IQ, Race, and Heredity
To the Editor:
In “ ‘The Bell Curve’ and Its Critics” [May], Charles Murray does your readers a disservice by using the same standards of evidence and scholarship that he adopted in the book he wrote with the late Richard J. Herrnstein. By portraying himself as a persecuted scholar held to higher standards than other social scientists by his critics, he hints darkly at a unified conspiracy against him by like-minded but intellectually slovenly social scientists.
The academic response to his book does not present a united front. There is much more subtlety to the criticism of his work than Mr. Murray’s broad-brush summary conveys. It is surely disingenuous for him to lump serious academic criticism with the inflammatory journalistic reviews that appeared in the popular press. As one of the critics mentioned by name, and the first person thanked in the acknowledgments section of The Bell Curve, I wish to report that Mr. Murray does not mention or respond to any of my substantive criticisms of his work, nor does he mention any of the fundamental points of agreement that have emerged in the literature that responds to his work. (My survey is scheduled for publication in the October 1995 issue of the Journal of Political Economy, the house journal of the University of Chicago. A more popular version appeared in Reason magazine, March 1995.)
The basic premise of the book is that g—or a single factor of intelligence—explains behavior, and that it is immutable. An essential flaw in this argument is that it equates skill with general intelligence, contrary to the findings of a large body of research in social science. No one denies that “one factor” accounts for “a lot” of the variability in test score’s across persons. But using the methodological standards adopted by Herrnstein and Murray themselves, more than one factor is required to explain test scores and wages. In reality, as many as four factors are required to explain wages.
Mr. Murray’s appeal for support to the work of Raymond Carroll is misleading. Carroll’s reanalysis of test-score data does not support the single g model—contrary to the claims made by Mr. Murray (see p. 706 of Carroll’s Human Cognitive Abilities, 1993). In fact, Carroll finds evidence that multiple skills explain social performance, and account for correlations among tests.
None of this denies that ability tests predict something valued in society. But Carroll’s evidence and the evidence cited in my review indicate that many skills affect outcomes and not all can be equated with native intelligence. Once this is recognized, a core argument in The Bell Curve evaporates and a much more subtle analysis of social policy is required. Even if IQ cannot be manipulated, partial offsets for it are available because success in life depends on more than raw intelligence.
Mr. Murray seeks to avoid the hard job of evaluating social programs designed to boost skills by claiming that skills are synonymous with IQ; that IQ cannot be manipulated; and that nothing else can compensate for a low IQ There is ample evidence that Mr. Murray’s measure of “intelligence”—really a score on an achievement test—can be manipulated by educational interventions, that many skills besides raw intelligence are valued in society, and that these skills can be produced by environment.
James J. Heckman
University of Chicago
To the Editor:
Charles Murray’s brief attempted rebuttal of the article I published over six months ago in the New York Review of Books consists mainly of a distortion of my arguments and, even more unfortunately, a whitewash of the Pioneer Fund—the far-right organization which funded much of the pseudo-scholarship upon which The Bell Curve relies.
Mr. Murray writes that I “alleged” Pioneer was “established and run by men who were Nazi sympathizers, eugenicists, and advocates of white racial superiority.”
There is nothing alleged about it. The 1937 charter of the Pioneer Fund specified its goal as aiding “parents of unusual value as citizens,” who were defined as those “descended primarily from white persons who settled in the original thirteen states prior to the adoption of the Constitution of the United States.”
Wyckliffe Draper, whose fortune founded the Pioneer Fund, was an ardent eugenicist. The most ideologically influential of the fund’s five founding directors, Harry Laughlin, was superintendent of the Eugenics Record Office and campaigned in the U.S. for the sterilization of the “feeble-minded.” He served as honorary vice president (in absentia) of a eugenics conference in Berlin in 1935, and drummed up support in the U.S. for Nazi eugenics policy. It was Laughlin who persuaded Draper to undertake the Pioneer Fund’s first activity, in 1937: funding the distribution in America of an edited version of the German eugenics propaganda film Erbkrank (“Hereditary Illness”).
“Never mind,” Mr. Murray adds, that “the relationship between the founder of the Pioneer Fund and today’s Pioneer Fund is roughly analogous to the relationship between Henry Ford and today’s Ford Foundation.”
This is a spurious analogy. Yes, in 1918 Henry Ford published a series of anti-Semitic tracts in his Dearborn newspaper, including the Protocols of the Elders of Zion. Nine years later, he apologized (sincerely or not), retracted his statements, and shut down the newspaper. In 1936, he created the Ford Foundation to support the Henry Ford Hospital and other charitable activities in Michigan.
Henry Ford died in 1947; his heirs eventually surrendered all control over the selection of directors and the funding activities of the foundation. Not surprisingly, then, Henry Ford’s anti-Semitism and right-wing ideology have no lingering influence whatsoever over today’s Ford Foundation, which has an endowment of $6.8 billion and 400 employees worldwide. Ford grants, in fact, are frequently criticized for supporting left-wing minority organizations.
Today’s Pioneer Fund, by contrast, has made no apology, even a perfunctory one, for the extreme right-wing leanings of its founders. The only changes have been cosmetic, such as deleting the reference to “white” persons from its charter—in 1985. Such concessions to fashion aside, the Pioneer Fund remains true to the essential purposes of its founders. Virtually every project the fund has underwritten has had to do with “proving” the mental inferiority of black people.
This is largely because the $5-million fund consists of a group of five unpaid “directors,” of whom the only fully active decision-maker is Harry Weyher. Weyher, a New York lawyer, is the chosen successor of Wyckliffe Draper. Draper picked him because of his ideological reliability; Weyher shared Draper’s opposition to the Brown v. Board of Education decision.
According to Mr. Murray, the “main substantive issue” I discuss relative to his book is African IQ. In fact, my article also took on his arguments about Asian IQ, and his claims about the purported impact of immigration on America’s intelligence supply.
As for the matter of African IQ, however, my article did not, as Mr. Murray’s distorted summary would have it, “point to the many technical difficulties of knowing exactly what is going on.” Rather, I pointed out that Mr. Murray’s data contradict his own contention—which was that low African IQ scores suggest that the low scores of African-Americans are due to genetic and other factors, rather than a history of oppression in the United States.
Mr. Murray’s claim of extremely low African IQ derives from tests conducted in South Africa before the end of apartheid. This fact was not revealed in The Bell Curve—and small wonder. South Africa is hardly a place to find black people free from the effects of centuries of oppression. In fact, low IQ scores there can easily be construed to support the view that lower black intellectual achievement is a result of racism. This is anything but a “technical” issue.
Mr. Murray’s “facts” about Africa came from Pioneer-funded British psychologist Richard Lynn. Lynn, an editor of the notoriously racialist journal Mankind Quarterly (which, pace Murray, is neither “respected” nor “refereed”), is on record in support of the scientifically absurd view that the “proliferation” of the poor and other “weak specimens” needs to be discouraged in the interests of “the genetic quality of the group.”
