A new piece by Chris Fabricant and I is now posted on ssrn here.  Below is the abstract:

The U.S. Supreme Court’s ruling in Daubert v. Merrell Dow Pharmaceuticals, Inc., and subsequent revisions to Federal Rule of Evidence 702, was supposed to usher a reliability revolution. This modern test for admissibility of expert evidence is sometimes described as a reliability test. Critics, however, have pointed out that judges continue to routinely admit unreliable evidence, particularly in criminal cases, including flawed forensic techniques that have contributed to convictions of innocent people later exonerated by DNA testing. This Article examines whether Rule 702 is in fact functioning as a reliability test, focusing on forensic evidence used in criminal cases and detailing the use of that test in states that have adopted the language of the 2000 revisions to Rule 702. Surveying hundreds of state court cases, we find that courts have largely neglected the critical language concerning reliability in the Rule. Rule 702 states that an expert may testify if that testimony is “the product of reliable principles and methods,” which are “reliably applied” to the facts of a case. Or as the Advisory Committee puts it simply, judges are charged to “exclude unreliable expert testimony.” Judges have not done so in state or federal courts, and in this study, we detail how that has occurred, focusing on criminal cases.

We assembled a collection of 229 state criminal cases that quote and in some minimal fashion discuss the reliability requirement. This archive will hopefully be of use to litigators and evidence scholars. We find, however, that in the unusual cases in which state courts discuss reliability under Rule 702 they invariably admit the evidence, largely by citing to precedent and qualifications of the expert or by acknowledging but not acting upon the reliability concern. In short, the supposed reliability test adopted in Rule 702 is rarely applied to assess reliability. We call on judges do far more to ensure reliability of expert evidence and recommend sharper Rule 702 requirements. We emphasize, though, that it is judicial inaction and not the language of Rule 702 that has made the reliability test a myth.