Credit: Veer

Jennifer Thompson was an outstanding college student in North Carolina when someone broke into her home and raped her. During the terrifying ordeal, Thompson studied the perpetrator's face, noting numerous details about his appearance. Several days later, she picked Ronald Cotton from a series of police photos, then again at a lineup, then again at trial, and Cotton was convicted.

After the conviction was overturned on appeal, at the second trial Cotton's defense team presented Thompson with another suspect, Bobby Poole. Thompson said she had never seen him before, and again pointed her finger at Cotton. Cotton was found guilty and sentenced to two life terms in prison.

Cotton was incarcerated for eleven years before DNA testing proved him innocent -- and proved Jennifer Thompson (along with twenty-four jurors from two trials) wrong. The real perpetrator? Bobby Poole.

Thompson's mistaken identification -- and the years in prison it cost Ronald Cotton -- is unfortunately not an isolated occurrence. In fact, of the 223 wrongful convictions proven thus far by postconviction DNA testing (for which innocent people have spent over 2,754 combined years in jail), more than 75 percent involved faulty eyewitness testimony. Although prosecutors often present identification evidence as irrefutable and jurors frequently find it irresistibly persuasive, DNA has proven what social scientists, as well as U.S. Supreme Court justice Thurgood Marshall, have long understood: An "unusual threat to the truth-seeking process [is] posed by the frequent untrustworthiness of eyewitness identification."

Why does mistaken identification occur?

Although exploration of the phenomenon dates back at least one hundred years to Harvard University professor Hugo Münsterberg, the work of Elizabeth L. Loftus in the 1970s sparked a flurry of scientific research seeking to answer that question. In 1978, Gary L. Wells, a leading eyewitness-identification expert, divided the factors that affect identifications into two categories: estimator variables (those related to the event itself, such as lighting, distance, the witness's level of stress, and so on) and system variables (those related to police procedures). Since then, psychologists have conducted hundreds of studies on the manner in which these two sets of variables can contaminate a witness's perception, storage, and retrieval of observed events.

The factors that can lead to false identification are myriad: Stress may impede our ability to encode events. Witnesses often have trouble identifying people of other races. The way a lineup is conducted -- one person at a time or everyone side-by-side -- can affect an identification's accuracy, as can the degree to which lineup participants resemble one another. Police can give inadvertent clues to witnesses during lineups that aren't conducted double-blind (where the lineup administrator does not know who the police suspect is) or may fail to warn witnesses that the perpetrator might not be in the lineup.

For example, let's say a witness makes an identification and is then given confirming feedback by the police. ("Good job -- you've identified our suspect.") Her confidence in the identification becomes inflated -- even if she has selected an innocent person. That in turn can skew her recollection of the events. Months later, when that witness points her finger authoritatively at the defendant and tells a jury with great confidence that he was her attacker, more often than not her certainty will carry the day.

As a staff attorney at the Innocence Project, a nonprofit legal clinic founded by Barry Scheck and Peter Neufeld in 1992 that uses DNA testing to exonerate the wrongfully convicted and reform the criminal justice system, I am part of a nationwide effort to use psychological research to improve eyewitness-identification law and policy. Working with lawyers, legislators, and law enforcement, we aim to improve police procedures and courtroom assessments of reliability in order to reduce identification errors and their devastating consequences.

Slowly but surely, the criminal justice system is recognizing the precariousness of identification evidence and seeking to implement more meaningful safeguards to reduce its role in wrongful convictions. In 1999, under then U.S. attorney general Janet Reno's leadership, the National Institute of Justice developed a handbook aimed at improving eyewitness procedures for law enforcement. (Download a PDF of the handbook.) In 2001, New Jersey's attorney general, John Farmer, established guidelines for gathering identification evidence. (Download a PDF of the guidelines.) A growing number of states have passed legislation mandating best practices, courts are increasingly embracing scientific research on the topic, and defense lawyers around the country are mounting more sophisticated challenges to identification evidence.

If this trend continues, it should enhance our ability to get it right the first time, reducing wrongful imprisonments and helping to create a fairer, more accurate criminal justice system.

Go to the Edutopia.org article "Dead Wrong: Research on Eyewitness Testimony" for a list of online resources on this subject.

Ezekiel Edwards is a staff attorney and a Mayer Brown Eyewitness Fellow at the Innocence Project.