Science in Christian Perspective

 


The Race and Intelligence Controversy
CLAUDE E. STIPE
Department of Sociology and Anthropology
Marquette University Milwaukee, Wisconsin 53233

 

From: JASA 26 (December 1974): 155-164.

Jensen and Shockley have revived the argument that blacks are genetically inferior to whites in intelligence. Jensen has published a number of articles and books which have been responded to by scholars in such disciplines as anthropology, education, population genetics, and psychology. Three areas focal to Jensen's position are discussed here: the concept of race, the nature of intelligence and IQ tests, and heritability estimates. The conclusion is that Jensen uses inadequate data and questionable methods to support his position, which he admits is at best a "not unreasonable hypothesis."

Race and Intelligence

The idea that races differ genetically in intelligence and in their capacity to create and embrace civilization is hardly new. It is periodically presented as a scientifically valid position and each time is criticized because the conclusions are not warranted by the cited evidence.1 Historically, most advocates of the genetic inferiority of blacks have been so blatantly racist that few scientists have felt the necessity of systematically refuting them.

Two current major proponents of the relationship between race and intelligence are William Shockley, a physicist at Stanford, and Arthur Jensen, an educational psychologist at Berkeley. Shockley has been on the lecture circuit for a number of years warning us that our gene pool is being contaminated by genetically inferior blacks and poor whites. Among his suggestions for alleviating the problem is a voluntary sterilization program which would pay low IQ individuals a bonus of $1,000 for each 19 point below 100.2 Is is highly questionable whether Shockley would have received so much attention nationally had it not been that in many instances his scheduled lectures at prestigious universities such as Harvard, Yale and Princeton were either cancelled or broken up by shouting students. Shockley's lack of understanding of genetics, intelligence and race are profound, and few scientists take him seriously. His plea to the American Association for the Advancement of Science in 1967 for a crash pro gram to measure genetic differences in intellectual and emotional traits hetwen racial groups was rejected as not being feasible and as being unlikely to lead to any conclusive results. The AAAS response stated in part:

There is no scientific basis for a statement that there are or that there are not substantial hereditary differences in intelligence between Negro and white populations. In the absence of some now-unforeseen way of equalizing all aspects of the environment, answers to this question can hardly be more than reasonable guesses. Such guesses can easily be biased, consciously or unconsciously, by political or social views.3

It might be said that a new era in the discussion of race and intelligence began in 1969 when the Harvard
Educational Review
published Arthur Jensen's article titled "How Much Can We Boost IQ and Scholastic Achievement?" His article begins with the statement that compensatory education efforts have failed to produce any lasting effects on the IQs and achievement of children, and then suggests that it is necessary to reexamine the premises on which these programs have been based: that IQ differences are almost entirely a result of different environments of the subjects, and that there is a cultural bias in IQ tests. He then argues that genetic factors are much more important than environment in determining IQ scores, and that furthermore there exist "real average differences among the [racial] groups-differences in the population distributions of those characteristics which are indisputably relevant to educational and occupational performance."4 In supporting this conclusion he discusses the nature of intelligence and of IQ tests, estimates of the heritability of intelligence, and the bases for his conclusion that blacks score an average of approximately 15 points below whites on IQ tests. Contrary to the impression given by the mass media at that time, Jensen offers no new data to support his position, only a reorganization of existing data.

A major point made by Jensen is that he has discovered two genetically determined types of learning ability: associative (or rote) learning (Level I), and cognitive or conceptual learning (Level II). He claims that all children have the ability for associative learning, but that not all have the capacity to learn conceptually. Associative learning is seen as characteristic of low socioeconomic status (SES) children and conceptual learning of middle SES children. Since low SES subjects in one test were all black and middle SES subjects were all white, the differences are seen as racial differences. Jensen concludes that we need school systems which will teach children in terms of their abilities to learn, and not require conceptual learning, which could cause failure of those who are capable of only associative learning.

Needless to say, the reactions were immediate, and were both positive and negative. Although it actually comprised a minor portion of the 123 page article, Jensen's views on the relationship between race and intelligence have received the most attention. One discussant states that the article

is one of those signal events that are rare in any field of science: the appearance of a scholarly work that will for years, possibly for decades, he regarded as the watershed that divides a period of misunderstanding, error and myths from a new era when emergence of the true facts led to the formation of a solid theory upon which future scientific progress can be built.5

On the other hand, a black neuro-psychiatrist comments: "As I reviewed this elaborate assortment of truths, half-truths, falsehoods, exaggerations, faulty deductions and speculations, I experienced mixed emotions -including a generous portion of hostility."6 Martin Deutsch, who had co-edited an earlier book with Jensen,7 states that the article contains "many erroneous statements, misinterpretations, and misunderstandings of the nature of intelligence, intelligence tests, genetic determination of traits, education in general, and compensatory education in particular." He then comments that

Perhaps so large a number of errors would not be remarkable were it not for the fact that Jensen's previous work has contained so few, and more malignant, all the errors referred to are in the same general direction: maximizing differences between blacks and whites and maximizing the possibility that such differences are attributable to hereditary factors.8

In a paper of this length it is impossible to cover all the points of Jensen's article which are open to question.9 Instead the focus is on three major areas which are crucial to his hypothesis: the concept of race, the nature of intelligence and intelligence testing, and estimates of heritability.

The Concept of Race

If one is going to deal with the relationship between race and intelligence, one of the crucial concepts is obviously that of race. Almost all who posit such a relationship imply that the races they are comparing are distinguished by biologically relevant criteria. Although the criteria are never stipulated, they so facilely accept this supposition that they never consider it necessary to justify their assignment of a specific person to one race or the other. Jensen does not even consider the problem of definition, but simply states that races are said to be breeding populations which according to geneticists have different distributions of gene frequencies.10 Of course, in technologically advanced societies such as the United States, there are no such isolated breeding populations, so that cannot constitute an operational definition for his study. Neither does Shuey, whose work The Testing of Negro Intelligence Jensen relies on for much of his data, attempt any definition. She only states that "except for small groups of transitional types, the American Negro constitutes a recognizable and clearly defined group, the criterion of membership in which group being that of more-or-less African ancestry.11


Most advocates of the genetic inferiority of blacks have been so blantantly racist that few scientists have felt the necessity of systematically refuting them.


