Iq Testing and the Media

A psychologist argues that the press won’t report both sides of the IQ controversy

BY R. J. HERRNSTEIN

IN JUNE OF 1973, I SPOKE AT WELLESLEY COLLEGE TO A small audience of students, visitors, and at least one reporter. My topic was press coverage of the IQ controversy, which I thought distorted to the point of newsworthiness. After my talk, the reporter, from The Boston Globe, asked me to substantiate my criticisms of the press. I sent him a bundle of clippings that I thought amply documented my charge. A few weeks later his article, putatively about my Wellesley talk, appeared. It was 300 column lines long, but it contained no mention of press coverage of the IQ controversy. Instead, the article, obviously based on interviews with my critics, attacked what I had said or written about IQ in other places. The article said nothing about my answers to these critics, and nothing about the content of my actual talk at Wellesley College.

When I called him to object, the reporter said that his editors intended to invite me to write a rejoinder. Later on, pleading a shortage of space, the reporter told me I would have no chance to answer. When I complained to one of that newspaper’s senior editors, he told me he had just rejected an article highly critical of me and my position. At some point, he implied, arguments about IQ, whatever their position, become tedious. I wrote again nevertheless, disputing his decision and his reasons for it. A few days later, the Globe published an article by a sociologist teaching at Princeton that accused me, falsely, of racism.

Unfortunately, this encounter with the press resembles a great many others. A few weeks before my talk at Wellesley College, I had been in New York at the invitation of Social Policy, a leftist magazine with a record of antipathy toward IQ testing and its social implications. My condition for accepting the invitation to participate in a debate on IQ was that I be shown in advance any press release advertising the debate, for I feared being misrepresented. The press release the editors sent me was acceptable, but I may have been the only person to get it. Reporters from The New York Times, The New York Post, and other media received a different press release, attributing to me racist and sexist beliefs. In my talk, I described how my hosts had set the scene, but only The New York Post told the story of the counterfeit press release.

One would hope that America’s newspaper of record, The New York Times, had higher standards. Perhaps it does in general, but my impression is that it doesn’t on the IQ controversy. On December 17‚ 1975, for example, Christopher Lehmann-Haupt, in his daily book review, praised two now largely forgotten books, both attacking IQ testing.

Since Lehmann-Haupt mentioned me for special disapproval, I thought I had a right to correct some of his errors. An editor I called agreed and invited me to do an Op-Ed piece, which I promptly wrote. A week later, in a letter rejecting my article, the editor said:

Many thanks for your thoughts. . . . We think you’re right to be outraged by that review, but we don’t think it’s right for us simply to engage in a direct polemic with a book review in the daily paper. Also . . . the basic argument is by now familiar. . . .

But we love the way you write and think and we would love to capitalize on your approach to us to go beyond this topic ... In any case, the purpose of this letter is not to be negative—quite the contrary. We want to sign you up for a good long piece in which you could make not only this point but many others, depending on the flow of your work.

Instead of 1,200 words on the Op-Ed page, I was to get about 6,000 words in the Sunday magazine: not a bad exchange, I thought. After a further conversation defining the scope of the longer article, I set to work, and a few weeks later I mailed a draft in for comment. Six weeks passed without a word from the Times. I then called the Sunday magazine’s editor, who bluntly rejected the article because, he said, everyone already knew what, in substance, it had to say about IQ. Several months later, the Sunday magazine published an article by an economist criticizing IQ tests, not as harshly as Lehmann-Haupt had, but making some of the very same errors.

Incurably addicted to quantification, I have now searched the daily and Sunday New York Times from 1975 to November of 1981 for all book reviews dealing with IQ. Of the fifteen reviews that I found, every one denigrated IQ tests, often vitriolically. All but two of the books reviewed were opposed to testing, as far as one can tell from the reviews, and were praised for their position. One of the exceptions was a book by Arthur Jensen, which happened also to be the only book by a trained psychometrician (psychometrics is the psychological specialty concerned with testing). Jensen’s book was panned by a philosopher with no detectable expertise in the subject.

