20 Sep Department of Justice Directs Forensic Trial Witnesses to Stop Claiming “Scientific Certainty.”
Last week the U.S. Department of Justice issued directives to all DOJ agencies regarding accurate analysis and presentation of forensic evidence. The directives can be found here. Directive Number 1 provides that agencies must now “ensure that forensic examiners are not using the expressions ‘reasonable scientific certainty’ in their reports or testimony.” Yes, you read that right. The Department of Justice is telling its forensic expert witnesses to stop claiming “scientific certainty.” Why? Because for most forensic disciplines, there never was any, and DOJ is—after decades of resistance—admitting it.
For many years, real scientists have protested, in real scientific publications, that most “forensic science” is not scientific at all. But their protests have gone unheeded, and our juries are primed instead by endless depictions of infallible forensic precision in the movies and t.v., what’s often called the “CSI Effect.” The National Academy of Sciences, which issued a scathing report a few years ago about the appalling absence of real science in much forensic analysis—have long known that. But law enforcement agencies, with the federal DOJ front and center, have historically been notoriously obstructionist when it comes to putting the science into forensic science, so the new directives are a welcome step.
Here’s one quick example of the importance of Directive Number 1: courtroom testimony from fingerprint examiners. For most of the last century, prosecutors have called fingerprint examiners into court to testify that two prints, or sets of prints, were made by the same individual. They do not provide a statistical likelihood; they assert complete, 100% certainty. What is that based on? Science, right? Not at all. It turns out that fingerprint matching is nothing like DNA identification. With DNA, the analyst looks at specific locations on the chromosome. Those locations were chosen because they exhibit a known degree of variation among the population, as determined by analysis of millions of samples. Thousands of scientists around the world research human DNA and publish their results in peer-reviewed journals. The scientific community’s knowledge of the genome and its permutations is public and ever-improving. A professional DNA examiner giving forensic identification testimony in court will be able to explain how the test was done, what locations on the chromosome were examined, why those locations were chosen, how many matches were found, and what the statistical likelihood is that those matches would be found if the samples came from different people.
It turns out that none of that is true for fingerprints. Fingerprint examiners look for “matching points” in prints, but believe it or not, there are no general standards for which points to look at, how many points to look at, or even what counts as a “point.” Not only are there no established standards, there isn’t even general agreement within the forensic analysis community. Some people like eight points, others ten, others twelve. Many examiners insist they can make an identification with just a single point.
Even more amazingly, in stark contrast to DNA matching, no one knows what the statistical likelihood is of two fingerprints sharing particular points, or whether that likelihood is different for different regions or features of the print. This is the crucial question for any identification methodology, because while each person’s fingerprints may be unique, the examiner doesn’t look at every molecule—the examiner looks at whatever five (or eight, or ten) “points” he or she chooses to look at. To make such a procedure scientifically valid, we need an answer to this basic question: What is the statistical likelihood that two different individuals will share (say) five particular ‘points’ in a given print? What is the answer to that question? No one knows. Yet print examiners routinely take the stand and testify that they are 100% confident, to a scientific certainty, that the same individual made both prints.
Why does no one know? It would actually be fairly easy to establish good statistical data on the relative frequency of various physical features of fingerprints. But the big fingerprint databases are controlled by DOJ, and DOJ has steadfastly refused to let researchers use them for the types of analyses geneticists do with DNA. That’s what makes print analysis so frustrating: the data exists, so fingerprint analysis could be a genuine scientific discipline, with publicly-available databases, peer-reviewed research, known error rates, and accepted methodologies. It could be a real body of knowledge about the differential rates of occurrence among populations of particular physical features of our fingerprints. Hopefully one day it will be. But it’s not now, as the DOJ directives finally acknowledge.
At Brown White & Osborn LLP, we deal with scientific evidence in many of our cases, and we know to use—and attack—forensic evidence. We’ve seen firsthand how good science can advance the cause of justice and how junk science can pervert it. We hope forensic science one day lives up to its name. Until then, it’s a step in the right direction for the legal process, at least, for DOJ examiners to stop claiming “scientific certainty.”
Latest posts by Caleb Mason (see all)
- Reflections From The Past Year of Criminal Defense Practice:Lesson Two - January 12, 2017
- Reflections From The Past Year of Criminal Defense Practice: Lesson One - January 5, 2017
- Department of Justice Directs Forensic Trial Witnesses to Stop Claiming “Scientific Certainty.” - September 20, 2016