Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Eyewitness Identification Expert Summary Report Prepared by Dr. Jennifer Dysart I. Brief Overview of Eyewitness Memory Research Over a period of several decades, researchers have established that when we experience an important event, we do not simply record it in our memory as a video recorder would. The situation is much more complex. Most theoretical analyses of the memory process divide it into three major stages. First, the event is perceived by a witness and information is entered into the memory system. Next, some time passes before a witness tries to remember the event. Finally, the witness tries to retrieve the stored information. Psychologists who conduct research in this area try to identify and study the important factors that play a role in each of the three stages. Numerous factors at each stage can affect the accuracy of an eyewitness account. Some of the factors that have been shown scientifically to affect eyewitness performance include: distance, the duration of an event, high levels of stress and arousal, and length of the retention interval. These factors, which are related to the event and the witness, fall under the umbrella of “estimator variables”. As it relates to law enforcement procedures, research has shown that the procedures and practices used by police during the third (retrieval) stage of the memory process can also influence the reliability of an eyewitness’ identification and subsequent testimony. These factors are referred to as “system variables” because they are under the direct control of the criminal justice system. Examples of system variables include pre-lineup/photoarray1 instructions, type of lineup administered (simultaneous or sequential), whether the identification was conducted using a double-blind administrator, and the type of post-event information provided to a witness after their identification decision. Over the past four decades, over 2000 scientific research articles, book chapters and books have been written on the subject of eyewitness identification accuracy. It is not possible in this report to detail or summarize this extensive body of research. A thorough summary can be found in “Eyewitness Testimony: Civil and Criminal, 5th Edition” by Loftus, Doyle and Dysart (2014, Lexis Nexis). At the end of each section describing a specific estimator or system variable, I will present information on the issue of general acceptance within the field of eyewitness identification on that variable. The sources that contribute to my opinions with respect to general acceptance on each variable include, but are not limited to, published research in the field, conference presentations, conversations with other eyewitness experts, and published surveys of eyewitness 1 The terms “lineup” and “photoarray” are used interchangeably in this report. 1 experts’ opinions about the reliability of these topics. In particular, I will be citing an article published by Kassin and his colleagues (2001) titled On The “General Acceptance” of Eyewitness Testimony Research: A New Survey of the Experts (hereafter Kassin survey, Attachment 2). The Kassin survey is the most recent published survey of expert opinions on eyewitness research and is frequently used in courts around the United States to support opinions of general acceptance on the topics describe below. Although this survey was published several years ago, the opinions of eyewitness experts have remained consistent on these topics. There are no recent surveys of expert opinions that would undermine the validity of the Kassin survey. Further, with respect to system variables in particular, many criminal justice agencies and practitioners also support the best practices described below (e.g., IACP, NIJ, several states and numerous police departments). For instance, the National Institute of Justice (US Department of Justice) published a research report in 1999 titled “Eyewitness Evidence: A Guide for Law Enforcement” (https://www.ncjrs.gov/pdffiles1/nij/178240.pdf), hereafter NIJ Eyewitness Guide. The NIJ Eyewitness Guide was a consensus document that was created through the work of a Task Force that was constituted of law enforcement (50% of the membership), prosecutors, defense attorneys and eyewitness experts. After its publication, the NIJ Eyewitness Guide was distributed to all law enforcement agencies in the United States. NIJ then published an Eyewitness training manual for law enforcement (2003), which can be downloaded from the internet. • Research Methodologies Used in Eyewitness Research In general, eyewitness identification researchers employ several techniques to come to the scientific conclusions that will be discussed in this report. The three most common research techniques are laboratory research, archival research and meta-analyses. The most common form of eyewitness identification research is experimental laboratory research. The primary reason for conducting experimental research is that it gives researchers the ability to make cause and effect statements, such as “Y happened because of X.” For example, a well-conducted experiment can tell us that using a specific identification procedure will cause an increase in identification accuracy. Archival research involves the examination of existing records or data from actual cases. This type of research is important for understanding how witnesses in actual cases behave. For example, archival studies have demonstrated that approximately 20% of witnesses in real cases select a lineup filler or stand-in, rather than the suspect. Thus, these records show us that nearly one in five eyewitnesses make an identification error by selecting a known innocent person. These results are consistent with results from laboratory studies, which have found very similar rates of erroneous filler selection. The study of DNA exonerations in the United States, discussed below, is an example of archival research. A third research technique that psychologists and other researchers employ is the meta-analysis. Generally, a meta-analysis is a statistical summary of research that has already been conducted, as opposed to the collection of new data with participants in a new experiment. Although the specific procedures employed by researchers in one meta-analysis may differ from those used in another meta-analysis, there