February 22, 2005

Summers's logic shouldn't be overlooked

Lawrence Summers of Harvard University and Ward Churchill of the University of Colorado at Boulder have each sparked media fervor with remarks that were, to say the least, incendiary. Summers provoked liberals by suggesting that the gender disparity in tenured positions in math, science, and engineering could be due, in part, to innate differences in aptitude; Churchill provoked conservatives by suggesting that America deserved 9/11 and the terrorists behaved nobly.

It is notable, then, that the mainstream media have responded—I am speaking very broadly here, in aggregate; reactions have been quite heterogeneous—by condemning Summers's "insensitivity" but by defending Churchill's right to "free speech." Some liberals have justified this with the claim that, as Harvard's president, Summers represents the University and has a responsibility to watch his words, whereas Churchill, as a tenured professor, speaks only for himself. A number of conservatives have argued that the opposite reactions to Summers and Churchill is simply due to the media's liberal bias, that only liberal academics' speech is free.

While there is some merit to both arguments, I believe the difference is deeper than that.

Lies can be irritating; they can be unpleasant; they can cause difficulties. Unsupported and unsupportable statements can be offensive; they can be annoying. But the truth—the truth is deadly.

The fact is that no one is particularly concerned about Churchill—outraged, yes, but not threatened—because his remarks comprise extravagant hyperbole, dramatic metaphors, pure opinion, and clear lies. He believes the hapless victims of 9/11 were "little Eichmanns" while he praises the "gallant sacrifices of the combat teams" that committed the acts of terror. But he provides no new evidence to prove that America is wrong; he does not threaten our patriotic paradigm.

Churchill claims that settlers deliberately gave Indians smallpox-infested blankets specifically to spread the disease—a good two centuries before scientists began to reconsider the immaculate-conception, spontaneous-creation theory of disease origination and transmission. Churchill's claim is easily refuted; therefore, it is not dangerous.

In contrast, Summers's speech was dispassionate, technical, unexaggerated, and fairly equivocal. The furor over his remarks is because of, not despite, the legitimacy of his suggestions.

Suppose Summers had limited his remarks to anecdotes, sweeping generalizations, and inappropriate Bible quotations. Suppose the truck story had been the crux of his argument: "So, I think, while I would prefer to believe otherwise, I guess my experience with my two-and-a-half-year-old twin daughters who were not given dolls and who were given trucks, and found themselves saying to each other, look, daddy truck is carrying the baby truck, tells me something."

Feminists would have been upset, naturally, but there would not be extensive, outraged efforts to discredit and oust Summers. He could have easily been dismissed as a fanatical misogynist.

The truck story was, however, merely a tangential anecdote; the substance of his argument is scientific in nature. As he prefaces his speech, "I think it is important to try to think systematically and clinically about the reasons for under-representation." His discussion is indeed systematic and clinical.

Summers delves quickly into a statistical discussion: "It does appear that on many, many different human attributes…there is a difference in the standard deviation, and a variability of a male and a female population." He proceeds to look at the demographics of the top one-in-20 12th-graders, where there is one girl for every two boys, and extrapolates to a 20 percent difference in "implied standard deviations." Methodologically, using a purported difference in standard deviations as evidence for innate difference in abilities—he does not quite go that far, but the insinuation is there—is an extremely strong claim. There are legion possible explanations for differing standard deviations, some social, some discriminatory.

Further, the statistical extrapolation to the "20 percent difference" is, as Summers readily admits, very questionable. He is making claims about, as he explains, adults "three-and-a-half, four standard deviations above the mean." He makes these bold extrapolations based on data for the top five percent of 12th-graders—not even two standard deviations above the mean. Economists (like Summers) like to assume that everything is normally distributed, but when the analysis turns to 99.9- to 99.99-percentile data points, as Summers does, a lot of that falls apart.

Summers cites Chicago's very own Gary Becker for his second main point; he makes a very simple, very elegant economic argument. Summers explains: "If it was really the case that everybody was discriminating, there would be very substantial opportunities for a limited number of people who were not prepared to discriminate to assemble remarkable departments of high-quality people at relatively limited cost simply by the act of their not discriminating, because of what that would mean for the pool that was available." This argument applies to old-fashioned, explicit discrimination much better than the newer, insidious, socially determined variety, but as an argument against the existence of explicit discrimination it is devastating.

There are, in fact, scientific responses to his sweeping claims. The fact that the methodological arguments exist but that the mainstream media have not bothered to explore them underscores the fact that the media and his opposition at Harvard do not want to engage his argument on its merits. Rather, they are racing to discredit him as an insensitive misogynist. Summers' remarks are no more offensive than Churchill's. Summers is being attacked and Churchill is not because Summers' remarks are plausible; therefore they threaten the reigning gender-equality paradigm.