Neuron Culture

Science found wanting in nation’s crime labs,” says the headline at the NY Times, which ran one of many stories on the upcoming National Academy of Sciences report on forensic science. This kind of front-page attention is long overdue, as shabby science that claims to be infallible has jailed many an innocent (and probably freed a few guilty). As the Knight Science Journalism Tracker notes,

The main surprise, upon reflection, … is that this news was not dug up and given heavy attention by media already, starting years ago.

Well, it was dug up, and it did get some press attention — though perhaps not of appropriate prominence. Michael Specter questioned fingerprint matching in a 2002 New Yorker article [download pdf], and the New Yorker visited forensics again in 2007 — though, in my view, through glasses a bit rosy. And — after failing to interest a couple high-profile magazines in the subject — I published in 2006 a story in Popular Mechanics that covered this ground.

As so often happens, the data was out there. But in this case it was — of all things — a government report that put it on the front pages and got wide play.


  1. #1 Pierce R. Butler
    February 7, 2009

    Surprise, surprise – certain crazies are already twirling this report for their own agenda:

    The National Academy of Science Has Been Perfecting Double-speak … How does creating one giant bureaucracy that oversees all forensics make the industry more independent? … The solution is not to further consolidate our judicial and legislative systems, but rather to realign the duties of police forces towards protection of private property and separate these duties from the state itself. … taxpayers forked over $1.5 million to fund this particular study.

  2. #2 David Dobbs
    February 7, 2009

    Pierce R. Butler wrote:

    “How does creating one giant bureaucracy that oversees all forensics make the industry more independent? …”

    The NAS report (along with many others) recommends that forensic work be done not by police but by outside, unaffiliated labs. If you want to call it science, which presumably connotes a disinterested evaluation of evidence, then you can’t let the police do the job, for the police are highly interested in finding someone to charge. They will be biased toward finding matches of evidence to suspects rather than determining whether a match exists.

  3. #3 Pierce R. Butler
    February 7, 2009

    Please note that I was quoting Lew Rockwell’s absurdity while labeling it for what it is.

    If LR is any example, the libertarians seem to think that the best way to improve this problem is to have all the labs deliver whatever their clients want, unsupervised & unregulated – which would be okay, because the cops themselves would also be operating in a gun-for-hire-to-the-highest-bidder mode…

    Seems to me – as a nonscientist – if we wanted forensics labs to perform in a truly scientific way, we’d require not only independent operation but replication of each analysis by a separate facility. This isn’t financially feasible, especially in the post-Bushonomics era, but it would be quite interesting to fund a test project to do exactly that in a range of cases. (Or is that what the NAS just did?)

  4. #4 David Dobbs
    February 8, 2009

    Pierce — I quoted you too quickly and did indeed in effect put Rockwell’s words in your mouth (even though I myself recognized they were his, not yours). Apologies! And thanks for the correction.

    You’re right: replication would be ideal. But having disinterested labs do the work in more blinded fashion would be a big improvement on the present practices.

  5. #5 Zenneia McLendon
    February 18, 2009

    My name is Zenneia McLendon and I work for the National Academies. We appreciate your comments on the National Academy of Science report, “Strengthening Forensic Science in the United State: A Path Forward”. We encourage your readers to visit us at to get a copy of the pre-publication of this widely anticipated report, or to read it online.

New comments have been disabled.