7 people saw the same shooting. But on some details, they couldn’t disagree more.

Charlotte-Mecklenburg Police investigate the scene of an officer involved fatal shooting in the 6200 block of Albemarle Road in east Charlotte on Thursday, January 26, 2017.
Charlotte-Mecklenburg Police investigate the scene of an officer involved fatal shooting in the 6200 block of Albemarle Road in east Charlotte on Thursday, January 26, 2017. dtfoster@charlotteobserver.com

Seven people witnessed the same shooting.

On many details, they agreed.

On others – not so much.

They described the two men involved by the wrong race. They disagreed over which man was the aggressor and who fired the first shot. One witness told authorities that he left the scene, got a pint of whiskey, went home and then called police to tell them what he saw.

Police and prosecutors use eyewitness statements to help make arrests and convict suspects in thousands of cases.

But January’s fatal shooting of a motorist by an undercover officer along a busy east Charlotte street reflects what researchers have warned for decades: Eyewitness accounts are often unreliable.

Josue Javier Diaz, 28, died on Albemarle Road in east Charlotte on Jan. 26. Charlotte-Mecklenburg police said his truck sideswiped the officer’s unmarked vehicle, then stopped. Diaz got out of the truck, police said, and fired a pistol at the officer and the officer, who is still unidentified, fired back, striking him.

Prosecutors decided last month not to seek criminal charges against the officer, but acknowledged that witness accounts conflicted and physical evidence could not determine who fired first.

A 120-page report from the Mecklenburg County District Attorney’s Office shows some witnesses observing the same events gave vastly differing descriptions – sometimes contradicting physical and scientific evidence.

The case illustrates an ongoing debate about how much police, prosecutors and juries should weigh eyewitness statements. It is significant because mistakes can lead to wrongful convictions or let criminals go free.

Of the first 250 cases where DNA evidence exonerated wrongfully convicted people, 190 involved inaccurate eyewitness accounts, said Brandon Garrett, a law professor at the University of Virginia and author of “Convicting the Innocent.”

“Eyewitnesses are much less reliable than people think,” Garrrett said. “They have real problems identifying faces.”

Garrett is among a growing number of legal scholars who argue eyewitness testimony significantly influences jurors and that more steps should be take to ensure accounts are trustworthy.

The American Psychological Association has said that “research shows that juries tend to ‘over believe’ eyewitness testimony.”

But retired Mecklenburg County senior resident Superior Court Judge Richard Boner and others say prosecutors sometimes must rely on eyewitness testimony or risk letting dangerous suspects go.

Jurors should judge for themselves the credibility of witness testimony, the same as they would for other evidence.

“That’s the jury’s call,” Boner said. “The defense has a chance to cross-examine the witness.”

Human weakness

Experts say most eyewitnesses are trying their best to recall what they saw, but are limited by the frailties of the human brain.

A 2014 report by the National Academy of Sciences examined 30 years of research on memory and identification. The study confirmed previous research, which found that memory is subject to bias, distortion and contamination under the best circumstances, and becomes less reliable under stress.

Researchers found witnesses sometimes pick up inadvertent clues from officers who conduct photo lineups of suspects. They said witnesses confidence in their identification is falsely inflated by confirming feedback from when they talk to investigators to the time of the trial.

In response, North Carolina, Oregon, New Jersey and other states have passed laws in recent years regulating how police conduct suspect lineups. CMPD, for example, tries to avoid bias by using a double-blind administration of lineups, in which neither the person conducting the procedure nor the witness knows the identity of the “right” suspect.

But some legal scholars and activists have called for more steps. Judges should hold preliminary hearings, outside the view of jurors, to determine if the accounts are credible, some argue.

North Carolina courts ask the jury generally to consider “the opportunity of the witness to see, hear, know, or remember the facts or occurrences about which the witness testified,” but doesn’t have specific, detailed instruction, according to the UNC School of Government, which provides guidance to state and local government bodies.

“I have more faith in juries,” Bill Stetzer, assistant district attorney for Mecklenburg County, told the Observer. “They’re pretty good at picking out the truth.”

In any case, Stetzer said his office carefully vets its witnesses. “If we don’t believe them, they won’t see a jury,” he said.

Sorting fact from fiction

In the Diaz case, the officer said he never had time to identify himself as a police officer before the shooting. “He was either gonna shoot me or I was gonna shoot him,” the officer told detectives.

Diaz was shot six times to the chest, abdomen and hand, all on the left side of his body, an autopsy report showed.

Police said investigators recovered a .22-caliber revolver at the scene and identified a bullet hole and projectile in the vehicle the officer had been driving. Prosecutors ruled that the officer feared for his life and was legally justified in shooting Diaz.

Some eyewitness statements seemed to bolster the police account, while others contradicted it.

One witness told police that she heard two shots, and then Diaz fell to the ground. She said the undercover officer walked over and shot him again.

Prosecutors said that none of the physical evidence is consistent with Diaz being shot while he was on the ground.

Another witness said after hearing the first gunshot, he saw an SUV pull up behind a pickup truck, and somebody get out of the SUV the officer was driving. That person was “yelling and screaming.”

Prosecutors say evidence suggests both vehicles were already stopped when the shooting occurred.

The witness said the officer pulled a pistol out of his shirt and walked up to Diaz’s vehicle, and “the next few shots came as (the officer) ‘stanced himself, ... leveled his gun, and he let at least three or four more rounds go.’ ”

Another witness said he heard a gunshot and then looked in his side view mirror to see a man exit the car behind him. He heard three more shots – “pap, pap, pap” – and saw the person fall to the ground.

Then, according to the District Attorney’s report, “Witness #2 stated that he left the scene, went to get a pint of whiskey and went home. He had a couple of drinks and then called the police about what he saw.”

Stetzer, the prosecutor, said his office evaluated the credibility of eyewitness statements by trying to determine whether they were consistent with other credible evidence. Prosecutors also tried to assess whether the witness was in a position to see and hear what they reported.

“I don’t think they were making up stuff,” Stetzer said. “They reported what they saw. ... Under stress, people don’t always see accurately. People remember things differently.”

Clasen-Kelly: 704-358-5027; @FrederickClasen