TAMPA — The thought of putting an innocent person in jail is horrific, says Tampa Police Chief Jane Castor.
“The officers, what they do out there every day to solve crimes and keep the city safe and put the bad guys behind bars is important to them,” Castor said, “but putting the wrong person behind bars is just devastating to them.”
Nationally, incorrect witness identification has been cited as a factor in 75 percent of all convictions later proven wrong through DNA testing, according to the Innocence Project. In Florida, that percentage is the same. Of the 12 DNA exonerations in the state, 9 involved eyewitness misidentification.
One of the Florida exonerations was Alan Crotzer, who was released in 2006 after spending 23 years behind bars for Tampa rapes and robberies he didn’t commit. The victim identifications of Crotzer were flawed, officials said. Initially, four of the five victims did not select Crotzer’s picture in a photo lineup. When they learned one of the women did, they changed their minds and selected Crotzer.
On Thursday, the National Academy of Sciences released a report recommending the best ways for law enforcement and courts to improve the accuracy of eyewitness identification.
“Eyewitness identification can be a powerful tool,” the report says. “However, the malleable nature of human visual perception, memory and confidence; the imperfect ability to recognize individuals; and policies governing law enforcement procedures can result in mistaken identifications with significant consequences.”
New police training policies, the report adds, as well as research and better data collection can improve the accuracy of the identifications.
Tampa Police already incorporate several of the report’s suggestions that previously were recommended by the Innocence Project, which cites the department as a model.
“In law enforcement, it’s incumbent upon us to make sure that we take every step, every measure, turn over every rock, look under every rug to make sure that we are holding the right people accountable for the crimes that they commit,” Castor said.
Four years ago, Castor made changes in how the department deals with eyewitnesses after attending a program at Harvard University that brought together law enforcement officials and academics to discuss various topics. At one session, they heard a presentation on problems with eyewitness identification and the consequences of wrongful identification.
“I won’t say it was a surprise,” Castor said. “It was more of a validation of what you knew, that witness identification really wasn’t as reliable as we have thought of in the past and that it an be influenced unintentionally.”
Castor noted that science has shown that while eyewitness identification is an important tool for law enforcement, it needs to be supplemented.
“I see that where we have grown as a society, but specifically in law enforcement, we put less and less weight on that to ensure you have DNA or other evidence or just other evidence in a case and not rely on that,” she said. “Years and years ago, that may have been good enough.”
Castor said she doesn’t want to criticize her predecessors who likely felt like they were doing the right thing but have been proven mistaken by research on human perception.
Castor said the Tampa Police Department joined with the Hillsborough Sheriffs Office and the Hillsborough State Attorney’s office to develop a uniform policy designed to enhance the accuracy of eyewitness identifications.
It’s an issue that law enforcement in Florida have struggled with for years. The same year Tampa changed its procedures, the state Legislature failed to pass a bill that would have mandated some of the recommended practices.
The Florida Department of Law Enforcement also implemented some - but not all - of the procedures recommended by the Innocence Project. And a state Innocence Commission created by the Florida Supreme Court issued guidelines for law enforcement agencies that suggested the same procedures adopted by FDLE.
Seth Miller, executive director of the Innocence Project of Florida, said the FDLE policy represents a good effort that falls short.
“Anytime there is a concern and folks make an effort to address it, you have to be thankful,” Miller said. “So obviously, we’re thrilled that amongst the many issues that are challenging law enforcement at any given time, we’re happy they’ve spent time to address this issue.”
Among the procedures adopted by FDLE is ensuring that the “filler” pictures in lineups - that is the photos of people who are not suspects - are similar to the suspect. So, for example, if the suspect has a beard, all the photos in the lineup show people with beards.
FDLE also gives witnesses instructions before viewing the lineups that are designed to alleviate any pressure they might feel to pick someone.
Miller said FDLE has “gone about 60 percent of the way,” and he hopes the agency goes further, which would also encourage other police departments around the state to do the same.
One policy recommended by the Innocence Project and incorporated in Tampa is called “blind” administration of a photo lineup.
In this method, the officer who shows the photos to witnesses is not involved in the investigation and doesn’t know which picture is the suspect or even if a suspect is among the photos being shown.
The method is aimed at guarding against any kind of suggestion, however inadvertent, being given to the witness, who might feel directed or pressured to pick a certain picture.
An alternate suggestion would allow a detective involved in the investigation to administer the photo lineup but have the detective positioned so he can’t see what pictures the witness is viewing.
The practice was cited in a Florida Senate legislative analysis as the one step most scientific studies agree is “critical to untainted witness identification.” And the National Academy of Sciences report recommended blind administration of photo lineups “to avoid the unintentional and intentional exchange of information that might bias an eyewitness.”
But it is not included in FDLE’s policy. Nor is another recommendation of a “confidence statement,” which would require a record of a witness’s words about how confident he or she is of the identification. According to the Academy of Science report, this would help to assess the reliability of witness testimony during a trial.
When asked about the fact that FDLE does not include these steps in its policies, a spokeswoman issued a statement saying the agency follows procedures required by the Commission on Accreditation for Law Enforcement Agencies and the Commission for Florida Law Enforcement Accreditation. Miller said the Academy of Science report shows it’s time for the procedures to be universally adopted by law enforcement.
“There can’t be any more dispute about what’s what scientifically” valid, Miller said. “We have to update our policies because it’s the right thing to do and it’s going to lead to the best possible criminal justice outcomes.”
The Thursday report also went further than the Innocence Project’s core recommendations to suggest other procedures, including video recording eyewitness identifications from lineups.
Tampa doesn’t do that, and spokeswoman Laura McElroy said the department is awaiting the results of studies into whether that would be effective.
“I think we have enough safeguards in our policy,” Castor said. “I feel very confident in the witness identification.”
Castor said the changes were more subtle than dramatic and that there was minimal resistance in the department. “The only negative feedback we got was that it would be too time-consuming and personnel-intensive because you have to bring someone else in, but that hasn’t proved to be an impediment,” she said.
Castor laughed. “We’re police officers,” she said. “We feel we have to complain about something.”