Stuart Buck points to this piece from Science Daily
Using data across a wide range of research, Sherman shows that most crime is committed by a small fraction of all criminals, at a tiny fraction of all locations, against a tiny fraction of all victims, during a few hours a week. By focusing police, probation, parole, rehabilitation, security and prison resources on these “power few” units with the most crime, the study shows how society could stand a far better chance at crime prevention without raising costs.
and asks:
I wonder how one is supposed to identify the "small fraction of all criminals" and the "tiny fraction of all victims" ahead of time, rather than in hindsight.
I wonder what we're supposed to do with them if we do manage to pick them out. Can one preemptively deploy the power of the state against people who seem likely to commit a lot of crimes? I sure hope not.
"... Can one preemptively deploy the power of the state against people who seem likely to commit a lot of crimes? ..."
This already occurs to some extent. Prosecutorial discretion means (among other things) devoting more effort to convicting people who are a continual danger to the community. Three strike laws (and other measures giving harsher sentences to repeat offenders) are an attempt take high frequency offenders off the streets. So are specialized organized crime or gang units.
Can one preemptively deploy the power of the state against people who seem likely to commit a lot of crimes? I sure hope not.
Wasn't this the plot of of that snooze-fest Minority Report?
But seriously, I think the fundamental idea, that there is a pretty strightforward method of crime prevention, is true. A very visible and engaged police presence in historically high-crime areas is really the key.
I think what Megan was asking about was something like CompStat whereby the police analyze which areas (based on past incidents) have higher rates of crime and deploy their resources accordingly. If you know that an area is more likely to have break-ins, you patrol the area more rather than just patrol each area equally. At least that’s how it worked on the District and on the Wire ;)
Isn't deploying the power of the state against people who seem likely to commit a lot of crimes more or less the theory behind attaching conditions to parole, preventing felons from legally buying guns, lighter sentences for first-time offenders and so forth? If the data are correct, I don't think you need any kind of morally troubling science-fictional style predicting of likely criminals before they've ever committed a single crime...you just wait until they've been found to be repeat offenders based on their prior convictions and you circumscribe their future freedom commensurate with how bad they've been so far. I'm just not alarmed by the idea that it's better to concentrate the resources of the state on monitoring, rehabilitating, imprisoning or what-have-you the worst apples of all rather than spread it more generally around the population of people once convicted, or worse, society at large.
focusing police, probation, parole, rehabilitation, security and prison resources on these “power few”
Too bad this isn't what is happening in airport security.
If you know that an area is more likely to have break-ins, you patrol the area more rather than just patrol each area equally.
Um, this does not stand a chance of passing the Diversity test.
focusing police, probation, parole, rehabilitation, security and prison resources on these “power few”
On the flip side, this seems somewhat contrary to the Broken Window theory....
It also fails to pass the laugh test; because as any True American knows, we need MORE police with MORE intrusive surveillance powers and discretion in using force, not less. Nothing less will suffice to keep us safe from the brown people who want to rob and kill us.
Why not? We've done it before. For that matter, we're doing it now with the subclass of all potential criminals called potential terrorists.
\For that matter, we're doing it now with the subclass of all potential criminals called potential terrorists.
How so, in particular? Airport security, for one, seems to go to great pains to avoid treating particular groups too harshly based on statistics, and instead treats everyone as potential terrorists-- and isn't that popular for it. I suspect that you're right that popular opinion would prefer more racial profiling (though perhaps without that name), and more than I'd like, but we certainly aren't getting that much, contrary to your claims.
Can one preemptively deploy the power of the state against people who seem likely to commit a lot of crimes? I sure hope not.
This was essentially my response to a Cornell professor who was attempting to claim that, based on brain chemistry, none of us are really responsible for our moral choices. For some reason, he seemed to think that this would make us easier on criminals. To the contrary; we're willing to be easier on a small minority of criminals who "can't help it" because we believe that most can. It seems to me that a society that believed that all crime was merely a result of brain chemistry, genes, and upbringing, but without personal responsibility would not shy away at all from preemptive punishment for crimes likely to be committed.
