True Crime, Wrong Story: Orleans Parish Episode Research Brief
Executive Overview
Of all the tools the criminal justice system uses to determine guilt, eyewitness testimony is the most trusted by juries, and one of the most unreliable in practice. The science of memory has established beyond reasonable doubt that human recollection is not a recording device but a reconstruction, a process susceptible to contamination, suggestion, stress, cross-racial bias, and post-event manipulation. The Anderson case deploys two of the most dangerous testimony archetypes in criminal law: the coerced or unreliable eyewitness and the rewarded jailhouse informant. Both categories carry enormous weight with juries. Both have catastrophic error rates.
Part I: How Memory Actually Works: and Why It Fails in Court
Memory Is Reconstructive, Not Reproductive
The foundational error in the legal system’s treatment of eyewitness testimony is treating human memory as reliable storage, assuming that what a witness recalls is an accurate copy of what they observed. Decades of research by cognitive psychologist Elizabeth Loftus and others have demolished this assumption. Memories are reconstructed each time they are recalled, not played back. During reconstruction, memories are vulnerable to incorporation of external information, including information provided by police officers, lawyers, prosecutors, or media coverage of the case.
Loftus’s misinformation effect experiments demonstrated that participants exposed to misleading information were reliably more likely to report seeing things that were not present. More disturbingly, individuals often cannot detect that their memories have been altered, they report contaminated recollections with the same or greater confidence than accurate ones. As Loftus has stated: “Without corroboration, it is almost impossible to tell if a memory is false or if it is true”.
The Confidence Illusion
Jurors consistently give more weight to testimony delivered with high confidence, even though research shows that a witness’s confidence in their identification is only slightly correlated with accuracy, and sometimes not correlated at all. A confident eyewitness can be entirely wrong. An uncertain eyewitness can be entirely right. Yet in court, confidence reads as credibility, and credibility drives verdicts.
The procedures of criminal investigation actively compound this problem. During police lineup procedures, detectives who have already identified a suspect can inadvertently, or deliberately, signal approval when a witness selects their preferred choice: a smile, a nod, a grunt of confirmation. Witnesses pick up these cues. Their uncertainty is resolved. Their confidence solidifies. And when they testify months or years later, they do so with conviction, often reporting an identification they made partly on social cues as though it were pure, uncontaminated memory.
Part II: The Eyewitness in the Anderson Case
A Witness Whose Story Changed
The eyewitness at the center of the Anderson case, the woman who was “found” within six hours after DA Jordan spent sixteen months unable to produce her, gave prosecutors a videotaped statement in October 2007. In that video, she said things that contradicted the testimony she would later give at trial in 2009. Prosecutors never disclosed that tape to the defense. The jury never saw it.
This is not just a Brady violation. It is a witness testimony problem of the highest order. The witness’s account changed between her initial statement and her trial testimony. The earlier version, the version closest in time to the events, was the one prosecutors found damaging enough to hide. This raises the fundamental evidentiary question: which version of her memory was more accurate? The reconstructed account delivered on the witness stand in 2009, after years of police contact, legal proceedings, and preparation with prosecutors, or the statement she gave within months of the crime?
Witness Cooperation Under Duress
The broader context of how Leon Cannizzaro’s office obtained witness cooperation is essential to understanding the reliability problems at stake. Cannizzaro’s DA office was ultimately sued and found to have fabricated subpoenas to coerce crime victims and witnesses into private, out-of-court interrogations, illegal interrogations not approved by a court. Witnesses who failed to appear were jailed on material witness warrants issued based on falsified information. One client, Lazonia Baham, was jailed for 8 days on a $100,000 bond for failing to respond to a fake subpoena.
Research on memory and witness compliance establishes clearly that testimony obtained under coercion or fear is inherently less reliable. When a witness knows, or fears, that non-cooperation means incarceration, their testimony is shaped by that fear as much as by their actual recollection. This creates a testimonial product that may satisfy the prosecution’s evidentiary needs while bearing only a distorted relationship to what the witness actually saw.
Part III: The Jailhouse Informant: A Systemically Corrupted Witness Category
What Makes Jailhouse Informants Different
Jailhouse informants, incarcerated individuals who testify that a fellow inmate confessed to them, occupy a uniquely dangerous position in criminal evidence. Unlike eyewitnesses, who are at least present at the scene of a crime, jailhouse informants are testifying about private conversations in custodial settings: conversations without recordings, without independent verification, and without any structural check on fabrication. Their testimony is powerful precisely because it mimics confession evidence, and juries treat it like a confession.
The incentive structure is catastrophically compromised. Informants receive leniency in their own cases, sentence reductions, improved prison conditions, or immunity from prosecution in exchange for their testimony. The promise or expectation of benefits from prosecutors creates a strong incentive to lie, and the secretive nature of the jailhouse informant system makes cross-examination and other legal safeguards ineffective. Defense attorneys cannot effectively challenge an arrangement they do not know exists.