It is important to emphasize that race is a statistical concept, which means that it is extremely difficult to positively assign a given individual to a specific race. As Mayr notes, when one looks at different populations he can see that there are different races, but "how to delimit them, how to draw the line between them is not only difficult, it is impossible."12 In contrast to a racial analysis, Livingstone has suggested the use of a clinical analysis, which would describe all gene frequencies. If, for example the variability of a specific gene is continuous from north Africa to south Africa, the variability cannot be described in terms of race. Also, if two genes vary discordantly, racial classifications based on the one gene will not describe the variability in the distribution of the other.13 Each of us probably differs from his neighbor by approximately 400 genes, yet not more than about a dozen genes can be specified as occurring in one race and not in others .14 When dealing with polymorphic frequencies between populations, it becomes obvious that the extent of variation within any population is usually far greater than the average difference between populations, and that there is a great deal of overlap.

One problem with assigning names to races is that it facilitates the error of typological thinkingthat is, the assumption that individuals in a given group are alike, or at least very similar. It is this concept that the typical man in the street has when he talks about races, and it also seems to be in the back of the minds of those who posit a relationship between race and intelligence. "Physical type, heredity, blood, culture, nation, personality, intelligence, and achievement are all stirred together to make the omelet which is the popular conception of race."15 This attitude is illustrated when Ingle asks whether there are "any biological bases for the failure of nations governed by Negroes and mulattoes to become self-sufficient and creative." He answers his own rhetorical question by maintaining that the "genes representing traits that are important in human affairs are not randomly and equally distributed among racial groups."16

Because of the problems listed above, there are many who feel that the concept of race should be abandoned. As it is commonly used, it is clearly arbitrary, undefinable, and without biological meaning, and as Dobzhansky notes, it is far from being a self-evident cliche.17

Since race is actually a statistical concept relating to groups and not to individuals, how then are individuals assigned to different racial groups in the United States? Obviously they are assigned primarily by social rather than genetic criteria. For evidence we need only note that children of black/white marriages are regarded as black rather than white. It is as if children are considered to have received their genotype from a population rather than from their particular parents. The gene pool of American blacks (defined socially) has been estimated to be approximately 30% derived from white ancestry.18 It has also been estimated that 70% of blacks have at least one white ancestor and that 30% of American whites have at least one black ancestor.19 If we choose to call the white individual with a black grandfather a Negro, then logic would require us to call the "average" black in Baltimore or New York a Caucasian. It has been suggested that the classification which a person gives of himself should be used, but if race is considered to be a biological construct, the lay person's view of his own racial identity is both incompetent and irrelevant.

An apocryphal story is told about an American newspaperman who had an interview with the President of Haiti.

They started to talk about Haiti and its population, and most indiscreetly the American newspaperman asked the President of Haiti what percentage of the people were white. And the President of Haiti said, "Oh, about 95 per cent." The American newspaperman looked a little puzzled and said, "Well, how do you define white?" And the President of Haiti said, "Well, how do you define colored?" And the American newspaperman said, "Well, of course anybody with Negro blood is colored." Said the President: "Well, that's exactly our definition too: anybody with white blood is white."20

The same situation obtains in Brazil-anyone who is not "pure black" is a Caucasian-21


Race is a statistical concept, which means that it is extremely difficult to positively assign a given individual to a specific race.


We do not have an accurate count of the so-called Negro population in the United States, because census takers are no longer allowed to ask the race of the interviewee. This means that the taker is asked to make a clinical judgment concerning race that a trained physical anthropologist would hesitate to make. If he is in doubt, his instructions are to assign race according to the prevailing racial composition of the neighborhood.22 Since skin color is the major basis for classification, studies imply a direct relationship between the genetics of skin color and the genetics of intelligence.

In light of the above problems, is it correct to say that race is important in our society? Unfortunately the answer is yes. Races are real "because people believe they are, and social reality-the human worldis determined by human belief."23
The crucial point in this discussion of race is that most of the studies cited by Jensen and others have used a social rather than a genetic definition of race. Under such conditions, how can one demonstrate a genetic relationship between race and intelligence?

Intelligence Testing

A second major concern is the nature of intelligence and the validity of intelligence test scores. Jensen insists that there is no point in arguing the question to which "there is no answer, the question of what intelligence really is,"24 but that it is possible to measure intelligence, for "intelligence, by definition, is what intelligence tests measure."25 He limits the term intelligence to that which is represented, by Spearman's g, which he claims is the factor common to all tests of complex problem solving.26 Although he considers it difficult to define intelligence in so many words, he considers that "it is probably best thought of as a capacity for abstract reasoning and problem solving."27

The Binet Test was originally devised to help Paris educational authorities separate out the "dull" children who would not benefit from the education offered at that time, Both Binet and Simon warned that the results were useful only when the tested children came from similar environments. This warning has been often repeated, but seemingly more often ignored. Jensen acknowledges that IQ tests have been developed for the express purpose of determining success in school, but insists that they not only measure school learning or cultural advantages making for scholastic success, but also tap fundamental psychological characteristics.28 He maintains that although intelligence has been singled out as especially important by educational demands, it is nevertheless "a biological reality and not just a figment of social convention."29
Jensen's argument that most intelligence tests measure g is not universally accepted. Jastak maintains that most IQ scales include a balanced number of associative and conceptual tests, and that furthermore, the associative tests are more highly correlated with IQ than are the conceptual ones.30 Deutseh strongly objects to the use of Spearman's g as the basis for intelligence, stating that

g represents only one theory of intelligence, among many others. It is by no means a universally accepted concept among psychologists and others who work in this area. Yet from Jensen's paper, the general reader would never know that there are competing theories, several of which are more widely accepted and based on more recent information and data than Spearman's.31

Deutsch suggests as an alternative Piagct's theory, which is based on intellectual development interwoven with the child's experiences. Basically it is a stage theory with each new level building on the previous one. This approach reflects process, whereas Deutsch evaluates Jensen's notions of level as categorical and static. He claims that Jensen's whole line of reasoning is inextricably linked with the concept of g, and questioning g throws doubt on his whole system.