This discussion dovetails with your comments about why prosecutors and police don't necessarily want to rush to free people convicted "on a technicality" that "are likely guilty of something else or would do something else in the future anyway."
If the areas of likely crime are inhabited primarily by members of any minority group (based on race, religion, etc...) or the perpetrators of most crimes are more commonly members of any particular minority group, an effort like this would be shot down as discriminatory and would not be permitted.
A good first step would be to focus less effort on victimless crimes and more on violent crimes.
EI
I've wondered for a while why we can't have an equivalent to the Megan's Law database for crime in general, at least felony and violent crime. Surely there's a good case to be made that anyone who has seriously harmed others should be identified -- for the safety of others?
I don't have kids, so I don't personally care if I'm living near a child molester (except inasmuch as it might affect my resale value). But I'd certainly like to know if the guy up the street from me is a repeat drunk driver or has a murder conviction.
Nothing less will suffice to keep us safe from the brown people who want to rob and kill us.
I realize you were being sarcastic, but 54% of the murders in America are committed by "brown people", as you call them. If you count white Hispanics as "brown" -- and liberals usually do -- the percentage is higher still.
Half of the murders in the United States consist of black men killing other black men. Food for thought.
JT quoting LR: "'For that matter, we're doing it now with the subclass of all potential criminals called potential terrorists.'
How so, in particular? Airport security, for one, seems to go to great pains to avoid treating particular groups too harshly based on statistics [.] . . . we certainly aren't getting that much, contrary to your claims."
Um, what do you think wiretapping outside of FISA rules is if not "treating [a] particular group too harshly based on statisics"? And if you have any basis for believing that we "aren't getting that much", I'd sure like to see it.
I'd also like to see stats (beyond "9/11, 9/11, 9/11") that provide a reasonable basis for profiling that could realistically be applied by the TSA. Can you tell me with extreme accuracy who is muslim and who isn't by their appearance?
Christina wrote: Wasn't this the plot of of that snooze-fest Minority Report?
In fairness, it was worth one DVD rental. Paying thater prices would have been a rip-off, though.
Dan wrote: I've wondered for a while why we can't have an equivalent to the Megan's Law database for crime in general, at least felony and violent crime. Surely there's a good case to be made that anyone who has seriously harmed others should be identified -- for the safety of others?
Because the database would be overwhelmed with useless information.
One of my friends is an employee of Denver district courts, and every now and then decompresses by sharing job-related stories that are on the public record. One recent and spectacular tale was that of a guy who had an argument with his girlfriend which turned violent, and so he dumped her pet fish down the garbage disposal. The prosecutor charged him with, among two other lessser domestic violence counts, felony cruelty to animals.
IIRC everything associated with that one was plea-bargained down to a single class 1 misdemeanor domestic violence charge, but the fact that there are even laws which make a felony charge possible for that situation is, IMO, pretty much the height of judiciary idiocy. Your database of felony convictions would include crap like that and thousands of "caught with 2g of Mary Jane" entries. Good luck sorting through it.
Um, what do you think wiretapping outside of FISA rules is if not "treating [a] particular group too harshly based on statisics"? And if you have any basis for believing that we "aren't getting that much", I'd sure like to see it.
I'd like to see your basis for claiming that the NSA wiretapping program "treated a particular group too harshly based on statistics". The exact targets and nature of the wiretapping program haven't been made public, after all.
Can you tell me with extreme accuracy who is muslim and who isn't by their appearance?
Why is "extreme accuracy" called for? Isn't "increased accuracy" a worthy enough goal? And aren't you ignoring the possibility of profiling based on something other than appearance, such as name, accent, or nationality?
