Eyewitness Identification: Challenging a Confident Witness and Common Misconceptions
Published in Voice for the Defense, May, 2002

As she was being raped, Jennifer Thompson was possessed by a single thought: She would study her assailant's face, remember it, and bring him to justice. The prosecution found her an exceptional witness-educated, articulate, sympathetic, and confident. But she was wrong. Ronald Cotton served eleven years in prison until exonerated by DNA evidence.

The National Institute of Justice has recognized the problem with the publication of Convicted by Juries, Exonerated by Science: Case Studies in the Use of DNA Evidence to Establish Innocence After Trial. Of 28 wrongful convictions, 23 were convicted on eyewitness testimony-one defendant by five separate witnesses. Studies have shown that misidentifications rates in laboratory studies are typically 20-40%, and can be raised to 95% with aggressive techniques. Common sense suggests the eyewitness identification process is straightforward. Closer analysis and research indicate it is not.

Information Processing in Eyewitnesses
Errors may creep into the identification process in many places. Eyewitness accuracy is affected by some of the same factors that allow for better photography, such as lighting, distance, and viewing angle. But human information processing is not like a video camera, and memory is not like replaying a tape of the scene. Elizabeth Loftus and others have shown that what people report that they remember is influenced by their expectations at the time they observed the event, how they are questioned about it, and other information thatthey learn may support or invalidate their memory.

Expectations influence what people see-or think they see. Fans watching a football game see many more infractions from the opposing team than they do from their own. If a witness has a negative view of a local homeless man, and sees a vague figure generally similar in appearance commit a crime, there is a strong tendency to put two and two together.

The importance of leading questions (and nonverbal attitudes) comes into play when the witness gives a report to the police. It is crucial to learn about this process in each individual case. What was the verbal description given? Did the interviewer have a suspect in mind while asking the questions? Did the witness look at mug shots? In a book or were they handed a picture and asked, "Is this the guy?" Every occasion that the witness was exposed to the defendant's face should be discovered. Ideally, every verbal and nonverbal communication, before and after each exposure, would be discoverable-because it may subtly influence the witness' subsequent choice and confidence in the choice. What questions were asked of the witness? Leading questions can create biased descriptions. For example, witnesses to a motor vehicle accident who are asked to estimate how fast the cars were traveling when they smashed into each other provide higher estimate that those who are asked how fast the vehicles were traveling when they hit each other.

Next, examination of the line-up and the way it was conducted is required. Did the line-up foils know they were foils? (If so, they may act less guilty or nervous than actual suspects) Were they matched to the verbal description of the witness or to the suspect? (For a particularly bad example in Texas, see http://www.psychology.iastate.edu/faculty/gwells/badandgoodlineups.htm) Were the police who conducted the line-up "blind" to who was the suspect? We expect clinical trials of new drugs to be conducted "double-blind," so that doctors' and patients' expectations don't taint the results. The same standards should apply to criminal identification, where subtle (and not so subtle) communications may influence an eyewitness' choice and his or her confidence in that choice.

The National Institute of Justice recently issued guidelines regarding eyewitness evidence handling. The publication (available at http://www.ncjrs.org/pdffiles1/nij/178240.pdf) advises police agencies to impress upon witnesses the importance of clearing an innocent suspect, and that the perpetrator may not be present in the line-up that is shown. Still, witnesses may reasonably assume that the police would not go through the trouble of creating a lineup unless they had a legitimate suspect. Before the line-up, police should ask how confident a witness is about being able to identify the perpetrator, if he is present. However, witnesses are rarely asked if they could state, confidently, that the perpetrator is not there, if that is the case. A recent study suggests that adding this simple question can dramatically reduce rates of false identification.

Research suggests that many subjects, when faced with a line-up or photo spread, pick the suspect who looks most like their recollection of the perpetrator. In other words, the judgment is not that this IS the guy, but rather that this guy looks most like the guy I saw. This relative judgment is fine if the perpetrator is in the line-up. If not, an innocent person will be identified. For this reason, psychologists who study eyewitness processes recommend that the traditional photo spread or line-up be replaced with a sequential presentation of suspects and foils. The witness would be shown one photo at a time and asked, "Is this the guy?" This forces the witness to directly compare the each picture to the image in memory. If there is a match, one does not need to see other defendants to know this. Recognition is a rapid, automatic process. An identification that takes more than ten seconds is probably not based on recognition, but other factors. Numerous studies have shown that sequential presentation of mugshots substantially reduces false identifications and has little impact on accurate identifications. Although recognized in the National Institute of Justice recommendations, sequential presentation of suspects is not included as a formal standard. Apparently, prosecutors on the committee were concerned that their hometown police departments would not meet published standards. Sequential identification has been officially adopted by the state of New Jersey, but is not widely used elsewhere.

Another cause of misidentification, related to the previous issues is called "transference."

Transference occurs when an eyewitness remembers the suspect, but from another context. The subject is picked because he seems "familiar." Sometimes bystanders at a crime will be identified in this way-sometimes the person just knows he or she has seen the suspect before and the unfortunately bears a resemblance to the perpetrator.

