If you read this blog you know I'm not a big fan of the way forensics is used in criminal cases. Many of the problems are the result of using techniques and processes that were not designed for determining a person's guilt or innocence. Another problem is that the scientific method does not parallel a criminal trial.
Under the scientific method you develop a hypothesis, and then try to prove it wrong. In most criminal cases it is the opposite. You try to prove someone is guilty, even if there is evidence to the contrary.
What made me think about this a decision out of New Hampshire (New Hampshire v. David McLeod) criticizing John Lentini, one of the leading arson experts in the country. Mr. Lentini's position is that you should start off with the presumption that the fire was an accident. He has a couple of reasons for that. One is that most fires are accidents. The more important one is that everyone is presumed innocent, so shouldn't you follow that in an investigation. The New Hampshire Court didn't like that approach, and prevented him from testifying.
The decision tries to portray Mr. Lentini as having an agenda - which he doesn't have. He has a presumption. What is the alternative - that you go in with an open mind - that's contrary to the scientific method, which the court believes it is following.
The problem with Mr. Lentini is that he's honest. He's admitting how he approaches the investigation. Those on the other side are not going to admit they assume the defendant is guilty, because that wouldn't look fair. Instead, they claim they will start of neutral.
The court relies on NFPA 921, which is the National Fire Protection Agency's guide to fire investigation. There's nothing wrong with that, and in fact it should be accepted - as long as you except everything in it, and not merely that which supports your position.
The NFPA calls for a systematic approach, and the court equates that with the scientific method. While that's part of it, so is testing a hypothesis, which you can't have if you are neutral. The end result in an arson case is that it was either an accident or arson. If you cannot disprove it was an accident a defendant should not be found guilty.
I don't know if the scientific method works or not in criminal cases. You are not dealing with absolute truths like you are in science. Instead, you are dealing many times with possibilities - what is more likely. Yet science still must play a role. If it is going to be the determining role - as it is in arson cases - I think fairness dictates Mr. Lentini's approach. Unfortunately, since that approach makes it more difficult to obtain a conviction, I doubt the court's will agree.