Although Jensen claims that the dispute over whether or not intelligence is fixed is a spurious one, 32 he writes as though IQ is essentially fixed. He defines intelligence as 'performance estimate," but explicitly and implicitly keeps falling into the error of treating it as a measure of "potential ability." IQ tests were constructed with the assumption that intelligence is fixed. Since this assumption resulted in an attempt to find test items which yield constant scores, the constancy may be due more to the tests than to the subjects tested. Although Jensen claims that it is very difficult to change IQ scores by compensatory education,

It seems quite wrong to attach so much importance to a change-or an absence of change-in IQ scores, when the test is designed both on the assumption that what is measured is a fixed and unchanging characteristic, and with the aim of producing constant IQ scores. Such a use of intelligence tests suggests insufficient appreciation of the assumptions implicit in their construction.33

There certainly is some irony in the fact that people who see IQ as essentially static in nature should use it as a measure of intellectual change.

Bereiter worked in programs for disadvantaged children in which the main purpose was to directly teach academic skills,

Nevertheless, in spite of the fact that the program was never intended to raise IQ and that twothirds of it was devoted to reading and arithmetic instruction having littie or nothing to do with the skills called for on IQ tests, significant IQ gains have been regularly obtained. Over the last four replications they have averaged about 15 points. This seems to he too much of a gain to write off to test-wiseness and things of that sort, especially since the children's IQs were in the middle nineties to begin with and thus rose to substantially above average.34

Rex suggests that the field of psychometrics is possibly the least sensitive of empirical human studies. Psychometricians claim that no assertions are made about essential intelligence, only about measured intelligence.35 Although they may pretend that this measured intelligence has no reference to practical, social and political implications, the repercussions for the tested individual can be crucial, as the following example illustrates,

The junior-high-.school counselor urged Ralph to drop some of his academic subjects on the ground that he was not "college material." The boy's dissatisfactions led his parents to see the counselor, who reported to them that Ralph's test results showed him to be "average" in ability. When questioned about the accuracy of the results, he argued that the boy's grades were consistent with the IQ test scores. This seemingly ironclad logic fell apart about eighteen months later. After the use of private reading and psychological specialists, Ralph's grades ruse from C's to B's and his IQ test scores from approximately 100 to 120 on comparable tests.36

Although it is not claimed that every student can be helped in this way, it does show the fallacy of supporting the validity of the IQ score by reference to school grades and vice-versa.

It was mentioncd above that Jensen claims to have isolated two genotypically distinct basic processes which underlie learning. Since conceptual (Level 11) abilities have been most important for scholastic performance under traditional methods of instruction, he postulates that the genetic factors involved in each of these types of ability have become differentially distributed in the population. 37 He also states that because of this genetically different ability, ordinary IQ tests tap only one part of the spectrum of mental abilities of disadvantaged children. Therefore the low scores they obtain on IQ tests are not unfair to them because they yield inaccurate or invalid measurements, but because the tests measure abilities which they basically lack.38 Because of this situation

The ideal of educational opportunity should not be interpreted as uniformity of facilities, instructional techniques, and educational aims of all children. Diversity rather than uniformity of approaches and aims would seem to be the key to making education rewarding for children of different patterns of ahility.39

Jensen fails to mention that in many cases children are now taught in one or the other style of learning. Brazziel notes that in the American south both black and white schools use an associative (rote) learning style. Since most of the urban northern black parents were reared in the south and therefore learned in this way, they pass this learning style on to their children. However, he states that if conceptual learning is used early in their educational experience, children can be conditioned to think in that manner. He is afraid that if teachers accept Jensen's position, many children will never have a chance to be exposed to conceptual learning styles.40

It is very interesting to note that Jensen and Rohwer have collaborated on studies of paired-associate learning, and yet their interpretations of the results are diametrically opposed. In these studies a student is presented with 20 word pairs, such as "elephant and pillow," and in subsequent trials is supposed to recall the second word when given the first. It was found that if a child was taught to elaborate on a pair of words, such as "the elephant is looking for his pillow," the rate of performance increased sharply. When this method was used with children from varied backgrounds, there was no class distinction in performance, even though the IQs differed. Jensen attributes the results to the fact that associative learning is genetically more evenly distributed in the population. Rohwer, on the other hand, argues that this experiment stresses the importance of an active process of learning which includes not a rote principle, but a highly inventive process of imaging.41 It was also found that simply coaching lower-class children to make up "meaningful" associations for word pairs brought them up to the middleclass rate of learning.42

Since Jensen has concluded that the heritability of intelligence is 0.80, he leaves very little room for the influence of environment. On the other hand, most researchers realize that there are many ways in which environment may be a factor. It is impossible to study cognitive abilities in isolation from their social and motivational determinants, and if this were explicitly recognized, it might remove much of the mystery surrounding the interpretation of IQ test scores. Even the experienced black interviewer raised in Harlem gets only monosyllabic, defensive answers when questioning Harlem boys lie knows in an indicidual setting, but in discussion with their peers, the same children reveal intricate verbal reasoning skills in talking about their experiences.43

A large number of statistically significant correlates with IQ have been discovered, only a few of which are; height, weight, anxiety level, race and warmth of examiner, mother's attitude toward achievement, mother's concern for language development, home cultural level, father's occupation and years of schooling, and desire to master intellectual skills.44 In the testing of laboratory animals there are many examples of situations in which the apparent strain differences in learning, activity levels, memory, etc., were found to stem from differences in fearfulness or some other circumstance peculiar to the test situation. Certainly this would be especially crucial in testing human behaviors, where the same testing conditions may tap different functions in different subjects or groups of subjects.45

