Kings of Men: Introduction to a Special Issue of the Journal of INTELLIGENCE (1998)
by DOUGLAS K. DETTERMAN
Case Western Reserve University
This special issue is dedicated — to Arthur Jensen. It has become apparent that he is unlikely to receive the recognition his work merits. The issue begins with a statement by Jensen which discusses some of his work people are less familiar with. His bibliography is also reprinted. A number of persons selected for their diversity in outlook then comment on the work of Arthur Jensen and the impact it has had.
This special issue is devoted to Arthur Jensen. Several years ago it became apparent to me that Arthur Jensen would probably never receive the kind of recognition others with even lesser accomplishments have been given. He will not receive the honors his work merits from organizations like the American Psychological Association, the National Academy of Science, or the National Association for the Advancement of Science, to name a few. The reasons for this lack of recognition are obvious. He has taken controversial and politically unpopular stands on issues that are important to the study of intelligence.
To attempt to rectify this situation, I contacted him and asked that he allow Intelligence to do a special issue in his honor. My plan was to have him write an introduction that would describe his career. This would be followed by commentary on his work by a wide assortment of researchers in the field. Not surprisingly, the one addition he requested was that a list of his publications be included. He has had a highly productive career and, I must say, I was not ful1y aware of the extent of his interests until I saw the list of publications.
The plan originally formulated has been carried out in this issue of Intelligence. In his personal statement, you will see that Jensen emphasizes work that may be less familiar to many readers but is no less important. The commentaries on Jensen’s work are well worth reading. They tell you as much about the commentators as about Jensen. You will also see that even those who strongly disagree with Jensen have high respect for his intellectual integrity and what he has accomplished. Though this volume stands as a tribute to Arthur Jensen by his contemporaries, like all scientists, his ultimate recognition will be the degree of acceptance of the ideas he has developed.
What has Jensen accomplished in his career? By any measure, Arthur Jensen has made substantial contributions to the study of human intelligence. While opinion may differ about what his most important contributions have been, I have three that I think would rank high on anyone’s list.
Genetics. It is hard to remember how dominant environmental thinking was just 30 years ago. Many thought that even mental retardation could be cured by purely environmental interventions. The genetics of intelligence were seldom discussed. Jensen’s Harvard Educational Review article and subsequent publications on the genetics of intelligence had an enormous effect on the zeitgeist. It was no longer possible to ignore genetic influences when discussing intelligence. He paved the way for the many behavior geneticists who were to follow in his footsteps.
We have still not realized the full ramifications of this change. For example, not everyone yet fully appreciates that genetics will have to be taken into account when considering environmental variables. Much of the research that was done on the effect of so called environmental variables is worthless because those studies failed to partial out genetic influences and, so, are hopelessly confounded. Among the most discerning, there is a new appreciation for the methodologies that will have to be used to truly understand the environment. Year’s from now we may realize that our appreciation of genetic influences was the first step over the threshold to a better understanding of the environment.
Cognition and the Brain. Jensen’s early work on reaction time and intelligence focused interest on cognition and its relationship to brain processes and how cognition and brain processes relate to intelligence. A better appreciation of the cognition-brain-intelligence relationship is critical to ultimately understanding intelligence. Though Jensen was not the only one to appreciate this, his work was certainly pivotal. What seemed to impress people most was that such a simple task as choice reaction time correlated with intelligence.
While we are now in only the earliest phases of understanding the cognition-brain-intelligence relationship, there is no question that is where the field is heading. Techniques for studying this relationship like averaged evoked potentials, PET, and fMRI are increasingly being used and reported upon. The future is promising for these tools. But without an appreciation of the brain-cognition-intelligence relationship it is unlikely they would have
been used at all.
g. Without question, Jensen’s most significant contribution has been to show the importance of general intelligence, or g. The idea introduced by Spearman, while still a graduate student in 1904, reached its highest pinnacle yet in Jensen’s (1998) recent book, The g factor: The science of mental ability. Besides being beautifully written, I predict the book will be the foundation for research for decades to come. It is a summary, that nearly anyone can read, of the research that makes g such an important concept. No one can fully understand individual differences without an understanding of general intelligence and its implications.
Jensen has been the major champion of the concept of g over the last two decades. He has shown that of all individual differences so far demonstrated, g is the most powerful both as a scientific construct and in the prediction of every day performance. He has crystallized methods for studying g and applied those methods. His work on g alone is sufficient for a distinguished scientific career.
Why has Arthur Jensen accomplished what he has accomplished? It seems reasonable to ask why he has accomplished so much if for no other reason than to understand scientific achievement. I think there are a number of personal characteristics that were important in his success. I have had the chance to observe him closely over the last 25 years and these conclusions are based on those observations.