You're also overlooking the fact that profiling works two ways. You aren't just playing "spot the potential Muslim terrorist", you're also playing "spot the person who clearly is NOT a Muslim terrorist". Every minute spent wanding an old white lady from Dubuque is a minute not spent checking up on somebody who actually has a snowball's chance in hell of being a Muslim terrorist. Even if there was no possible way to know with near-certainty that someone was a Muslim there are plenty of ways to know with near-certainty that they aren't -- and eliminating those people from the pool lets you focus more effort on the remaining folks, increasing your chances of finding the bad guy.
One recent and spectacular tale was that of a guy who had an argument with his girlfriend which turned violent, and so he dumped her pet fish down the garbage disposal. The prosecutor charged him with, among two other lessser domestic violence counts, felony cruelty to animals. IIRC everything associated with that one was plea-bargained down to a single class 1 misdemeanor domestic violence charge
Um... if that's your example of "useless information" then I'm afraid I don't see your point. Anyone who would grind up a helpless living animal in a garbage disposal just because he was pissed off at its owner is one sick son of a bitch. Yeah, the fact that it was bargained down to an unrelated violent crime means that I'm not getting a clear picture of what the guy did, but the fact that he's not someone you want to trust around defenseless people remains clear.
Besides, linking to the details of the crime they were actually convicted of -- or pled guilty to -- should be easy enough. Those are already public information, after all, just not public information that can easily be gotten at.
Your database of felony convictions would include crap like that and thousands of "caught with 2g of Mary Jane" entries. Good luck sorting through it.
First of all I said "felonies and violent crimes", a list that misdemeanor drug possession obviously isn't on. That being said, making a website that lets you search a database based on specific criteria (such as type of crime) is child's play, even for the government. Even if they lumped in every crime from mass murder to speeding tickets it wouldn't be unmanageable.
Um... if that's your example of "useless information" then I'm afraid I don't see your point. Anyone who would grind up a helpless living animal in a garbage disposal just because he was pissed off at its owner is one sick son of a bitch. Yeah, the fact that it was bargained down to an unrelated violent crime means that I'm not getting a clear picture of what the guy did, but the fact that he's not someone you want to trust around defenseless people remains clear.
I said "fish", not "kitten". Unless you're a vegan (in which case your views are definitely NOT representative of mainstream views regarding use and utility of animals), that stretches the definition of felony and cruelty to animals like a bowstring. Charges related the theft and destruction of property, sure. But that wasn't what was employed. It was the use of an axe to chop parsley.
Besides, linking to the details of the crime they were actually convicted of -- or pled guilty to -- should be easy enough. Those are already public information, after all, just not public information that can easily be gotten at.
So rather than the doing the hoofwork on the ocassion that there is a genunine risk to your safety, you'd like the details of everyone's criminal history neatly made available for your consumption in a Rolodex? You must have been devestated when police radios switched to digital spread spectrum.
First of all I said "felonies and violent crimes", a list that misdemeanor drug possession obviously isn't on.
Okay, specifying a precise quantity was evidently a mistake in this context. I wasn't referring to misdemeanor possession charges, I was referring to the slew of felony possession charges that, for all practical purposes, should have been misdemeanors with fines, not felonies with associated rights forfeitures and possibility of incarceration.
That being said, making a website that lets you search a database based on specific criteria (such as type of crime) is child's play, even for the government. Even if they lumped in every crime from mass murder to speeding tickets it wouldn't be unmanageable.
Actually, it's child's play for everyone except the government. Worse, good luck getting your name cleared if a transcriber ever makes a copy error, or if (as happens often) you have the same name as a criminal. Thankfully my name is reasonably unique, but I've had friends run into problems for having a common first name/last name/both that got associated with someone else's.
anony-mouse is obviously right here. Moreover, Megan's Law is the same thing. It basically serves to wind up hysterical citizens, and it turns people who are mostly harmless into pariahs. Most sex offenders are guilty of dropping their pants in public or downloading nasty photos online, and will never do anything else. If the law only applied to violent sex offenders, it would make some sense.
