SJC Reverses Murder Conviction After Unreliable Gunshot Residue Evidence Admitted at Trial: Commonwealth v. Williams (SJC-12702, September 22, 2026)

Keith Williams was convicted of murder after a 2014 shooting during a Dorchester parade that killed Dawn Jaffier and wounded Lealah Fulton. No witness identified him as the shooter. The Commonwealth’s case rested on his disposal of the murder weapon seconds after the shooting and on gunshot residue evidence that the Supreme Judicial Court has now ruled unreliable. The court vacated all four convictions and authorized a new trial.

The error centered on expert testimony about gunshot residue particles found on Williams during testing. His test results were negative under the Massachusetts State Police standard requiring at least three particles for a positive result. But the Commonwealth’s expert testified that one particle was found on Williams’s hand and another on his T-shirt, while none were found on his two companions, Jordan Reed and Brian Joyce. The prosecutor urged the jury to infer from this that Williams was the shooter.

The Facts

On August 23, 2014, Williams, Reed, and Joyce were outside a convenience store on Blue Hill Avenue. Security footage showed Williams entering the store around 7:26 a.m., making a purchase, and leaving. Forty-five minutes later, Reed, Joyce, and two others entered. Shortly after, Tevan Williams (no relation to the defendant) and Wesson Colas, a codefendant, also entered. The two groups eyed each other tensely but exchanged no words.

After both groups left the store and walked toward McLellan Street, Colas pulled a silver handgun from his waistband and pointed it at the group that included Reed and Joyce. Williams, who had rejoined Reed and Joyce, then fired at least five rounds at Colas and Tevan. None of those bullets hit their apparent targets. One struck Jaffier in the head as she crossed the street with friends, killing her. Another hit Fulton in the leg as she watched the parade.

Williams, Reed, and Joyce ran down McLellan Street. Lashonda Osorio heard the shots from her apartment 500 feet away, looked out her window, and saw the three men walking down the street acting nervous. She heard Reed tell Williams to “get rid of it.” She watched Williams go to a house across the street and dispose of a firearm under the porch. As they left, Reed told Williams to walk, not run.

Osorio alerted police. Officers recovered a .357 revolver containing six empty shell casings from under the porch. Ballistic testing confirmed the weapon fired the bullets that killed Jaffier and injured Fulton. Police stopped and handcuffed Williams, Reed, and Joyce nearby, then brought them to headquarters.

The Gunshot Residue Testing

Police collected gunshot residue samples from the hands of all three men at 11:53 a.m., more than three and one-half hours after the 8:16 a.m. shooting. An officer collected Williams’s T-shirt the next day, and a criminalist tested it for gunshot residue thirteen months later. All tests came back negative under the three-particle threshold.

John Drugan of the State police crime laboratory explained at trial that gunshot residue particles are released when a firearm is fired. Each particle consists of three components not generally found together in nature: barium nitrate, lead styphnate, and antimony sulfide. A positive test requires detection of at least three such particles from a single sample. The threshold accounts for contamination risk and ensures accurate interpretation.

Drugan testified that one gunshot residue particle was found on a stub from Williams’s hands and another on the stub from his T-shirt. No particles were found on Reed or Joyce. Despite these findings, Drugan agreed on cross-examination that there was “nothing to be drawn from” the negative test results. His written reports stated that “no conclusions can be made as to whether these individuals may have discharged a firearm.”

Drugan also testified that gunshot residue particles are microscopic, resilient, transferable, and known to persist on handcuffs and surfaces inside police cruisers and stations. The Commonwealth acknowledged at trial that all three men had been handcuffed, transported in police vehicles, and detained for hours before testing.

The Trial

The defense argued that no witness positively identified Williams as the shooter. Off-duty Detective Arthur Hall-Brewster was outside the barbershop at the scene. He noticed Reed and Joyce, some of whom had hands in their waistbands, suggesting they might be armed. He sensed tension between groups and saw Colas with a gun. He heard gunshots but never saw Williams.

Amanda Griffith testified that the shooter was the largest of three men running from the scene. Williams was distinctly taller and larger than Reed or Joyce. But Griffith also described the shooter as about five feet eight or nine inches tall, not heavy set, and wearing a green shirt and dark pants. Williams was over six feet tall, heavy set, and wearing a bright blue T-shirt with green cartoon characters, blue shorts, and bright blue and orange sneakers.