There are a number of studies which show that the IQ of twins is systematically lower than that of non-twins by about five points,46 and that firstborn children have higher IQ scores than their siblings. Firstborn children are overrepresented among high IQ students, and also among a wide variety of both intellectual and career high achievers. More than half of the National Merit Scholarship finalists belonging to families with from two to five children were firstborn. If one considers only those from five child families, 52% were firstborn, and only 6% were fifthborn.47 A common explanation is the reduced attention that parents can give when they have more than one child, but in keeping with his general approach, Jensen discounts sociopsyehological explanations. He attributes the differences to biological factors, although he does concede that it "almost certainly is not a genetic phenomenon."48

Although sensory deprivation in the black population is often cited, Jensen is not concerned with it. He only cites the black writer Kristin Hunter to the effect that ghetto babies must be the most thoroughly loved in the world, for they are cradled, cuddled, tickled, and passed from one set of loving arms to another, and then comments that "this does riot sound like sensory deprivation."49 Yet Ward's study of a black community in the south shows that when a child starts talking it is no longer considered to he a baby, and after that time there is little verbal interaction with adults.50

Another possible environmental factor is the dearth of institutions in the black community that function routinely at advanced cognitive levels. For example, there are too few black colleges, professional people, newspapers, etc. As the children grow older, "their environment is progressively less rich in the frequency of use of the next developmental cognitive structure."51 Certainly the situation in which IQs drop solely as a function of age points to the environment as a major factor. 52

Rosenfeld has graphically demonstrated that the attitudes of educational personnel in Harlem schools affect the performance of black children, not only on standardized tests, but everyday in the classroom.53 There seems to be little question that the self-fulfilling prophecy is at work the belief that blacks are intellectually inferior can cause both blacks and whites to behave in such a manner as to yield confirmatory evidence. Most of the experienced faculty are assigned to superior classes, whereas the poor students are saddled


A large number of statistically significant correlates with IQ have been discovered, only a few of which are: height, weight, anxiety level, race and warmth of examiner, mother's attitude toward achievement, mother's concern for language development, home cultural level, father's occupation and years of schooling, and desire to master intellectual skills.


year after year with less qualified teachers.

Statistics are available to support the contention that poorly supported schools and poor pupil performance tend to be correlated. In 1963-1964 the mean expenditure per pupil in the United States was $455. In Kentucky the average was $300, and in depressed rural areas it was even less. In the High School Achievement Test on which the national average score is 100, 69% of the school districts of eastern Kentucky had average scores below 80, and none averaged as high as 100.54

Most researchers recognize that it is not valid to compare IQ scores of lower and middle class children because of their different socio-cultural environments, Therefore Socioeconomic Status (SES) measures which use schooling, occupation and income of parents have been devised in an attempt to equalize environments. The implication is that when there is a matching of SES between a sample of blacks and whites, any differences in IQ scores are due to genetic differences. Although Jensen assumes that SES measures actually match environments, this is not generally accepted. It is extremely difficult to accomplish standardization, and the fact is that "comparable groups" have never been standardized even for simple physical health or for nutrition during pregnancy.55

Even if one takes a given cut off point for the middle class, those few blacks who are middle class will be mostly very near the cut off, whereas many whites will be quite far from it. It follows that the mean class position of middle class blacks is likely to he much lower than the mean class position of middle class whites. Thus even if the class measures were perfect, the groups would not be equated.56 It should he obvious that black and white families with identical incomes do not have equal access to economic options. Housing discrimination is only the most outstanding example of the fact that a dollar may be worth less to a black family than to a white one. With the disparity in quality of educational institutions, equal numbers of years of education do not imply equal educations, nor do they imply equal access to future benefits such as income levels or jobs. Therefore, blacks and whites have not been successfully matched for social class.

Bodmer concludes that the only real approach to determining the effect of environment on IQ would be to adopt black children into white families and vice versa, after which a comparison would he made of the IQs of black children adopted into white families with white children adopted into comparable white families. He doubts that even this kind of experiment could be controlled, because it still would not remove the effects of race prejudice against blacks which exists in most white communities.57 Rex maintains that any studies in which one attempts to equalize environments should be supplemented by an experiment in which 

the peoples of Africa conquer, capture and enslave sonic millions of European and American whites under conditions in whirls a very large proportion of the white population dies and in which the white culture is systematically destroyed, and in which finally a group of emancipated whites living in "good neighborhoods" are then compared to their Negro masters.58

Washburn has long argued that when comparing IQs between groups of people the same criteria which are applied to comparisons between white groups should be applied to comparisons between blacks and whites, If one looks at the literature, he finds that when two groups of whites differ in their IQ scores, the explanation is immediately sought in SES differences, but when blacks and whites differ in precisely the same way, the differences are said to he genetically determined."59

In attempting to correlate race and IQ, Jensen relies primarily on the results obtained by past studies of IQ. He seems to have no qualms in accepting the results of some 380 investigations of black intelligence reported in Shuey, which provide the basis for the conclusion that blacks average 15 points lower than whites on IQ tests. Since Jensen states that he puts very little confidence in a single test score-especially if it is the first test of a child from a poor background and of a different race from the examiner-it is surprising that he does not question the validity of the test results in Shuey. In his own testing programs Jensen usually spends from two to four sessions of one-half hour each to ensure that the child feels at ease in the testing situation. There is regularly an increase of from 8 to 10 IQ points between the first test and the subsequent one given after the orientation sessions.60 Jensen certainly most be aware that these conditions were not present in the studies cited by Shuey, and yet he accepts the data as being valid.

The amount of credence Jensen places in test results seems to depend on whether or out they agree with his suppositions. For example, in commenting on the studies which have demonstrated that high birth rates among lower socioeconomic status people have not led to a general decline in average IQ, he states that the studies are far from adequate to justify complacency, and that they cannot be generalized beyond the specific generation studied or the white population upon which the studies were made.61 On the other hand, studies made in the 1920s are used to substantiate his arguments, and he even applies to black IQ the heritability estimates standardized no the white population.

Even if one were to accept the data as correct, it is difficult to understand the importance of the average racial differences in IQ upon which Jensen places so much emphasis. He sees the 15 point average difference between blacks and whites as crucial, yet acknowledges that the average between siblings is about 12 points, and that some 20% of siblings differ by more than 20 IQ poiots.62

Heritability

The pivotal facet of Jensen's argument is his estimate of the heritability of intelligence, which has been derived primarily from the studies of moonzygotic and dizygotic twins raised apart. Estimates of broad heritibility answer the question: "What fraction of the variance of a phenotypic trait in a given population is caused by (or attributable to) genetic differences?"63 Jensen denies that the problems in measuring intelligence affect determining its heritability.

