Could You Read This First?
I have received an increasing number of calls and e-mails regarding the sequential lineup in the last two months or so. It is the hot issue right now. The questions revolve around my view of the evidence regarding the superiority of the sequential lineup and what this means for issues of policy. [This little essay assumes that the reader knows what a sequential lineup is and how it is different from a simultaneous lineup.] This recent flurry of questions arises from several factors. New Jersey’s switch to sequential lineups and recent court battles in New York (Brooklyn and Bronx) represent one focal point. In addition, the recent meta-analysis by Steblay et al (published in Law and Human Behavior) and the manuscript of Ebbe Ebbesen (on the internet) have led to questions regarding the consistency and interpretation of the sequential-superiority effect. Prosecutors, defense attorneys, and researchers have wanted my analysis of the sequential lineup. Frankly, I’ve grown tired of answering my phone and e-mails on this question, so I prepared this statement. I hope that future calls and e-mails to me on this topic occur after reading this statement, not before.
Because I was a co-developer of the sequential lineup procedure, some have assumed that I would press for the sequential procedure above all other causes. In fact, however, I think that it is unfortunate that the sequential procedure has come to dominate so much of the discussion regarding lineup procedures. Most of my research and writing over the years has been addressed at problems with lineup procedures that are independent of the simultaneous versus sequential lineup issues. Regardless of whether one uses a simultaneous or sequential procedure, there are other important problems with lineups that have to be addressed. These other problems include: instructions to eyewitnesses, the selection of lineup fillers, how witness certainty is assessed, how to eliminate inadvertent influences from the lineup administrator, what records must be kept, and so on. Fixation on the sequential procedure is creating a certain degree of myopia with regard to seeing the broad problems with lineups.
As for the sequential lineup itself, I recommend the sequential lineup when it is properly conducted (e.g., using double-blind procedures). Overall, I believe that the scientific literature shows that the sequential lineup, although perhaps a conservative test, helps to make the identification evidence more reliable. The foundations of my conclusion are contained in the following analysis.
What is the empirical evidence?
The sequential-superiority effect is obtained when the sequential lineup procedure produces a higher ratio of accurate identifications to mistaken identifications than does the simultaneous procedure. There are many published studies showing the sequential-superiority effect (each study relying on dozens or even hundreds of research eyewitnesses). Observation of the effect is not unanimous in the eyewitness literature, however. Accordingly, it is best to turn to a meta-analysis of the data on this question. A meta-analysis is a statistical review technique that combines the results of a large number of previous studies to observe overall conclusions from a body of research. A meta-analysis of studies comparing the simultaneous and the sequential procedures was published recently. This meta-analysis shows that the sequential procedure reduces the chances of mistaken identification with a slight cost in the rate of correct identifications (Steblay et al, 2001).
Ebbesen (internet manuscript) also reviewed the scientific literature on the sequential-superiority effect with an emphasis on the question of whether the sequential/simultaneous effect is due to a "criterion shift" or a "change in discrimination." A criterion shift means that the witness becomes more cautious (less willing to make a choice, more conservative) with the sequential lineup than with the simultaneous lineup procedure. A change in discrimination means that the eyewitness is able to more clearly discern differences between an innocent person and the actual culprit, thereby making the eyewitness less likely to confuse the two. Ebbesen concludes that he cannot rule out the idea that the sequential-superiority effect is a criterion shift rather than a change in discrimination. Ebbesen feels that we need to know the answer to this question before we can make policy decisions.
My interpretation of the data that exists to date is that the difference between the simultaneous and simultaneous procedures is likely to be due to both a criterion difference and a difference in discrimination. In general, it appears that the simultaneous lineup seduces witnesses to use a low criterion (lax criterion or low threshold of making a mere relative judgment) compared to the sequential lineup. The sequential lineup probably manages to raise the criterion by the fact that the eyewitness knows that there could be another person in the sequence (not yet viewed) who is even a closer match than anyone viewed thus far, thereby forcing the witness to abandon the low criterion involved in mere relative judgments. This same process, reduced reliance on relative judgments, is also responsible for enhancing discrimination. The best evidence for the latter is the fact that the sequential lineup often enhances the ratio of correct rejections to false rejections, which is difficult to explain by a mere criterion difference. Hence, the sequential procedure helps reduce confusion between the culprit and an innocent person (increased discrimination) and also raises the bar for positive identification (requires a higher criterion).
