"Faith-Based Science"

In the May 7, 2007 New Yorker's Annals of Law column, Jeffrey Toobin writes about "The CSI Effect", focusing on the hair and fiber analysis performed at the NYPD crime lab. In what I consider the highlight of the column, Arizona State professor of law and psychology Michael J. Saks says:

There are really two kinds of forensic science. The first is very straightforward. It says, "We have a dead body. Let's see what chemicals are in the blood. Is there alcohol? Cocaine?'" That is real science applied to a forensics problem. The other half of forensic science has been invented by and for police departments, and that includes fingerprints, handwriting, tool marks, tire marks, hair and fibre. All of those essentially share one belief, which is that there are no two specimens that are alike except those from the same source. . . . There is no scientific evidence, no validation studies, or anything else that scientists usually demand, for that proposition--that, say, two hairs that look alike came from the same person. It's the individualization fallacy, and it's not real science. It's faith-based science.

I'd always wondered how fingerprint examiners could possibly assert with a straight face that no two people have the same fingerprints. Now I understand: it's a matter of faith.