Except for Jensen’s book, none of the major works on testing written by professionals during the period was reviewed. Most remarkable, however, the Times published no review by a trained psychometrician. Dozens of literate psychometricians might have commented on the shallowness of the books the Times usually chooses to review. But psychometrics is forbidden territory in the Times—books in the field are mostly unreviewed, its discoveries are unreported, and its experts are, apparently, unconsulted. Rarely, if ever, in more than a decade, has a specialist in psychometrics published a review of a book on testing in the Times, The New York Review of Books, The New Republic, or other national publications that occasionally comment on testing. For no other subject of public concern—not for economic policy, disarmament, welfare reform, nuclear-power plants—has the professional outlook on a controversy been so shut off from a voice in the national press. Yet, while public policy on testing may not have the immediacy of a tax cut or a nuclear accident, it ultimately affects everyone, as I will try to show.

The one-sided debate about IQ misleads the public and trivializes the subject. Books conceived in ignorance or misunderstanding are evaluated by reviewers who usually know even less. For example, it is often said that testing assumes that intelligence can be quantified by a single number, as in a much-praised recent book‚ The Mismeasure of Man, by a paleontologist, Stephen Jay Gould. To such authors and their newspaper and magazine reviewers, so wrongheaded an assumption must stem from defects in the thought—and perhaps also the character—of testers. The truth, however, is that psychometrics has plainly shown that intelligence is multi-dimensional: people with the same IQ may have different mixtures of abilities. Consequently, for many purposes, tests provide profiles, not a single IQ. The Wechsler Adult Intelligence Scale, for example, which is widely used as an IQ test, can be scored to yield a profile of eleven subscales. Other test batteries result in still larger numbers of dimensions of intellectual variation. To tell the public, as the critics often do, that testers assume intelligence to be one-dimensional is simply to misinform.

At the same time, the public hears little about a surprising and important discovery—not an assumption—of psychometrics. Tests show that many intellectual activities are highly correlated. People who are good at seeing the figure-ground relationships in pictures tend also to be good at sentence-completion exercises, word analogies, problem solving, and so on. Superficially the activities differ, but at a deeper level they converge. Indeed, about 50 percent of the variation from person to person in mental activities can be attributed to variation in the strength of a single attribute, which, because of the great diversity of its expression, is called g, for general intelligence. A person’s level of g has ramifications for everyday life—in school, at work, and in personal matters. Because a standard IQ score is usually a good measure of g, it efficiently tells us something important. The structure of intelligence reveals itself spontaneously in the way people behave intellectually; it is not imposed by the assumptions of testers, as the national press would have it.

My most recent experience with the Times was in early 1979, after it published its fifth article, by my count, about the case of Sir Cyril Burt. Repetitions of the Burt story have appeared several times in its pages since. Not just in the Times but in the newsweeklies, in Science magazine (official publication of the American Association for the Advancement of Science), on national TV news programs, even on the hourly news programs on our local radio station, I kept running into the Burt story, again and again. Everyone has no doubt read, seen, or heard a Cyril Burt story or two, but the true story is much harder to find. When Science published a psychologist’s statistical analysis of Burt’s data, concluding that it revealed traces of deliberate fraud, The New York Times reported the story. When, shortly thereafter, Science published a rejoinder by a statistician, reducing the psychologist’s analysis to shambles, the Times reported nothing. The familiar story is that Sir Cyril invented data to prove that IQ scores are influenced by heredity; once the hoax is exposed, say Burt’s detractors, the genetic case collapses.

The fifth telling of this tale in the Times, by its education editor, prompted me to submit a more accurate account, pointing out that most psychometricians had stopped trusting Burt’s data years before, partly because of inconsistencies first noted in a 1974 article by Arthur Jensen. In his later years, Burt published articles that claimed to be presenting new data but that were probably just rehashes or fictitious confirmations of his earlier results. We can believe to some extent in his earlier findings because they were often the first of their kind and have since been confirmed in many other studies, but for that very reason Burt’s data are expendable. Removing them from the scientific literature would change no important conclusions about the genetic contribution to intellectual development, for if Burt fabricated data, he apparently knew enough to guess correctly. Obviously, it matters that many studies beyond suspicion support Burt’s arguments on behalf of the hereditarian position, but my attempt to say so in a short essay, after countless misleading renditions in all the popular media, was rejected without explanation by the Times.

