Emily Skarbek  

The Institutions of Forensic Science

Sympathy for the devil?... Friedman and Reynolds on Saez ...

Without posting any spoilers, it's fair to say that the new Netflix sensation, Making a Murderer, has sparked a significant discussion on the effectiveness of the US criminal justice system. Bravo. A much more critical attitude towards US incarceration practices is a welcome development, as the US is outstanding in respect to how much of the population we lock up. The graph below shows the prison use rate per 100,000 people in OECD countries.

Incarceration OECD Resized.png

Most of the discussion the show has stirred up centres on prosecutorial misconduct, police corruption, inadequate defence, false confessions, jury tampering, etc. All of which are important issues. But much less has been said of the role of forensic science in both producing convictions and overturning prior convictions.

Roger Koppl argues the institutional structure of police forensics is an inadequate system for producing good approximations of truth. Why? The institutional organization of forensic science is not a competitive system, where knowledge is tested, contested, voluntarily adopted, and subject to open-scrutiny and revision. Forensic science is political, not a set of institutions designed for tracking Truth.

What features of the US system of conducting police forensics are cause for concern? Koppl gives eight.

1) Monopoly. In most jurisdictions today, including those in the US, each laboratory has a monopoly on the evidence it analyses. No other lab is likely (or sometimes allowed) to examine the same evidence.

2) Dependence. Forensic labs are often organized under the police and are thus dependent on the police for their budgets. This introduces a clear pro-prosecution bias.

3) Poor quality control. Quality control systems tend to be weak.

4) Information sharing. Forensic scientists are privy to information that may be crucial to a criminal proceeding, but extraneous to the questions put to the forensic scientist.

5) No division of labor between forensic analysis and interpretation. The same scientist who, say, performs a test to establish blood type judges whether his test results exclude the police suspect.

6) Lack of forensic counsel. Indigent defendants rarely receive aid and counsel from forensic scientists hired to help them. In common law countries this creates an asymmetry whereby the prosecution has forensic counsel and, indeed, sometimes great batteries of forensic specialists, whereas the defence has no forensic counsel and, often, an attorney unable to adequately understand and challenge forensic testimony.

7) Lack of competition among forensic counselors. From the absence of forensic counsel for the indigent, it follows rather trivially that there is no competition among forensic counselors for their custom.

8) Public ownership. Forensic laboratories are publicly owned. In the US often organized under police agencies such as the State police or the FBI.

This institutional arrangement creates the conditions that can induce unconscious bias, conscious bias, and even give some an incentive to lie. Koppl outlines a system of "competitive self-regulation" with the potential to ameliorate many of these issues. Such a system would utilize private firms to conduct much of the testing, randomization in deciding which labs gets different pieces of evidence and when to send the same evidence to more than one lab, de-coupling forensic analysis and interpretation, statistical review for quality control, and a voucher system for supplying low income defendants with adequate forensic counselors.

Other calls for reform focus on laudable piecemeal changes (independence of forensic labs, double-blind proficiency tests, the use of evidence line-ups, etc) but do not discuss fundamental reforms in the institutional structure of forensic science in the US criminal justice system. Ultimately, the institutions determine both the knowledge and incentives facing actors within the system. Often failures in criminal justice are attributed to corruption, discrimination, or inequalities. But much of the systematic effects of the US judicial system and incarceration practices require examining the rules that structure the system.

COMMENTS (5 to date)
RPLong writes:

Wow, excellent post. Forensic science was definitely in my blind spot in this regard. But you're right, and Koppl is right. When have forensic scientists ever published their data and invited close scrutiny? Where is the competition?

Jesse C writes:

Very good post. Some thoughts:

1. It would be interesting to see some categorical analysis of the countries OECD countries using grading of the 8 points listed.

2. While I agree with the overall post, here's a thought experiment:

Imagine we had an oracle instead of a jury, and we still tried the same number of people but never jailed an innocent person. You'd probably feel much better about the things from this blog post, but how much lower would you expect the USA spike to be? Keep in mind that in order to be as low as the second highest bar, we'd have to be currently convicting mostly innocent people.

It's important to distinguish between

  • People being incarcerated for things that are petty, relative to the punishment.
  • Innocent people are being incarcerated.

IMO, the second point is more tragic, but I don't think it shows up in data in any eye-popping way.

I hate trivial laws. And the USA is FULL of trivial laws. But we're also tougher on actual perps where many(most?) countries are too lax. E.g., India has a very low incarceration rate at the cost of rape, mandatory bribes, etc. (I could be wrong with my characterization of India, maybe that's a weak example.)

Roger Koppl writes:

Jesse C: I've read over Emily's post carefully, and I don't think she says problems in forensic science cause over-incarceration. She is saying that if you look at that graph, you may start to see that we should have a "much more critical attitude towards US incarceration practices." She then points out that when people do start to tick off problems, forensic science is often neglected. I think she's right about that.

Emily says, "Ultimately, the institutions determine both the knowledge and incentives facing actors within the system." The incentives in the American criminal justice system in general are skewed inappropriately toward getting convictions independently of the guilt or innocence of those convicted. And that's true of forensic science in particular. Crime labs are often funded in part *per conviction*. This point has been discussed by Radley Balko:


Jesse C writes:

Roger - I see that Emily did not claim a causal relationship. I did read it as loosely implied: ("high incarceration rates... lots of factors, one seldom considered factor is bad state of forensic science"). I see that incarceration practices need a critical look is the the bridge - my mistake.

Re your 2nd paragraph: The institutional problems with forensic science are disturbing and far more important than the incarceration rate data, IMO.

BTW, I'm grateful to all of you working to improve something so important to a free society.

Urstoff writes:

How do researchers in forensic science feel about this? I work at a university with one of the top forensic science departments in the country (not too difficult, I gather, given that there aren't very many of them; the building is pretty impressive though), and would they agree with this critique and say that the actual science of forensic science gets radically distorted in its actual typical use?

Comments for this entry have been closed
Return to top