WSJ: The Bell Curve: The Inequality Taboo - race WSJ: The Bell Curve: The Inequality Taboo - race
The thing that caused the most rucous eleven years ago when Charles Murray et al. first published "The Bell Curve" was their finding that IQ differed by race and ethnic group. In particular, they found that Blacks in this country scored approximatly one standard deviation below the norm, and Jews, one standard deviation above. This was the extreme of political incorrectness, which is why, to this day, the book is cause for immediate scorn by those invested in the equality theory.
A couple of years ago, when the human genome project was significantly further back than it is now, scientists were telling us that the amount of genetic material that differs by race was insignificant compared to the human variations in the remainder of the genome.
However, since then scientists have come a long way. At present, they can predict within less than 1% error rate the race of a person based on their genes - or I should say that in better than 99% of the cases tested, the race predicted for a person matched their racial self-identification.
But the white/black IQ difference stubornly remains. It really hasn't changed in that eleven years. Interestingly though, it seems to be larger for mathematical reasoning, and smaller for verbal reasoning. And I should note that this racial difference remains even when controlling for income level.
However, knowing how controversial this whole thing is, Murray has suggested a way to test this. Start with the fact that the average African-American today is approximately 80% black and 20% white, but that there are significant numbers of people ranging all the way from 0% to 100% black, and that this can be determined fairly accuarately by genetic testing. His proposal then is to test both the IQ and the racial genetics of a population. Then, for each subject in the study, determine the percentage of black genes or ancestry from the genetic results. And finally, see if subpopulation means do drop accordingly as the percentage of black ancestry increases. The supposition is that if race really is a factor here, the curve will be fairly smooth, running from 100 down to 84 or so as the percentage of black ancestry runs from 0% to 100%.
Will this proposed study be done in the near future? I think it highly unlikely, given the size of the population that would have to be tested, that genetic testing must also be done for each participant, and, most importantly, the political incorrectness of the entire subject matter.
A couple of years ago, when the human genome project was significantly further back than it is now, scientists were telling us that the amount of genetic material that differs by race was insignificant compared to the human variations in the remainder of the genome.
However, since then scientists have come a long way. At present, they can predict within less than 1% error rate the race of a person based on their genes - or I should say that in better than 99% of the cases tested, the race predicted for a person matched their racial self-identification.
But the white/black IQ difference stubornly remains. It really hasn't changed in that eleven years. Interestingly though, it seems to be larger for mathematical reasoning, and smaller for verbal reasoning. And I should note that this racial difference remains even when controlling for income level.
However, knowing how controversial this whole thing is, Murray has suggested a way to test this. Start with the fact that the average African-American today is approximately 80% black and 20% white, but that there are significant numbers of people ranging all the way from 0% to 100% black, and that this can be determined fairly accuarately by genetic testing. His proposal then is to test both the IQ and the racial genetics of a population. Then, for each subject in the study, determine the percentage of black genes or ancestry from the genetic results. And finally, see if subpopulation means do drop accordingly as the percentage of black ancestry increases. The supposition is that if race really is a factor here, the curve will be fairly smooth, running from 100 down to 84 or so as the percentage of black ancestry runs from 0% to 100%.
Will this proposed study be done in the near future? I think it highly unlikely, given the size of the population that would have to be tested, that genetic testing must also be done for each participant, and, most importantly, the political incorrectness of the entire subject matter.
Labels: Science/Techology
11:19 PM Display: Full / Chopped / Footer
Display:
Full /
Chopped /
None
Display:
Chopped /
Footer /
None
Posted by
Bruce Hayden at 11:19 PM
1 Comments:
Let me suggest the following calculation. Toss 64 coins 1000 times and calculate the relative frequencies of the number of heads. It will be a bell shaped curve. Imagine a gene for IQ for every head and you have the basis for a calculation. In this distribution the total number of coins is (mean/Std Dev)^2 The mean is clearly 32 and the std dev 4. So that (32/4)^2=64.If genes for intelligence have different probabilities of occurance(For a head it is 1/2) and the contribution to intelligence varies you will get for the number of genes k*(mean/std dev)^2. Now look at the Black case. (mean=85, std dev ranges from 11 to 14 say 12.5) the no of genes =k*(85/12.5)^2 ie k*(6.8)^2.For whites Mean/std dev =100/15 = 6.67(the same essentially within a margin of error) One does not doubt the honesty of the researchers who obtained the Bell shaped curves especially since it appears to affirm the belief that both races are equal at least on the average. I would like an opinion on this approach.
Post a Comment
<< Home >>