Reporters from Science and Time magazines, in spite of my urgings, declined to mention in their articles that Burt’s major conclusions have been and continue to be confirmed. For reasons to be outlined later, these reporters appear to want to discredit an entire scientific literature, when the true Burt story would discredit, at most, a single contributor to it.

A STORY NEGLECTED SOMETIMES DAMAGES THE truth more than a story mistold does. On July 30, 1981, Richard Heber and an associate were convicted in a federal court in Madison, Wisconsin, of numerous counts of diverting institutional funds; later they were sentenced to three years in prison. A few months later, a state court sentenced Heber to four additional years. Heber is serving his three-year term in a federal prison in Bastrop, Texas. Thus ended an episode of incomparably greater meaning for testing than all of Burt’s lapses, proven or suspected. But not a word about the Heber scandal has appeared in the Times, the newsweeklies, Science magazine, or on national TV.

Heber was, until the scandal broke in December of 1980, director of the Center on Mental Retardation and Human Development at the University of Wisconsin. He had chaired President Kennedy’s Panel on Mental Retardation in the early 1960s, and had held other responsible positions in the field of intellectual development and compensatory education. Heber is better known, however, for his direction of the Milwaukee Project, an intensive educational and child-care intervention program supported by the federal government, in which the IQs of preschool ghetto children were supposedly raised by more than thirty points. When the President’s Committee on Mental Retardation proclaimed in the early 1970s that the intellectual similarity of parents and children was “mainly because of the environment that the parents create for the young child,” the decisive evidence was Heber’s Milwaukee Project. Here, the committee claimed, was evidence that by intervening directly in the environment, education specialists could raise the IQs of ghetto children to an average of over 125 points. These contentions in the committee’s report were echoed in the national press, which consistently asserts that the environment is primarily, perhaps solely, responsible for the lower IQs of economically deprived groups.

Needless to say, the scientific community has been eager from the start to scrutinize Heber’s work. Reading about it in the popular press is no substitute for the scrutiny that follows the disclosure required by technical journals. But more than fourteen years after the Milwaukee Project started, and long after its spokesman’s claims have been absorbed into the consciousness of editors, politicians, judges, school boards, parents, and social theorists, as well as into countless textbooks and lecturers’ notes, no scientific account has been published. A few unpublished annual reports and fragmentary descriptions are all we have to weigh against the claims made by Heber and those who cite him.

From the sketchy results available, some experts think they see flaws in the study, in the sampling and testing procedures especially. But the plain truth is that we cannot say what was really done for the children or what the results were. The Milwaukee Project broke into the national news when a writer named S. P. Strickland implied, in the July 1971 issue of American Education, a federally sponsored magazine, that the IQs of children in Heber’s study had been raised by 33 points. Soon thereafter, Time magazine told its readers that the Milwaukee Project “offers persuasive evidence that mental retardation in the offspring of mentally retarded mothers can be prevented.” In a syndicated article that ran all over the country, The Washington Post summarized Strickland’s disclosures:

Prof. Rick Heber’s group at the University of Wisconsin may have settled once and for all the question of whether the disproportionate mental retardation of slum children is the result of heredity or environment.

Heber’s study, the Post said,

revealed not that mental deficiencies are passed on genetically, but that mentally retarded mothers tend to create an environment that is less conducive to mental development than that created by slum neighbors of normal intelligence.

The New York Times told its readers:

The Milwaukee Project, an experiment in intensive preschool education for [potentially retarded] children, has proved that they can be raised more than 30 test points higher than other children from the same environment and with same type of mother.