Whether we can or cannot measure intelligence ... let it be emphasized that it makes no difference to the question of heritability. We do not estimate the heritability of some trait that lies hidden behind our measurements. We estimate the heritability of the phenotypes and these are the measurements themselves. Regardless of what our tests measure, the heritability tells us how much of the variance in these measurements is due to genetic factors.64

However, tests of IQ differ from measurements of conventional phenotypic characters in two different ways.
IQ tests are analogous to physical readings made with a black box-a device whose internal working is unknown. Since it is not known what an IQ test or a black box measures or how it works, it is impossible to know to what extent the measurements carried out on different subjects are comparable or to what extent they are influenced by extraneous factors. "Thus IQ scores contain uncontrollable, systematic errors of unknown magnitude."65 IQ scores also differ from conventional phenotypic characters in that they have no strict quantitative meaning. IQ is a rank order on a standardized test, and the intervals have been chosen so that the frequency of test scores in a reference population will be approximately normal.

The heritability index used by Jensen has been widely questioned. Reservations have been expressed about both the quantitative validity of the methods and the reality of the necessary assumptions. It is recognized that the choice of a particular statistical procedure with which to handle data to a great extent also chooses the result to be obtained .66 Since heritability measurements are somewhat arbitrary, alternative methods of computing the fraction of genetic variance for intelligence might he less striking. Jensen concludes that the heritability of intelligence is 0.80,67 but using the same data Cavalh-Sforza suggests an estimate of between 0.40 and 0.60,68 and Jencks has stated that the chances are about two out of three that the heritability of intelligence is between 0.35 and 0.55.69

Since Jensen emphasizes the studies of monozygotic twins raised apart to support his estimate of heritability, it is instructive to note the actual IQ scores rather than just the overall average. In the four existing studies of monozygotic twins raised apart, average differences range from 6 to 14 points. In one sample of 38 pairs, at least 25% had within-pair differences of 16 IQ points on one of the tests,70 and the average difference in those 38 pairs was 14 points.71 This is only one point under the average difference between blacks and whites, and mnnozygntic twins have identical genotypes!

It might seem likely that a comparison between monozygotic and dizygotic twins raised apart would lead to a measure of the relative importance of genetic and environmental factors. However, there are two problems with such a method. First, the differences between the dizygous pairs represent only a fraction of the genetic differences which can exist between two individuals, that is, they are more related than two individuals taken at random from the population at large. Second, the environmental differences between monozygous twins encompass only a fraction of the total environmental differences which can exist between two individuals, that of the family. "In short, whereas the contrast between monozygous and dizygous twins minimizes genetic differences, it also tends to minimize environmental differences."72

A major problem is that heritability estimates have been derived only from studies on European and North American white populations. Jensen is aware of this and states that such estimates are "specific to the population sampled, the point in time, how the measurements are made, and the particular test used to obtain the measurements. "73 He also recognizes that estimates represent average values in sampled populations and do not necessarily apply either to differences within various subpopulations or to differences between subpopulations.74 Nevertheless, he uses those estimates to discuss the differences in "inherited" intelligence between blacks and whites, and states that when used in conjunction with other information about the amounts of "relevant environmental variations within groups and overlap between groups," they can be used to formulate testable hypotheses that could "reduce the heredity-environment uncertainty concerning group differences."75

There are at least two implications of the fact that heritability is a population statistic. Since estimates depend on the extent of genetic and environmental variation in the population at the time it is studied, they are invalid not only for other populations, but also for the same population at a different time.76 Also, since estimates represent an average of the individuals who make up the population, there is no way in which they can be used to predict how much any given individual will he affected by a change in his environment. There may possibly be a few relatively rare optimum environments in which a given individual's performance would be extensively modified.77 Jensen carefully notes that heritability is a population statistic, and therefore has no sensible meaning with reference to a measurement or characteristic in an individual, which makes it impossible to partition a given person's IQ into hereditary and environmental components, e.g., 80% due to heredity and 20% due to environment. However, he does claim that one can make a probabilistic inference concerning the average amount of difference between the obtained IQ of an individual and the "genotypic value" of his intelligence.78

A second major problem is that of genotype-environment interaction. Since differences in IQ are certainly related to some extent to genetic differences, it is safe to assume that the genotype-environment correlation is significant in subpopulations composed of children raised by their biological parents or close


Since it is not known what an IQ test measures or how it works, it is impossible to know to what extent the measurements carried out on different subjects are comparable or to what extent they are influenced by extraneous factors.


relatives. If this is true, then estimates of heritability based on data which refer to such subpopulations cannot he valid, and yet the bulk of all the available data is precisely of this kind. In one set of twins data, two-thirds of the separated pairs were placed with members of the family, 79 and in the largest and most homogeneous of the four major twin studies (that of Burt), one member of each of the 53 pairs included in the study was raised by his or her natural parents.80 Correlations between separated monozvgotic twins would be highly sensitive to distortion by genotype-environment correlation, and yet in the published studies, no serious attempts have been made to minimize the effects of genotype-environment interaction.

