The Tainted Sources of ‘The Bell Curve’

For all the shock value of its assertion that blacks are intractably, and probably biologically, inferior in intelligence to whites and Asians, The Bell Curve is not quite an original piece of research. It is, in spite of all the controversy that is attending its publication, only a review of the literature—an elaborate interpretation of data culled from the work of other social scientists. For this reason, the credibility of its authors, Charles Murray and Richard J. Herrnstein, rests significantly on the credibility of their sources.

The press and television have for the most part taken The Bell Curve’s extensive bibliography and footnotes at face value. And, to be sure, many of the book’s data are drawn from relatively reputable academic sources, or from neutral ones such as the Census Bureau. Certain of the book’s major factual contentions are not in dispute—such as the claim that blacks consistently have scored lower than whites on IQ tests, or that affirmative action generally promotes minorities who scored lower on aptitude tests than whites. And obviously intelligence is both to some degree definable and to some degree heritable.

The interpretation of those data, however, is very much in dispute. So, too, are the authors’ conclusions that little or nothing can or should be done to raise the ability of the IQ-impaired, since so much of their lower intelligence is due to heredity. Murray and Herrnstein instead write sympathetically about eugenic approaches to public policy (though they do not endorse them outright). It is therefore interesting that Charles Murray recently expressed his own sense of queasiness about the book’s sources to a reporter from The New York Times: “Here was a case of stumbling onto a subject that had all the allure of the forbidden,” he said. “Some of the things we read to do this work, we literally hide when we’re on planes and trains. We’re furtively peering at this stuff.”¹

What sort of “stuff” could Murray mean? Surely the most curious of the sources he and Herrnstein consulted is Mankind Quarterly—a journal of anthropology founded in Edinburgh in 1960. Five articles from the journal are actually cited in The Bell Curve’s bibliography (pp. 775, 807, and 828).² But the influence on the book of scholars linked to Mankind Quarterly is more significant. No fewer than seventeen researchers cited in the bibliography of The Bell Curve have contributed to Mankind Quarterly. Ten are present or former editors, or members of its editorial advisory board. This is interesting because Mankind Quarterly is a notorious journal of “racial history” founded, and funded, by men who believe in the genetic superiority of the white race.³

Mankind Quarterly was established during decolonization and the US civil rights movement. Defenders of the old order were eager to brush a patina of science on their efforts. Thus Mankind Quarterly’s avowed purpose was to counter the “Communist” and “egalitarian” influences that were allegedly causing anthropology to neglect the fact of racial differences. “The crimes of the Nazis,” wrote Robert Gayre, Mankind Quarterly’s founder and editor-in-chief until 1978, “did not, however, justify the enthronement of a doctrine of a-racialism as fact, nor of egalitarianism as ethnically and ethically demonstrable.”⁴

Gayre was a champion of apartheid in South Africa, and belonged to the ultra-right Candour League of white-ruled Rhodesia.⁵ In 1968, he testified for the defense at the hate speech trial of five members of the British Racial Preservation Society, offering his expert opinion that blacks are “worthless.”⁶ The founders of Mankind Quarterly also included Henry E. Garrett of Columbia University, a one-time pamphleteer for the White Citizens’ Councils who provided expert testimony for the defense in Brown v. Board of Education;⁷ and Corrado Gini, leader of fascist Italy’s eugenics movement and author of a 1927 Mussolini apologia called “The Scientific Basis of Fascism.”⁸

Mainstream anthropologists denounced Mankind Quarterly. “It is earnestly hoped that The Mankind Quarterly will succumb before it can further discredit anthropology and lead to even more harm to mankind,” G. Ainsworth Harrison wrote in a 1961 article in Man, the journal of Britain’s Royal Institute of Anthropology.⁹ Bozo Skerlj, a Slovene anthropologist who had survived Dachau, resigned in protest from his post on the editorial advisory board of Mankind Quarterly, saying that he had joined unaware of the journal’s “racial prejudice.”¹⁰ Undaunted, Mankind Quarterly published work by some of those who had taken part in research under Hitler’s regime in Germany. Ottmar von Verschuer, a leading race scientist in Nazi Germany and an academic mentor of Josef Mengele, even served on the Mankind Quarterly editorial board.¹¹