are common elements to all meta-analyses. A benefit of using the 2 meta-analysis technique is that it informs researchers about eyewitness performance over the course of a large number of studies, from many researchers and from different laboratories (and perhaps from many different countries around the world). • Eyewitness Error Rates in Actual Cases By examining actual cases, we have learned some important facts about eyewitness errors. According to the Innocence Project, there have been mistaken eyewitness identifications in over 75% of DNA exonerations – which currently number 337. See www.innocenceproject.org. Brandon L. Garrett (2011), a law professor at the University of Virginia, has systematically examined the first 250 DNA exoneration cases in the United States. Professor Garrett reports that the leading contributing factor in these wrongful convictions was erroneous eyewitness identification, which occurred in 76% of the cases. In a quarter of all wrongful convictions studies, eyewitness testimony was the only direct evidence against the defendant. Further, in the 190 cases where there was an erroneous eyewitness identification of the defendant, 36% included mistaken identification from more than one eyewitness. In fact, some of the cases had as many as five eyewitnesses who mistakenly testified that the defendant was the perpetrator. • Eyewitness Error Rates in Archival Studies Archival studies also show that eyewitness identifications are often unreliable. In recent years, researchers have begun analyzing records of actual eyewitness identifications and attempted identifications. Unfortunately, when using archival data and police records, it is not possible for researchers to determine when a suspect identification is correct, but it is possible to determine error rates as reflected in the false identification of non-suspect fillers. A properly constructed lineup includes only one suspect (who might or might not be the actual perpetrator) and everyone else is a known innocent filler who would most likely not be charged with the crime in question even if they were identified by the eyewitness. According to scientific psychological research and the NIJ Eyewitness Guide (described above) it is critical to have only one suspect in each lineup so that law enforcement can assess whether a particular eyewitness is reliable. When an eyewitness identifies a lineup filler (which law enforcement should know to be an innocent person), police will know that that witness is unreliable. If all of the lineup members were potential suspects, it would be impossible for police to determine if an eyewitness is accurate or merely guessing – as any identification would be categorized as a “positive ID”. This is particularly important when one considers the findings from field studies with real witnesses presented below. While false identifications of innocent fillers do not necessarily send innocent fillers to prison, these are identification errors and provide useful information about the accuracy of eyewitness identifications and the reliability of lineup procedures. In a 2006 paper, Drs. Gary Wells, Amina Memon and Steven Penrod summarized the findings of four studies of actual eyewitnesses to serious crimes: 3 • • • • Wright and McDaid (1996) analyzed 1,561 lineup outcomes in London and found filler-identification rates of 19.9%. These data are similar to the 21% filler identification rate reported by Slater (1994) in a study of 843 lineups conducted by the Metropolitan Police in London. Behrman and Davey (2001) reported that 24% of identifications from live lineups in Sacramento, California, were identifications of fillers. Valentine, Pickering, and Darling (2003) analyzed 119 lineups in the greater London area and found that 21.6% of the eyewitnesses identified fillers. Wells, et al., underscored that these archival results are “a very important complement to the experimental studies of eyewitnesses” for the following reasons: • • • First, they indicate filler-identification results that are quite consistent with rates obtained in experiments (Ebbeson & Flowe, n.d.; Steblay, Dysart, Fulero, & Lindsay, 2001). Second, these archival results address a common criticism of experiments— namely, that participant witnesses in experiments are not as cautious as actual crime witnesses are, because the consequences of a mistaken identification in an experiment are not serious. But the witnesses in the archival studies were actual witnesses to crimes and yet mistakenly identified fillers in one third of their positive identifications. Third, the filler-identification rates in the archival studies permit us to make conservative estimates of the risk that an innocent suspect would face in these lineups. The archival studies (of actual cases) thus underscore several important points: (1) Filler identification rates are fairly high—approximately 20% across studies; (2) To the extent witnesses are guessing fillers, we have to expect other witnesses are guessing suspects—some of whom are innocent (research by Penrod and his colleagues confirms this finding); (3) Innocent suspect misidentification rates - as a result of guessing - are actually greater than the assumed one-in-six risk because police arrays tend to be biased against suspects. II. Estimator and System Variable Research The following sections are summaries of the scientific literature on the most common estimator and system variables that are relevant to adult eyewitness identification cases. These variables include: Estimator Variables: 1. Effects of brief exposure on eyewitness accuracy; 2. Weapon focus effect; 3. The effects of stress/fear on memory; 4. Cross-race identification; 5. Effects of distance on accuracy; 6. Effects of delay on accuracy; 4 System Variables: 7. Pre-identification lineup instructions; 8. The effects of filler bias on eyewitness accuracy; 9. Non-blind lineup administration; 10. Use of simultaneous lineup rather than a sequential lineup; 11. Unconscious transference and commitment effects; 12. Witness confidence and accuracy; 13. The post-identification feedback effect; • Effects of brief exposure on eyewitness accuracy Common sense might suggest that even a brief opportunity to view a stranger allows us to form a mental snapshot of that person, but research shows that the amount of time that a witness