An ISU professor has been awarded a $330,000 grant that may keep innocent people out of prison.

The National Science Foundation awarded Gary Wells, professor of psychology, with the grant last month to support his study on the effect of facial composites, which are used when investigators do not have a suspect. These pictures can be drawn by a sketch artist or a computer system.

Wells said his interest in this area began when he was a graduate student at Ohio State University where he heard a lawyer was searching for someone to help prove that his client was mistakenly identified from a lineup.

The year was 1975, and controlled studies had yet to be done on eyewitness identification.

“Having no science in place at the time, I was unable to help. The client was convicted,” Wells said.

This case sparked Wells’ passion for studying eyewitness identification, he said. Later that year he conducted an experiment using a staged crime. Wells was surprised to find how likely people were to make mistaken identifications.

After a series of experiments, he said he began to realize that certain precautions could reduce the chances of mistaken identification in lineups.

DNA analysis recently freed the 100th innocent death row inmate, many of whom were convicted because of mistaken eyewitness identification, according to Wells’ Web site at www.psychology.iastate.edu/faculty/gwells/DNAcasesAPstudy.htm. A 1998 study on 40 DNA-exoneration cases by Wells and five colleagues found 90 percent of the cases involved one or more mistaken eyewitness identifications, according to his Web site. Wells’ work has decreased the chances of mistaken identification.

The results of his research have been adopted by various agencies, including the U.S. Department of Justice, according to its Web site at www.usdoj.gov. Wells was part of 34-member group that developed the department’s first national guide on eyewitness identification for law enforcement in 1999.

Other agencies in the state of New Jersey, Iowa and New York have also used Wells’ findings in developing policies, Wells said. The Illinois Governor’s Commission asked to adopt Wells’ procedures before the consideration of reinstating the death penalty.

Wells said he has conducted experiments in this field for more than 25 years and has worked on 50 murder cases and about 100 other felony cases.

His continued interest led him to submit a proposal to the National Science Foundation to expand upon his work with eyewitness identification, he said.

Wells said he believes composites tend to change the memories of eyewitnesses.

“Because composites are often a very poor likeness of the culprit, we believe that witnesses who develop composites will perform more poorly on later lineups than will control witnesses who do not develop composites,” he said.

Ames Police Sgt. Mike Johns was not able to offer a prediction.

“Typically when we do an eyewitness identification, we already have a suspect,” he said. “It isn’t often that we do a computer composite, and I don’t have any statistics on their effectiveness.”

ISU Police Chief Jerry Stewart believes the composite picture system can be helpful in certain cases.

“In crimes that involved armed robberies at local banks and grocery stores, the composites produced were nearly identical to the people who were ultimately convicted,” he said.

As with photographs, negative results can occur if these composites are inaccurate, or implicate innocent people, Stewart said.

“Additional research needs to be conducted in a variety of law enforcement-related issues; I applaud Dr. Wells’ work in this important area,” he said.

The project is expected to take about three years to complete.

In these experiments, Wells and his assistants will ask participants to view a simulated crime video. Afterward, some will be asked to create a facial composite of the culprit. Both groups will be asked to identify the culprit from a lineup later.

Although Wells expects a few obstacles, he is optimistic the team will find ways to further increase the reliability of eyewitness identification evidence.

“Even if we don’t fully solve the central problem, we will discover other important things that we have not yet even thought about,” he said.