But that's got little to do with Megan's post. The answer to the post is that the police are entirely capable of concentrating on people who are likely to commit crimes, and the courts are entirely capable of prosecuting people who are likely to commit more crimes more harshly than people who probably just stepped over the line once. And they do.
I said "fish", not "kitten". Unless you're a vegan (in which case your views are definitely NOT representative of mainstream views regarding use and utility of animals), that stretches the definition of felony and cruelty to animals like a bowstring.
Huh? Fish feel pain too. Making something which CAN feel pain, feel pain, for no reason other than your own personal satisfaction, is the very definition of "cruelty". There's no stretching involved. You don't have to be a vegan to oppose the senseless killing of animals.
But even if it doesn't bother you, the fact is that it would bother me and a whole hell of a lot of other people, including -- I'm guessing -- approximately every last one of the 150 million women in America. Given that the guy is in fact a convicted criminal, which is in fact a matter of public record, giving those of us who care about that information convenient access to it seems perfectly reasonable. The right to privacy does not and should not cover the right to conceal past criminal activity.
So rather than the doing the hoofwork on the ocassion that there is a genunine risk to your safety, you'd like the details of everyone's criminal history neatly made available for your consumption in a Rolodex?
Of course not. I want them made neatly available for me on the Internet. :)
As for the comment about "genuine risk to my safety", part of the point here is for me to identify potential threats. If you are a thief and you live next door to me, my property is at risk. The only question is whether or not I know it. Why should I have to go down to the police station to find out that you're a thief? I deserve to know that fact; everybody does. If you have a problem with that, maybe you shouldn't have become a thief.
You must have been devestated when police radios switched to digital spread spectrum.
Accused criminals -- i.e., the people being discussed by police -- are presumed innocent. Convicted criminals are, by definition, guilty. The former have a reasonable right to privacy regarding the fact that the police suspect them of criminal activity. The latter have absolutely no right to privacy regarding the fact that they broke the law.
wasn't referring to misdemeanor possession charges, I was referring to the slew of felony possession charges that, for all practical purposes, should have been misdemeanors with fines, not felonies with associated rights forfeitures and possibility of incarceration.
The "slew of felony possession charges" that "should have been misdemeanors" are usually the result of harsher charges being plea-bargained down. Prosecutors are not in the habit of throwing a felony charge at every pothead who gets caught with a bag of weed. In cases where felony possession is among the charges filed, it is the most serious charge in less than 5% of cases.
Actually, it's child's play for everyone except the government.
No, it is child's play for the government, too. They already have a lot of sites that do that sort of thing, ranging from the existing Megan's Law sites to the US Census Bureau's table generator. I myself have written software to deal with bigger databases than the one I'm talking about here.
Worse, good luck getting your name cleared if a transcriber ever makes a copy error, or if (as happens often) you have the same name as a criminal.
You appear to be trapped in the 1970s. Court databases are highly accurate, aren't copied by "transcribing", and rely on more than just the name when matching a person to his felonies. Which is why every John Smith in America doesn't get informed of his 9372 outstanding warrants when pulled over by the cops.
Most sex offenders are guilty of dropping their pants in public or downloading nasty photos online, and will never do anything else. If the law only applied to violent sex offenders, it would make some sense.
Hm, I just punched in two local middle-class zip codes for here in San Diego. I got 40 hits and checked the first dozen. The least crime on the list was "lewd/lascivious conduct with a 14 or 15 year old child". Ten were for continuous child abuse, child molestation, or lewd/lascivious conduct with a child under the age of 14. The twelfth was for lewd/lascivious conduct with a dependent adult.
No people who just "dropped their pants in public". No internet porn wankers. Just ten child molesters, a Humbert Humbert wanna-be, and whatever the hell that "dependent adult" guy is.
I'd be interested to hear the evidence for your claim that most of the people on the registry are on it for innocuous reasons. Keep in mind that it might only be true for your state, since different states have different conditions for putting people on the list.
