Lecture 4: Profiling and Lie Detection (PDF)

Identification and Evaluation of Criminal Suspects I: Profiling and Lie Detection Associate Professor Helen Paterson Phone: 9036 9403 Email: [email protected] The Plan… Profiling  Background  What is criminal profiling?  The origins of criminal profiling  Different met...

Identification and Evaluation of Criminal Suspects I: Profiling and Lie Detection Associate Professor Helen Paterson Phone: 9036 9403 Email: [email protected] The Plan… Profiling  Background  What is criminal profiling?  The origins of criminal profiling  Different methods of profiling  How accurate are criminal profilers? Lie Detection  Background  Types of polygraph tests  How accurate are polygraph tests?  Current directions Learning Outcomes At the end of this lecture you will be able to:  Define profiling.  Describe the history of profiling.  Name, describe and critically evaluate different methods of profiling.  Define deception.  Describe what a polygraph is and what it measures.  Name, describe and critically evaluate three different types of polygraph tests.  Discuss current directions in lie detection Profiling Background “Offender profiling is the stuff myths are made of.” Grubin (1995) Profiling is probably what most people imagine forensic psychologists spend their time doing Very few people in Australia or UK are employed as profilers. What is Criminal Profiling? Criminal Profiling: “A technique for identifying the major personality and behavioral characteristics of an individual based upon an analysis of the crimes he/she has committed” (Douglas et al., 1986) Often includes information about demographics, personality, and behavioural characteristics Used in a range of contexts, but most commonly cases of serial homicide and rape. The Origins of Criminal Profiling Early attempts at Criminal Profiling: Late 1400s: A text was written by contractors to the Catholic Church for the purpose of accurately identifying and eradicating witches 1888: A police surgeon attempted to profile ‘Jack the Ripper’, a murderer who mutilated a number of women in London 1956: Most famous early case was that of Dr. James Brussel’s (1968) profile of the “Mad Bomber” The Origins of Criminal Profiling The FBI and Beyond 1970s: The development of a criminal profiling program at the FBI Today: Similar programs have been developed internationally Different Methods of Profiling 1. Deductive Criminal Profiling 2. Inductive Criminal Profiling  The FBI Approach: Organised-Disorganised Model  Statistical Profiling  Geographic Profiling 1. Deductive Criminal Profiling Deductive Criminal Profiling: Profiling the background characteristics of an unknown offender based on evidence left at the crime scene(s) by that particular offender Relies on logical reasoning… but the underlying logic can sometimes be faulty 2. Inductive Criminal Profiling Inductive Criminal Profiling: Profiling the background characteristics of an unknown offender based on what we know about other solved cases Relies on a determination of how likely it is an offender will possess certain background characteristics given the prevalence of these characteristics among known offenders who have committed similar crimes The FBI Approach First attempt to develop systematic approach was by Behavioral Science Unit in the 1980s Interviewed 36 convicted sexually motivated murderers. This led to the distinction between Organised and Disorganised crime scenes The FBI Approach FBI categorisation of sexual murders Organised Disorganised Planned in advance Spontaneous offence Specific victim targeted Depersonalises victim – is “personalised” Minimal conversation Controlled conversation Chaotic scene Crime scene shows signs Sudden violence to victim of control Little use of restraints Restraint used Body left in view at scene Body moved and hidden Evidence and weapon Evidence and weapon present removed The FBI Approach Once we can categorise crime as organised or disorganised we can then move to describe perpetrator  Organised – intelligent, socially competent, skilled job, sexually competent, lives with partner, follows details of crime in news, may move following crime etc  Disorganised – average IQ, immature, poor work history, sexually incompetent, lives alone, lives near crime scene, little interest in media, little change in lifestyle after murder etc The FBI Approach- Problems Little research has examined the model, but that which has been conducted raises serious doubts (Canter et al., 2004) Cannot account for offenders who display a mix of organised and disorganised features (Douglas et al., 1992) Statistical Approach Data collected from solved crimes and analysed using complex statistics The idea is that groups of actions will tend to occur together