Psychology of Eyewitness Memory Research Paper

1. Scientific Approaches To The Study Of Eyewitness Testimony

The scientific investigation of verbal descriptions of brief events was initiated by experimental psychologists in Germany and North America. Surprisingly, the same cautions and warnings that these researchers expressed 100 years ago concerning the low reliability of eyewitness testimony appear frequently in today’s research journals and the criminal justice system. That is, although much has been learned about the variables related to eyewitness testimony, researchers have provided relatively little information as to how to improve its reliability and accuracy. Instead, the lion’s share of their efforts has served to maintain as much reliability as possible in the face of potentially contaminating investigative procedures that may follow a witnessed event; for example, the use of suggestive questioning and biased identification procedures.

The types of information of interest to investigators may relate to the actions, words, and characteristics of the perpetrator (including voice and movement) as well as to the details of the context and the roles played by other people. A focus upon the acoustic qualities of voice is sometimes referred to as ear witness testimony (Thompson et al. 1998). Witnesses may also be victims and, as such, their recollections may have been influenced by the demands of the perpetrator and the degree of threat leveled at them. Indeed, in comparison to our recall of more mundane everyday events, it is the surprising, threatening, and emotional impact of criminal acts that renders them of special interest to eyewitness testimony researchers. As a result, one focus of their research has been the level of arousal experienced by witnesses to and victims of crime, particularly when they are children (Goodman and Bottoms 1993). For historical reviews of the growth of the field and the results of early eyewitness testimony research, the reader is referred to Cutler and Penrod 1995, Loftus 1979, Ross et al. 1994, Shepherd et al. 1982, Thompson et al. 1998, and Wells 1993.

1.1 Methodological Limitations To The Study Of Eyewitness Testimony

Current understanding of eyewitness testimony derives from two sources: real-world observations of witnesses victims and research simulations of criminal acts. All would agree that the recollections of real victims and witnesses is the most relevant if we wish to understand the quality and quantity of eyewitness testimony. All would also agree that to assess the accuracy of eyewitness testimony it is necessary to have independent knowledge of the events as they occurred. Unfortunately, in most real-world events this ground truth can never be known because independent records of events in natural contexts are rarely made. To circumvent this problem and to be assured of access to the ground truth, eyewitness researchers often present simulated events to witnesses (staged or on videotape) that can be later recalled. When scientists control the events to be experienced by witnesses, they also have invaluable control over most other variables relevant to the scenario. For example, researchers can manipulate the type and duration of the event, the specific features of the perpetrator(s), the age, gender, and race of the witnesses, the presence of weapons or threat, and the types of recall and identification tests. In real crimes of course, all of these are outside the control of the researcher and assessment of the credibility of a witness’s allegation or a defendant’s statement about an event falls to triers-of-fact (judges and jurors) in the courtroom.

Depending upon the researcher’s creativity, a crime simulation may be very realistic; however the very features that distinguish eyewitness events from mundane real-world events are, for the most part, missing in the laboratory simulations; specifically, surprise, threat, and high emotional involvement. Research participants must give their informed consent before taking part in scientific experiments and ethical considerations preclude the possibility of their being exposed to highly emotional, disturbing, or life-threatening events. For these reasons, virtually all of the simulation work on eyewitness testimony is open to the criticism that it incorporates, to a greater or lesser degree, insufficient ecological validity; that is, the results may not necessarily be generalizable to eyewitness events in the real world. Nonetheless, with appropriate safety and ethical precautions, threat, arousal, and fear have been manipulated in laboratory research (Thompson et al. 1998).

2. Nature Of Eyewitness Testimony

The vagaries of human perception, memory, and decision-making were well known to philosophers, scientists, police, and judges long before experimental psychologists initiated their research. However, psychologists did make a unique contribution to this knowledge base: systematic data collection that allowed measurement of the magnitude and frequency of human errors in the recollections of brief events. Most researchers in the 1970s and 1980s were sufficiently impressed by the frequency of these errors that much energy was directed to their documentation. Whereas none would anticipate perfect performance in either the perception or recall of such events, an unexpected finding was the influence a variety of nonperceptual and nonmemorial variables had upon performance, for example, those of social influence and the form of questions asked of witnesses.

2.1 Recall Of Details And Perpetrator Identification

Two categories of eyewitness information of special interest to researchers are the details of the crime and of the perpetrator. It is recognized that the measurement of accuracy of recalled details is highly dependent upon the types of questions used to elicit responses. For example, police investigators regularly interviewed witnesses and victims, yet their interviews were usually not based upon techniques known to enhance recall in cognitive and social psychology. As of 1983 this situation began to change as researchers applied cognitive principles to police interviewing. One result was a specific interview protocol known as the cognitive interview (Fisher and Geiselman 1992) that has well demonstrated its superiority in gathering more correct information from both adults and children without increasing the numbers of errors.

