The Reliability of Fingerprint Identification - A Case Report

by Andre A. Moenssens
Douglas Stripp Professor of Law
University of Missouri at Kansas City

17 January 2002

       When it comes to determining the validity or reliability of a long accepted technique of personal identification, it has become fashionable to disparage anything done by researchers of the past, and to substitute for accumulated wisdom and experience the word of individuals on the margins of a professional discipline who do a literature search on a topic in which they have some interest, and who profess to have found a state of affairs lacking in merit. In some cases, the disparaging view is largely arrived at because the critics lacked sufficient familiarity with the techniques used in a discipline and because they have never actually "done" the work; in other situations, it is spawned by a frustration of what academics see as a lack of what we in the post-Daubert world like to refer to as "empirical validation" of a technique. But the reason for the disparagement is often of secondary importance. Of crucial import in hammering home the criticism is to find that some people in the discipline under attack -- in reality a miniscule number of them -- have made a mistaken identification! 

        Never mind that these few mistaken identifications were exposed by other law enforcement experts and resulted in the removal of incompetent, negligent or fraudulent  individuals from the profession. Never mind that qualified defense expert testimony to establish such errors is readily available. Never mind that a few of the cases used to illustrate "unreliability" involved actual fraud, which could of course not be prevented by requiring proper training and experience of the examiners. All that matters is that critics attacking a discipline find one that can be attacked as non-scientific, where they can point to "mistakes" or "fraud" to taint a profession, and find a willing audience of one:  a judge deciding an admissibility-of-expert-evidence question. And to find a forum within which to litigate such issues has of course become easy in the aftermath of United States Supreme Court decisions modifying the rules under which the admissibility of expert opinion evidence is to be judged. And some critics are all too willing to argue that all expert opinion evidence that cannot be expressed with a negligible mathematical error rate -- as the DNA model supposedly does (although it really doesn't!) --  fails to meet Daubert criteria.

        After 20 or more similar attempts, in the last two years, to unseat fingerprint evidence, all unsuccessful, critics found such an audience. The case to which we refer is United States v. Carlos Ivan Llera Plaza, a district judge's ruling rendered on January 7, 2002, following a Frye-Daubert submission of documents by the parties on the admissibility of fingerprint identification evidence. The documents submitted consisted of the testimony in the case of United States v. Byron Mitchell, CR No. 96-00407, wherein Judge J. Curtis Joyner of the Eastern District of Pennsylvania ruled that fingerprint evidence was admissible under Federal Rule of Evidence 702 and met the requirements of the Daubert and Kumho Tire v. Carmichael decisions.  Judge Joyner also took judicial notice, in the Mitchell case, that human friction ridges are unique and permanent, including those visible in small friction ridge areas, and granted the government's request to exclude the testimony of three defense experts who would have testified that fingerprint comparisons "are not scientific evidence" under Daubert. 

        Judge Pollak, the capable jurist in the United States v. Llera Plaza case, who considered the Mitchell evidence and perhaps also that of one of the Mitchell defense witnesses in another case (Frye hearing testimony in People [of the State of New York] v. James Hyatt), rendered an opinion on January 7, 2002, that differed from that of Judge Joyner in the Mitchell case. Judge Pollak ruled:  (1) that qualified  fingerprint examiners could testify to the principles on which fingerprint identification rests (uniqueness and life-time permanence, principles of which the court took judicial notice), (2) to the methods used in discovering, visualizing, and comparing latent fingerprints with prints of known individuals, (3) to the similarities and dissimilarities between the prints so compared, (4) that defense experts are be permitted to rebut the prosecution evidence, but (5) experts were not going to be permitted to testify that the questioned latent and the known inked print were produced by the same digit -- testimony of a "match" that has been routinely admitted in all state and federal courts since the Illinois Supreme Court decided People v. Jennings, 252 Ill. 534, 96 N.E. 1077 (1911). 