Smart. Jensen comes up with more good ideas than anyone I have known. In the course of reviewing for Intelligence, I have seen countless times when he has suggested ideas for studies to people that have been successfully carried out. He receives no credit for this and, in most cases, is never acknowledged for his contribution. I once asked him why he didn’t carry out some of these good ideas himself. He told me he had more things than he could possibly do and the important thing was to see the work done to advance the field.
His published work is testament to his clear and insightful thinking. The work is always understandable and usually makes a fundamental point. There may be people who publish more, but I am sure there are very few whose publications reflect as many good ideas as Jensen’s do.
Tenacity Developed from a Love of His Work. Jensen sticks to an idea when he knows he is right. There are few people who have a firmer sense of what is right and what is wrong and who are willing to follow their own instincts about what is right. His tenacity is propelled by a real enjoyment of the work he does. If you talk to him about intelligence, it is hard not to become excited yourself because of the enthusiasm he shows. It is clear he really loves the pursuit of answers in this field for their own sake.
Agnosticism and Open Mindedness. In his own statement in this issue, he admits to a healthy agnosticism about everything. This is particularly true about intelligence. He has no investment about how a question comes out, he simply wants it answered correctly. For years, his critics have called him every name in the book and have accused him of all kinds of biases and prejudices. In fact, I have never known anybody with fewer prejudices. The biggest prejudices scientists usually have are those in favor of their own ideas. Such prejudices are very hard to avoid and the notion of the “objective” scientist is, for most of us, a goal we fail to achieve. However, Jensen has no loyalty whatsoever to any theory or hypothesis even if they come from his own ideas. He would gladly know the truth even if it proved him wrong. In fact, he would be excited to know the truth.
His agnosticism is one of the characteristics that it took me longest to identify. I think that is because it is such an unusual one and not typical of most others I have known. When I first met him personally, I wondered what his biases and prejudices really were and tried to identify them for many years. My effort was wasted. I finally came to the conclusion that he just doesn’t have any. I think this may be a point that is impossible for his critics to understand. On the other hand, it is the very reason he has stood up so well against his critics. He has invested himself in pursuit of the truth, not any particular set of ideas.
Thick Skin. I doubt that there have been few people in the history of science who have suffered more criticism than Jensen. There are other examples, of course, including Galileo, Darwin and others. But I doubt if any of them had to have police guards or were regularly threatened with acts of physical violence. I have heard all sorts of rumors about Jensen. One of the most interesting was that he conspired with the Nixon Whitehouse to kill Headstart. I asked him about this and he had a recollection of someone asking him about his research but there was no conspiracy. What is ludicrous about this rumor is that Headstart spending increased dramatically during the Nixon administration (Caruso, Taylor, & Detterman, 1988). There have been many other rumors and gossip, but the ones I have been able to check out have all been false.
Besides vociferous attacks from organized opposition, Jensen has also had to suffer the indignity of seeing his research and writing systematically misrepresented in the popular press. Many of the articles that I have read in the popular press have made me wonder how much of Jensen’s work the author had actually read. I am sure that this misrepresentation would be the most difficult part for me to withstand. However, in the years I have known him, I have never heard him complain about this treatment or express any sentiment of unfairness. I always wondered why. It was not long ago that I figured it out. Because he has no commitment to any particular outcome, Jensen finds it amusing, and perhaps humorous, that people become so exorcised about ideas, ideas that could be right or wrong. Instead of applying their intellect to finding out if these ideas hold water, they express their emotions against the message bearer. The saddest part of the whole thing is that the criticisms that have been directed against Jensen have led to little, if anything, of lasting scientific value. Viewed in this way, the effort expended in futile activity is rather ironically humorous.
One of the incidents that typifies many of Jensen’s personal characteristics occurred when he came to Case Western Reserve University to give a colloquium. The talk was open to the university community and drew a large crowd. Among those in the crowd were several members of the local Communist Party. They had come to hear him talk about race and intelligence but, instead, he had just begun his reaction time research and was talking about that. During his presentation they listened attentively and politely, as is the custom of all Midwesterners, even members of the Communist Party. At the end of the presentation, there was time for questions and they asked a few pointed ones showing that they had studied up for Jensen’s appearance. (In fact, they had probably read more of Jensen than most in the audience.)
