A University of British Columbia law professor is arguing that Canadian courts have leaned too heavily on three long-accepted investigative techniques: firearms identification, bloodstain pattern analysis and drug-detection dogs. Her research finds the scientific support for each is weaker than the legal system has assumed.
Sara Gordon, an associate professor at UBC's Peter A. Allard School of Law, examined how courts treat the evidence in three recently published papers. Gordon said she noticed a gap between advances in the scientific understanding of forensic evidence and the way courts handle it.
“Just because a technique has been used in court for decades doesn’t mean we should take its reliability for granted, especially when someone’s liberty is at stake in a criminal trial,” said Gordon. “My hope is that this research will help judges and lawyers better understand the limitations of these techniques and ensure that forensic evidence is presented in a way that accurately reflects what the science can—and can’t—tell us.”
Firearms evidence
Canadian courts have accepted ballistics evidence for more than a century. Gordon found that the research does not support the level of certainty with which examiners often testify. In practical testing, examiners misidentified bullets 37 percent of the time.
Her firearms paper focuses on Canada, where the evidence is still routinely admitted. In 2023, a Chicago court became the first in the United States to bar ballistics-matching evidence in a criminal case.
Bloodstain pattern analysis
Canadian courts have long treated bloodstain pattern analysis as a physical science. Gordon traced the discipline to a chemist who experimented in his basement and later helped establish its journal and early textbooks, which gave the field the outward markers of an established science. She said practitioners often cite concepts such as fluid dynamics with little or no formal training in physics.
Emerging studies point to significant error rates. In one, analysts correctly classified patterns about 70 percent of the time on rigid surfaces and 62 percent of the time on fabrics. Another found that different analysts frequently reached different conclusions on the same evidence, and additional research showed that what examiners were told about the case or scene could sway their conclusions.
Drug detection dogs
In Canada, police can use a sniffer dog to search when they have reasonable suspicion of a drug offense, a lower bar than most warrantless searches require. The Supreme Court of Canada has justified that on the grounds that such searches are minimally intrusive, narrowly targeted and highly accurate. Gordon's review of research published since the Court's 2013 rulings points the other way.
“They're not nearly as reliable as many courts seem to assume,” she said.
She cited an Australian study in which no drugs were found after nearly 75 percent of police dog alerts. Dogs can alert in the absence of drugs because of poor training, residual odors or even scents associated with some common baking ingredients. A handler can also unintentionally cue the animal, for example by slowing down or acting differently near a person he or she suspects. Searches that follow an alert can also cause fear, anxiety, anger and humiliation, Gordon noted, and may disproportionately affect racialized people.
Her proposed reforms include clearer standards for training, certification and performance monitoring of detection dogs, and applying the same standard used for other warrantless searches to sniffer dog searches.
Taken together, the papers urge judges to take a more active gatekeeping role, ensuring forensic evidence is reliable and that experts stay within what the science can support. Gordon also calls for greater scientific literacy among lawyers and judges.
The papers are published in The Alberta Law Review, The Canadian Bar Review and The Dalhousie Law Journal.