"For many years now, too many scholars on left and right alike have pretended they live in a Lake Wobegon world where everyone can be above average. It is time for policy analysts to stop avoiding the reality of human inequality, a reality that neither equalization of opportunity nor a freer market will circumvent."

So concludes a thin monograph just published by the American Enterprise Institute under the title Income Inequality and IQ. The author? You guessed it, Charles Murray, who, along with the late Richard Herrnstein, wrote the incendiary 1994 bestseller The Bell Curve. Murray is back on message—chastened perhaps, but unrepentant. While studiously avoiding any claims about the innate intellectual inferiority of blacks this time around, he is still preaching his gospel about a social hierarchy based on sheer mental candlepower. He still claims that inherited human differences dictate disparities of income so stubborn that we can have a more equal society only if we tolerate government intrusions that would surely curtail liberty. "No realistic assessment of our empirical experience," he declares, "can yield grounds for concluding that our repertoire of social interventions, augmented with greater funding and energy, may be expected to narrow the national income inequality statistics."

I like that "our"—Murray having never met a "social intervention" he much cared for. But never mind. The fact is that "our empirical experience" compels no such categorical judgment. Murray persists in making sweeping claims of the sort that were effectively rebutted years ago, while ignoring the most serious criticisms of his earlier work. One cannot imagine his coauthor, Richard Herrnstein, who had been a respected professor of psychology at Harvard before his untimely death, behaving in so unscholarly a fashion. Indeed, one wonders whether Murray, who has never shown mastery of statistical technique in his work (his gift is exposition), has a sufficient grasp of the methodological issues involved to engage in an expert discourse.

Certainly, Income Inequality and IQ does nothing to dispel this suspicion. The monograph uses a subsample of the National Longitudinal Survey of Youth, consisting of matched pairs of siblings raised in the same households. The survey looks, first, at how these siblings fared on cognitive skills tests that they took when they were in their late adolescence. It then looks at how those siblings do in their late 20s and early 30s—that is, how much money they earn, how much education they receive, how much job prestige they gain. Murray uses these data to establish a causal relationship: By comparing persons with middling test scores to siblings whose scores were either much better or much worse, Murray finds that, among youngsters raised in the same household, those who did better on the cognitive test were substantially more successful later in life. Ergo, he says, innate cognitive differences explain inequality.

Of course, the story is a lot more complicated than that. Nobody disputes that people with better mental skills will, on average, perform better in our society. And, yes, this view is rather more broadly accepted now than it had been before the appearance of The Bell Curve. But Murray still hasn’t managed to answer the academic community’s chief complaint: His theory does not adequately account for the role social environment plays in determining one’s lot in life.

As you may recall, in The Bell Curve Herrnstein and Murray used an extremely crude measure of social background, then contrasted its effects with those of a test of mental skills given to most of the sample in late adolescence. Murray claims his new study answers this criticism: "The Bell Curve’s method of controlling for SES [socioeconomic status] and the sibling method of controlling for everything in the family background yield interpretations of the independent role of IQ on income that are substantively indistinguishable," he writes. But the test Murray uses, the Armed Forces Qualification Test, is not an environment-free measure of intelligence, so it does not identify "the independent role of IQ." Scores on the afqt have been shown to vary significantly with the quantity and quality of education to which a young person has been exposed.

Moreover, comparing siblings, while helpful, does not come close to "controlling for everything in the family background." Environments can differ within families, too—because of differences in the sex, personality, or birth order of the children, for example. In any case, Murray’s conclusion—that improving the environments of unrelated children will do little to reduce inequality—is a non sequitur. Finding a correlation between intelligence and success within families says nothing about the extent to which inequality in a population is driven by differences between families. After all, incomes are much more equal among siblings than among unrelated individuals, which attests to the equality-enhancing effects of a common family environment. Variance in IQ explains at most one-fifth of the variance of incomes; so, most inequality is caused by other factors. It is by now well-established that, holding ability constant, more education raises earnings, and well-designed, early childhood interventions can improve later-life outcomes for disadvantaged youths in a cost-effective way. But Murray seems utterly unfazed by these results.

Several months after The Bell Curve appeared, the University of Chicago arranged a workshop in which Murray was to face one of his most distinguished critics, the renowned econometrician James Heckman, in a discussion on the technical merits of his work. Murray backed out of that engagement. "I am canceling out of the session," he wrote to the organizer. "My experience of the last few months leads me to this position.... I will no longer deal with academics in groups." A few weeks later, Murray appeared at Harvard’s Kennedy School of Government, before a rather large group of academics, to debate a nonspecialist under a strictly regulated format. Murray obviously prefers nonexpert audiences that can’t ask him overly technical questions. Is it any wonder why?