The misidentification of an innocent person is a particularly dramatic kind of eyewitness error because of its serious consequences for a person’s freedom. Indeed, legal opinion in most countries has emphasized the dangers of convictions based upon eyewitness testimony evidence alone (Cutler and Penrod 1995) and media attention given erroneous convictions has raised the issue’s profile in the public mind. As increasing numbers of demonstrably faulty convictions were studied, it became readily apparent that simple manipulations of photos and test instructions could have large effects upon the identification decisions of eyewitnesses. For example, Loftus (1979) showed that presenting a photo with a unique quality or different alignment in a photospread significantly raised the number of misidentifications of the person depicted in the photo. Analyses of both live and photo lineups in real cases have shown many to be poorly constructed and when so constructed, heavily biased against the police suspect (Wells and Bradfield 1998). Further, minor wording changes in the instructions given witnesses before they view a lineup dramatically raise the rate of false positive errors or selections of an innocent person. Indeed, false positive rates of 50 percent or more are not uncommon and Wells and Bradfield (1998) recently reported that when instructions encouraged witnesses to believe falsely that the perpetrator was in the photospread, all witnesses selected someone.

Similarly, social pressure from an authority figure or from other witnesses can influence lineup decisions. In particular, information may be subtly conveyed from investigators to witnesses by the kinds of questions asked of the witnesses and the investigator’s responses to answers given. Wells and Bradfield (1998) have shown that when research participants were informed falsely that their identification choices had been correct, their descriptions of the perceptual qualities of the event itself were dramatically altered, for example, how much time and attention they had directed to the perpetrator, the ease and confidence of their identification, and their willingness to testify in court.

Factors of interest to researchers may be classified as either estimator or system variables (Wells 1993). In the real world, estimator ariables are not under the control of either the criminal investigator or the criminal justice system and their effects upon eyewitness identification accuracy may only be estimated, for example, the age, sex, and race of the witnesses and the lighting conditions at the scene of the crime may be related to reliability of eyewitness testimony but they are inherent to the crime itself. Knowledge of the effects of these variables may assist us to better characterize the average performance levels that are obtained by specific types of people in particular environmental and viewing conditions. On the other hand, system variables are under the control of the criminal justice system, for example, the size, type, and quality of the lineup or photospread, the instructions given the witnesses, and the temporal interval between the crime and the identification task. Wells (1993) has argued vigorously that researchers should dedicate more effort understanding system variables because the accuracy and reliability of eyewitness testimony can be improved by manipulating features of the system itself.

A striking example of this kind of improvement may be seen in the recommendation (Lindsay and Wells 1985, Wells 1993) that lineup or photospread members be presented one at a time to an eyewitness rather than together. Although simultaneous presentation of lineup and photospread members has been normative in North America and the UK for many years, doing so is well known to encourage the use of a relative judgment strategy by which witnesses attempt to choose the person who best fits their memory of the perpetrator. The difficulty with relative judgments is that every lineup or photospread necessarily includes someone who looks most like the perpetrator and, therefore, someone will virtually always be chosen. Mistaken identifications will necessarily comprise a subset of these choices. Indeed, the high levels of false positives reported for identification tests are a likely consequence of just such a strategy.

In contrast, a sequential lineup presents each member one at a time and the goal is to force witnesses to rely upon an absolute judgment strategy. For each lineup member shown, the witness must indicate whether this person matches the representation of the perpetrator in memory. Once a person or photo has been presented, it may not be viewed again. To further reduce the opportunity to make relative judgments, witnesses are not informed of the number of people in the lineup. Thus, for the witness who attempts to use a relative judgment strategy, there remains the possibility that there may yet be someone who is even more like the perpetrator among the as yet unseen lineup members. In research comparing sequential to simultaneous lineup performance the two presentation procedures yield identical hit rates (correct identifications of the target) for target-present lineups (when the guilty person is included). However, in target absent lineups (in which the perpetrator is not present), significantly fewer false positives are made with the sequential than the simultaneous lineup. Thus, when police investigators employ a sequential procedure, innocent persons are better protected from misidentifications than with the simultaneous presentation.

3. Relationship Between Accuracy And Confidence

Analyses of jurors’ decision-making processes have indicated that of all the factors influencing their judgments of credibility of a witness, the confidence expressed by the witness is most influential. When witnesses claim to be absolutely certain of their statements their evidence is given more weight than that from other witnesses who expresses less certainty. Assignment of greater evidential weight to the first than to the second group of witnesses would be reasonable if subjective confidence was predictive of accuracy. Although it is clear that the triers-of-fact do believe there is a strong relationship between accuracy and confidence, the research evidence has been equivocal. Whereas a minority of studies has demonstrated moderate to strong relationships between the two variables, the majority have not. Indeed, the obtained relationship has been generally so low as to provide virtually no predictive value from confidence to accuracy. However, as research evidence accumulates from experiments with higher ecological validity, it is possible that the earlier research will be seen as having underestimated the strength of the accuracy-confidence relationship (Lindsay et al. 1998).

Witnesses’ subjective confidence is an estimator variable because it reflects a witness’s characteristics, such as age and gender. Nonetheless, its expression by a witness may be seen also as a system variable because it can be altered by the criminal justice system. For example, witnesses may be encouraged by others to assign high confidence to their in-court opinions, often much higher than they provided at the time of the identification. As a result, many researchers have argued that some standardized assessment of confidence should be taken by the investigating officer at the time of a witness’s identification decision (Wells et al. 1998).

