Mark Kleiman asks a good question:
Having worked on the problems of crime control for almost thirty years, I tend to be much more sympathetic to the viewpoints and operational needs of law enforcement agencies than the average of the people I usually agree with politically. But on one point, I find myself utterly unable to understand what my friends in the law enforcement biz could possibly be thinking: why isn't it as obvious to them as it is to me that clearing innocent people is just as important a goal of law enforcement as nailing guilty ones?
I agree 100% that this should be a coequal goal with convicting the guilty; but it doesn't surprise me that it isn't. Human beings are such irrepressible optimists, so naturally aversive to meditating upon their own failures, that psychologists have a technical term for the rare people who are predisposed to clearly and accurately assess their achievements: "depressives". When we fail, the natural urge is to cover it up--to others, in order to preserve our status, and to ourselves, in order to preserve our peace of mind. Undoubtedly, the folks at the FBI who decided not to notify people that they'd been convicted on faulty evidence reasoned that those people were all probably guilty anyway, and no real harm had been done, so why rock the boat?
The more important question, I think, is why the rest of us don't spend more time worrying about false convinctions. What I've read about the Jeffrey MacDonald case, for example, makes it clear that at the very least, prosecutorial misconduct and dubious forensic testimony played some role in his conviction. This should bother us whether or not he's guilty, since presumably the kind of games the prosecutors played with the evidence have been inflicted on other, less notorious, defendants who may have been innocent. Yet there's been little interest from any quarter.
...why isn't it as obvious to them as it is to me that clearing innocent people is just as important a goal of law enforcement as nailing guilty ones?
Clearing innocent people is important, but it's hard to imagine why (for law enforcement people, anyways) it would be just as important, as they're not, you know, dangerous, like criminals.
Well many of us do. Especially those who have suffered at the hands of the criminal justice system "for crimes they did not commit."
But would guess that the just-world effect underlies much of the problem. We have a hardwired tendency to align with authority figures like the police (and government overall) and assume that people arrested are guilty - so proper criminal procedure becomes just an unnecessary formality.
I suspect that this one reason the death penalty is unpopular amongst many libertarians (Ron Paul also, and presumably The A-Team as well).
This effect is probably exacerbated by all the "cops and courts" tv dramas with the frequent cheap storylines about bad guys who get away because of some procedural technicality or restraint that keeps the 'noble' cops and prosecutors from getting them.
As to depressive realism, I would note that depressives tend to be at a disadvantage when it comes to rising in the ranks of most organizations (especially those which lack the discipline of outside competition, such as government agencies).
Blinding extroverted optimism and confidence works much better for climbing the ladder. So there is probably a selection effect adding to this problem.
Also, I can't help but think of your previous post about public schooling - how people seem to value the appearance of proper means (criminal procedure or public schooling) more than actually achieving the proper ends (convicting the right people or educating children).
It's not that easy to run a successful prosecution. The bias starts with the presumption of innocence. It is evident with the first steps, i.e the police, who have some pretty weighty regulations to follow, in terms of procedure and evidence. It continues on with trial motions, where most dismissed help the defense. It is up to the prosecution to prove its case beyond reasonable doubt; the defense defaults to a prima facie stance, and, with sufficient funds, pay for an edge in litigation strategy and tactics (Johnny Cochrane comes to mind)..
Our system is founded on the errors tradeoff of a thousand guilty beating the rap before a single innocent is wrongly convicted. I think the current results approximate that ratio, from the jay walkers and casual marijuana smokers all the way up to Marc Rich.
Admitting mistakes undermines a LEO's authority, which, unfortunately, is more important to many of them than actual justice.
My question is why is the FBI not legally obligated to provide notification?
Because being so committed (and I absolutely wish we were) would require a realization that is very difficult for people to accept: the trial by jury system is one that fails all the time. We like to think it doesn't, but there is ample evidence, both in terms of false convictions and unjust acquittals, that it doesn't. It's the sort of truth we can't admit because it is too uncomfortable; our system depends far too heavily on the assumption that our legal system works.