Not surprisingly, then, this “scholar” accepted South African IQ test results at face value, as a fair test of the inherent intellectual capabilities of blacks not only in South Africa but in the continent as a whole. And Mr. Murray replicated his error.
In 1968, Richard Nixon had a secret plan to end the war in Vietnam. Charles Murray now assures us that he has “many more details” that would clinch his argument about African IQ, but these were omitted from The Bell Curve because—well, Mr. Murray doesn’t say why. He does say that I “want this literature to be weak and racist.” How can I “want” it to be anything, when I don’t even know it exists?
Perhaps Charles Murray did not know all the facts about the Pioneer Fund and Mankind Quarterly when he wrote his book. But now that he does, he must cease to deny them. He must deal openly and honestly with their implications for The Bell Curve. For a man who depicts himself in the pages of COMMENTARY as a brave struggler for unpleasant but vital truth, this is an appropriate test of intellectual courage. The Pioneer Fund and Mankind Quarterly are not mainstream, modern scientific institutions. They are scientific racism’s keepers of the flame. And scientific racism is one of the scourges of this century.
To the Editor:
There are few things more predictable than an author’s response to reviews of his work, so it is not surprising that Charles Murray found Peter Brimelow’s review to be “the best published synopsis of The Bell Curve.” Brimelow likened the book to Darwin’s Origin of Species, “the intellectual event with which it is being seriously compared.” Brimelow, who wants immigration laws changed so that America’s “racial balance” will be “shifted back . . . where it was in 1960: almost 89 percent white . . .,” shares Mr. Murray’s abhorrence of affirmative action and “quotas.” Those quotas favor blacks and Latinos.
What was not so predictable was the enormously successful marketing campaign for The Bell Curve engineered by the American Enterprise Institute (AEI) and the Free Press. The barrage of reviews in major newspapers and magazines, just as the book was published, propelled it to the bestseller lists and the talk shows, and into public consciousness. The reviews were carefully cultivated by AEI. Mr. Murray, in COMMENTARY, described an AEI-sponsored “conference of academics and journalists from various points on the political spectrum [held] soon after the book’s publication.” The authors of many journalistic puff pieces, including Brimelow, had been invited to that conference, expenses paid, and had been provided with advance copies of the book.
The Wall Street Journal—it occupies a point well to the Left on Murray’s political spectrum—indicated that the book had been “swept forward by a strategy that provided book galleys to likely supporters while withholding them from likely critics.” The Journal suggested that AEI “tried to fix the fight when it released review copies selectively, contrary to usual publishing protocol.” That charge was denied by AEI president Christopher DeMuth, who in a letter to the Journal indicated that the conference had in fact been held “several weeks before publication.” DeMuth asserted that he had “made a particular effort to attract likely critics to the conference . . . with the [deliberate] exception of Leon Kamin.” That most unkindest cut smarts, but I will try now to rise to the critical occasion. My allotted space permits adequately detailed discussion of only a single topic. I will focus here on the Herrnstein-Murray treatment of race and IQ. (My detailed critique of the book appears in The Bell Curve Debate, edited by R. Jacoby and N. Glauberman, Times Books/ Random House.)
Mr. Murray describes his book as “relentlessly moderate—in its language, . . . filled with ‘on the one hand, on the other hand’ discussions. . . . Anchored securely in the middle of the scientific road.” But the book’s description of affirmative action as “a poison leaking into the American soul” suggests passion, not moderation; and the even-handedness of Mr. Murray’s middle-of-the-road position is exemplified by his comments on the relationship of genes to race differences in IQ. On the one hand, “it is scientifically prudent at this point to assume that both environments and genes are involved, in unknown proportions.” But, on the other hand, “Herrnstein and I did not make nearly as aggressive a case for genetic differences as the evidence permits.”
The nonaggressive approach taken in The Bell Curve was to examine
a hypothesis Herrnstein and I heard frequently, that the test scores of American blacks have been depressed by the experience of slavery. We briefly summarize the literature indicating that African blacks in fact have lower test scores than American blacks . . . on standardized mental tests, including ones especially designed for illiterate non-Western subjects.
The logic here is: if slavery made American blacks dumb, then African blacks, who have merely been colonized, not enslaved, should be smarter than their genetic kinfolk in America. But in fact they are dumber! So American blacks must be dumb because of their genes. There is still a problem, however. Why should American blacks be smarter than their African cousins? Herrnstein and Murray maintain that “The IQ of ‘colored’ students in South Africa—of mixed racial background—has been found to be similar to that of American blacks.” Miscegenation appears to have paid off, to the tune of some 15 IQ points for American blacks. That follows since Richard Lynn, described in The Bell Curve as “a leading scholar of racial and ethnic differences,” after summarizing the literature, estimated the average IQ of Africans to be 70. That is to say, half of all Africans are mentally retarded. That finding evidently struck Herrnstein and Murray as reasonable. What, one wonders, would Mr. Murray’s more aggressive case for genetic differences look like?
But let us examine the test, widely used in Africa, which was “designed for illiterate non-Western subjects.” The test is Raven’s Progressive Matrices. Average scores on the Matrices, like those on other “IQ” tests, have been rising steadily over time throughout the world. A massively large study of Dutch draftees, using the Matrices, found that average IQ scores in Holland had risen by about 25 IQ points between 1950 and 1982! Richard Lynn concluded that
requirements for a culture-fair test are far from being met by the Progressive Matrices. . . . The testee has to decipher the code and then solve the progression problem. These largely arithmetical skills are of course taught in schools. Dutch adolescents in the 1980’s have enjoyed significantly more schooling . . . no doubt they have picked up a few more arithmetical skills. . . .
These observations did not prevent Lynn, four years later, from tabulating, in a review article, numerous African studies using the Matrices. The Bell Curve depended upon that review-article for its estimates of genetically debased African IQ.
Lynn himself felt that “the best single study of the Negroid intelligence” was performed in South Africa by Kenneth Owen, using the Junior Aptitude Tests. Zulu schoolchildren did very poorly on the test, so much so that Lynn judged their average IQ to be 69. Owen (but not Lynn, or Herrnstein and Murray) pointed out that “the knowledge of English of the majority of black testees was so poor that certain [of the] tests . . . proved to be virtually unusable.” To do well, Owen wrote, Zulu pupils would have had to have been familiar with electrical appliances, microscopes, and “Western type of ladies’ accessories.”
Charles Murray has written that “the social science that deals in public policy” has become “in a word, corrupt.” Pithy, and the shoe does fit The Bell Curve.
Leon J. Kamin
To the Editor:
Charles Murray takes the stance that many people have made an inappropriate fuss over The Bell Curve’s treatment of race and IQ, and professes himself bemused that “the critics have been obsessed—no hyperbole here—with genes . . . ,” inasmuch as the book simply makes the point that
a legitimate scientific debate is under way about the relationship of genes to race differences in intelligence; that it is scientifically prudent at this point to assume that both environment and genes are involved, in unknown proportions; and, most importantly, that people are getting far too excited about the whole issue.
