Once Bitten, Courts Aren’t Shy About Using Bitemark Testimony Again and Again

Uncertainties in Bitemark Comparison

Shortly after the Marx case, bitemark comparison became famous due to its role in the Ted Bundy serial killer case and 1979 conviction. Bundy had bad teeth—“unusual dentition” in the parlance of forensic odontologists. Four separate odontologists agreed that Bundy could be associated with a bitemark on a deceased victim. Those four examiners went on to become leaders in the field and each worked hundreds of cases. Two became president of the American Academy of Forensic Sciences (AAFS). And all four eventually contributed to wrongful convictions based on their bitemark testimony in other cases.

It’s a remarkable record. No other AAFS president has ever contributed to a known wrongful conviction. Overall, there have been at least 27 wrongful convictions associated with bitemark comparison errors. The total number of cases involving bitemark comparison and the actual error rate associated with the discipline are completely unknown. It is clear that the discipline is based on assumptions with weak scientific foundations and uncertainties that undermine its reliability.

Even at the time of the Marx case, the uncertainties in bitemark comparison were well known among forensic odontologists, many of whom opposed its use in criminal trials. There is limited data about the uniqueness of the features of human dentition. Human skin is a poor registration medium for dentition. The flesh moves and distorts under the pressure of a bite. The pattern of a bite may be due more to the healing process in living victims or postmortem changes than the direct effects of the original bite. The collection process may introduce its own set of distortions. None of this is made easier by the messy facts associated with real crimes in which an assailant will not take the time to leave easily identifiable impressions on a victim.

Due to these concerns, many scientific review panels have called for the limitation or exclusion of bitemark evidence from criminal trials. In its 2009 report on forensic science, the National Academies of Sciences said, “The ability of the dentition, if unique, to transfer a unique pattern to human skin and the ability of the skin to maintain that uniqueness has not been scientifically established. . . . [I]t is reasonable to assume that the process can sometimes reliably exclude suspects.” The 2016 report of the President’s Council of Advisors on Science and Technology (PCAST) said that “bitemark analysis does not meet the scientific standards for foundational validity and is far from meeting such standards. To the contrary, available scientific evidence strongly suggests that examiners cannot consistently agree on whether an injury is a human bitemark and cannot identify the source of bitemark with reasonable accuracy.”

In 2016 in response to the Steven Chaney wrongful conviction and mounting research evidence concerning the unreliability of bitemark comparison, the Texas Forensic Science Commission used its authority to recommend that “bitemark comparison not be admitted in criminal cases in Texas.” The courts in Texas have deferred to the commission in this regard, and Texas remains the only jurisdiction in the United States that bars the admission of bitemark evidence.

Even the ABFO has recognized the issues. They revised the standards for evaluating bitemarks in 2013, 2016, and 2018. The current standard allows for only three types of conclusions that would relate a biter to a mark on human skin: Excluded, Not Excluded, and Inconclusive. This has turned out to be a judicious change. Appeals courts have recognized the ABFO as an authority to support that bitemark comparison has general acceptance under the Frye or Daubert standards. For example, the U.S. Court of Appeals for the Sixth Circuit issued the Hill opinion in 2023 citing the new ABFO guidelines extensively, but it declined to overturn the conviction, which had been based on a match conclusion by the original examiner. They stated:

We note that if we came out Hill’s way, most convictions involving forensic evidence would never be final. The end of Hill’s logic is that a petitioner could raise a new due-process claim, unmoored from the gatekeeping provisions, anytime a relevant scientific standard changed.

Scientific Basis for Bitemark Comparison

Even the dissent in Hill, which took great exception to this extraordinary logic, cited the ABFO as the “relevant scientific community” to determine the scientific basis for bitemark comparison testimony. The courts may have erred in this regard because the ABFO has significant limitations as an objective source of scientific consensus on bitemark comparison. First and foremost, it is run by forensic odontologists, many of whom perform bitemark comparison and have a vested interest in the unimpeded continuation of the discipline. They are a classic example of “regulatory capture,” the tendency of oversight bodies to be influenced or even operated by those who are subject to their authority. That is not to demean the role of the ABFO, which provides the only substantive governance in the field. Nonetheless, the ABFO has a clear conflict of interest when it comes to judgments about the scientific basis for bitemark comparison.

This is a problem recognized in the Frye decision itself, where the D.C. Circuit pointed out that general acceptance must be gained “in the particular field in which it belongs.” Pointing out that the blood pressure deception test at issue had not gained “such standing and scientific recognition among physiological and psychological authorities as would justify” its admission as evidence (as opposed to the sole testimony of the person who had invented the test), they excluded the novel test.

