ditions, those mitigating against the likelihood of highly reliable testimony, . indeed find evidence of a direct, positive relationship between accuracy and exp when confidence ceases to track memory accuracy, an eyewitness might then. Under those conditions, mock-crime studies and police department field studies . eyewitness confidence informs accuracy under more pristine testing conditions spread impression that eyewitness memory is unreliable. the relation between accuracy and confidence in memory retrieval matters for . Most cases of eyewitness identification come from people who are highly.
For example, nowadays, a typical photo-lineup identification procedure consists of the simultaneous or sequential presentation of one photo of the suspect the person the police believe may have com- mitted the crime and five or more fillers photos of peo- ple who are known to be innocent but who physically resemble the suspect.
Such a lineup offers protection to an innocent suspect because a witness who chooses ran- domly is far more likely to land on a filler than the sus- pect. However, before the dangers of eyewitness misidentification were understood, an investigating offi- cer might present a lineup consisting of only suspects with no fillers and tell a witness who had just identified one of the suspects with low confidence that it was clearly the right decision, resulting in a higher expression of confidence the next time the witness was asked about it.
By the time of the trial, the jury would see the witness honestly misidentify the suspect with high confidence and convict on that basis alone, often sending an inno- cent person to prison.
Practices like these help to explain why, in every one of the DNA exoneration cases involv- ing eyewitness misidentification examined by Garrettwitnesses who mistakenly identified innocent defendants did so with high confidence when the case was tried in a court of law.
(PDF) The Relationship Between Eyewitness Confidence and Identification Accuracy: A New Synthesis
But what about the confidence expressed by an eye- witness tested using the scientifically validated proce- dures that have been developed over the years by eyewitness-identification researchers? That is the ques- tion we focus on here, and the answer will undoubtedly come as a surprise to many. Understandably, the dispro- portionate role played by eyewitness misidentification in the DNA exoneration cases has helped to create a wide- spread impression that eyewitness memory is unreliable even under the best of circumstances i.
How- ever, over that same period of time, the legal system has increasingly come to interpret the scientific literature as indicating no meaningful relation between confidence and accuracy. As a result, some courts now advise juries to disregard eyewitness expressions of confidence and to focus instead on a variety of other factors when trying to assess the reliability of an ID. The purpose of our article is to explain why a blanket disregard for eyewitness con- fidence not only is at odds with what has been learned in recent years but also can contribute both to the wrongful conviction of innocent suspects and to the unwarranted removal from suspicion of a guilty suspect.
Our article is organized as follows: We first document a growing trend within the legal system to disregard eye- witness confidence, with no distinction drawn as to whether the eyewitness-identification procedures were appropriate or not and with no distinction drawn between witness confidence at the time of the initial identification versus witness confidence at a later time.
The general idea is that a strong relation between confidence and accuracy is the natural state of affairs, but there are various things that can contaminate that relation. We then consider the nontrivial issue of how best to measure the confidence- accuracy relationship, followed by a detailed review and reanalysis of the empirical literature on the confidence- accuracy relation.
The results will show that when pris- tine identification procedures are used, eyewitness confidence is a highly informative indicator of accuracy, and high-confidence suspect identifications are highly accurate. We go on to demonstrate that the confidence- accuracy relationship can be compromised when certain non-pristine identification procedures are used, and we enumerate priorities for future research on the confi- dence-accuracy relationship.
Understood in the Legal System In the legal system, eyewitness confidence is increasingly distrusted. When discussing confidence, no distinction was drawn between identification procedures that are pristine and 12 Wixted, Wells those that are not.
These jury instructions are, of course, accurate when applied to problematic eyewitness-identi- fication procedures, but our question concerns the confi- dence-accuracy relationship when pristine procedures are used early in the investigation and prior to any mem- ory contamination. As our review will demonstrate, there are known conditions under which confidence clearly informs accuracy and other known conditions under which it clearly does not.
A bleak view of eyewitness confidence is not in any way limited to New Jersey. Other jurisdictions have revised their jury instructions so as to encourage juries to place little faith in eyewitness confidence. These instructions appro- priately focus on the importance on the initial identifica- tion, but they do not appropriately communicate the high information value of an initial statement of confidence obtained from a pristine identification procedure. Next, consider this recent statement made by the Con- necticut Supreme Court in State v.
No distinction was made between the confidence of the witness at the time of identification and the confidence of the witness at trial. Similarly, in Brodes v. Again, no distinction was made by the court between the confidence of the witness at the time of identification and the confidence of the witness at trial.
Along the same lines, in State v. Mitchellthe Utah Supreme Court recently stated, In the end, we agree with the Connecticut Supreme Court that the available studies are not definitive on the question whether there is a significant correlation between certainty and accuracy. But we are also mindful that the literature suggests certainty may not always be as reliable an indicator of accuracy.