If this were all the book had really said on the topic of race and IQ, the furor would indeed be a puzzle. But in fact, Herrnstein and Murray make three assertions about race and IQ pointing to the likely biological basis of race differences in IQ and to the probability that they are not eradicable by any means currently known. . . . (I have recently criticized the way Herrnstein and Murray make these points in a chapter in The Bell Curve Wars, edited by Steve Fraser, Basic Books.)
First, as to Herrnstein and Murray’s purported “review” of the direct evidence on heritability of race differences in IQ. This consists of presenting, at substantial length, only one of seven extant studies on the question. This is an adoption study showing that black children adopted into white families have lower IQ’s than white children adopted into white families. Despite the original investigators’ cautions that their study could not be taken as evidence for a genetic basis for IQ differences between the races—for a variety of reasons ranging from the possibility that white children might have been placed with more intelligent families to the possibility that emotional and adjustment difficulties would have been present for the black adoptees—Herrnstein and Murray declare the study to constitute strong evidence of the heritability of race differences.
The remaining six studies are all more consistent with the alternative position that genetic differences are negligible or that they favor blacks slightly. Herrnstein and Murray dismiss one of these studies in a single paragraph, citing interpretive difficulties of the same sort they neglect to mention for the study they favor. Another study is dismissed on the same sort of grounds in a note in the appendix. The other four studies are not mentioned at all.
The second major point of the case for genetics is that one cannot hope that low cognitive abilities of either blacks or whites can be much improved by intervention of any kind. They review two studies on intervention in infancy, both of which had very positive results but which they reject on methodological grounds. They ignore a dozen studies not subject to such criticisms but having results consistent with the two they reject. These additional studies include a recent, very large study conducted at eight different sites.
Herrnstein and Murray conclude that although vigorous post-infancy, preschool interventions boost IQ by 7 points or so, this is of little importance because gains fade by third grade. Yet it should be obvious (except to those holding dubious “ballistic” or “critical-period” theories of early intervention) that one would expect gains to be maintained only if enrichment were maintained. Herrnstein and Murray do not mention the evidence that this is, in fact, the case. More importantly, with one exception, they do not mention the interventions that have been initiated in elementary school. William J. Bennett, Ronald Reagan’s Secretary of Education, has described a score of such programs that have proved effective. Detailed reports on many of them appear in the education literature. The single elementary-school study reported on in The Bell Curve is an intervention that produced a gain of between 1.5 and 6.5 IQ points in a single year. This they discount because it is only one study and data could not be obtained on the question of the maintenance of the gains.
Though it is known that intermediate-school and high-school math programs can have a very dramatic effect on minority children’s math scores, this is also not mentioned. Finally, college programs exist that have a marked impact on minority achievement in several different specific fields, and on overall grade-point average. Yet again none of these demonstrations is mentioned.
In short, the review of intervention programs is so highly selective as to be misleading in the extreme, and so negative about the programs it does discuss as to forfeit any claim to represent balanced analysis.
Now, as for the 15-point gap in IQ between blacks and whites, Herrnstein and Murray refer to 10 studies of the current gap in IQ (or in a composite of abilities that is highly correlated with IQ). Summarizing these studies, they allow that the gap might have narrowed to 12 or 13 points now, though elsewhere in the book and several times in recent public statements by Mr. Murray, including his article in COMMENTARY, this gap is restored to a full 15 points. In fact, however, the current gap indicated by the median value of the studies reviewed by Herrnstein and Murray, and using the numbers they supply, is 9 points. Their description of these data goes beyond dubious analysis, beyond irresponsibly selective choice of evidence, to become outright misrepresentation.
In short, the burden of my critique is that the treatment of the question of race, heredity, and IQ in The Bell Curve is so selective, eccentric in interpretation, and factually incorrect that it could not be published in any reputable journal. In his article, Mr. Murray responds to these assertions only by attempting to discredit my statement that there is a large literature showing effective intervention with infants which is not subject to the criticisms in the Herrnstein and Murray book, and which shows substantial IQ gains.
Since I give only one example of this literature, Mr. Murray writes as if only one exists. He asserts that the large gains reported for this study did not persist past age three, when IQ’s are still unstable, and that the gain at age five had declined to a mere 2.5 points. Once again, he is trying to make the data say what he wants them to say rather than what they actually say. The 2.5 figure refers to the entire sample including an unusual group of very low-birth-weight infants. The near-normal-weight infants in the sample gained 4 to 6 points, or rather retained them because the intervention ended two years before the children were five years old, and the gain was even higher for children born to mothers with little education. These data are highly encouraging and entirely in line with the rest of the literature on normal-weight infants.
Most of your readers will have recognized the shrill and panicked nature of much of the reaction to the Herrnstein and Murray assertions about race and IQ. This is likely to create in any dispassionate person the presumption that that someone’s ox is being deservedly gored. The hysterical response to The Bell Curve is, in my opinion, the major reason the book is being paid any attention on the question of race and IQ. The scholarly critiques that will be coming out in the years to come will show how terrible the science in the book is. In the meantime, a great deal of damage has been done—to rational policy discussion and to relations of trust between blacks and whites.
Richard E. Nisbett
Institute for Social Research
University of Michigan
Ann Arbor, Michigan
To the Editor:
. . . In “ ‘The Bell Curve’ and Its Critics,” Charles Murray argues that criticisms of the Herrnstein/ Murray measure of background effects in the form of a weighted average of the education, occupation, and income of parents (called SES or socioeconomic status) are unfounded. He asserts that he and Herrnstein “deliberately constructed an SES index that uses the same elements that everybody else uses” and “throws down a challenge . . . to come up with another means of measuring the environment.”
Readers unfamiliar with the research literature on the effects of social environment on individual achievement may assume that Mr. Murray accurately describes the current state of research in this field. In fact, nothing could be farther from the truth. Measuring social background using only occupation, education, and income of parents characterizes only the earliest studies of the late 1960’s and early 1970’s. Since then a vast body of work has appeared in the leading academic journals of economics, sociology, and child development in which more subtle and sophisticated methods of assessing social-background effects have been put forward. Even a cursory examination of this work would reveal that Mr. Murray’s “challenge” to professional social scientists—that we “come up with another means of measuring the environment”—has long since been met.
Characterizing how social environment impacts on individual achievement is a difficult and complex process. A central aspect of this difficulty involves identifying whether or not a particular correlation—say, between a youngster’s churchgoing and her subsequent avoidance of early pregnancy—reflects a “causal” relationship, or is merely an artifact of some unmeasured factor which influences both churchgoing and early child-bearing (e.g., intensity of parental supervision). Attention to this kind of difficulty is what separates high-quality professional work in the field from the more pedestrian efforts.
Thus, contrary to Mr. Murray’s claim that the book uses the same SES index that everyone uses, the literature of the effects of social background on children’s academic achievement has identified a long list of factors which significantly influence the outcomes for children.