Since 1978, the journal has been in the hands of Roger Pearson, a British anthropologist best known for establishing the Northern League in 1958. The group was dedicated to “the interests, friendship and solidarity of all Teutonic nations.” In 1980, Pearson resigned from the ultra-right World Anti-Communist League in a struggle with members who said he was too far to the right.¹² But Mankind Quarterly didn’t change. Pearson published eugenically minded attacks on school integration by two American academics, Ralph Scott and Donald Swan, who were alleged to have pro-Nazi affiliations; reports on a sperm bank in which geniuses have deposited their superior genetic material; elaborate accounts of the inherited mental inferiority of blacks; and the fact that Jews first came to South Africa because its gold and diamonds were “attractive” to them.

Pearson’s Institute for the Study of Man, which publishes Mankind Quarterly, is bankrolled by the Pioneer Fund, a New York foundation established in 1937 with the money of Wickliffe Draper. Draper, a textile magnate who was fascinated by eugenics, expressed early sympathy for Nazi Germany, and later advocated the “repatriation” of blacks to Africa. The fund’s first president, Harry Laughlin, was a leader in the eugenicist movement to ban genetically inferior immigrants, and also an early admirer of the Nazi regime’s eugenic policies.¹³

The Pioneer Fund’s current president, Harry Weyher, has denied any Nazi or white supremacist connections. But the fund’s current agenda remains true to the purpose set forth in its charter of 1937: “race betterment, with special reference to the people of the United States.” In a letter in 1989, the fund proposed that America abandon integration, on the grounds that “raising the intelligence of blacks or others still remains beyond our capabilities.”¹⁴ The fund not only underwrites Mankind Quarterly and many other Pearson publications, but has also provided millions of dollars in research grants to sustain the “scholars” who write for it and serve on its editorial board.¹⁵

Which brings us back to Murray and Herrnstein. They cite in their book no fewer than thirteen scholars who have benefited from Pioneer Fund grants in the last two decades—the grants total more than $4 million. Many of The Bell Curve’s sources who worked for Mankind Quarterly were also granted Pioneer money.¹⁶

Most of The Bell Curve does not explicitly address the relationship between race, genes, and IQ—as Murray has taken great pains to point out. Rather, the book couches its arguments about the impact of IQ on social behavior in terms of class, mostly using examples drawn from data on whites. But in view of the characteristic overlaps between race and class in American society, the insinuation is that all the connections between social pathology and low IQ which the authors find for whites must go double for blacks. It is only after one factors in their argument that IQ itself is mostly inherited (however hedged that argument may be), that the racial connotations of their policy prescriptions become evident.

And many of The Bell Curve’s most important assertions which establish causal links between IQ and social behavior, and IQ and race, are derived partially or totally from the Mankind Quarterly—Pioneer Fund scholarly circle. The University of California’s Arthur Jensen, cited twenty-three times in The Bell Curve’s bibliography, is the book’s principal authority on the intellectual inferiority of blacks. He has received $1.1 million from the Pioneer Fund.¹⁷ To buttress Jensen’s argument, Murray and Herrnstein draw on a book edited by University of Georgia psychologist R. Travis Osborne (the book, co-edited by former Mankind Quarterly editorial advisory board member Frank McGurk, is also cited by Murray and Herrnstein as an authority on the link between low IQ and criminality: pp. 277, 339). Osborne, the recipient of $387,000 from Pioneer, once testified as an expert witness for plaintiffs in a federal suit to overturn the Brown v. Board of Education decision.¹⁸

Other scholars who have received substantial amounts of money from Pioneer include Robert A. Gordon, a Johns Hopkins sociologist cited by Murray and Herrnstein on the causal link between low IQ and black criminality (pp. 321, 327, and 338); Linda Gottfredson of the University of Delaware, cited on the disproportionate representation of lower-IQ blacks in the professions; and University of Pennsylvania demographer Daniel Vining, Jr., a former Mankind Quarterly editorial advisory board member, cited on incipient “dysgenesis,” or biological decline, in America, owing to the falling birthrate among the most intelligent members of society.¹⁹

2.

