'I Don't Think You Did This, But I Can't Fix It': How To Improve Eyewitness Evidence
Advocates for the wrongly convicted often say they’d much rather prevent mistakes than fix them after years of incarceration. But how to do that is up for debate.
Defense lawyers in Massachusetts say police and prosecutors have made progress in how they gather eyewitness evidence, but the courts have not quite caught up with the science.
'I kind of went kicking and screaming'
In 1999, the U.S. Department of Justice put out a report on how to collect eyewitness testimony using science-based methods, from asking open-ended questions to designing more neutral lineups.
At first, legal experts say, not much changed in courtrooms. But the report did lead to police trainings, including one, in 2003, attended by an officer named Bill Brooks.
“I kind of went kicking and screaming,” Brooks said.
At the time, Brooks was deputy police chief in Wellesley, Massachusetts. He was skeptical about eyewitness science because he felt police were the experts on criminal investigation.
“So I went to the training prepared to be unimpressed,” he said.
But Brooks was impressed. He persuaded his own chief to let him write a new science-based policy for the department. The district attorney then sent that policy to other departments. Brooks went on to develop training manuals and videos.
Brooks is now police chief for Norwood, Massachusetts. He’s known nationally for his work on eyewitness identification. He was on a 2013 Massachusetts Supreme Judicial Court committee that studied the issue and has worked with the National Academy of Sciences on eyewitness recommendations. He’s even been invited to speak with members of the Innocence Project, which is often at odds with police.
“Don’t tell my friends!” Brooks likes to joke.
'No one wants the wrong guy to go to prison'
Brooks also goes around the country giving workshops to police officers. At one training he led in Maryland, he showed a video of a mock robbery, and then a lineup of suspects who all looked similar. Some of the trainees picked one guy, some another, others picked no one.
“And how certain are you?” he asked the group. “Seventy-five percent? OK, how good a look did you get at the guy?”
After the first lineup, he showed the group a second lineup with different people in it.
“See anybody who looks familiar in that second lineup?”
That’s when he admitted to the group that the actual culprit was in the second lineup, and had they seen that one first, “we wouldn't have sent number one in the first lineup to prison.”
Brooks sees himself as a link between science and law enforcement. He said police are often under immense pressure to zero in on a suspect, and shortcuts are tempting.
But at the same time, “I think if you approach the police in the right way, you can get the police on board,” he said. “Nobody wants the wrong guy to go to prison.”
“That's everyone's nightmare, that's a tragedy,” said Steven Gagne, an assistant district attorney for Hampshire and Franklin counties. “And prosecutors, as much as anybody else, want to avoid that.”
Not only is wrongful conviction horrendous for the person who loses their freedom, Gagne said, “but secondly, if you get the wrong guy, that means that the actual perpetrator is still out there and possibly committing more offenses.”
That said, he’s not sure recent judicial rulings on eyewitness science have been what he calls “earth shattering.” He said good prosecutors have always been aware of the pitfalls in eyewitness memory, and that’s why his office tries never to hinge a case entirely on what a witness thinks he or she saw.
“Witnesses tend to dig in their feet and perhaps be reluctant to admit that maybe they're wrong,” he said.
Still, some of Gagne’s cases do rely heavily on eyewitness testimony, but he said they always have at least some corroboration, like a confession or forensic evidence.
“And I think that I'm more wary of that now than I was earlier in my career, where you figured, well, if you have one or ideally two eyewitness identifications, you're good to go,” he said.
Sometimes, Gagne said, defense attorneys try to stir up too much doubt over eyewitness testimony. For instance, in a recent gunshot case, the suspect admitted being at the crime scene and the victim testified he bit his attacker during the struggle.
“And sure enough, the defendant had a nice set of teeth marks on his forearm when he was arrested a day later,” Gagne said. “Still, the defense went with a mistaken identification defense, and I pretty much called it out for what it was during my closing argument. I said, ‘That insults your intelligence, ladies and gentlemen.’”
The surprising science of memory
But Gagne said occasionally his office is uncomfortable with how police get a witness to pick a suspect — and he won’t use the evidence at trial. For instance, police may put together a photo array that makes the suspect really stick out from the so-called fillers because of their clothing or skin color, or the photo background.
“We see those as opportunities not only to slow down and pump the brakes on that particular case,” Gagne said, “but to work with that detective or that police department, use it as a teaching moment.”
But even when all parties want to follow best practices, the science sometimes evolves faster than law enforcement can keep up.
“Some of the findings are very intuitive,” said Jeffrey Starns, a psychology professor at UMass Amherst who studies memory and eyewitness testimony.
For instance, juries generally understand why it’s important to consider how far a witness was from a crime, or whether it was dark outside.
“Some things are much more surprising,” Starns said. “Like one thing that's been an element of jury instructions in courts is you should pay attention to how confident the witness is.”
Starns said that’s an outdated assumption. In fact, he said a witness can become more confident after they testify multiple times, or get interviewed repeatedly by police. The process itself reinforces the belief, even if it's wrong. But he said jurors are rarely taught how to make that distinction.
One question that has not been resolved is whether to show photos to witnesses at the same time or one after the other — a simultaneous vs. sequential lineup or array.
The simultaneous lineup — where a suspect and fillers are displayed in a row — used to be the gold standard. (Think of the movie poster for “The Usual Suspects.”)