views a person (perpetrator) is positively associated with the witness’s ability to subsequently identify him. Further, what is critical with respect to accuracy is the witness’ opportunity to see the perpetrator(s) at the time of the event. In their 1986 meta-analysis, Shapiro and Penrod found a systematic relationship between exposure time and identification accuracy. Since this meta-analysis others (e.g., Memon, Hope & Bull, 2003) have replicated the positive correlation between exposure time and identification accuracy. The effects of exposure time are well illustrated in a study by Memon, Hope and Bull (2003) in which mock witnesses viewed a video of a realistic crime which lasted either one minute, forty seconds (with the perpetrator’s face in view for 45s) or one minute and seven seconds (with the perpetrator’s face in view for 12s). Witnesses were tested with target-present and target-absent arrays 40 minutes later. As shown in the following table, the proportion of correct identifications and correct rejections in target-absent arrays increased substantially when exposure time increased. (Note, however, that mistaken identifications in target-absent arrays remained relatively high regardless of the exposure time.) Performance of Young Adults (ages 17-25) and Seniors (ages 59-81) in the 12s and 45s Exposure Groups with Target-Present an Target-Absent Lineups Target Present Young Seniors Target Absent Young Seniors 12 Seconds Exposure Hits False NonAlarm Choice 45 Seconds Exposure Hits False NonAlarm Choice .29 .35 .42 .45 .29 .20 .95 .85 .05 .10 .00 .05 --- .90 .80 .10 .20 --- .41 .50 .59 .50 Note. Identification Errors are bolded 5 The results of the Memon et al. study above show that in circumstances where witnesses viewed the perpetrator’s face for 45 seconds, approximately 45% of all witness made a mistake and misidentified an innocent person from a lineup in which the actual perpetrator was not shown. When the exposure time was reduced to 12 seconds, the false identification of innocent people increased to an average of 85%. Further, the ability of witnesses to correctly identify the actual perpetrator when he was shown dropped by over 50% when the exposure time was reduced from 45 seconds to 12 seconds. The topic of exposure time was addressed in the Kassin survey. A full 81% of experts believed that the proposition that the less time an eyewitness has to observe an event, the less well he or she will remember it was reliable enough to present in court. • Weapon focus effect The phenomenon where witnesses look at a weapon during an event is referred to as the "weapon focus effect”. As the witness focuses on the weapon, her ability to adequately remember and later recall details such as characteristics of the perpetrator is lessened. Researchers have assessed the ability of eyewitnesses to recall various crime details in an attempt to establish the parameters of weapon focus effects on perception and memory. The first meta-analysis of the weapon focus effect research was published by Steblay in 1992. The review included 19 studies with a total sample of 2082 participants. The weapon focus effect was statistically significant and demonstrated impairment of identification accuracy. Finally a recent meta-analysis confirms the findings of the Steblay 1992 report (Fawcett et al., 2012); the presence of a weapon during an event reduces eyewitness accuracy. Research by Pickel (e.g., Pickel, 1999) and Shaw and Skolnick (1999) further indicates that any surprising object can draw attention away from the perpetrator’s face, thus reducing description and identification accuracy. In summary, although it can certainly be true that a victim pays close attention to a weapon, the research results indicate that attending to the weapon impairs memory for the characteristics of the person wielding the weapon and reduces eyewitness accuracy, especially when the opportunity to see the perpetrator is short or limited (e.g., due to concealment of the face). The topic of weapon focus was addressed in the Kassin survey. A full 87% of experts believed that the proposition that the presence of a weapon impairs an eyewitness’ ability to accurately identify the perpetrator’s face was reliable enough to present in court. • The effects of stress/fear on memory In their research, Deffenbacher, Bornstein, Penrod, and McGorty (2005) published a metaanalysis on the effects of stress on eyewitness performance. This meta-analysis reviewed 27 tests of the effects of heightened stress on identification accuracy and 36 tests of its effect on recall of crime-related details. The researchers found that high levels of stress negatively impact both types of memory. The effect of stress was notably larger for target-present than for target-absent lineups—that is, stress particularly reduced correct identification rates. Significantly, the effect was also considerably larger for eyewitness identification studies which simulated eyewitness conditions—e.g., staged crimes—than for studies involving simple face recognition activities. 6 These effects are illustrated in a study by Morgan, et al, (2004) who studied the eyewitness capabilities of more than 500 active-duty military personnel enrolled in a survival school program. After twelve hours of confinement in a mock prisoner of war camp, participants experienced a high-stress 40-minute interrogation with real physical confrontation and then, four hours later, experienced a low-stress 40-minute interrogation without physical confrontation. (The high-stress and low-stress interrogations were conducted by different individuals.) A day after release from the prisoner of war camp, after the participants recovered from food and sleep deprivation, each participant viewed a live lineup or photoarray. Regardless of the testing method, participants’ memory and identification accuracy of the high-stress interrogator was significantly less than the memory and identification accuracy of the low-stress interrogator. See the following table showing percentage of accurate identifications decisions. High Stress HITS [Target-Present] Live line-up method Photo-spread method Low Stress 27% 36% 62% 76% FALSE ALARMS [Target-absent] Live line-up method 45% Photo-spread method 48% 50% 61% It should also be noted that the length of exposure to the interrogators in this study was a relatively long period of time (40 minutes). And yet identification accuracy was still relatively low in the high stress conditions. The topic of stress was addressed in the Kassin survey. In 2001, 60% of experts believed that the proposition that very high levels of stress impair the accuracy of eyewitness testimony was reliable enough to present in court. It should be noted that the Deffenbacher eta l. (2005) metaanalysis and other important research in this area were published after the 2001 survey. It is my opinion that there is now general acceptance on the topic of stress with eyewitness experts. • Cross-race identification When a witness/victim and a perpetrator are from different racial or ethnic backgrounds, it is referred to in the scientific community and literature as a cross-race identification. Research indicates that mistaken identifications are more prevalent under cross-race conditions. Meissner & Brigham (2001) published what eyewitness researhcers consider to be the best review of research on the problem that has been called other-race or cross-race identifications or own-race bias (ORB). Meissner and Brigham analyzed data were from 39 research articles, using 91 independent samples and involving nearly 5,000 witness participants. Overall, they reported that the chance of a mistaken identification is 1.56 times greater in other-race conditions than in same-race conditions. Moreover, witnesses were 1.4 times more likely to correctly identify a previously viewed own-race face as compared with performance on other-race faces. (Overall, participants were more than 2.2 times as likely to accurately identify own-race faces as new versus old, when compared with identification performance involving other-race faces). 7 The topic of cross-race bias was addressed in the Kassin survey. A full 90% of experts believed that the proposition that eyewitnesses are more accurate when identifying members of their own race than members of other races was reliable enough to present in court. • Effects of distance on accuracy Research conducted by Professor Geoffrey Loftus and his colleagues (Loftus & Harley, 2005) examines the effects of distance on a person’s ability to view the details of a person’s face. In his “distance-as-filtering hypothesis”, Loftus explains that as a face is viewed at further and further distances, there is less ability to detect the details of the face and the facial details are coarser. The image below from Loftus’ research recreates the detail lost in viewing a face from 20 feet (left) to 100 feet (right). Research by Wagenaar and van der Schrier (1996) tested witnesses on their ability to recognize a stranger’s face from a range of distances. Participants viewed faces from distances between 3 and 40 meters and were then immediately asked to make an identification from a six-person lineup. The results showed that the proportion of correct responses to errors was too great at distances over 15 meters (about 16.4 yards) for an identification to be probative or reliable. Accordingly, the authors recommended a 15-meter distance as a useful “rule of thumb” for courts. A replication of this study using photos of famous people led to the same conclusion (De Jong et al., 2005). Similarly, in a more recent study (2008), Lindsay and his colleagues found a decrease in identification decision accuracy as distance increased. The topic of distance was not addressed in the 2001 Kassin survey. It is my opinion, however, that researchers support the proposition that distance decreases the ability of witnesses to encode details of a person’s face and consequently reduces identification accuracy. My opinion is supported by multiple sources, including the research cited above. • Effects of delay on accuracy Although the generally accepted principle amongst memory and eyewitness researchers is that memory decreases with the passage of time, eyewitness researchers have typically examined relatively short periods of delay from a few minutes to a few weeks or months. One of the earliest eyewitness studies to investigate longer periods of delay was a study conducted by Egan, Pittner, and Goldstein (1977). After a delay of 2 days, 21 days, or 56 days, participants were asked to make an identification of two targets that they had viewed during a live exposure. The researchers found no significant decrease in correct identifications of the target over the delay; however, the rate of false alarms of innocent people increased from 2 days (48% errors) to 21 days (62% errors) to 56 days 8 (93% errors). Other research conducted over the past three decades has confirmed the deleterious effects of delay on identification accuracy and in particular the misidentification of innocent suspects. This research has been summarized in a book chapter by Dysart and Lindsay (2007). The topic of delay was addressed in the Kassin survey. A full 83% of experts believed that the proposition that the rate of memory loss for an event is greatest right after the event and then levels off over time was reliable enough to present in court. • Pre-identification instruction bias Informing the witness that the police have a suspect or failing to tell a witness that the actual perpetrator may or may not be present in a lineup is suggestive because it implies that the perpetrator is in the identification task. Implying in any way to eyewitnesses that the perpetrator is in the photoarray (or that their task merely is to find the perpetrator among the set) encourages witnesses to make a selection from the array. Instead, eyewitnesses should be told explicitly that the person in question might not be in the photoarray and that they should not feel compelled to make an identification. This pre-lineup instruction follows from decades of empirical data showing that eyewitnesses are less likely to identify an innocent suspect when they are warned that the actual culprit might not be present. Further, witnesses should also be told that the person administering the photoarray does not know which person is the suspect in the case (i.e., that the photoarray is double-blind). The NIJ Eyewitness Guide (1999) outlined several methods for reducing mistaken eyewitness identification when collecting eyewitness evidence. These best practices recommend, among other things, that cautionary instructions be provided to the eyewitness that the