These factors include: (1) peer influences in the form of perceived peer education plans; (2) parental expectations and aspirations for their children’s schooling; (3) the income and racial composition of the community of origin; (4) the amount of time mothers spend in the labor market; (5) family structure—two parents versus a single parent, and whether parents are separated or divorced; (6) number of siblings and birth order; (7) religious denomination and church attendance; (8) grandparents’ schooling; (9) age of the mother at birth; (10) measures of the quality of stimulation found in the home environment, including emotional and verbal responsivity of the mother, provision of appropriate play materials, time and quality of maternal involvement with the child . . . , parental instigation of and participation in intellectual activities, parental affection, rejection, and nurturance . . . etc.; (11) language spoken at home; (12) discussions about college plans with teachers and other school officials; (13) parental emphasis on self-direction versus conformity; (14) ethnicity and immigrant status; (15) parental involvement in school activities; and (16) parental wealth and receipt of welfare income.
This long list is hardly exhaustive. I enumerate at such length only to stress how grossly inaccurate is Mr. Murray’s assertion that, by including their simple index of parents’ education, occupation, and income, he and Herrnstein were using “what everybody else uses.”
A list of the researchers who have contributed to the statistical analysis of various bodies of data in an effort to identify how social background, broadly construed, affects academic achievement would include scores of names and read like a “who’s who” of the fields of applied economics, sociology, and child psychology. The same is true of the other dimensions of achievement which Herrnstein and Murray investigate in The Bell Curve—poverty, welfare dependency, criminal behavior, parenting effectiveness, etc.
Finally, as a professional economist working in this area, I am taken aback by Mr. Murray’s insinuation that academic investigation of the effects of social background on offspring’s achievement has been somehow stunted by the purported liberal political commitments of researchers in this field. Contrary to Mr. Murray’s claims, we do not all think alike. I often find myself at odds with more liberal colleagues on policy questions. What I share with these colleagues, though, are a commitment to using the appropriate techniques of investigation and a reliance on peer review and critique before putting research findings which bear on important matters of policy in the public domain. Reading The Bell Curve, and now Mr. Murray’s response to his critics, makes me wish that this commitment were more widely shared.
There is a nearly universal consensus among professional analysts who have reviewed Herrnstein and Murray’s statistics that they grossly underestimate the relative effect of environment versus intelligence in accounting for individual differences in various dimensions of achievement. It is genuinely puzzling that Herrnstein and Murray failed to include a richer array of background factors in their analysis, since many of these items were unavailable in the data set they employed. Studies now going on and using the same data (e.g., one by Jonathan Crane of the University of Illinois) suggest that these richer measures of social background may account for much of the racial gap in cognitive test scores.
Thus, in his response to critics, Mr. Murray is calling on the experts to do what we have, for many years now, already been doing while leaving us no credible explanation for why he and Richard Herrnstein failed to do it themselves.
Linda Datcher Loury
To the Editor:
Given the motivations of the critics of The Bell Curve, Charles Murray’s compendious and lucid response will not silence them. To suggest that individuals and groups may differ in innate abilities, specifically intelligence, and that these differences are not reducible to socioeconomic causes, or entirely malleable, inevitably draws a hysterical response from dogmatic egalitarians who believe that au fond all men must be “created equal” in a far more literal sense than Jefferson ever dreamt of. Those in the media, literary intellectuals, and academics accept as common sense the para-Marxist view propagated in sociology courses that life chances differ according to the economic class into which one is born. (Although dead elsewhere, Marxism leads a merry afterlife in academe.) . . .
But The Bell Curve proves that life chances depend far more on one’s cognitive ability than on one’s class and, worse, that cognitive ability is mostly the cause rather than the effect of one’s position in life. Class, the wealth or poverty of one’s parents, matters far less than one’s IQ, which largely (about 60 percent) is an independent variable.
Much of the negative reaction to The Bell Curve can be attributed to the fact that its conclusions are inconsistent with what is taught in sociology courses and with what, wittingly or unwittingly, is accepted by most literary intellectuals. It is much easier to repudiate The Bell Curve than to part from the egalitarian dogma and notions which hitherto have explained everything.
Ernest van den Haag
New York City, New York
To the Editor:
Charles Murray is correct when he states in his article that he and Richard Herrnstein downplayed the evidence for the genetic basis of ethnic/race differences in cognitive ability (including the very high IQ of Jews). They even equivocated on whether the term “race” can be applied to African-Americans who may be more “white” than “black,” genetically speaking. . . .
Researchers like me are greatly indebted to The Bell Curve (and its critics) for getting the “genie out of the bottle,” i.e., the idea that individual and racial differences are due, at least in part, to genetic differences. Human differences can be fully understood only in a wider (evolutionary) context. Such understanding will show why the “American dilemma” is international in scope.
J. Philippe Rushton
University of Western Ontario
London, Ontario, Canada
To the Editor:
Charles Murray has once again demonstrated why he is among the most perceptive social scientists and social critics in America today (along with James Q. Wilson and Peter L. Berger). The incredibly hostile reception accorded The Bell Curve would have thrown a lesser person into paroxysms of vituperation. Instead, Mr. Murray coolly and clinically diagnoses the causes and foresees some of the consequences of the hysterical treatment his exemplary study has provoked.
The Bell Curve (like its predecessor, IQ in the Meritocracy, by Mr. Murray’s co-author Richard J. Herrnstein) simply draws inferences from the disparities in intelligence that are readily observable among individuals and groups. That these inferences are not conducive to utopian designs for society is, as Mr. Murray contends, a wholly “modest” finding.
What is so infuriating is the abuse to which Mr. Murray has been subjected. It is one thing to attempt to refute his arguments; quite another to assassinate his character by impugning his motives, sources of funding, and even his high-school pranks of decades ago. I cannot for the life of me understand why some of the most respected scholars in the country stoop to ad-hominem attacks. It is one thing for TV talk-show host Phil Donahue to violate elementary canons of civil discourse by accusing Mr. Murray, without any justification, of having been a proto-Nazi in adolescence. It is quite another for Harvard scientist Stephen Jay Gould to make only slightly more sophisticated but equally spurious and damning allegations.
I have watched in despair as Christina Hoff Sommers, Michael Levin, Carol Iannone, Steven Goldberg, Elizabeth Fox-Genovese, Thomas Sowell, and dozens of other first-rate scholars have been vilified for following where evidence and logic lead. Without minimizing their suffering, I would say that Charles Murray has been the subject of an academic inquisition which is without parallel in recent memory. It makes me ashamed of my profession to realize that smear tactics are the weapons of choice in the battle over ideas for people trained and paid to lead a “life of the mind.” Truly, in the words of Thomas Jefferson, “I weep for my country when I reflect that God is just.”
Lloyd B. Lewis
Savannah College of Art and Design
To the Editor:
I write to corroborate, by way of personal anecdote, Charles Murray’s proposition that “the critics of The Bell Curve are going to produce the very effects that their attacks have been intended to avert.”
