How reliable is Eyewitness testimony, and what can be done to improve it
To answer this question in depth is beyond the scope of this work, which will explain briefly how human memory works; factors effecting the memory process; how a memory is recalled, with emphasis on eyewitness testimony; and concluding with steps to improve the quality and quantity of information retrieved from eyewitnesses.
“Only a fraction of the signals that reach us from the outside world can be registered by our senses. Even fewer are converted into memories.” (Lloyd-Bostock, 1988: p5). When we experience an event, all of our senses are receptors of information, which we then process for memory, however we are selective. We cannot remember everything as if videotaping our lives (Ainsworth, 1998). This filtering of information is recognized in Broadbent’s ‘Filter Theory’ (Milne & Bull, 1999).
Memory begins with ENCODING (Milne & Bull, 1999), where sense is made of the representation of the event. Attention is paid to the important and relevant. This attention-worthiness can be affected by our life experiences, attitudes, and expectations, causing each of us to selectively attend to different aspects of the same event. (Fruzetti, Toland, Teller & Loftus, 1992). Next, the experience is STORED (Milne & Bull, 1999). The encoded information is placed in memory where it will remain, barring some destructive trauma, until required for the final stage of the process, RETRIEVAL, the recall of the event (Milne & Bull 1999). At any stage in the process, one or more of the components can fail.
Memory is constructive, in that we build a meaningful picture of what our senses are receiving, filling in the gaps, an example of which was provided by Trankell (1972). A lawyer travelling in a taxi, had reported seeing the door of a car ahead open, an old man fall out, and lay in the road. It was later found that the old man was a pedestrian who had been knocked down, not a passenger in the car ahead. Based on fragmented visual information, he had constructed a feasible reality. (Lloyd-Bostock, 1988)
Generally, it has been found that we recall an event better if it did not involve violence (Clifford & Scott, 1978). Experiments using film of a fictitious robbery, depicting varying levels of violence, resulted in viewers of the violent version being 7 times less likely to accurately recall significant detail than those viewing the non-violent version (Loftus & Burns, 1982). It is suggested that the shock of witnessing a real crime interferes more with the processing of information ( Lloyd-Bostock, 1988). Stress is also shown to interfere with recall (Holmes, 1974). Some researchers argue that participants are better at recalling an event than those who are bystanders (Cohen & Faulkner, 1988; Yuille, Davies, Gibling, Marxsen & Porter, 1994) while others have found that involvement has no measurable impact on the accuracy of recall (Saywitz, Geiselman & Bornstein, 1992; Farringdon & Lambert, 1993; Roberts & Blades, 1998)
Ebbinghaus (1885/1964) conducted early experiments into learning and retention, describing the ‘Serial Position Effect, where items at the beginning – ‘primacy effect’, and the end ‘recency effect’ of a list, were more likely to be remembered than those in the middle. This was developed and expanded in the 1960’s by Broadbent, and by Atkinson and Shiffrin (1971) in their ‘Buffer Theory’. Which describes information entering a ‘sensory store’ where it remains briefly. If attended, the information is then passed to the ‘short term memory'(STM) which acts as a buffer to prevent ‘long term memory'(LTM) having to process too much information at one time. Short term memory can hold 7 (+- 2) chunks of information. Early chunks stay longer in STM and are therefore more likely to be passed to LTM, and later chunks are still in STM and easily recalled. (Ainsworth, 1998)
There are many factors which can negatively effect recall, (Milne & Bull, 1999):-
Sensory Information gained from one sense, can be distorted by another. Loftus (1977) demonstrated how visual information could be contradicted by auditory signals. The brain selectively makes sense of the conflict to make it fit with our expectations. Loftus and others expanded further with work on Compromise memory, making sense of the conflict between what we saw, and what we are told happened. The effect is compounded the closer the introduction of the conflicting information, to the requirement to recall (Hall, Loftus & Tousignant, 1984)
Inferences The words used as prompts for recall, are shown to be important in Loftus and Palmer’s (1974) experiments. Subjects shown a series of slides of a car crash and asked whether they saw broken glass when the vehicles smashed, and when the vehicles hit, 32 percent and 16 percent respectively, saw broken glass when there was none. However, people are more resistant to suggestion relating to core detail than peripheral detail (Yuille & Cutshall, 1986). This is important to note as defense lawyers will often challenge errors in peripheral detail, to discredit main details in testimony.
Stereotypes Encoding is contaminated by stereotypes. Hollin, (1989) experimented with a picture of a suspect with blond hair and green eyes. When asked to describe the suspect, 92 percent got the hair colour right, but of those, half stated the eyes were blue.
Partisanship Memory of events which have personal significance can be distorted by unintentional rationalization, to fit our expectations. Hasdorf and Cantril (1954) and later, Boon and Davies (1996), showed a number of sport fans a series of on pitch clashes in a game. Partisan fans wrongly judged who was responsible for the infringement.
Scripting We develop a series of scripts or schemata. How we deal with common recurring events is simplified by using our previous experience and tried and tested routines. Where an event does not fit the script, we fill in the gaps (Trankell, 1972). In an experiment where a number of people were given a very short story of a visit to a restaurant, when asked to recall it, some has people eating and paying for a meal, when the story had contained no such detail (Bower, Black & Turner, 1979).
