The Anatomy of a Case: Marvin Williford

County convicted in: Lake County
Convicted of: Armed robbery, murder, and unlawful restraint
Year convicted: 2004
Sentence: 80 years (17 years served)


In North Chicago on January 20, 2000, three people conducted a home invasion of Delvin Foxworth. As they did so, they ransacked the house. All three of them took turns beating Foxworth with a 2×4, and one perpetrator bound Foxworth with duct tape and sprinkled him with gasoline before the entire room went up in flames. Foxworth knew the perpetrator that bound him with tape and sprinkled the gasoline — a person he referred to as “T.” Foxworth’s girlfriend, Delia Conners, was the sole eyewitness, but T and the other perpetrators were unknown to her. Foxworth lived for two years after the incident.
In 2004, following extremely troubling identification procedures, Marvin Williford was mistakenly identified as “T,” convicted, and sentenced to 80 years in prison. Conners had been shown a single picture of Williford just before walking into the grand jury (in 2000), shown that picture again in 2002, right after Foxworth passed away. After some hesitation, she ultimately chose Williford from a photographic lineup in 2003. At trial, Willfiord testified and maintained his innocence, and no forensic evidence connected him to the messy crime scene. Nonetheless, Williford was wrongly convicted based on the misidentification.
Williford has consistently challenged his conviction on appeal and in post-conviction proceedings. Between 2013 and 2014, new DNA testing was conducted on the 2×4 and other items from the crime scene. None of the DNA found on the items belongs to Williford. In addition, the DNA matched another suspect—a DNA match from the rape kit of a 1992 rape and murder of a Waukegan child, Holly Staker.
Williford’s legal team at the Exoneration Project filed an amended post-conviction petition and a petition for relief from judgment in 2015. This petition asserted Williford’s claims of actual innocence in light of DNA testing but also pointed to new developments in social science confirming the unreliability of witness identifications made like those in the circumstances presented in this case. Williford also presented new evidence concerning other constitutional issues that, the petition alleges, deprived him of a fair trial. Williford’s efforts, to date, have been unsuccessful but litigation is ongoing in federal court.

Key Factors In Williford’s Wrongful Conviction

Lack of evidence implicating Williford but pointing to other suspects:
• The DNA on the duct tape, gas can, and wooden 2 x 4 used to beat Foxworth was not a match for Williford. He was also not a match for any other evidence at the crime scene, However, one of the assailants’ DNA matched DNA found at the scene of a previous crime: 11 year old Holly Staker’s rape and murder in 1992 (pages 9-10).
• Fingerprint evidence taken from the scene, including from the soot related to the fire, did not belong to Williford

Mistaken Witness Identification

The problem with eyewitness identification: Advances “in DNA testing have confirmed that ‘eyewitness misidentification is now the single greatest source of wrongful convictions in the United States, and responsible for more wrongful convictions than all other causes combined.’” People v. Lerma, 2016 IL118496, ¶25 (quoting State v. Dubose, 699 N.W.2d 582, 591-92 (Wis. 2005).
• The Second District Court held that Conner’s “testimony and her identification of [Williford] made up the majority of the State’s case, and ‘absent her testimony identifying defendant as one of the perpetrators, there is a reasonable probability that [Williford] would have been acquitted’” (page 34).
Suggestiveness and contamination: Conners was shown a single photo of Williford just before her Grand Jury testimony in 2000 where she misidentified him as an assailant. Conners was reminded of her prior misidentification and shown this same picture again in 2002. Conners was then shown a photo array in 2003, where Williford was also included. He was the only suspect included in all three confrontations, and she was never asked to identify the other two assailants (pages 7-9).
An unreliable, though confident, misidentification. Social science now shows that the confidence a witness has in their own testimony can be strengthened “by repeat questioning or by information that the witness receives during, or after, the identification process” (page 22). In other words, showing Conner a photo of Williford multiple times before her testimony likely influenced her identification of him as an assailant. As a matter of social science, the eyewitness identification in this case was unreliable (pages 41-45).


  • Amended post-conviction petition
  • Appeal opening brief
  • Orders denying relief
  • Update: EP attorneys are currently working on a habeas corpus petition


RECENT FILINGS (as of 1/30/24)