The Numbers Are Damning
The Innocence Project reports that informant testimony was a factor in 19% of post-conviction DNA exoneration cases. Research by Warden (2004) found that informants, most of whom were jailhouse informants, provided false information in 45% of 111 wrongful convictions in the U.S. in which defendants had been sentenced to death. The Canadian Registry of Wrongful Convictions confirms the same pattern domestically: evidence from jailhouse informants contributed to over 15% of registered wrongful convictions.
The Innocence Project identifies jailhouse informant testimony as one of the leading contributing factors to wrongful convictions nationally, “playing a role in nearly one in five of the 367 DNA-based exoneration cases”.
Ronnie Morgan: The Professional Snitch
In the Anderson case, jailhouse informant Ronnie Morgan claimed Anderson had confessed to him in jail. Prosecutors secretly rewarded Morgan for his testimony, deals, benefits, considerations, without disclosing this to the defense or the jury. This is the classic jailhouse informant problem in its most concentrated form: a repeat player in the system, a “professional snitch” who had testified in multiple New Orleans cases, incentivized and rewarded for producing a narrative the prosecution needed.
The Innocence Project recommends that defendants be given complete information about a jailhouse informant’s criminal history, all benefits received, and every prior case in which they testified or offered to testify. None of this was provided in Anderson’s trial. The failure was not accidental, it was the operational method of the Cannizzaro office.
Part IV: Cross-Racial Identification: A Compounding Risk Factor
The Own-Race Bias in Witness Identification
The Anderson case involved a Black defendant in a predominantly Black neighborhood. The reliability problems with eyewitness identification are significantly amplified in cross-racial identification contexts, when the witness and the identified individual are of different racial backgrounds. Research establishes that eyewitnesses attempting to identify a stranger are over 50% more likely to make a misidentification when the stranger and witness are of different races. Approximately 42% of wrongful convictions can be attributed to cross-racial misidentification.
This phenomenon, known as the “own-race bias” or “cross-race effect”, operates through what researchers call Social Categorization: the mental shortcuts individuals use to rapidly process information can prevent them from attending to individual differences in the facial features of people of different races. This effect is seen across all racial groups, but is most pronounced when white witnesses attempt to identify people of other races. The problem is compounded by the fact that jurors, attorneys, and police remain largely unaware of its extent and continue to give undue weight to cross-racial identifications.
A 2024 study published in PubMed found that racial disparities in evidence that police have before placing a suspect in a lineup likely explain racial disparities in mistaken identifications that lead to wrongful convictions, meaning the bias enters the system before the lineup even takes place.
Part V: When the System Treats Unreliable Testimony as Reliable
Juror Psychology and the Weight of Testimony
Research consistently shows that jurors place heavy weight on eyewitness testimony when deciding guilt, and that this weight is largely independent of its actual reliability. The confidence displayed by a witness on the stand, the directness of their identification, the emotional conviction with which they point to a defendant: these are powerful signals that override analytical caution. Expert psychological testimony about the limits of memory is one of the few tools available to counter this bias, but courts have historically been reluctant to admit it.
A 2024 survey of 76 scientists on eyewitness memory phenomena confirmed broad expert consensus that eyewitness testimony reliability is substantially lower than the legal system treats it as being. The National Academy of Sciences issued a landmark report calling for tighter controls on eyewitness testimony and recommending that lineups be treated as double-blind scientific experiments, neither the witness nor the presiding officer should know in advance whether the suspect is in the lineup. These reforms have been adopted piecemeal, but many jurisdictions, including Louisiana, were slow to implement them during the period of the Anderson trial.
The Lesson of Anderson’s Witness
The eyewitness in the Anderson case was not necessarily lying. She may have genuinely believed, by the time she testified in 2009, that what she said was true. But between her initial statement and her trial testimony, years had elapsed. She had been subjected to prosecutorial preparation, police contact, and lived through the intense social pressure of a city demanding accountability for five murdered teenagers. Her memory was not stable. It was, almost certainly, reconstructed, shaped by post-event information, by the pressures of the legal proceedings, and by the consequences she knew would follow if she didn’t cooperate.
This is not a story about a lying witness. It is a story about what the justice system does to witnesses, and why treating their testimony as a reliable window into historical events is one of the most dangerous habits in criminal law.
Key Distinctions for Discussion
Episode Thesis Anchor
The witness in the Anderson case was “found in six hours”, and this was presented as proof that Jordan was incompetent. But finding someone and having a reliable witness are not the same thing. A witness found under pressure, whose contradictory prior statement was hidden, whose testimony was shaped by years of prosecutorial contact, is not a witness. She is a narrative construction.
The question this episode should force the audience to confront: When we say we have a witness, what do we actually have?