Wrongful Convictions and Forensic Error

Between 1989 and 2025, DNA testing exonerated more than 375 people in the United States who had been convicted of serious crimes, according to the National Registry of Exonerations. Invalid or improper forensic science was a contributing factor in approximately half of those cases. That figure understates the problem: it counts only cases where post-conviction DNA testing was possible and applied, and the forensic disciplines most implicated in the errors are not always the ones where DNA can provide a clean answer.

The Innocence Project's analysis of its own exoneration cases consistently identifies forensic science issues in roughly 44 percent of them. The disciplines most frequently cited are microscopic hair comparison (present in about one-third of hair-related convictions later overturned), bite-mark analysis, serology misinterpretation, and fire investigation based on discredited burn-pattern theories. These are not fringe techniques. All of them were admitted in court across hundreds of cases over decades, with expert witnesses testifying that their conclusions were scientifically reliable.

The pattern is international. The UK's House of Commons Science and Technology Committee published a report in 2005 that raised similar concerns about unvalidated forensic disciplines in English courts. Australia's Victorian Parliament produced a report in 2013. Canada's Goudge Inquiry reported in 2008. The problems they identified were not unique to any one legal system. They arose wherever courts admitted discipline-generated conclusions without requiring the discipline to demonstrate that those conclusions were reliable.

Microscopic hair comparison was introduced as a forensic technique in the mid-20th century and was admitted in US courts for decades. Examiners compared morphological features of hair under a microscope, such as the scale pattern, medulla structure, and pigment distribution, and testified that two samples were consistent with originating from the same person. In practice, testimony frequently went further, with examiners asserting that a match was highly significant or that a particular hair was unlikely to have come from anyone else.

In 2012, the FBI, in collaboration with the Innocence Project and the National Association of Criminal Defense Lawyers, began a systematic review of cases in which FBI examiners had provided microscopic hair testimony. By 2015, the review had examined 268 criminal cases. It found that FBI examiners had provided testimony with erroneous statements in 257 of those cases, that is, in 96 percent. Of 26 FBI examiners whose testimony was reviewed, 26 provided erroneous statements at some point. The errors were not random: they systematically overstated the significance of a hair match in ways that favoured the prosecution.

The review identified three recurring error types. First, examiners stated or implied that a hair match was a positive identification, when the technique cannot establish identity, only consistency of features. Second, examiners cited a population frequency statistic, such as claiming a match occurred in one in ten thousand people, when no validated population database existed to support such a figure. Third, examiners failed to disclose that microscopic hair comparison had a significant false-positive rate. Among the reviewed cases, 33 involved capital convictions; 14 of those defendants were executed before the review findings were released.

The FBI review led to notification of defendants in affected cases and, in some states, triggered new hearings. It did not by itself overhaul the admissibility of hair evidence: because the technique had been admitted for decades, courts were slow to treat the review findings as grounds for automatic exclusion. Reform required case-by-case litigation, often decades after the original conviction.

Bite-mark analysis rests on two foundational claims: that human dentition is unique to individuals, and that this uniqueness is reliably transferred and recorded in bite wounds on skin or other materials. Neither claim has been validated by controlled, independent studies. The American Board of Forensic Odontology imposed a moratorium on bite-mark evidence in 2016, acknowledging that the discipline lacked the scientific foundation to support the source-attribution conclusions routinely given in courts.

Among the most prominent cases in the United States is that of Keith Harward, convicted in 1982 of murder and rape in Virginia partly on the basis of bite-mark testimony linking him to the victim. He was exonerated by DNA testing in 2016 after 33 years in prison. Another frequently cited case is that of Ray Krone, convicted of murder in Arizona in 1992 based on bite-mark evidence; he was exonerated by DNA in 2002. The Innocence Project has documented over two dozen exonerations in which bite-mark evidence was a factor.

The litigation response has been uneven. In some US states, defendants have successfully excluded bite-mark evidence under Daubert by presenting the post-2000 research on its unreliability. In others, courts have held that bite-mark evidence is admissible because it has historically been admitted, a circularity that the NAS report specifically criticised. Texas created a Forensic Science Commission with authority to review past cases where specific techniques have been found unreliable. No equivalent body exists at the federal level.

England and Wales created the Criminal Cases Review Commission under the Criminal Appeal Act 1995, following the Royal Commission on Criminal Justice that itself arose partly from a series of high-profile miscarriages involving forensic evidence, including the Birmingham Six and Guildford Four cases. The CCRC reviews applications from convicted persons who believe their convictions may be unsafe, and refers cases to the Court of Appeal if it identifies a real possibility that the conviction would not stand. Between 1997 and 2023, the CCRC reviewed tens of thousands of applications and referred several hundred cases. A number of CCRC referrals have involved forensic science, including cases where pathology evidence, contact trace analysis, and fire investigation conclusions were later found to be flawed.

