In response to exonerations of the wrongfully convicted, both researchers and laypersons alike tend to agree that the legal sy

Running Head: In Response to the Illinois Pilot Program

In Response to the Illinois Pilot Program on Simultaneous v. Sequential Lineups[1]

Ebbe B. Ebbesen[2] and Kristin M. Finklea

University of California, San Diego

Address Correspondence to:

Ebbe B. Ebbesen or Kristin Finklea

Department of Psychology, 0109

University of California, San Diego

9500 Gilman Drive

La Jolla, CA 92093

Abstract

The utility of replacing traditional simultaneous lineup procedures with a proposed sequential double-blind procedure was investigated in the Illinois Pilot Program across three jurisdictions. Results suggest that witnesses select more suspects and identify fewer fillers under the traditional simultaneous method of conducting lineups than in the sequential double-blind procedure. In this review, we take a closer look at the data to determine whether the observed differences in identification outcomes across lineup procedures can be accounted for by investigator bias. Suggestions for future field research are also discussed.

In Response to the Illinois Pilot Program on Simultaneous v. Sequential Lineups

The true level of eyewitness accuracy in the legal system has been debated by researchers and laypersons alike. Specifically, inaccurate eyewitness identification is thought, by some, to be the primary cause of false convictions.[3] The goal of the legal system is to maximize the number of convicted guilty suspects while minimizing (and in theory eliminating) the number of convicted innocent suspects. To minimize the rate of mistaken identifications, researchers have suggested that eyewitness ID evidence be collected using modified lineup procedures. Two of the major changes in protocol include: 1) using blind lineup administrators and 2) displaying the lineup photographs sequentially.[4] In order to effect policy change in a scientifically reasonable manner, researchers should compare a proposed new policy against the established policy already in place. Without this comparison, we will never know whether the new policy is any better than the old. Such a comparison is no different than that used in the medical field to evaluate the effectiveness of a new treatment against the current method. The Illinois Pilot Program[5] was designed utilizing this philosophy to compare the proposed sequential double-blind lineup procedure against the traditional simultaneous (non-blind) lineup procedure.

To date, findings from laboratory research suggest that witnesses make fewer selections from lineups presented sequentially than they make from lineups presented simultaneously.[6] As a consequence, both fillers and “suspects” are identified less often in sequential compared to simultaneous lineups. Based on these findings, some researchers have advocated that policy makers adopt sequential over simultaneous lineups to minimize the potential for eyewitnesses to mistakenly identify innocent suspects. Unfortunately, research has not adequately addressed whether adopting a sequential procedure in practice would also reduce the rate at which guilty suspects are identified.[7] If the results of laboratory studies can be generalized to witness responses in actual criminal cases, both filler and suspect choice rates should decrease as a result of moving from a simultaneous to a sequential lineup procedure.

Summary of Major Findings from the Pilot Program

To assess these predictions, we can look at the results from the Illinois Pilot Program[8] (Table 1). The entire sample contained a total of 367 different cases, in which researchers identified a total of 741 lineups. Of these, a total of 521 unique lineups were identified, as some investigators presented the same suspect in the same position with the same fillers to more than one witness. Across three jurisdictions investigators conducted a total of 366 standard simultaneous, single-suspect lineups and a total of 271 sequential, double-blind, single-suspect lineups. Witness/victims chose the suspect in 244 (or 67%) of all of the simultaneous lineups and in 154 (or 57%) of the sequential/blind lineups. Witness/victims chose fillers a total of 8 times (or 2.2%) when viewing the simultaneous lineups and 18 times (or 6.6%) when viewing sequential lineups (the difference in choice rates between the simultaneous and sequential lineup procedures was statistically significant). Including multiple suspect lineups in the analysis did not change the basic pattern of results. Thus, overall, the suspect choice rate was higher and the filler choice rate was lower for the simultaneous than sequential lineup procedure.