Genotype-environment interaction can only be controlled by randomizing environments, and heritability estimates would be completely valid only if each possible genotypic child were placed randomly in each conceivable environment. At the least one would have to include black children in a representative range of environments. Hirsch is concerned with the number of possible genotypes and the number of potential environments which may influence trait expression. If one takes the next to smallest ease-two genotypes in three environments or three genotypes in two environments-there are 60 types of interactions. He concludes that

it is ridiculous to attempt to characterize an environment as generally favorable or unfavorable, or any genotypes as generally superior or inferior. Some average measure of an environmental influence is applicable only to those genotypes affected by it in the same way. Similarly, any rank ordering of genotypes can be applied only to those environments which preserve the ranks of their phenotypes.81

Layzer maintains that the only potentially useful data are phenotypic correlations between foster children raised together, which could possibly provide the lower limits for the effect of environment. The available data on such children suggest a broad heritability between 0.0 and 0.5.82

It seems to be generally agreed that certain studies of genotype-environment interactions vitiate heritability studies. For example, it has been noted that amino acid excretion patterns of monozygotic twins was less affected by separation than was the pattern of dizgotic twins, and a suggested explanation is that monozygotic twins select similar diets and environments. If this is a general phenomenon, the limits of behavioral genetics are very broad, because one would not ordinarily consider amino acid excretion to be a behavioral trait.83

A third major problem is that of gene-gene interaction. Although Jensen is aware that some traits may be truly polygenic, he prefers the assumption that intelligence is not polygenic, but rather is due to a number of genes with additive effects.

It may be more heuristic . . to work on the hope that the trait in question, though seemingly polygenic, is po
tentially analyzable into a number of Mendelian characters . . . If it is a false hope, we can find out only by trying. I know of no scientific laws or principles which a priori make it a false hope.84

However, the greater the number of genes which contribute to a given trait, the more likely it is that nonadditive genetic effects will play an important role. It is therefore possible that human intelligence depends on the total genotype in a way that is too complex for the application of conventional heritability analyses.

Even if it could be demonstrated that there is a high heritability of intelligence in both black and white populations, that would still not be convincing evidence that the differences between the groups are genetic.

No matter how high the heritability (unless it is 1), there is no assurance that a sufficiently great environmental difference does not account for the difference in the two means, especially when one considers that the environmental factors may differ qualitatively in the two groups.85

Conclusion

In reacting to his detractors, Jensen argues that reasonable hypotheses concerning questions which are socially and educationally relevant should he appropriately investigated and that the findings should be published and widely discussed not only by the scientific community, but the "general public as well."86 This raises the question of the educational relevance of his hypothesis that blacks as a group are genetically inferior in intelligence. Jensen reiterates the caveat that we must maintain the distinction between the individual and the population when discussing racial differences in mental abilities, and that all persons must be regarded in terms of their individual qualities and merits. However, the implications of his hypothesis prevent the drawing of this distinction between individuals and groups. If each child is to be taught according to his potentialities, he would have to be tested individually for the ahilty to learn in one way or the other, and his position in a racial or ethnic group would be irrelevant. Yet Jensen includes no suggestions for identifying a potential conceptual learner other than by noting skin color. He states that

The question of race differences in intelligence comes up not when we deal with individuals as individuals, but when certain identifiable groups or subcultures within the society are brought into comparison with one another as groups or populations. It is only when the groups are disproportionately represented in what are commonly perceived as the most desirable social and occupational roles in society that the question arises concerning average differences among groups.87

Since learning is essentially an individual rather than a group process, it is not clear how one's membership in a racial or ethnic group is educationally relevant.

It is not surprising that Jensen's work is seen by many as an attack on special education programs for blacks.88 They would agree with a statement written before publication of Jensen's article that

With indecent haste, evidence is adduced to raise suspicions about the alleged inferiority of a people before society has completed even the early stages of correcting the inequities and the consequences of centuries of inequality in every form. Their inferiority is established be/ore they have a chance to prove otherwise,89

And, in fact, Jensen's arguments have been used in opposition to the funding of various poverty programs.90

Jensen accuses his critics of demanding almost impossible criteria of certainty before proposing and investigating genetic hypotheses as opposed to environmental ones,91 and complains that no one has yet produced any evidence on a properly controlled sample to demonstrate that it is possible to equalize the intellectual ability in representative samples of black and white children through statistical control of environment and education.92 Situations which tend to support the importance of environment are interpreted in genetic terms. For example, he refers to a 6th grade class in the Windsor Hills Elementary School in Los Angeles, which was 90% black and in which the mean IQ score was 115. Rather than considering this as possible evidence that environment affects performance on IQ tests, he asks "why should anyone be surprised that there are Negro children having IQs of 115 or higher, or that they should he concentrated in the affluent integrated neighborhood of Los Angeles?"93 He then "explains" the situation in terms of the superior genetic endowment of the parents rather than as a result of a good environment.

It is relatively easy to gain the impression that Jensen feels that his hypothesis has been validated by the data. Yet this is not so, for he states:

So all we are left with are various lines of evidence, no one of which is definitive alone, but which, viewed all together, make it a not unreasonable hypothesis that genetic factors are strongly implicated in the average Negro-white intelligence difference.94

Since he is convinced that the research to test his hypothesis is entirely possible, but just has not been done,95 one wonders why he did not conduct the research rather than publish so many articles and books based on the same inadequate data. It would seem that if Jensen were really concerned with an unbiased testing of the heritable component in intelligence differences between human groups, he would have attempted to develop an operational definition for equal social conditions. He then could have devised scientific experiments in which a systematic effort would he made to ensure that his subjects fit that definition.

One attempt to explain Jensen's approach is that he has "girded himself for a holy war against 'environmentalists,' 96 and that his position is must understandable when seen as a reaction against the "iniquities" of his professional colleagues who take an essentially environmentalist viewpoint,97 and whom he accuses of having branded him as a "moral pariah" for his approach.98

Jensen sees little validity in the many criticisms of his work. He refers to the "storm of ideologically, often politically, motivated protests, misinterpretations, and vilifications" which his article prompted,99 and characterizes the majority of the letters and articles in the summer 1969 issue of the Harvard Educational Review as "only masquerading as serious critiques of my article."100 Eysenck, who is Jensen's British counterpart, claims that "truly competent judges" have not found any serious error in Jensen's statements,101 and in 1972 Jensen saw no reason to substantively revise any of the main points of his 1969 article.102

Jensen is unhappy with critics who are concerned with the moral, political and social policy dimensions of a general acceptance of his hypothesis. He insists that such issues should be kept clearly distinct from the scientifically answerable aspects of the question,103 giving the impression that he sees the situation only as representing a normal scientific exchange between the representatives of two alternative scientific theories. But as Fried points out:

