Actuarial Justice

January 2, 2005 10:15 AM

The State of Virginia(nyt) has provided judges with a checklist to determine whether or not nonviolent offenders should go to jail. 40 year old woman with a job and husband = no jail. 21 YO man without job or wife = see you in 3-5. Here are the official guidelines (pdf) for sex offenders with a detailed explanation of the process.
posted by jmgorman (38 comments total)

I couldn't find the regular non-violent guidelines. My google-fu is poor.
posted by jmgorman at 10:16 AM on January 2, 2005

It's important to know something that the Times doesn't point out, which is that the Republicans abolished parole here in Virginia.
posted by waldo at 10:25 AM on January 2, 2005

Sure, checklists. How much discretion is left to the judges, though? Otoh, how inclined are the judges to exercise discretion?
posted by PurplePorpoise at 10:25 AM on January 2, 2005

First link doesn't seem to be working on the New York Times Link Generator...
posted by runkelfinker at 10:26 AM on January 2, 2005

OK, why does this make sense? (remember that higher scores = more severe punishment):

10. Most Serious Victim Injury
(for index offense, the most serious injury is scored)
Death -2
Hospitalized 0
Treated & Released +1
None or slight +2

(page 32 of the PDF)

Why would you PUNISH offenders who don't injure their victim?
posted by aberrant at 10:33 AM on January 2, 2005

Leaving aside, for discussion of this question, the fact that the sexual assault in itself is a fairly severe injury to the victim.
posted by aberrant at 10:35 AM on January 2, 2005

This article is just... ugh. Segue from the Minority Report name dropping to an account of how "for years, this was junk science! But not anymore - trust me, I'm a journalist, I can tell." Then we move onto the confidence-boosting fact that this new system was born from a budget crunch. As if the fact that judges are now resorting to the equivalent of a Cosmo personality quiz to do their sentencing for them wasn't worrying enough. You'd figure we'd at least stop paying them as much.
posted by mek at 10:41 AM on January 2, 2005

The point of this isn't punishment, it's prevention.

Figure out who is most statistically likely to commit a future crime, and lock 'em up for longer so they have less opportunity to commit the crime.

I don't agree with it, but I think that is the intent.
posted by selfmedicating at 10:43 AM on January 2, 2005

"Still, if the state has a compelling reason to discriminate based on age and sex, it sometimes can. And public safety is almost always deemed to be such a reason."

Public safety? But these are nonviolent crimes they're talking about.
posted by The Great Big Mulp at 10:45 AM on January 2, 2005

aberrant, agreed, number 10 seems meaningless; and I don't know what a DSM-III Diagnosis is, but numbers 3 and 9 look like they contradict each other...
posted by runkelfinker at 10:48 AM on January 2, 2005

So can they be like major corporations and trade points for crimes they could have committed, but didn't? "Well, I didn't murder anyone from 1982 to 2003, so that's gotta be worth five or six points right there!"
posted by XQUZYPHYR at 11:04 AM on January 2, 2005

Mulp, the PDF deals with sexual offenses -- hardly nonviolent crimes. I think what they're getting at is whether the offender has a history of other nonviolent crimes, and developing a methodology to base sentencing for sexual offenses taking that history into account.

of course, I could be way off....
posted by aberrant at 11:06 AM on January 2, 2005

runkelfilter, the DSM-III is the standard diagnostics manual for psychiatric disorders (actually, the DSM-IV and its variants are the most recent). It has its problems. I think it would be pretty easy for anyone who commits violent rape to get marked off with a personality disorder, meaning pretty much everyone would be getting +3 for this question if they bothered to go get diagnosed.

Note that in the "other factors considered" section they take "race of offender" and "race of victim" into account.
posted by painquale at 11:09 AM on January 2, 2005

I think any judge in a sentencing situation does this kind of +/- stuff in his/her head anyway. All this number crunching is a constructive exercise, and the the fact that the judges are not required to follow the guidelines set forth make it a worthy endeavor. It's a tool in a judges toolbox, nothing more.
posted by boymilo at 11:10 AM on January 2, 2005

Just so I'm not vague... it looks like they considered and analyzed race, found it didn't make any statistical difference in rates of recidivism, and so didn't include it in their test. The Cosmo +/- quiz is kind of offputting, but it does look like some real statistical analysis went to finding out what types of people are likely to rape again. If done well, information like this can only be a good thing, though it's probably not right to let it dictate our sentencing.

