Lie Detection Psychology & Law Lecture Slides PDF

Lie Detection  History of Lie Detection Idea that physiological reactions indicative of lying has long history China - suspects forced to chew handful of rice, if dry following assumed to be lying West Africa - people suspected of a crime pass a bird's egg to one another. Whoever broke the egg w...

Lie Detection  History of Lie Detection Idea that physiological reactions indicative of lying has long history China - suspects forced to chew handful of rice, if dry following assumed to be lying West Africa - people suspected of a crime pass a bird's egg to one another. Whoever broke the egg was accused of the crime, believing that their nervousness caused them to drop the egg  Early Precursors to Polygraph William Moulton Marston credited with early development (1938), which lead to polygraph Looked at measurable change in blood pressure as signs of lying Under pen name Charles Moulton created comic book character Wonder Woman, whose Lasso of Truth trapped those in it’s coils to reveal the truth How accurate are we? Bond & DePaulo (2006) – meta analysis of deception detection studies Average accuracy is 54% 47% for lies 61% of truths How accurate are we? Police vs students (Kassin et al., 2005) Had 10 inmates give a confession to their crime, and to a crime they did not commit Asked students and police officers to watch 5 true and 5 false confessions Asked to rate whether each person was telling the truth or a lie, and provided confidence on each from 1 – 10 Kassin et al. (2005) Results – Overall Accuracy 70 65 60 55 50 45 40 35 30 Overall Students Police Kassin et al. (2005) Results – Confidence 10 9 8 7 6 5 4 3 2 1 Overall Students Police Does Training Improve Deception Detection May depend on the type of Training Reid Technique relies on faulty assumptions about verbal and non-verbal cues to deception Training that debunks misconceptions & gives feedback on performance may improve accuracy (Porter, Woodworth, & Birt, 2000) Risks of Lie Detection Training Investigator bias: Training & experience increase the chance that police will believe a person is lying. Accuracy is not improved, but confidence is higher Dangerous Decisions Theory – initial judgments (gut reactions) about credibility can affect how later information is interpreted (confirmation bias) Verbal & Non- verbal Cues to Deception  Facial Cues to Deception Paul Ekman Pioneering work on facial expressions of emotion Six basic emotions recognized worldwide  Microexpressions Occur in a fraction of a second Can they reveal a lie? Debated in literature Which is the real smile? Which is the real smile? Are there facial cues to deception? Can you tell when someone is lying? How? Liar’s stereotype – belief that when people lie, they show facial specific facial and behavioural cues  Verbal Cues to Deception? ten Brink & Porter (2012) Studied 78 emotional pleas to the public for the return of a missing relative, ½ eventually convicted of murder (i.e., lying) Coded for speech, body language, facial expressions Results Liars used fewer words and more tentative words (higher cognitive load) Liars failed to show sadness, occasionally let happiness slip through  Cry Me a River (ten Brinke & Porter, 2012) Sample of 78 emotional pleas to the public for the return of a missing relative, half of whom were eventually convicted of murder Coded for speech, body language, and emotional facial expressions Liars used fewer words, but more tentative words Micoexpressions – very rare and didn’t distinguish What about physiological changes? Changes in respiration (speed, volume) Muscle tension Pupil dilation Pulse Blood pressure Individual Differences in Deception Detection Are some people better at detecting lies? What personality traits might be useful? Bond & DePaulo (2008) – meta analysis of 142 studies, almost 20,000 decision makers Mean accuracy only 54% No reliable, replicable individual differences Perceived credibility of the target may matter more than the ability of the perceiver Wizards O’Sullivan & Ekman tested lie detection abilities of more than 12,000 people Identified 42 lie detection “wizards” (out of 12,000) with 80% accuracy ‘Wizards’ can notice more cues, and more subtle cues than the rest of us Lying -- Can we identify this complex behaviour that is related to emotions? “The