After the talk ended, there was a reception in the lobby. As he was drinking his wine and eating his cheese, Jensen slowly made his way around the room working toward the Communist Party members who were bunched in a corner. He was probably drawn to them because it was clear that they were among the best informed about his work even though they had philosophical differences. Maybe it was the philosophical differences’ that attracted Jensen to them. I will never be sure. They began asking him questions about intelligence which he enthusiastically answered. The conversation went on for some time. The rest of the audience drifted away and the caterers began cleaning up. Jensen carried on enthusiastically and, at least in my opinion, his opponents were loosing badly. Looking for a way out, the Communist Party members slowly began backing towards the door. But Jensen was just getting started and for every step backward they took toward the door, he took one forward both figuratively and literally. Feeling a bit sorry for them by this time, I told Dr. Jensen that we had to leave for dinner. Taking the opportunity, the representatives of the Communist Party bolted for the door and began walking east on Euclid Avenue. You could see that Jensen was disappointed to loose his sparring partners. He quickly asked which way we were going to dinner. I said we first had to return to my office which was east on Euclid. Quickly, he proceeded me out the door and caught up to his victims. The discussion proceeded down Euclid Avenue until our ways had to part. I think this encounter was the highlight of Jensen’s visit.
Solitude. Finally, I think one thing that Jensen has enjoyed as a result of his notoriety is a kind of solitude in which to think, work, and write. Even those who become moderately successful in this business are asked to do many things they don’t really want to do and that don’t contribute to their scientific accomplishment. Because many have regarded him as a social outcast, he has been spared many of these nearly meaningless activities that he would have had to carry out if he had been in the good graces of those in power. Add to this a wife who Jensen acknowledges has left him totally free to pursue his research and you have what seems to me a nearly ideal circumstance for a scientific career. My one concern in doing this issue is that we could ruin all of that by giving him the recognition he deserves. Let’s hope that doesn’t happen.
The following quote is one of my favorite from Galton. It describes general intelligence and those who possess it in high quantity. Galton never knew Jensen but I am sure that he had men like him in mind when he wrote this:
“People lay too much stress on apparent specialties, thinking over rashly that because a man is devoted to some pursuit he could not possibly have succeeded in anything else. They might just as well say that because a youth had fallen desperately in love with a brunette, he could not possibly have fallen in love with a blonde. He may or may not have more natural liking for the former type of beauty than the latter, but it is as probable as not that the affair was mainly or wholly due to a general amourousness of disposition. It is just the same with special pursuits. A gifted man is often capricious and fickle before he selects his occupation, but when it has been chosen he devotes himself to it with a truly passionate ardour. After a man of genius has selected his hobby, and so adapted himself to it as to seem unfitted for any other occupation in life and to be possessed of but one special aptitude, I often notice, with admiration, how well he bears himself when circumstances suddenly thrust him into a strange position. He will display an insight into new conditions, and a power of dealing with them, with which even his most intimate friends were unprepared to accredit him. Many a presumptuous fool has mistaken indifference and neglect for incapacity; and in trying to throw a man of genius on ground where he was unprepared for attack, has himself received a most severe and unexpected fall. I am sure that no one who has had the privilege of mixing in the society of the abler men of any great capital, or who is acquainted with the biographies of the heroes of history, can doubt the existence of grand human animals, of natures pre-eminently noble, of individuals born to be kings of men. (Galton, 1869, pp. 24-25)”
I think if you read this issue cover to cover, you will find that no matter what your opinions are on the issues, no matter who is right or wrong, Arthur Jensen is a man to be respected not only for what he has accomplished but for who he is. Thank you, Professor Jensen.
Acknowledgements: Parts of this work were supported by Grants No. HD07176 from the National Institute of Child Health and Human Development, Office of Mental Retardation.
Intensive, Detailed, Exhaustive by THOMAS J. BOUCHARD, JR., University of Minnesota
Arthur Jensen’s bibliography is characterized as breathtaking and his scientific work as intensive, detailed, exhaustive, fair-minded, temperate, and courageous. Specific articles and books are targeted as must reading. I argue that Jensen’s characterization of the influence of the Berkeley psychology department in the 1940′s reflects his own intellectual biases rather than those of the department. Jensen’s work is praised as an extension of the British Biological-Theoretical Tradition which attempts to integrate psychological, biological, social genetic, sociological, and cultural processes in a coherent theoretical framework. A new definition of Jensenism. based on the Jansenist heresy, is provided.
