15.3 EYEWITNESS IDENTIFICATION AND HUMAN MEMORY
To be remembered, information must first be encoded by the eyewitness at the crime scene. An eyewitness may be unable to remember some aspect of an event because they did not attend to relevant detail and therefore it was not encoded in memory. The information must be stored for the intervening period without being lost or corrupted. Finally it must be retrieved at the appropriate time, either by recall or recognition.
PHOTO 15.1 Case histories show that mistaken eyewitnesses are often confident in their identification and more than one eyewitness can make the same mistaken identification.
Source: © John Gomez/Shutterstock
It is a mistake to think of human memory as like a video recording, which allows the viewer to freeze a frame and examine in close detail some aspect not previously noticed. Human memory is an active process of reconstructing an account of an event from incomplete information encoded in memory. When remembering, all available information is used including the witness’ own prior knowledge, expectations and assumptions. We tend to fill gaps in our memory with expectations of what “must have happened”. The person remembering is usually unaware of this process, and so is unable to distinguish genuinely remembered information from new information included in the memory. Expectations are derived from memory of typical everyday events (scripts), such as buying goods in a shop or going out for a drink (Schank & Abelson, 1977) or stereotypes (e.g. that the driver of a van was a man). When remembering, we may also use information acquired since the event (post-event information – see also Chapter 6 ), which may be derived from an investigator’s questions, or another witness’ account. For example, a witness may be asked: “Another witness mentioned that the man had a knife. Did you see a knife?” In summary, human memory is an active process that involves a creative component. This is why it is vulnerable to suggestion (from post-event information) and bias (from prior knowledge or assumptions).
Eyewitnesses are asked to perform a feat of episodic memory. They are required to remember what they saw at a specified place and time (i.e. an episode). When recalling an event a witness may misattribute a detail from a different event to the one being recalled. For example, a car seen earlier in the day may erroneously be remembered as being at the crime scene. This is known as a source attribution error.
Psychologists distinguish explicit memory, such as giving a free recall of an event, from implicit memory. When memory has an effect of which one is not aware it is said to be implicit memory. An example may be the influence of a script in memory used to fill in details of an everyday event.
The principles of memory described up to this point apply to eyewitness testimony in general (see also Chapter 6 ). When asked to identify the face of a perpetrator, visual memory is involved. The general principles of memory described apply equally to visual memory. However, recall of visual memory, especially memory for faces, is very difficult, partly because we do not have adequate language to describe each face uniquely. Therefore when attempting to identify the face of a perpetrator we rely on recognition. Recognition yields superior memory performance to recall. The purpose of organising an identification procedure (such as a lineup) is to test whether a witness recognises the face of the suspect. A witness may recognise a face because they can recollect seeing that person on a specific occasion (episodic memory). However, they may recognise the face merely because it is seems familiar. The influence of implicit memory may make a face seem familiar, in the absence of any recollection of when the face was seen previously. Recognition memory for a face exceeds the ability to recall the context of encounter. Familiarity may then be misattributed to presence at a crime scene and potentially may lead to a mistaken identification.
15.4 DESIGN REQUIREMENTS OF IDENTIFICATION PROCEDURES
It is useful to consider an identification procedure as a scientific experiment (Wells, 1993). The police have the hypothesis that the suspect is the perpetrator. The purpose of an identification procedure is to test the hypothesis, by establishing whether the witness can identify the suspect as the perpetrator. Such an experiment should enable a compliant but unreliable witness to be distinguished from a witness who makes a reliable identification of the perpetrator.
The simplest identification procedure is to allow the witness to see the suspect or their photograph and ask if this person is the perpetrator. This procedure is known as a show-up in the United States and as a confrontation in the UK. It is most commonly used at or near the scene shortly after a crime has been reported, when there is insufficient justification to arrest of a suspect. In these circumstances the procedure would be referred to as a street identification in the UK. Essentially the same procedure is also used as a dock identification during a trial under Scots Law, when a witness is asked whether they see the perpetrator present in the courtroom (see Case Study 15.1). The problem with these procedures is that there is no means of establishing whether a witness who makes an identification is mistaken. There are only two possible outcomes: a witness may decline to identify the person or they identify the suspect. The procedure may lead the witness to believe that there is other evidence that the person committed the offence. The procedure is suggestive and may bias the witness’ response.
CASE STUDY 15.1 THE MURDER OF JUSTIN MCALROY
Justin McAlroy was said to owe £50,000 in connection with a drugs deal. After visiting serious criminals in Perth Prison on 7 March 2002 he returned to his home shortly before 10pm, where a gunman was waiting for him. His wife was in the house. She heard three or four loud bangs and ran to the front door. She looked out of the window in the door and saw a man in her driveway. She described him as wearing a blue-green hooded bomber jacket with the hood up and a similar coloured scarf or snood covering his nose and mouth. In her first statement, made about 40 minutes after the shooting, she said she would not be able to identify him.
