Lawyers often practice law to avoid science. But science has an enduring role to play in revealing truth at trial. And lax science in the courtroom is dangerous. It had led judges and juries to accept bogus methods as accurate, and overstated conclusions as irrefutable fact. In the criminal context, lax forensic science has led to convictions that are questionable or wrongful. This was the warning cry of a former judge and two law professors who spoke at this past fall’s Appellate Judges Educational Institute in Atlanta.
Twenty-five years after Daubert v. Merrell Dow Pharmaceuticals, Inc. instructed courts to use Federal Rule of Evidence 702 to keep unreliable expert testimony from juries, the reliability of many forensic science practices, in the panelists’ view, remains poorly understood and too infrequently examined in robust admissibility hearings.
Panelist and former Fourth Circuit Court of Appeals Judge Andre Davis described the “rocky marriage” between law and science. Both seek accurate answers to questions. But while scientists have the leisure to test and retest their hypotheses—and the freedom to say that there are no answers today—lawyers need answers in the moment to resolve disputes. Science can give the illusion of certainty in what might otherwise be a messy trial. It can coax a factfinder to go “beyond a reasonable doubt.” But science does not offer certainty, noted panelist Jessica Cino, Professor and Associate Dean at Georgia State University College of Law. No amount of rigor can reduce error rates to zero. Thus science must be used in court with care and nuance.
Further complicating the marriage, as Professor Cino noted, many forensic sciences appearing in courts, though bearing the name “science,” are not built on basic research establishing their reliability and are not frequently practiced by classic research scientists with peer review. Instead, these fields were developed for the purpose of crime solving and are often practiced by law enforcement officials. The lack of rigor Cino observed while working with forensic scientists spurred her to attend law school with the goal of serving the wrongfully convicted.
To illustrate the problem more broadly, Cino highlighted two forensic pursuits that have come under recent scrutiny: microscopic hair comparison and fingerprinting. In 2015, the Department of Justice and the FBI acknowledged that testimony regarding hair comparisons given in thousands of cases since the 1970s lacked scientific underpinnings, throwing the convictions in those cases into doubt.
Fingerprinting, which for so long seemed to offer certitude, has its own, not-fully-explored problems. Take Brandon Mayfield, who in 2004 was wrongfully pursued for involvement in a terror attack based on a fingerprint match until authorities located another suspect who also matched the fingerprint. Typically, fingerprint examiners will testify to a match if 5-15 points of comparison line up, but a fingerprint has 75-125 points that can be compared. Research has not been done to ask whether a match of 5-15 points is sufficiently reliable, or whether more comparisons should be made to decrease errors and reduce false positives.
Two seminal reports—one in 2009 by the National Academy of Science and one in 2016 by the President’s Council of Advisors on Science and Technology—discussed a wide variety of forensic science fields and trumpeted the need for tests to assess the reliability of opinions proffered by practitioners in these fields. Such tests, advised panelist Jay Koehler, Beatrice Kuhn Professor of Law at Northwestern Pritzker School of Law, should not simply assess how well practitioners perform their assigned tasks, but whether their opinions are accurate in real-world scenarios and subject to low error rates.
Professor Koehler noted the most problematic forensic science fields are those that rely on subjective opinions—for example, that two items look similar or share certain traits. Think of fields that aim to match guns with casings, teeth with bite marks, or tires with tracks. Saying that two items appear similar is one thing; saying they both come from the same source is another. To usefully testify about items sharing a source, a witness skilled in the appropriate field would need to tell the trier of fact how often items with the observed similarities are actually a match. There is no research correlating an expert’s confidence in a match with whether there really is a match. The expert’s confidence is not a proxy for reliability testing. So if an expert testifies with “100% certainty,” that is a misleading overstatement that unduly suggests the expert’s forensic field is infallible.
Other overstatements are more transparent, such as when an expert who was not properly certified as a medical examiner testified, in a case Professor Cino handled, that a bullet’s path indicated “intent.” But that testimony led to a conviction that was only vacated years later by the Mississippi Supreme Court.
Forensic evidence’s reliability, the panel believed, should be established up front at 702 hearings. If the reliability of proffered evidence cannot be established, it cannot be helpful to the trier of fact, and it should be kept out of court. Too often, Professor Cino noted, these hearing do not occur at all in criminal cases or are shorter than they should be. Leaving challenges to a field’s reliability to cross examination is fraught; by then, the expert’s bona fides and opinion have already been too-well established. Jurors can easily conjure plotlines from their favorite television crime procedurals that make forensic experts the heroic purveyors of certainties. And if unreliable testimony is allowed before a factfinder, unwinding that error is difficult. Even when a forensic science field is discredited, the post-conviction process is often too little too late.
Appellate judges, and lawyers, the panel said, have a crucial role to play in fashioning rules of law applicable to forensic evidence and in assuring that in each case, any forensic testimony is based on reliable methods, reliably applied, by reliable experts. Professor Cino warned that appellate precedent has “baked in” reliability determinations from decades ago that may be dubious, and that these assessments should not be blindly accepted, but rather revisited based on current knowledge.
The panel asked judges to ensure that 702 gatekeeping hearings were being held whenever required. Professor Cino said these hearings should be the norm. Judge Davis said courts should slow down and hold these hearings and get things right. Professor Koehler noted that more frequent 702 hearings would incentivize forensic scientists to conduct the necessary reliability testing and to show their data, which might make their evidence useful. Though 702 rulings, the panel agreed, are typically reviewed for abuse of discretion, it was suggested that legal issues regarding the rule’s interpretation would get de novo review, and that judges seeing such legal errors should act to correct them. Judge Davis asked appellate judges who felt constrained by the standard of review to nonetheless use dicta to guide the trial courts on the importance of 702 hearings.
Ultimately, judges and lawyers have to critically think about the science in their cases and work towards a system of justice that embraces science, but only so far as it goes.