Table 1: Number and Percent of Suspect and Filler Choices for Known Single Suspect Lineups as a Function of Lineup Procedure

Lineup Procedure	Number of Choices			Percent of Total Choices
Lineup Procedure	Suspects	Fillers	No Choice	Suspect	Filler	No Choice
Simultaneous	244	8	114	66.7	2.2	31.1
Sequential	154	18	99	56.8	6.6	36.5

The tendency for witnesses to choose suspects more frequently and fillers less frequently given a simultaneous compared to a sequential lineup presentation was replicated for two of the three jurisdictions, Chicago and Evanston. In Chicago, suspects were chosen 64% of the time with simultaneous lineups and 49.5% of the time with sequential lineups. In addition, no fillers were chosen with simultaneous lineups, but with sequential lineups 6.3% of the choices were of fillers. Similarly, in Evanston, suspects were chosen 72.1% of the time with simultaneous lineups and 44.2% of the time with sequential lineups. Again, no fillers were chosen with simultaneous lineups, but 13.5% of the choices in sequential lineups were fillers. In Joliet, the pattern was slightly different. The suspect choice rates were 61.7% for simultaneous and 69.4% for sequential lineups, with filler choice rates at 4.4% and 6.5%, respectively. In summary, across all three jurisdictions, the known error rate (i.e. foil identifications) was higher with sequential than simultaneous lineups, and suspect choice rates were higher in simultaneous than sequential lineups for two out of three of the jurisdictions. It is clear that the results from this field program are in direct contradiction to the generalizations from laboratory research findings.

Alternative Explanations for the Findings

Two different classes of explanations might account for the differences in the pattern of results seen in the field project and that in laboratory simulations. We can focus on the differences in how lineups are constructed in laboratory research and how they are typically constructed in the legal system. Alternatively, we can focus on the fact that the two lineup procedures differed not only in terms of the presentation of lineup “alternatives” (simultaneously v. sequentially) but also in terms of whether the investigators conducting the lineups knew the suspect’s location in the lineup.

Lineup Construction

The discrepancy between the Illinois Pilot Program[9] results and the laboratory findings can be explained by considering that the characteristics of the lineups employed in the field might have differed from the characteristics of the lineups employed in the laboratory. First, the base rate of guilty suspects appearing in lineups might have been higher in the field compared to laboratory studies. In order to accurately assess the generalizability of laboratory research to eyewitness identifications in actual criminal cases, researchers should evaluate the true rate of target present (TP) and target absent (TA) lineups.[10] In the typical laboratory study, there are equal proportions, 50:50, of TP and TA lineups. However, researchers have yet to determine if this proportion is representative of the rate at which guilty and innocent suspects appear in actual lineups. The burden of suspect choice errors is often overlooked in applied research, assuming that suspect = culprit. The most problematic error an eyewitness can make is that of selecting an innocent suspect (not that of selecting a known-innocent foil) from a lineup. Some have concluded that in the laboratory, the largest difference in identification outcomes across the two lineup procedures occurs in the identification of “innocent” compared to “guilty” suspects.[11] That is, although both guilty and innocent suspects are chosen more often in simultaneous lineups, the difference found in innocent suspect identifications is larger.[12] Therefore, some have suggested that switching to a sequential lineup will reduce innocent suspect choices more than guilty suspect choices. However, if guilty suspects are present in real world lineups more often than are innocent suspects, using a sequential lineup procedure will suppress the hit rate more than the false alarm rate in actual cases. As such, the higher rate of suspect choices in simultaneous compared to sequential lineups conducted in Illinois may have resulted from the fact that the large majority of suspects presented to witnesses for identification were actually guilty culprits rather than innocent suspects.

Second, the relatively low rate of filler choices in actual lineups compared to laboratory studies might be explained by the difference in the way fillers are selected for actual lineups compared to lineups constructed in a laboratory. In a typical controlled experiment, the guilty suspect is removed and replaced with an innocent look-alike. Foils, consequently, remain high in similarity to both the guilty (TP) and innocent (TA) suspects. In the real world, however, an innocent suspect who is apprehended may look nothing like the actual culprit. Hence, when foils are selected for the lineup based on their degree of similarity to the innocent suspect, these foils will have a high degree of similarity to the innocent suspect, but a low degree of similarity to the actual culprit. As a result, a witness may be less likely to select a foil from such a TA lineup than a lineup containing the culprit. This could explain why the filler choice rates are so much lower in the Illinois Pilot Program[13] than in laboratory studies.

