Problems in Using Eyewitness Testimony

The criminal justice system heavily relies on eyewitnesses and eyewitness testimony. Eyewitnesses are used to identify culprits and remember conversations or various details, and sometimes provide the only evidence there is in terms of identification. In such cases, these accounts are generally accepted by police, prosecutors, judges, and juries. And aside from a “smoking gun,” nothing carries as much weight as the testimony of an actual witness. However, implicit in the acceptance of this testimony is the assumption that the mind precisely records and stores events. While people tend to believe that our memories are preserved intact and our impressions are never forgotten, the truth is that even if we are careful observers and take in an accurate picture of some object or experience, it doesn’t necessarily remain intact. And with the advent of new DNA testing, it is becoming more and more obvious that this is the case. This paper will discuss the problems of using eyewitness testimony in light of new technological and psychological developments.

Even before DNA testing, psychologists were questioning the validity of eyewitness reports. Hugo Mu¨nsterberg’s research in the early part of the 20th century strongly advocated the involvement of psychological science in terms of understanding the vagaries of eyewitness testimony (Wells, Memon and Penrod, 2006). However, it wasn’t until the 1970s that psychologists started conducting experiments aimed at understanding the extent of error found in eyewitnesses accounts, many of which resulted in articles containing strong warnings that such accounts were being overvalued by the criminal justice system (ibid).

Using various methodologies, including filmed events and live staged crimes, researchers realised that mistaken identification rates were extremely high and that eyewitnesses would often express certainty even when they were wrong (Wells and Olson, 2003). Tests conducted in 1979 by Elizabeth Loftus showed a 54% swing from a non-guilty to a guilty verdict with the addition of an eyewitness, which illustrates the strength of such testimony (Kennedy and Haygood, 1992). Around the same time, eyewitness research was evolving in the UK, prompted by the Devlin Committee (Devlin, 1976).

Much of the research focused on the notion that, with the passage of time, or with the introduction of contradictory facts, memories change or become transformed, often without conscious awareness (Loftus and Ketcham, 1991). In fact, individuals can actually come to believe in memories that never happened. For example, child psychologist Jean Piaget, in his book Plays, Dreams, and Imitation in Childhood, describes a personal story about the malleability of memory:

“One of my first memories would date from my second year. I can still see, most clearly, the following scene, in which I believed until I was about fifteen. I was sitting in my pram, which my nurse was pushing in the Champs Elysees, when a men tried to kidnap me. I was held in by the strap fastened around me while my nurse bravely tried to stand between me and the thief. She received various scratches, and I can still see vaguely those on her face. Then a crowd gathered, a policeman with a short cloak and a white baton came up, and the man took to his heels. I can still see the whole scene, and can even place it near the tube station. When I was about fifteen, my parents received a letter from my former nurse saying that she had been converted to the Salvation Army. She wanted to confess her past faults, and in particular to return the watch she had been given as a reward on occasion. She had made up the whole story, faking the scratches. I, therefore, must have heard, as a child, the account of this story, which my parents believed, and projected into the past in the form of a visual memory” (Piaget, 1962).

Another theme that emerged during this time was the distinction between system variables and estimator variables. The argument was that some of the variables that affect the accuracy of eyewitness reports were under the control of the justice system (system variables) while others were not (estimator variables) (Wells, 1978). Additionally, researchers were concerned about the fact that while there were scientific procedures in place to handle physical evidence (fingerprints, fibers, blood, etc.) there were no corresponding procedures to handle eyewitness evidence. The worry was that, like physical evidence, “memory trace evidence” could be lost, contaminated, destroyed or misinterpreted, leading to incorrect results (Wells and Loftus, 2003).

Although defense attorneys seized on the growing amount of research dispelling the validity of eyewitness accounts, it was largely ignored by prosecutors, judges and law enforcement until the 1990s, when forensic DNA testing began to uncover numerous cases in which innocent individuals were convicted based on mistaken eyewitness accounts (Scheck, Neufeld and Dwyer, 2000). However, because most crimes do not have DNA evidence, reliance on eyewitness identification is still widely prevalent, and has not been significantly diminished by the use of DNA tests. And so, many people are still being convicted even though they are innocent, with estimates as high as 45% (Loftus and Ketcham, 1991).

