My article, "Expert Witnesses, Adversarial Bias, and the (Partial) Failure of the Daubert Revolution," has just been published in the Iowa Law Review. You can find an on-line version here. Here is the abstract:

This article raises two questions that have been surprisingly missing from the voluminous law review literature on expert evidence since the landmark Daubert decision. First, what is the underlying rationale for the replacement of the old qualifications-only, let-it-all standard for expert testimony with Daubert/Federal Rule of Evidence 702's requirement that all expert testimony be subject to a stringent reliability test? Second, once we have identified this rationale, has the Daubert revolution succeeded on its own terms?

I conclude that the implicit rationale for the reliability test is to preserve the perceived advantages of the adversarial system, while mitigating the harms to the courts' truth-seeking function by the inevitable strong biases that accompany adversarial expert testimony. These biases include the conscious biases of hired guns, the unconscious biases of other paid experts, and the selection biases that result from the fact that attorneys shop for their experts from a large pool of qualified individuals.

Rule 702 thus attempts to serve a worthy goal, but it far from fully succeeds in efficiently achieving this goal. First, in the context of forensic expertise in criminal cases, Rule 702 does nothing to address the huge gaps in resources between the prosecution and most defendants that severely inhibit defendants' ability to challenge unreliable prosecution expert testimony.

Second, Rule 702, applied correctly, does succeed in barring junk science causation evidence in toxic torts cases. However, it does so at the expense of excluding speculative evidence supporting causation, even when most experts in the field would conclude that the relevant evidence is a sufficient basis from which to find causation by a preponderance of the evidence. While Rule 702 is easily preferable to the prior overly permissive regime, it likely goes too far in insisting on a reliability test that makes the courtroom stricter about causation evidence than is the scientific community itself. The way around this problem is to amend Rule 702 to allow courts to admit educated guesses about causation, but only when nonpartisan experts, not subject to adversarial bias, are willing to make such guesses.

Finally, Rule 702 puts severe restrictions on the testimony of experience-based testimony by "connoisseurs". Such experts may only testify if their field of expertise is a legitimate one, and they have proven to the court that they truly have the expertise they claim. Rule 702 also properly prevents attorneys from shopping for outlier and hired gun connoisseurs, given that there is no objective way for a jury to determine whether an experience-based expert's views are correct or representative of other experts in the field. Therefore, in the context of connoisseur testimony, courts should either replace adversarial experts with a panel of nonpartisan experts, or only allow an adversarial expert to testify if his conclusions are consistent with those of a nonpartisan advisory panel.

In my Iowa Law Review article, discussed in my previous post, I point out that Daubert/amended Rule 702 haven't done much as yet to improve the state of forensic science, and are unlikely to do so in the future.

One reason that forensic science (itself a misnomer, as there is little science involved, for the most part) has emerged mostly unscathed from the upheaval in expert evidence law, I suspect, is that the "Daubert Trilogy" of expert evidence cases--Daubert v. Merrell Dow Pharmaceuticals, General Electric Co. v. Joiner, and Kumho Tire v. Carmichael--on which Rule 702 is an elaboration, were all civil cases. Moreover, despite the obvious relevance to criminal cases, the criminal law bar was virtually silent when these cases were being decided. For example, when Daubert was before the Supreme Court, out of over thirty amicus briefs, not a single one was filed by an organization concerned with criminal justice issues. Almost all the energy opposing "junk science" has come from corporate America and its attorneys, with the opposition coming from the plaintiffs' bar.

Political liberals, especially elected officials (who, not coincidentally, get much funding from the plaintiffs' bar, and very little from public defenders and their clients) have generally opposed attempts to crack down on bad science (and other forms of expertise) because of the latter dynamic, and have mostly ignored the ongoing abuses of expert testimony in the criminal justice system. Prosecutors, for their part, have fought tooth and nail against attempts to restrain their use of dubious evidence, especially at the state level--I've heard from sources in several states that the main barrier to adopting a rule requiring reliable expert testimony in their states has been the strong opposition of prosecutors who fear that such a rule will make it hard for them to get convictions. Of course, their job, properly, is not to "get convictions," but to convict the guilty; and even though Daubert has actually had very little effect on prosecutors' ability to get convictions, the mere idea that their experts may be held to some reasonable standard of reliability is enough to raise prosecutors' hackles.

For reasons discussed in my Iowa article, mere tinkering with the rules of evidence is unlikely to change the current dynamic in which quackspertise is almost routinely admitted in criminal cases. But I wonder if more progress would have been made by now if quackspertise in the forensic contest had received a fraction of the attention of junk science in civil litigation, from the Supreme Court and otherwise.

For essential reading on the subject, see Paul C. Giannelli, Wrongful Convictions and Forensic Science: The Need to Regulate Crime Labs, 86 N.C. L. Rev. 163 (2007) (detailing many of the problems with forensic science) and Roger Koppl, How to Improve Forensic Science, 20 Eur. J.L. & Econ. 255 (2005) (providing the best ideas I've seen to improve forensic science, using economic reasoning). Todd recently linked to a popular version of Koppl's ideas, published in Forbes.