The tainted funding of some of the scholars Murray and Herrnstein cite does not by itself invalidate those scholars’ findings. After all, history is full of examples of scientists who were pilloried as crackpots in their own times but are hailed as geniuses today. However shocking it may be that some of Murray and Herrnstein’s sources have chosen to affiliate themselves with such organizations, their work—and those parts of The Bell Curve that draw upon it—must be judged on the scholarly merits.

Take the case of Richard Lynn. A professor of psychology at the University of Ulster in Coleraine, Northern Ireland, Lynn was particularly influential in guiding the two authors of The Bell Curve through their review of the literature. In the book’s acknowledgments, they say they “benefited especially” from the “advice” of Lynn, whom they identify only as “a leading scholar of racial and ethnic differences” (pp. xxv, 272).

Lynn is an associate editor of Mankind Quarterly, and has received $325,000 from the Pioneer Fund.²⁰ One of his articles expressed support for the view that “the poor and the ill” are “weak specimens whose proliferation needs to be discouraged in the interests of the improvement of the genetic quality of the group, and ultimately of group survival.”²¹ He has also written that the genetic mental superiority of the Jews may be a happy Darwinian byproduct of “intermittent persecutions which the more intelligent may have been able to foresee and escape.”²²

Lynn’s work is cited twenty-four times in The Bell Curve’s bibliography.²³ It is used to support three important claims: that East Asians have a higher average IQ than whites; that most immigrants come from groups with subpar IQs; and that the IQ score of blacks in Africa is “substantially below” the American black average. Each of these seemingly discrete claims has a key role in the formulation of The Bell Curve’s broader suggestions about the relationship among race, heredity, IQ, and social structure.

The assertion about inferior black African intelligence has particularly far-reaching implications. If it can be shown that low IQ predicts social ills such as crime, poverty, and unstable families, current views of Africa and of the sources of its tragic problems would have to be significantly revised. The finding would also support the claim that the IQ superiority of whites is genetic, because the African-American edge over blacks in Africa could be attributed to their admixture of white genes. (Murray and Herrnstein note pointedly that South African “coloureds” have about the same IQ as American blacks.) And lagging African IQ could also be taken to refute the claim that black Americans’ lower IQ is a legacy of racism—assuming, as Murray and Herrnstein put it, that “the African black population has not been subjected to the historical legacy of American black slavery and discrimination and might therefore have higher scores” (p. 288).

Setting up their discussion of Lynn’s data, Murray and Herrnstein contend that the comparison between black Americans and black Africans is a valid exercise because IQ scores have been found to predict job and school performance of black Africans as well as those of black Americans (p. 288). They also attribute the paucity of published estimates of an overall average IQ score for blacks in Africa to the fact that these scores have been extremely low—the implication being that researchers are reluctant to publish such politically incorrect findings (p. 289).

These assertions are based on a highly selective reading of the article Murray and Herrnstein cite to support them: a comprehensive 1988 review titled “Test Performance of Blacks in Southern Africa,” by the South African psychologists I.M. Kendall, M.A. Verster, and J.W.V. Mollendorf (p. 289). The main point of these three researchers’ argument is to question sweeping comparisons such as the one Lynn attempts, and Murray and Herrnstein repeat. The three South African psychologists write:

It would be rash to suppose that psychometric tests constitute valid measures of intelligence among non-westerners. The inability of most psychologists to look beyond the confines of their own culture has led to the kind of arrogance whereby judgments are made concerning the “simplicity” of African mental structure and “retarded” cognitive growth.²⁴

Given the host of environmental and cultural factors that hamper black Africans’ test performance, they also say, “one wonders whether there is any point in even considering genetic factors as an additional source of variance between the average performance levels of westerners and Africans.”²⁵