Shortly after the book’s publication I was invited to attend a lecture at MIT. Neither the topic nor the speaker was specified, but I inferred that the subject was significant and the speaker’s reputation weighty along the Cambridge corridor.
I arrived at the overfilled lecture hall with my critical canvas a tabula rasa: I knew of Stephen Jay Gould (the speaker, it turned out) only vaguely; I knew of Charles Murray (part of the topic) even more vaguely; I knew of The Bell Curve (the balance of the topic) not at all. I was an innocent.
If I harbored any expectation, it was that I would be exposed to reasoned, objective, critical, and informative discourse, based on sound scholarship. What I was exposed to instead was an academic hour of arch, snide, and conclusory attack on a book and its absent authors.
Perhaps with some residual respect for the newly deceased Richard J. Herrnstein, Gould concentrated his sharper ad-hominem weapons on Charles Murray. Directly and indirectly Gould made him out to be an intellectual lackey . . . of the far Right, which, . . . to Gould’s correct-thinking audience, meant that he and his work were probably infected with racism. In Gould’s view, Mr. Murray’s scholarship was, of course, flimsy, based on discredited psychometrics, the antediluvian g, and the outmoded IQ concept. Using clever innuendo, as well as facial expressions and patronizing chuckles to indicate his far-superior intellect, Gould left little doubt as to the charlatanry of the authors. . . .
At the end of Gould’s talk, I felt I had not heard a critical lecture on behavioral science but a sermon on theology, complete with a postmodern edict of excommunication. . . .
In the weeks that followed, the deluge of equally vitriolic reviews directed at The Bell Curve confirmed my impression that much more was at work here than the disputation of scholars. What I was witnessing was a jihad in which Charles Murray was the infidel. . . .
It has been intellectually refreshing finally to encounter in the pages of COMMENTARY the real Charles Murray rather than a devil-person. I can now explore both sides of a legitimate question Stephen Gould so decidedly intended to remain unexamined.
Charles Murray writes:
I am grateful for the letters from Ernest van den Haag, J. Philippe Rushton, Lloyd B. Lewis, and W.H. Ryan. But I must pass up further comment on these letters to focus on the critical ones from James J. Heckman, Charles Lane, Leon J. Kamin, Richard E. Nisbett, and Linda Datcher Loury.
In “ ‘The Bell Curve’ and Its Critics,” I discussed four issues that have been the subject of intense attack: the idea of a general factor of mental ability; the possibility that genes play a part in ethnic IQ differences; the statistical power of the results reported in The Bell Curve; and the attempts to raise IQ through program interventions.
James J. Heckman’s letter deals primarily with issues involving the general factor of mental ability. “The basic premise of the book,” Mr. Heckman writes, “is that g—or a single factor of intelligence—explains behavior, and that it is immutable.” But this characterization, which provides the rationale for the criticisms in his letter and for much of his longer critiques of The Bell Curve published elsewhere, constitutes a straw man that bears no resemblance to the spirit or content of the book. At no point do the late Richard Herrnstein and I hint that IQ is an all-powerful determinant of behavior, or that IQ is immutable.1
Let me offer some examples of our plainly stated view on these issues.
Although we regard intelligence as helping to “explain” behavior in a statistical sense, we repeatedly emphasize how much IQ scores leave unexplained. In the introduction, when we first describe our view of intelligence, we conclude with this passage:
All of this is another way of making a point so important that we will italicize it now and repeat it frequently throughout the book: measures of intelligence have reliable statistical relationships with important social phenomena, but they are a limited tool for deciding what to make of any given individual (p. 21).
On page 68, after displaying a scatterplot of two variables with a correlation of .33, we again put our message in italics to make it hard to miss:
For virtually every topic we will be discussing throughout the rest of the book, a plot of the raw data would reveal as many or more exceptions to the general statistical relationship, and this must always be remembered in trying to translate the general rule to individuals.
And in the introduction to Part II, opening the analyses of the relationship of IQ to social behaviors:
High cognitive ability is generally associated with socially desirable behaviors, low cognitive ability with socially undesirable ones. “Generally associated with” does not mean “coincident with.” For virtually all of the topics we will be discussing, cognitive ability accounts for only small to middling proportions of the variation among people (p. 117).
When Mr. Heckman further writes in his letter that “. . . many skills affect outcomes and not all can be equated with native intelligence,” may I suggest that we said exactly the same thing?:
Perhaps a freshman with an SAT math score of 500 had better not have his heart set on being a mathematician, but if instead he wants to run his own business, become a U.S. Senator, or make a million dollars, he should not put aside those dreams because some of his friends have higher scores. The link between test scores and those achievements is dwarfed by the totality of other characteristics he brings to his life (p. 66).
In the light of such statements, I only wonder how Mr. Heckman, a careful scholar, can write, “Once [the role of many skills] is recognized, a core argument in The Bell Curve evaporates.”
Mr. Heckman is equally wrong about our position on the immutability of intelligence. In the introduction we state, as one of six general points well-established in the literature, that “IQ scores are stable, although not perfectly so, over much of a person’s life,” adding immediately that “All six points have an inverse worth noting. For example, some people’s scores change a lot . . .” (p. 23). With regard to actual changes in intelligence, we credit the establishment of universal education with having major effects on intelligence (pp. 396-97, 589-592); similarly with adoption at birth (pp. 411-13). With regard to other potential means of changing intelligence, we are more aggressive: “Limitless possibilities for improving intelligence environmentally wait to be uncovered by science. . . . In principle, intelligence can be raised environmentally to unknown limits” (p. 390). This is not the work of authors who believe that intelligence is immutable.
In the last two lines of his letter, Mr. Heckman makes two large statements: first, that one of the measures of intelligence we use is really an achievement test, not a test of mental ability; and second, that it can be manipulated. The test he refers to is the Armed Forces Qualification Test (AFQT).
Concerning the AFQT as a measure of intelligence: all IQ tests are designed for a reference population. In the case of the AFQT, that population is people in their late teens who have been exposed to the American school system. But the fact that items on a test depend on some past education does not necessarily mean that the test is no longer a measure of general intelligence. The brief story, told at length in Appendix 3 of The Bell Curve, is that the AFQT is one of the most highly g-loaded tests in current use.
Would The Bell Curve’s results have been different if we had access to one of the many other standardized mental tests? Unlikely. The AFQT is more highly correlated with other major IQ tests than other major IQ tests are correlated with each other (pp. 584-85).
Can the AFQT scores be manipulated? Yes—on the computer, by econometricians using complex statistical models. What about by school administrators and teachers trying to keep someone in school for another year to raise his AFQT score? Not much; but more on this below.
As for Mr. Heckman’s statement that I cite John Carroll’s work misleadingly, I can only say that it will come as a surprise to Carroll, who at my request reviewed and approved my citation of his work.
Charles Lane, Leon J. Kamin, and Richard Nisbett all weigh in regarding genes and ethnic differences in intelligence.