This is not a question of digging the "Mohole" or not, or whether Homo habilis is or is not an Australopithecus. It is more like dividing on the question of whether or not to exterminate six million Jews: one side says no and presents its arguments, and the other side says yes and presents its arguments, and this too becomes a debatable scientific question.104

REFERENCES

1Cf. Klineberg, Otto. 1935. Race Differences. New York: Harper and Brothers; Comas, Juan. 1961. " 'Scientific Racism' Again?" Current Anthropology 2:303-14; Stipe, Claude E. 1964. "Race and Culture: A Valid Basis for Segregation?" Journal of the American Scientific Affiliation 16:36-41.
2Tobach. E. 1972. "The Meaning of the Crytanthroparian," in Genetics, Environment, and Behavior: Implications for Educational Policy, Edited by Lee Ehrman, Gilbert S. Omenn and Ernst Caspari. New York: Academic Press. p. 229.
3Craw, James F., James V. Ned and Curt Stern, 1967. "Racial Studies: Academy States Position on Call for New Research," Science 158:893.
4Jensen, Arthur R. 1969a. "How Much Can We Boost IQ and Scholastic Achievement?" Harvard Educational Review 39:79.
5Freeman, Roger A. 1969. "Letter to the Editors," Harvard Educational Review 39:599.
6Nelson, James D. 1969. "Letter to the Editors," Harvard Educational Review 39:615.
7Deutsch, Martin, Irwin Katz and Arthur R, Jensen (Eds.). 1968. Social Class, Race, and Psychological Development. New York: Halt, Rinehart and Winston,
8Dentsch, Martin. 1969. "Happenings an the Way Back to the Forum: Social Science, IQ, and Race Differences Revisited," Harvard Educational Review 39:524,
9A list of 117 articles and books written in response to Jensen's l969a article is given in Jensen, Arthur R. 1972b. Genetics and Education. New York: Harper and Row. pp. 356-364.
10Jensen, Arthur R. 1969a:80.
11Shuey, Audrey M. 1966. The Testing of Negro Intelligence (Second Edition). New York: Social Science Press, p. 3,
l2Mayr, Ernst. 1968. "Discussion," in Science and the Concept of Race, Edited by Margaret Mead, et al. New York: Columbia University Press. p. 103.
13Livingstane, Frank B. 1964. "On the Nonexistence of Human Races," in The Concept of Race, Edited by Ashley Montagu. New York: Free Press. p. 54.
14Glass, Bentley. 1968, "The Genetic Basis of Human Races," in Science and the Concept of Race, Edited by Margaret Mead, et. al. New York: Columbia University Press. pp. 88-89.
15Maotago, Ashley. 1962. "The Concept of Race," American Anthropologist 64:920.
16Ingle, Dwight J. 1968. "The Need to Investigate Average Biological Differences among Racial Groups," in Science and the Concept of Race, Edited by Margaret Mead, et al. New York: Columbia University Press. pp. 113-114.
I7Dobzhansky, 'l'hcodosius. 1968. "Discussion," in Ibid. p. 78. 'Cf. Glass, Bentley. 1968:92; Badmer, W. F. 1972. 18Race and IQ: the Genetic Background," in Race and Intelligence: the Fallacies Behind the Race-IQ Controversy, Edited by Ken Richardson and David Speers. Baltimore: Penguin Books Inc. p. 90.
19Downs, James E. 1971. Cultures in Crisis. Beverly Hills: Gleocne Press. p. 3.


Even if it could be demonstrated that there is a high heritability of intelligence in both black and white populations, that would still not be convincing evidence that the differences between the groups are genetic.