4. Effects Of Post-Event Information And Cognitive Activities

Many factors influence a person’s recall of a brief event and because human memory is assumed to involve both constructive and reconstructive processes, information considered or received by the witness following the event may also contribute to the specifics of what is ultimately recalled. For example, in postevent misinformation studies, participants view a video event, then hear a narrative about it that contains incorrect information about details in the film (e.g., the getaway car was blue rather than green). Later, they are asked to recall details from the original video they viewed. The typical finding is that participants often incorporate information from the narrative by recalling (or recognizing) details that are consistent with the misleading information. Going well beyond distortion of minor details, research participants have also constructed complete but false autobiographical events as a result of similar suggestive misinformation techniques. The ease with which such memories may be manipulated or constructed has contributed to the development of an entire new field of false memory research, a field whose topics often overlap with those of eyewitness testimony research.

Other post-event mental activities such as rehearsal, verbal coding, and image generation can similarly contribute to altered recollections of the event or person. The Wells and Bradfield (1998) research dramatically demonstrated these kinds of changes as do the detrimental effects of both postevent verbal (Schooler and Engstler-Schooler 1990) and conceptual (Read 1995) rehearsal of events and people. These kinds of retrospective reconstructions or reframing of events are likely to form the basis of much additional research in the field.

5. Impact Of Eyewitness Testimony Research On The Criminal Justice System

The fruits of eyewitness testimony research since the 1970s may be seen in the publication of two significant documents. The first was a scientific brief commissioned by the American Psychology-Law Society (Wells et al. 1998) concerning best practice guidelines for the collection of eyewitness identification evidence. The second document (NIJ 1999) surveyed eyewitness testimony evidence more broadly and was the out-growth of an ongoing US National Institute of Justice project reviewing convictions of men for whom DNA testing of forensic evidence could not be completed at the time of their convictions, but for whom such testing could now be accomplished. The first summary of the project (Connors et al. 1996) described 28 men who had been incarcerated for 10–20 years but were exonerated by DNA testing. In 90 percent of these cases one or more eyewitnesses had falsely identified the person convicted. Many more incarcerated individuals have been similarly exonerated since the Connors et al. (1996) paper was published (Wells et al. 1998). It appears that the cautions expressed 100 years ago by early researchers finally have had an impact upon investigative practices.

Bibliography:

1. Connors E, Lundregan T, Miller N, McEwan T 1996 Convicted by Juries, Exonerated by Science: Case Studies in the use of DNA Evidence to Establish Innocence after Trial. National Institute of Justice, Alexandria, VA
2. Cutler B L, Penrod S D 1995 Mistaken Identification: The Eyewitness, Psychology, and the Law. Cambridge University Press, New York
3. Fisher R P, Geiselman R E 1992 Memory-enhancing Techniques for Investigative Interviewing. Charles C. Thomas, Springfield, IL
4. Goodman G S, Bottoms B L (eds.) 1993 Child Victims, Child Witnesses. Guilford Press, New York
5. Lindsay D S, Read J D, Sharma K 1998 Accuracy and confidence in person identification: The relationship is strong when witnessing conditions vary widely (as they do across real-world witnesses). Psychological Science 9: 215–8
6. Lindsay R C, Wells G L 1985 Improving eyewitness identifications from lineups: Simultaneous versus sequential lineup presentations. Journal of Applied Psychology 70: 556–64
7. Loftus E F 1979 Eyewitness Testimony. Harvard University Press, Cambridge, MA
8. National Institute of Justice (NIJ) 1999 Eyewitness Evidence: A Guide for Law Enforcement. US Department of Justice, Washington DC
9. Read J D 1995 The availability heuristic in person identification: The sometimes misleading consequences of enhanced contextual information. Applied Cognitive Psychology 9: 91–121
10. Ross D F, Read J D, Toglia M P 1994 Adult Eyewitness Testimony: Current Trends and Developments. Cambridge University Press, New York
11. Schooler J W, Engstler-Schooler T Y 1990 Verbal overshadowing of visual memories: Some things are better left unsaid. Cognitive Psychology 22: 36–71
12. Shepherd J W, Ellis H D, Davies G M 1982 Identification Evidence: A Psychological Evaluation. Aberdeen University Press, Aberdeen, UK
13. Thompson C P, Herrmann D, Read J D, Bruce D, Payne D G, Toglia M P 1998 Eyewitness Memory: Theoretical and Applied Perspective. Erlbaum, Mahwah, NJ
14. Wells G L 1993 What do we know about eyewitness identification? American Psychology 48: 553–71
15. Wells G L, Bradfield A L 1998 ‘Good, you identified the suspect’: Feedback to eyewitnesses distorts their reports of the witnessing experience. Journal of Applied Psychology 83: 360–76
16. Wells G L, Small M, Penrod S, Malpass R S, Fulero S M, Brimbacombe C A E 1998 Eyewitness identification procedures: Recommendations for lineups and photospreads. Law and Human Behavior 22: 603–47