However, he does link to real government sources (near the botton), so you can check to see if he weighs by ideology when posting numbers.
Sometimes they end up in disagreement with me, but I always prefer informed and still wrong to uninformed and wrong[1]. :)
[1]Which is not to say these numbers are wrong or right. I'm speaking generally.
My point is simple. You have a group A of 10 million that on average score 450 on the math SAT. You have a group B of 10 milllion that score 500 on average. If you randomly move 5 million from group A to group B, then statistically, the average score of group B will fall to ~481.
Good job, you just discovered that our standardized tests scores are biased in favor of white kids.
A better way to show that standardized tests mean nothing is to have every student take them, but then store them away for the next 40 years. No one knows the results. And then in 40 years correlate the results of these tests with "life success". We've already seen with one similar measure "IQ test", which has little bearing on academic placement, that it turns out to have very little utility in predicting "life success". UPDATE: Link on IQ -- http://www.halfsigma.com/2006/06/high_iq_does_no.html
The problem is that a young black kid raised where many are raised, SAT scores not withstanding, likely won't have the opportunities that those I listed had. Let me give a glimpse of Michael Dell's early life, from Wikipedia: "The son of an orthodontist and a stockbroker, Dell attended Herod Elementary School in Houston, Texas. In a bid to enter business early, he applied to take a high school equivalency exam at age eight. In his early teens, he invested his earnings from part-time jobs in stocks and precious metals. Dell purchased his first calculator at age seven and encountered his first teletype machine in junior high, which he programmed after school."
If you go into the inner-cities, this type of early life, even from the bright and ambitious will likely look very different.
And lastly, Asians are disciminated against in a very different way than Blacks are. They are discriminated in the belief that they excel at math and engineering. I've seen first hand, in elementary school, how powerful expectations are (and there is good data around this too).
That article comes right out and says that IQ does not have an effect on life outcomes, if life outcomes are normed for educational attainment. It then goes on to say that IQ does not predict career success, if career success is measured by money and normed for career type. That's like saying air temperature has no effect on snow cover, if snow cover is corrected for season. And altitude.
Two grad students at CalTech were having a discussion about what Feynman's IQ must be. They bandied 180... maybe even to 200. When coincidentally Feynman walks right past them.
One of the students pipes up and meekly asks Feynman, "Professor Feynman, what is your IQ?"
Feynman turns around and responds, "120". The grad students have a look of disbelief. To which Feynman adds, "but its all in Physics".
But sticking with just the article. The point he makes is that its the educational attainment that people pay for, not the IQ score. Now you're saying, "But educational attainment is a product of IQ score". Possibly, but in controlling that factor you can attempt to tease out the extent at which it really is the educational attainment rather than simply the IQ.
A better way to put is that if IQ was in itself the key to life success then it wouldn't matter if you went to college or not. Rather, what the data seems to show is that those with high IQs seem to at least have the wisdom to know that educational attainment will be the key their life success. Quite possibly being tracked into by virtue of a higher IQ.
BUT the data should show that regardless of credential, higher IQ yields more success. But it doesn't show that.
Now there are a lot of factors that potentially come into play. And that was my original point. One can't look at SAT scores and then say, "those with high SAT scores do better in life", when there are so many comingling factors that complicate the equation, and are ridiculously difficult to control for in the real world.
Any ethnicity can have a scholarly culture, but for people starting out at the bottom, in a new society, it's less common.
There's also a selection bias against those who have not moved up in society. E.g. take a depressed area -- if the people with the gumption to improve their situation do so and leave, what's left? Right, concentrated fail.
For the change in average levels of education and knowledge, the cause seems pretty obvious to me. If you extended the border of California 400 miles south into Mexico, the average state scores on tests of math, history, and English would obviously go down, simply because of the composition of the population. Similarly, if you move 10 million Mexicans into California, the average test scores of the people living within the borders of California will fall (even if there is no change in the amount of learning for one individual person). In general if you look into correlations between education spending and test scores, there is very little correlation. If you look into the correlation of ethnicity and test scores, its very significant.