Reforming Forensics

Wednesday, May 14th, 2008

Fairleigh Dickinson Professor Roger Koppl argues for a significant overhaul of forensics in the U.S. in the current issue of Forbes. Forbes editor William Baldwin was alarmed enough at Koppl’s examples of forensics malfeasance to write a sharply-worded editorial of his own.

Koppl wrote a study on forensics reform for the Reason Foundation, and wrote a summary of the study for the November issue of reason. Koppl and I have also co-written an article touching on similar themes that will appear in an upcoming issue of Engage, a journal published by the Federalist Society.

Koppl’s work deserves more attention. Controlled studies have shown that the bias forensic experts absorb even by such seemingly innocuous interactions as speaking with police and prosecutors before running tests can have a disturbingly significant impact on their results. This bias exists even in well-intentioned, professional scientists. That’s bias that’s independent of the more egregious examples such as Dr. Hayne in Mississippi, or cases where prosecutors put ethically dubious pressure on forensics experts to tailor their findings to help the prosecution’s case.

Koppl’s proposals employ competition, proper incentives, and strategic manipulation of information (that is, separating information about the crime from the analysis of the evidence) to produce more accurate results—results less likely to be influenced by unintended bias, and that also would also go a long way toward uncovering the more egregious offenders.  Koppl estimates that the cost of implementing his ideas would be less than the cost of just a couple wrongful convictions.

The most urgent of Koppl’s reforms is the idea of giving forensic vouchers to indigent defendants. We need a Gideon v. Wainwright for forensics. Until defendants are given access to their own experts, far too many criminal cases will feature testimony only from state forensic scientists, and all the problems that come with that.  When only one guy with letters after his name is testifying, jurors are going to tend to put quite a bit of faith in what he says.  We’ve seen this even when what the expert is saying is absurd, and scoffed at by just about everyone else in the scientific community.  When poor defendants aren’t given access to their own experts, then, it calls into question whether we really have an adversarial criminal justice system.

Unfortunately, too many people think all of the country’s forensic labs work like the ones they see on CSI.  I’m not sure it’s enough to merely ask that judges take a more aggressive approach to weeding out the frauds.  First, judges can be duped, too.  Second, even competent, professional forensic scientists can make mistakes.  The changes need to be more radical.  Another of Koppl’s suggested reforms essentially applies the idea of peer review to the criminal forensics process. That would go a long way toward cutting down on mistakes, intentional and otherwise.

Digg it |  reddit | |  Fark

9 Responses to “Reforming Forensics”

  1. #1 |  Ryan The Sea Lion | 

    This is unrelated, but last night I discovered MySpace has blocked all links to

    You had earned a place in my ‘heroes’ section, and I htmled in a link to your blog (along with Tom Palmer’s, etc). Of the six links, yours is the only one blocked. This is the message MySpace gives:

    You have reached a link that is no longer in service. That means the link was very naughty, and, much like head lice, had to be eliminated before it spread.

    You may be asking yourself, “Hey, what was it about that link that got it in trouble?” An excellent question! Usually, it’s one of the following reasons:

    – The link was spam! No one likes spammers, and we don’t like their links.

    – You almost got phished! There are people out there who want to steal your MySpace password…

    – Viruses are not fun!

    click ‘Ryan The Sea Lion’ to see for yourself!

  2. #2 |  annemg | 

    An additional off topic… I was surprised when I opened my textbook today and see your name in it! I wonder if it’s bad that I’ve had it since January and I never noticed before….

  3. #3 |  Jethro | 

    I’m afraid police and prosecutors will leap on the idea of competition and twist it so that the forensics experts compete to give the answer most genial to the prosecution.

  4. #4 |  crack | 

    I’ve mentioned it before, and maybe I saw it here first, but Malcolm Gladwell’s take down of profiling is worth mentioning again.

  5. #5 |  Michael Chaney | 

    I’ve been saying for a long time that forensics labs need to employ basic scientific method, such as we all learned in high school.

    Examples of “how it’s done” vs. “proper method”:

    1. See if these two blood samples match
    Proper method: see if any of these 5 blood samples match, which includes one from the crime scene, one from a suspect, and three control samples which came from random people. Perhaps grab some from the blood bank that they’re getting ready to toss.
    The only slight problem is that the sample from the crime scene will be obvious, but if this is done double-blind it won’t be a big problem.

    2. Our drug dog hit on his money, so he’s dealing
    Proper method: Ladies and gentlemen of the jury, we have here 5 samples of cash marked a-e. They include one from the defendant, as well as 4 that were submitted by the judge, the prosecutor, and two of you. They have been randomly labeled by our expert who has left the room, and only he knows which is which. We will now see if the drug dog hits on any of this cash. After noting which ones he hits on, we’ll bring our expert back in to reveal which cash belongs to which person.

    That’s proper scientific method. I cringe, especially on #2. Without a control and double-blind methodology, it’s utterly meaningless. This goes for a *lot* of what we see.

    This brings to light, also, that the juries need to have investigative powers.

  6. #6 |  supercat | 

    The problem with double-blind studies is that they don’t work. Which is to say, they don’t give the results they’re supposed to.

  7. #7 |  Against Stupidity | 

    The double blind protocol is to prevent the person who conducts the testing from having any preconceived notion about what the results should be. It does work, as long as the tester does not try to circumvent the protocol.

  8. #8 |  Michael Chaney | 

    I think that’s supercat’s point. To spell it out (and tell me if I’m wrong), prosecutors can game the system now to get the results they want. If done scientifically, it’s impossible or at least way more difficult.

  9. #9 |  Frank | 

    This is nothing new. Remember what happened in Houston? The crime lab director that got fired for not cooking the evidence for police and prosecutors? The one that sued for wrongful termination and won, getting her job back, back pay, legal fees and a small amount of punitive damages?

    How is it that there was no outcry then? Could it be that this is what people really want — convictions at all costs?