Other witnesses provided similar descriptions of a padded jacket. One witness, Mr Madden, reported seeing a man in similar clothing get into a white car and remove a ski mask. At about 10.30pm a white Saab car was found abandoned some miles away. An attempt had been made to set it on fire. Inside was clothing including a hooded jacket and a drinks bottle on which William Gage’s DNA was found. His DNA was also on gloves and a snood, with traces of the DNA of at least two other individuals. Small quantities of firearms discharge were found on the jacket and the snood.
William Gage was arrested and charged with murder. The prosecution case depended on establishing that the white Saab was the getaway car and on identification evidence. The defence contended that the white Saab was not connected to the offence, and that William Gage had spent the evening with a girlfriend.
In the course of the investigation, the police showed Mrs McAlroy a tailor’s mannequin dressed in the clothing recovered from the Saab. She identified the clothing as similar to that worn by the gunman, although it differed from her earlier description. For example, the jacket was a thin waterproof cagoule rather than a padded jacket. An identification parade had been arranged but abandoned because Gage objected to the selection of foils. Neither Mrs McAlroy nor Mr Madden attended the parade, but instead, the prosecution sought to rely on dock identifications.
At trial, Mrs McAlroy identified Gage by his eyes: she said that the man running away had “scary eyes” that she would never forget, but she had never mentioned his eyes in her statements. Mr Madden did not identify Gage at trial. The jury convicted Gage of murder and he was sentenced to life imprisonment.
The case went to appeal in 2006 but was dismissed. The court noted that in a case involving circumstantial evidence, it is necessary to look at the evidence as a whole. Each piece of evidence might not be incriminating in itself, but the concurrence of testimony was critical. It was for the jury to decide how to interpret this evidence and they were entitled to reject inconsistent evidence if they so choose. In June 2009 the Scottish Criminal Cases Review Commission referred William Gage’s convicti
on back to the Court of Appeal on the ground that the absence of a specific direction to the jury in relation to the dock identification may have given rise to a miscarriage of justice. The second appeal was dismissed.
(Gage v. Her Majesty’s Advocate 2006); Gage v. Her Majesty’s Advocate (2012)
A much better procedure is to ask the witness whether the perpetrator is present in a lineup of people, which includes one suspect. The foils in the lineup are not suspects. This procedure has three possible outcomes. The witness may identify the suspect, providing evidence that the suspect is the perpetrator. The witness may make no identification. Or the witness may identify a foil, making a known mistaken identification. Lineups are widely used, especially in the UK and the United States. In the United States lineups may consist of a live line of people or an array of photographs (a photo-spread). Historically, in the UK identification has been made by use of live lineups, known as identity parades, but since 2011 video identification procedures have been the preferred method of identification.
There is extensive research on the factors that affect the ability of eyewitnesses to identify a perpetrator from a lineup. It is useful to distinguish between factors that are under the control of the criminal justice system (known as system variables) from those that are not (known as estimator variables; see Chapter 6 and Wells, 1993; Box 15.1). Understanding the influence of estimator variables is important in order to evaluate the likely performance of eyewitnesses under various conditions. However, understanding the effects of system variables enables recommendations about best practice to be made.
BOX 15.1 ESTIMATOR AND SYSTEM VARIABLES
Estimator variables, which concern the circumstances of the witness’ opportunity to view the culprit and are not under the control of the criminal justice system, include:
the time to view the perpetrator, the distance of the witness, the lighting and other circumstances of the incident
whether the perpetrator is known to the witness
the presence of a weapon
the stress induced in the witness
differences in ethnicity between the witness and perpetrator
the age of the witness.
System variables, which generally are under the control of the criminal justice system, include:
the selection of identification method (e.g. street identification, lineup procedure, dock identification)
the mode of presentation (e.g. photographs, live, video)
instructions given to the witness
“blind” vs. “non-blind” administration of a lineup
use of prior identification procedures (e.g. showing photographs prior to a lineup procedure)
method used to select foils for a lineup
method of lineup administration (e.g. simultaneous vs. sequential presentation of lineup members)
feedback given to witnesses.
15.5 ESTIMATOR VARIABLES
Is it possible to judge from the circumstances of an eyewitness’s view of the culprit whether their identification is likely to be accurate? In this section the effects of some estimator variables will be briefly reviewed. To begin, estimator variables that have a special status in English Law will be considered from a psychological rather than a legal perspective. The section concludes by examining the effect of estimator variables in archival studies of real criminal cases.
15.5.1 The Turnbull Guidelines
English case law is based on the premise that a distinction between good and poor eyewitness identification evidence is possible. Following a landmark ruling in the Appeal Court (R v Turnbull, 1976), when identity is disputed a trial judge must advise the jury to consider carefully the circumstances of an identification (see Box 15.2).