Mr. Lane is shocked and horrified that I cite scholars who have received funding from the Pioneer Fund. The one paragraph I devoted to the subject in my article was not intended to rebut his allegations but to express my disdain for them. As I have written elsewhere, the attempt to discredit a book at third hand—by tracing the funding history of some of the scholarship it refers to—is a form of McCarthyism. I do not choose that adjective lightly.
In that sense, I also do not care whether Lane’s accusations about the history of the Pioneer Fund are true, although I have read enough of the Pioneer Fund’s side of the story to doubt them. To me, the key point is that for some decades the Pioneer Fund has given money to legitimate scholars to work on topics of legitimate scientific inquiry, and it makes no attempt whatsoever to influence the course of that work.
As for the sources themselves in The Bell Curve, the one criterion Herrnstein and I followed was whether the work had scientific merit and was relevant to our presentation. So let us examine Mr. Lane’s and Mr. Kamin’s charges on that count.
As a preliminary, let me take up Mr. Kamin’s complaint about the logic linking African IQ, the legacy of slavery, and African-American IQ, and simply say that the logic he finds objectionable is not ours. Our strategy in Chapter 13 was to answer all the common questions about ethnic differences in intelligence. In the course of writing the book, it quickly became apparent that many people assumed scores of blacks in Africa would be higher than scores of blacks in the U.S., on the theory that colonialism or apartheid was less destructive than the legacy of 250 years of slavery. We did not try to assess the relative oppressiveness of these systems; we simply told readers that such expectations were wrong. African IQ has been found to be substantially lower than African-American IQ.
How valid are the studies that show African scores on mental tests as markedly low? Richard Lynn, a careful scholar who has been the target of almost as much unwarranted criticism as J. Philippe Rushton, converted scores on the existing studies of African cognitive ability to IQ scores, and found that they fall in a range from the 60’s to the 80’s, with a mean of about 70 and a median of 75.2 His rationale for converting the scores is defensible but also unnecessary to his argument. One may instead use standard deviations or percentiles to make the same case: the African mean on cognitive tests is in the region of two standard deviations below the white mean, or somewhere below the fifth percentile of the white or European distributions on the same tests.
Might it be, as Mr. Kamin argues, that these studies are invalid because the tests were administered to illiterates, or to Africans who could not be expected to be familiar with culturally specific bits of information? No. The largest and most careful of the studies have sometimes been limited to urban populations, to persons who have graduated from middle school, to students still in school at the secondary level, or to employed persons. Such samples may if anything tend to overestimate, not underestimate, the national mean by overloading the sample with persons who have had the ability and persistence to remain in school or hold jobs.
Many of the best studies have also used Raven’s Standard Progressive Matrices (SPM), which is an excellent measure of the nonverbal component of IQ, is highly g-loaded, and is not bound by culturally specific information. Does it help to have been in school to do the SPM? Probably—and the samples in the best African studies have been students in school.
To illustrate how troubling the results have been, let me turn to two studies postdating Lynn’s review. One, from South Africa, was led by Kenneth Owen (not the study mentioned by Mr. Kamin). The results were published in the refereed British journal, Personality and Individual Differences.3 Its sample consisted of enrolled seventh-grade students and included 1,056 whites, 778 “coloreds” (mixed-race), 1,063 Indians, and 1,093 blacks. The SPM was administered without time limits. Except in the case of the Indians, subjects were tested by school psychologists of the same ethnic group.
Owen presents the full psychometric profile for the test results (distributional characteristics, reliability, item difficulty, item discrimination, congruence coefficients, and discriminant analysis), demonstrating that the test did indeed measure the same thing in each of the various ethnic groups. The discrepancies in mean scores? Expressed in standard deviations, they were as follows: Indian-white: -.52; colored-white: -1.35; black-white: -2.78. This black-white difference is larger than Lynn’s earlier estimates.
The second recent study was conducted by a black scholar, Fred Zindi.4 It took 204 black Zimbabwean pupils and 202 white English students from London inner-city schools and matched them according to age (12-14), sex, and educational level; both samples were “working-class.” Despite the fact that the white sample was well below average for whites, with a mean IQ—as measured by the test known as WISC-R—of only 95, the difference between whites and blacks was 1.97 standard deviations on the SPM and 2.36 standard deviations on the WISC-R. Mr. Zindi reported these as IQ’s of 72 for the SPM and 67 for the WISC-R—consistent with Lynn’s earlier estimates. (There is reason to think that the WISC-R score was somewhat depressed by language considerations, but not much.)
What should one make of these results? Above all, we must proceed cautiously, for the same reasons that guided us in The Bell Curve. Our view was that the differences between groups will narrow over time, probably dramatically, as nutrition and the quality of schools improve for black Africans. Changes in black African culture may also provide an environment more conducive to cognitive development among young children. But that does not mean that the current differences, as measured through these samples, are figments of the imagination, or that differences in test scores do not represent real differences in cognitive functioning. They do, and those differences are extremely large, much larger than the differences separating American blacks and whites.
You may choose to believe that Owen is a white racist who wittingly or unconsciously rigged the results, although his scholarly reputation belies it. (Zindi would seem exempt from the same charge.) You may choose to believe that the poor black performance is itself a consequence of colonialism and apartheid and will soon vanish. But these are just assumptions. When data are as carefully collected and analyzed as these, attention must be paid.
Richard E. Nisbett takes a different tack, focusing on studies in the American context. He begins by complaining that Herrnstein and I discussed only one of seven studies that bear on racial ancestry at sufficient length to suit him. We alluded to two others, he acknowledges, but dismissively. Here, however, is our sentence evaluating one of them: “The study is inconclusive but certainly consistent with the suggestion that the B/W [black-white] difference is largely environmental” (p. 310). Of the other, we wrote:
If the whites who contributed this ancestry were a random sample of all whites, then this would be strong evidence of no genetic influence on black-white differences. There is no evidence one way or the other about the nature of the white ancestors (p. 729).
These sentences sound to me like the phrasing of careful social scientists trying to give as much credit to studies as the data warrant. What does Mr. Nisbett find unjustifiably dismissive in either of them?
In his letter, Mr. Nisbett does not cite the remaining four studies, but in his article in The Bell Curve Wars he cites studies from before 1975 that are also discussed in a book by John Loehlin, Gardner Lindzey, and J.N. Spuhler.5 Here is what we said about these studies in The Bell Curve:
Several smaller studies bearing on racial ancestry and IQ were well summarized almost two decades ago by Loehlin, Lindzey, and Spuhler. They found the balance of evidence tipped toward some sort of mixed gene-environment explanation of the B/W difference without saying how much of the difference is genetic and how much environmental.
If Mr. Nisbett is going to claim that we misled our readers by skipping over literature we did not like, he must show that we misrepresented Loehlin, Lindzey, and Spuhler’s review of the evidence; or that their conclusions were themselves wrong; or that some decisive change in their evaluation has been prompted by subsequent work. He does none of these things, instead offering a one-sentence summary of the (unspecified) literature. I submit that our one-sentence summary is more balanced and prudent than his. Interested readers with access to a university library can look up Loehlin, Lindzey, and Spuhler, Chapter 5, and decide for themselves.
