The Validity of Latent Fingerprint Identification: Confessions of a Fingerprinting Moderate

Abstract:     
The validity of latent fingerprinting identification as a science has been receiving growing scholarly attention. The courts, however, have almost uniformly not only held that fingerprint identification withstands scrutiny under Daubert, but have largely ignored or sidestepped the scholarly arguments on the other side. In this essay, I argue that fingerprint evidence ought to fail Daubert, but not for the reasons that many might think.

This essay is written in response to a forthcoming article in Law, Probability and Risk, which argues that latent fingerprint identification has not only never been subjected to genuinely scientific testing, but that at present, such testing would literally be impossible, given the state of latent fingerprint identification as a technique. I agree with the claims this article makes about the lack of testing (and the impossibility of certain kinds of testing, given our present state of knowledge), but challenge the conclusion that these failures ought necessarily to mean it fails Daubert. Rather, latent fingerprint identification fails Daubert for the failure to do various other, more modest, kinds of testing - especially proficiency testing of examiners; research into biases and observer effects; and scaling back claims about the certainty of identifications. Daubert, I suggest, should be understood not to require the impossible, but instead, should merely require that evidence of validity that can reasonably be produced under the circumstances. Unfortunately, even under this more modest and moderate understanding of what ought to be required, fingerprinting still fails.

Keywords: fingerprint identification, evidence, expert evidence, forensic evidence, Daubert, admissibility, criminal

Accepted Paper Series