allowing one to reveal a typology Problem is that the analysis is only as good as the data fed into it. It is not always obvious what characteristics of a crime to record. Also, this type of analysis can be very unstable (statistically) if we do not have a large enough sample size David Canter Geographic Profiling Geographic profiling: Involves an analysis of crime scene locations in order to determine the most probable area of offender residence Assumes that offenders do not travel long distances from home to commit the majority of their crimes In many cases geographic profiles can be quite accurate (Rossmo, 2000) How Accurate are Criminal Profilers? Surveys of police officers indicate profiling is useful in investigations (Copson, 1995) Benefits of profiles include:  Provide a greater understanding of the case (60.9%)  Reinforce judgements about the offender (51.6%) Only 2.6% of respondents indicated that profiles led to the identification of the offender How Accurate are Criminal Profilers? Profiles often contain ambiguous advice and such advice can be interpreted to fit a wide range of suspects (Alison et al., 2003)  Examined the content of 21 profiling reports and found that 24% of the profiling opinions could be considered ambiguous  Ambiguous profiles can be interpreted to fit more than one suspect, even when the suspects are quite different from one another. How Accurate are Criminal Profilers? Experimental Evaluations: Pinizzotto et al (1990)  Gave solved cases (one murder, one sex offense) to profilers, detectives, psychologists, and students.  Profilers did better with the sex offense, but not the murder.  Profilers did not use the information any differently to other groups. How Accurate are Criminal Profilers? Experimental Evaluations: Kocsis et al. (2000)  Gave solved murder case to profilers, police, psychologists, students and psychics.  Psychologists were slightly better than police and psychics.  Some evidence that profilers were better than psychologists but very few agreed to participate (5).  Psychics were no good at all – relied on stereotypical views about murderers Deception Background Deception: “A successful or unsuccessful deliberate attempt, without forewarning, to create in another a belief which the communicator considers to be untrue” (Vrij, 2000) Three Ways to Catch a Liar Examine their physiological responses (this lecture and tutorial) Observe their nonverbal behaviour Analyse the content of what they say Polygraph: A Brief History 19th Century: Idea put forward by Lombroso, an Italian Criminologist Early 20th Century: Polygraph machine first used  1917: William Marston claimed to detect lying by measuring systolic blood pressure  1932: John A. Larson built a forerunner of the modern polygraph that measured pulse rate, blood pressure, and respiratory changes Polygraph: Not a “Lie Detector” Polygraphs measure physiological change, not lies. Assumption: Telling a lie is more stressful than telling the truth Polygraphs measure:  Blood pressure  Heart rate  Respiration  Sweating (GSR) WHY LEARN ABOUT POLYGRAPH TESTING???? American diary study: College students told 2 lies/day and community members told 1 lie/day. (DePaulo, 1996) Uses of Polygraph Helps in criminal investigations Verify a crime has occurred Pre-employment screening for security agencies and police Polygraph Use Around the World Polygraphs used to be quite common in the US, but now mostly used for investigative purposes (Honts & Perry, 1992) In Australia, UK and most of Europe the polygraph is not used as part of formal legal process – but may still be used in some contexts including security services Polygraph has been prohibited from use in criminal investigations in NSW since the Lie Detectors Act 1983 (NSW) was passed Types of Polygraph Tests There are three main types of polygraph tests: 1. Relevant/Irrelevant Test 2. Control Question Test (CQT) 3. Guilty Knowledge Test (GKT) 1. Relevant/Irrelevant Test Asks questions relevant to the crime and irrelevant to the crime Idea is that guilty people will respond more strongly when lying about the relevant questions Problem is that innocent person will know which are the relevant questions and will worry about their response to them...would lead to many False positive errors No longer used in law enforcement, but used for employee screening (e.g., drug use, honesty) ← Irrelevant Questions → Relevant Question 2. Control Question Test (CQT) Designed to overcome problems with the Relevant- Irrelevant Question test Includes three types of questions: irrelevant, relevant, and control The CQT Process – Phase 1 The CQT begins with a pre-test interview Yes/No questions are formulated and discussed with the suspect 1. Irrelevant questions: No arousal 2. Relevant questions: