The fallibility of eyewitness identification

Former United States Supreme Court Justice William J. Brennan Jr. wrote in a dissenting opinion nearly 35 years ago, “There is almost nothing more convincing than a live human being who takes the stand, points a finger at the defendant, and says, ‘That’s the one!’”

Convincing yes — reliable, maybe not. More than 75,000 prosecutions every year are based entirely on eyewitness identification. Some of those identifications are erroneous. Advances in the social sciences and technology have cast a new light on eyewitness identification.

Hundreds of studies on eyewitness identification have been published in professional and academic journals. One study by University of Virginia Law School professor Brandon L. Garrett, found that eyewitness misidentifications contributed to wrongful convictions in 76 percent of the cases overturned by DNA evidence.

A current member of the U.S. Supreme Court, Justice Sonia Sotomayor, has acknowledged the shortcomings of eyewitness testimony. She wrote, “eyewitness identifications’ unique confluence of features — their unreliability, susceptibility to suggestion, powerful impact on the jury, and resistance to the ordinary tests of the adversarial process — can undermine the fairness of a trial.”

What can cause an eyewitness to misidentify a suspect? There are a number of factors: Poor lighting, the crime occurred quickly, the presence of a gun, and the fact that the perpetrator is a different race than the witness. The police can, as well, intentionally or unintentionally influence an eyewitness’ identification. This week Louisiana Gov. John Edwards signed into law legislation requiring all police agencies in Louisiana to adopt eyewitness identification procedures aimed at preventing mistaken identifications and wrongful convictions.

Louisiana joins New Jersey, Massachusetts and North Carolina, among other states, to review and revise state rules for how judges and jurors treat evidence from police lineups and photo arrays. States are utilizing practices supported by years of research.

There are four basic rules proposed by researchers to help promote valid police identifications: Who conducts the lineup; instructions on viewing the lineup; the structure of the lineup or array, and immediately obtaining a confidence statement for the eyewitness.

When it comes to the “who,” research supports double-blind lineups administered by a police officer who is not familiar with the suspect and who is not one of the primary investigators on the case. The instructions are equally important. For instance, a photo array should be presented sequentially rather than as a group without comment by the officer displaying the array. Research studies have revealed that both practices decrease the pressure on witnesses to pick someone and guards against undue influence. The structure of the lineup is another area where bias can seep into the process. If the eyewitness described the suspect as a white male with long hair, approximately 6 feet, 4 inches tall with a thin build, it would not be fair to have the suspect and four short, overweight, bald men in a lineup.

Finally, a confidence statement taken from the witness immediately after the array or lineup will provide the police, the suspect and ultimately jurors with a clear understanding of just how sure — or confident — the eyewitness is in her identification of the suspect.

The Louisiana legislation, according to the New Orleans Times-Picayune, mandates all police agencies in the state adopt the Louisiana Sheriff’s Executive Management Institute model policy on eyewitness identification procedures, or write its own policy that adapts best practices. The model policy calls for, among other procedures, blind administration of photo lineups and instructions that the suspect may or may not be in the lineup.

What was once the gold standard of evidence — the disinterested eyewitness — is now subject to intense scrutiny. Standardized identification procedures are a start, however, acknowledging that witnesses can be unintentionally biased or that the human mind is fallible may be more difficult.