20Mayr, Ernst. 1968:104.
21Erlich, Paul B. and Richard W. Halm. 1964. "A Biological View al Race," in The Concept of Race, Edited by Ashley Montagu. New York: Free Press. p. 173.
22Gottesman, I. I. 1968. "Biogeneties of Race and Class," in Social Class, Race, and Psychological Development, Edited by Martin Deutsch, Irwin Katz and Arthur R, Jensen. New York: Holt, Rinehart and Winston, p. 21,
23Brace, C. Loring. 1971. "Introduction to Jensenism," in Race and Intelligence, Edited by C. Loring Brace, George R. Gamble and James T. Bond. American Anthropological Association: Anthropological Studies No. 8. p. 5.
24Jeosen, Arthur R. 1969a:S-6.
25Jensen, Arthur R. 1969a:9,
26Ihid.
27Jensen, Arthur R. 1969a:19.
28Jensen, Arthur R. 1969a:8.
29Jensen, Arthur B. 1969a: 19-20.
30Jastak, Joseph F. 1969, "Letter to the Editors," Harvard Educational Review 39:609.
31Deutsch, Martin. 1969:542.
32Jensen, Arthur R. 1969a: 17.
33Ryan, Joanna. 1972. "The Illusion of Objectivity," in Race and Intelligence: the Fallacies Behind the Race-IQ Controversy, Edited by Ken Richardson and David Speers. Baltimore: Penguin Books Inc. p. 45.
34Bereiter, Carl. 1969. "The Future of Individual Differences," Harvard Educational Review 39:315.
35Rex, John. 1972. 'Nature versus Nurture: The Significance of the Revived Debate," in Race and Intelligence: The Fallacies Behind the Race-IQ Controversy, Edited by Ken Richardson and David Speers. Baltimore: Penguin Books Inc. p. 168.
36Schwebel, Milton. 1968. Who Can Be Educated. New York: Grove Press. p. 74.
37Jeosen, Arthur R. 1969a: 114.
38Jenseo, Arthur B. 1961a:113.
39Jensen, Arthur R. 1969a: 117.
40Brazziel, William F. 1969. "A Letter from the South," Harvard Educational Review 39:355.
41John, Vera P. 1971. "Whose is the Failure?" in Race and Intelligence, Edited by C. Loring Brace, George B. Gamble and James T. Rood. American Anthropological Association: Anthropological Studies No. 8. p. 38.
42Cronbaeh, Lee J. 1969.." Heredity, Environment, and Educational Policy," Harvard Educational Review 39:342.
43John, Vera P. 1971:37.
44Gottessnao, I. I. 1968:25.
45Erlenmeyer-Kimling, L. 1972. "Gene-Environment Interactions and the Variability of Behavior," in Genetics, Environment and Behavior: Implications for Educational Policy, Edited by Lee Ehrman, Gilhert S. Omenn and Ernst Caspari. New York: Academic Press. p. 198.
46Bodmer, IV. F. 1972:107-108.
47Sowell, Thomas. 1973. "The Great IQ Controversy," Change 5:35 (May).
48Jeosen, Arthur R. 1969a:74.
49Jensen, Arthur B. 1969b. "Reducing the Heredity-Environmental Uncertainty: A Reply," Harvard Educational Review 39:473.
50Ward, Martha Coonfield. 1971. Them Children: A Study in Language Learning New York: Holt, Rinehart and Winston.
51Stiischeomhe, Arthur L. 1969. "Environment: The Cumulation of Effects is Yet to he Understood," Harvard Educational Review 39:521.
52Gottesosao, I. I. 1968:27.
53Rosenfeld, Gerry. 1971. "Shut Those Thick Lips!": A Study of Slum School Failure. New York: Halt, Rinehart and Winston.
54Schwebel, Milton. 1968:157-158.
55Lederberg, Joshua. 1969. "Letter to the Editors," Harvard Educational Review 39:613.
56Stinchcombe, Arthur L. 1969:514-515.
57Bodmer,W. F. 1972:111.
58Rex, John. 1972:170-171.
59Washburo, Sherwood. 1963. "The Study of Race," American Anthropologist 65:529.
60Jensen, Arthur R. 1969a: 100.
61Jensen, Arthur R. 1969a:83.
62Jensen, Arthur R. 1969c. "An Embattled Hypothesis: An Interview with Arthur R. Jensen," The Center Magazine
2(5):79.
63Layzer, David. 1974. "Heritability Analyses of IQ Scores: Science or Numerology," Science 183:1259. Genotype refers to the totality of factors that make up the genetic complement of the individual, and phenotype refers to the totality of physically or chemically observable characteristics of an individual that result from the interaction of his genotype with his environment.
64Jensen, Arthur B. l969a:44.
65Layzer, David. 1974:1262.
66For different statistical procedures see: Layzer, David. 1974; Daniels, John and Vincent Houghton. 1972. "Jensen, Eysenck and the Eclipse of the Galton Paradigm," in Race and Intelligence: The Fallacies Behind the Race-IQ Controversy, Edited by Ken Richardson and David Speers. Baltimore: Penguin Books Inc. pp. 68-80; and Light, Richard J. and Paul V. Smith. 1969. "Social Allocation Models of Intelligence: A Methodological Inquiry," Harvard Educational Review 39:484-510.
67Jensen, Arthur R. i969a:S1.
68Anandalakshmy, S. and Janice F. Adams. 1969. "Letter to the Editors," Harvard Educational Review 39:585.
69Layzer, David. 1974:1259.
70Deutsch, Martin 1969:549.
71Kagan, Jerome 5. 1969. "Inadequate Evidence and Illogical Conclusions," Harvard Educational Review 39:275.
72Bodmer, W. F. 1972:96.
73Jensen, Arthur B. 1969a:43.
74Jensen, Arthur B. 1969a:64.
75Jensen, Arthur B. 1969b:460.
76Bodmer, W. F. 1972:98.
77Workman, Peter L. 1972. "Comment," in Genetics, Environment, and Behavior: Implications for Educational Policy, Edited by Lee Ehrman, Gilbert S. Omenn and Ernst Caspan. New York: Academic Press. p. 26.
78Jensen, Arthur B. 1969a:42-43.
79Fehr, F. S. 1969. "Critique of Hereditarian Accounts of 'Intelligence' and Contrary Findings: A Reply to Jensen," Harvard Educational Review 39:575.
80Layzer, David. 1974:1263,
81Hirsch, Jerry. 1968. "Behavior-genetic Analysis and the Study of Man," in Science and the Concept of Race, Edited by Margaret Mend, et. al. New York: Columbia University Press. p. 42.
82Layzer, David. 1974:1265.
83Morton, N. E. 1972. "Human Behavorial Genetics," in Genetics, Environment, and Behavior: Implications for Educational Policy, Edited by Lee Ehrman, Gilbert S. Omenn, and Ernst Gaspari. New York: Academic Press. p. 251.
84Jensen, Arthur R. 1972a. "Discussion," in Ibid. p. 241.
85Grow, James F. 1969. "Genetic Theories and Influences: Comments on the Value of Diversity," Harvard Educational Review 39:308.
86Jensen, Arthur B. 1969b:461.
87Jensen, Arthur B. l969a:78-79.
88Cf.Smith, Paul M. Jr. 1969. "Letter to the Editors," Harvard Educational Review 39:628.
89Sehwebel, Milton 1968:6.
90Rex, John. 1972:176.
91Jensen, Arthur B. 1971. "Can We and Should We Study Race Differences?" in Race and Intelligence, Edited by C. Luring Brace, George R. Gamble, and James T. Bond. American Anthropological Association: Anthropological Studies No. 8. p. 10.
92Jensen, Arthur B. 1969a: 82-83.
93Jensen, Arthur B. 1972b:43-44.
94Jensen, Arthur B. l969a:82.
95Ibid.
96Cronbach, Lee J. 1969:338.
97Hudsun, Liam. 1972. "The Context of the Debate," in Race and Intelligence: The Fallacies Behind the Race-IQ Controversy, Edited by Ken Richardson and David Speers. Baltimore: Penguin Books Inc. p. 14.
98Jensen, Arthur B. 1972b:55.
99Jensen, Arthur B. 1972b: 1.
100Jensen, Arthur B. 1972b: 28.
101Eysenek, H. J. 1972. 'The Dangers of the New Zealots," Encounter 39:88 (December).
102Jensen, Arthur B. 1972b:60.
103Jensen, Arthur B. 1971:10.
104Fried, Morton H. 1968. "The Need to End the Pseudoscientific Investigation of Race," in Science and the Concept of Race, Edited by Margaret Mead, et. al. New York: Columbia University Press. p. 129.