Eyewitness (Mis?) Identifications

The New York Times had an article over the weekend concerning eyewitness identification. What struck me about the article was the generalization of all cases based on eyewitness identification without providing any data to support the claims. I realize my argument will not be based on scientific studies either. I will only discuss my experience.

For instance, the author states that "every year 75,000 eyewitnesses identify suspects in criminal investigations. Those identifications are wrong about a third of the time, a pile of studies suggest."

One of the reasons I wanted to start this blog was to address issues like this from a prosecutor's standpoint. My caseload consists of hundreds of felonies every year. For easy math, let's say I handle 99 a year. Based on this assertion, 33 of those cases are based on mistaken identification! While there are a handful of cases a year that I am not confident in the identification, my office disposes of such cases before they are even presented to the grand jury. We are not interested in prosecuting anyone when we have a doubt about.

The key is corroboration in cases. As a prosecutor, corroboration is the lifeblood of an eyewitness ID. What does this mean? A few examples:

1) A robbery victim calls 911 and reports to the police that the robber was a white male, six feet tall, with a shaved head, wearing a red hooded sweatshirt with white letters. A half hour later a police officer stops a white man, 5'11" tall, bald, and wearing the sweatshirt described. He's only one block away from where the robbery took place, but he doesn't have the purse on him.
2) The suspect has the watch that was stolen from the victim on him when stopped by police
3) The suspect says they were in the area when the robbery happened. Maybe they even say they saw it. This coupled with an ID is good corroboration.
4) Three gas stations are robbed in three weeks. The identified suspect lives smack in the middle of all three.
5) Two separate witnesses are shown a series of photos of the suspect in separate rooms at the same time and both identify the suspect's photo as the robber.

Should the suspect be charged with the robbery if the victim identifies him and that's the only evidence? The prosecutor should look at single eyewitness identifications with a critical eye as it is abhorrent to prosecute an innocent man. If there is nothing to contradict the ID though, justice for the victim and the community demand the person be charged.

Are the police supposed to let the suspect go because their is no forensic evidence? As I tell juries, the defendant decides where the crime happens and how it happens. He decides whether to leave any forensic evidence behind. We cannot reward smart criminals and only punish the sloppy ones.

There is no DNA or fingerprints in the majority of cases. There's not even an opportunity to gather some as the defendant was wearing gloves or threw the only evidence in a river never to be recovered. Justice demands that crimes be prosecuted, not just the easy ones. I applaud and encourage the work being done by the Innocence Project and have a friend working hard to exonerate the wrongly accused. I subscribe to some of their suggestions as well like a double blind line-up (where the detective conducting the line-up doesn't know who or where the suspect is in the line-up so they can't even inadvertently suggest who it is). I don't agree that these exonerations should require more evidence in cases than the law currently requires.

In New York State, a judge determines whether a police arranged identification was "unduly suggestive" during pre-trial hearings. The law recognizes that all identification procedures are somewhat suggestive. The court is looking for those procedures where it is so suggestive that the witness picked the defendant based on the improper police conduct. Also, the identifications are then subjected to cross-examination by defense attorneys in front of juries. These are two levels of already established review and the judge instructs the jury on how to judge the credibility of a witness and the factors to consider in weighing an identification.

The debate will rage on and eyewitness experts will keep gaining in popularity. Experts like that, however, are a topic for separate blog. I'm trying to find statistics of identifications that have been confirmed through forensic testing after new testing is ordered. Does anyone know of any?

Should cases based solely on eyewitness identifications require additional evidence to proceed? Should charges be barred if the only evidence in the case is the victim saying with certainty the defendant committed the crime?