Error Rates for Human Latent Fingerprint Examiners

Abstract

Fingerprint comparison evidence has been used in criminal courts for almost 100 years to identify defendants as perpetrators of crimes. Until very recently, this evidence has been accepted by the courts as infallibly accurate. We review four kinds of available data about the accuracy of fingerprint comparisons made by human latent fingerprint examiners: anecdotal FBI data; published data on the accuracy of consensus fingerprint comparisons made by groups of examiners working in crime laboratories; the proficiency and certification test scores of latent fingerprint examiners tested individually; and the results of controlled experiments on the accuracy of fingerprint comparisons. We conclude that anecdotal data are useless and misleading; consensus judgments of fingerprint comparisons show either indeterminant or quite large error rates; the proficiency and certification procedures in current use lack validity and cannot serve to specify the accuracy or skill level of individual fingerprint examiners; and t here is no published research evidence on error rates. It is impossible to determine from existing data whether true error rates are miniscule or substantial.