Mr. Nisbett is correct, however, in saying that we devote most of our attention to one study, known as the Minnesota Transracial Adoption Study. Why? Because the Minnesota Transracial Adoption Study was for more than a decade Exhibit No. l for optimists on the nature of black-white differences, and was widely cited as close-to-definitive empirical proof that such differences in intelligence were environmental.
As it happens, the early data from the Minnesota Transracial Adoption Study were not as unambiguously positive as they were often reported to be in the press.6 At age seven, the scores of the black adoptees were below those of the white adoptees, and even further below those of white biological children of the adoptive parents. Still, the mean of the black and interracial adoptees was 106; which was not only higher than the Minnesota black mean of about 89 but higher than the national white mean. This was good news by any standard.
In the late 1980’s, however, when the same children were tested in adolescence, the IQ means were as follows: 109 for the biological children of white parents; 106 for the adopted white children; 99 for the adopted children with one black biological parent; and 89 for the adopted children with two black biological parents.7 As we put it in the book: “The bottom line is that the gap between the adopted children with two black parents and the adopted children with two white parents was 17 points, in line with the B/W difference customarily observed” (p. 310). This is bad news by any standard.
If the national black-white difference is in the region of 15 points, and if, after growing up in white homes, adopted black children still are just as far behind, then the first, parsimonious explanation of such differences must be that they are largely genetic. I have no problem with the attempt of the authors of the Minnesota study to search for alternative, environmental explanations for the results, which Mr. Nisbett cites enthusiastically. Thus, when those authors—Richard Weinberg, Sandra Scarr, and Irwin Waldman—came to offer conclusions, they wrote:
The results of the longitudinal follow-up continue to support the view that the social environment maintains a dominant role in determining the average IQ of black and interracial children and that both social and genetic variables contribute to individual variations among them (p. 133).
Other scholars argue (as Herrnstein and I thought) that Weinberg, Scarr, and Waldman retreated too far from parsimony, trying too hard to squeeze the last ounce out of an environmental explanation. But here is the key point: this is an ordinary scholarly difference of opinion, bounded by common agreement. Weinberg, Scarr, and Waldman acknowledge that their data indicate that some genetic component is probably involved—as indeed their data make it very hard to deny.
I have rehearsed the African IQ data and the Minnesota Transracial Adoption data at such length because, in the letters printed above as in the published reviews of The Bell Curve, our discussion of the possibility of genetic racial differences in intelligence has been attacked so relentlessly. What has all the fuss been about? Let me, for the umpteenth time, repeat our concluding paragraph here:
If the reader is now convinced that either the genetic or environmental explanation has won out to the exclusion of the other, we have not done a sufficiently good job of presenting one side or the other. It seems highly likely to us that both genes and environment have something to do with racial differences. What might the mix be? We are resolutely agnostic on that issue; as far as we can determine, the evidence does not yet justify an estimate (p. 311).
At present, this is the scientifically prudent position.
One final aspect of the black-white difference is raised by Mr. Nisbett, who claims that I continue to refer to a black-white gap of 15 IQ points while our own account in The Bell Curve reveals a median current gap of only 9 IQ points. “Herrnstein and Murray’s description of these data goes beyond dubious analysis, beyond irresponsibly selective choice of evidence, to become outright misrepresentation of a state of affairs.”
Mr. Nisbett has it exactly wrong. Far from ignoring evidence of convergence, Herrnstein and I were, if anything, guilty of downplaying important evidence that the current black-white gap is not closing at all but diverging, and that the actual current gap is not just 15 points, but some larger figure.
Mr. Nisbett’s use of a 9-point gap comes from our review of recent IQ test data (pp. 289-90) as reported in an article by Ken Vincent.8 There, Vincent argued that studies of children in the 1980’s indicate a smaller difference than is observed among adults. But (a familiar story) the studies Vincent used are beset by problems of interpretation: unrepresentative samples, results that in the technical literature are said to be artifactual, and, in five of the studies, a control for socioeconomic background that is guaranteed to reduce the black-white difference by about 40 percent. Nonetheless, we called Vincent’s evidence encouraging, as indeed it is.
But another, much larger source of data on the black-white difference among today’s children, not used by Vincent, is also available. Every two years since 1986, the children of the women in the National Longitudinal Study of Youth (NLSY) have been tested on the Peabody Picture Vocabulary Test, a nationally normed and widely used measure of child IQ. When we wrote The Bell Curve, we had access to the data through the 1990 testing cycle. We alluded to it at the end of our review of Vincent’s data, and presented the actual numbers in Chapter 15. Restricting the comparison to pairs of mothers and children who had been tested, we reported the gap separating the black and white mothers as 13.2 IQ points, while the gap separating the children was 17.5 points. We also said that “There are technical reasons to hedge on any more specific interpretation of these data” (p. 356).
I have looked since then at the data through the 1992 testing cycle. The same trend persists, even after correcting for the ways in which current samples are unrepresentative. Nor should it be surprising that the NLSY is yielding these results. They are consistent with the expectations one may draw from the national birth data, which show a marked black-white divergence in births based on educational level of the mother (which in turn has a reliable correlation with maternal IQ).
There are still reasons to hedge on what will eventually happen to the black-white difference, but the notion that the balance of the data demonstrates a brighter future, let alone only a 9-point difference, ignores reality. The Bell Curve’s summary of the situation struck the right note, I think:
Many of you will be wondering why we have felt it necessary to qualify the good news [on convergence of scores]. A smaller number of readers who specialize in mental testing may be wondering why we have given so much prominence to educational achievement trends and a scattering of IQ results that may be psychometrically ephemeral. The answer for everyone is that predicting the future on this issue is little more than guesswork at this point. We urge upon our readers a similar suspension of judgment (p. 295).
In “ ‘The Bell Curve’ and Its Critics,” I argued that the book undermined the importance of socioeconomic background (SES) as an explanation for social outcomes. Linda Datcher Loury takes me to task for not acknowledging a large and expanding literature on the many factors other than traditional measures of SES that are known to affect such outcomes.
Her criticism is not without merit, but also not entirely on point. I did not argue that no one has explored other factors. I pointed out, rather, that (1) SES has for many years been a staple in academic analyses of why children of some families become poor, why they drop out of school, why they commit crimes, etc.; (2) a major purpose of The Bell Curve was to add IQ to this standard set of explanatory variables; (3) our use of SES has come in for much frivolous criticism; and (4) The Bell Curve has diminished the persuasiveness of the role of SES in explaining social problems. Nothing in Mrs. Loury’s letter disputes the last three points. As for the first, does she really want to contend that in academic treatments of poverty, welfare, crime, and other social pathologies discussed in Part II of The Bell Curve, socioeconomic background has not (along with racism) been the dominant explanatory construct since the 1960’s?
But Mrs. Loury is quite right that I should have acknowledged the growing literature on other explanatory variables. When I write about this issue in the future, I will amend my presentation accordingly.