culprit may not be in the lineup and that the police will continue to investigate the case even if no identification is made, so as to minimize natural inclination to guess or to be guided by suggestion simply because the witness believes that the police suspect must be in the lineup or photoarray. Empirical research shows that it is essential to tell eyewitnesses that the person who committed the crime may or may not be in the lineup. In the classic study of “instruction bias”, Malpass and Devine (1981) staged an act of vandalism during a lecture and 100 participants were asked to identify the vandal from one of two live lineups within the next three days. Half of the eyewitnesses in each condition were given the following biased instructions (p. 484): "We believe that the person . . . is present in the lineup. Look carefully at each of the five individuals in the lineup. Which of these is the person you saw . . ." The form on which these eyewitnesses were to indicate their decisions contained the numbers 1 through 5 (for eyewitnesses to circle their choice) but no option for rejecting the lineup. The remaining eyewitnesses were given an unbiased instruction: "The person . . . may be one of the five individuals in the lineup. It is also possible that he is not in the lineup. Look carefully at each of the five individuals in the lineup. If the person you saw . . . is not in the lineup, circle 0. If the person is present in the lineup, circle the number of his position." Among eyewitnesses who viewed a vandal-present lineup, 100% of those who received biased instructions made a positive identification, 75% of whom correctly identified the vandal. Of the eyewitnesses who received unbiased instructions, 83% made a positive identification, all of whom picked the vandal and thus were correct. Among eyewitnesses who viewed a vandal10 absent lineup, 78% of those who received biased instructions made a positive identification – all of whom were incorrect. In contrast, only 33% of those who received unbiased instructions made an incorrect positive identification from the vandal-absent lineup. Thus, significantly more false identifications – 45% – were obtained with biased instructions than with neutral instructions. The instruction bias research was reviewed by Steblay in a 1997 meta-analysis in which she cumulated the results of 22 different experimental studies of the effects of biased instructions involving nearly 2600 witness-participants. She found that biased instructions were particularly harmful in target-absent lineups in which witness accuracy declined from 60% (unbiased lineups) to 35% (biased lineups). Strikingly, the magnitude of the biasing effect was just as large when witnesses were simply not given a “don’t know” or “not present” option as it was when instructions also included some pressure to make a selection. The topic of instruction bias was addressed in the Kassin survey. A full 98% of experts believed that the proposition that police instructions can affect an eyewitness’ willingness to make an identification was reliable enough to present in court. • The effects of filler bias on eyewitness accuracy The scientific research on “filler selection and bias” shows this factor has a significant impact on the reliability of the identification outcome – particularly when simultaneous lineups are used. Researchers use the terms “functional size” (Lindsay & Wells, 1980) and “effective size” Malpass et al., 2007) to refer to the number of viable lineup members, or the number of lineup members who plausibly match the eyewitness’ description of the perpetrator. The presence of lineup fillers who resemble the perpetrator in physical appearance and the witness’s description helps protect innocent suspects from the eyewitness’ tendency to use a relative judgment comparison and merely choose the lineup member who most closely resembles their description. For example, if an eyewitness had a poor memory for the perpetrator but remembered some general characteristics, such as the perpetrator’s facial hair, then having other lineup members with similar facial hair safeguards the suspect from identification by deduction. The quality and the number of fillers in an array clearly influence the fairness of the array — as reflected in the tendency for witnesses to make lineup identifications, particularly false identifications. Research also shows that sequential lineups help “protect” innocent suspects from mistaken identification when poor fillers have been chosen for an array or when the suspect stands out for other reasons from the fillers (e.g., background color in photo is different; Lindsay, Lea, Nosworthy, et al., 1991). The reason for this is that when the lineup members are viewed one-at-a-time, the differences between them are less noticeable than when they are viewed all at once. Thus, a difference in the shade of the background will be less likely to influence a witnesses’ decision when a sequential lineup is used rather than a simultaneous lineup. It should also be noted that it is critical in an investigation for law enforcement to select lineup fillers that are known to be innocent of the crime so that witness are not being presented with multiple-suspect lineups (see NIJ Eyewitness Guide, 1999). 11 The topic of filler bias was addressed in the Kassin survey. A full 70% of experts believed that the proposition that the more members of a lineup resemble the suspect, the higher is the likelihood that identification of the suspect is accurate is reliable enough to present in court. • Non-blind lineup administration Contemporary guidelines (e.g., IACP), and laws in some states (CT, NC, TX), for conducting identification procedures state that the police officer conducting the proceedings should not know who the suspect is—this completely eliminates the possibility that the officer can influence the witness to pick the suspect. We need not assume that a lineup administrator’s influence is conscious or deliberate in order to see the value of the “double-blind” procedure. In other words, the influence by the administrator may be unintentional and it may be outside of the officer’s awareness (for example, nodding and smiling), or it may be purposeful and explicit. We know that police sometimes conduct lineups in a manner that clearly shows how their knowledge of which person is the