Discussed with suspect until they agree they are unambiguous. Should be stressful for guilty but not for innocent. The CQT Process – Phase 1 cont. 3. Control questions: Related to issues involved in the case, but not the crime itself. Control questions are general, vague and cover long periods of time. Should embarrass both innocent and guilty. Questions require an honest “yes” response, but suspect is told they must lie and that the polygraph will detect this fact. Control Question Test: Sample Questions Irrelevant Used to obtain a Are you left handed? baseline. Control Deal with prior Before age 25, did behaviour. Designed you ever verbally to provoke anxiety. threaten to hurt anyone? Relevant Deal with the crime. Did you assault Sam Smith the evening of November 11th? The CQT Process – Phase 2 Stimulation Test: Idea is to convince suspect that the polygraph will be able to detect lies Often involves a card trick Designed to make the guilty more fearful and the innocent less fearful The CQT Process – Phase 3 Control, relevant, and irrelevant questions are asked several times over The CQT Process – Phase 4 Score using either: 1. Global approach (subjective impression) 2. Numerical scoring approach  If Relevant > Control = negative score  if Control > Relevant = positive score  All comparisons are summed to give a total score (often +5 = truth, –5 is deceptive. In-between = inconclusive) Sometimes a 5th phase is added– tell suspect the result in hope he/she will confess! Criticisms of the CQT Formulation of control questions can be difficult Scoring is not sufficiently quantified Ethics? Strong element of deception How do we know that the suspect is “taken in” by the stim test? Why should innocent show stronger responses to control than relevant questions? Why should guilty show stronger response to relevant than control questions? 3. The Guilty Knowledge Test (GKT) The GKT (aka Concealed Knowledge Test) assesses if the suspect has information that only the criminal would know Asks suspects multiple choice questions, one option is correct Assumes if the suspect is guilty they will react strongly to correct information An innocent person would not recognize the crime- relevant details and would therefore not respond GKT Precautions The examiner should be unaware of correct answer – suspect may “detect” correct answer from examiner Need to check that alternatives are equally arousing for innocent person Problems with the GKT Applicability: Cannot be used unless lots of details are known only to the perpetrator. Knowledge of perpetrator: Assumes that perpetrator knew or remembered details – may not be true How Accurate Are Polygraph Tests? Accuracy of the CQT Majority of guilty suspects correctly identified  84% to 92% guilty correctly identified Relatively large number of innocent suspects falsely identified as guilty (false positives)  9% to 24% false positive errors How Accurate Are Polygraph Tests? Accuracy of GKT Very accurate at identifying innocent participants  Around 95% correctly identified Less accurate at identifying guilty participants (false negatives)  Around 85% correctly identified What Do the Experts Say? Question % Agree CQT is based on scientifically sound theory 36 GKT is based on scientifically sound theory 77 Would advocate admitting a failed CQT as 24 evidence in court Conclude that an individual who fails 8/10 GKT 72 items has guilty knowledge CQT can be beaten by increased response to 99 control questions Source: Iacono & Lykken, 1997 Current Directions: Voice Stress Analysis Voice Stress Analysis (VSA) Advocates suggest that lying is considered more stressful, and this changes your voice Most research on this area has found results no greater than chance at being able to distinguish lying and truth telling (e.g., Damphousse, 2008; Hollien et al., 2008) Lie Detectors Act 1983 (NSW) restricts use of VSA instruments Current Directions: Brain-Based Lie Detection Recent research has investigated the use of electroencephalography (EEG) and brain scanning by means of functional magnetic resonance imaging (fMRI). The research is has generated a lot of interest, but results of studies have not shown sufficient consistency to be admissible in court. Take Home Message There are a variety of approaches to criminal profiling, but there is limited evidence that profilers actually contribute much to investigations However, profiling is becoming increasingly evidence- based (from psychological research), so there is promise it will be more useful in future Physiological measures do not detect deception – they measure physiological changes associated with stress or recognition The polygraph does not work reliably enough to be regarded as a ‘lie detector’ New lie detection techniques are being investigated