.... Is what A.N. Whitehead would have to say about the conundrum in which you find yourself Saletan. It is a form of reification. Before I explain either of these two terms I need to explain one other indispensible tool of all scientific research, namely, abstraction. "Abstraction in mathematics is the process of extracting the underlying essence of a mathematical concept, removing any dependence on real world objects with which it might originally have been connected, and generalising it so that it has wider applications." (Wikipedia)

So, all the data you have cited is in fact a mathematical abstraction from the real world. In other words the data is used to create a model. The problem with models of course is that they are by definition gross simplifications of reality. The problem with simplification of a principle is that although a principle may be incontrovertibly true, the language that we use to describe it is never detailed enough to explain it fully. (So a man jumping out of a plane should fall at a rate which increases by 23 ft. per second every second. Conclusion: a man jumping out of a plane will die ... wrong! Our language has left out the caveat of a man jumping out of a plane without a parachute; or a man jumping out of a plane when it is airborne, etc.)

Abstractions are particularly useful in the social sciences because we don't have laboratories in the social sciences, and there's the rub. According to A.N. Whitehead, we get so accustomed or so (sucked in) familiar with our models that we begin to think of them as the real thing. We ascribe concrete values to abstract ideas and form real life conclusions based on these values which are nothing more than highly simplified abstractions from reality. Do they have any use? Sure they do, for people dealing with large numbers who are forced to make "best fit" decisions.

The problem is that people who deal with individuals, like doctors, policemen, social workers and judges cannot afford to use these generalisations.. What is called for is individual analysis. Let's take a department of child welfare in a mythical city for example. Even if 95 percent of all applicants to adopt have been found to be decent parents, the department would still be guilt of a gross dereliction to send off a child to live with an applicant without having made a proper screening.

For this reason, I would say that most of the data you handed out was of little use to us in real world situations, except when used in a cautionary manner. This is exactly how most (good) doctors use the data on genetic predispositions and susceptibility to certain medications, simply as a weak reminder of a likelihood that a person might have a particular problem. I say weak reminder because such information doesn't do much good for people whom it excludes. So my wife's cousin died of breast cancer some time back, because he was a man and it never occured to him that the carbuncle he felt under the skin near his armpit was a cancer and not just a swollen gland.

It all boils down, Saletan, to your essential humanity. Faced with the certainty that such generalisation will ultimately do harm to a significant percentage of all people of a particular type, (and even 1% of a thousand is significant), wouldn't you rather use other more traditional tools of assessment rather than race? What is the spirit that make a person want to callously use such an admittedly clumsy tool to assess an individual's performance? What use is it anyway?

(And don't think I don't understand why you want to re-introduce this tired topic to the discourse.)