More than ten years passed before a reporter at the Capital Times of Madison discovered that Mr. Strickland, in mid-1971, had become a stockholder and officer in a corporation set up by Heber to sell the very techniques of educational intervention that Strickland credited in his article. But this was a minor revelation, compared with the ones that convicted Heber. A series of Capital Times articles, from January to October of 1981‚ paints a picture of flamboyant misconduct. Heber and an associate, Patrick Flanigan, diverted at least $165,000 of their center’s money into personal bank deposits. Heber’s legal residence was just outside Colorado Springs, though he was paid for “full-time" duties in Madison, Wisconsin. Colleagues were quoted as saying that Heber was rarely seen on campus. He owned hundreds of acres of land in Wisconsin, Iowa, Florida, and Colorado, and was busy in the horse-raising business, which must have involved substantial time, travel, and capital.

Heber’s grants from the federal government for raising IQs totaled in the millions of dollars, much more than the usual federal funding level, and the results he claimed were comparably out of the ordinary. While other workers have published careful studies showing modest, often transitory, benefits of training for IQ, Heber’s project stood as a beacon of hope for dramatic and lasting benefits, and also as the prime evidence against those who say that the IQ has proved disappointingly hard to budge. But the national press, which found Cyril Burt interminably fascinating, has published nothing about the ambiguity surrounding the Milwaukee Project. The media seem unwilling to publish anything that might challenge the certitude with which editors, politicians, judges, and others insist that we know how to increase measurable intelligence, or that test data “prove,” to use The New York Times’s word, that a poor environment causes familial retardation.

SOMETIMES THE PRESS CREATES CONTROVERSY where none exists. An example is an article, published in Newsweek in 1973, reviewing a book by Arthur Jensen. The review was a mild one, but it contained the following sentence: “And [Jensen] predicts that education will eventually be realigned as he suggests— with different curricula and methods for poor and black children.” But Jensen has always advocated individualized education, tailored as much as possible to the abilities and interests of each child. The greater the individual variation in abilities and interests, the greater the benefits of individualized instruction and the greater the harm of educational uniformity, within any social class or ethnic group. Newsweek’s characterization could hardly have been more alien to Jensen’s views; more important, it might easily be taken as a suggestion that Jensen was guilty of elitism and racism.

I wrote to Newsweek’s reviewer (an education editor), Jerrold Footlick, to try to explain the error he had made and to persuade him to correct it. His response granted my point by illustrating it with passages from Jensen, but added:

It seems clear, from these passages and others, that Professor Jensen believes in individualizing education. But from his discussions of our over-emphasis on egalitarianism and from his discussions of different patterns of learning among different social classes, it seems fair to suggest that he believes many poor and black children might need different curricula and methods. I trust that the fact that sotne of them might not fit this description is too obvious to require saying.

Thus, to repeat‚ I do not consider this a serious error. I do not think the review did an injustice to Professor Jensen, and as far as I know, he did not think so either.

The reviewer evidently knew Jensen’s actual position, but apparently not that Jensen objects to “egalitarianism” for scientific, not political, reasons. If children differ, and if education is to be individualized, then schools cannot be “egalitarian.” Moreover, the reviewer’s suggestion that Jensen was not bothered by the review surprised me, and I checked it with Jensen. Not only was he upset about the quoted passage, Jensen said, but he had heard from a halfdozen colleagues around the country who had complained to Newsweek and had, as I had, sent a letter to the editor. None of these letters was ever published, nor was any correction by Newsweek’s writers. I wrote an additional letter to Newsweek’s editor-in-chief, also unanswered. The false record thus stands, obscuring Jensen’s views in a smoke screen of controversy.

By itself, the Newsweek review was not momentous, but multiply its theme a thousand times in a thousand news stories and the result may help explain why American public education is in trouble. To avoid the stigma of “Jensenism,” as the press depicts it, educators must deny significant individual differences in test scores if they are correlated with race or class. But those differences persist, and to deny them is to deny a reality that educational systems cannot ignore without penalty. When education is turned over to those who either do not know how children vary or are willing to ignore that variation, at least in public, the result works about as well as would a public-health policy that ignored individual differences in health. In the case of education, the cost of ignoring individual differences falls most heavily on the disadvantaged, the very people whose interest the Newsweek editor doubtless thought he was serving.

