My thoughts on the NAS Report on Forensic Science

Well, now I've read the full summary of the report and I've also gotten through the first two chapters. From the summary, I can already reach some conclusions, though I plan to read it all the way through.

There is much about it that I like. It does somewhat bother me that they sort of tiptoe around the notion of the strong pro-prosecutor bias in forensics, but I'm sure that is politics. What counts is the recommendations they make, and what they recommend would do much to eliminate that bias. (They do mention something known as "drylabbing" - which simply put is writing a forensic report, complete with "matches" to a defendant, without doing any actual analysis or any actual lab - thus, the lab stays "dry" - others may call this perjury, fraud, or a felony. That there is a word for it, to me, indicates how widespread it must be - sort of like police "testilying")

Ok, they do note the strong pro-prosecution bias in courts, noting that "ironically, the appellate courts appear to be more willing to second-guess trial court judgments on the admissibility of purported scientific evidence in civil cases than in criminal cases." (Page S-8) For this, they cite a study that notes that courts are more lax in their role as gatekeeper on science for criminal trials than civil - and they are much stricter with civil plaintiffs. In other words, they are much more careful about what gets admitted when large sums of money are on the line that might have to be paid by corporate interests than they are when peoples' very freedom is at stake. Big surprise there, right?

So in the end, they recommend what I'd do - having independent labs who serve everyone, not just law enforcement. They further recommend that to implement this at the state level, Congress use the power of the purse - offering funds for forensics only if states adopt an independent forensics lab system (matching what they recommend for the federal level - A NIFS - National Institute of Forensic Science) - The NIFS would be an entirely new agency, not beholden to any other agency, fully independent. They make a good case for why a new agency is needed. First, most existing agencies that could do it are already in bed with law enforcement or are part of law enforcement. None of them really has the focus anyway. It helps that this was what I would do, so I'm biased in favor of it, but that doesn't change the fact that it is an excellent idea.

One thing that struck me in reading the summary (I haven't gotten through the rest so maybe this is spelled out more specifically later) - was that even if you have independent labs, and take other measures, you still have the problem of the courts themselves and what they might or might not allow in. Further, I somehow doubt we'll ever get the independent labs - that would threaten prosecutors too much and prosecutors always get what they want these days (though maybe that will change, I see no sign of it now). But then something else they recommend might be a way to deal with that:

They recommend scientific research on the effects of human bias, intentional or not, in forensic analysis. While they don't go out and say this (at least in the summary) - to me, the main benefit of such research is you could use it to totally destroy the credibility of any forensics lab that is not independent. Because I suspect that if you really did such studies, you'd find out what we have been finding out since DNA evidence has begun releasing lots of innocent people from prison - that the forensic evidence used to put them there with its "100% matches" and certainties really was anything but. And much of that could be laid at the feet of bias for being in bed with law enforcement - some deliberate, maybe other just "helping" things along in more subtle ways. If we have a solid body of scienctific research that shows that non-independent forensic laboratories are hopelessly skewed in favor of prosecutions, that is cause for excluding it from trial altogether. Now, the right-wing jurists may not go along with this, but the intellectually honest ones will have to. So that would be another pressure point to get independent forensics labs - either you do it or you might as well flush your current lab down the toilet for all the good its reports will do you in court.

Also in the report is a strong emphasis on science for training and techniques. Right now, many forensic lab positions are based on an apprenticeship model. There are no scientific papers or underlying, peer-reviewed, scientifically proven techniques for much of what is used - for instance hair matching or even fingerprinting (if you can believe that - it is true - there is almost no peer-reviewed scientific research on how accurate or useful fingerprinting really is for identification). The report advocates eliminating all apprenticeship for forensics and making it 100% science based, starting early, with classes in undergrad onwards. They also advocate science for lawyers, judges, and law enforcement, as there is much scientific illiteracy there as well.

The strong emphasis on science and peer review also probably turns off current forensic experts, as the report notes that many of them were loathe to admit, when testifying before the committee, that their techniques provided anything less than 100% matches. Thus, they are also against another recommendation in the report that the actual scientific margin of error be prominent on reports. Even with DNA, this can greatly reduce the "match" below 100% - because while DNA matching is very good in the lab, there are still possible human errors - mislabeling vials, contaminations, etc. The actual error rate for matches would take all of this into account. So now instead of being told that there is a 99.9999999999% match, jurors would be told that there is a 85% match or whatever it would be, taking EVERYTHING into consideration. That puts the evidence back into the proper perspective, and also leaves room for reasonable doubt where there should be room for it.

Related to the desire to make science the backbone of the whole process is the recognition that it really isn't there, currently. The report notes that forensic labs everywhere are underresourced - you'd think that means not properly funded or manned, and that is part of it, but it goes deeper. The lack of a scientific foundation underlying much of the work means that there is no "there" there - the labs can't properly do their jobs even with more money and people because what they do simply isn't accurate or useful absent an underlying scientific process. That backbone needs to be created, from the ground up, in the world of science, not law enforcement.

As I finish reading the report, getting into the details of all of the above, I may write more on this. Suffice it to say that I think there are a great deal of excellent suggestions in the report. It saddens me to know that it is a virtual certainty that none of the significant recommendations will ever be implemented.

I have already heard criticism - as I mentioned before, law enforcement doesn't want non-independent labs. Well, tough shit for them - they need them. I think they know that they'll actually have to do their job with them and won't be able to take shortcuts, drylabs, or anything like that, and that is the impetus behind much of the resistance. I've also heard criticism that this shouldn't be a federal thing or some new federal agency, with the usual tired complaints about bureaucracy. Well, I think those complaints are bullshit. The lack of uniform standards is a big problem and the only way to get them is to have something national, not state by state. This will take strong leadership at the top to ever get implemented.

The important thing is to put scientists (who are good administrators, but strong in the science) in charge of the office and to keep politics out of it. Measure the results based on science and peer-review. And stick to that. Measure its success by accuracy in the science, rather than convictions or other bullshit numbers, and it may stand a chance.

Here's hoping.