Therefore, we hold it is error to instruct the jury on the degree of certainty factor, and we discourage its future use. Undeniably, eyewitness certainty at pretrial hearings or at trial should be highly suspect for reasons we will discuss. But when a lineup is conducted under pristine testing conditions and the confidence statement of the witness is taken at the time of identification, the data indicate that confidence is a reliable indicator of accuracy.
The fact that courts increasingly distrust eyewitness confidence is not altogether surprising, given that expert witnesses and concerned organizations routinely paint a gloomy picture of the confidence-accuracy relationship.
It is not just the Innocence Project that has a generally pessimistic view of the confidence-accuracy relationship.
A recent amicus brief filed by the American Psychological Association painted a similarly bleak picture of the situation: Another slightly older analysis. Importantly, error rates can be high even among the most confident witnesses. In one article reporting results from an empirical study, researchers found that among witnesses who made positive identifications, as many as 40 percent were mistaken, yet they declared themselves to be 90 percent to percent confident in the accuracy of their identifications.
This confirms that many witnesses are overconfident in their identification decisions. American Psychological Association, Claims like this do not accurately inform the legal sys- tem. If these claims of an untrustworthy confidence-accu- racy relation had been restricted to specific non-pristine testing conditions that have been shown to compromise the information value of eyewitness confidence or to confidence statements taken later rather than at the time of the initial identification, then they would be defensible claims.
One of the key points we will emphasize is that the only time that confidence is known to be a highly reli- able predictor of accuracy is when memory is first tested, before there is much opportunity for memory contamination to occur.
An expression of low confi- dence on that first test is a glaring red flag because it is almost always an indication that the risk of error is high. If the witness is assumed to be honest, and if the ID was made with low confidence, then it is an unreliable ID. In fact, most of the DNA exonerees who were misidentified by an eye- witness were, at the outset of the investigation, identi- fied with low confidence Garrett, It was only later, in court and in front of the jury, that the initial low-confidence ID somehow morphed into a high- confidence ID.
If it had been understood that confi- dence is indicative of accuracy only on an initial mem- ory test i.
Or, if prosecutors had understood that low confidence at the initial identification is indicative of a high risk of error, then the innocent suspects in these cases might not have been indicted in the first place.
Thus, far from being a problem, initial eyewitness confidence is part of the solution to eyewitness-based wrongful convictions Box 1.
To appreciate how important it is to take into account not ignore an initial expression of low confidence by an eyewitness, imagine an eyewitness-identification case involving an innocent suspect that is adjudicated using an approach in which eyewitness confidence is ignored but various factors known to affect eyewitness memory are taken into consideration by a jury.
Many of these fac- tors are estimator variables—that is, variables that affect memory but are outside of the control of the legal sys- tem Wells, Some common estimator variables include: Cotton was convicted largely on the basis of her testimony, but he was later exonerated by DNA evi- dence after spending more than 10 years in prison.
However, after confir- matory feedback from the police, Thompson became increasingly confident that Cotton was the rapist. From this perspective, the mistake was to rely on confidence expressed at the time of the trial after it had become improperly inflated instead of relying on confidence expressed at the time of the initial ID before memory contamination had a chance to play a significant role. Indeed, in a very real way, it was the legal system—not Jennifer Thompson—that made the key mistake by ignor- ing her initial low confidence.
From this perspective, the time has come to exonerate her, too. Race cross-race IDs are less accurate than same- race IDs 2. So what are pristine eyewitness-identification procedures? Wixted and Wells note that a lineup should contain only one suspect. The inclusion of fillers makes it more likely that a random choice will be a wrong one, helping to prevent mistaken identifications.
Additionally, the suspect should not stand out in the lineup: Both the fillers and the suspect should match the description of the perpetrator. Researchers have found high rates of mistaken identifications — and increased eyewitness confidence in such identifications — when innocent suspects who match the description of the perpetrator are included in lineups alongside fillers who do not.
The Relationship Between Eyewitness Confidence and Identification Accuracy: A New Synthesis
Because eyewitnesses often approach lineups with the goal of finding the offender, eyewitnesses should be cautioned that the offender may not be in the lineup. Loftus and Rachel L. Finally, a confidence statement should be collected at the time of the identification. Confidence in an identification is a known reliable indicator of accuracy only at the time of the initial identification. Confidence statements at trial and other retrospective accounts of confidence are not indicative of accuracy because, after the initial identification, confidence can be altered by a variety of factors.
Given that a suspect has been identified with a certain level of confidence, what is the probability that the identification is correct?