Nevertheless, Murray and Herrnstein venture an estimate of African IQ, drawn mainly from an article by Lynn that appeared in Mankind Quarterly in 1991. It should be noted, for a start, that the authors of The Bell Curve misreport Lynn’s data. They say he found a median IQ of 75 in Africa (p. 289). But in his article, “Race Differences in Intelligence: The Global Perspective,” Lynn said that the mean African IQ—not the median—was 70.²⁶

In any event, how did Lynn arrive at his number? First, he assembled eleven studies of the intelligence of “pure African Negroids,” drawn from different tests of several different peoples and widely varying sample sizes in the years from 1929 to 1991. Then, he decided which was the “best”: a 1989 study from South Africa. In this test, he says, 1,093 sixteen-year-old black students (who had been in school for eight years and were therefore familiar with pencil-and-paper tests) scored a mean of 69 on the South African Junior Aptitude Test. Finally, Lynn rounded this result up to 70, and declared it a valid approximation of black IQ in the continent of Africa as a whole.²⁷

This methodology alone invites skepticism. But Lynn also seems to have misconstrued the study. Its author, Dr. Ken Owen, told me his test was “not at all” an indication that intelligence is inherited. He blamed the low performance of blacks on environmental factors such as poorer schooling for blacks under apartheid and their difficulty with English. Owen said his results “certainly cannot” be taken as an indication of intelligence among blacks in Africa as a whole.²⁸

Lynn further defends his choice of 70 as a “reasonable” mean for Africa on the grounds that 70 was the median of the average IQ scores reported in the eleven studies he had found. This statistical artifact aside, his list of studies is dubious. It includes what he calls “the first good study of the intelligence of pure African Negroids”: an experiment in 1929 in which 293 blacks in South Africa were given the US Army Beta Test, and got a mean score of 65.²⁹

The test was administered by M.L. Fick, whom Kendall, Verster, and Mollendorf call an “extreme protagonist” of the view that blacks are inherently inferior to whites.³⁰ The Beta test, which was developed for illiterate recruits in the US military, shows blatant cultural bias. One question presents a picture of people playing tennis without a net; respondents are supposed to sketch in the net to get full credit. In 1930, just a year after the Beta test was given in South Africa, C.C. Brigham, who had been its leading proponent in the US, finally admitted that the test was invalid for non-Americans. Lynn does not mention this fact.³¹

Far from refuting the thesis that the legacy of racism is to blame for black Americans’ lower IQ scores vis-à-vis whites, as Murray and Herrnstein contend, Lynn’s data actually support it (to the extent they have any meaning at all). Of Lynn’s eleven studies, five were conducted in South Africa under apartheid (and one in the Belgian Congo in 1952).³² If any country oppressed black people more than the United States, it was South Africa. Indeed, as the modern South African psychologists now acknowledge, one of the main uses of IQ tests under apartheid was to provide “scientific” justification for that system.

The assertion of an East Asian IQ advantage over whites, though essentially a success story, also plays a subtle, but crucial, supporting role in The Bell Curve’s overall argument about the connections among IQ, social achievement, and race. Coming before the discussion of black-white differences, it helps prepare the reader to accept racial categories as units of social analysis. It also conforms to readers’ preconceptions, shaped both by the media and by everyday experience, about the amazing brilliance of Asian immigrants and their offspring.

The authors would seem to be on firmer ground invoking Lynn here, since his specialty is the inherited mental superiority of East Asians, or “Mongoloids,” as he refers to them. In Mankind Quarterly, he has contended that the Japanese “have the highest intelligence in the world.”³³ In an article in Nature in 1982, Lynn claimed the Japanese enjoy a ten-point IQ advantage over European whites, and that this difference is growing. He suggested that this helps to explain the postwar economic miracle in Japan.³⁴

But two American psychologists, Harold W. Stevenson and Hiroshi Azuma, pointed out in a rebuttal in Nature that the Japanese sample Lynn used was made up of children of relatively well-off urban parents—a fact Lynn failed to disclose in his article. Lynn’s result was thus fatally flawed: he had tried to compare this socially skewed sample with a much broader and more representative American one.³⁵ Murray and Herrnstein’s sole mention of this is a footnote: “For a critique of Lynn’s early work, see Stevenson and Azuma 1983” (p. 716).

