Week 7.2 ID evidence Part 2 - Lineups.pdf

The psychology of identification evidence Why is ID evidence so dangerous? • Because it is very convincing – Jurors find the testimony of an eyewitness very convincing. After a juror has heard the evidence of an eyewitness they are likely to convict regardless of what other evidence is presented...

The psychology of identification evidence Why is ID evidence so dangerous? • Because it is very convincing – Jurors find the testimony of an eyewitness very convincing. After a juror has heard the evidence of an eyewitness they are likely to convict regardless of what other evidence is presented • Because research show us that identification evidence is often inaccurate • Thus, eyewitnesses can be sincere, confident, convincing…. And wrong. How do we know EW ID is often inaccurate? 1. Cases studies of wrongful conviction 2. Empirical studies of ID performance DNA Exoneration • Since the 1980s advances in DNA analysis have allowed investigations to link a person to a crime with a very high degree of certainty. • DNA analysis has also been used to prove the innocence of convicted inmates • Much of this work has been done by “The Innocence Project” in USA • Analysis of cases of DNA exoneration give us a very valuable insight into the dangers of various forms of evidence. innocenceproject.org An early example of a DNA exoneration Kirk Bloodsworth • Following rape and murder of 9 year old girl, Bloodsworth was identified by a total of 5 eyewitnesses. Convicted 1985. • Sentenced to death. • 1993 released after serving 9 years – 2 on death row. • "I was in total shock. I couldn't believe that all these people were saying I was a brutal murderer." Facts on Post-Conviction DNA Exonerations • • • • • • • • As of July 2021 were 375 post-conviction DNA exonerations in the United States. The first DNA exoneration took place in 1989. Exonerations have been won in 37 states 21 of the people exonerated through DNA served time on death row. The average length of time served is14 years. The total time served is approximately 5,284 years. The average age of exonerees at the time of their wrongful convictions was 27yrs. 288 have been financially compensated. The true suspects and/or perpetrators have been identified in 165 DNA exoneration cases. While innocent person was in prison true offenders committed an additional 154 violent offences including 36 murders and 163 sexual assaults Factors in DNA exoneration cases (2021) Percentage of exoneration cases 100 90 80 70 60 50 40 69 30 43 20 29 10 17 0 ID Forensic False Confess Science Evidence type Informant Contributing causes in US DNA exoneration cases as of July 2021 (note percentages sum to more than 100% because most cases have more than one contributing cause). Facts about the ID exonerations • • • • • • • • • • • • • • • 34% of these misidentification cases involved an in-person lineup 52% involved a misidentification from a photo array 7% involved a misidentification from a mugshot book 16% involved a misidentification from a show-up procedure 5% involved a misidentification from a one-on-one photo procedure 27% involved a misidentification through the use of a composite sketch 11% involved a voice misidentification 2% involved a misidentification through hypnosis 54% involved an in-court misidentification 29% involved a misidentification through some other procedure (e.g., mistakenly “recognizing” someone on the street and reporting them to law enforcement) 77% of the misidentification cases involved multiple procedures 84% of the misidentification cases involved a misidentification by a surviving victim 42% involved a cross-racial misidentification 32% involved multiple misidentifications of the same person by different witnesses 18% involved a failure to identify the exoneree in at least one procedure Wells et al (1998) note: “It is important to note that (…these cases…) were not selected because they happen to have eyewitness identification evidence as the primary evidence. Instead, these cases are simply the first instances in the US in which DNA was used to exonerate a previously-convicted person. Hence, the kind of evidence that led to these wrongful convictions could have been anything. The fact that it happens to be eyewitness identification evidence lends support to the argument that eyewitness identification evidence is amongst the least reliable forms of evidence and yet persuasive to jurors” Empirical evidence • There are two types of studies: – Laboratory studies – Field Studies • Lab studies allow more control but typically lack ecological validity. Field studies are high in validity but lack control – especially over factors such as amount of contact, attention etc. In archival studies never know whether suspect was perpetrator Evidence from Lab studies • • Typically either watch a video or stage a mock-crime in front of participants. participants are then required to identify the perpetrator from either photographs or a live lineup. Wells (93) reviewed many of these studies and concluded that