You might not think six eyewitnesses could all be wrong, but they were.
The case was a 2010 murder. A man named Aaron Scheerhoorn was stabbed to death outside a nightclub in Houston. A crowd gathered, and six people told detectives that Lydell Grant was the murderer.
There was no physical evidence against Grant, which was amazing, considering how violent the crime had been. Moreover, Grant had an alibi.
Nevertheless, the jury felt the weight of the six eyewitnesses and convicted Grant. He was sent to maximum security prison in Gatesville, Texas, despite insisting upon his innocence.
Grant had been in prison for almost ten years when he finally caught a break. In 2019, a new method of DNA analysis allowed for testing of samples found under Scheerhoorn’s fingernails. That test implicated another man, who then confessed to the murder. Grant was exonerated and released from prison.
How could six eyewitnesses be so wrong? And how common is it for eyewitnesses to be completely mistaken?
According to the Innocence Project, over 375 people have been exonerated after new DNA testing. Of those, 21 were on death row. And, in the overwhelming majority of those cases, the original convictions were based on eyewitness testimony.
Human memory is not like video
Since the 1970s, experimental studies have shown that people’s memories are actually quite malleable and subject to manipulation. We tend to think that our memories are true records of the events, much like a video was taken. In reality, our experiences come back to us steeped in assumptions, prejudices and reconstructions.
In Lydell Grant’s case, it seems likely that the six eyewitnesses received small clues from the detectives about which person was considered the suspect. This could have been intentional, but it’s just as likely that the clues were unintentional and even subconscious.
Specifically, three of the eyewitnesses reported that the detective had said they picked the same person that others had. Two discussed their recollections with each other, confirming each other’s memories. The last witness said the detective praised him after he identified Grant.
Good practices can reduce mistakes
Ideally, police officers would engage in “double blind” identification processes, where neither the officer nor the eyewitness is told the identity of the suspect.
The witnesses should also view the photographs or in-person lineup sequentially, one person or photograph at a time, and answer whether or not they believe that person to be the perpetrator. The traditional “simultaneous” lineup, where six people walk out on a stage together, or an array of several photos is presented to a witness, has been shown to be prone to error.
Studies show that witnesses tend to view a simultaneous array and decide which person among them looks most like the suspect from their memory. This “comparative” judgment by a witness may not lead to the real perpetrator, as we have seen in many tragic wrongful convictions.
In addition, there should be some standardized instructions that help witnesses resist being primed to select any particular person. After they make the ID, the witness should be asked their level of confidence in the choice. And, the whole process should be videotaped.
Following these procedures can minimize the possibility of suggestion or bias in the eyewitness ID process.