At the opening of their section headed “Do Asians Have Higher IQs Than Whites?” Murray and Herrnstein seem to be struggling to salvage some meaning from Lynn’s data. The basic problem is the enormous difficulty of drawing conclusions about the relative intelligence of people who come from vastly different civilizations. They cite a string of Lynn’s comparisons that suggest East Asians are superior, but eventually back off, conceding that the various test results he has assembled are not really comparable. Finally, the authors note: “Given the complexities of crossnational comparisons, the issue [of relative East Asian-white-black intelligence] must eventually be settled by a sufficient body of data obtained from identical tests that are comparable except for race” (pp. 272–274).

Murray and Herrnstein write that they “have been able to identify three such efforts.” In the first, they say, “samples of American, British, and Japanese students ages thirteen to fifteen were administered a test of abstract reasoning and spatial relations”—the British and American students did far worse than the Japanese, naturally. In the second “set of studies,” they write, nine-year-olds in Japan, Hong Kong, and Britain, drawn from comparable socioeconomic populations, were administered the Ravens Standard Progressive Matrices. Once again, the British children lost out by “well over half a standard deviation” (p. 274).

Only by checking the footnotes (at the back of the 845-page book) can readers discover that the author of both these studies is Richard Lynn. With regard to the first case, The Bell Curve’s text leaves the impression that the tests were conducted with similar samples in the three countries at more or less the same time. This is not quite what happened, as one learns from reading the 1987 Mankind Quarterly article from which these data are drawn. Lynn and his assistants gave the test in 1985 to 178 Japanese children. The tiny sample was not checked to reflect the social makeup of Japan as a whole (some 57 percent of the test-takers were boys). The test-givers merely showed up at two schools, one rural and one urban, and gave the tests to whoever was present. Lynn then compared this result to results from an American test that had been given thirteen years earlier to 64,000 subjects screened for their representativity, and to the results of a test given in 1978 to a similarly representative sample of 10,000 students in Britain. His conclusion that Japanese children do better was arrived at by distributing extra points among the three groups to “adjust” for the time lag among the three tests. ³⁶

The second “set of studies” is in the same 1991 Mankind Quarterly article in which Lynn presented his claims about “pure African Negroids.” He says that a group of 118 Hong Kong nine-year-olds scored a 113 IQ, a sample of 444 Japanese children got a 110 IQ, and a sample of 239 British children got a 100 IQ. He asserts that all three samples were “representative” and drawn from “typical public primary schools,” as Murray and Herrnstein report. But in the article Lynn does not explain how he assured the “representativity” of the samples, or the “typicality” of the schools.³⁷

Murray and Herrnstein then go on to describe a third set of studies done by Harold Stevenson in Minnesota. In contrast to their seeming circumspection about Lynn’s identity, they mention Stevenson’s name in the main text of the book. As they note, he “carefully matched the children on socio-economic and demographic variables”—and found no difference at all between the IQs of Japanese, Taiwanese, and American children (pp. 274–275).

“Where does this leave us?” Murray and Herrnstein then ask. On the one hand, we have two methodologically dubious studies by Lynn, a professor who believes, as he wrote in the Mankind Quarterly article, that “the Caucasoids and the Mongoloids are the only two races that have made any significant contribution to civilization.” ³⁸ On the other hand is a rigorous study by a social scientist with no known axe to grind, who finds no IQ disparity between whites and Asians. But Murray and Herrnstein portray this as a debate among a large number of contentious and equally reputable experts. “We will continue to hedge,” they write; and simply split the difference. They venture that East Asian IQ exceeds that of whites by three points, a figure which “most resembles a consensus, tentative though it still is” (pp. 276).