Interestingly, Kleiman is the first to complain about this issue and offer up a statistic of some sort. The numbers are hard to come by and typically arrive with the "one is too many" cliche.
Kleiman estimates an error rate of 2% for incarceration. I'll say it's 4% for argument's sake. What is an appropriate rate and how much do we have to spend to get there? I'm guessing that to get down to 0.05% (a 8X improvement), it would cost a lot more that we already spend.
Speaking as someone who actually worked in law enforcement, I can say that the main reason why LEO's and prosecutors aren't as concerned about clearing the innocent is due to 2 reasons.
One is that the public doesn't care, and the second is that it doesn't advance their careers.
Everyone has indicators that their bosses look at to see if someone is doing a good job. For the uniformed officers, it is the number of arrests no matter if the person taken in is released later or not.
For detectives, it is clearing cases. What this means is finding a specific person to arrest, as well as gathering enough evidence for the prosecutor to take the case to trial.
For the prosecutor, it is the number of people he manages to get convicted.
Please note that every step has to be what is known as "righteous". ("That was a righteous arrest.") What that means is that the cop, detective or prosecutor can't abuse their authority so much that they are found guilty of committing a crime themselves, or even violating department policy too badly. To do so is a career ending move.
I have mentioned before that people who have never worked in the law enforcement field before have some sort of bizarre idea that the cops, detectives and prosecutors are making the rules up as they go and setting policy on the fly. This is simply not true in the vast majority of cases, since all of them get their marching orders from elected officials. Everyone involved does what they do because the voters want it that way, or at least don't care.
So, if you don't like the way things are, either elect someone who will change the system or campaign for office yourself. Blaming the police for following what the voters want is silly in the extreme.
James
[t]hose people were all probably guilty anyway
I think that this reason should not be underestimated. It includes the belief that "well, if they didn't do this, they probably did something else."
I support the Innocence Project 100%. At the same time, a decent number of the people that they claim are "exonerated" or demonstrated "innocent" are not actually innocent of the crime in the moral sense, but still should be found not guilty because of procedural errors. (They continue to claim some of them as "innocent" despite knowing for a fact that the person committed the murder.)
I suspect that there's also a sense in some cases of "not only is the suspect probably guilty, but he'll likely go on to commit X more crimes if released. So we balance a possible injury to one person against a possible injury to many."
Question: is there any way of judging the reliability of the judicial system in the first place?
If there is no system to assess this, there should be.
(I do not mean the reliability of any individual verdict, but the reliability of the system in general.)
ad: I've seen proposals by David Friedman for a way to try to assess this, but it honestly looks hard.
We have some conflicting information here:
a. There are a fair number of high-profile murder cases where the prosecution's story of what happened (which led to the conviction) is contradicted later by DNA evidence.
b. There are a bunch of cases where we've found out that the "rock solid" evidence used by the police/prosecutors to get a conviction is pretty weak and poorly used. We also get stuff like eyewitness evidence that turns out to be seriously wrong, prosecutors playing games like in the Duke Lacrosse case (if those guys hadn't had money, good lawyers, and connections, they'd all be in prison right now), etc.
c. We get people agreeing to testify in exchange for a reduced sentence, and the ugly game-theoretic side of plea bargaining where prosecutors and judges more-or-less explicitly tell people that they'll pile all the available charges onto the accused criminal, if he doesn't agree to plead guilty to something.
All those make it obvious that the system is noisy as hell. On the other hand, we have:
d. If you talk to people in prison (I have, a bit), very few of them claim innocence once they're inside and not up for any kind of appeal. I think this is a pretty common observation.
e. If you talk to defense attorneys, they will typically tell you that they very rarely see innocent people in their practice.
f. Decreasing the number of people we were locking up seems to have tracked broadly with big rises in crime rates, while increasing that number seems to have tracked broadly with falls in crime rates. This gives at least some circumstantial evidence that the police are arresting the right people, albeit maybe for the wrong crimes.