One of the reasons for this has to do with the highly unscientific nature of police lineups. When conducting a lineup, nonsuspect (filler) members are chosen. Ideally, these fillers would be chosen so that an innocent suspect is not mistakenly identified merely from “standing out,” and so that a culprit does not miss being identified merely by blending in (Brewer and Wells, 2006). In experiments where fillers did not at all resemble the culprit, eyewitnesses tended to mistakenly identify an innocent suspect who resembled them; when the suspect was the culprit, however, the manipulation of fillers had little effect on the number of correct identifications (Lindsay and Wells, 1980).

In such demonstrations, researchers used their knowledge of the culprit’s identity in choosing fillers, whereas in actual cases, the identity of the culprit is unknown. Using the suspect as a proxy for the culprit will have different effects on the rates of accurate or inaccurate identification (depending on the guilt or innocence of the suspect), and while selecting fillers who are similar to the suspect can help protect them in a culprit-absent lineup, they can also reduce accurate identifications in a culprit-present lineup (Luus and Wells, 1991). Other areas of research have shown that using the suspect as the reference point to select fillers can create a “backfire effect” in which the innocent suspect, being the “origin” of the lineup, actually has an increased chance of being chosen as the culprit (Clark and Tunnicliff, 2001).

An alternative strategy to selecting fillers based on their resemblance to the suspect is to choose them based on their similarity to the verbal description provided by the eyewitness; this fit-to-description strategy has several practical advantages and has worked well in some experiments (Wells, Rydell, and Seelau, 1993). However, biases against the innocent suspect can remain when the description is especially sparse or when the innocent suspect happens to show a high resemblance to the culprit (Clark and Tunnicliff, 2001).

Another problem has to do with interference from the lineup administrator. Commonly, such a person knows which member is the suspect and which members are fillers, and could inadvertently communicate this knowledge to the eyewitness through verbal and nonverbal means, including casual remarks and subliminal body messages (Chartier, 1997). This kind of feedback can artificially increase eyewitness confidence level by leading to false perceptions (“if the police believe it, then they must be right;” “after all, if they did not have a suspect they would not require the assistance of the witness”) (ibid).

For example, in one experiment, researchers manipulated lineup administrators’ assumptions about the identity of the culprit and found that this manipulation affected the choices that eyewitnesses made (Phillips, McAuliff, Kovera and Cutler, 1999). In another, it was found that post-identification suggestions to eyewitnesses from lineup administrators led them to develop high levels of false certainty that they had made a positive identification (Bradfield and Wells, 1998). In addition, tests conducted by the National Science Foundation revealed that once memory has been distorted in such a way, a straightforward cross-examination will often fail to distinguish inaccurate recollections, and eyewitnesses will even change their original opinions (Chartier, 1997). One reason for this is that:

“When you place a victim of a serious offense together with a police investigator, there is a unification of goals. Both have a desire to solve the crime, but if the victim has a very strong desire to solve that crime – as most victims of serious crimes do – they become especially sensitive to any feedback they may receive. And so, if the investigator has a suspect in mind, there remains the possibility that they may communicate that idea to the victim unwillingly” (Loftus, 1998).

The results of this kind of interaction are referred to as suggestibility, which is the tendency of an individual to accept uncritical information during questioning, or to simply comply with what they think the interviewer wants to hear (Sadava and McCreary, 1997). In addition, some interviewers adopt “self-fulfilling prophecy techniques,” in which case questions are specifically designed to generate flawed results, and any information that does not meet these expectations is ignored (ibid). A 1986 study by Kwock and Winer found that suggestibility depends on the identity of the person asking the questions, along with assertiveness and reactive responses, and that those designated as experts are more likely to influence responses (ibid).

Not surprisingly, researchers have invested a great deal of effort in trying to identify the variables that sometimes act as independent markers of identification accuracy (Sporer, 1993). One important variable in this regard is the eyewitness’s confidence in his or her identification response; police, attorneys and judges all share the view that confidence is an important factor in determining eyewitness accuracy (Potter and Brewer, 1999). In addition, mock-juror studies have found that confidence has a major influence on the assessment of witness credibility and verdicts (Bradfield and Wells, 2000).

Eyewitness researchers, however, have generally viewed eyewitness confidence as not a particularly useful indicator of identification accuracy, and most analyses label the average CA (confidence-accuracy) relationship as around 0 to .3 (Kassin, Tubb, Hosch, and Memon, 2001). Various arguments have been advanced to explain this. For example, one study detected:

“Greater observer confidence for faces shown brighter at test than at encoding although accuracy was higher when brightness at test matched that at encoding. This was attributed to observers’ confidence judgments being influenced by analytic (or metacognitive) cues when there was a mismatch between encoding and test stimuli. In a similar vein, superior exposure conditions at the time of the crime may increase the likelihood of an accurate identification but (to the extent that the witness relied upon such metacognitive cues when assessing confidence) may produce an even more marked increase in witness confidence” (Busey, Tunnicliff, Loftus and Loftus, 2000).

