The IQ Mythology: Class, Race, Gender and Inequality
reviewed by Lloyd G. Humphreys - 1992
Title: The IQ Mythology: Class, Race, Gender and Inequality
Author(s): Elaine Mensh, Harry Mensh
Publisher: Southern Illinois University Press, Carbondale
ISBN: 0809316668, Pages: 214, Year: 1991
Search for book at Amazon.com
This book is made up of opinions of the present authors and of the authors of their references. Quantitative data are almost entirely absent. When data are presented, they are likely to be poor, for example, those of Rosenthal and Jacobsen.1 The pejorative term pseudoscience is used frequently and dogmatically to describe testing, and the principal justification offered is that psychological constructs cannot be measured.
The opinions in the book are also highly one-sided. References to scientific journals or to books by qualified scientists are few in number. For the most part, the authors refer to newspapers, journals of opinion, and books of opinion by persons who are typically ill-informed concerning test theory and the many correlates of scores on intelligence tests.
The book does have the appearance of scholarship in the form of almost twenty-five pages of footnotes, appearing inconveniently in the back, and twenty pages of references, but these do not represent scientific scholarship. Humanists who believe that psychological measurement is worthless, if not impossible, will accept the opinions presented. As one prominent example, consider the almost complete dichotomy between the evaluations by humanists and scientific reviewers of Goulds The Mismeasure of Man.2
The first chapter criticizes the report of a committee, drawn from the behavioral and social sciences, of the National Research Council. The claim is made that the Educational Testing Service (ETS) was instrumental in preparing the report and faults the National Academy of Sciences for this association. One of nineteen members was a black psychologist from ETS. The report is characterized as a sophisticated manual on how to defend mental testing, but the committee overlooked a vital point: mental measurement is a pseudoscience whose practice arouses revulsion. 3
Another opinion expressed, related to a discussion in the committees report and to a quotation from Ravitch,4 is that tests do not locate children of merit and do not promote equality of opportunity. Rather, because of their class and racial biases, they [the tests] sort the test takers in a way that conforms to the existing allocation [of privilege], thus justifying it (p. 5). These opinions are standard fare for most test critics and actually make their first appearance in the present volume on page 1, referring to the correspondence of IQ scores to race and class. Similar statements appear repeatedly, and gender is quickly added to race and class as a third category in which a group is abused by tests.
Chapter 2 criticizes Gould and others who reject so-called biological determinism, but accept deprivation as the basic cause of intellectual differences among persons. A strong environmental position is not enough in the eyes of the present authors; testing should be rejected. Goulds title, The Mismeasure of Man, is the clue to the books position because it suggests that man can also be properly measured (p. 13). Lewontins criticism of Goulds book is quoted favorably on page 16: The Mismeasure of Man remains a curiously unpolitical and unphilosophical book. Gould departed from Lewontins articles of faith in the eyes of Elaine and Harry Mensh as well.
Nothing is to be gained by continuing to quote opinions from this book. Instead, I shall briefly present the quantitative, empirical basis for test use. Intelligence tests are misused, but antibiotics are also misused by physicians and patients. It is a misuse to interpret the score on an intelligence test as a measure or estimate of fixed capacity or genetic potential, but neither intuition nor ideology is a valid alternative for such inferences. Standard intelligence tests do have many correlates of sufficient size for effective use in diagnosis and prediction. Correlations as such are independent of attributions of causality and of beliefs about whether constructs can be measured. Use in an actuarial manner of correlations with intelligence tests that are based on large samples from definable populations is neither pseudoscience nor science. It is more aptly called technology. Theory based on sound empirical research in accordance with the standards for construct validation is helpful, but not essential. The only recourse for a dogmatic test critic is to invoke a value judgment that the measurement of differences among human beings is immoral.
The actual correlations of scores on intelligence tests with socioeconomic or intellectual status of the examinees parents or with being black or white are quite moderate in size. The correlation is slightly more than .40 with a continuous measure of social status and about .50 with parental intelligence test scores in wide ranges of talent. It is slightly less than .40 with race when the ratio of whites to blacks is 4 to 1. The correlation with gender is essentially zero. For many uses of tests the causal bases of these correlations are not important and can be safely ignored.
Persons in the power structure of a stratified and segregated society who want to maintain the privileged status of their children and grandchildren would be fools to depend on intelligence tests for their purpose. Our democratic society can be faulted because we allow the economic and intellectual status of the family to play too large a role in the education of children. This, in turn, is a constraint on occupational access. The other side of the coin is that we pay too little attention to individual merit. Performing effectively in higher education, occupations, and military service represents an important kind of merit that is predicted much more accurately by intelligence and related tests than by either parental social and intellectual status or race.
It is not generally realized how large a stake liberal intellectuals have in criticizing or abolishing intellectual tests. Their children regress half-way, on average, back to the population mean (grandchildren even further) from the superior scores of the parents and show almost as much variability as the population at large. Use of test scores in decision making threatens many of the children of privileged parents.