polygraph”  Polygraph Method Based on belief that deception is related to physiological changes: Increased heart rate Increased blood pressure Changes in breathing rhythm Perspiration Compare against baseline Three Types of Polygraph Test Relevant/Irrelevant Test (RIT) Control Question Test (CQT) Concealed Information Test (GKT)  1. Relevant/Irrelevant Test John Larson develops 1st systematic questioning procedure First, convince the suspect that the test is accurate with practice questions (investigator knows answers) Then ask irrelevant and relevant questions Looks at difference in physiological responses to relevant & irrelevant questions If guilty suspect denies involvement in crime, reactions to relevant question should be stronger Problems with RIT ◦ Little research, but agreement among polygraph researcher that RIT is inappropriate ◦Irrelevant question provides no control ◦High rate of false alarms – 71% (innocent assessed as guilty) (Horowitz, Kircher, Hons & Raskin, 1997) 2. Control Question Test Begins with a pre-test interview: Control questions are developed Examiner attempts to convince the suspect of the accuracy of the polygraph Explains how polygraph works, stressing impt of being truthful, that test detect lies Control/Comparison Question Test During test different types of questions asked alternately. 1. Irrelevant questions (their name, eye color) 2. Control questions (things many people will lie about) 3. Crime relevant (questions relevant to the crime)  Comparison Question Test Type of Question Purpose Example Irrelevant Used to obtain Are you left-handed? baseline Control Provoke anxiety, give Have you ever illegally measure of arousal downloaded not related to crime software? Relevant Measure reactions to Did you assault Sam the crime Smith on the night of May 30? Scoring Numerical scoring: 0 No diff between Relevant and Control 1 Noticeable diff 2 Strong diff 3 Dramatic diff Reactions to Relevant stronger: negatively scored Reactions to Control stronger: positively scored  Questioning the Validity of CQT David Lykken: ◦Some people emotionally nonreactive ◦Innocent can react strongly ◦Test is not standardized, quantifiable ◦Decisions are subjective ◦Opponents allege that test can be held in an interrogation-like fashion that elicits nervous response from innocent & guilty suspects alike Types of Polygraph Studies Laboratory studies (mock crimes) ◦ Ground truth is known ◦ Limited application to real-life situations Field studies ◦ Real-life situations and actual suspects ◦ Ground truth is not known Accuracy of CQT in Lab Studies Review of 97 studies using CQT 77% 70% 21% 16%  Validity of CQT Summarize these results Good at detecting real lies (~90%) Fairly good at detecting real honesty (~80%) BUT high false positive rate (15-25%)  Can you beat a polygraph? Try to give a false negative – raise your arousal during control questions Physical measures (tongue biting, foot tensing) Mental measures (counting backwards, meditation, visualizing) Chemical measures might work (e.g., Valium) One study found that 50% of participants using countermeasures were able to beat the CQT Psychopaths are not (reliably) able to beat the polygraph in lab studies 3. Concealed Information Test (CIT) Tests suspect on knowledge of crime that would not be known to an innocent person Asks suspects multiple choice questions, one option is correct Assumes if suspect is guilty, they will react strongly to correct info Concealed Information Test The victim was murdered. How was he killed? …was he drown …was he shot …was he beaten to death …was he stabbed …was he hit on the head with something …was he strangled Accuracy of CIT Very accurate at identifying innocent participants in lab studies (Lykken, 1998) ◦ Around 97% correctly identified (3% false positives) Less accurate at identifying guilty participants ◦ Around 88% correctly identified (12% false negative)  Limitations of CIT Must have sufficient number of crime facts available to construct questions Crime facts must not be public knowledge Guilty person must remember the details of the crime  Legal Status of the Polygraph  In Canada: Polygraph evidence NOT admissible (R v. Beland, 1987, SCC)  In US: Courts more divided Admissible in 19 states Discretionary admission in some, requiring mutual agreement by both sides for admission