Two other concerns have been expressed recently by parties that are resistant to changing to the sequential procedure. One is a concern that the scientific studies of the sequential procedure have been restricted to photo lineups. How do we know it works the same way with live lineups? Eyewitness researchers point out that no important differences have emerged between live versus photo lineups for other variables (e.g., effects of instructions, lineup fillers), so there is no reason to expect a live versus photo lineup difference for the sequential-superiority effect. Furthermore, our understanding of the psychological process involved (sequential reduces reliance on relative judgments) applies equally well to live and photo lineups. Nevertheless, those who are resistant to the sequential lineup will likely latch on to this as an excuse to not use the sequential procedure with live lineups. It might be a short-lived excuse, however, as there will undoubtedly be some live lineup studies soon on this issue. Meanwhile, it is no excuse at all for resisting the sequential procedure for photo lineups.
The other concern recently raised stems from a finding in our original study of the sequential versus simultaneous lineup (Lindsay & Wells, 1985). In that article we reported that the eyewitnesses who were tested with the sequential lineup were slightly less willing to testify about their identification than were those who were tested with the simultaneous lineup. This could have been an anomaly in our original findings because I am not aware of subsequent studies finding this same thing. Nevertheless, in theory it would not be surprising to find that eyewitnesses tested with the sequential procedure are somewhat less overconfident than eyewitnesses tested with the simultaneous lineup. Most eyewitness experts would consider that to be a good thing.
Policy Considerations
Mistaken identification is the largest single factor leading to convictions of innocent people, so it can be good policy to set the bar a bit higher for witnesses by adopting the sequential procedure. Using the sequential procedure, innocent suspects will be less likely to be identified, police investigators will waste less time pursuing additional evidence against the wrong person, prosecutors will be able to rely more on the identification evidence, and jurors will be less likely to encounter mistaken identifications in the mix of evidence at trial. Accordingly, I contend that changing to the sequential procedure represents good policy that helps all parties.
Again, however, let’s not become myopic because of the sequential lineup. Let’s not forget that we need to institute a broad range of improvements to lineup procedures that make the terms "biased procedure" and "motions to suppress based on suggestive procedures" mere memories from the past. I continue to be amazed as to why prosecutors are not at the forefront of this effort to jettison suggestive procedures. Do prosecutors like to have these "motion to suppress identification" hearings? Of course not. Would not prosecutors prefer to have cleaner identifications to carry forward to trial? Of course they would. Are prosecutors at the forefront of efforts to: (1) improve instructions to witnesses, (2) select better fillers for lineups, (3) implement procedures that eliminate influence from the lineup administrator, (4) assess eyewitness certainty in an unbiased manner, (5) make sure that full records are kept? The answers are (1) no, (2) no, (3) no, (4) no, (5) no. I apologize to my prosecutor friends in Clinton, Iowa, in New Jersey, and in other places who are clear exceptions to this characterization of prosecutors not assuming leadership roles on identification issues. Still, the failure of law enforcement to significantly improve their lineup procedures is primarily because prosecutors have not pressed for these improvements.
My Bottom Line
Several improvements need to be made to lineup procedures. Many of these improvements have nothing to do with the sequential procedure. Making these other improvements would be progress regardless of whether the sequential procedure is adopted or not. Although I recommend the sequential procedure and I believe that to be good policy, I will continue to help jurisdictions improve their identification procedures even if the sequential procedure is not going to be adopted for whatever reason. There is plenty to be done independently of the issue of whether or not to implement the sequential lineup procedure. Furthermore, I believe that New Jersey’s experience with the sequential lineup is working out quite well and that their success will spread use of the sequential procedure naturally to other local jurisdictions and states.
So, that’s my view as of the date of this writing (December 2001).