Investigator Bias

i) Conditions that Might Suppress or Enhance Investigator Bias

Based on the observed difference in choice rates between the data from the Illinois Pilot Program[14] and the predictions grounded in laboratory research, we wanted to examine whether specific variables that might make it easier or harder for investigators to influence witness choices (in the non-blind simultaneous lineup) had the predicted effects on choice rates. One such variable is the relationship that existed between the witness and the suspect prior to the crime. One might expect investigators would find it more difficult to influence witnesses who knew the suspect prior to the crime than witnesses who were attempting to identify a stranger. Therefore, not only should strangers be identified less often than acquaintances (regardless of lineup procedure), but the investigators conducting simultaneous lineups should be able to influence witnesses to pick the suspect more often and the fillers less often when the suspect and witness were strangers. This should not occur with blind, sequential lineups.

We examined the prior relationship predictions in photo lineups, as most live lineups contained stranger relationships. With simultaneous photo lineups, 90.3% of the witnesses chose the suspect when a prior relationship existed but only 53.6% chose the suspect when they were strangers. With sequential photo lineups, these percentages were 76.3 and 43.8 respectively. Thus, the difference in choice rates between sequential (blind and no influence) and simultaneous (with influence) lineups was larger for the acquaintance choices (76.3% v. 90.3%) than for the stranger choices (43.8% v. 53.6%), exactly opposite to prediction.

Considering filler choices from simultaneous photo lineups, none of the witnesses or victims chose a filler when a prior relationship existed; but, 1.3% of the witness/victims chose a filler when they were strangers. With sequential photo lineups, these percentages were 2.6 and 9.4 respectively. As expected, fillers were chosen more often when a “stranger” relationship existed in both lineup types, but were investigators who conducted simultaneous lineups better able to influence witnesses to avoid choosing fillers in stranger lineups (compared to acquaintance lineups) than were investigators who conducted sequential lineups? The shift from sequential to simultaneous lineups caused a small decrease (from 2.6% to 0%) in filler choices when the suspect was an acquaintance but a bigger decrease (from 9.4% to 1.3%) when the suspect was a stranger. These results are consistent with the investigator bias explanation.

Consequently, although the filler choice rates for strangers and acquaintances might be explained by investigator bias, the suspect choice rates are inconsistent with this hypothesis. Since the investigator influence hypothesis assumes that investigators would be simultaneously directing witnesses away from fillers and towards suspects, the pattern of results seems inconsistent with the investigator bias explanation.

We also analyzed whether the status of witnesses (as a victim of the criminal act or simply a witness to the action) had any effect of choice rates. Because the consequences of making a choice are different for the two types of witnesses, we might expect those who were victims of the crime to be more likely to make a selection purely for the sake of conviction. Investigators could take advantage of this tendency when they know who the suspect is in the lineup, an argument for a potential benefit of instituting a [sequential] double-blind procedure. If so, we may expect victims to be less likely to select foils and more likely to choose suspects, but only when presented with the simultaneous procedure – a procedure in which the investigators knew who the suspect was. Results of the analysis are inconsistent with this view. Given a simultaneous lineup procedure, victims and witnesses selected suspects at equal rates (63.49% and 67.69%, respectively) and chose fillers at equal rates (2.07%, 1.83%). The same pattern held in sequential lineups for victims and witnesses selecting suspects (51.17%, 53.96%) and fillers (8.72%, 6.47%). In essence, the effect of lineup procedure on choice rates was unchanged for victims and witnesses.

ii) Witness Confidence

The double-blind procedure was included in the Illinois Pilot Program,[15] in part, because researchers have suggested that without it, there is a possibility that investigators may consciously or inadvertently influence witness selections from a lineup.[16] Were this to happen, we might expect those witnesses who agreed with the investigator to be more confident that they were right in selecting the suspect (the same person the investigator believed was guilty) and less confident when they disagreed with the investigator and selected a filler. To examine this notion, we first analyzed suspect and filler choice rates as a function of the confidence that witnesses expressed in their identifications (Table 2). Confidence could be assessed for 31% of the simultaneous lineups and 63% of the sequential lineups based on the investigators’ written assessments of witness confidence at the time the lineup ID was conducted. High and moderate confidence choices were more associated with higher suspect choice rates than were low confidence choices. Moderate and low confidence choices were associated with higher rates of filler choices. This pattern remains constant across those lineups known to contain only a single suspect.