Recent history demonstrates the importance of scientific independence when assessing general acceptance. The National Institute of Standards and Technology (NIST) is an independent scientific authority. That is why NIST now operates the Organization of Scientific Area Committees (OSAC), the primary mechanism for the development of draft consensus standards in the forensic sciences. The courts would do well to defer to NIST and OSAC guidance concerning scientific issues such as the validity of bitemark comparison. The OSAC Forensic Odontology Subcommittee states that the foundational research into the “reliability and credibility” of bitemark comparison has not been conducted and constitutes a “major gap in current knowledge.” Separately, NIST conducted a review of the scientific foundation of bitemark comparison. That report concluded that there was a lack of empirical support for three key assumptions of bitemark comparison:

1. Human dentition is unique at the individual level.
2. This uniqueness can be accurately transferred to human skin.
3. Identifying characteristics can be accurately captured and interpreted by analysis techniques.

The American Academy of Forensic Sciences (AAFS)—the primary professional association for the field in the United States—engaged its Odontology Section of the Academy to respond to the NIST report as “the subject matter experts of the discipline.” The Odontology Section collaborated with the American Society of Forensic Odontology (ASFO) and the ABFO, while stating that the resulting “response is not a position taken by the AAFS or the Forensic Odontology Section.” The three entities have substantial overlap. Most of the current ASFO leaders are certified by the ABFO, as is the current chair of the Odontology Section. The result may reasonably be called the “practitioner response” of the forensic odontology community.

The practitioner response agrees that “the accuracy of skin pattern transfer remains uncertain” but holds that there can be sufficient information in bitemark patterns to support exclusions. They analogize to the situation of partial fingerprints or tire marks. But there is no set threshold for what constitutes an acceptable “partial” bitemark to support an exclusion. It is possible that distortions prevent reliable exclusions, and the available research studies support that view. Further, fingerprint examiners are not permitted to make class conclusions on the basis of partial prints, though they may produce exclusions on the basis of prints that are of lower quality than those used for source identification conclusions.

The practitioner response also noted that practitioners show substantial variability in their interpretation of presumed bitemarks on skin. Although the practitioner response acknowledged that this is a “major issue,” they attribute the problem to poor research methods and “no published criteria for minimum quality of photographic evidence to confirm their evidentiary value.”

The “Evidentiary Value” of Bitemarks

This last issue was not highlighted by the NIST report but is very important. There is no clear standard to determine “evidentiary value” of bitemarks. In contrast, latent print examiners have produced extensive guidance concerning the issue of “suitability,” i.e., whether a particular partial print is suitable for comparison, database searching, or exclusion. ASFO and ABFO training does provide some guidance. For example, bites that don’t break the skin may produce only bruising that lacks sufficient detail for analysis. Also, extreme bites that tear flesh will not leave behind useful features for analysis. Bites that penetrate the skin but don’t disturb the underlying flesh are considered the most useful for bitemark comparison. Nonetheless, examiners may use bitemarks with questionable utility. In many wrongful convictions, examiners used bitemarks that had been distorted by postmortem degradation. When examining features that are known to lack persistence like bitemarks, it is critical to have strict criteria for forensic utility. The ABFO recognizes this—as clearly stated in the practitioner response to NIST—but its criteria for determining evidentiary value emphasizes the quality of photographs. Additional criteria relate whether one of the top (maxillary) or bottom (mandibular) arches can be found and matched to the general size and shape of human dentition and whether “some marks caused by individual teeth” can be recognized.

This guidance is grossly insufficient. Examiner variability can extend to the assessment of whether particular injuries are bitemarks and certainly includes determinations of evidentiary value. Examiner variability has arisen in many wrongful convictions, including widely varying interpretations by examiners even when the suspect’s dentition was just as “unusual” as Ted Bundy’s. In some cases, examiners disagreed concerning whether particular patterns on the skin were bitemarks. Presumably, some of these patterns looked like the arch from a human jaw and showed some characteristics that looked like tooth marks to trained examiners. It has become routine to dismiss the experience of wrongful convictions among odontologists because the original convictions were from 2006 (the Crystal Weimer case, which involved an uncertified examiner who did not follow ABFO standards) or before. Nonetheless, wrongful convictions are an important source of data for a field that lacks sufficient research basis, and it is a source of data that has arguably been more influential than any other set of information for improvement of the field. It is a mistake to assume that all the issues related to wrongful convictions have been addressed by the ABFO—clearly, issues related to suitability and evidentiary value remain unresolved. Latent print examiners (and automated fingerprint identification system programmers) have been able to develop models for suitability based on types of distortion and examiner opinions. That is because the examiners and distortions are consistent enough to be reliably deconvolved by humans or algorithms. Clearly, bitemark distortions and expert judgments about those distortions are not sufficiently consistent to develop such models or incorporate appropriate guidance into the ABFO standards. From a scientific perspective, this problem undermines the essential basis of bitemark comparison, even if the courts may never recognize it.