suspect can lead them to say things that focus the eyewitness on the suspect (see Wells & Seelau, 1995). We also know that what the person administering the lineup says to the eyewitness at the time the eyewitness makes a selection has strong effects on the confidence of the witness, easily leading a “tentative identification” eyewitness to become positive in their identification, even when the identification is of an innocent person (Luus & Wells, 1994; Wells & Bradfield, 1998). The most effective method of eliminating police bias or suggestion is to have an officer who does not know the identity of the suspect conduct the identification procedure. The potential impact of administrator knowledge on witness behavior is illustrated in a study by Kovera and Greathouse (2009) in which 234 witnesses viewed a videotaped speech, that was interrupted by a man trying to take the projector and were later administered a photoarray to see if they could recognize the thief. The “administrators” were an additional 234 people who viewed a lineup training video and received further instruction on how to administer the photoarray to the witnesses. The administrators were given some background on the “case” and were told they would receive a $20 reward if the witness chose the suspect (but that they would not receive the award if they blatantly led the witness). Half of the administrators knew who the suspect was (non-blind presentation) and half did not (double-blind presentation). Unbeknownst to the administrators, half of the time the suspect was the perpetrator (target-present arrays) and half the time the suspect was not the perpetrator (target-absent arrays). In the double-blind administrator/target-absent condition 9% of the witnesses chose the innocent suspect. In the non-blind target-absent condition 21% of the witnesses chose the innocent suspect—thus, the non-blind administrators were able to subtly steer a large number of witnesses to the suspect. The non-blind administrators were most successful in steering witnesses to the suspect when the witnesses were given biased instructions (see discussion below) and photos were presented simultaneously—under these conditions 36% of witnesses chose the innocent suspect. The topic of double-blind administration was not addressed in the 2001 Kassin survey. It is my opinion, however, that researchers support the proposition that double-blind administration of lineups increases the reliability of the identification outcome. My opinion is supported by multiple sources, including the 1998 official position paper of the American Psychology-Law Society (Division 41 of the American Psychological Association) on eyewitness identification 12 procedures where double-blind administration was one of the best practice recommendations (Wells et al., 1998). • Use of simultaneous lineup rather than a sequential lineup More than twenty-five years of research indicates that sequential lineups – when compared to simultaneous lineups – can significantly reduce the rate of false identifications. In simultaneous lineups (live and photographic), the suspect and fillers are presented at the same time and the eyewitness identifies which (if any) is the suspect. A large body of scientific research demonstrates that sequential lineups, in which a witness views the suspect and fillers one at a time and makes a judgment about each face as it is presented, is generally superior to simultaneous lineups. The dominant explanation for this difference is that witnesses who view simultaneous lineups are more likely to engage in a relative judgment process and choose the lineup member who most closely resembles their memory for the perpetrator. Witnesses who view the images one-at-a-time are less able to engage in this relative or comparison process and therefore are more likely to make an identification based on their memory, rather than a combination of their memory and choosing the person who is the best answer of those presented. In 2001, Steblay, Dysart, Fulero, and Lindsay published a large-scale meta-analysis, examining 23 papers, containing 30 comparisons of sequential and simultaneous procedures, based on 4,145 research participants. These experiments were conducted in the United States, Canada, the United Kingdom, South Africa, Germany, and Australia and employed a wide array of research participants, formats (photo, video, both), types of crime (robbery, theft, other, no crime), and event stimuli (video, live, slides, transparencies). Steblay et al. reported that, across 25 studies, witnesses were nearly half as likely to make a false identification from a target-absent array (28% mistaken identifications) than from simultaneous arrays (51% mistaken identifications). This advantage holds up even in studies in which research teams have further tested the perpetrator-absent scenario by placing a particular suspect in the lineup who closely matches the description of the true target. As with the more general results, sequential lineups foster significantly fewer false identifications of the innocent suspect compared to simultaneous lineups (9% vs. 27%). Steblay, Dysart and Wells updated the 2001 meta-analysis in 2011. In the revised meta-analysis there are over 70 experiments and 10,000 participants. Interestingly, the pattern of findings is almost identical to those reported above in the 2001 article. A recent field study on eyewitness identification procedures using real witnesses and real identification decisions in ongoing criminal investigations (Wells, Steblay, & Dysart, 2011, in press) found that sequential lineups – where the witness views one suspect photograph at a time – produce fewer mistaken identifications than lineup procedures that present all of the photographs simultaneously. It should be noted that all of the lineups conducted in this study were done in a double-blind manner, where the administering officer does not know which lineup member is the suspect and which the fillers (see next Section below). The study also found that double-blind sequential lineups (compared to double-blind simultaneous lineups) as administered by police departments across the country resulted in the same number of suspect identifications (27.3% for sequential and 25.5% for simultaneous) and fewer known-innocent 13 filler identifications (12.2% for sequential and 18.1% for