Posts
When a DNA match is presented in court, invariably the scientist proclaims that the odds of the DNA matching are something astronomical, like 1 in 4 quadrillion. Now, I'm no mathematician, but I do know that there are only about 6-7 billion people in the world. So those odds seem pretty slim. I think, in fact, that there probably have never even been 4 quadrillion people in human history. So the odds seem pretty good that the DNA really does belong to the defendant.
But what if the odds we've been given have been a bunch of bull? The standard for a DNA match has been 13 loci. I can't explain it that well, but every person's DNA has these loci, or locations of chromosomes. There are something like hundreds (or thousands or millions, but way more than 13). But in statistical analysis, the theory has always been that a small sampling size is all you need. That's why you can do valid phone polls with fewer than 1,000 voters. So it was always thought that 13 loci were a sufficient sample size to rule out the possibility of random matches.
A scientist in Arizona's crime lab unwittingly discovered the fallacy of that assumption. The LATimes published this article on the discovery. Out of her database, she found two people who matched at 9 of the 13 loci. Now, they weren't complete matches, of course, but they were remarkably close, closer than we thought was statistically possible. After that first discovery, the analyst combed through the Arizona DNA database and found dozens of other similar matches.
So defense attorneys involved in a DNA-based prosecution in Maryland sought a court order to do a similar search in their database. The judge agreed, at which point the Maryland crime lab folks, in conjunction with the FBI, dug in their heels. They did everything they could, including just flat saying no to a court order, to prevent the defense from looking through the state DNA database to see what kinds of random matches they might find. Now they didn't get away with it in Maryland, but they just might elsewhere. And a lot of defense attorneys around the country probably aren't yet aware of this identified problem. And even if they are, they may not have the means to do much about it.
What does the FBI not want us defenders to find? Are they afraid we'll uncover evidence that their oft-cited statistical analysis for random DNA matches are bunk? Or do they know we will? I'm always suspicious whenever police or crime labs do all they can to keep us from looking into things. If they're so sure of their evidence, I can think of no reason for them to hide it from the defense. I can't find a way around the thought that they know their evidence isn't really as irrefutable as they want the defendant and the jury to believe.
Think of all the kinds of forensic evidence that we now know have credibility problems. Hair evidence is absolute junk without DNA. Anyone who claims that s/he can say the hair found at the scene is a "microscopic match" with the defendant's hair is full of it. By merely looking at two hairs side by side under a microscope, one cannot even tell the difference between a human hair and a dog hair. Fire science is also coming under, well, fire now. I'm not as familiar with that area, but from what I do know, we all ought to be suspicious of the claims made by arson experts. Handwriting analysis has always seemed questionable to me. Should I now be even warier of firearms analysis? We know that the lead analysis they used to do on bullets was just a bunch of bunk. Blood spatter analysis, footprint analysis. Even fingerprint analysis is far more questionable than the general public realizes.
My rapidly growing theory is that the forensic experts at the FBI and around the country just make this stuff up as they go along. They make pretty graphs, use big words, sound authoritative, and hope no one ever gets the resources to delve too deeply into their claims. I think they might even really believe it themselves and I don't think it's necessarily malicious; I think they just really believe they've got the right guy and have to find some scientific way to prove it. They're just not that good at fully accepting, or even realizing in their zeal to nail defendants, all of the limitations of the science they want so desperately to use.
One of the main, ongoing problems with forensic evidence is the defendant's lack of access to independent experts. My clients are all indigent. They can't hire an outside lab to retest the evidence or review the crime lab's testing protocols. They can't hire a statistician to review the crime lab's claimed odds. So the defense attorneys ask for funds. They ask the court to fork over the money or they ask for it from their own over-taxed defender system. Most often, those funds are denied. The response is, "You can use the state crime lab." Well, no we can't. And the efforts to stymie defense investigation into the Maryland DNA database, backed by the FBI for crying out loud, ought to prove that. Those crime lab scientists do not consider themselves equal opportunity employees of the state, working for both sides. No, they consider themselves investigators working for the prosecution. Defendants simply do not have access to those labs and it's way past time for courts to stop hiding behind that excuse.
Prosecutions rely on forensic evidence, especially DNA, and most defendants aren't in a position to fully meet that evidence. But it's the evidence that juries most want to hear, and often the evidence they most rely on. If we can't get the resources to challenge that evidence properly in court, we have to find some way to get the message out to the public that this evidence isn't all we'd thought it was.
So for the defense attorneys out there, if you had not heard about this yet, now you have a new issue to raise in a case involving DNA.
And for those of you who might someday wind up on juries, remember that DNA is not the infallible evidence we thought.
5 comments:
DNA matching is like the Powerball. There are a discrete number of "balls", aka gene sequences, that are looked at and compared. The smaller the number of balls compared, the greater the chance two non-identical twins will "match" DNA. As we get better tests and processing power behind the databases, the number of balls should be high enough that there can be very, very little doubt that the DNA came from someone else.
This is why modern electronic fingerprint scanners suck. They only look at discrete points on the print for comparision. Therefore, they match too many people they shouldn't.
One day there will be the processingpower to grind through a database of fingerprints to compare scanned images. But with DNA there won't be much of a need, except to settle long gone cold cases with collected prints but no DNA.
So what we now know is that the number 13, of loci to compare, isn't as good as we thought. So how many loci do we need to go up to? And how will we know when we really have reached the number to compare that will sufficiently eliminate the possibility of random matches?
The main problem facing defenders now is we do not have access to the database to do our own tests. The FBI controls CODIS, the national database. They don't want defense experts accessing that database. They tell states that opening up the state database to defense experts will disqualify them from access to CODIS which scares off state agencies from cooperating with the defense. So we're still left at the mercy of the state's experts to tell us when they've really compared enough loci to eliminate random matches. I'm not comfortable with that, anymore, since I've heard that before and it's been shown to be questionable.
Actually I have seen a recent article that points out that finding 9 point matches isn't that unlikely. It's like winning a minor prize in the lottery - OK but doesn't prove anything. Finding non twins with a 13 point match would be a bombshell however.
That would be the bombshell! But we (the defenders) are never going to find it because the other side controls access to the information and they really, really don't want us to find a bombshell.
So we're like the winners of that minor lottery prize: we have renewed hope that we really can win that big prize, but we can't afford to keep buying the lottery tickets.
It is worth contemplating whether you should ask the next questions: "Was my client the only person whose DNA matched at all 13 points? Did you check for all matches? Did you check other databases for other matches?"
Of course you might not like the answers.
Post a Comment