THE CLAIM THAT IQ IS HERITABLE APPEARS TO PROvoke controversy at every turn. But, as controversial as that claim may seem to readers of the national press, it is an ersatz controversy, a creation of the press itself. In the technical literature, virtual unanimity reigns: most of the variation among individual IQs is due to variation in genes. In public discussion, however, the idea that genes account for most variation among IQs seems not only controversial but also associated with a few men of allegedly questionable character. To transform a scholarly consensus into something that appears the obsession of a disreputable fringe group requires more than accidental bias. The transformation, in fact, seems to involve three sets of people, each promoting something it values:

1.The national press favors “sociological” explanations of society, perhaps because when most editors and senior reporters went to school, optimism about the potentialities for social reform was high. Insights from the growing science of human behavioral genetics have yet to reach the top offices of the media. Because the national press is sympathetic to social reforms that are threatened by intractable individual differences, arguments stressing the weight of genetic factors are not welcome, but arguments that minimize the significance of genetic factors are.

2. A few professors are available to provide the antigenetic arguments. They rarely publish their arguments in the technical literature; when they do, the arguments usually fare poorly. They gather little or no data of their own, but instead tend toward ad hominem charges against the scholarly consensus, an example of which is the greatly overblown Burt story. Mostly, they are not psychologists; few of them have genuine psychometric expertise. The more extreme attacks on the heritability of IQ tend to be political, not scientific.

3. The professors just characterized could not so influence the national press without the tacit cooperation of many of the scholars who study, and the merchants who sell, tests. An unspoken agreement grants peace and prosperity, respectively, to scholars and publishers who stay out of the public debate. To explain the timidity of these people, we need look no further than the example of Jensen, who, after devoting himself to the study of educationally significant individual differences, is declared by the press to be an advocate of segregated schooling for “poor and black children.” Rather than try to describe to a journalist the subtle but crucial distinctions between individual and group differences, these scholars describe their results in technical journals and allow merchants to sell their products quietly. When they are contacted by reporters following up on a story, they typically say something equivocally critical about IQ’s heritability, so as to quarantine themselves from controversy and, above all, to protect themselves from false charges of racism and elitism. They usually do not even risk saying, as they truthfully could, that though groups may differ on the average, group differences are small relative to individual differences, and that every sociological stratum and ethnic group contains the full range of mental ability. A familiar example is the professor who assures a reporter that scientists cannot say “exactly” how large a role genes play in determining IQ‚ as if to criticize those who do assign a numerical weight to heritability. What the professor does not tell the reporter is that, although exact values are unknown, a range of possible values, given the data, is quite well established and almost uniformly accepted by experts.

Virtually all specialists on the heritability of IQ estimate a value somewhere in the range of 50 percent to 80 percent. Yet, in mid-1975, CBS produced a documentary, The IQ Myth, that in its narration and selective argumentation denied any evidence of a significant heritable component in IQ scores. The documentary presented snippets of testimony from professors in the second and third groups above, emphasizing the most extreme environmentalist position. I wrote to CBS to protest, and quoted from the 1970 Encyclopaedia Britannica to illustrate the true consensus of experts:

The usually accepted figures, based on the soundest and most extensive studies . . . employing verbal intelligence tests, assign 75% to 80% of the observed test variation in the general population to genetic factors and the rest to environmental physical and mental conditions and opportunities in home and society. Later work . . . using the Culture-Fair tests and improved statistical designs, gives a value of 78:22 to the heredity:environment variance ratio . . .