The final topic in my article was my prediction that the attacks on The Bell Curve will backfire by forcing policy-makers to confront just how difficult it is to make substantial, long-term improvements in cognitive functioning. As promised earlier, I will now return to this point.
Here, again, Herrnstein and I are accused of dismissing studies we did not dismiss. In this case, Richard Nisbett cites two studies of intervention in infancy, “both of which had very positive results but which they reject on methodological grounds.” He refers to the Milwaukee Project and the Abecedarian Project. This, to be tiresome about it, is what we actually wrote:
In summary, the two experiments contain some promising leads. But it is not obvious where to go from here, for they differed in possibly important ways. The Abecedarian Project evaluated day care; the Milwaukee Project provided numerous interventions besides day care, including parental payment and training. It is hard to tell whether the former found enduring IQ benefits, given the very early divergence in test scores for experimental and control groups, but it found some academic benefits; the latter found an enduring IQ gain, but it has not yet shown comparable intellectual gains in school work (p. 409).
This is what he calls a rejection? Is he prepared to claim more for these two projects than we did?
Mr. Nisbett also asserts that we “ignore a dozen studies not subject to such [methodological] criticisms but having results consistent with the two they reject.” Once again he has wrongly accused us of ignoring studies that we covered, in this case using the respected synthesis of the literature conducted by the Consortium for Longitudinal Studies.9 Overall, the Consortium found about a 7-point gain in the exit test for such interventions, a gain that faded out within a few years. In the end, concluded the Consortium, “The effect of early education on intelligence test scores was not permanent.”
In the next part of his letter Mr. Nisbett refers to a variety of interventions that conflate improvements in educational instruction and achievement (which neither Herrnstein nor I would dispute) with improvements in cognitive ability. This distinction is crucial, however. Do we know how to take a set of youngsters with a given tested IQ and reliably improve their educational achievement? Yes. Do we know how to take a set of youngsters with a given tested IQ that would not (for example) allow them to become engineers, and reliably raise their cognitive functioning so that they can become engineers? No. Do we know how to sustain gains in IQ if we sustain the enriched curriculum? Mr. Nisbett says that we do. In my article, I observed that we have no credible study offering evidence of significant, long-term effects on cognitive functioning that Herrnstein and I did not consider in The Bell Curve. I am still waiting for a citation. Mr. Nisbett offers none.
I grant that he is one-twelfth correct when he complains of the dozen studies we failed to consider in The Bell Curve. The eight-site study he mentions, in which preschool, social, and pediatric services were supplied to low-birth-weight (less than 2,500 grams) infants, postdates the Consortium’s analysis, and we did indeed fail to include it. Mr. Nisbett’s citation of it tells an instructive story.
In his article in The Bell Curve Wars, Mr. Nisbett had pointed to a follow-up of this study, published in 1992, claiming that the infants had gained an average of 9 IQ points at age three. More accurately, the study administered two IQ tests. Mr. Nisbett mentioned only the one showing a 9.4-point gain. The other showed a 6.4-point gain. I pointed out in my May COMMENTARY article that there was also a subsequent follow-up, published in 1994, which showed those gains dropping to near-zero.
Let us pursue this issue just a little further. The 1994 follow-up, published in the Journal of the American Medical Association (JAMA), also broke down the results according to whether the babies were “lighter LBW [low birth weight],” meaning less than 2,000 grams, or “heavier LBW” (2,001-2,500 grams).10 In that comparison, the heavier LBW group showed an advantage over the control group of 6.0 IQ points on one test and 3.7 points on the other IQ test, as Mr. Nisbett now says. I could stop here: this is precisely the kind of gain—modest in size, showing a decline over time—that Herrnstein and I described as characteristic of the literature. But there is more. Mr. Nisbett neglects to mention that the lighter LBW sample, just as much a part of the study as the heavier LBW sample, showed no effect at all from this ambitious attempt to help them.
Now, imagine that Richard Herrnstein and I had found a study of children that showed the results we “wanted” at age three, and reported them in The Bell Curve for the entire sample. Suppose we ignored the already-published results from a follow-up at age five which contradicted those results. Suppose we then responded to criticism of our presentation by focusing on a subsample of subjects at age five who still showed the “wanted” results, saying that this was the meaningful sample, ignoring the subsample that showed no effect. Had we done any of that Mr. Nisbett would have been enitrely correct in charging us with playing fast and loose with the interpretation of data. But he is the one who has done each of those things, not Herrnstein and Murray in The Bell Curve or Murray in COMMENTARY.
Mr. Nisbett writes, “Mr. Murray is trying to make the data say what he wants them to say.” I ask the reader to judge whether, given all the results I have just described, I was correct when I wrote in my article: “The only study in ‘the very large literature’ that we missed does not contradict our conclusion that such interventions have provided promising leads but no more.”
It is appropriate to close with this episode because it characterizes the nature of far too many of the attacks on The Bell Curve. I am happy to report that since I wrote “ ‘The Bell Curve’ and Its Critics,” I have begun to see the first examples of the kind of tough but collegial criticism which Richard Herrnstein and I had hoped for. That said, I still find the most frightening part of the The Bell Curve experience to be not the frenzied reaction in the popular media, but the intellectual irresponsibility of so many of the book’s academic critics.
1 A technical correction to Mr. Heckman's statement: we report the role of g only once in the main text of the book, with regard to grades in military schools (pp. 76-77). Apart from that, we report the relationship of all the outcome variables to IQ scores, accepting (as apparently Mr. Heckman does) that the general factor, g, accounts for the bulk of whatever explanatory power IQ scores have.
2 “Race Differences in Intelligence: A Global Perspective,” Mankind Quarterly 31 (1991), pp. 254-96.
3 K. Owen, “The Suitability of Raven's Standard Progressive Matrices for Various Groups in South Africa,” Personality and Individual Differences, 13 (1992), pp. 149-59.
4 “Differences in Psychometric Performance,” Psychologist, 7 (1994), pp. 549-54.
5 Race Differences in Intelligence. (W.H. Freeman, 1975).
6 S. Scarr and R.A. Weinberg, “IQ Test Performance of Black Children Adopted by White Families,” American Psychologist, 31 (1976), pp. 726-39.
7 R.A. Weinberg, S. Scarr, and I.D. Waldman, “The Minnesota Transracial Adoption Study: A Follow-Up of IQ Test Performance at Adolescence,” Intelligence, 16 (1992), pp. 117-35.
8 “Black/White IQ Differences: Does Age Make the Difference?,” Journal of Clinical Psychology, 47 (1991), pp. 266-70.
9 I. Lazar and R. Darlington, “Lasting Effects of Early Education: A Report From the Consortium for Longitudinal Studies,” Monographs of the Society for Research in Child Development, 47 (1982), pp. 2-3 (Serial No. 195).
10J. Brooks-Gunn, et al., “Early Intervention in Low-Birth-Weight Premature Infants,” JAMA, 272 (1994), pp. 1257-1262.