Confirmation bias played a role in the RI State Crime Lab's review of invalid tool tags, report says • Rhode Island Current

Two Rhode Island State Crime Lab employees who performed a specific type of forensic testing on evidence in gun cases are no longer working at the lab, while another has been “permanently suspended,” the lab director confirmed to the Rhode Island Current on Wednesday.

However, the staffing shortage is likely to be temporary, said Dennis Hilliard, director of the crime lab. The nationally accredited laboratory has outsourced tool tag testing to two former NYPD examiners for the next six months, Hilliard said in an interview after a meeting of the five-member State Crime Laboratory Commission at the Rhode Island attorney general's Cranston office.

The commission met to review the findings an independent report commissioned after officials discovered a discrepancy in the identification of a weapon allegedly used in a 2021 Pawtucket murder case and paused all toolmark testing on August 20. The break in toolmark testing at the laboratory in Fogarty Hall on the University of Rhode Island's Kingston campus was made public on September 3rd at the AG’s office.

The report by California-based consultant Ronald Nichols, who formerly worked for the U.S. Bureau of Alcohol, Tobacco, Firearms and Explosives, found a lack of care and confirmation bias in all three forensic examiners who conducted toolmark analysis. which forensic examiners use to determine whether a cartridge or grenade was fired from a particular weapon.

“If the investigations and reviews had been conducted as critically as each auditor stated in their interviews, this error should not have occurred,” Nichols wrote.

In 15 gun cases re-examined by Nichols, no further misidentifications were found by the three examiners. But Nichols disagreed with the results of two cases that the lab deemed inconclusive — including the gun in a 2021 Pawtucket case that was first reported because it matched bullet casings fired from a Glock pistol was fired that was in the possession of the Boston Police Department.

In the Pawtucket case, Karel Martinez-Scarlet was charged with the murder of Keshaudas Spence, who was a running back at Sacred Heart University in Fairfield, Connecticut, between 2011 and 2014. The case is pending in Providence Superior Court, where a trial is scheduled to begin April 28, 2025. The discrepancy with the evidence review was discovered during a routine access to the National Integrated Ballistic Identification Network (NIBIN) – the digital database that captures and compares images of ballistic evidence to help police departments determine whether there are connections between incidents.

The cases Nichols reviewed were not among the 22 gun cases that were lab-tested and still in court and sent to state crime labs in Connecticut, Massachusetts and New Hampshire for toolmark retesting. At least four re-investigations were completed and all confirmed the initial results from the Rhode Island crime lab, AG spokesman Timothy Rondeau said in an emailed statement Wednesday.

Defense attorneys are awaiting responses to motions as the RI State Crime Lab's review continues

Nichols interviewed all three inspectors involved in tool mark testing but did not include their names in his report. Hilliard also declined to disclose them. The report identifies the employees as Investigator 1, Investigator 2 and Investigator 3. The report notes that Investigators 1 and 3 have now left the laboratory and recommends that Investigator 2 be “removed from comparative analysis casework.”

In the Pawtucket case, Nichols' report said all three toolmark examiners made an error when they attributed 13 fired shell casings to a Glock pistol that could not have actually fired them.

Nichols found that there was no critical examination of the marks created by the ignition of the powder when the cartridge was fired, commonly referred to as “breech marks.”

“Based on the case notes, it appears that the 13 cartridge cases were first compared to each other and a correct identification of these cartridge cases was made,” he wrote. “If a critical examination of the scissors and at least one of the other two markings (firing pin impression or breech surface markings) was performed when comparing the evidence cartridge cases with test fires from the submitted Glock, the misidentification should never have occurred. ”

As for confirmation bias—the principle that knowing what you're looking for makes you more likely to find it—Nichols faulted Examiner 2. Examiner 1 told Nichols that the verification was not truly blind because Examiner 2 “innocently communicated during an interview that this case was an identification.”

“Simply put: In the eyes of the auditors, it must have been a false lead,” Nichols wrote.