A CBS spokesman responded that his executive producer knew the Britannica quote and also knew the views of its author (R. B. Cattell, a highly respected psychometrician), but dismissed them as an unsupported “genetics-isall theory,” to use the spokesman’s expression. I answered that Cattell neither believed nor wrote that genetics is all, and I sent two quotes from the 1974 Britannica (also in the 1980 edition), neither by Cattell. The first one below is from a behavior geneticist named W. R. Thompson; the second, from a psychometrician named H. J. Butcher:

Concerning the extent of genetic determination in human intelligence, most investigations have yielded heritability estimates between 70-80 percent. Since such values are relative to the population studied and to the method of estimation, some disagreement should be expected. It seems most unlikely, however, that genotype contributes less than 50 percent of the variability and it is conceivable that the figure is closer to 80 percent.

It has been demonstrated repeatedly that intelligence as assessed by the most highly regarded tests reflects substantial genetic influence. A common finding in European and North American research is that about 70-80 percent of human variation in measured intelligence is ascribable to genetic differences, only 20-30 percent to environmental factors.

I also.sent the spokesman nineteen other quotes from standard textbooks, monographs, and widely cited empirical studies, to indicate the overwhelming consensus on IQ’s substantial heritability. I told him that no comparable set of statements could be found to support the CBS position and concluded with the following paragraph:

My reading of the fairness doctrine suggests that CBS has failed to live up to its legal obligation to “scrupulously eschew intentional and deliberate falsification‚" to quote from a relevant FCC document. In its program on IQ‚ someone at CBS was as deliberate a liar as the topic permitted. You would be wise to chase him down, for he is serving well neither CBS nor the public.

The spokesman’s response was to take offense at my intemperate paragraph and to close our correspondence. Three months later‚ The IQ Myth was aired a second time, without alteration.

Millions saw the CBS TV documentary, but an article by Thomas J. Bouchard, Jr., and Matthew McGue, in Science magazine, in May of 1981, went unnoted by the national press, even though it was a unique contribution to the scientific literature on IQ’s heritability. The brief article presented the world’s data on IQ correlations between relatives—twins, siblings, half-siblings, cousins, parents and children, and adoptive relations. The data were gathered by dozens of scientists and reported in 1ll studies representing more than 225,000 cases, which may be the largest sample ever assembled bearing on the heritability of a complex human trait, physical or psychological. To be included, a study had to meet standards of measurement and publication that, incidentally, excluded Burt’s data. The pattern of correlation fell short of answering all the questions one may have about the inheritance of IQ, as Bouchard and McGue pointed out, but it unambiguously implicated genes. “The pattern of averaged correlations is remarkably consistent with polygenic theory,” said the authors. As far as the public controversy over IQ is concerned, this article should have been big news. Instead, it has become simply one more entry in the technical literature.

Most people would have trouble conceiving of the enormity of the bias against testing in the national press. I am willing, indeed eager, to believe that so severe a bias is atypical; that most aspects of a story usually get told, because of the sheer anarchy of the national press, if for no other reason. But after more than a decade, and scores of contacts with the press, I can count on the fingers of one hand all the journalists I have met who are committed to telling the truth about my field as well as they can discover it. No one could organize censorship so effectively in America, so the distortions in the press must reflect countless more or less independent decisions by editors and reporters which lead to the perpetuation of a misleading and intellectually fallacious understanding of a serious scientific subject.

FIVE YEARS AGO, A NEW YORK TIMES REVIEWER, praising a how-to book for raising IQ, called the claim that IQ cannot be raised a “counsel of despair.”Although the book’s claims are not taken seriously by professional psychologists, the reviewer probably named the major source of bias against testing. What seemed to him a counsel of despair threatens people’s hopes and expectations for themselves, their children, and others they care about. People sense that the range of inequality in intelligence is awfully important; they may even overestimate its importance. They do not want to hear that the range cannot be shrunk. Science paints in shades of gray, which laymen sharpen into black and white and then find depressing. The data say not that intelligence cannot be raised but that raising it much is hard, given existing methods of teaching and measurement. The data say that objectively measured intelligence contributes to, but does not fully determine, success in life’s tasks. People are not much comforted to know that intelligence may be less important for happiness than for success.

