April 2007
"Expert Witnesses, Adversarial Bias, and the (Partial) Failure of the Daubert Revolution"
In Daubert v. Merrell Dow Pharmaceuticals (1993), the Supreme Court ruled that expert testimony is only admissible in court if it passes a strict reliability test, and assigned the role of evidentiary "gatekeepers" to federal trial judges. This standard, later codified as Rule 702, has undoubtedly provided significant protection against the worst abuses of junk science since its inception. But has it created a better overall environment for sound scientific evidence? Are courts misusing the rule to bar legitimate scientific evidence? Do judges administer Daubert standards effectively? Are there lingering problems caused by experts being chosen and paid by the parties to the case? What are the future opportunities for reforming the use of scientific expert testimony in adversarial litigation? At an April 23 AEI event, David Bernstein presented a paper concerning these questions and discussed it with a distinguished panel.
David E. Bernstein
George Mason University School of Law
Until Daubert was decided in 1993, expert testimony was treated as a free-for-all in American courts. The legal environment created by Daubert, General Electric Co. v. Joiner (1997), and Kumho Tire v. Carmichael (1999) has undoubtedly been better than the relative chaos that preceded it, but it is far from ideal. The seldom-explored rationale behind Daubert was to root out and correct for the biases in expert testimony that arise in our adversarial legal system. Those biases include conscious bias (opinions for hire), unconscious bias (wanting to help the side paying the bills), and selection bias (choosing unrepresentative witnesses in the first place). Daubert and Rule 702 try to limit the negative effects of these biases, but they do so in a way that creates its own problems. First, in the criminal context, the prosecution is able to use forensic evidence of low quality because public defenders rarely have the resources or wherewithal to even raise a Daubert motion. This might be remedied by severing state-run forensic labs from the prosecution and opening them up to greater competition. Second, in toxic tort cases, Daubert has done well by excluding the worst "junk science." But there are legitimate questions as to whether too much is being excluded, and as a result some judges are hesitant in applying Daubert or are ignoring it altogether, limiting its effectiveness. Finally, applying Daubert's exacting standards to connoisseur evidence keeps jurors from hearing some testimony that would be useful to them. Using court-appointed experts, or panels of experts, is a positive step courts should take toward limiting bias from these and other experts.
Edward K. Cheng
Brooklyn Law School
Expert testimony is not the only type of evidence that can be unreliable. Much about the Daubert standard is hard to justify compared to alternatives, including the Frye standard that was in place prior to it. Specifically, Daubert focuses on reliability and admissibility of particular bits of evidence. Focusing on the sufficiency of all the evidence presented in terms of the overall burden of proof makes more sense, however. By taking the approach it did and making judges the gatekeepers of evidence, Daubert has put judges in a position for which they are ill-prepared. Bernstein seems conflicted about what judges should be doing with connoisseur testimony--on the one hand, he seems to be saying that rigorous and systemic epidemiological studies should be the sort of thing jurors ultimately rely on, but on the other hand, he is rightly reluctant to prevent jurors from using other potentially useful forms of evidence. The logical conclusion to draw from this is that such systemic evidence should be given a privileged place, but a mosaic of other sorts of evidence should also be presented to jurors, especially where such evidence simply does not exist.
Deborah Runkle
American Association for the Advancement of Science (AAAS)
Several years ago, the AAAS launched a program called Court Appointed Scientific Experts (CASE), which maintains a registry of qualified scientists to work as neutral experts on behalf of courts and facilitates such arrangements. Of the types of biases Bernstein outlined, selection bias is the most important and the one most difficult to correct for within an adversarial framework. But with Rule 706, courts already have the authority to go beyond the confines of that system to appoint experts, and they should take advantage of this more often than they do. When the court retains experts, they can present a fuller picture of evidence to the judge or jurors, rather than the fractured and often incoherent picture that emerges from the testimony of dueling experts. Although there are challenges that arise when scientists are brought into an often hostile legal system, the results of CASE have generally been positive, suggesting that greater reliance on neutral scientific expertise in the future could improve the results our courts produce.
David Michaels
George Washington University
Most people thinking about Daubert or issues of expert scientific testimony assume that the distinction between "junk science" and "good science" is easy, but in practice it is incredibly difficult. The most predictive factor for whether a particular theory is treated as one or the other is the amount of money that has been invested in proving it. As a result, indigent defendants are treated unfairly and wealthy parties to litigation can sway things through expert testimony more than they should be allowed to. But, as Churchill said about democracy, the adversarial system may be the worst legal system we know of except for all the others that have been tried from time to time. Selecting independent experts to avoid adversarial bias sounds wonderful in theory, but in practice it is difficult and can be costly. In addition, an expert's supposed independence from the parties to the litigation does not mean that the person will not have some professional axe to grind that will lead to bias. It may be more practical to have judges consult experts in order to make more informed decisions on Daubert motions, and leave it at that. It is also worth pointing out that there is nothing particularly special about the procedures leading to a statistically significant finding from epidemiology that should lead courts to treat them as qualitatively distinct from other forms of evidence.
Joe G. Hollingsworth
Spriggs & Hollingsworth
The admissibility standard put in place by Daubert is far from arbitrary. It is deeply rooted in the same English jury tradition that excludes hearsay evidence and seeks to present jurors only with reliable information. This reliability standard is intentionally strict, and it should not be displaced by a sufficiency standard or ideas about jurors sorting out a "mosaic" of evidence, which in practice would be far less exacting. When it comes to admitting evidence, the "best available evidence" may simply not be good enough--"best available" is not a defensible standard. This often means that "law lags science"--and that is how it should be. This might lead to someone's legitimate search for justice being delayed, but there is no cure for that ill not far worse than the disease. There are many examples of plaintiffs destroying corporations with allegations of misconduct later conclusively shown to be without scientific merit--silicone breast implants and Bendectin being the most noteworthy. There are few, if any, cases of errors in the opposite direction, indicating that a too-rigorous standard is not our problem.
AEI research assistant Philip Wallach prepared this summary.
