The question many crime victims face when they are asked to identify their attacker: How accurate is our memory?
The year was 1985. A series of men was marched into a police lineup and detectives asked Penny Beerntsen to identify the man who had raped and brutally beat her.
"When I came to Steven Avery in the lineup the hair on the back of my next stood up," said Beerntsen. "I could feel the color drain from my face."
She says she was 99 percent certain Avery was the man who attacked her. Her testimony helped put him away. He spent 18 years behind bars before DNA proved he didn't do it.
"That was the worst day of my life," said Beerntsen. "I was absolutely devastated."
Beerntsen picked Avery out of what's called a simultaneous lineup, in which possible suspects are lined up right next to each other.
A new round of research, though, shows that when eyewitnesses are presented with simultaneous, conventional lineups, they pick an innocent member of the group 18 percent of the time, almost one in every five attempts.
"They are assuming that he is probably there somewhere, that whoever looks most like him, must be him, and that is a problematic assumption," said Gary Well of Iowa State University.
The study showed sequential line ups, in which suspects are viewed one at a time, resulted in an innocent person being selected just 12 percent of the time.
In Penny Beerntsen's case, DNA connected Gregory Allen to the crime, but only after 18 years, after he committed at least one other rape.
"If these reforms are implemented, it will correct many of the problems we've had with erroneous eyewitness identifications," said Rob Warden of the Northwestern Center on Wrongful Convictions.
While wrongly convicted in Penny Beerntsen's rape case, two years after he was released, Steven Avery murdered a woman, and he is now back behind bars.
Monday night, Chicago Police and the Cook County Sheriff's Department say that while they are always willing to look at studies like this, they have no immediate plans to change their lineup procedures.