By the time Murray and Herrnstein get around to talking about immigrants, their “tentative consensus” on the East Asian-white IQ gap has grown by two points, and hardened into a datum firm enough to be factored into immigration policy. Drawing, once again, on Lynn’s 1991 article in Mankind Quarterly, they assign East Asians a mean IQ of 105, whites 100, “Pacific” populations a score of 91, and blacks 84. Without reference to Lynn or any other source, Murray and Herrnstein give “Latinos”—a designation empty of meaningful “racial” content—a mean IQ of 91. They give no data on IQs of South Asians and Middle Eastern people, who supplied 11 percent of the immigrants in the 1980s. They’re just “omitted from the analysis,” as the authors put it. From this hodgepodge of assumptions Murray and Herrnstein produce the “basic statement” that 57 percent of legal immigrants in the 1980s came from ethnic groups with average IQs less than that of American whites, and therefore the mean for all immigrants is probably below that of all native-born Americans (pp. 359–360).

Even if their “basic statement” is true, it says nothing at all about the scores of the individuals who actually did immigrate to the US. Thus Murray and Herrnstein must deal with the common-sense notion that immigrants generally represent the brightest and most energetic members of their former societies, by virtue of their willingness to get up and go to the US. This the authors try to do by citing numbers from the National Longitudinal Survey, or NLSY. They find that foreignborn NLSY members had a mean IQ “.4 standard deviation” lower than the rest of the NLSY sample (p. 360).

But the NLSY began in 1979, as a survey of people who were fourteen to twenty-two years old at the time, and have then been re-examined and re-interviewed each succeeding year. Thus it has no bearing at all on people who arrived in the United States after 1979, when immigration from the third world reached its height—as Murray and Herrnstein themselves report. (Probably for this reason, the sample did not include a statistically significant percentage of East Asians.) The authors also acknowledge that the slightly poorer IQ performance of those Latino immigrants who were interviewed in the NLSY probably reflects their weak command of English. That normally improves in a few years, and IQ rises along with it. Finally, Murray and Herrnstein find that foreignborn blacks in the NLSY score five points higher than native-born blacks (p. 360)—a fact they are utterly at a loss to explain, perhaps because some of the immigrants must have come from Africa, and they have just finished alleging that black Africans are even stupider than American blacks.

“Nonetheless,” Murray and Herrnstein assert, “keeping all of these qualifications in mind, the kernel of evidence that must also be acknowledged is that Latino and black immigrants are, at least in the short run, putting some downward pressure on the distribution of intelligence” (pp. 360–361). One hundred eighty-nine pages later, this strained contention is used to justify their inclination toward a more eugenically minded—and, hence, restrictive—US immigration policy. Yet other than Lynn’s flawed survey, and their own bald assertion that Latinos have a mean IQ of 91, there is no “kernel of evidence” of the kind they refer to (p. 360).

3.

Murray and Herrnstein aren’t answerable for every belief of every member of the racialist crowd they rely on for so much of their data. (And they didn’t get any money from Pioneer.) Still, there are two matters on which their book and the intellectual mission of the men who founded Mankind Quarterly overlap: both sought to restore the scientific status of race, and to reintroduce eugenic thinking into the public policy debate.

The more pertinent issue here is full disclosure, or what used to be called intellectual honesty. Just as Murray blushingly covered some of his materials on the Delta shuttle, so The Bell Curve tiptoes around facts that might have an inconvenient influence on its readers’ evaluation of the book’s sources and data—not to mention the judgment of its authors in choosing those sources. Geoffrey Cowley of Newsweek, in a sympathetic review of the book, pronounced its scholarship “overwhelmingly mainstream.”³⁹ Would he have done so if Murray and Herrnstein had provided a full account of the provenance of their data? Indeed, would this heavily marketed book have achieved the same sales success and as much respectful press attention if it had leveled with readers about all of its sources?

There is no way to isolate the scholarship of Richard Lynn, and that of the other Mankind Quarterly contributors, from their racial and political views. Social science is not so easily insulated from ideology, as Murray and Herrnstein are quick to emphasize when railing against their critics. The scholarly subcultures on which the authors of The Bell Curve depend for information are hardly less biased than those they are summoned to rebut. The bias of the Mankind Quarterly contributors, however, is much nastier. And as we have seen, some of the scholars Murray and Herrnstein rely on distort the evidence, which in key cases does not support The Bell Curve’s contentions.