What's Wrong With Forensic Science

In a culture smitten by CSI, NCIS, and Bones, it wasn't surprising that the jury bought the blue-jeans testimony. Testifying in a 1989 murder trial in upstate New York, a forensic scientist explained that the imprint on a pickup owned by the accused could have been made only by the victim's jeans—of which a mere 200 had been sold in the area. The scientist described her meticulous analysis of fabric and stitching; soil from the truck also matched dirt at the crime scene. The jury bought it: guilty.

If this had been a TV show in which forensic science solves murders through bite marks, footprints, even sand grains (a match between the grit on a suspect's shoe and the beach where the victim was killed!), that would have been the end of it. In real life, the convicted man was innocent, as DNA analysis revealed (he was freed in 2008). It was no aberration. According to the Innocence Project, of the 252 DNA exonerations since 1989, half the convictions were based at least partly on "unvalidated or improper forensic science." The surprise is that the rate isn't higher: a 2009 report by the National Academy of Sciences found that, in contrast to DNA matching, "for many other forensic disciplines—such as fingerprint and toolmark analysis—no studies have been conducted" to determine how many shoes, teeth, fibers, sand grains, or anything else "share the same or similar features" and so might be linked to the wrong person. As a result, invalid forensic science "may have" helped convict innocent people.

There is no "may" about it. An Idaho man was convicted of rape and murder based on shoe prints; he spent 18 years on death row. A Louisiana man was convicted of rape when a bite-mark analyst testified that he "is the person who bit this lady." Both were exonerated by DNA.

The readership of NAS reports, unfortunately, is hardly enough to fill a voir dire, and little has changed. There is no progress toward creating a National Institute of Forensic Science, as the report recommended, and the only forensic field that is formally trying to be more scientific is fingerprinting, which is already better than analyses of bite marks, handwriting, lip prints, shoe prints, and other variants of "pattern recognition." Individual prosecutors may be declining to bring cases based on dubious forensics, says staff attorney Nina Morrison of the Innocence Project, and—in a first—a federal judge in Massachusetts ordered lawyers to actively challenge forensics that have long been presumed valid, including ballistics and fingerprints, and prosecutors to be prepared to have it scientifically validated. Now the reformers are getting help from a perhaps unexpected source: Kathy Reichs, anthropologist and author of the Temperance Brennan novels that inspired Fox TV's Bones.

In her 2009 book, 206 Bones, Reichs tears into pop-culture portrayals of forensic scientists "as knights in shining lab coats." She warns that jurors cannot "spot junk science" (blue jeans, anyone?) and that some forensics disciplines are rife with "bad methodology, sloppy performance, or intentional misconduct." (Points for prescience: in March, the chief of a CSI lab in Nebraska was convicted of planting a murder victim's blood in a car linked to a suspect, to support a false confession.) In her next book, Spider Bones, which is due in August and centers on the misidentification of a body, she underscores the shoddy evidentiary basis for many forensic sciences.

When I caught up with Reichs, she was unsparing. "Jurors have been misled by Bones and CSI and the idea that science is always going to solve the case," she said. Although the NAS report has been topic A at the last two annual meetings of the American Academy of Forensic Sciences, and many members agree that "we need to firm up our protocols and the replicability and reliability" of forensic techniques, she said, there are holdouts. Those who specialize in the pattern-recognition techniques that have little empirical basis have been "resistant," Reichs told me. Yet the "science" of these fields is so dodgy that "sometimes they can't even agree if, for example, it is a bite mark," let alone whose.

The NAS report recommended that forensic scientists be board-certified, like doctors; as Reichs noted, unqualified people, many inspired by the media, "are hanging out a shingle." But instead of rigorous, impartial certification, companies are springing up to take almost anyone's money "and let you add some initials to your name," said Reichs—impressive to juries, but meaningless. Like too much forensic "science."

Join the Discussion