The law was intended to reduce wrongful convictions after several were tied to errors in eyewitness testimony.
“The real scandal is how few departments have developed best practices after policymakers have bent over backwards to let police do this themselves. It just hasn’t worked,” said Garrett, whose 2011 book, “Convicting the Innocent: Where Criminal Prosecutions Go Wrong,” tracked the causes of the first 250 DNA exonerations.
Of 144 Virginia police departments and sheriff’s offices that responded to a survey earlier this year, only 6 percent had fully applied a policy approved in 2011 by the Virginia Department of Criminal Justice Services.
More than one-fourth of the agencies relied on an outdated policy from two decades ago. Forty-one other agencies had no written policy at all, apparently violating the 2005 state law, according to a paper by Garrett to be published in the Virginia Journal of Criminal Law.
Dana G. Schrad, executive director of the Virginia Association of Chiefs of Police, said her members have been working on adopting changes, most recently at a June symposium, and will be surveyed again next month.
“Clearly we want them to adopt best practices” and apply them through training and supervision, Schrad said. “I’m not going to make excuses for them. They need to do this,” Schrad said. “Our goal is to get the vast majority in compliance by the end of this calendar year, before the next General Assembly session.”
At issue is a problem found in 13 of 16 wrongful convictions discovered through DNA testing in Virginia since 1989 — eyewitness misidentifications. Nationwide, such mistakes have played a role in more than 70 percent of DNA exonerations, often following police procedures that intentionally or unintentionally bias witness recollections.
Last year, Virginia Gov. Robert F. McDonnell (R) approved more than $1 million in compensation to Thomas Haynesworth, who spent 27 years in prison for rapes he did not commit, based largely on a misidentification by a rape victim in 1984. DNA test results years later cleared him and pointed to a convicted rapist.
Social scientists say that despite their persuasive power for juries, witness recollections are wrong about one-third of the time.
Potentially suggestive practices, such as repeated showings of a suspect, can be reduced by simple measures such as conducting “blind” lineups. In such lineups, the police officer running the procedure does not know which person is the suspect.
Small police agencies say that blind lineups could take more time and personnel, but the Virginia policy includes a cheap solution. The “folder shuffle” method places photographs of individuals in numbered folders, shuffles them, and allows eyewitnesses to examine each one without police looking inside.
Only nine agencies reported allowing the folder option. Many also do not offer uniform, meaningful instructions to eyewitnesses or train officers on recording lineups, Garrett said.