Similarly, another study found that with longer exposure to the offender, identification accuracy improved, but “greater confidence for correct relative to incorrect identifications was no longer evident for targetpresent lineups” (Memon, Hope and Bull, 2003).

Other factors include the tendency for people to seek out only confirmatory information about their judgments (confirmatory bias), thereby producing overconfidence (Koriat, Lichtenstein, and Fischhoff, 1980); the fact that judgments of uncertainty (quantitative or qualitative) cannot be made reliably because of the inevitable failure to take into account the influence of unavoidable possibilities or scenarios that should guide these judgments (Tversky and Koehler, 1994); the difficulty that individuals have in accurately translating their internal or subjective states of confidence onto a numerical scale (Gigerenzer, Hoffrage, and Kleinbölting, 1991); the possibility that social factors, such as feedback from lineup administrators or other witnesses, may affect confidence judgments for identification responses, independent of any effect on accuracy (Semmler, Brewer and Wells, 2004); and the possible contribution of individual difference variables, such as gender and personality on the realism of people’s assessments of their own abilities (Furnham and Thomas, 2004).

However, it must be noted that the majority of research in this field does not subscribe to the view that people are always overconfident or underconfident in their judgments. For example, it has been reported that there is considerable agreement between “participants’ self-assessments of fluid and crystallized abilities and a diverse array of knowledge measures and objective measures of those abilities and knowledge domains. Not only did participants recognize their areas of strength, but they also discriminated those from their areas of weakness” (Ackerman, Beier and Bowen, 2002).

In general, there are several ways to remedy false identification. For example, in suspect lineups, the eyewitness might be informed that there is a possibility that the culprit is not present, which will remove the “one of these individuals is certainly guilty, otherwise I wouldn’t be here” mentality (Williams, 1999). In doing so, identification would no longer be based purely on similarity defaults, and would put a stop to the selection of innocent individuals simply because they looked like the culprit (ibid). In addition, because a wealth of scientific evidence suggests that the interviewer’s expectations and hopes unintentionally influence the respondent, it would be wise to use interviewers who are unaware of the facts of the case (Wells, 1998). This will remove the possibility of the interviewer inadvertently making suggestions as to who is guilty and who is not via direct or indirect vocal or body language.

Many alternatives to the traditional lineup have been proposed and tested. One of the first was the blank lineup control method, in which there are only fillers and no suspect (Wells, 1984). If the witness identifies someone from this lineup as being the culprit, they can be discarded, whereas if they do not make an identification, they can then be shown the actual lineup, which would include the suspect (ibid). Data indicate that eyewitnesses who do not make an identification from the blank lineup are more reliable on the second (actual) lineup than are those who were not screened during the blank lineup (Wells and Olson, 2003).

Another possible alternative is the sequential lineup. Here, instead of the lineup members being shown to the eyewitness simultaneously, they are shown only one member at a time, in which case the eyewitness is required to make a decision (“Is this person the culprit?”) prior to viewing the next member (Lindsay and Wells, 1985). Theoretically, the sequential lineup prevents eyewitnesses from selecting the person who looks most like the culprit relative to the other lineup members (relative-judgment decision) (Wells 1984). Thus, unlike the simultaneous lineup, the sequential lineup prevents eyewitnesses from making a relative-judgment decision, because at any point a person has not yet been seen may turn out to resemble the culprit more than any person previously viewed; eyewitnesses must compare each member of the sequential lineup to their memory of the culprit, not to each other (Wells and Olson, 2003). An analysis of twenty-five studies comparing simultaneous and sequential lineups showed that the sequential lineup reduced the chances of mistaken identifications in culprit-absent lineups by almost half (Steblay, Dysart, Fulero and Lindsay 2001).

The guidelines for suspect lineups are also applicable to photospreads:

“At no time should the tester initiate any form of interaction with the witness who is attempting to distinguish the true offender, and comments such as ‘Take a closer look at this one’ should never be used; in fact, the only comment that should ever be made would be ‘Remember, the suspect may not be present within this photospread.’ This will remove any pressures from the witness, and ultimately lead to a better selection. As with lineups, it is important to question the witness as to their conclusion. Accurate witnesses tend to have difficulty in providing descriptions as to how they reached their decisions, whilst inaccurate witnesses tend to waffle about their choice, explaining their process of elimination (e.g. ‘I compared photos to each other in order to narrow down my selection’)” (Dunning and Stern, 1994).