More interesting is that "difference in age between offender and victim" showed no difference in rates of recidivism. In other words, pedophiles are no more likely to rape again than normal rapists. This is actually a weird result... pedophilia is usually touted as a disorder that one just cant help, so one would think if that a pedophile were willing to rape, they would rape again. Maybe I'm missing something and this report does not apply to child rape?

Seeing as the resolution passed in the Senate that led to this report (see page 13) specifically mentioned protecting children, many were probably not happy with this result. The senators wanted harsher sentencing for child abusers who were threats, but the statistics didn't give it to them.
posted by painquale at 11:31 AM on January 2, 2005

aberrant: Ah, true. I was referring only to the article, as I hadn't yet read the PDF.
posted by The Great Big Mulp at 11:36 AM on January 2, 2005

Although, from the table on page 20, it looks like they only considered a single study that analyzed child molestation (Schram and Milloy (1995)), and according to page 18, that report included juveniles. I really don't know what to make of this data. Prentky, Knight, and Lee (1997), and Rice, Quinsey, and Harris (1997), on the other hand, consisted of only child molesters, and they seem like they might show a higher rate of recidivism than other reports (this information is hard to eyeball from the scanty information given). I assume that those reports didn't consider child molestation a variable because they internally had no non-child molesters to compare the data to. The big list on pages 20 and 21 doesn't compare the results of individual tests with one another.

I have a feeling that the report might have systematically missed a potentially important variable. But maybe not; maybe rates of recidivism for child molesters really are the same as those for rapists. I just don't know. Anyone have any outside information?
posted by painquale at 11:54 AM on January 2, 2005

As a Virginia resident, I feel I must say:
Oh man, time to get a wife.
Or job.
posted by dougunderscorenelso at 12:06 PM on January 2, 2005

Aberrant, I'm utterly confused. Do sexual offenses in your mind really equate to violent crimes? Like all of them do? Most do?
posted by wolftrouble at 12:25 PM on January 2, 2005

Offender's relationship to victim:

Stranger: 15%
Other: 1%
Unknown: 3%.

That's 19% of cases where the offender does not know the victim, or the victim was unable to recognize them.

In my opinion, that means that chances are 9 to 1 that someone you know is going to touch you in the bad place. If I lived in Virginia, I'd be married, employed and ESTRANGED from everyone I know.
posted by jsavimbi at 12:53 PM on January 2, 2005

Selfmedicating: The point of this isn't punishment, it's prevention.

Yes, I thought so too. This is why I found it surprising that the following issue isn't addressed at all: How does time spend in jail affect the chances of committing another felony?

In other words, if the jobless single young male gets away with a slap on his wrists, he just might find a job after all. But his chances of finding a job (and staying out of trouble) will be severely reduced after serious jail time. Thus, I would expect a correlation between time served for non-violent offenses and future criminality.

Thinking this through to the end, if the only goal were future crime prevention, then you might just wind up with a scheme where the statistically "best" way to fight future crime is letting everyone go home on parole below a certain threshold (when their personal history indicates chances of future criminality of lower than 50%) and locking them away for good above that threshold.

Of course, jail is not only about future crime prevention, but also about deterrence.
But I'm surprised that the proposed scheme seems to be missing this angle entirely.
posted by sour cream at 1:18 PM on January 2, 2005

In the days before mandatory minimums and sentencing guidlines, this was how sentencing was always done -- albeit relying upon a judge's gut rather than upon a series of studies and formulae.

I really do hope that liberals don't manage to frustrate this program. When they forced a shift to determinate sentencing once before, they defeated their own point. Voters won't accept laxity for dangerous offenders, so, deprived of the ability to determine who is dangerous, they rightly insisted upon denying laxity to all. Bad for the tax bill, bad for offenders who were clearly less of a danger to society, but couldn't access leniency.
posted by MattD at 1:47 PM on January 2, 2005

Aberrant, I'm utterly confused. Do sexual offenses in your mind really equate to violent crimes? Like all of them do? Most do?