It also has become routine for bitemark examiners to say that reliable research is unavailable due to problems with experimental methods and human-subjects’ approval of any study based on actual bitemarks on living humans. The practitioner response to the NIST report makes this point repeatedly. The objection has limited validity. Research has been done on animals and human cadavers, and it is not favorable to the use of bitemark comparison for any purposes. These research studies may or may not have reflected the ABFO photography standards, but it is not credible to dismiss them on that basis. Bitemarks on cadavers in research studies should be easier to assess than bitemarks on living victims in casework, but examiners in those research studies still showed poor reliability. The most important empirical research was conducted by Mary Bush at the School of Dental Medicine under the State University of New York at Buffalo. Her studies on the subject were highlighted in the NIST report. Bush recently prepared a summary of her research for the Journal of the California Dental Association, including example photographs that appear to be clear and in conformance with ABFO standards. The results: “Bitemark transfer to skin is not reliable. Significant distortion occurs, and the distortion is not predictable.” Bush reinforces that cadaver studies present easier problems than bitemarks from criminal casework, in part because cadavers don’t bruise and living victims will. Thus, the two main sources of information—wrongful convictions and empirical research—both suggest that bitemark comparison is unreliable even for exclusions.

Moral Panic and Claims of Forensic Scientist Cognitive Bias

Why do examiners argue so forcefully against the inevitable conclusions about that information reached by NAS, PCAST, the Texas Forensic Science Commission, and NIST? As Bush relates, her work was acclaimed by forensic odontologists until it was cited by defense lawyers in court, after which she states that she and her colleagues “were openly criticized, bullied, [and] attacked verbally” by pro-bitemark advocates. In recent years, many researchers have raised issues about cognitive bias among forensic scientists, and the experience of the ABFO and bitemark comparison examiners may be understood to some extent in that framework. As Upton Sinclair said, “It is difficult to get a man to understand something when his salary depends upon his not understanding it.” There may also be a moral panic issue. Cases involving presumed bitemarks often include sexual assault, violence against children, and other heinous crimes. By its nature, bitemark comparison exposes the examiner to terrible images of human cruelty and violence. At the workshop convened to support the preparation of the NIST report, the bitemark advocates showed numerous slides that reflected the tragic victimization associated with presumed bitemarks.

Similar biases may go some way to elucidating the reasons why the courts have been so reluctant to block the admission of bitemark testimony. For example, in the Hill case decided by the federal appeals court last year, the bitemark comparison was a critical element to determine whether the defendant was a bystander to the torture, rape, and murder of a child or an active participant. Like many other courts, they could have taken the view that there was “changed science” that fundamentally altered the basis for the admissibility of the forensic science testimony, as in the Eddie Lee Howard case. They could also have recognized that even the ABFO guidelines would deem the original match testimony to be plain error, not just out of date. It cannot have been lost on them that such a judgment could have resulted in the false exoneration of a violent psychopath. One can hardly fault the Hill court if it might have been motivated by a sense of moral panic, but the assessment of scientific reliability cannot be based on such sentiments.

Precedence for the Admissibility of Scientific Evidence

Of course, the courts also have a conscious bias in favor of precedent. Although scientists must certainly work within the paradigms of their field, they are constantly updating their state of knowledge. Scientific journals are full of review papers that attempt to synthesize research in specific fields and update views based on new empirical findings. The courts have no similar process. The 1975 Marx decision that kicked off the era of bitemark comparison was more limited than what followed; the Marx court recognized that bitemarks often lacked clear detail and that the specific case provided an unusually clear injury for interpretation. It has nonetheless been used as a precedent for 50 years for a much broader use of bitemarks than the Marx court intended. With respect to bitemarks and almost any other forensic method, the courts default to admissibility, meaning it is usually up to trial judges or lay juries to weigh the value of the forensic evidence. Judges are idiosyncratic to say the least about the source of scientific knowledge they use to inform decision-making. Juries must rely on gut judgments about the narrative coherence and seeming trustworthiness of the experts they see in the courtroom. It is likely not an accident that the four examiners in the Bundy case—and many bitemark examiners—have displayed remarkable communication skills and charisma. In short, the history of bitemark comparison does not reflect kindly on the ability of the judiciary to be reliable gatekeepers of scientific evidence.

Bitemark Determinations Used to Guide DNA Sampling

The issues are only becoming more complex. Many forensic laboratories now rely on bitemark determinations solely to guide DNA sampling. A presumed bitemark may be an excellent place to find biological material from an assailant, and modern methods are able to amplify small quantities of DNA from such locations. That said, this practice inherently implicates activity-level propositions that have questionable reliability. In other words, if a suspect’s DNA is found at that location, then the analyst or a bitemark examiner is likely to testify that the DNA got there because the suspect bit the victim. This kind of testimony has weak foundations with or without the presence of a presumed bitemark because the scientific foundations have not been adequately established to support most activity-level propositions. Nonetheless, the courts have confused the high reliability of DNA analysis to characterize the source of a biological stain with the uncertain reliability of these activity-level propositions. The courts are also now set to allow questionable determinations concerning presumptive bitemarks without any gatekeeping about the scope of activity-level testimony based on those determinations. There are no standards by the ABFO, NIST, or anyone else to guide the courts on the matter. Hopefully, the courts will resolve to “do better” on these matters going forward.