Brenda Robinson felt a “heavy” shooter lean on her as he fired but could not describe his appearance. The defense’s sole witness, Akeisha Lewis, described the shooter as wearing black pants and a white shirt with black sleeves, consistent with Reed’s clothing that day, not Williams’s.

No forensic evidence definitively tied Williams to the shooting. DNA and fingerprint tests on the revolver were inconclusive. The gunshot residue tests were negative.

In closing argument, the prosecutor acknowledged the negative test results but told the jury that the three-particle threshold was “not a jury standard of practice, that’s a Massachusetts State Police standard of practice.” The prosecutor urged the jury to interpret the evidence however they wanted and to rely on the fact that particles were found only on Williams, not on Reed or Joyce, to conclude Williams was the shooter. The jury convicted Williams of murder in the first degree on a theory of deliberate premeditation, along with the other charges.

The Court’s Analysis

Justice Dewar, writing for the court, held that the Commonwealth failed to establish the reliability of using subthreshold gunshot residue particles as evidence that Williams fired the weapon. The proponent of expert testimony must show that the methodology is reliable when applied to the specific issue before the court. Here, that issue was whether the defendant, rather than his companions, fired the weapon.

The Commonwealth offered no evidence that detection of one particle in each of two negative test results reliably supported that inference. The court noted that the Commonwealth’s own expert testified that nothing could be drawn from the negative results and that the three-particle threshold exists to guard against contamination and ensure accurate interpretation. Without any metric for assessing reliability, the testimony should not have been admitted.

The court distinguished cases where expert testimony is admissible even when not definitive, such as Commonwealth v. Torres, 469 Mass. 398, 407-408 (2014), where the expert led the jury through the factors applied in analyzing a bloody shoeprint. Here, the Commonwealth provided no analysis to support the reliability of the subthreshold particles as proof of firing a weapon.

The court also clarified that its decision in Commonwealth v. Johnson, 463 Mass. 95 (2012), did not establish that detection of particles below the scientific threshold necessarily supports an inference that a person fired a gun. In Johnson, the defendant newly raised arguments on appeal that gunshot residue analysis was junk science and that testimony about two detected particles was false. The court rejected those arguments but said nothing about whether such evidence is reliable proof of firing a weapon. To the extent Johnson suggested otherwise, the court disavowed that suggestion.

Prejudice and Retrial

The court found the error prejudicial. The disputed testimony figured prominently in the Commonwealth’s case. In opening statement, the prosecutor mentioned that gunshot residue particles were found on Williams without noting the test was negative. In closing, the prosecutor highlighted that particles were found only on Williams and undercut the defense argument about contamination by pointing out that Reed and Joyce also had been handcuffed and transported but tested clean.

The Commonwealth’s evidence identifying Williams as the shooter was not overwhelming. He was seen disposing of the weapon and fit one witness’s description of being the largest of the three fleeing men. But he was not seen in the area immediately before the shooting, no witness or forensic test positively identified him, and another witness’s description matched Reed. Reed was also observed with his hand in his waistband before the shooting, and Osorio heard Reed directing Williams’s actions after the shooting.

The court rejected Williams’s argument that insufficient evidence supported deliberate premeditation, meaning he may be retried for murder in the first degree. Viewing the evidence in the light most favorable to the Commonwealth, a rational jury could find that Williams formed a decision to kill Colas after Colas displayed and pointed a firearm, then carried out that decision by firing at least five rounds. Although the interval was brief, deliberate premeditation may be formed in seconds.

Why It Matters

This decision clarifies the Commonwealth’s burden when introducing forensic evidence that falls below established scientific thresholds. Prosecutors cannot simply present subthreshold results and argue that the jury may draw whatever inferences they wish. The Commonwealth must establish that the evidence reliably supports the inference it asks the jury to draw.

The ruling also limits the scope of Johnson, which prosecutors had cited as authority for admitting subthreshold gunshot residue evidence. The court made clear that Johnson involved different issues and did not foreclose challenges based on a more developed record or advances in science.

Going forward, the Commonwealth must provide a foundation showing that detection of particles below the three-particle threshold reliably proves a contested issue at trial. The decision does not categorically bar such evidence, but it requires proof of reliability that the Commonwealth failed to provide here.

The court also separately vacated Williams’s conviction for carrying a firearm without a license because the jury were not instructed that the Commonwealth bore the burden of proving he lacked a valid license, an error under recent precedent. All convictions were set aside, and the case was remanded for further proceedings.

Read the full opinion: Commonwealth v. Williams (SJC-12702, September 22, 2026).