The effect of stress on an eyewitness' accuracy has been studied extensively. Several studies found a "weapon focus" effect, where the presence of a gun or knife decreased the accuracy of identifications. Although it was initially believed the loss of accuracy was due to fear and stress, the weapons focus effect has been duplicated with a celery stick replacing the gun or knife. An exceptionally attractive, or ugly, accomplice can also reduce accuracy. The new and improved formulation is that reducing a witness' attention to the perpetrator's identifying characteristics is likely to decrease accuracy of the identification.

The eyewitness research literature now composes about one thousand studies and is growing rapidly. A typical issue of Law and Human Behavior has two new studies. Fortunately, a helpful summary expert opinion was published recently. Table 1 identifies factors were identified as reliably demonstrated in the scientific literature by at least 80% of 64 "blue ribbon" experts in eyewitness testimony (Kassin, et al., 2001), and the percentage of experts agreeing that the position articulated is adequately established in the scientific literature.

Table 1

Factor	Percentage agreeing	Is this common sense?
Wording of questions will influence accuracy of identifications	98	25
Line-up instructions will influence identifications	98	39
The eyewitness's confidence is malleable	95	10
Showing the defendant's mug shot will increase subsequent lineup identifications	95	13
Information the witness receives after the crime can influence his identification	94	17
Children are more suggestible	94	73
Eyewitnesses' attitudes and expectations influence what they see, remember, and report	92	31
Hypnosis increases suggestibility and susceptibility to false information	91	19
Intoxication reduces eyewitness accuracy	90	95
Cross-race identifications are less accurate	90	65
A weapon reduces accuracy of identifications (weapon focus)	87	34
There is little relation be a witness's confidence and the accuracy of his identification	87	5
Memory deteriorates rapidly; most information lost over first few hours and days	83	29
Greater exposure time leads to greater accuracy	81	97
Sequential presentation format less prone to false identifications	81	0
Witnesses may falsely identify someone who seems "familiar" (transference)	81	19

Eyewitness Confidence
A confident eyewitness can be highly persuasive. Everyday experience tells us that, for normal adults with no incentive to lie, a person's confidence in his or her ability to fix your air conditioner, for example, is related to actual competence. However, this assumes that people have experience and feedback about their performance. Eyewitnesses do not. The conclusion of most psychological research, until very recently, was that there is no relationship between witnesses' confidence and their accuracy. This position has evolved into a more conservative claim: that the relationship between confidence and accuracy is modest-but positive.

It is common for an eyewitness' confidence to increase substantially between the time of the first identification and the trial. In the Dror Goldberg case, victim Roberta Ingrando initially told police she was 80% sure of her identification. In two subsequent court appearances, she described herself as "totally sure." Repeated, suggestive exposures to the suspect will likely lead to increases in feelings of familiarity, which may inflate confidence and be confused with a correct identification. The eyewitness may be told that his or her testimony is vital to convict a criminal who the police "know" is guilty. An identification that is immediately followed by a cheer in the next room, within earshot of the witness, is likely to have a dramatic effect on the witness' confidence in the identification. Once on the witness stand, it would take a courageous witness to say he or she isn't sure. Further, eyewitnesses whose confidence is artificially inflated (by cues of the experimenters) tend to overestimate all the factors that lead up to good recognition, such as the time they viewed the suspect and lighting conditions. In other words, they seek reasons to justify their level of confidence.

Admissibility
Eyewitness factors testimony should provide information to the trier of fact that is not available through common sense. Some have argued that eyewitness research does not meet this criterion. However, it is also true that some of the best-supported findings flatly contradict common perceptions. Table 1 shows that of the sixteen eyewitness factors deemed most reliably demonstrated, only four (accuracy of children, intoxication, cross-racial identification, and exposure time) approach being "common-sense."

Conclusion
An eyewitness' recollection is fragile forensic evidence. It will degrade over time, more rapidly at first, and can easily be contaminated by suggestive investigative procedures. Even without these problems, an eyewitness will sometimes be wrong. There are now clear guidelines. in place established by the Department of Justice, and these represent standards of practice for police investigators. Other recommendations from research psychologists (sequential presentation) are not implemented in Harris County or most other jurisdictions. An energetic defense lawyer should be able to identify multiple factors that make an eyewitness identification less convincing than it once was.

Essential References
Loftus, E. & Doyle, J. (1997). Eyewitness Testimony: Civil and Criminal. Charlottesville: Lexis Law Publishing. This book reviews the scientific literature and then walks the defense attorney through the implications for every step of the adjudication process. Expensive ($100) but worth every penny.

Kassin, S. M., Tubb, V. A., Hosch, H. M., & Memon, A. (2001). On the "General Acceptance" of Eyewitness Testimony Research: A New Survey of the Experts. American Psychologist, 56 (3), 405-416.

http://www.psychology.iastate.edu/faculty/gwells/homepage.htm Web page of Gary Wells, Ph.D., renowned expert on eyewitness issues, has tons of information, publications, examples and demos.

Eyewitness Evidence: A Guide for Law Enforcement. U.S. Department of Justice publication; can be downloaded from http://www.ncjrs.org/pdffiles1/nij/178240.pdf