There is much that can go wrong with an eyewitness identification of a criminal suspect. The original perception of the perpetrator may have been distorted by expectations and uncertainties. Memories may be forgotten, reconstructed, updated and distorted. Asked if the perpetrator is among a set of photos or people in a lineup, a witness may be tempted to affirm the one who best fits a remembered general appearance.
Unsurprisingly, 15 percent of people identified by eyewitnesses are innocent, studies have found. Worse, 70 percent of those later cleared by DNA evidence were convicted based on some type of misidentification.
New Jersey’s Supreme Court has been a leader in setting standards to reduce eyewitness misidentification. In a unanimous 2011 decision, it set a legal framework for admitting eyewitness identification evidence and advising juries how to evaluate it. Now it has refined those standards in response to a March case in which an eyewitness identification was the only evidence of a suspect’s guilt.
Two days after an armed robbery attempt, the victim chose the defendant from an array of photos and said he was very confident of his choice. But police made no record of the procedure, even though N.J. law requires one for identifications outside the courtroom. Nor did the judge in the case advise the jury that police hadn’t adhered to the recording requirement.
A 2014 National Academy of Sciences study of eyewitness identifications recommended that videotaping them should be standard practice and that witnesses should immediately express their level of confidence in their identifications in their own words.
Since no video, audio or written record of the suspect’s identification was made, the Supreme Court returned the case for a hearing on the admissibility of it as evidence. Previously, the court had required defendants to show that police behavior might have influenced a witness to merit such a hearing.
The court also required that juries be informed of a failure to record an identification procedure, and made it clear that the preference for such a record was video first, then audio and if neither was available, then a verbatim written record.
Associate Justice Barry T. Albin agreed, but wanted the conviction reversed and a new trial ordered. That makes sense since without a record of the identification, it will be impossible years later to determine whether the witness “betrayed even the slightest doubt before he made a ‘confident’ identification,” as Albin put it.
The police did, in this case, use the NAS-recommended double-blind method of witness testing — with neither the test administrator nor witness knowing which photo if any in the assortment was the suspect. That should be standard in all procedures as well.
The lack of a recording in the case suggests New Jersey also needs far better training for officers and agencies involved with witness identifications. There are so many ways that inaccuracy can be introduced into the procedure that every step must be handled precisely if it is to continue to have the power to convict on its own.