Now, according to Bill Brooks, most police departments consider sequential arrays the most reliable, because witnesses are comparing faces to their memory, not just to the other faces.
But Jeffrey Starns said there are drawbacks to that approach too.
“It certainly makes you more cautious for whatever reason,” Starns said. “So you're less likely to identify innocent suspects, but you're also less likely to identify guilty suspects.”
By now, Starns said, the scientific field has gone beyond courtroom dramatics and onto “probability models.” For instance, researchers want to let juries know what percent of the time certain kinds of memories are likely to be true. But often judges don’t want experts to testify on those details.
“And one of the concerns is that detracts from the role of the jurors,” Starns said. “If you're giving them a mathematical procedure to have to evaluate this evidence, it's not letting ... the juror be the one that evaluates the evidence.”
'They wanted us to hand it off to someone who knew nothing about the case'
Police chief Bill Brooks said he had to come to terms himself with the less intuitive aspects of eyewitness science.
Even when there’s immense public pressure to make a quick arrest, Brooks tells his detectives it’s better to slow down the process to get it right.
“Once your witness picks the wrong guy, you've lost your witness,” he said. “It's important your witness not pick the wrong person because you'll never solve that case. And that's a point to be made in training.”
For instance, he said composite sketches — when an artist draws a suspect’s face from the witness’s description — can be problematic.
“The witness walks away from the composite procedure with an image of the composite, not an image of the offender, and the research shows that doing a composite or a sketch damages memory,” Brooks said.
Among the most game-changing eyewitness procedures is called blind administration. That’s when the detective who’s working a case leaves the room and a different officer shows the witness photos. At first, Brooks was wary.
“They wanted us to hand it off to somebody who knew nothing about the case and have them show it,” he said, “which I took as insulting, like we were trying to throw an ID or something.”
But now Brooks requires blind administration in his own police department because researchers have found the identification is more reliable without the case detective watching.
“The reason for that is so you don't get any body language feedback,” said defense attorney Lisa Steele, a Shrewsbury-based defense attorney who has written about eyewitness protocol. “The witness isn't going to be subconsciously swayed by anything that the (detective) says.”
Even though the research is clear, Steele said, blind administration doesn’t always happen.
As one example, she pointed to a 2016 case in which Steele’s client, Luther Williams, was accused of a shooting in Springfield.
As seen in a police video, two detectives sat at a table while a woman who witnessed the shooting looked through a pile of photos. According to court documents, both men knew who the chief suspect was, so they noticed when the witness picked someone else.
“Do you want to look through them again?” said one detective.
“Looks like he’s the man,” she said.
When she seemed to settle on the photo, a detective working the case continued to question her.
“So do you think that you would be able to say, 'This is the guy' or do you think maybe he looks like that? You know what I mean? Because I don't want you just to go by, like, the expression,” he said.
“We're not putting pressure on you to pick someone out if you don't know for sure,” he added.
Steele has watched this video many times.
“It's not ... that I think they're wanting to influence her directly,” she said, “but they know this ain't the guy.”
In the end, the witness shrugged and backed off her pick. Later, the police said another witness did pick Williams out, and he was eventually convicted of the crime.
A spokesman for the Springfield police would not discuss this case. He said the department follows state guidelines for eyewitness interviews.
But those guidelines include blind administration, which was clearly not followed here.
'Holy crap, I don't think you did do this, but I can't fix it'
Bill Brooks said he does try to keep an eye on which police departments in Massachusetts are following best practices, and he’s urged the Municipal Police Training Commission (MPTC) to make sure eyewitness evidence is part of the standard training.
“Could departments have missed the mark and not done the training and not be using an adequate policy?” Brooks said. “That is possible.”
Reached by phone, the director of the MPTC said he could not talk to the media without permission from a state government representative. That representative did not return emails requesting permission.
Brooks said Springfield is among the departments he’s called in the past after finding problems in their eyewitness procedures, but he did not want to give details.
Without consistency across departments, Steele said the risk of wrongful conviction goes up.
Steele has worked on cases that changed the evidence rules going forward, but she’s never won an appeal for a client by challenging the eyewitness testimony at trial. That’s why she wants police, prosecutors and attorneys to improve their game from the start.
“Among the hardest things in the defense attorney's life is when your client looks up at you and says, ‘Miss Steele, I didn't do this,’” she said. “And you'll look at the record and you bang your head on the desk and you say, ‘Holy crap, I don't think you did do this, but I can't fix it.’”
She would like more law schools to teach eyewitness science. And she wants procedures better legislated in Massachusetts, including mandatory, state-funded police training and videotaping of all witness interviews.
Late last year, the Massachusetts governor signed a massive police reform package. Among the components, it created a civilian commission for police certification, which includes some oversight of training and conduct. But it made no mention of eyewitness testimony.
Frankly, Brooks doesn’t think legislation is the answer, because police don’t like being told how to do their jobs. A better solution, he said, is just teaching them the best way to figure out who committed the crime.
Brooks knows there are “awful” cases where an innocent person goes to prison because the police messed up.
“And then there are the wrongful conviction cases where the witness just picked the wrong person. The police did everything right. They've been properly trained by their department. The witness just made an error,” Brooks said. “And the point that I make is if, God forbid, if you ever have a wrongful conviction case, you want it to be that second kind.”
This is the third in a three-part series, Questioning The Witness — a look at the evolving science of eyewitness testimony and the people with the greatest stake in how it’s used.