UVA Law Professor on Convicting the Innocent
When DNA evidence began springing people from prison, prosecutors discovered just how unreliable eyewitnesses can be.
Here in Virginia, 13 out of 16 cases of wrongful convictions involved inaccurate identifications. That led the state to issue model procedures for dealing with witnesses, but after nearly two years, very few have put those recommendations into practice.
The police lineup is a staple of film and television crime stories, but police departments don’t use that technique anymore. Instead, they ask people who’ve witnessed crimes to study photographs, and UVA Law Professor Brandon Garrett says that process can play with people’s memories.
“Our memory of the faces of strangers is so fragile that every time we recall it we can affect that memory, and there have been these amazing cases where DNA has freed someone. We know the person is absolutely innocent, but the eyewitness has the wrong person’s face in their mind, and some of them have even said afterwards, ‘It was what happened at the lineup that convinced me. It’s what the police did when they repeated that one photo, or when they told me, “good job,’ and gave me feedback.”
That’s not to say police officers try to influence the identification.
" Well obviously it’s bad if you do something blatant like you sometimes see on cop shows – and you tap your finger on the picture that you care about or you say, ‘Take another look at number three,’ That would be egregious suggestion, but what the studies have found is that just non-verbal cues, body language – people are really powerful non-verbal communicators, and so the administrator may just sort of wince if the eyewitness is lingering over one of the wrong photos.”
So Garrett and other experts advised the state on a new, model policy for assuring that eyewitness identification would be as accurate as possible. They assumed police departments would voluntarily adopt best practices.
“It doesn’t cost anything to just make sure you give the eyewitness careful instruction, make sure the photos are shown one at a time, and make sure the photos are shown by someone who doesn’t know which photo is the suspect’s photo.”
But after nearly two years, Garrett found only 6 percent of police departments had implemented the policy. Knowing that people are jailed because of mistakes in witness identification, Garret thinks it’s time for the legislature to require adoption of model procedures. Two neighboring states, North Carolina and West Virginia, have already done so.