I'm confused that you would find this confusing. No, really. When have sexual crimes ever not equated to violent crimes? I mean, except in Saudi Arabia maybe, where they're considered much worse...
posted by Jimbob at 2:02 PM on January 2, 2005

the problem with this is that it will create the perception that one gets put in jail for what one is, rather than what crime was committed ... that could be a dangerous perception for people to have
posted by pyramid termite at 3:02 PM on January 2, 2005

I see this as requiring that the penal system take a more active role in rehabilitation. They seem to be saying that the 40yo woman doesn' t need rehabilitated as much as the 20 yo man. The subsequent drop in prison costs could therefore lead to an increased investment in counseling, education, and vocation training.

Not that I think this will happen. Just pipe dreams probably.

Oh - and sorry I could only produce the sex crimes stuff. I'm not sure how that figures into the nonviolent category, but it's what I could find.
posted by jmgorman at 3:27 PM on January 2, 2005

that means that chances are 9 to 1 that someone you know is going to touch you in the bad place.

I think it actually means that if someone sexually assaults or abuses you, the chance that it was someone you know is 85%. To figure out the risk of anyone being assaulted by someone ze knows, you would have to know the percentage of the population that is assaulted.
posted by hippugeek at 3:58 PM on January 2, 2005

(Commonwealth of VA.)
posted by Hankins at 4:09 PM on January 2, 2005

selfmedicating:

The point of this isn't punishment, it's prevention.

Figure out who is most statistically likely to commit a future crime, and lock 'em up for longer so they have less opportunity to commit the crime.

I don't agree with it, but I think that is the intent.

And a good reason NOT to agree is that it can only reinforce the trends that it already observes. Are blacks (or teenagers, or men, or whatever) statistically more likely to re-offend? Well, if one of the reasons for that is that being black and having served time make it really hard to get a job and reintegrate into society, then that will only be worsened by these guidelines.

If we had some kind of genuine rehabilitative program that people could be directed to based on their likelihood of needing rehabilitation, I would (tentatively) be for that. But our present prison system is manifestly no such thing.

I really do hope that liberals don't manage to frustrate this program.

I kind of hope reasonable people of whatever political persuasion do frustrate it, for the reasons I just mentioned.
posted by rkent at 4:17 PM on January 2, 2005

This proposal completely misses the effect of the experience of jail time itself on recidivism. Once jailed, for any reason, a person will become more likely to be jailed again, for any reason.

Consider a first-time offender (actually a large set of first-time offenders, averaged). If he is caught, his probability of re-offending is p1; if he is not caught, it is p2. It is fair to expect that p2 > p1. Now assuming he is caught, if he is jailed, his probability of re-offending is p3. If he is not jailed, his probability of re-offending is p4. I expect that p3 > p4.

The problem, as usual in these discussions, is jail itself. It's a stupid idea to confine criminals to one area without anything productive for them to actually do. It costs a huge amount of time and money to make it happen, and its effects on criminals are generally to make them worse criminals, thus on the average increasing crime in the community. Getting out of jail is not the end of the punishment; a released prisoner is subjected to extrajudicial punishment in the form of social stigma. Drastically reduced employment opportunities, etc.

Fines are an equally stupid idea, as the economic impact on the offender is directly dependent on the offender's personal means, and any dependents of the offender (as with jail) are innocent parties who share in the punishment.

A just method of dealing with prisoners would meet four objectives: (1) an appearance of karmic balance, in that it involves a "bringing low" or "shaming", which is to vary in degree according to the public perception of the severity of the crime; (2) preservation of free will, which is necessary to keep the idea of personal responsibility, which justifies karmic balance; (3) rehabilitation of the individual criminal into a non-criminal, as much as is possible; (4) aggregate reduction of further crime in the community.
posted by aeschenkarnos at 4:20 PM on January 2, 2005

I don't think I want "karmic balance" to play any role whatsoever in my models of punishment.

