Sunday, January 17, 2010
Texas should strengthen standards for admitting forensic evidence
The Texas House Criminal Jurisprudence Committee hearing last week included an interesting debate on the subject of the proper standard for admitting forensic testimony into evidence. Committee Chairman Pete Gallego opened the hearing by declaring he was looking for a "milepost or guidance" for when and how forensic evidence is admitted that would "rebuild some of the faith, some of the credibility of the system."
UT Law professor Gary Wellborn was first up, and his testimony framed the debate for many of the speakers after him. Elaborating the case law governing admission of evidence, he noted the the Kelly standard used in Texas is essentially similar to the federal Daubert case, and our rules of evidence are also the same on the subject.
However, Daubert usurped an earlier 1923 case known as Frye which required that a forensic technology mush have reached "general acceptance" among scientists. Daubert used that as one part of a multi-pronged weighing test which also evaluated the "reliability" of forensic methods, an innovation praised by prosecutors as a more "flexible" (read: permissive) approach because they didn't have to wait until scientists agreed to use a new technology or forensic method.
Between 12 and 16 states (different numbers were given by different presenters at the hearing) still operate under the Frye standard, but those include most of the larger states, said Wellborn, so that a majority of the US population actually lives in jurisdictions where forensics can't come into court until they've achieved "general acceptance" among scientists. Generally, Frye is considered a somewhat stricter standard among legal scholars.
Texas operated under the Frye standard before 1992, and Wellborn suggested the Legislature consider reenacting it by statute. He said that though the same standards theoretically govern civil and criminal evidence, in practice evidence was vetted much more thoroughly in civil court because all parties have access to resources, whereas in most criminal cases (especially those with indigent counsel) only the state has access to forensic evidence, expert witnesses, etc.. Chairman Gallego lamented that the Court of Criminal Appeals seemingly doesn't apply the same level of "rigor" when evaluating evidence as civil courts.
Wellborn said he didn't consider enacting the Frye standard as the end-all-be-all of reform. The change "could make very little difference" in practice, he said, but it's something reasonable the Lege could try to tighten up the system. Since most of the US population already lives under this standard, he emphasized, the risk of serious unintended consequences is low.
Williamson County DA and Forensics Science Commission Chairman John Bradley testified later that he held the "opposite point of view" from Mr. Wellborn and that Frye is not the right test, but the reasons he gave for saying so don't hold up to scrutiny. Bradley told the committee that if the Frye standard had been in place, Texas courts would not have admitted DNA evidence for another 5-10 years compared to Texas' more permissive case law. He went so far as to suggest that innocent people would have been forced to sit in prison longer if Prof. Wellborn's suggestion were implemented.
That's an excellent argument and there's only one real problem with it: It's simply untrue. Just four years after Texas changed its standard, the National Research Council completed a comprehensive review of DNA forensics in a 1996 report, The Evaluation of Forensic DNA Evidence, which found the Frye standard had not delayed admission of DNA evidence: "To the best of our knowledge, no state or federal court has held that VNTR profiling is inadmissible on the grounds that it is not scientifically accepted or sound." What's more, in jurisdictions following Frye, they found, "The biological and technological principles underlying the forensic methods for characterizing DNA variations have generated little controversy in court."
Jeff Blackburn from the Innocence Project of Texas added an excellent point that IMO should receive just as much attention as Wellborn's idea of reenacting Frye. Blackburn suggested the committee should strengthen by statute the so-called Nenno standard, which further relaxed requirements for forensic testimony to eliminate expectations that "soft" sciences must be subjected to peer review or independent, external evaluation. Opined the court, "hard science methods of validation, such as assessing the potential rate of error or subjecting a theory to peer review, may often be inappropriate for testing the reliability of fields of expertise outside the hard sciences." This ruling opened the door to all sorts of "soft" forensic testimony and has been the source of much mischief, as evidenced by the fiasco over faulty dog-sniff lineups.
For more background, see a good discussion in this CCA decision of the relationship between Frye, Kelly and Nenno under current Texas case law.
Whether Frye is the right standard, and it may well be the best approach, I certainly would like to see the Legislature step up to require greater rigor regarding forensic evidence admitted into criminal court, including for the "soft" sciences. Concerns that stronger standards might keep out genuinely probative evidence are either overblown or manufactured - that didn't happen with DNA and it won't happen with other legitimate scientific advances. But there has to be a way to distinguish between evidence like DNA and the guy who uses dogs to match scents to suspects. One is legitimate science, the other is backed by no science at all and finds its primary support in past use by Communist totalitarian police states.
Since the Court of Criminal Appeals overtly weakened standards by which Texas judicial gatekeepers judge forensic evidence, it's up to the Legislature to enact a stronger standard for them. I thought the discussion at the Criminal Jurisprudence Committee on Monday (see the archived video here) was a good first conversation about what that stronger standard might look like.