"Can one preemptively deploy the power of the state against people who seem likely to commit a lot of crimes? I sure hope not."
If the individual has not given the state sufficient reason to resort ot the criminal justice system, then there is a good chance that resorting to education and job training opportunities would be beneficial. They have the added benefit of being constitutional.
Every minute spent wanding an old white lady from Dubuque is a minute not spent checking up on somebody who actually has a snowball's chance in hell of being a Muslim terrorist.
Well, there have to be some searches made of the little old ladies, otherwise the bad guys will plant their weapons on them or in their carryon baggage. The TSA should be trying to keep explosives off planes, as well as tools that could be used to crash a plane or get into the cockpit. Everything else is political theater, because if there is an attempted takeover the crew will land the plane ASAP - a lesson of 9/11. The 3 oz /quart bag for fluids was actually a reasonable restriction - I had already realized I could get enough fluids on a plane in my carryon to build many dangerous devices.
BTW, when I travel with my mom, we have to budget extra time because she always gets wanded - she has two artificial knees and sets off the metal detectors.
This research is not that new. Back in college in the mid-70s, I was dual majoring in astronomy and sociology and took a SOCI course on "Deviant Behavior", which concentrated on crime and drug abuse. Some of the key facts I took away from the course: the vast majority of violent crime is committed by 15 to 25 year old males (Hmmm. Same as the high risk pool for auto insurance.), that career criminals aren't that hard to spot, and that US drug policy was seriously screwed up.
I asked in class one day: "Does this mean that the rise in crime rates will continue as the baby boomers move through this age group and then drop as they get older?" "Of course!" was the answer. From looking at FBI crime data, rates peaked in 1980, 15 years after the end of the baby boom, declined until the early 90s (there was a birth spike in the early 70s) and are dropping since 1993. In this case, demography is (partly) destiny.
dan: "I'd like to see your basis for claiming that the NSA wiretapping program "treated a particular group too harshly based on statistics". The exact targets and nature of the wiretapping program haven't been made public, after all."
Of course, the defense of wiretapping on the basis that we cannot know whether there has been any abuse because we don't know who has been wiretapped. Thank you, administration stooge.
How 'bout you provide a plausible explanation for the administrations refusal to comply with FISA that does not involve a too-intrusive government?
ech:
The interesting question that raises: Is it that the criminals between 15-25 are mostly not criminals by age 30 or so? Or is it that those inclined to serious crime have mostly managed to end up in prison, crippled, or dead by the time they get past age 25?
The main place this matters is on parole decisions. I'd love to see some scheme for providing someone the right incentives for predicting future likely crime. Like, list the people up for parole, and have insurance companies bid on, say, a $10M policy to be split between the state and their victims if they are convicted of another crime after being released. Then have the state make parole decisions based on minimizing the cost of those insurance policies (so they guy nobody will insure never gets out till his full sentence is done). You could even include in the policies requirements for conditions of parole--maybe the insurance company provides job training or something to decrease the chances of the guy going back to robbing convenience stores.
The goal of all this would be to get good decisions being made on parole, by giving people good incentives to make the decisions. You could even decide on parole based on the relationship between the cost to keep the guy in prison N more years and the cost to buy his insurance policy.
The interesting question that raises: Is it that the criminals between 15-25 are mostly not criminals by age 30 or so? Or is it that those inclined to serious crime have mostly managed to end up in prison, crippled, or dead by the time they get past age 25?
Those factors are part of the explanation.
In addition, by age 25 most men realize they aren't bulletproof and won't live forever, damping some of their wild behaviors. This is seen in auto insurance rates, which is why I mentioned it above.
Second, there is the effect of getting into a serious relationship - women are the great tamers of men, more or less. The need for a stable income with lower risk of incarceration leads to "going straight".
I'd also suspect a change in hormone levels, causing a moderation of anger.
