March 30, 2002

"Analysis Finds Race Disparity in School Tests"

"Analysis Finds Race Disparity in School Tests" reports the NYT in what I call America's Longest Running Series of 'Dog Bites Man' Stories. Has anyone ever seen a newspaper article where a school district (not a single school) didn't come out with this general ranking of Asian and whites doing better on tests than Hispanics and blacks? The renegade statistician who calls himself La Griffe du Lion explains how to convert these kind of test scores into standard deviation terms here. In school districts all over the country, the white-black testing difference comes out to right around 1.0 standard deviations. La Griffe finds this same result in all sorts of different fields, such as the racial difference in rate of uncountable ballots in the 2000 Florida election. He calls the one standard deviation gap The Fundamental Constant of Sociology.

March 29, 2002

Why not put nuclear weapons missile defense interceptors?

Some Answers to My Dumb Question about why don't we put nukes on our missile defense interceptors:

- "At a recent discussion held in DC by Jim Pinkerton's group, Bradley Graham, who just published a book on the subject, said that setting off nukes over either one's own territory or someone else's territory was just TOO controversial. That's why the book was entitled "Hit to Kill," that's what they have to do."

- "Sure. That's how the proposed Sprint/Spartan system worked back in 1970 and it's lots more effective - specially when you consider that the X-rays reach out quite a distance in vacuum. But we might, in the process of attenuating a massive nuclear attack on the US, create a little fallout - wouldn't want that. Better to die. This is one of a big class of important technical decisions that are fundamentally nuts."

Hmmmhmmm ... maybe it wasn't such a dumb question after all? Another reader suggested:

- "Here's my take. No one would probably have a problem if low-level fall-out happened during a response to an ACTUAL ATTACK; but people (the press, international community, citizens) would be up-in-arms having this fall-out happening, perhaps, 20-50 times over the next 10-15 years during the necessary TESTING of the system."

Excellent point. Clearly, "hit to kill" is a terrific way to practice. If we get good enough to "hit to kill," our nuclear-tipped interceptors would be that much more effective. The problem is we aren't developing nuclear-armed interceptors. The solution would be to build them, but test them with proximity-fused non-nuclear high explosive charges.