Crime witness ID method can affect error rate: study

Reuters News
Posted: Sep 20, 2011 11:29 AM
Crime witness ID method can affect error rate: study

By Kay Henderson

DES MOINES, Iowa (Reuters) - Showing photographs of suspected criminals to witnesses in sequence, rather than all at once, can produce fewer mistakes in identifications, according to new research.

Gary Wells, a psychology professor at Iowa State University, said presenting photos one at a time produced a lower error rate than when witnesses were shown a simultaneous array of photos.

"We believe these results go a long way toward instilling greater confidence in the sequential procedure as something that improves the reliability of eyewitness identification evidence," Wells, the lead researcher, said in a conference call with journalists.

Witnesses identified a "known innocent filler" photo 18 percent of the time under the simultaneous procedure compared to 12 percent under the sequential procedure, according to the study issued by the American Judicature Society, which acts as a clearing house for information about judicial ethics and discipline.

Wells said that when witnesses were asked to look at suspects simultaneously, they tended to compare the lineup members and decide who looked most like the perpetrator.

Sequential line-up is thought to prevent the side-by-side comparison process and instead forces witnesses to use a more absolute comparison of each photo to their memory, rather than compare lineup members to each other, Wells explained.

The national study, conducted between 2008 and 2011, involved police departments in Texas, Arizona, California and North Carolina.

Wells said mistaken witness identification remained a significant problem.

"In DNA exoneration cases, for instance, 75 percent of those who were exonerated with forensic DNA tests after being convicted by juries are cases involving mistaken eyewitness identification," Wells said.

He hopes the research will encourage more law enforcement agencies to switch from simultaneous to sequential lineups. The full study is posted online at (Writing and reporting by Kay Henderson; Editing by Mary Wisniewski and Cynthia Johnston)