the rate of false ID varied from about 0% to about 100%! Not a very helpful finding. More useful to note that false Ids often associated with very high witness confidence. Problems with lab studies • As Cutler & Penrod note, lab studies are not normally designed to estimate overall level of accuracy – rather are designed to investigate effect of some particular variable. • For this reason researchers tend to manipulate task difficulty so as to avoid floor and ceiling effects. • Thus must not try to estimate overall performance on basis of these studies. Field studies • • • Brigham (82) Krafka & Penrod (85) Platz & Hosch (88), all had stooges go into convenience stores and perform an unusual transaction. Piggott et al (90) – stooge attempted to cash a forged cheque at a bank Overall about 41.8% correct ID, but 35.8% false ID. No different for the bank clerks who clamed they had been trained to ID Valentine & Mesout (2009). Participants attending “London Dungeon” experience. Tired to identify a person seen in a ‘scary encounter’. Higher stress = fewer correct IDs form lineup and less accurate description. Factors affecting the accuracy of EW identification (Wells 1978) • Estimator Variables – Influence can only be estimated after the event. – Viewing conditions – Attention – Cross Race Effect – Weapon focus – Arousal – Fatigue, alcohol etc • System Variables – Variables the police, psychologists and legal professionals can influence – ID procedures – Instructions given to witnesses – Interview method etc Identification procedures • There are four types of identification procedure 1. 2. 3. 4. • Mugshot searches Showups Line-ups / ID parades Voice and other non-facial Identifications In addition Facial Composites are constructed on basis of a witness description. Photo-lineups • Procedure varies across jurisdictions. In NSW most are photo lineups, as is also the case in most US states. In UK used to be live but switched to video in early 2000s. • Cutler et al (1994) concluded that photospreads produced similar performance levels to live and video lineups. • However, Pike & Kemp (1995) showed that movement can improve performance. Lineups • Exact arrangement varies between jurisdictions. In US are typically 6 people in a lineup. In UK 9 is normal and in NSW 20 is common. In UK suspect chooses position in line. • Wells et al (94) “there is no threshold number below which the dangers of false ID are significant and above which they are not” • More important than real size may be the “functional size” Functional size • In this example lineup physical size is 5 but functional size is 1. • Perpetrator was described as a black male. • In lineup is only one black male. All foils are Hispanic. Fair lineup? • Police have covered birthmark noted by witness but look at ankle of No5! Psychologists recommend change • • • • Wells et al (1998) reviewed available research and produced a set of recommendations in the “White Paper” In 1999 US Department of Justice produced set of guidelines partly based on “white paper” In UK procedures are controlled by Police and Criminal Evidence Act (Pace) In NSW there is no equivalent set of guidelines but in practice police meet many of these requirements Psychologists recommend change The “White Paper” (Wells et al, 1998) made four recommendations: 1 2 3 4 Blind procedure – officer should not know identity of suspect Witnesses told that person “may or may not be present” and that is a blind procedure Foils should be matched to description of perpetrator and suspect should not stand out Measure confidence at time of identification Simultaneous and sequential lineups: The “fifth recommendation” • In a standard or “simultaneous” lineup witness sees all members at same time. • This encourages witness to pick person most-like perpetrator - relative judgement • Sequential lineups were developed by psychologists to reduce the reliance on these relative judgements Sequential lineups • In a sequential lineup the witness sees each person just once. • They must make a yes/no decision before moving on. • Once they choose someone or reach the end, then lineup stops. • Are not told how many images in the lineup Evidence regarding sequential lineups • Steblay et al (2001) Meta analysis of 30 studies. Found that for Target Absent lineups sequential procedure results in fewer False+ve errors. • For Target Present simultaneous results in more hits, but this effect is moderated by “realism” of study is considered. • Subsequent research (Memon et al 2005) confirmed that, relative to simultaneous, sequential lineups results in fewer correct IDs in target present lineups. • Kneller et al (2001) Decision processes – use of absolute strategy associated with correct decisions. In Sequential tend to use absolute, but in simultaneous report using both absolute and relative. • Lindsay et al (1991, 1999) Technique matters! Implementing sequential lineups • • • • Some jurisdictions starting to adopt sequential lineups – but in implementing the sequential lineup have introduced changes Kemp et al (2006) suggests that these changes to procedure (e.g. allowing witnesses to see all images) may eliminate or even reverse the sequential advantage. There have also been some field trails of sequential procedures. Illinois field test (Sheri, 2006) – showed rate of false ID much lower than expected, and sequential resulted in MORE false Ids than simultaneous. This data is currently the topic of much debate (e.g. see Innocence project website) Steblay (2010): describes use of freedom of information lawsuit to gain access info about how lineups in Illinois study were conducted. Argued that was not a fair test of sequential lineups Further research • Malpass (2006) conducted an analysis of the utility of the two lineup systems – argued simultaneous is BETTER unless: – a) place MUCH more weight on avoiding false Ids than on convicting the guilty – b) assume that a very high proportion of real lineups (eg 50%) are target absent This analysis strongly favours simultaneous lineups • • Gronland et al (2014) argue that Sequential merely shifts bias but does not increase sensitivity. Wells et al (2015) Report first results form a new trial (AJS trial) – researchers provide police with computer systems to administer lineups. – • Initial data from 500 eyewitnesses: Suspect identified in 25% of cases (same for Seq and Sim), but Seq resulted in fewer identifications of foils (11% vs 18%). About 60% make no ID. The consensus in favour of Sequential lineups has broken down in the last few years Figure 2. Percentages of Witnesses Identifying the Supect, a Filler, or Making No Identification for the Simultaneous and Sequential Procedures [N = 497] Wells et al 2011 White paper update • Wells et al 2020 • Confirmed original 4 recommendations and added 5 more: – Need for preliminary witness interviews – Evidence based suspicion prior to ID – Avoid repeat ID procedures with same witness – Avoid show-ups – Video record procedures Confidence and accuracy • • • • Confident witnesses are very persuasive to jurors But confident witnesses are not always accurate. Krug (2007) suggested correlation is about 0.3 Major problem is that of “confidence hardening” – witness confidence increases following feedback from police so that when in court appear very confident Brewer and Weber (2009) show that if confidence is measured BEFORE feedback is provide it can provide a useful predictor of accuracy. ‘Pristine’ Lineups and confidence • • • • Wixted and Wells (2017) Define ‘pristine lineups’ e.g., initial, uncontaminated memory tests using fair lineups, with no lineup administrator influence, and with an immediate confidence statement For such pristine lineups with adults they say both field and lab studies show confidence and accuracy are strongly related and highconfidence suspect identifications are remarkably accurate. However, with non-pristine conditions this relationship breaks down. They note “some jurisdictions have not yet made reforms that would create pristine testing conditions and, hence, our conclusions about the reliability of high-confidence identifications cannot yet be applied to those jurisdictions”. Unconscious transference • • • • • Identifying an innocent person from a lineup because they are familiar Real world exampleDon Thompson was giving a lecture on TV while woman was being raped. She later saw him and identified him as her attacker. In studies UT has been demonstrated 2 ways – either bystander at crime scene or from mugshot searches to lineup (Davis, Loftus, Vanous, & Cucciare, 2008) Cross Race Effect • People better able to recognise individuals of same “race” than individuals of different race. • Has been demonstrated across a variety of groups, age groups and recognition tasks (Meissner & Brigham, 2001) • Particularly likely to give rise to false identifications of innocent targets (“they all look the same”) But lineups are dangerous… • • • • Katzman and Bull Kovera (2023) undertook a very careful meta analysis of 54 published effect sizes of the cross race affect. The effect was not large enough to account for racial disparities in wrongful conviction data Concluded that was better explained by differences in the evidence that police have against suspects before placing them in an ID procedure. That is police appeear more willing to place a back suspect in a lineup putting them at risk of false ID. Recommended policy requirements for evidence-based suspicion before placing suspects in an ID Conclusions • ID evidence is potentially dangerous • Affected by System and Estimator variables • Psychologists have attempted to improve the accuracy of ID evidence by looking at impact of System variables. • Sequential lineup was heralded as a great advance, but recent evidence begins to throw some doubt on this. • Provides an interesting case study in how (and how not to) move form lab research to public policy. Revision Question • Which of the following is/are system variables: a) b) c) d) e) Age of suspect Distance from eyewitness Lineup procedure employed Race of suspect and witness All of the above