        The evidentiary basis for the ruling was, no doubt, in part the testimony of a young Ph.D. holder, defense witness in the Mitchell case, who had transformed his doctoral dissertation into a book-size manuscript published as Suspect Identities -- a book critical of the fingerprint identification process as  used by law enforcement, and who had testified or sought to testify in several fingerprint-challenge cases.  The book is comprised mostly of historical accounts (approx. 260 pages of its 311 pages of text), related in a manner that, if anything, demonstrates the superficiality of the author's understanding of that about which he writes. Other critics in similar challenges to either fingerprints or handwriting comparisons have included law professors who have a broad generalized knowledge in one or several forensic science disciplines, but who lack practical experience and training in conducting specific examinations of a technical nature, and who seek to translate their knowledge of the law of evidence and their literature study on a specific topic into judgments, however honestly held, on the "validity" of specific forensic techniques. 

        Looking at the jurist's opinion from the viewpoint of a legal scholar having taught "scientific evidence" for some 35 years, rather than from the perspective of one who has also been involved in fingerprint research and latent print comparisons for nearly half a century, I find the Judge Pollak's opinion unwarranted by the Supreme Court's decisions that the court supposedly follows. 

        The eminent trial judge's opinion in Llera Plaza  recognizes that the Supreme Court has indicated in Daubert v. Merrell Dow Pharmaceuticals, Inc. (1993) that the test for reliability of a scientific technique was to be a flexible one, and that the "factors" enunciated in Daubert were not the only ones that might be appropriate. In fact, the Court also emphasized the same points in Kumho Tire v. Carmichael (1999), stressing, again, that the Daubert "factors" are not the sole criteria to be used and that more flexible requirements may be appropriate in some cases. Despite the court paying lip service to these principles, it then nevertheless proceeded to apply rigidly ONLY the Daubert factors to fingerprint identification. 

        We can argue as to whether the judge's application of the criteria was correct, of course, but totally apart from that consideration, it is this writer's opinion that the court clearly gave an undue weight to "factors" that pertain only marginally to the validity of  fingerprint comparisons and identifications. The opinion places an undue emphasis on one Daubert factor ("peer review") that has been denounced in the scientific and legal literature as chimerical and not indicative of reliability at all. As a member for some 20 years of the editorial board of scientific journals (among them the Journal of Forensic Sciences) and a  reviewer of articles submitted for publication, I can attest that the so-called "peer review" involved in having an article approved for publication by a panel of editorial board readers does not, in the least, establish that the premises of the "approved-for-publication" article have any validity. But aside from that, of course, the "peer review" on latent print identification has been more stringent than that which has occurred in many other disciplines in addressing reliability issues. 

        Judge Pollak took pains to discuss the methodology used in fingerprint comparisons called ACE-V [Analysis, Comparison, Evaluation, and Verification], and found the "evaluation" part not to be scientific because it involved a subjective judgment, and perhaps also because the label ACE-V to describe the comparison process had come on the fingerprint scene relatively recently. Yet, the ACE-V process in which fingerprint examiners engage, rather than being untested, is nothing more than a refinement, a clarification, and an evolution of the comparison process that examiners have been using for many decades. It's an improvement in that it bases the subjective ultimate opinion of a latent print examiner upon the following of a described scientific methodology - a protocol designed to have the subjective opinion of the examiner depend on the following of objective criteria. Judge Pollak failed to appreciate the significance of this improvement in methodology and, instead, appeared to consider it as an untested novel principle.

        The court's application of the Daubert "error rate" factor as an indispensable ingredient of all "scientific evidence" would eliminate from the courts nearly all medical, psychiatric, psychological evidence, and much crime laboratory testimony based on diagnoses and opinions rendered in unique factual settings, all of which involve subjective judgments based upon the education, training, and experience of the experts. Certainly, there is no law that removes from the purview of jury consideration expert opinions based on subjective evaluations. Justice Blackmun, author of the Daubert opinion and its "factors,"  dissenting in another case [Barefoot v. Estelle, 463 U.S. 880, 896 (1983)] said that to suggest that no psychiatrist's testimony ought to be admitted on the issue of future dangerousness - a very subjective judgment and an issue on which the American Psychiatric Association had opined its professionals could not offer a more reliable opinion than that of an ordinary lay person - "is somewhat like asking us to disinvent the wheel." I think Justice Blackmun's opinion applies equally to the admissibility of latent print identification testimony, except that with fingerprints there is a plethora of highly reliable and verified data that indeed supports the validity of the evaluation techniques utilized. 

