Case Study 16.2 — Roy Brown: The Bite Mark That Excluded Its Own Defendant

Sourcing and tone. This case study draws on the public record of a U.S. case (New York; conviction 1992; exoneration 2007) and is used to teach a single, devastating methodological point: a bite-mark "identification" not only pointed at an innocent man, but the very evidence used to convict him appears to have contained, on its face, a difference that — in any disciplined comparison — should have driven toward exclusion. We treat a real murder soberly and stay within documented, public facts. The defendant was exonerated by DNA; the case is presented to show how a comparison discipline can fail by ignoring an exclusion, not to relitigate the underlying tragedy.

Background

In May 1991, a woman named Sabina Kulakowski was murdered in upstate New York; the attack included multiple bite wounds on the body, and the scene involved a fire. Roy Brown, a local man, was charged with and convicted of the murder in 1992. He maintained his innocence and would spend roughly fifteen years in prison before DNA evidence secured his exoneration in 2007.

As in the Ray Krone case (Case Study 16.1), bite-mark testimony was central to the prosecution. An odontologist linked the bite injuries on the victim to Brown. But this case carries an extraordinary detail that makes it, in this book's terms, even more instructive than Krone's — because the bite-mark evidence here did not merely fail to identify the right man; by one account it appears to have positively pointed away from the man it was used to convict.

The forensic evidence — and the difference inside it

The bite-mark comparison was offered to link the marks to Brown. Yet by accounts in the public record, the bite injuries reportedly displayed features inconsistent with Brown's dentition — including, by one account, more upper teeth than Brown actually possessed (he was missing teeth). In a disciplined comparison of any kind, that is not an ambiguous datum. It is a difference, and an unexplained difference is the single most powerful thing impression comparison can produce: grounds for exclusion (§16.1).

Recall the symmetry the chapter draws with DNA (§16.1, §16.5): agreement only associates, but a genuine, unexplained difference can exclude. The whole of honest impression comparison is built on that asymmetry. A feature in the questioned mark that the known source could not have produced means the known source did not produce it — full stop. By the logic the chapter teaches, a bite mark showing more upper teeth than the defendant has is, on its face, an exclusion of that defendant.

Instead, the trial testimony minimized the discrepancy rather than treating it as the exclusion it arguably was. The difference was explained away to preserve a match the prosecution wanted. This is the chapter's nightmare scenario stated plainly (§16.6): a method so unmoored from the exclusion-over-proof logic of forensic science that an apparent mismatch was spun into a match.

The inversion, named. §16.1 establishes a strict hierarchy of conclusions, with exclusion as the cleanest and most defensible rung. Bite-mark analysis as practiced in the Brown case did not merely overstate the top rung (claiming individual identification it could not support); it inverted the bottom — taking evidence that pointed toward exclusion and reading it as inculpatory. A discipline that can convert a difference into a match has abandoned the one thing comparison evidence does best.

What the evidence did — and didn't — establish

It is worth being exact about what failed here, because the failure is layered.

1. The method had no validated basis to identify the biter at all (§16.5). On skin — elastic, mobile, changing over time — and with no measured error rate, bite-mark identification could not reliably say whose teeth made the marks even in the best case.

2. Worse, the specific evidence appears to have contained an exclusion that was overridden. This is the additional, aggravating failure. Even a court inclined to admit bite-mark testimony should have confronted a mark that, by one account, showed more upper teeth than the defendant had. The honest reading of that feature is "not this man."

So the bite-mark evidence established, reliably, nothing about Brown's guilt — and on its face appears to have contained a reason to doubt it.

Brown did not wait passively for the system to correct itself. From prison, he pursued another suspect through public-records requests, building a case that pointed elsewhere. After his release was secured, DNA testing — of saliva on the victim's clothing, by accounts in the record — excluded Brown and implicated the other man he had identified. In 2007, Roy Brown was exonerated.

Connect this to the chapter's ⚖️ In the Courtroom beat in §16.6. Roy Brown's case is "the courtroom embodiment of §16.1's hierarchy of conclusions." A genuine, unexplained difference should drive toward exclusion; when an examiner instead explains it away to preserve a wanted match, the discipline has inverted its own logic. The first question on cross in any comparison case is therefore not "what agrees?" but "what differs, and why isn't that an exclusion?"

The lesson

Three lessons, all central to this chapter:

1. Exclusion is the surest forensic voice — and silencing it is the gravest error. Bite-mark identification at the discredited end of the spectrum is bad enough; overriding an apparent exclusion is worse, because it discards forensic science's most defensible conclusion in favor of its least defensible one. Brown's case shows the tragedy of an exclusion ignored: a difference was available at trial, and it was overridden (§16.6, Theme 1).

2. The first cross-examination question is "what differs?" A competent attack on comparison testimony (Chapter 30) does not begin by asking what features agree; it asks what features differ and why each difference is not an exclusion. Understanding the hierarchy of conclusions tells you exactly where the discipline is most likely to have cheated — at the place where an inconvenient difference was minimized.

3. Bite marks deliver the worst of both worlds. As §16.5's Cognitive-Bias Watch puts it, bite-mark analysis is a cognitive-bias case study layered on top of a validity failure: a weak, unvalidated method plus a procedure (shown the suspect, asked to confirm) that invites the examiner's expectation to do the matching. Brown's case adds the final turn — the expectation was strong enough to read a mismatch as a match. That combination — invalid method, biased procedure, exclusion overridden — is precisely what convicts the innocent.

Discussion questions

1. By one account the bite marks showed more upper teeth than Brown possessed. Using §16.1's exclude-vs-associate asymmetry, explain why this single feature should, in a disciplined comparison, carry more weight than any number of "points of agreement" the examiner also reported.

2. The trial testimony "minimized the discrepancy rather than treating it as the exclusion it arguably was." Name the cognitive mechanism (§16.5, previewing Chapter 31) that makes an examiner inclined to explain away a difference once they "know" the answer the case wants.

3. Compare the bite-mark failures in Brown and Krone (Case Study 16.1). Krone teaches "confidence is not validity"; what additional failure mode does Brown add that Krone does not? Why is overriding an exclusion arguably worse than a confident false identification?

4. Brown, from prison, identified the true perpetrator through public-records requests, and later DNA confirmed it. What does this say about how the wrongful conviction was corrected — and about how rarely the forensic system corrects such errors on its own (preview Chapter 34)?

5. Place the two methods that decided Brown's fate on the validity spectrum (§16.1, §16.5): the bite-mark "identification" at trial and the DNA at exoneration. State the single question (the NAS/PCAST foundational-validity question) that separates them, and what the honest answer was for each.

6. Ethics and reform tie-in (Chapter 31, previewed). Suppose lab policy had required the odontologist to record and characterize the questioned bite mark — including any teeth count — before ever seeing Brown's dental models. Explain how that sequence (context management / blind analysis) might have forced the difference to be confronted rather than minimized.