Table 2. Number and Percent of Suspect and Filler Choices as a Function of Expressed Confidence for all Lineups with Known Suspect Structure

Confidence	Number of Choices			Percent of Choices
Confidence	Suspect	Filler	No Choice	Suspect	Filler
High	186	7	64	72.4	2.7
Moderate	32	9	2	74.4	20.9
Low	7	4	18	24.1	13.8
Not Known	204	12	174	52.3	3.1

We recorded a total of 257 high confidence identifications (including all suspect structures and relationship categories). Of these 186 were suspect choices, 64 were no choices, and seven were filler choices. Thus, the rate of known errors for high confidence choices was 7/257 or 2.7%. When looking solely at those single-suspect lineups containing identifications of strangers, it is of some interest to note that there was only one known error (filler choice) out of a total of 65 high confidence identifications regardless of lineup type and procedure. Out of 81 high confidence stranger identifications (for both single and multiple-suspect lineups), there was also only one filler choice.

If witnesses were influenced by investigators in the simultaneous and not the sequential/blind procedure, then they should be more confident in their choices, on average, in the simultaneous than the sequential procedure. When we examined the percentage of highly confident witnesses ("That's him. I'm certain.", “100% sure.”, "100% absolutely positive.”, "I'm positive that's the one that shot me.", "Yep, that's him. I'm sure, 200%") for each lineup procedure, 78.3% and 77.1% were highly confident in their choices and 7.8% and 9.8% expressed low confidence (“I think that’s him, but I can’t be positive.”, “"He looks like the guy, but I'm not positive.", "#1 could have been the passenger.", “Only 45% sure.”) in the simultaneous and sequential lineup procedure, respectively. Thus, contrary to the investigator influence explanation, witnesses were not more confident in their choices under the simultaneous than sequential lineup procedure.

We can test the investigator influence explanation even more precisely by noting that if the administrator was leaking cues to pick the suspect (and not the fillers) during the simultaneous lineups, only those witnesses that picked the suspect would have the consensual validation of the their choices. Those who picked the fillers would actually be disagreeing with the administrator’s influence attempt. This reasoning predicts that the witnesses viewing the simultaneous lineup (and who chose the suspect) should be more confident in those choices than witnesses who chose the suspect from a sequential lineup. In contrast, those who chose the fillers from a simultaneous lineup should be less confident than those who chose fillers from a sequential lineup. We analyzed the percent of witnesses who expressed high confidence for suspect choices and witnesses who expressed high confidence for filler choices. For simultaneous lineups, 69 out of 87 (or 79.3% of the) witnesses who chose the suspect did so with high confidence. For sequential lineups, 118 out of 140 (or 84.3% of the) witnesses who chose the suspect did so with high confidence. Thus, if anything, contrary to the investigator bias explanation, witnesses were more likely to be confident in their suspect choices in sequential/blind lineups than in simultaneous lineups.

When the filler choices were examined, 66.7% of the filler choices made to simultaneous lineups and 21.5% made to sequential lineups were done so with high confidence. While the Ns are small, the trend is nonetheless opposite to the investigator influence explanation for the results. Thus, those who chose a filler from a simultaneous lineup were more confident even though their choices should have disagreed with the influence attempts of the administrator (assuming they existed).

iii) Investigator bias and cross-racial identifications

General consensus among experts is that an own-race bias exists such that witnesses are more accurate in their selections from lineups of their own race than in their selections from lineups containing individuals of a different race.[17] Consistent with this idea, results from the Illinois Pilot Program[18] (Table 3) indicate the probability that witness/victims chose a suspect increased when the suspect and witness were of the same race compared to when they were of different races. However, the filler choice rates were unaffected. Thus, when witness/victims attempted to identify suspects who were in a different racial group than their own, they were less likely to identify the suspect as the culprit and were no more likely to make a known error by identifying a filler. Another way to describe this result is that when confronted with an other-race lineup, suspects were less likely to choose someone as the perpetrator.