Nor do I think we actively have to try to maintain the preservation of free will... that will just take care of itself. Judge Posner has been guest blogging at philosophy professor Brian Leiter's blog this week, and has made some pretty interesting comments on this. (Not just interesting, I think, but correct. If something makes use of Quine, it's probably correct).
posted by painquale at 4:39 PM on January 2, 2005

Not all "sex offenses" are violent crimes--statutory rape (though wrong wrong wrong in my opinion) is, by definition of the charge, non-violent.
posted by Sidhedevil at 4:39 PM on January 2, 2005

painquale: I don't think I want "karmic balance" to play any role whatsoever in my models of punishment.

Maybe you're not happy with the squishiness of the language (nor am I), but you almost certainly want something to the effect that aeschenkarnos described if you're interested in incentives (as I presume you are from that linked Posner post). The point of shaming/shunning/fining is to provide a personal disincentive to the criminal.

Likewise, the important part of aeschenkarnos's point 2 is the "responsibility" part, regardless of what we think of the possibility of free will. The problem with prison is that it says to an offender "the next X years of your life are going to suck no matter what you do." This causes acting responsibly to become futile, a behavior and attitude which can be deeply learned in prison and reinforced later by the same social stigma aeschenkarnos brought up earlier in his post. Hence, the punishment imposed on an offender should provide a disincentive for the original bad behavior, but it should also provide immediate, constant incentives to return to normal, socially useful behavior.

I don't feel any particular need to save aeschenkarnos's post, really, but it kind of seems like if you tossed out the "karma" talk and sprinkled on some "utility", you guys aren't thinking that far apart.

What Quine are you into? I remember being unimpressed with The Web of Belief in undergrad, but it might have been because our prof kept bragging about her PhD studies with him rather than discussing his arguments.
posted by rkent at 5:08 PM on January 2, 2005

Is that so, Sidhedevil?

I thought that statutory rape means an adult having sex with a minor. For this, there are two possibilities: the sex is consensual (i.e. "non-violent") or not (i.e. "violent").

Thus, an adult raping (in the popular sense, i.e. violently) a minor, is still committing statutory rape, no?
posted by sour cream at 5:09 PM on January 2, 2005

It's extraordinarily unlikely that a jurisdiction would bring a statutory rape charge if an adult committed a violent rape against a minor, but I suppose anything is possible (if they were "charge-stacking" to get additional jail time or something).

I guess I should have said "someone who is convicted of statutory rape, and no other charges, is not necessarily a violent offender".
posted by Sidhedevil at 6:07 PM on January 2, 2005

Wolftrouble, I try to avoid extremes like "all" or "never", so no, I wouldn't say that "all" sexual offenses have to be violent. Certainly, sodomy is still against the law in several states, and is classified as a sexual offense, but that doesn't mean it's violent cases where two consenting adults practice it.

I would say, however, that *most* sexual offenses (that are prosecuted) have at least one unwilling partner, and by my definition, that makes it violent. I admit that I hadn't considered the statutory rape angle, but my opinion is that there is a very small age window between the parties that would qualify the encounter as non-violent.
posted by aberrant at 6:15 PM on January 2, 2005

I don't feel any particular need to save aeschenkarnos's post, really, but it kind of seems like if you tossed out the "karma" talk and sprinkled on some "utility", you guys aren't thinking that far apart.

I can agree with that.

Yes, the Quine I'm talking about is The Web of Belief's W. V. O. Quine. He's personally my favorite philosopher of the past century, or at least in the top three, so I loved seeing him cited. Gavagai!
posted by painquale at 11:45 PM on January 2, 2005

I admit that I hadn't considered the statutory rape angle, but my opinion is that there is a very small age window between the parties that would qualify the encounter as non-violent.

I was raped as a child. I also had a happy sexual relationship with a man more than 10 years my senior when I was 16-17. I can assure you that the experiences were very, very different.

Also, what about "sex offenses" like, say, having sex with a consenting adult partner in a public park? I know someone who was arrested for same and has to register as a sex offender. I'm not in favor of adults having sex in public parks, but I don't think of it as a "violent" crime in any way.
posted by Sidhedevil at 1:09 PM on January 3, 2005

Fines are an equally stupid idea, as the economic impact on the offender is directly dependent on the offender's personal means

But you can link the fines to income on an escalating scale to at least cut down on income inequalities, as is common in some jurisdictions.
posted by biffa at 5:00 AM on January 4, 2005