Emotional Factors Traumatic events particularly in childhood can be suppressed by an automatic defense mechanism (Freud, 1917). Repressed memory is controversial,with adults claiming memories were not available to them because of suppression (Memon, Vrij & Bull, 1988). Debate about the reliability of recovered memory, especially as subjects are often vulnerable is heated (Conway,1997). Research has shown that adults,[even law students] can be misled into believing that they have seen an event they have only read or heard about. Given a description of an air crash, after repeated interviews, subjects were giving clear accounts of what they saw (Crombag, Wagenaar & Van Koppen, 1996).
Retrograde Amnesia Events in the last few seconds or as long as the last few minutes, [depending on the level of sensory activity] immediately preceding a trauma which renders the subject unconscious can be lost. It is believed that as there is not enough time to process information entering STM for passing to LTM. Events are therefore lost because they did not reach LTM (Atkinson & Shiffrin, 1971)
Sincere Belief Confidence, not accuracy is the most influential factor in juries assessment of who to believe ( Wells, Lindsay & Ferguson, 1979; Cutler & Penrod, 1995), yet some studies show that there is no relationship between confidence and accuracy at all ( Loftus, 1979) A confident eye-witness can be sincere in their belief, but nevertheless entirely mistaken. This can happen with a number of witnesses in the same case. Huff (1987) attributed almost 60% of 500 cases of wrongful conviction to eyewitness error, among them:-
Deckinga – Robberies- 3 bank tellers wrongly identified him as responsible
Father Pagano – robberies – 7 victims wrongly identified him as responsible
Murphy – Bombing – 3 witnesses including a hijacked taxi driver wrongly identified him
In all of the above, they were jailed and only released when real perpetrator was later identified. Cutler and Penrod taking account of this, suggest that expert evidence should be presented in court, to point this out to juries.
Age The elderly are more likely to make mistakes, and take longer to assimilate information. Yarmley (1996) also suggests that the elderly are more likely to make mistakes if their attention is divided between two or more tasks.
Occupation Our occupation’s effect what we attend to. Fruzetti et al (1992) noted that five people gave five different accounts of the same event. Their perceptions were influenced by what was familiar to them in the event they witnessed. Police and Security guards are not conclusively better at remembering things, but when called upon to describe events encountered professionally, perform well. It has been found that if someone knows in advance that they will be questioned about an event afterwards, they pay more attention and will be able to provide more accurate information than someone who is not (Clifford & Richards, 1977; Lloyd-Bostock, 1988).
What can be done to improve it?
Developed in the 1980’s by Fisher and Geiselman, and revised and reviewed since, the Cognitive Interview Technique (CIT) changed the way witnesses and suspects were interviewed. In early tests, after only four hours of training, high school students were able to elicit twice as much information from witnesses as police officers with many years experience. (Fisher & Geiselman, 1992; Geiselman, Fisher, Cohen, Holland & Surtes, 1986).
The process puts the interviewee at the forefront rather than the interviewer. The subject is encouraged to recall everything about the event in question, without editing and not to guess at details. This includes feelings, smells, and the physical environment. The interviewee is encouraged to use relative rather than absolute judgements, by the interviewers use of questions such as “was he taller than you?”, rather than “how tall was he?” The use of props, for example pictures of different types of knives could be used to assist witnesses to detail a weapon more accurately than they could give a verbal description. Similarly this could be used to demonstrate a mannerism, better than it could be described. (Milne & Bull, 1999)
Recreating the context of the event, before questioning commences is suggested by Fisher and Geiselman (1992) as an aid to recall. How they were feeling, whether in a hurry, what they were doing prior to the event may assist in recalling detail of the event.
Recalling detail from a different sense modality, can often trigger further visual recall (Farringdon & Lambert, 1997). They also suggest repeating interviews, encouraging the witness to try to recall detail they thought they could not. The fact that a witness is not able to retrieve on demand, a piece of information, does not necessarily mean it was not stored. However, there is a danger that the witness, if badgered, will make up a plausible response, inventing detail in order to please the interviewer ( Ainsworth, 1998).
While Memon, Bull and Smith (1995) suggest that some of the improvements attributed to CIT may be due to increased interviewer motivation, evaluation suggests it has led to an improvement in recovery of detail, with little increase in inaccurate information (Ainsworth 1998). However, CIT does not improve witness ability to identify suspects (Fisher, Quigley, Brock, Chin & Cutler, 1990). One possible explanation suggests recognition uses different memory skills than recall (Bahrick, Bahrick & Wittlinger, 1975; Lloyd-Bostock, 1988)
It is suggested (Baldwin & Bedward, 1991) that all witness interviews are recorded, to eliminate the dangers of the interviewer summarizing when writing the final statement. Few are written contemporaneously, and contain added jargon, unreported information, inconsistencies and omissions which the defense can exploit when questioning the witness. This is noted by experienced barristers as one of the primary causes of wrongful acquittals. ( Heaton-Armstrong & Wolchover, 1999).
An acceptance of the evidence of psychological research into eyewitness reliability, has driven improvements in the interview techniques employed by the police, but there is still a need for training to educate interviewers about the factors which can effect recall.