In terms of research, one problem that needs to be remedied is the fact that the eyewitness identification literature has been driven more by practical perspectives than by theoretical frameworks. The focus on application and forensic relevance has not only reduced the sharing of ideas between researchers and their counterparts in psychology and social psychology, but inadvertently created gaping holes in the research; thus, better theory is necessary to “generalise” this body of literature and to cover every possible situation that arises in actual cases (Wells and Olson, 2003).

Another concern is that while laboratory data on eyewitness identification are extensive, some key forms of real-world data are lacking:

“Certain estimable rates of eyewitness identification behavior and lineup conditions from actual cases could assist the design and interpretation of laboratory work. For instance, there have been no empirical estimates of the base rate for culprit-present versus culprit absent lineups in actual cases. Although it is difficult to establish the ground truth (actual guilt or innocence) needed for precise estimates of this base rate in actual cases, methods exist for estimating upper limits” (Wells and Olson, 2002).

Future research needs to address the gap between psychological science and the traditional practices of the legal system. One way to do this would be to examine the theories by which the latter uses in collecting eyewitness evidence (Wells and Loftus, 2003). An understanding of these theories, no matter how implicit, can help researchers to better communicate their findings, along with having them better received. In addition, this will allow for a better chance to actually affect how the justice system works (ibid).

Eyewitness testimony has an enormous impact on the outcome of a trial. Unfortunately, because it relies on human memory (which if fallible), its use is highly problematic. Analyses of recent and numerous DNA exoneration cases reveal that mistaken eyewitness identification was involved in the vast majority of these convictions, accounting for more convictions of innocent people than all other factors combined. However, as long as police procedures are accurately followed and research and technology continue to be developed and properly implemented, eyewitness testimony can still play a beneficial role in the criminal justice system. Nearly one hundred years ago, Hugo Mu¨nsterberg stated that psychology could help the legal system sift through the “chaos and confusion” of eyewitnesses. His claims were dismissed by many, including legal scholar J.H. Wigmore, but only because he felt that psychology did not yet have the tools to properly handle the problem of evaluating eyewitness accounts. He did believe, however, that in the future, that time would come. And by all accounts, it has.

Bibliography

Ackerman, P. L., Beier, M. E., and Bowen, K. R. “What we really now about our abilities and our knowledge.” Personality and Individual Differences, 33, 2002.

Brewer, N. and Wells, G.L. “The confidence-accuracy relationship in eyewitness identification: Effects of lineup instructions, foil similarity and target-absent base rates.” Journal of Experimental Psychology: Applied, 12, 2006.

Busey, T. A., Tunnicliff, J., Loftus, G. R., and Loftus, E. F. “Accounts of the confidence-accuracy relation in recognition memory.” Psychonomic Bulletin and Review, 7, 2000.

Charman, S.D. and Wells, G.L. “Eyewitness Lineups: Is the Appearance-Change Instruction a Good Idea?” Law and Human Behavior, Nov, 2005.

Chartier, G. “False identification: New research seeks to inoculate eyewitnesses against errors.” National Science Foundation, 1997.

Clark, S.E., Tunnicliff, J.L. “Selecting lineup foils in eyewitness identification: experimental

control and real-world simulation.” Law and Human Behaviour, 25, 2001.

Devlin, L. “Report to the Secretary of State for the Home Department of the Departmental Committee on Evidence of Identification in Criminal Cases.” London: Her Majesty’s Stationery Office, 1976.

Dunning, D., and Stern, L. B. “Distinguishing accurate from inaccurate eyewitness identifications via inquiries about decision processes.” Journal of Personality and Social Psychology, 67, 1994.

Furnham, A., and Thomas, C. “Parents’ gender and personality and estimates of their own and their children’s intelligence.” Personality and Individual Differences, 37, 2004.

Gigerenzer, G., Hoffrage, U., and Kleinbölting, H. “Probabilistic mental models: A Brunswikian theory of confidence.” Psychological Review, 98, 1991.

Kassin, S. M., Tubb, V. A., Hosch, H. M., & Memon, A. “On the ‘general acceptance’ of eyewitness testimony research: A new survey of the experts.” American Psychologist, 56, 2001.

Kennedy, T. D., Haygood, R. C. “The discrediting effect of eyewitness testimony.” Journal of Applied Social Psychology, 1992.