BOX 15.2 THE TURNBULL GUIDELINES (R V TURNBULL AND OTHERS, 1976)
In a case that relies substantially on disputed eyewitness identification evidence, the trial judge must warn the jury about the special need for caution before relying on the accuracy of eyewitness identification evidence to convict the defendant. The judge should make some reference to the possibility that a convincing witness may be mistaken and that a number of witnesses who make the same identification may all be mistaken. The judge should direct the jury to consider carefully the circumstances of each witness’s identification. The relevant factors are often summarised by the acronym ADVOKATE:
Amount of time the perpetrator was in view
Distance of the witness from the perpetrator
Visibility of the perpetrator; how good was the lighting?
Obstruction to the witness’s view?
Known to the witness? Has the witness seen the suspect before? How often?
Any reason to remember? If only seen occasionally before, did the witness have any reason to remember the suspect?
Time delay between the incident and the formal identification procedure
Error. Is there any material discrepancy between the description given to the police at the time of the incident and the appearance of the suspect?
The judgement in Turnbull arose from the Devlin enquiry (Devlin, 1976). Subsequent laboratory research has confirmed that most of the factors mentioned in the judgement do affect the accuracy of eyewitness identification. For example, witnesses who had 45 seconds to view a culprit were more likely to identify him from a lineup than witnesses who had only 12 seconds (Memon, Hope, & Bull, 2003). The ability to identify faces viewed under different levels of lighting and at a range of distances has been investigated (Waganaar & Van der Schrier, 1996). Faces of people known to the viewer are remembered in an episodic memory task with much greater accuracy than unfamiliar faces, when a different view of the face is presented in the study and test phase of the experiment (Bruce, 1982). Fewer correct identifications are made after long delays (Deffenbacher, Bornstein, McGorty and Penrod, 2008). The absence of errors in a verbal description is positively associated with the accuracy of a subsequent identification (Meissner, Sporer, & Susa, 2008; Gabbert & Brown, 2015).
A recent Supreme Court judgement in New Jersey provided a landmark for evaluation of eyewitness identification evidence in US courts (New Jersey v. Henderson, 2011). If the identification procedures can be shown to have been suggestive, all relevant estimator and system variables should be considered to reach a judgement of whether the eyewitness identification evidence is admissible. The judgement in the Henderson case required that enhanced directions to the jury on eyewitness identification must be developed.1
The Henderson judgement considered the reliability of the scientific literature of many system and estimator variables. With respect to estimator variables, the Court found that the reliability of an eyewitness’s identification can be affected by: high levels of stress; the presence of a weapon when the interaction is brief; cross-ethnicity identification; and witness interaction with co-witnesses or other sources of information. The Court also concluded that research showed that people do not intuitively understand all of the relevant scientific findings. Therefore it is necessary to educate jurors on the psychological science of eyewitness memory.
15.5.2 Laboratory Studies of Estimator Variables
In this section the effect of several factors on the accuracy of eyewitness identification will be reviewed. The issues that have attracted research attention in recent years will be highlighted.
15.5.2.1 Weapon focus
When confronted by a perpetrator wielding a knife or a gun, the weapon captures the attention of a witness (see also Chapter 6 ). Under these circumstances there is believed to be a narrowing of attention, meaning that less attention is paid to other aspects of the scene. Therefore eyewitnesses are very capable of describing the detail of the knife or gun, but may be less able to recognise the face of the perpetrator. Hope and Wright (2007) argued that both the novelty (unusualness) and perceived threat of an object captures the attention of witnesses and therefore contribute to the weapon focus effect. Steblay (1992) conducted a meta-analysis of 19 tests of the hypothesis and found a re
liable but small effect (effect size = .13) on identification. Witnesses were less accurate in identifying a perpetrator when a weapon was present. From a more recent meta-analysis, Fawcett, Russell, Peace and Christie (2013) concluded that the effect of weapon focus was greater for recall of the appearance of the culprit than for accuracy in identification tasks (see also Chapter 6 ).
15.5.2.2 Stress
The presence of a weapon may exert its influence through inducing fear or stress in a witness, rather than directly by the capture of attention per se. Are witnesses who experience a very frightening or stressful event less reliable than witnesses who experience less stress? Morgan et al. (2004) examined the ability of solders to recall a person present at an interrogation. The solders had been detained for 12 hours in a mock prisoner of war camp. Each solder then underwent a high stress interrogation involving physical confrontation and a low stress interrogation. Twenty-four hours later the soldiers took part in an identification procedure. Identification was more accurate for the target person seen during a low stress interrogation (67%) than for the person seen during a high stress interrogation (29%).
Forensic Psychology Page 69