Table 3. Number and Percent of Suspect and Filler Choices as a Function of Racial Similarity of Witness/Victim and Culprit for All Lineups Containing Race Information

Racial Similarity	Number of Choices			Percent of Choices
Racial Similarity	Suspect	Filler	No Choice	Suspect	Filler
Other Race	126	13	127	47.37	4.86
Same Race	303	18	138	66.01	3.92
Total	429	31	265

Examining the cross-racial data yields findings inconsistent with the notion of stronger investigator bias in simultaneous lineups. Investigator bias should be predictably stronger when memory for the culprit is weak. Research on the own-race bias would suggest that witnesses have weaker memory for culprits of another race than for culprits of their shared race. It would be hypothesized then, that effects of investigator bias would be more sizeable between simultaneous and sequential lineups given cross-racial identifications than would be seen in same-race simultaneous and sequential identifications. Suspect choice rates in same- and other-race simultaneous lineups were 188/267 (70.41%) and 74/140 (52.86%) respectively, while filler choice rates were 3/267 (1.12%) and 4/140 (2.86%). These rates can be compared to suspect choices in same- and other-race sequential lineups at 115/192 (59.9%) and 51/125 (40.8%) respectively, while filler choice rates were 15/192 (7.81%) and 9/125 (7.2%). Data is included from both single and multiple-suspect lineups.

As can be seen, the increase in suspect choice rates from sequential to simultaneous lineups was not larger when witnesses had presumed weaker memories for other-race suspects (40.8% v. 52.9%) than stronger memories for same-race suspects (59.9% v. 70.4%). In addition, the decrease in filler choices was not larger for other-race fillers (7.2% v. 2.9%) than same-race fillers (7.9% v. 1.1%); In short, these results are inconsistent with the investigator influence explanation.

Addressing critics of the Illinois Pilot Program

Some critics may argue that the results of the Illinois study are compromised by a confound between lineup procedure and blind testing. That is, the sequential lineups were conducted by investigators blind to the identity of the suspect in the lineup, whereas the investigators administering simultaneous lineups were not blind to the suspect’s identity. These conditions, however, were proposed as a package deal, meaning that the double-blind sequential lineup represents the policy as it would and should be instantiated over the traditional lineup. Thus, in evaluating the proposed policy change, we compared the components of the proposed change with the standing policy. Had the effects of lineup procedure and blind testing been evaluated independently, researchers would not be able to draw conclusions about the suggested policy change as a whole. The Illinois Pilot Program[19] was not designed to test the varying options of simultaneous/blind, simultaneous/non-blind, sequential/blind, and sequential/non-blind. These four procedures would need to be evaluated separately for the most comprehensive assessment. The primary conclusion researchers can make is that the sequential double-blind procedure, as tested in Illinois, is not superior to traditional simultaneous lineups.

Previous field studies that promote the success of sequential double-blind lineups, such as that in Hennepin County, MO,[20] have not included traditional simultaneous control groups to fully examine the proposed policy change. Regardless, results of the Hennepin County program closely mirror those of the Illinois Pilot Program[21] with a 54% suspect identification rate and 8% filler choice rate for all known single-suspect, double-blind sequential lineups. Had a control group been included, is it not possible that they would have also seen a higher rate of suspect identifications and lower rate of filler choices in traditional single-suspect, simultaneous lineups? The error on which researchers have clearly focused is that of reducing false identifications, while omitting the consequences of a witness failing to select the suspect.

With respect to any policy, the details must be clearly articulated. Simply indicating that a sequential lineup procedure is to be instituted is insufficient. There are several procedural variations of a sequential lineup, which if altered and implemented in different combinations, may dramatically affect any resulting eyewitness decision. One such variation is the inclusion of a stop rule, in which witnesses are told they will not be allowed to view the remaining alternatives in the lineup once they have made a positive identification. If no stop rule is included (following the sequential lineup procedure of the present study), an interesting psychological phenomenon arises. Suppose a witness positively identifies a foil in the lineup before the suspect is shown. What do witnesses do when they then see the suspect? Do they raise their decision criterion at that point? Do they dismiss the suspect’s photo because now that they believe they have completed the task of making a selection, are no longer paying careful attention to the lineup? In other words, we need to determine where witnesses set their decision criterion for making a selection from the lineup and whether this decision criterion is set differentially for simultaneous and sequential lineup procedures.[22]

These issues beg the question regarding the decision strategy witnesses use in their selections. One key variable in this decision process is the witness’ strength of memory for the culprit. Also, a witness’s ability to discriminate between the actual culprit and innocent foils may depend on: how similar innocent foils appear to the culprit, how similar the culprit’s looks in the lineup are to his looks at the time of the crime, and how similar an innocent suspect’s looks in the lineup are to the culprit’s looks at the time of the crime. It is also important to know what witnesses use as their standard for determining this degree of match. When witnesses are presented with items is sequence, rather than all at once, we raise the distinct possibility that witnesses will use different decision criterion for different items as they progress through the sequence.

