August 30, 2005
Intellectual Garbage Pickup: Andrew Sullivan Needs to Do Some Remedial Work
www.AndrewSullivan.com - Daily Dish: One of my proudest moments in journalism was publishing an expanded extract of a chapter from [Herrnstein and Murray's] "The Bell Curve" in the New Republic before anyone else dared touch it. I published it along with multiple critiques (hey, I believed magazines were supposed to open rather than close debates) - but the book held up, and still holds up as one of the most insightful and careful of the last decade...
The impeccably right-wing Thomas Sowell certainly doesn't think so. And Thomas Sowell knows a hell of a lot more about these issues than Andrew Sullivan does:
Notes: What Thomas Sowell Thinks of The Bell Curve: Archive Entry From Brad DeLong's Webjournal: Strangely, Herrnstein and Murray refer to "folklore" that "Jews and other immigrant groups were thought to be below average in intelligence. " It was neither folklore nor anything as subjective as thoughts. It was based on hard data, as hard as any data in The Bell Curve. These groups repeatedly tested below average on the mental tests of the World War I era, both in the army and in civilian life. For Jews, it is clear that later tests showed radically different results--during an era when there was very little intermarriage to change the genetic makeup of American Jews.
My own research of twenty years ago showed that the IQs of both Italian-Americans and Polish-Americans also rose substantially over a period of decades. Unfortunately, there are many statistical problems with these particular data, growing out of the conditions under which they were collected. However, while my data could never be used to compare the IQs of Polish and Italian children, whose IQ scores came from different schools, nevertheless the close similarity of their general patterns of IQ scores rising over time seems indicative--especially since it follows the rising patterns found among Jews and among American soldiers in general between the two world wars, as well as rising IQ scores in other countries around the world.
The implications of such rising patterns of mental test performance is devastating to the central hypothesis of those who have long expressed the same fear as Herrnstein and Murray, that the greater fertility of low-IQ groups would lower the national (and international) IQ over time. The logic of their argument seems so clear and compelling that the opposite empirical result should be considered a refutation of the assumptions behind that logic....
A man who scores 100 on an IQ test today is answering more questions correctly than his grandfather with the same IQ answered two-generations ago, then someone else who answers the same number of questions correctly today as this man's grandfather answered two generations ago may have an IQ of 85.
Herrnstein and Murray openly acknowledge such rises in IQ and christen them "the Flynn effect," in honor of Professor Flynn who discovered it. But they seem not to see how crucially it undermines the case for a genetic explanation of interracial IQ differences. They say:
The national averages have in fact changed by amounts that are comparable to the fifteen or so IQ points separating blacks and whites in America. To put it another way, on the average, whites today differ from whites, say, two generations ago as much as whites today differ from blacks today. Given their size and speed, the shifts in time necessarily have been due more to changes in the environment than to changes in the genes.
While this open presentation of evidence against the genetic basis of interracial IQ differences is admirable, the failure to draw the logical inference seems puzzling. Blacks today are just as racially different from whites of two generations ago as they are from whites today. Yet the data suggest that the number of questions that blacks answer correctly on IQ tests today is very similar to the number answered correctly by past generations of whites. If race A differs from race B in IQ, and two generations of race A differ from each other by the same amount, where is the logic in suggesting that the IQ differences are even partly racial?
Herrnstein and Murray do not address this question, but instead shift to a discussion of public policy:
Couldn't the mean of blacks move 15 points as well through environmental changes? There seems no reason why not--but also no reason to believe that white and Asian means can be made to stand still while the Flynn effect works its magic.
But the issue is not solely one of either predicting or controlling the future. It is a question of the validity of the conclusion that differences between genetically different groups are due to those genetic differences, whether in whole or in part. When any factor differs as much from Al to A2 as it does from A2 to B2, why should one conclude that this factor is due to the difference between A in general and B in general?...
A remarkable phenomenon commented on in the Moynihan report of thirty years ago goes unnoticed in The Bell Curve--the prevalence of females among blacks who score high on mental tests. Others who have done studies of high- IQ blacks have found several times as many females as males above the 120 IQ level. Since black males and black females have the same genetic inheritance, this substantial disparity must have some other roots, especially since it is not found in studies of high-IQ individuals in the general society, such as the famous Terman studies, which followed high-IQ children into adulthood and later life. If IQ differences of this magnitude can occur with no genetic difference at all, then it is more than mere speculation to say that some unusual environmental effects must be at work among blacks. However, these environmental effects need not be limited to blacks, for other low-IQ groups of European or other ancestries have likewise tended to have females over-represented among their higher scorers, even though the Terman studies of the general population found no such patterns.
One possibility is that females are more resistant to bad environmental conditions, as some other studies suggest. In any event, large sexual disparities in high-IQ individuals where there are no genetic or socioeconomic differences present a challenge to both the Herrnstein- Murray thesis and most of their critics.
Black males and black females are not the only groups to have significant IQ differences without any genetic differences. Identical twins with significantly different birthweights also have IQ differences, with the heavier twin averaging nearly 9 points higher IQ than the lighter one. This effect is not found where the lighter twin weighs at least six and a half pounds, suggesting that deprivation of nutrition must reach some threshold level before it has a permanent effect on the brain during its crucial early development.
Perhaps the most intellectually troubling aspect of The Bell Curve is the authors' uncritical approach to statistical correlations. One of the first things taught in introductory statistics is that correlation is not causation. It is also one of the first things forgotten, and one of the most widely ignored facts in public policy research. The statistical term "multicollinearity," dealing with spurious correlations, appears only once in this massive book.
Multicollinearity refers to the fact that many variables are highly correlated with one another, so that it is very easy to believe that a certain result comes from variable A, when in fact it is due to variable Z, with which A happens to be correlated. In real life, innumerable factors go together. An example I liked to use in class when teaching economics involved a study showing that economists with only a bachelor's degree had higher incomes than economists with a master's degree and that these in turn had higher incomes than economists with Ph.D.'s. The implication that more education in economics leads to lower incomes would lead me to speculate as to how much money it was costing a student just to be enrolled in my course. In this case, when other variables were taken into account, these spurious correlations disappeared. In many other cases, however, variables such as cultural influences cannot even be quantified, much less have their effects tested statistically...
Posted by DeLong at August 30, 2005 03:32 PM