The federal government has established a National Commission on Forensic Science (NCFS) to "improve the practice of forensic science by developing guidance concerning the intersections between forensic science and the criminal justice system." (DOJ press release) The Commission follows a highly critical National Academy of Sciences report that asserted major problems within the forensic sciences, especially with respect to the subjective disciplines, such as fingerprint examination. NCFS has now met twice, and I was privileged to be able to attend one day of the second meeting while I was in Washington on a recent trip.
The Commission seeks a laudable goal: the improvement of the administration of justice through higher standards in the crime laboratory. The commission room was full of well-meaning, obviously intelligent leaders of the same type that one sees at all such DC-based committee meetings. The room was also filled to the brim with the type of naivete that too often inflicts such affairs. I fear that high-mindedness is not a substitute for a practical sense of how law enforcement, the crime laboratory and science interact to produce forensic results in the criminal justice system. They may end up doing more harm than good.
To start, let's take an honest look at the actual problem. NCFS members and advocates for forensic science reform often refer to the post-conviction exonerations that have been aided by the advent of DNA over the last twenty year. Over 300 such cases have resulted in exonerations (http://www.innocenceproject.org/Content/DNA_Exonerations_Nationwide.php). We can surmise that there are many more innocent individuals in prison today, because the exonerations are usually limited to cases in which DNA is available to clear someone and advocates available to champion the innocent person's cause. About half of the DNA exonerations involved the use of faulty forensic science ( http://www.innocenceproject.org/docs/FSBreakdownDiscipline.pdf). That said, the vast majority of these cases would not recur in current practice. The top two sources of error have been serology and hair evidence. Serology is the examination of blood type, which can be confounded by mixtures or misinterpretation. There is little danger of improper serology these days. In such cases, DNA analysis is always possible, so serology is merely a preliminary step to more detailed and specific analysis. Hair analysis matches the morphology of hair from a crime scene to hair from a suspect. Although morphological hair examination continues, there is no reason it should result in a wrongful conviction because it can now be checked using mitochondrial DNA analysis (or even nuclear DNA analysis on occasion). Given the legitimate concerns about hair analysis, no well-run crime laboratory in 2014 should be reporting morphological hair matches by themselves; such matches should be used as screening tools only and always lead to DNA analysis to confirm or exclude an individual.
The next two categories of forensic errors are instructive: DNA and bitemark. Bitemark evidence is in the same category as hair evidence. Except in isolated cases involving unusual impressions, bitemark evidence is so unreliable that it should not be relied upon in criminal cases. The DNA cases all involved improper testimony or analysis, i.e., analysts who performed poor analysis, reported it incorrectly or were just plain fraudulent. In other words, the top four sources of forensic errors--covering 93% of the postconviction cases involving bad forensic science--can be avoided by good laboratory management. A competent crime laboratory director knows to restrict hair analysis, serology, and bitemark evidence to very limited circumstances. And a good director is also more than capable of making sure that other analyses are performed with high standards of scientific practice and ethics.
Thus far, the NCFS has focused almost exclusively on the scientific basis of the forensic sciences, particularly with respect to fingerprint evidence and cognitive bias. Also, only a small minority of the NCFS members have ever performed a forensic analysis, so their deliberations have tended to be so high-level as to prevent them from engaging in practical analysis and leadership.
However, given the actual lessons from exonerations, they would do much more good examining other issues. They should advocate for professional standards in crime laboratories, especially among crime laboratory directors. I have known crime lab directors from across the country. I have never witnessed sustained problems in a lab run by an experienced and well-educated crime laboratory director who had solid leadership skills.
Of course, there are problems in certain practice areas. The NCFS should immediately stake a claim on the limitations of hair and bitemark evidence. That said, they should avoid poorly-based criticisms of fingerprint evidence and other disciplines. There is a clear, well-established, and scientific basis for the individuality and persistence of fingerprint patterns. In addition, research has shown that practitioners have a bias to exclusion of suspects when latent prints are ambiguous, especially in light of high-profile errors such as during the Madrid bombing investigation. Forensic scientists represent the most objective corner of the criminal justice system. Police investigators, lawyers, and judges engage in wholly unscientific and wildly biased judgments all the time. That's why false confessions, poor detective work, and prosecutor and defense misconduct are very often the heart of false convictions, including in most cases in which forensic science may have been improperly applied. To be sure, that makes it all the more important for us to ensure the reliability of forensic results. Nonetheless, I argue that critics may contribute to future false convictions by undermining the confidence of the criminal justice community in forensic results. Further, I maintain that trial judges are often loath to exclude what may be questionable forensic results because they are faced with much shakier eyewitness or other evidence every day. In comparison to that reality, fingerprints are indeed the Holy Grail of reliability.
The NCFS should engage with police leaders and state and local governments to address these broader issues. In Houston, Detroit and elsewhere, police departmetns and local governments failed to prioritize the funding of crime laboratories. In many jurisdictions to this day, crime laboratories lack training and infrastructure. Pay is also an issue. Personally, I have known many, idealistic young people who chose to study forensic science in college only to find that their income in the crime laboratory did not match their CSI dreams and was a small fraction of what they could make in private-sector chemical or biomedical laboratories.
In the real world thus far, concerns about reliability have led to much more documentation and paperwork. The NCFS appears headed in the direction of mandating even more elaborate protocols and paperwork, though there is scant evidence that the "solutions" being discussed will actually result in improved reliability of results. Lab practitioners know this and live it every day, and they are increasingly cynical about federal mandates in which they have little voice and for which there is no funding. The NCFS could advocate for additional federal funding to back up their mandates. Forgive me my own cynicism in this regard. The likely result is a few more Coverdell dollars (or none) and state and local labs having to implement difficult mandates within their already constrained resources. Again, we may be faced with the unintended consequences that follow: crime labs that are stressed to deal with higher backlogs and greater delays that inevitably lead to sloppier work and reliability problems.
Of course, there have been some excellent studies in recent years on cognitive bias in forensic practice. I am a particular fan of Itiel Dror's work and recommendations. (See, e.g., The forensic confirmation bias: Problems, perspectives, and proposed solutions.) The NIST scientific advisory committees would be well-served by a close examination of that work, followed by practical recommendations that implement the most workable ideas without adding to the time and expense of forensic identification. On such matters, it is sufficient for the NCFS to urge on the work, not substitute its own biases.
In fact, that is exactly my point. If forensic scientists are susceptible to bias, then so too are the members of the NCFS. They come with points of view that are not necessarily aligned with absolute truth and justice. In this, we may hope that they are more self-aware than other federal advisory committees, because they know that bias is part of any human-centered business. With such humility, they may yet end up doing some good.
For more information about John Morgan and his consulting business, see www.coptechllc.org.