One of the key problems with the nature of a sequential protocol containing a stop rule is that it prevents the witness from being able to choose the best lineup member when there is more than one that is above the witness’s decision criterion for a good match. Thus, in low similarity lineups where the witness’ ability to discriminate (between the suspect and foils) is high, and the witness’ criterion for choosing is high, one would see a small difference in selection choices using different decision strategies.[23] Conversely, in high similarity lineups, we might expect that one of the foils presented before the suspect might be a “good enough” match for the witness to pick him. However, were this foil and the suspect presented side by side, the witness might choose the suspect because the suspect is an even better match to the witness’ memory than is the foil. One consequence of this is that more foils will be chosen when the suspect is placed later in the lineup. All foils that meet or exceed the witness’ criterion for making a positive ID will be chosen before the witness even gets to see the suspect.

From the Illinois Pilot Program[24] data, we were able to examine differences in the rate of sequential lineup fillers selected before and after the suspect appeared in the lineup. Of the 21 sequential lineup foil choices (compared to a total of 6 simultaneous lineup foils chosen), 13 were selected by witnesses before the witness had the opportunity to view the suspect in the lineup, while 8 were chosen after the suspect was viewed. Eliminating the pre-suspect foil choices in sequential lineups, the difference in filler choice rates between the two lineup procedures is virtually eliminated. Interestingly, these results suggest that the higher rate of foil choices seen in sequential lineups may be driven by pre-suspect fillers that meet or exceed a witness’s decision criterion for making a positive identification.

Future Research

Evident from the current discussion, more laboratory and field research is needed to examine the efficacy of varying methods of conducting eyewitness identification lineups. Sequential double-blind lineups do not appear to yield fewer known errors than traditional simultaneous lineups. No single study can yield definitive results or subsequent recommendations that are to be widely applied.[25] These findings should encourage more law enforcement agencies to conduct further research to scientifically investigate the costs and benefits that would be associated with instituting a given policy change. As seen from the different pattern of findings in Chicago and Evanston compared to those in Joliet, any policy changes must be evaluated against jurisdictional differences in the outcome and perhaps adjusted accordingly.

Furthermore, procedural variants of sequential lineups should be clearly defined and examined before any policy recommendations should be enacted, and the same holds for details of the double-blind procedure. In addition to the aforementioned option of including a stop rule, sequential lineups can differ in protocol based on several factors: what witnesses believe about the size of the lineup, what happens to lineup items after they are viewed, how many passes through the lineup a witness is given, and where in the lineup the suspect is positioned. The specifics of the sequential procedure may play an important role in eyewitness accuracy. For example, providing witnesses with information on the number of alternatives in the lineup could create pressure for the eyewitness to select someone before the end of the lineup is reached, or it may build expectations toward the end of the lineup that the culprit is not present. Conversely, if witnesses are not told how many faces are in the lineup, they may withhold making a selection, believing there are always more alternatives to be seen. If each alternative remains in view after it is presented, witnesses may utilize these as comparisons to the current item upon which they are deciding. Another possibility is to remove each item from view after a yes/no decision has been made, forcing an absolute decision for each item. Also, allowing multiple passes through a lineup may lead witnesses to withhold a selection until all lineup members have been viewed and compared to one another.