simultaneous). Thus, the results of the laboratory research were replicated in real criminal investigations. One of the most interesting and surprising results from the field study was that witnesses in these real criminal cases who made positive identifications (“yes, that is the person I saw commit the crime”) from a simultaneous photoarray made an identification error and chose a lineup filler 42% of the time. That is, 4 out of every 10 positive identifications that were obtained from double-blind simultaneous lineups were mistaken identifications of innocent people (see Figure below). Even with the double-blind sequential procedure, 3 of every 10 identifications were of an innocent filler. Thus, even when the best identification procedures are used, identification procedures are not entirely eliminated and witnesses can still be unreliable (most likely due to the effects of estimators on eyewitness accuracy). The topic of simultaneous/sequential lineups was addressed in the Kassin survey. A full 81% of experts believed that the proposition that witnesses are more likely to misidentify someone by making a relative judgment when presented with a simultaneous (as opposed to sequential) lineup was reliable enough to present in court. • Unconscious transference and commitment effects The phenomenon of unconscious transference has likely plagued most people at one time or another as evidenced in the question “where do I know that face?” Witnesses that view a person in multiple identification procedures or in multiple contexts (e.g., in a photoarray and then in court) are faced with a similar question. The correct answer is for the witness to say “I saw that face from several different contexts”, but the erroneous conclusion is that the face is familiar only because it is the face of the perpetrator. The concern is that this sense of familiarity on the part of the witness may lead to an increased feeling of confidence in subsequent identification procedures. In fact, a meta-analysis on transference from viewing mugshot photographs confirms that witnesses are more likely to pick from a lineup a person previously viewed (Deffenbacher, Bornstein, & Penrod, 2006). 14 If a person has been identified in an earlier procedure, he is considerably more likely to be identified in a subsequent procedure regardless of whether or not he is the actual perpetrator (Behrman & Vayder, 1994; Brigham & Cairns, 1988; Deffenbacher et al., 2006; Dysart, Lindsay, Hammond, & Dupuis, 2001; Godfrey & Clark, 2010; Gorenstein & Ellsworth, 1980; Haw et al., 2007); this is known as “commitment”. Identification of an individual from a mugshot (Brigham & Cairns, 1988; Deffenbacher et al., 2006; Dysart et al., 2001; Gorenstein & Ellsworth, 1980), as well as from a showup (Behrman & Vayder, 1994; Godfrey & Clark, 2010; Haw et al., 2007), has been found to increase the probability that participants will make a positive identification of the individual selected from a subsequent lineup. The commitment effect was also found in Behrman and Davey’s (2001) archival study, where 72% of eyewitness presented with two identification procedures identified the suspect in both procedures. The topic of unconscious transference was addressed in the Kassin survey with two questions. A full 95% of experts believed that the proposition that exposure to mug shots of a suspect increases the likelihood that the witness will later choose that suspect in a lineup was reliable enough to present in court. A full 81% of experts believed that the proposition that eyewitnesses sometimes identify as a culprit someone they have seen in another situation or context was reliable enough to present in court. • Witness confidence and accuracy Decades of research now show that there is a small to moderate relationship between the accuracy of an eyewitness’ identification decision and her confidence in that identification. When unbiased identification procedures (discussed above) are used and a witness chooses the suspect, confidence and accuracy are moderately related. This relationship, however, can be affected by post-identification factors. Researchers have studied many variables related to witness confidence but in general are concerned with the following issues: • How diagnostic of accuracy is confidence if measured at the time of the identification? How good are people at generating confidence levels that appropriately indicate their levels of accuracy? • If only highly confident witnesses testify at trials, how could witness confidence possibly assist jurors in differentiating accurate from inaccurate witnesses? If all the witnesses are highly confident, then confidence will be useless—all the witnesses will look the same— jurors will/should rely on other indicators of accuracy. Unfortunately, the problems relating to witness confidence in the accuracy of their identifications and the actual accuracy of those identifications are manifold. Some of these problems relate to jurors’ over-reliance on witness confidence as a guide to witness accuracy and some relate to the correlation between confidence and accuracy. In addition, witness confidence can be strongly influenced by suggestive procedures and post-identification factors such as repeated questioning, briefings in anticipation of cross examination, and feedback to the witness. The most useful expression of confidence is one made at the time the initial unbiased/non-suggestive identification. Research demonstrates that jurors over-believe eyewitnesses, have difficulty reliably differentiating accurate from inaccurate eyewitnesses, and are not adequately sensitive to aspects of witnessing and identification conditions that affect witness performance. A major 15 source of juror unreliability is reliance on witness confidence and this reliance appears to be largely unaffected by traditional safeguards such as cross-examination and judges’ instructions in eyewitness cases. Although there is a presumption among many actors in the legal system that there is a significant and positive confidence-accuracy relationship, serious questions about whether that is true have been raised by researchers within the law and psychology community. a) Confidence and accuracy in jury studies There is significant evidence indicating that the confidence that an eyewitness expresses in his or her identification during testimony