June 30, 2005

More of James Q. Wilson on Steven D. Levitt's Freakonomics

from the July issue of Commentary:

The abortion-crime connection [is] … symptomatic of a general disposition on the part of Levitt or his coauthor to avoid being thought politically incorrect. For just as this book's discussion of abortion ignores race, so its chapter on the gap in educational achievement between blacks and whites soft-pedals some indisputably important facts.

The chapter begins by considering how little influence parents may exert over their child's personality given that half of the difference among personalities can be attributed to genes. This is quite correct. But genes also account for well over half (in some studies, as much as three-quarters) of differences in intellectual ability. If we are explain the black-white gap in educational achievement, we cannot turn away from the fact that on average, African Americans have a lower IQ than white Americans.

There are, of course, many highly talented blacks and many really stupid whites.

[Using stylized data from the normal probability distribution, I've estimated that about 33 million non-Hispanic whites score below the average black on IQ tests, and about 6 million blacks score above the 100 million whites who are below the white average. -- Steve]

But these important individual differences are not relevant to explaining the average difference between black and white school achievement. That difference is not the product of racist innuendo; the matter has been measured for decades, often by means of tests that do not require the use of words.

It is true enough that black IQ scores have risen -- owing, one suspects, to improvements in the social condition of blacks over the last several generations. But the black white gap in educational attainment has not narrowed. In Freakonomics, the authors assert that this gap is the result of differences in incomes between blacks and whites. Such differences certainly exist. But income differences are themselves in large measure the result of differences in intelligence, so one cannot explain the gap in IQ-based school scores by "controlling" for income.

The best test of this was done by Sandra Scarr and Richard Weinberg. They looked for changes in the IQ scores of black children who had been adopted by white families, mostly middle-class and well-educated. Over a ten-year period, there was no significant gain in the IQ's of the adopted black children. (Not was there any gain in the IQ's of adopted white children.) The data strongly suggest that parental environment, even in well-to-do families, has only a modest and probably short-lived effect on educational ability.

If you bothered to look up Levitt's original paper on the black-white test gap, written with Roland G. Fryer, you would find that that the authors are indeed aware of the many other studies that have been done of this issue. But they also think that once one "controls" for socioeconomic status, black and white schoolchildren become "observationally equivalent." Observationally, perhaps, but not actually.

As they themselves note, moreover, the gap between black and white test scores increases as children get older, and this widening gap cannot be explained by socio-economic differences in the quality of the schools the children attend. Experts on genetics have long known that heritability increases with age, and so, as a result, will the average gaps in school achievement between white and black children.

Wilson belongs to the heroic first generation of neoconservatives, who were primarily heterodox social scientists largely concerned with explaining domestic issues involving race and ethnicity: Wilson, Moynihan, Glazer, Sowell, Herrnstein, Coleman, Banfield, and Murray. Basically, I'm an Old Neocon in that tradition.

By a strange process, the second generation of neocons became primarily propagandists for a foreign political party and largely gave up research and analysis.

By the way, I wrote about Levitt and Fryer's IQ paper last October:

http://www.iSteve.com/Oct04.htm#levitt

Steven Levitt "eliminates" the black-white test score gap! Levitt is the imaginative U. of Chicago professor who recently was named the top young economist in America...

Levitt has now turned his peripatetic attention to the IQ racial gap, with the goal of statistically showing that it can be eliminated just by adjusting for various environmental factors. In a May 2004 paper in the Review of Economics and Statistics called "Understanding the black-white test score gap in the first two years of school," Levitt and Roland G. Fryer Jr. claim

"As in previous data sets, we observe substantial racial differences in test scores in the raw data: black kindergartners score on average 0.64 standard deviation worse than whites. In stark contrast to earlier studies (including those looking at kindergartners), however, after controlling for a small number of other observable characteristics (children’s age, child’s birth weight, a socioeconomic status measure, WIC participation, mother’s age at first birth, and number of children’s books in the home), we essentially eliminate the black-white test score gap in math and reading for students entering kindergarten."

However, I have thought of a simpler way to "essentially eliminate the black-white test score gap" than the one they describe. First, though, let's think of an analogy for what they are doing. Say you were attempting to "essentially eliminate" the Japanese-Lithuanian child height gap. You could control for household environmental measures such as the length of the beds, the height of the ceilings, the average inseam length of pants hanging in the closets, the size of the shoes, etc. Or you could simply control for the height of the parents. Much simpler! Occam would approve.

Similarly, if you controlled simply for the IQ of the parents, I bet you wouldn't have to laboriously include so many proxies for the parents' IQ like socio-economic status, WIC participation (i.e., get free formula from the government for being poor), mother's age at first birth, number of children's books in the home, and birth weight.

This generated the usual indignant replies from Levitt's ex-students about how my reductionist crudities were unfair to the ineffably brilliant sophistication of Levitt's thinking, which they couldn't quite explain because you had to be there to fully experience its wonderfulness. Levitt looks like he's well on his way toward becoming one of those charismatic U. of Chicago professors whose students take his every idea on faith, like Leo Strauss and Allan Bloom.


My published articles are archived at iSteve.com -- Steve Sailer

No comments: