Measuring Intelligence—Fact and Fantasy
Straight Talk about Mental Tests, by Arthur R. Jensen, New York: The Free Press, 1981, 269 pp., $12.95.
Probably no other scientific controversy has been so badly misrepresented in the public debate, nor allows arrant fantasy and myth to play such determinant roles, as that surrounding mental tests. Partly, this is because the central issues are technically complex and hard for nonspecialists to grasp; partly, it is because such intricate issues do not lend themselves readily to editorial paraphrase and summary without distortion; and partly, it is due to deliberate efforts at confusion and obfuscation. There is thus a great need for a book that discusses these matters clearly in terms accessible to the educated layman.
Such a book is now at hand, and it is written by a man who stands not only at the top of his profession but also at the center of the controversy. The book is Straight Talk about Mental Tests, by Arthur R. Jensen. The questions that constitute the core of the testing brouhaha are here carefully, though simply and clearly, reviewed, including the nature of intelligence as psychologists understand it and why it is socially important, how standardized tests are constructed, what they predict and how validly they do so, the degree of cultural and racial bias in these tests, what is known about environmental determinants of IQ and about raising IQ through specific treatments, the relation of race and social class to IQ, the relative importance of genes and environment to IQ, and the proper and improper uses of mental tests.
Of special interest to the general reader will be the behind-the-scenes discussion of "headline-grabbing" issues of recent years. For instance, a recent, greatly publicized claim by a Ralph Nader-affiliated group asserted that college entry tests are a fraud, since they predict the college grades of admitted students only slightly better than throwing dice would do. The statement is true, but there is a little confidence trick at its heart; admitted students are already a highly selected group whose range of academic ability is much narrower than that of the entire group of applicants. And for making distinctions within this broader group, the validity of college entry tests is solidly established.
Another item much in the news in the last decade was the so-called Miracle in Milwaukee. This was an attempt to intervene heroically in the cultural environment of selected Milwaukee children who were thought likely to become retarded (being born to ghetto mothers who had already produced one retarded child). These children, from birth until they entered school at age six, were given eight hours of intensive stimulation each weekday to speed their mental development. A comparison group of children, chosen by the same criteria, was reared conventionally.
The results? The early gain was spectacular, nearly 30 points on average, and this is what made news. But the gain did not last. Twenty points of it have vanished now that the children are well along in elementary school. More important, while these children are still above the comparison group in nominal IQ, the two groups show no difference at all in the most important academic task of the early school years, namely, learning to read. It seems that the intensive training succeeded in conveying only the specific skills for taking IQ tests without having any detectable effect on the underlying faculty of general intelligence that is normally signified by IQ and whose cultivation was, of course, the true objective of the program. This sobering follow-up evaluation has not been widely publicized.
A third matter of recent media attention is a California court case (Larry P. et al. v. Wilson Riles), where the court enjoined the use of standardized mental tests for the placement of black children in classes for the mentally retarded on the basis that the tests are racially biased. The noteworthy fact is that the court, evidently harking to wider political considerations, made this determination despite a preponderance of evidence to the contrary. What studies actually show is that the scholastic achievement of black children is just as accurately predicted by IQ tests as in the case of white children; a black with a low IQ has the same problems with scholastic material as a white child with the same low IQ. The court decision, be it noted, is under appeal.
Besides busting some widely propagated media myths, the book corrects many other common misperceptions. For instance, the charge is often heard that mental tests, since individual scores on them merely reflect cultural upbringing, reinforce the existing socio-cultural class structure. The truth is just the reverse. A study done in England showed that a greater proportion of working-class children were selected for college preparatory work when selection was based on IQ and achievement tests than when it depended on grades and teachers' recommendations; the opposite was true for middle- and upper-class children, whose deportment apparently biased the teachers' evaluations in their favor. And in the United States, the paramount role played by mental and achievement tests in forcing open educational opportunities for Jews, Negroes, Asians, and other minorities is a matter of history.
Moreover, social classes do not, as far as IQ is concerned, "breed true"; the correlation between a child's IQ and his parents' social class is a rather low .35 (on a scale of 0 to 1), and more than half of all children move to a different social class from that of their parents. On what basis do they move? It turns out to be related to IQ; bright siblings tend to move upward, slow ones downward. The result is that a person's IQ correlates with his own adult social class at a level of .6, which is much higher than to his parents' class. Modern societies, then, are actually characterized by turbulence and flux, not immobility, and mental tests, which can read through social veneers to identify the presence or absence of academic potential, promote rather than hinder this merit-based movement.
Speaking of inheritance, Jensen gives the most lucid explanation of the "heritability" of IQ ever written. Failure to comprehend this basic concept gives rise to many policy confusions, such as the egregiously mistaken belief that high IQ heritability necessarily implies IQ's eternal immutability and hence is cause for inaction and despair, or its inverse, that finding isolated instances where exceptional circumstances do raise IQ—such as Milwaukee's "miracle"—thoroughly invalidates the general finding of high IQ heritability.
Why does all of this matter? Because the efficient allocation of human resources is at least as vital to the functioning of the economy as the allocation of any other resource. The placement test most commonly used in the military, for instance, is the Armed Forces Qualification Test. Studies show that recruits scoring in the lowest third on this aptitude test need two to five times more training trials, and two to six times more prompting to learn basic military tasks than middle- or high-aptitude recruits. Proper placement of recruits is estimated to save more than $400 million per year. An estimate of the savings to the federal government of optimal selection procedures for its four million employees is $16 billion every year. On a national scale, the estimated annual loss in productivity if general ability tests were abandoned for selective purposes would be $80 billion. Clearly, much is at stake.
With the publication of this useful book, there is no longer any excuse for public commentators, to the extent they have professional integrity, to remain ignorant of this matter and to continue to spread misunderstanding.
Bill Havender is a research biochemist at the University of California, Berkeley.