is the most powerful single determinant of whether or not observers of that testimony will believe that the eyewitness made an accurate identification (e.g., see Cutler, Penrod & Dexter, 1990; Leippe & Romanczyk, 1987, 1989; Leippe, Manion, & Romanczyk, 1991; Lindsay, Wells, & O'Connor, 1989; Lindsay, Wells, & Rumpel, 1981; Turtle & Wells, 1988; Wells, Ferguson, & Lindsay, 1981; Wells, Lindsay, & Ferguson, 1979; Wells & Murray, 1984). Overall, it is clear that jurors do rely on witness confidence as an indicator of witness accuracy, but to what extent does witness confidence predict identification accuracy? Should jurors (and attorneys and judges) rely so heavily on witness confidence as a guide to witness accuracy? b) Confidence is highly malleable Confidence malleability refers to the tendency for an eyewitness to become more (or less) confident in his or her identification as a function of events that occur after the identification. The confidence malleability problem is particularly important because actors in the legal system can contaminate the confidence of an eyewitness in ways that can make an eyewitness’s in-court expression of confidence a meaningless indicator of the eyewitness’s memory. An eyewitness who expresses high confidence in their identification is expressing a strong belief that the identified person and the culprit are the same person. An eyewitness’s belief that the identified person is the culprit can arise out of pure memory judgments, i.e., a perception of remarkable resemblance between the identified person and the eyewitness’s memory of the culprit (Leippe, 1980; Wells, Ferguson, & Lindsay, 1981). But, significantly, an eyewitness may believe that the identified person is the culprit for reasons other than the eyewitness’s memory (Leippe, 1980; Wells, Ferguson, & Lindsay, 1981; Luus & Wells, 1994; Wells & Bradfield, 1998). For example Hastie, Landsman, & Loftus (1978), in an early demonstration of confidence malleability, found that witnesses who were questioned repeatedly grew more confident about the accuracy of details in their reports (see also Shaw, 1996; Shaw & McClure, 1996; Turtle & Yuille, 1994). Similarly, Wells, Ferguson, and Lindsay (1981) demonstrated they could increase witness confidence simply by briefing witnesses about the types of questions they might encounter in an upcoming cross-examination. When cross-examined, the briefed witnesses (who were no more accurate than the un-briefed witnesses) were significantly more confident about their identifications (than were un-briefed witnesses) and were believed more often by the jurors. Unfortunately, the briefing effect occurred among inaccurate eyewitnesses, whose levels of 16 confidence rose dramatically, whereas confidence levels among accurate witnesses were unchanged. The topics of witness confidence and confidence malleability were addressed in the Kassin survey with two questions. A full 95% of experts believed that the proposition that an eyewitness’ confidence can be influence by factors that are unrelated to identification accuracy was reliable enough to present in court. A full 87% of experts believed that the proposition that an eyewitness’ confidence is not a good predictor of his or her identification accuracy was reliable enough to present in court. • Post-identification feedback effect Even stronger and broader effects of confidence malleability have been shown to emerge when eyewitnesses are told after their identification that they identified the suspect (versus being told nothing see meta-analysis by Steblay, Wells & Douglass, 2014). In their research, Wells and Bradfield (1998) found that eyewitnesses who received confirming feedback (“Good, you identified the suspect”) were not only much more confident than the witnesses with no feedback and witnesses with disconfirming feedback - the confirming feedback witnesses also distorted their reports of their witnessing conditions by exaggerating how good their view was of the culprit, how much attention they paid to the culprit’s face while observing the event, and so on. The results of this study have been replicated many times in research labs (Steblay et al., 2014) and also with real witnesses in real ongoing criminal investigations (Wright & Skagerberg, 2007). In fact, the effects of post-identification feedback are among the strongest effects to be found in the field of eyewitness identification research. Research has also shown that the effects of post-identification feedback can be reduced significantly if a double-blind lineup administrator is used for the identification procedure (Dysart et al., 2011). One of the explanations that have been proposed to explain the post-identification feedback effect, and its strong and pervasive influence on eyewitness confidence, is the theory of cognitive dissonance (Charman, et al., 2010; Festinger, 1956; Festinger & Carlsmith, 1959). In essence, this theory, which is a long-standing and well supported theory in social psychology, states that people are in a state of discomfort when they have inconsistent or contradictory beliefs, or when they have beliefs and behaviors that are inconsistent. As it relates to eyewitness identification, a powerful example of cognitive dissonance is the DNA exoneration case of Dean Cage from Illinois. After Dean was exonerated in 2008, the victim refused to believe the accuracy of the DNA results and held on to her belief that Dean was guilty. Thus, cognitive dissonance was so powerful in that case that it was easier for the witness to believe that the DNA testing was flawed than to accept that she had made an error and identified an innocent person. Only after she was presented with independent results of the DNA testing did she come to accept that Dean was innocent and was not the man who had raped her in 1994. The topic of post-identification feedback was addressed in the Kassin survey. A full 94% of experts believed that the proposition that eyewitness testimony about an event often reflects not only what they actually saw but information they obtained later on was reliable enough to present in court. 17 III. Summary The reliability of an eyewitness identification can be evaluated by the existence or absence of factors known – empirically – to influence the strength of the witness’ memory and the reliability of the identification procedure. 18