Canada's response was triggered by the case of William Mullins-Johnson, David Milgaard, and others wrongly convicted of serious offences partly on flawed forensic pathology. The most consequential inquiry focused on the work of Dr Charles Smith, a pediatric pathologist at the Hospital for Sick Children in Toronto, whose reports were used to support charges of child abuse and murder. Smith's reports were later found to contain serious errors, overstatements, and in some cases conclusions that no properly qualified pathologist would have reached. The Goudge Inquiry (2008) examined 45 of Smith's cases, found systemic failures in oversight, and produced 169 recommendations, including mandatory accreditation of forensic pathologists, structured peer review of pathology opinions used in criminal proceedings, and reforms to expert disclosure obligations.

India's legal system handles post-conviction challenges primarily through the revision and appeal mechanisms under the Bharatiya Nagarik Suraksha Sanhita 2023 (which replaced the Code of Criminal Procedure). There is no equivalent of the CCRC or a forensic review body with a standing mandate to examine past convictions. The Bharatiya Sakshya Adhiniyam 2023 (which replaced the Indian Evidence Act 1872) retains broad judicial discretion over expert opinion evidence under sections analogous to the former Section 45, but does not impose a structured framework for assessing foundational validity. Post-conviction DNA testing has been ordered in individual cases by the Supreme Court and High Courts, but there is no statutory framework for systematic review of convictions affected by subsequently discredited forensic techniques.

The existence of formal admissibility standards, Daubert in the United States, the reliability requirement developed by English courts following R v Atkins and codified in R v Dlugosz, and the general relevance and weight framework in other jurisdictions, did not prevent wrongful convictions from forensic error. Understanding why is important for any forensic scientist who may be called to give evidence.

The primary failure under Daubert was judicial deference to discipline-internal consensus. Daubert requires courts to assess whether a technique has been tested, whether it has a known error rate, and whether it is generally accepted within the relevant scientific community. Courts applied these factors but accepted discipline-internal peer review as a substitute for independent validation, and accepted the absence of a published error rate as consistent with reliability rather than as evidence of an unresolved gap. The result was that bite-mark evidence and hair comparison were admitted under Daubert in jurisdiction after jurisdiction, because practitioners trained in and testified for those disciplines constituted the relevant community.

A second failure was the structure of adversarial challenge. Daubert contemplates that the opposing party will contest admissibility with counter-expert evidence. In practice, defendants in the cases now identified as wrongful convictions often lacked the resources to commission independent validation studies or retain qualified experts to critique a technique. The burden of demonstrating unreliability fell on the party least able to bear it. This structural imbalance has led some reformers to argue that the burden should be reversed: that proponents of pattern-comparison techniques should be required to demonstrate validated accuracy before admission, rather than opponents being required to demonstrate unreliability to secure exclusion.

The opinion evidence framework under the Bharatiya Sakshya Adhiniyam 2023 and its predecessor statute operates through a similar structure: courts admit expert opinion if the witness is qualified by experience or study, with reliability assessed as a matter of weight rather than admissibility. Opinion Evidence and Its Limits explains the distinction between admissibility and weight. That structure creates the same vulnerability: a technique that practitioners regard as reliable, even without independent validation, will ordinarily be admitted.

The 2009 National Academy of Sciences report 'Strengthening Forensic Science in the United States' is the most cited single document in the reform of forensic evidence law. It reviewed the scientific foundations of multiple forensic disciplines and found that, with the exception of nuclear DNA analysis, no forensic discipline had been rigorously shown to have the ability to consistently produce accurate results. It called for mandatory accreditation of forensic laboratories, a national forensic science commission, standardised reporting and testimony language, and research programmes to validate the foundational claims of unvalidated disciplines. Its direct legislative impact was limited: Congress did not enact most of its recommendations. But it changed the terms of litigation by giving defendants authoritative grounds for Daubert challenges that courts could not easily dismiss.

The President's Council of Advisors on Science and Technology (PCAST) followed in 2016 with a report specifically on forensic feature-comparison methods, including latent fingerprints, bitemarks, firearm toolmarks, footwear, and hair. For each discipline, the report asked whether there were valid studies demonstrating foundational validity. It found that bite-mark analysis lacked such studies entirely and recommended that it should not be admitted. For others, it identified studies of varying quality and called for additional research. The PCAST report has been cited in court by defendants challenging pattern-comparison evidence, with mixed results.

Legislative reform at the state level in the United States has included the creation of conviction review units in prosecutors' offices, post-conviction DNA access statutes, and, in Texas and a few other states, science commissions with authority to investigate claims of unreliable forensic methods. The Texas Forensic Science Commission investigated the Cameron Todd Willingham case, in which fire investigation conclusions used to support an arson conviction have been challenged by subsequent fire science, though the case remains contentious. England and Wales updated the Criminal Procedure Rules to impose structured expert disclosure requirements, applicable to all expert witnesses including forensic scientists. See The Expert Report: Structure and Duties for the resulting obligations.