This was the conventional wisdom is sociology circles in the mid-70s. It's possible that research has shown new factors at play. I know that the "crack wars" a few years ago strongly skewed some of the crime stats, especially murder rates in some urban areas.
It is interesting that the FBI stats show long term reductions in the violent crime rates to levels well below what they were in the 60s. Despite this, the popular perceptions is that crime is horribly out of control.
Lessee, the question is, how to identify career criminals. Flipping through my memory, I find an article in the San Francisco Chronicle last month about a gentleman who had been convicted of breaking into an auto for the purpose of involuntary property conversion, i.e. theft. Surprise, surprise, this gent was NOT unknown to the local strength, in fact, he had been arrested AT LEAST 50 times in the 10 YEARS he had been known to operate in the city. But this time, he got nailed good! Sentenced to the max, 3 years, which after time served and probable good-conduct (or whatever it's called these days) he was expected to have a restful 14 months in the county lockup.
The article went on to describe many many other gents with similar lengthy histories and similar null punishment records. Including mention of the typical number of HOURS between one being released and the next arrest.
The head copper in the relevant SFPD division was quoted at the end of the article, something like, "They are doing it for money, but THEY KNOW THE CONSEQUENCES!"
Oh-Kay, then. Consequences, I see now.
Face it: foresight is often not required to ID the "likely repeater". He has, to a 99% probability, already repeated many times, and this fact is known to the local justice system.
zaleriana,
Of course, the defense of wiretapping on the basis that we cannot know whether there has been any abuse because we don't know who has been wiretapped. Thank you, administration stooge.
I asked for support for your claim that the wiretaps were too-harshly targeting a group based on statistics. That you for letting me know that you had no evidence for the claim. I now know that your opinion can be ignored.
How 'bout you provide a plausible explanation for the administrations refusal to comply with FISA that does not involve a too-intrusive government?
The topic under discussion is "treating people too harshly based on statistics", not "intrusive government". Would you like me to provide two or three plausible explanations for the program that either do not treat anyone "too harshly" or don't rely on statistics?
(1): Intercepting all calls routed through the United States -- i.e., the majority of international calls -- and using an algorithm to scan them for keywords. This violates FISA (since the call is technically "in the United States"), but is harsh to nobody and doesn't target anyone based on statistics (although there may be statistical correlations between saying "I have the parts for the bomb" and being a terrorist, of course).
(2): Intercepting all calls made by people who have been called by known terrorists. This is also not "harsh" in any meaningful sense of the word, and relies on the fact that the person has been contacted by a terrorist, not on statistics.
(3): Intercepting all calls made to foreign countries. This may, perhaps, be considered "harsh", but wouldn't be based on targeting specific groups based on statistics.
There. Easy as pie.
"Can one preemptively deploy the power of the state against people who seem likely to commit a lot of crimes? I sure hope not."
This is a pretty naive question coming from someone who has lived in NYC. The NYPD has long given extra scrutiny and attention to young black men. The NY Times's Bob Herbert has complained about this, though arguably the black community has benefited the most from this extra attention and the resulting drop in crime rates.
It is interesting that the FBI stats show long term reductions in the violent crime rates to levels well below what they were in the 60s. Despite this, the popular perceptions is that crime is horribly out of control.
According to the FBI:
The violent crime rate per 100,000 people was 161 in 1960 and 473 in 2006. The property crime rate was 1726 in 1960 and 3334 in 2006. So, roughly speaking, you're about three times as likely to be the victim of a violent crime, and twice as likely to be the victim of a property crime, as you were in 1960. Crime did increase later in the 60s, but not to those numbers -- it reached around 325/100k violent and 3300/100k property. Furthermore, the late sixties and early seventies were ALSO a time when crime was "out of control" -- because of the Baby Boom we had a surplus of teens and 20-somethings, and that's the group that most criminals come from.
The perception that crime is very bad today is entirely accurate. What's wrong is the perception that it is getting worse. It isn't; it has been improving since the early 90s.