DNA: Is it exonerating the innocent or putting them away?

The ability to test biological evidence for DNA has been the greatest innovation in scientific evidence in the criminal justice system over the last three decades. How important is it? Just ask someone like Ronald Cotton who did over 10 years for a rape he did not commit. Better yet, ask Jennifer Thompson, the victim of that rape what she thinks of DNA. Thompson actually ID'd Cotton TWICE during trials. At his original trial and then, several years later when he got a new trial. Both times, Cotton was convicted and Thompson was sure he was the man who had raped her. It was not until DNA testing came about later, that Thompson had to face the fact that at that second trial, she had looked right at the man who had actually raped her, Bobby Poole, and never had any recognition that he was the guy. Folks like Cotton and Thompson can tell you first hand the power of DNA to save innocent people.

But a new article by Michael Bobelian for The Washington Monthly, DNA's Dirty Little Secret, points out that all is not kosher in the use of DNA to solve "cold case" files. Specifically, the article focuses on a 1972 rape/murder case from San Francisco which resulted in the 2008 conviction of a man named John Puckett. The article is is well worth a read.

The article basically points out one of major problems with DNA being used in courtrooms: the statistical probabilities that are thrown around by the State to show just how sure they are that this particular person committed the crime. In Puckett's case, the jury was told that a random person's DNA would match the DNA found at the scene just one in 1.1 million times. But that was not accurate. You see, in Puckett's case, the actual chance of a false match was a staggering one in three. Unfortunately for Puckett, his lawyers were specifically barred from even bringing this fact out to the jury. Why? Well...that's the question now, is it not?

Bascially, the FBI's own DNA advisory board and the National Research Council, a body created by Congress to advise the government and the public on scientific issues, came up with the formula Puckett's attorneys used to come up with that one in three figure. In fact, those bodies have specifically recommended that law enforcement and prosecutors calculate the probability of a coincidental match differently in cold-hit cases like Puckett's (ie., use the formula that those same boards came up with and that Puckett's attorneys used).

But here's where it gets interesting, not only do most law enforcement agencies not utilize the recommended formula, but the FBI specifically is attempting to bar states from going into their national database and performing tests to see just how accurate those probabilities are. That is bullshit.

We once attended a seminar where Brent Turvey, MS spoke about the problems in the Forensic Science community. If there was one thing we took away from Turvey's presentation it was this: you can't have science without the scientific method. And you can't have the scientific method if you don't put your hypothesis up and attack the ever living hell out of it. Only if it is left standing after all those attacks, can you call it good science. A scientist, by nature, seeks to disprove things. It is only through the inability to disprove, do we ever actually prove something. If the FBI's probabilities are "good science," then they will stand up to attack. If the FBI is afraid to let people test and attack those probabilities, what does that tell us? And if they can't stand up those attacks, why are they being allowed into the courtroom?