Out of personal concerns emerge political beliefs. If people differ genetically in their capacities for school and work, then at least some of society’s inequalities cannot be wiped out by merely equalizing opportunity. And if inequality persists in the face of equal opportunity, the argument against inequality loses at least one dimension of its moral force. Dedicated egalitarians rarely surrender their sense of moral rectitude quietly. Rather than face the resentment of aroused egalitarians (or perhaps because they share that egalitarian hopefulness), editors tend to neglect, and occasionally to denigrate, the findings and conclusions of psychometrics.

Egalitarians share the bias against testing with social scientists who do not like psychology. Some sociologists, economists, and political theorists share a strain of antipsychologism based on the premise that, in the study of society, the important direction of causality is from group to individual, rather than vice versa. Individual psychological variables, such as IQ scores, are a mote in the eye of social theorists whose vision takes in the human landscape as a whole. Such people may be scholars of quality and character, but few of them know much about psychometrics or behavioral genetics. They often cannot evaluate the controversies they read about in the national press, and are not impelled by curiosity or a sense of scholarly necessity to find out enough to do so. They exert little pressure on the press to present the data on individual differences accurately.

Tests are being banned all over the country by courts and legislatures, something that would make sense if they were as vicious and ineffectual as the press says they are. But because IQ is, in fact, so broadly correlated with other cognitive activities, in school and at work, it is virtually impossible to design objective measures of performance not correlated with tested intelligence, and hence subject to the same charges of ineffectuality and bias. Competency testing in high school, civil-service examinations, jobplacement examinations, college-entrance examinations, and diagnostic testing in school have, as a result, been restrained‚ banned, or revised so as to reduce test accuracy. The disclosures of answers to test items from the SAT (Scholastic Aptitude Test) demanded by Ralph Nader and his associates, and passed into law in some states, will probably make the test more expensive and less reliable in predicting college performance. Both the increased cost and the loss in predictive validity will most hurt the poor and disadvantaged, who will be paying more for a test that is less likely to spot unfulfilled native abilities. The more heritable a predictor is, the less it reflects the advantages or disadvantages of a student’s environment, and no predictor of potential academic ability is as heritable as the scores on intelligence tests and their close correlates, the aptitude tests. Hence selection without these tests—which means reliance on grades, conduct, or family connections—is likely to be more biased than selection with them.

Many public school systems have stopped routine intelligence testing, which means that most children will lose the benefits of a proper use of their scores in individualizing instruction. These schools may be spared the risks of improper use, but schools cannot tailor programs of study to children throughout the range of ability if ability cannot be assessed. The restrictions on job-placement tests may be costing billions of dollars annually in lost productivity.

Judges and legislators, relying on the national press and on “experts” who speak more often for political constituencies than for science, assume that the political benefits of banning tests are cheaply won. If tests measure nothing important about individuals, as the press often says, then they are a needless barrier to achieving social and economic equality. But if, as the data say, they measure socially significant individual differences, the cost of banning them is paid mainly by those who might have benefited from better job placement and school curricula. Historically, testing arose out of the desire to break down arbitrary barriers of class, race, and nationality; it was part of the democratization of society at the turn of the century. Better than any other instrument available to teachers, then or now, tests would cut through the unpredictable circumstances of a child’s cultural background to the relatively stable aptitudes on which education builds. The early testers believed that tests would open doors to disadvantaged people, not close them. And that, by and large, is what tests have done, enabling millions of people from poor or deprived backgrounds to develop their abilities better than the circumstances of birth would otherwise have allowed. As tests disappear, displaced by political requirements imposed on school and workplace, arbitrary barriers return; these barriers are new ones, perhaps, but no less unproductive and unjust.

Besides politicizing education and employment, the antitesting bias inhibits the search for useful knowledge about human differences. Powerful pressures—even legal restraints—inhibit analyses showing how individual traits correlate with education and job training, with the conditions of learning and working, and with the quality of people’s personal experience. Instead of learning more about how people function in a world whose intellectual demands are growing, America has all but outlawed the data, encouraged to do so by what it encounters in the national press. □