Yes, the specifics of a particular protocol can easily be defined as to which procedure will be utilized. However, research has not yet been conducted to determine how all procedural variations of sequential lineups interact to produce different results, nor has it been determined how foil choice rates and suspect (guilty suspect vs. innocent suspect) choice rates are affected by these combinations. Further research is needed regarding the double-blind procedure and its effects on accuracy as well. Options to be evaluated include witnesses making selections in private, in front of a blind administrator, or on a computer screen. The main argument for a double-blind procedure stems from the presumed possibility of investigators inadvertently sending signals to witnesses, thus influencing their selections.[26] Perhaps, though, the mere presence of an investigator (blind or not) may differentially affect witness selections. For this reason, the recommended blinding procedure should be evaluated with witnesses making their final selections without any investigator contact. Alternatively, simply telling the witnesses that they should not assume that the investigator knows who the suspect is might be more than sufficient to produce whatever benefits researchers believe might be achieved with the more complex double-blind procedure.

When research is proposed to compare polices, all recommendations for change should include methods and measures that will allow one to monitor which procedure is more successful; before commencing research, the measures of success should be defined and agreed upon before any policy change is instituted. In the present domain, work that uses DNA, for example, to establish the accuracy of suspect choices would avoid criticisms about the use of filler choices as the only measure of error. Future applied research will assuredly guide academics, experts, and policy-makers in an improved direction for recommending change.[27]

References

R. Brandon & C. Davies, Wrongful Imprisonment (1973).

E. Conners, et al., National Institute of Justice, Convicted by Juries, Exonerated by Science: Case Studies in the Use of DNA Evidence to Establish Innocence After Trial (1996).

D. Corey, et al., Parallelism in Eyewitness and Mock Witness Identification, 13 Applied Cognitive Psychol. Special Issue: Measuring Lineup Fairness, S41, S58 (1999).

Brian L. Cutler & Steven D. Penrod, Forensically-Relevant Moderators of the Relationship Between Eyewitness Identification Accuracy and Confidence, 74 J. Applied Psychol., 650, 652 (1989).

Kenneth A. Deffenbacher, Eyewitness Accuracy and Confidence: Can We Infer Anything About Their Relationship?, 4, Law. & Hum. Behav., 243, 260 (1980).

Ebbe B. Ebbesen & Heather D. Flowe, Simultaneous Versus Sequential Lineups: What do We Know? (2003) (unpublished manuscript, on file with the University of California, San Diego, <://www-psy.ucsd.edu/~eebbesen/SimSeq.htm>).

Ebbe B. Ebbesen & Vladimir J. Konecni, Eyewitness Memory Research: Probative v. Prejudicial Value, 5, Int’l Dig. Hum. Behav. Sci. & L., 2, 28 (1996).

Innocence Project (last modified May 8, 2006) <://www.innocenceproject.org/>.

Saul M. Kassin, et al., On the ‘General Acceptance’ of Eyewitness Testimony Research: A New Survey of the Experts, 56, Am. Psychologist, 405, 416 (2001).

Amy Klobuchar et al., Improving Eyewitness Identifications: Hennepin County’s Blind Sequential Lineup Project, Cardozo Pub. L. Pol’y & Ethics J. (forthcoming, 2006).

Rod C. Lindsay & Gary L. Wells, Improving Eyewitness Identifications From Lineups: Simultaneous Versus Sequential Lineup Presentation, 70, J. Appl. Psychol., 556, 564 (1985).

Office of Justice Programs, U.S. Dep’t of Justice, Eyewitness evidence: A Guide for Law Enforcement (1999).

Report to the Legislature of the State of Illinois: The Illinois Pilot Program on Sequential Double-Blind Identification Procedures by Sheri H. Mecklenburg, Program Director (Mar. 17, 2006) (on file with the author).

Siegfried L. Sporer, Eyewitness Identification Accuracy, Confidence, and Decision Times in Simultaneous and Sequential Lineups, 78, J. Appl. Psychol., 22, 33 (1993).

Nancy Steblay et al., Eyewitness Accuracy Rates in Sequential and Simultaneous Lineup Presentations: A Meta-Analytic Comparison, 25, Law & Hum. Behav., 459, 473 (2001).

Gary L. Wells, What Do We Know About Eyewitness Identification?, 70 Am. Psychologist, 619, 626 (1993).

Gary L. Wells & Rod C. Lindsay, On Estimating the Diagnosticity of Eyewitness Nonidentifications, 88, Psychol. Bull., 776, 784 (1980).

Gary L. Wells et al., From the Lab to the Police Station: A Successful Application of Eyewitness Research, 55, Am. Psychologist, 581, 598 (2000).

