The Bite Mark Controversy: NAS and PCAST Critiques

Ray Krone is probably the most cited bite mark exoneration case. An Arizona postal worker with no criminal history, Krone was convicted of the 1991 murder of a bar employee in Phoenix. The prosecution's forensic odontologist testified that marks on the victim's neck and lip matched Krone's irregular dentition. He was convicted, sentenced to death, retried, convicted again, and eventually cleared in 2002 when DNA from saliva on the victim matched a convicted sex offender who had been living near the bar. Krone had spent ten years in prison.

Krone is not an isolated case. The Innocence Project and its affiliated organisations have tracked multiple exonerations in which bite mark testimony was a contributing factor in the conviction. The cases share a pattern: a forensic odontologist testified with high confidence that the defendant's dentition uniquely matched a bite mark; no other strong physical evidence tied the defendant to the crime; and years later DNA testing pointed to a different person.

Other cases that feature in the literature include those of Robert Lee Stinson (Wisconsin, convicted 1985, exonerated 2009), Levon Brooks and Kennedy Brewer (Mississippi, both wrongly convicted in separate cases in the 1990s and both exonerated in 2008 when DNA linked both crimes to the same man), and several others in the Innocence Project database. The pattern across cases is consistent enough that the wrongful conviction record itself became part of the argument in the NAS and PCAST analyses.

In 2009, a committee of the National Academy of Sciences published 'Strengthening Forensic Science in the United States: A Path Forward.' It had been commissioned by Congress and was explicitly designed to evaluate the scientific basis of the disciplines used in criminal courts. The report covered everything from DNA to hair microscopy to bitemark analysis, and its tone was direct.

On bite marks, the committee found that the discipline lacked sufficient scientific foundation. It identified two specific gaps: first, there was no established research base demonstrating that human dentition is unique enough to allow individualisation from a bite mark in skin; second, there was no established research base on how reliably skin registers dental features. The committee called for research to fill these gaps before bite mark evidence could be considered scientifically validated.

Although the majority of forensic odontologists are satisfied that bite marks can demonstrate sufficient detail for positive identification, no scientific studies support this assessment, and no large population studies have been conducted.

The NAS report did not call for an immediate ban. It called for research. But the research it called for had not been done by the time the next major review arrived seven years later.

In September 2016, the President's Council of Advisors on Science and Technology (PCAST) released 'Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods.' PCAST was charged with examining the scientific foundations of feature-comparison forensic methods, which it defined as methods that compare features of an evidentiary sample to features of a reference sample to reach an identification conclusion.

Bite mark analysis was examined against two criteria: foundational validity (does scientific research establish that the method works?) and applied validity (is it applied reliably in practice?). PCAST's finding on foundational validity was direct: bite mark analysis had not been established as a foundational forensic feature-comparison method. Specifically, the two black-box studies that existed at the time both showed error rates that were too high to support reliable identification conclusions.

PCAST recommended that courts should not admit bite mark evidence for the purpose of identifying a specific individual until the field had conducted proper black-box studies with acceptable error rates. This was a stronger recommendation than the NAS report's call for more research: PCAST was saying that in the interim, the evidence should not be admitted for identification purposes at all.

The critiques converge on two specific scientific problems that are not procedural shortcomings an improved protocol could fix. They are structural problems with what bite mark analysis is trying to do.

• Uniqueness of human dentition: to identify a bite mark, an analyst must assume that the pattern of features in that mark is unique to one person. But no large-scale population study has measured how often two people could produce identical or near-identical bite marks in skin. Without such data, a statement that only one person could have made a particular mark has no measurable basis.
• Skin as a registration medium: skin distorts bite marks in ways that are not predictable or consistent. The same dentition biting the same body region on different occasions can produce different patterns. The same bite mark photographed at different time intervals shows different patterns. Studies attempting to correlate bite mark patterns with known dental casts have found that examiners frequently disagree and that error rates for both false positives and false negatives are high.

These two problems interact. Even if dentition were highly unique between individuals, skin's poor fidelity would degrade that uniqueness to the point where it could not reliably be measured from the mark. And even if skin registered detail well, without a population database it would be impossible to say how rare a given set of features is. Current bite mark analysis is trying to work both problems at once with no solid empirical ground under either foot.

The legal standard for scientific evidence in US federal courts and most state courts is the Daubert framework, which requires judges to act as gatekeepers and assess whether proffered expert testimony is based on sufficient scientific validation. The NAS and PCAST reports were explicitly written to inform courts as well as scientists, and several post-2016 Daubert hearings have applied heightened scrutiny to bite mark identification testimony.

Courts in Texas, New York, and California have held hearings in which bite mark identification was contested on PCAST grounds. Some courts have excluded definitive identification conclusions while still allowing the evidence in a more limited form (for example, to establish that a bite occurred and that the suspect could not be excluded). The Texas Forensic Science Commission issued a report in 2016 recommending a moratorium on bite mark identification evidence, though the moratorium was not ultimately adopted as a binding rule.

The ABFO responded to the PCAST report by noting that it had already been revising its guidelines and that the 2018 conclusion scale revision reflected a commitment to more conservative language. The tension between the professional body's measured response and the more categorical scientific critique has not been fully resolved.

The bite mark case illustrates a mechanism by which forensic methods can enter legal practice, accumulate decades of court acceptance, and face rigorous scientific scrutiny only later, by which point convictions have already been decided on their basis. Several features of this trajectory are generalisable.

• Court acceptance is not scientific validation. A judge admitting evidence is making a legal decision about reliability, not conducting an experiment. Decades of admission does not establish that a method works.
• Examiner confidence is not a substitute for error rate data. An expert who has examined thousands of bite marks and feels certain about a match is reporting a subjective state. That is not the same as knowing the method's false positive rate on controlled test sets.
• The absence of a known error rate should be disclosed to the jury. Without it, jurors cannot calibrate how much weight to give the conclusion.
• New technology (DNA, 3D scanning) does not automatically rescue an unvalidated technique. Applying better tools to a step in the middle of an unvalidated chain does not validate the chain.

These lessons apply to hair microscopy, tool mark analysis, handwriting comparison, and every other pattern-matching discipline that entered courts through expert testimony rather than through a prior scientific validation process. Bite marks are the field's most explicit cautionary tale precisely because the exonerations made the cost of the error visible.