Week Seven/Eight - Detecting Deception Flashcards
Defining deception key components?
- intent
- no forewarning
- success is not relevant
Types of deception?
- falsifications
- distortions / exaggerations
- omissions
3 ways to (try to) catch a liar?
- Physiological / formal techniques
• polygraph etc.
• thermal imagining
• brain fingerprinting - Nonverbal / body language cues
• believed & actual - Verbal cues
• “content” of the lie
• active interviewing
The Polygraph
Measures: – Heart rate – Blood pressure – Breathing (rate & depth) – Galvanic Skin Response
Detects physiological changes
– assumes that reactions are linked to lying
• Expected changes include: – Increased skin conductance – Increased blood pressure – Decreased respiratory activity – Decreased blood flow to fingers due to emotional response to lying (guilt, fear) which leads to the above
Polygraph techniques?
- Control (comparison) Questions Technique (CQT)
2. Guilty Knowledge Test (GKT)
- Control (comparison) Questions Technique (CQT)
• Compare responses for:
– Neutral questions
• establish baseline
– Relevant questions
• related to the specific crime under investigation
– Control questions
• related to the crime / “naughtiness” in general
• intended to embarrass and evoke arousal
Problems with CQT?
• Assumes control > relevant for innocent suspect (not convincing)
• Role of examiner
– can’t scare suspect too much or too little
• Ethical question of deceiving suspects
– Inadmissible evidence!
- Guilty Knowledge Test (GKT)
– Based on recognition response (not emotion)
– Responses triggered by personally relevant stimuli
• e.g., hearing own name at role call or party
– Increased EDA; decreased heart rate
• Questions are details that only guilty P would know
– e.g., ‘Where was the body found?’
• Several plausible alternatives for each question
– e.g., ‘in the kitchen’, ‘in the bedroom’ etc.
• Suspect gives same response to each
– e.g., ‘No’.
• If guilty, larger reaction to correct (recognised) alternative
Advantages of GKT?
• Better ethics, theory, validity than the CQT
Disadvantages of GKT?
Limited to cases where:
• Guilty suspect has the relevant knowledge
• Innocent suspect doesn’t
The Polygraph: Does it work?
Claims of up to 99% accuracy
– Dodgy stats: Manipulate ‘accuracy’ result via
comparison structure
need to look at guilty & innocent separately
CQT and GKT (false positives and negatives)
CQT = prone to false positives (saying innocent person is lying)
GKT = prone to false negatives (say telling truth when they aren’t)
Countermeasures to the Polygraph?
Increase response to baseline/control items
– Physical (e.g., bite tongue)
– Mental (e.g., imagine being slapped)
Decrease response to relevant items
– Mental (count sheep, count backwards)
Overall, very effective
– 50% beat CQT after 30 mins training
– Only 12% of P detected (physical countermeasures)
• Less effective for some techniques (“target” items)
Countermeasures with target items
– Used with GKT
– Administrator gives a list of items to memorize
– These appear as alternatives on the GKT
• An innocent person will recognize the target
items but not the guilty items
• Non-recognition of target items = possibly using
countermeasures
Summary of Polygraph testing?
– CQT (false positives)
– GKT (better, but false negatives)
– Countermeasures
– Less effective for some techniques (“target” items)
Thermal Imaging?
Based on instantaneous anxiety response
- increases in heat
Accuracy rate: 83% (Comparable to polygraph)
Quick, little training required
- May be suitable for mass-screening (e.g., airports)
Prone to false positives
“Brain-fingerprinting”: ERP analysis
• Event-related potentials (ERPs)
– Measures changes in P300 waveform
– Reaction to relevant item in series of irrelevant items
– e.g., own name in random series of names
• Orienting technique
– Use with GKT
– Increased P300 reaction to recognised item?
Brain-fingerprinting: ERP analysis pros and cons?
Pros:
– Accuracy rate: 82-88%
– High face validity
Cons:
– Same problems as other orienting techniques
(Applicability; culprit’s knowledge of answer;
recognition ≠ guilt)
– Validity in naturalistic settings
“Brain-fingerprinting”: fMRI scanning
• Functional magnetic resonance imaging
(fMRI)
– Measures changes in blood flow
– Neural activity in brain and spinal cord
Changes in:
Prefrontal cortex activity
• Monitor & control/inhibit own responses
• Monitor reactions of listener
Anterior cingulate cortex activity
• Risk/reward assessment
• Anticipation of consequences
“Brain-fingerprinting”: fMRI scanning pros and cons
Pros:
– Accuracy rate: 78-93%
– Measures central cf. peripheral nervous system
– Taps into complexity- and control-related markers
Cons: – Lack of replication of results • Different patterns for different P & types of lies (spontaneous vs. rehearsed) – Cost – Ethical and legal guidelines
What do people think are signs of deception?
Non-verbal:
• Signs of nervousness
• Gaze aversion (i.e., looking away)
• Increased body movements, fidgeting
Verbal:
• Reduced detail in speech content
• Errors in speech content
How good are people at detecting deception?
Ordinary: 57%
Police: 56%
Secret Service: 64%
CIA Agents: 74%
Why detecting deception is no better for professionals?
Relying on the wrong cues
• e.g., increased gaze aversion and fidgeting are NOT signs of lying
Giving incorrect explanations for behaviours
• Acting nervous = lying or worried about proving
innocence
• Avoiding eye contact = guilty? shy? cultural reasons?
Emotional markers of lying?
- Higher voice pitch
* Micro-expressions
Cognitive markers of lying?
Mentally harder to lie so:
• More errors
• Less detail
• More structure
Behavioural markers of lying?
- More rigid
* Less expression
Can lie detection be improved?
Train people to look for specific markers
• Ignore gaze aversion; monitor pauses & errors; look
for foot & leg movement; etc
Active Interviewing Techniques
Active Interviewing Technique?
- Increase cognitive load
- Unanticipated questions
- Strategic use of information
Increasing cognitive load (AIT)
- Reverse order
- Maintain eye contact
- Increase amount of information generated
• Example of “evidence” from another witness in unrelated
case (Leal et al., 2013) - Different reporting modes
Potential drawbacks of increasing cognitive load?
May increase amount of inaccurate information generated by truth tellers during interview
• Especially when memory for event is poor
Effectiveness might depend on working memory capacity
(WMC) of person being interviewed
• Ineffective for high-WMC liars?
• Problematic for low-WMC truth tellers?
Unanticipated questions (AIT)
- Liars prepare well for interviews
- Benefit reduced for unanticipated Qs
- Cognitively easier for truth tellers than liars
• People anticipate some Qs
“Tell me what happened?” “What topics were discussed?”
…but not others (e.g., spatial and temporal details)
• “In relation to the front door, where was the nearest table?”
• “What order were the topics discussed in?”
• “Who finished their meal first?”
Increases errors and lie detection (up to 80% accuracy)