Koriat, A., Lichtenstein, S., and Fischhoff, B. “Reasons for confidence.” Journal of Experimental Psychology: Human Learning and Memory, 6, 1980.

Lindsay, R.C.L., and Wells, G.L. “Improving eyewitness identification from lineups: Simultaneous versus sequential lineup presentations.” Journal of Applied Psychology, 70, 1985.

Lindsay, R.C.L., and Wells, G.L. “What price justice? Exploring the relationship between lineup fairness and identification accuracy.” Law and Human Behavior, 1980.

Loftus, E. Eyewitness Testimony (2nd ed.). Cambridge, MA: Harvard University Press, 1996.

Loftus, E. and Ketcham, K. Witness For the Defense: The Accused, the Eyewitness. New York: St. Martin’s Press, 1991.

Loftus, E. “What Jennifer saw.” Frontline, 1998.

http://www.pbs.org/wgbh/pages/frontline/shows/dna/interviews/loftus.html

Luus, C.A.E., & Wells, G.L. “Eyewitness identification and the selection of distracters for lineups.” Law and Human Behavior, 15, 1991.

Memon, A., Hope, L., and Bull, R. “Exposure duration: Effects on eyewitness accuracy and confidence.” British Journal of Psychology, 94, 2003.

Phillips, M.R., McAuliff, B.D., Kovera, M.B., Cutler, B.L. “Double-blind photoarray administration

as a safeguard against investigator bias.” Journal of Applied Psychology, 84, 1999.

Piaget, J. Plays, Dreams and Imitations in Childhood. New York: W.W. Norton and Co., Inc, 1962.

Potter, R., and Brewer, N. “Perceptions of witness behaviour-accuracy relationships held by police, lawyers and jurors.” Psychiatry, Psychology and Law, 1999.

Sadava, S., and McCreary, D. Applied social psychology. Saddle River: Prentice Hall, 1997.

Scheck, B., Neufeld, P., Dwyer, J. Actual Innocence. New York: Random House, 2000.

Semmler, C., Brewer, N., and Wells, G. L. “Effects of postidentification feedback on eyewitness identification and nonidentification confidence.” Journal of Applied Psychology, 89, 2004.

Sporer, S.L. “Eyewitness identification accuracy, confidence, and decision times in simultaneous

and sequential lineups.” Journal of Applied Psychology, 78, 1993.

Steblay, N.M., Dysart, J., Fulero, S., and Lindsay, R.C.L. “Eyewitness accuracy rates in sequential and simultaneous lineup presentations: A meta-analytic comparison.” Law and Human Behavior, 25, 2001.

Tversky, A., and Koehler, D. J. “Support theory: A nonextensional representation of subjective probability.” Psychological Review, 101, 1994.

Wells, G.L. “Applied eyewitness testimony research: System variables and estimator variables.” Journal of Personality and Social Psychology, 36, 1978.

Wells, G.L. and Bradfield, A.L. “Good, you identified the suspect: Feedback to eyewitnesses

distorts their reports of the witnessing experience.” Journal of Applied Psychology 83, 1998.

Wells, G.L. and Loftus, E. “Eyewitness Memory for People and Events” in A.M. Goldstein (ed.) Handbook of Psychology. Vol 11 Forensic Psychology. New York: John Wiley & Sons, 2003.

Wells, G.L., Malpass, R.S., Lindsay, R.C., Fisher, R.P., Turtle, J.W., and Fulero, S.M. “From the lab to the Police Station: A Successful Application of Eyewitness Research.” American Psychologist, Vol 55, No. 6., 2000.

Wells, G.L., Memon, A., Penrod, S.D. “Eyewitness Evidence: Improving its Probative Value.” Association for Psychological Science, Vol 7 No. 2, 2006.

Wells, G. L., and Olson, E. A. “Eyewitness identification: Information gain from incriminating and exonerating behaviors.” Journal of Experimental Psychology: Applied, 8, 2002.

Wells, G.L. and Olson, E.A. “Eyewitness Testimony.” Annual Reviews Psychology, Iowa State University, Psychology Department, 2003.

Wells, G.L. “The psychology of lineup identifications.” Journal of Applied Social Psychology, 14, 1984.

Wells, G.L., Rydell, S.M., Seelau, E.P. “On the selection of distractors for eyewitness lineups.”

Journal of Applied Psychology, 78, 1993.

Williams, K. “Psychology and Law: Eyewitness Testimony.” University of Sydney, 1999. http://www.psy.unsw.edu.au/~kipw/Lectures/3rdYearPsychLaw/PL7-Eyewitness_files/v3_document.htm