DUI on a curve

I am an inveterate reader of the Courthouse news that appears regularly in The Appalachian News-Express and among the features that catch my attention are the Police Blotter, and the indictments that are listed periodically. One thing I often notice is that whenever an arrest is made for driving under the influence (DUI), the offence will not always be the first one for the arrestee. In fact, it is not uncommon to find that the driver so cited will also be driving with a suspended license, and that the DUI offense may be the third, or even fourth for that person.

What I wonder about more than anything, aside from the fact that people are actually let near a car after their second, or even third arrests for DUI, is what percentage of DUI arrests these people constitute? And I have to wonder what measures are taken to punish such offenses, and to insure that those arrested are not repeat offenders? I mean, it would seem to me that a second, or third arrest for driving with a suspended license would tell the authorities that these people apparently are not too concerned with following the advice they’re being given during their court appearances.

Well, according to an article in The New Yorker magazine by Malcolm Gladwell, these repeat offenders may constitute only a small percentage of DUI arrests. The article, called Million-Dollar Murray, concerns problems such as homelessness, and how the nation deals with them, specifically the cost of managing such problems versus actually trying to solve them with unorthodox methods. And to suggest that perhaps the more unorthodox approach might sometimes work better that a traditional approach and actually solve the problem, the author cites both anecdotal and statistical evidence.

The title character of the story (whose choice of lifestyles has since led to his demise) was a homeless former Marine whose hard-drinking kept him on the street despite being given help on an on-going basis. The surprising thing the authorities discovered was that, while he was in a structured environment, Murray thrived. He held a job, and he didn’t drink. He even managed to save up respectable amounts of money. But the very minute he was let out of treatment, he went back to his old way of life, and in no time, he would be broke, and off the wagon.

The author goes on to say that everyone involved with trying to solve the homeless problem assumed that the homeless population had a normal distribution, or one that would mimic a bell curve. A bell curve is what it sounds like it is. It is a graph that represents a population that is grouped largely in the middle. The smaller parts of the population are represented at either end of the graph. So, trying to graph the period of time people are homeless, by the normal distribution, you would suspect that a small amount of people stay homeless for a short period of time, and a small amount are chronically homeless, while the majority are somewhere in between. A similar graph should emerge when it comes to graphing the amount of money it costs to deal with these people. There should be a few who are relatively inexpensive to help, some who are very expensive to help, and the majority would again be between the two extremes.

But what the author reveals is that not every long-term problem fits the normal distribution model. Take the homeless, for instance. The viewpoint has been for some time that the majority of street people are in a more semi-permanent state of homelessness than they actually are. The author cited the work of a Boston College graduate student, Dennis Culhane, who did research on the homeless problem in Philadelphia, and who compiled a data base on their comings and goings. It was Culhane who discovered that the homeless population does not have a normal distribution. In fact, according to Culhane’s figures, a full 80% of the homeless were very short term users of shelters. Another 10% were episodic users, showing up, usually only in the winters. The remaining 10% were the chronic users. This type of distribution, by the way, is referred to as power-law distribution.

In the case of costs, it is this last 10% that is also the most expensive to deal with. Murray, for instance, had, over the course of ten years, cost the state of Nevada over one million dollars in medical bills alone. The solution to taking care of these chronically homeless people, a viewpoint advocated by Philip Mangano, President Bush’s director of the Interagency Council on Homelessness, is to give these people an apartment paid for by the local agencies charged with dealing with homelessness. Given a structured environment, the most of these so-called hard corps homeless are easier to take care of, do better, and, in the long run, cost far less than the help the traditional approaches to homelessness provided, and with better overall results.

Back to the problem of those who habitually collect DUI arrests, I wonder if anyone has done any similar studies to determine what percentage get cited for that second or third offense? Does this problem have a normal distribution, or is it, like the problem of the nation’s homeless, a power-law distribution? And if there is a small number of repeat offenders who are not responding to the normal methods to keep them from drinking (or using drugs) and driving, has anyone attempted to find an alternative solution? By the looks of the Police Blotter, a lot of drunk drivers are being let out, only to be re-arrested, and at this point, perhaps an unorthodox solution of some sort may be called for.