In response to a Washington Post investigation, published back in April, the Justice Department has asked courts to free two men convicted of violent crimes largely because of testimony from hair and fiber experts in the FBI Crime Lab. They are also undertaking a "major review" of the role played by crime lab data in other convictions. According to the Post, the FBI has known for years that their lab regularly produced false results, but they never shared this knowledge with judges or defense attorneys.
Even though this will be the "largest post-conviction review ever undertaken by the FBI," it will not include the only thing that would tell us whether these methods work: a randomized, double-blind trial of the lab overseen by outside scientists. It frustrates me profoundly that nobody involved will acknowledge that this is necessary. Instead we get a lot of crap about "better training" and "stricter oversight" and such, when what we really need is to know how accurate the lab's results are. I would be willing to bet that when it comes to fibers, ballistics, and fingerprints, no lab in the world is right more than 80 percent of the time. But of course I am guessing, since nobody has ever done the necessary tests. Judges and juries need to know what the real numbers are.