DMC is not devastating mic control in this case. It's Disproportionate Minority Contact - a regime that seeks to answer with statistical reporting the following questions.
- Are there differences in the rates of contact (e.g., arrest) based on race/ethnicity? If so, at what stages of the justice system are these differences more pronounced?
- Are there differences in the processing of juveniles within the justice system based on race/ethnicity? If so, at what stages of the justice system are these differences more pronounced?
- Are the racial/ethnic differences in contact and processing similar across jurisdictions within a state? If not, in which jurisdictions are these differences more pronounced?
- Are the differences in contact and processing similar across all racial and ethnic groups? If not, which groups seem to show the greatest differences?
- Are racial/ethnic differences in contact and processing changing over time?
Now here's the opening qualification taken directly from the same manual. I'm going to put it in bold so that you don't overlook it.
It is important to note what is not included at this stage: any attribution about the reasons for the differences. Therefore, the identification phase of information neither describes the reasons for any differences that occur nor creates strategies to reduce those differences.
In other words, although they can say with great precision that they are observing race, they cannot and will not say at all whether or not they are observing racism. So therein may be answers to what and perhaps how, but not why? Except that why is a presumption that plays into the politics of counting noses by race anyway. Essentially people are invited to speculate why and your guess is as good as mine.
Me? I was trolling for data. It is my job actually to make the meaning of such numbers plain and accessible, so I may as well have some fun doing it. The problem is that this data is dirty. They don't say that in so many words, they say it with too many words. Take the following paragraph as an example:
Studying More Jurisdictions and More Categories of Youth and Offenses
States may use the basic RRI method described above to extend the number of jurisdictions to be studied, subdivide the types of youth being studied, and subdivide the types of offenses (and other features) being studied to broaden their analysis of DMC issues. Each such refinement adds analytic power and specificity to the search for ways in which to address DMC issues. A few examples of such refinements would include separate identification analysis for males and females or for older and younger age groups. The logic that jurisdictions might use to justify such endeavors would be that there is some additional contact risk that attaches to younger (or older) male youth. Likewise, jurisdictions might add additional stages to the basic RRI model to track the implementation of specific additional statutory provisions such as the application of determinate sentencing or of automatic transfers to adult court for some offenses. For such policies to be fruitful for analysis, states would have to demonstrate that the policies actually apply to a substantial number of youth. In a similar fashion, it might be feasible to conduct the RRI analyses separately for various classes of offenses, such as those involving crimes against persons, property, drug offenses or public order. Again, the need is to ensure that a sufficient number of cases are processed to make the search for patterns potentially fruitful. If one is engaged in analysis of subsets of offenses, it is also necessary to recognize that the processes of plea-bargaining and diversion programming may lead to situations in which the classification of an offense changes as the case proceeds through the systems.
In short they know race but they don't know gender. They also don't know crime, nor do they have a good taxonomy for the crimes. They don't know age, nor do they have a taxonomy for aging. They don't have attributes for charges or sentencing.
Now it's true that a brother like me gets 250 an hour building analytical systems. Now you know why I get no municipal government work. Their data is weak. You cannot make sound analytical decisions on data this dirty and arbitrarily qualified. I know that sounds like a dismissal but you do grow a sense about these things after 20 years in the business. More's the pity. It almost wants to make me join Connorly.
Connorly's quest of eliminating all racial data collection is fraught with the peril of knowing to little and disabling analysis altogether. Yet there is the peril on the other side which is that of 'knowing' too much about very dense and well-qualified data sets. These aren't data these are people. And as much as I'd love to march every human on the planet through a 48 byte universal identifying system I know that runs the serious risk of treating people like things we think we can all too easily abstract. Of course there are greater risks in the world, and somehow I think we'll end up doing that anyway.
I'm for adding more and more data to a singly authenticatable person. This one of the reasons I don't blog anonymously. And I think people should be able to assume multiple pseuds which link (under their control) to their root, unchangeable one.
When you really recognize how difficult it is to get simple demographic information correct it makes you wonder how much we think we know about each other's digital information is just wrong, wrong, wrong.
Recent Comments