregory told BuzzFeed News by email.
Joshua Lott I Getty Images
There's no evidence that the TSA's list of "suspicious behaviors" - including yawning, whistling, staring, and complaining about security too much - actually help spot terrorists, a
0. Sections =
mbe tllasbington ost
Dcmocrucy Dies in Durkncss
GAO says there is no evidence that a TSA program to spot terrorists
is effective
In a report to be presented to a House subcommittee Thursday, the Government Accountability Office says there is no evidence that it's effective for Transportation Security Administration officers to scan crowds for telltale signs someone might be a terrorist.
Critics have suggested that the stress and exhaustion that often accompany air travel are too easily misread as suspicious behavior.
The GAO report recommends that Congress stop funding for the program, which has cost more than $878 million since its launch in
2007.
Times Square bombing suspect Faisal Shahzad was arrested after he boarded a plane headed for Dubai, though the government is spending millions each year on a program that's supposed to spot terrorists before they reach the gate. As CBS News chief investigative correspondent Armen Keteyian reports, the program doesn't seem to be working.
There's a hidden layer of airport security most people don't know about. It's called "behavior detection," and involves specially trained Transportation Security Administration employees whose primary mission is to spot terrorists.
They look for unique facial expressions and body language that may identify a potential threat. About 3,000 of these officers work at 161 U.S. airports -- costing taxpayers nearly $200 million in 2009. This year, the TSA asked Congress for $20 million more to expand the program.
60 Minutes: Expert Says TSA Screening Is Theater
But CBS News has learned that the program is failing to catch terrorists. It's never even cau g ht one.
Facial Action Coding System
Paul Ekman and Wallace Friedman, in the 1970s, set up a classification system for human facial expression, part of which is lettered in this photo
The Facial Action Coding System identifies 43 sets of muscles, which often work in tandem to facially express emotion
From Physiological Arousal to Cognitive Load
Initial polygraph system
Stressed-based, arousal-based lie detection
Fooled by emotions and anxious behaviors
Cognitive load approach
Mental resource drain during task performance
Lying requires performance of simultaneously demanding cognitive tasks
fMRI detects brain activity where lying occurs
Different types of lies and truth telling produce different brain activation patterns
Brain Activation in fMRI Studies of Deception
Red areas are typically found to be active in neuroimaging studies of deception:
A, middle frontal gyrus; B, inferior parietal lobule; C, ventrolateral prefrontal and insular cortex; D, anterior cingulate and surrounding medial prefrontal cortex.
The red outlines indicate areas with relatively high rates of activation — that is, areas that tend to be active across many different cognitive tasks, making it uncertain whether activity found in those areas in studies of deception is actually associated with deception.
57
From Physiological Arousal to Cognitive Load
Brain
fMRI detects brain activity where lying occurs
Different types of lies produce different brain activation patterns
Limitations of fMRI
Not portable, expensive, and requires patient motionless behavior
Slow readout compared to actual speed of neural transmissions
United States v. Semrau (2010); Federal court in Tennessee excluded fMRI expert testimony
From Physiological Arousal to Cognitive Load
Limitations of fMRI and lying:
Not portable, expensive, and requires patient motionless behavior
Slow readout compared to actual speed of neural transmissions
United States v. Semrau (2010); federal court in Tennessee excluded fMRI expert testimony
From Physiological Arousal to Cognitive Load
Another method of seeing lies in the brain
EEG detects when lying associated brain activity occurs (event-related potentials)
Reads neural impulses continuously in milliseconds
May detect intent to suppress truthful response and replace it with lie
Brain fingerprinting (Brainwave Science/iCognative)
Limitations of brain fingerprinting
Shares limitations of GKT
Lacks accuracy confirmations by other scientists
Statistical measurements not available to evaluate validity
Emerging Technologies for Lie Detection
Newer methods:
Using high-definition infrared thermal imaging to monitor miniscule shifts in human face heat
Eye movement memory assessment (National Center for Credibility [NCCA])
Laser Doppler vibrometry
Deception Analysis and Reasoning Engine (DARE)
Telling lies from the eyes:
Searching for clues to deception on eye movements, changes in pupil size, and blinks per second
Tracking of subtle heat changes in skin near eyes
Thermal images of a participant’s face before telling a lie (left) and after telling a lie (right).
As reflected in these images, research results suggest that lying causes the temperature of the tip of the nose to drop and the temperature around the eyes to rise (red regions are colder than orange and yellow regions).
62
