Why Eyewitness Confidence Was Not Proof

What the campaign asked the public to understand

Eyewitness memory is not a recording. Research and court decisions developed after JJ’s trial documented how stress, weapon focus, cross-racial perception, time, suggestion, and feedback can reshape a witness’s memory and confidence.

The original campaign cited the New Jersey Supreme Court’s Henderson decision, Innocence Project data, and the New York State Bar Association’s work on wrongful convictions. Those materials showed that eyewitness misidentification was a major contributor to DNA exonerations and that multiple witnesses can repeat the same error when they are exposed to the same flawed procedures or outside influence. The Innocence Project’s 2012 case coverage placed JJ’s identification evidence in that broader wrongful-conviction context.

System variables in JJ’s case

Police control how photo arrays and lineups are administered. Safeguards such as double-blind administration, sequential presentation, matched fillers, immediate confidence statements, and instructions that the suspect may not be present reduce the risk of suggestion. The U.S. Department of Justice later adopted department-wide procedures for blind or blinded photo arrays and immediate confidence documentation.

In JJ’s case, a witness viewed roughly 1,800 photographs over many hours. The later lineup placed a Hispanic suspect among five non-Black fillers even though the first descriptions identified a Black shooter. Witnesses were reassured that police had the right person. These choices made it easier for uncertainty to become a selection.

Estimator variables in JJ’s case

The killing of retired NYPD officer Albert Ward occurred during a sudden, armed robbery inside an illegal Harlem gambling joint. Witnesses were frightened, several had limited viewing conditions, and some were elderly or affected by substance use. The identifications were also cross-racial: the witnesses were Black, while JJ is Hispanic.

What the 2024 expert review added

Psychologist Jennifer Dysart reviewed the identification record for the 2024 reinvestigation. As summarized by the National Registry of Exonerations, she identified the combined effects of stress, unfamiliarity, cross-racial identification, repeated exposure, and lineup fillers who did not sufficiently match the original descriptions. She also addressed the risk created by asking a witness to search an enormous mugshot collection: after hundreds of faces, recognition can shift from the crime to a recently viewed photograph.

That expert analysis was not testimony the original jury heard. It gave prosecutors a modern framework for assessing why several people could express confidence even when their first descriptions, viewing conditions, and selection histories did not align.

Confidence developed over time

Not every identification was immediate. Dorothy Canady selected two other people from a photo array and did not choose JJ in the February 2 lineup; she later said she recognized him after reflecting for roughly two weeks. Lorenzo Woodford initially reported that he could not make an identification, then later identified JJ while explaining that fear had affected his first response. The evidentiary question was therefore not simply whether a witness eventually pointed to JJ, but when, under what procedure, and after what intervening information.

What happened afterward

JJ’s lawyers later presented recantations from two identifying witnesses and argued that other witnesses had expressed doubts or qualified their certainty. The 2016 Appellate Division decision rejected the recantation evidence then before it. The post-conviction record traces those disputed reversals, the DNA exclusion, and the eventual dismissal without repeating the full chronology here.