Gary L. Wells et al., Eyewitness Identification Procedures: Recommendations for Lineups and Photospreads, 22, Law & Hum. Beh., 603, 647 (1998).

[1] See Report to the Legislature of the State of Illinois: The Illinois Pilot Program on Sequential Double-Blind Identification Procedures by Sheri H. Mecklenburg, Program Director (Mar. 17, 2006) (on file with the author).

[2] To whom correspondence should be addressed at 9500 Gilman Drive, Department of Psychology, University of California, San Diego, La Jolla, CA 92093; e-mail: eebbesen@ucsd.edu, kfinklea@ucsd.edu.

[3] See, e.g. R. Brandon & C. Davies, Wrongful Imprisonment (1973); Gary L. Wells, What Do We Know About Eyewitness Identification?, 70 Am. Psychologist, 619, 626 (1993); E. Conners, et al., National Institute of Justice, Convicted by Juries, Exonerated by Science: Case Studies in the Use of DNA Evidence to Establish Innocence After Trial (1996); Innocence Project (last modified May 8, 2006) <://www.innocenceproject.org/>.

[4] Office of Justice Programs, U.S. Dep’t of Justice, Eyewitness evidence: A Guide for Law Enforcement (1999); Gary L. Wells et al., Eyewitness Identification Procedures: Recommendations for Lineups and Photospreads, 22, Law & Hum. Beh., 603, 647 (1998).

[6] See Nancy Steblay et al., Eyewitness Accuracy Rates in Sequential and Simultaneous Lineup Presentations: A Meta-Analytic Comparison, 25, Law & Hum. Behav., 459, 473 (2001).

[7] See, e.g. Rod C. Lindsay & Gary L. Wells, Improving Eyewitness Identifications From Lineups: Simultaneous Versus Sequential Lineup Presentation, 70, J. Appl. Psychol., 556, 564 (1985); Siegfried L. Sporer, Eyewitness Identification Accuracy, Confidence, and Decision Times in Simultaneous and Sequential Lineups, 78, J. Appl. Psychol., 22, 33 (1993); Steblay et al., supra note 6.

[10] See, e.g. D. Corey, et al., Parallelism in Eyewitness and Mock Witness Identification, 13 Applied Cognitive Psychol. Special Issue: Measuring Lineup Fairness, S41, S58 (1999); Gary L. Wells & Rod C. Lindsay, On Estimating the Diagnosticity of Eyewitness Nonidentifications, 88, Psychol. Bull., 776, 784 (1980).

[11] Steblay et al., supra note 6. But see Ebbe B. Ebbesen & Heather D. Flowe, Simultaneous Versus Sequential Lineups: What do We Know? (2003) (unpublished manuscript, on file with the University of California, San Diego, <://www-psy.ucsd.edu/~eebbesen/SimSeq.htm>) for a different conclusion after reanalysis of the same data.

[12] Ebbesen & Flowe, id, suggest that the relative size of the difference in suspect error rates and filler suspect rates will depend on the criteria witnesses use to determine how much their memories of the suspect must match the presented alterative before they claim that the alternative is the culprit. Somewhat counter to intuition, the higher their criteria, the bigger the difference between simultaneous and sequential lineups in guilty suspect choices and the smaller the difference in innocent suspect choices.

[16] See, e.g., Wells et al., supra note 4; Gary L. Wells et al., From the Lab to the Police Station: A Successful Application of Eyewitness Research, 55, Am. Psychologist, 581, 598 (2000).

[20] See Amy Klobuchar et al., Improving Eyewitness Identifications: Hennepin County’s Blind Sequential Lineup Project, Cardozo Pub. L. Pol’y & Ethics J. (forthcoming, 2006).

[25] See Ebbe B. Ebbesen & Vladimir J. Konecni, Eyewitness Memory Research: Probative v. Prejudicial Value, 5, Int’l Dig. Hum. Behav. Sci. & L., 2, 28 (1996).

[26] See, e.g., Wells et al., supra note 4; Wells et al., supra note 15.

[27] Acknowledgements: The authors would like to thank Dr. Heather Flowe for her immense contribution to the project as well as undergraduate research assistants, Joseph Hicken and Jeffrey Osborne.