Blackstone and the balance of eyewitness identification evidence.

• Author: Clark, Steven E.

1. INTRODUCTION

   Blackstone's Commentaries on the Laws of England is comprised of four books, written in two volumes, running well over 1500 pages in length. Within this enormous work there may be no more well-known or more memorable line than that which has come to be known as the Blackstone Ratio: "the law holds that it is better that ten guilty persons escape than that one innocent suffer." (1)

   The Blackstone Ratio clearly acknowledges that there are two kinds of errors that can be made within the criminal justice system, and that one, the false conviction, is far worse than the other, the false acquittal. Two other points are also implicit within the Blackstone Ratio, specifically that there is a trade-off relationship between false convictions and false acquittals, and the criminal justice system has some control over that trade-off in terms of the kinds of errors, false acquittals or false convictions, it will allow. The implication of the Blackstone Ratio is that the criminal justice system could reduce the number of false convictions, but in the process would lose some correct convictions. Alternatively, the criminal justice system could increase the number of correct convictions, but would likely convict more innocent people as well.

   Without these two assumptions, regarding the trade-off between false convictions and false acquittals, and regarding the criminal justice system's role in determining that trade-off, the Blackstone Ratio would be meaningless. There would be little point in considering the proper relationship between false convictions and false acquittals to the extent that it was possible to reduce both errors simultaneously, or if the criminal justice system was powerless to control the relationship between the two. Of course, much has happened since Blackstone wrote those famous words. Technological advances in forensic science in just the last thirty years do make it possible to reduce both errors simultaneously, DNA evidence being the obvious example. For other forms of evidence, however, the trade-off implied by the Blackstone Ratio may operate today much as it did almost 250 years ago.

   The focus here is on eyewitness identification evidence for three reasons. First, the connection between eyewitness identification and wrongful conviction is well-established. In approximately 75% of the 261 cases in which innocent people were convicted and later exonerated through DNA evidence, the original conviction was obtained in whole or in part through mistaken eyewitness identification. (2) These and other archival analyses of false convictions have led to a consensus among legal scholars that mistaken eyewitness identification is one of the primary causes of wrongful convictions in the United States. (3) This dubious distinction begs the question, why does eyewitness identification contribute to so many wrongful convictions? An answer to this question brings us to the second reason for focusing on eyewitness evidence: it is extremely malleable, both in terms of the underlying information in memory and in the decision processes that operate on that information. This malleability not only affects the outcome of the identification procedure, in terms of whether a witness fails to identify the guilty or falsely identifies the innocent, but also influences the way eyewitness evidence flows through the criminal justice process. Third, this malleability of information and decision processes connects eyewitness identification to the two components of the Blackstone Ratio--the trade-off between false convictions and false acquittals, and the criminal justice system's control over that trade-off.

   The relationship between mistaken identification and wrongful conviction is often described retrospectively by starting with the wrongful conviction and tracing the path backwards to the cause of that wrongful conviction. The statistic linking wrongful convictions to mistaken identifications is such a retrospective analysis. Of course, evidence flows forward in the criminal justice system, from eyewitness identification to case dispositions, not in the reverse. Thus, in order to better understand the relationship between false identification and false conviction, one must look at the forward flow of identification evidence. (4)

   The forward flow of eyewitness identification evidence is considered here as a series of filters created by the decisions of witnesses, police, attorneys, judges, and jurors. Witnesses first must decide whether to report the crimes they have seen and whether to participate in criminal investigations. (5) Evidence from a number of sources suggests that as many as half of all serious crimes are not reported to police. This aspect of the filter/decision process lies somewhat outside the scope of the present paper; however one aspect of this initial filtering of eyewitness evidence is worth noting. The witnesses who choose to participate in criminal investigations are not necessarily those who are in the best position to provide reliable evidence. (6)

   Not all witnesses who come forward to participate in a criminal investigation will be presented with an identification procedure. Police must make the decisions as to whether and how to conduct an identification procedure. Those witnesses may be presented with a suspect through a one-person showup or a multiple-person lineup, and they must decide whom to identify or whether to make any identification at all. Prosecutors and defense attorneys must make decisions regarding each case based on their evaluation of the evidence and its presumed impact on the jury. For the small subset of cases that are not resolved through plea bargain agreements, the trial judge must decide whether to allow the identification evidence to be heard by the jury. Finally, for that small minority of cases that go to trial, the jury must decide whether to render a guilty or not-guilty verdict. The spirit of Blackstone's Ratio looms over each of these decisions.

   These various decision-makers and the decisions they make will be examined here within a theoretical framework loosely based on Detection Theory. At the core of this theoretical foundation is the proposition that all of these decisions--by witnesses, law enforcement, judges, and juries--may involve trade-offs between the guilty who escape and the innocent who suffer.

2. THEORETICAL FRAMEWORK

   Detection Theory was initially developed and applied to sensory processing tasks that have little in common with the decision-making tasks described here. (7) However, at its simplest level (the level that is of interest to legal decision-making) Detection Theory describes decision-making in terms of two components, the information available to the decision-maker and a simple binary decision rule. The diagnosticity of the information determines the overall accuracy of the decisions. The decision processes have no effect on overall accuracy, but determine the kinds of decisional errors that will be made.

   An example from eyewitness identification illustrates the Detection Theory. When a suspect is presented to a witness, the witness may compare that suspect to his or her memory of the perpetrator. If that match between suspect and memory is high, above some decision criterion, the witness will say "yes, that's him." However, if the match is low, below that decision criterion, the witness will say "no, that's not him." (8) The probability of an identification error of some kind depends on the diagnosticity of the match information. The pattern of errors depends on the placement of the decision criterion. To the extent that the witness's decision criterion is low, the likelihood of a false negative, saying "no" when the suspect is guilty, will be low, but the risk of a false positive, saying "yes" when the suspect is innocent, will be high. In other words, few guilty may escape, but many innocents may suffer. If the witness raises his or her decision criterion, requiring higher matches between suspect and memory in order to say "yes," the false positive rate will decrease, and the false negative rate will increase. By raising the decision criterion, the number of guilty who escape will increase, but the number of innocent who suffer will decrease.

   This theory has also been extended to the task of the juror. (9) The underlying information is the evidence of guilt. If the evidence of guilt is high, above the decision criterion, the juror will render a verdict of guilty, whereas if the evidence of guilt is lower, below that decision criterion, the juror will render a verdict of not guilty. As in the previous example, if the juror's criterion is low then the conviction rate will be high. Few guilty men will escape justice, but many innocent will be falsely convicted. If the juror raises the decision criterion, more guilty men will escape justice and fewer innocents will suffer.

   These examples illustrate the core components of Detection Theory. Variation in the quality of the underlying information leads to changes in accuracy, whereas variation in the decision criterion has no effect on the overall level of accuracy, but instead determines the particular kinds of errors that will be made. Thus, raising one's decision criterion does not make the decision-maker more accurate, but only shifts the distribution of errors, reducing false positives in exchange for false negatives. Detection Theory provides a simple framework for separating the diagnosticity of the underlying information from the decision-making process, and for understanding the trade-offs in errors, such as the trade-off between guilty men who escape and the innocent who suffer.

   Section III of this paper examines the information that underlies eyewitness identification as well as the trade-off between correct identifications of the guilty that may be lost in exchange for false identifications of the innocent that are avoided. Section IV examines the flow of...