        I was struck by the fact that the judge's opinion apparently gave more credence to the opinions of critics, including law professors, who had not actually worked with latent print techniques, and made no reference to the opinions of the many respected scientists who have authored peer reviewed -  which the recently published Suspect Identities book is NOT - works during the past 100+ years. Perhaps the judge was under the mistaken impression that scientists and skilled researchers have not investigated fingerprint identification adequately. Perhaps the court believed that this method of identification was one invented by "mere" police officers. Certainly no mention was made in the opinion of the studies of scientists like Dr. Harold Cummins, of the Tulane University medical school and a renowned anatomist who, assisted by Dr. Charles Midlo, authored many scientific papers on fingerprints and wrote respected books and articles on "dermatoglyphics," including the still relevant treatise, "Fingerprints, Palms and Soles" (1943). Ignored also by the court are "peer reviewed" publications that no one in the scientific community has ever exposed as erroneous, such as the extensive writings of foreign scientists such as Dr. Edmond Locard, Dr. Okros' book examining the heredity of papillary patterns, the American scientists Bert Wentworth and Harris H. Wilder, and countless others. Indeed, the author of the very first book on the subject, titled Finger Prints (1892), Sir Francis Galton, a nephew of Charles Darwin, was not a  "cop" either. Galton was a respected and enlightened scientist. He was not only medically trained (King's College, London, 1839), he also studied mathematics at Cambridge (1840-1843), especially the mathematics of probability with Belgian astronomer and mathematician Dr. L.A.J. Quetelet. Galton became known as a renowned anthropologist after the major emphasis of his research turned to genetics, biological variation, and human heredity. Apparently all of this scientific work done by seasoned and well credentialed researchers with academic standing, none of whose findings contradict or disprove the accuracy of current latent print identification practices, can compare to the eight-year part-time experience of a person who wrote a dissertation on a topic he only read about. 

        What I believe is the most indefensible part of the opinion is that, after agreeing expressly that fingerprints are unique and permanent - a fact of which Judge Pollak took judicial notice - and that testimony about fingerprint practices, observations, and comparisons may be presented to the fact finder, the court then holds that an expert is not permitted to state an opinion that two prints "match" or do not match because it involves a subjective judgment. The very reason why experts are called to state their opinions is to provide the fact finder with their evaluation of the meaning of factual information which the jury cannot properly understand without such expert input. The Supreme Court certainly never has required that only opinions that can be mathematically tested, or that are "guaranteed to be perfectly accurate, can support expert testimony. There is no legal prohibition in the law against subjective "opinion" evidence. This is certainly true when opposing experts are routinely permitted to testify to contrary opinions. The subjective determinations made by properly trained experts in fingerprint identification can be and have been validated, though perhaps not according to the satisfaction of theorists for whom the mathematics of probability is the only acceptable basis of opinion evidence. The Supreme Court in Daubert specifically recognized the process of cross-examination and the presentation of opposing testimony as the hallmark of proper challenges to the accuracy of scientific opinions. 

        The insistence upon mathematical precision ands calculable error rates has gained currency largely as a result of the astronomical probability calculations used in DNA examinations, which have lulled fact finders (including judges and lawyers) into a false sense of security about the purported accuracy of an "identification."  Yet, statistics calculating random match probabilities of DNA characteristics, which are often said to be as slight as one in several billion, are generally misunderstood by fact finders and tend to overwhelm the jury. Witness the duplication which has already occurred in fairly limited databases of DNA "matches" wherein as many as 8 different alleles were found to be shared by more than one individual, despite testimony that the random probability of such occurrence is calculated  to be negligible. DNA scientists predict even today that such duplication of allele characteristics in two or more individuals  is bound to occur even when 13 different alleles (loci) on the DNA strand are tested. 

        Fingerprint identification does not rest, as does DNA analysis, on class characteristics, but on individual ridge detail. The DNA of identical twins is identical according to methods of analysis now available. By contrast, it has been empirically established that the fingerprints of identical twins are different in their individual ridge characteristics (now referred to as "Level 2 detail" or Galton details). Which is the more "unique"?

        Let me suggest that the Havvard (2 "v"'s, not "w") opinion (United States v. Havvard, 117 F.Supp. 2d 848 [S.D.Ind. 2000], aff'd 260 F.3d 597 [7th Cir. 2001) represents the correct view on the admissibility of testimony about fingerprint matching evidence; it also tracks the position taken by the courts in all the other cases which have been asked to decide the Daubert issue in fingerprint evidence cases.