Upon reading both Arthur Jensen’s bibliography and his new book, The g Factor: The Science of Mental Ability, in the same week only one word comes to mind-”breathtaking”. Reading his bibliography is as much a delight as reading his books and papers, truly an intellectual feast. As a fellow Galtonian I will point out a few facts the casual reader might miss if they did not count items and have not read much of Jensen’s work; a) he is the first author on 357 of the 384 items, b) he is the sole author of 319 of the 384 items, c) he has four citation classics, d) he has published nearly 10 items a year (including books) since 1962, e) there is no indication that he is slowing down, and f) the quality is not only superb, it is getting better! One disconcerting feature of the bibliography is the paucity of items that have been reprinted. I was stunned, for example, to see the classic 1977 article, “Cumulative deficit in IQ of blacks in the rural South”, reprinted only once. It is still the definitive paper on the topic. I suspect that the reason so few papers have been reprinted is the same one that has resulted in his not having been given the numerous honors other scientists of his stature have already received. He has dared to study and speak straight forwardly about important issues that most other social scientists only whisper about — race and class differences in IQ, lack of bias in intelligence testing, the biological basis of general intelligence, genetic influences on intelligence, and fallacious research methods in developmental psychology. This point can be nicely illustrated by comparing the way Jensen was treated when he visited the University of Minnesota in 1976 and the way Todd Risely was treated on a recent visit. Jensen had been invited to speak on his new work dealing with test bias by the Institute of Child Development (ICD). I had been asked by Scarr, the invitee, to sit in the front row of the auditorium with her because she had heard that he might be attacked. Attacked he was. He was overwhelmed on stage by some extremely hostile members of the audience. She, I, and the police in attendance had to escort him out to safety. He was able to make a presentation to a small audience at ICD later in the day. What the University community was not allowed to hear was a synopsis of work that has now become the definitive statement on test bias, work which has completely reversed professional opinion on this issue. Almost everything which has followed is derivative. In 1997, Risley was invited to the University of Minnesota by the Institute of Child Development where he expounded on his findings reported in the book, “Meaningful differences in the every day experiences of young American children” (Hart & Risley, 1995). Hart and Risley reported on a long-term within-family correlational study in which they show a high correlation between parental language diversity and children’s IQ. This work was cited by President Clinton during the 1997 White House Conference on Children (UPI, 1997). In his work Jensen has repeatedly emphasized the behavior genetic dictum that correlations between parental behavior and child behavior computed on biological relatives reared together are completely uninterpretable. This fundamental methodological flaw, repeatedly committed by many psychologists, is a simple variation on the argument that “correlation does not mean causation”. For reasons, that I cannot fathom, warnings about this elementary flaw have still not been incorporated into many introductory statistics and methodology textbooks (an exception is Ellis (1994)). One has to ask about the viability of a science that allows the consistent repetition of a serious methodological flaw pointed out and solved by Galton (by the use of the adoption design) over 150 years ago. It is not as if no one noticed Galton’s admonitions. The problem was discussed in great detail by Burks (1928a, 1928b, 1938). In recent years it has been written about in great detail by Meehl (1970, 1971, 1978), Scarr (198 1, 1992, 1997, 1978) and in other guises by Plomin (1994).
WHAT EVERY PSYCHOLOGIST SHOULD READ
Upon examining his bibliography I am embarrassed at the number of Jensen’s publications that I have not read. That will not, however, keep me from making some recommendation to readers who are much less familiar with his work. From the early work read, “The Stroop Color-Word Test: A review” (Jensen & Rohwer, 1966). The 1969 Harvard Educational Review (HER) article, “How much can we boost IQ and scholastic achievement?”, (Jensen, 1969) is still a gem as are the replies to critics. Some critics have argued this article is a citation classic because it is often cited solely for purposed of refutation. I have no doubt that many who cite it for the purpose of refutation have not read it. I recommend it, however, because it is a true classic. Better yet read his book Genetics and Education (Jensen, 1972) in its entirety as it contains the HER article and numerous other superb papers. Jensen, of course, makes a few mistakes now and then as Kamin (1975) points out in his review of this book. The history of one the mistakes is fascinating. Jensen reprinted a graph that included a data point, for dizygotic twins reared apart — a sample of IQ kin data that did not exist at the time. According to Kamin this kind of error reflects the bias of those who take a genetic position. Locurto (1991), however, informs us that the graph came from an article by Heber, Dever, and Conroy (1968). The senior author of that paper was in fact a well known environmentalist (see pages 63-66 in Locorto’s book for a discussion of Heber).
If you are somewhat interested in behavior genetics and don’t know much beyond high school genetics, and would like a primer in quantitative genetics read, “Genetic and behavioral effects of nonrandom mating” (Jensen, 1978). If you want to know something about psychometrics and the issue of bias in mental testing the definitive work is still “Bias in mental testing” (Jensen, 1980a). If you are short on time the Behavior and Brain Science summary of “Bias in mental testing” (Jensen, 1980b) will give you a very good overview of the bias issue. If, like me, you have wondered about Stephen J. Gould’s veracity and competence in the mental ability domain you must read Jensen’s review of “The missmeasure of man”. The title of the review is “The debunking of scientific fossils and straw persons” (Jensen, 1982) and it is among Jensen’s very best book reviews. I would recommend it be followed up with Phil Rushton’s review of the revised edition of the same book (Rushton, 1997). If you still need more criticism of Gould read Dennett’s (1995) assessment of Gould. Alas as I write these words I find that S. J. Gould has been elected president of the American Association for the Advancement of Science (AAAS). The only solace I can garner from this event is that the AAAS once elected Margaret Mead as its president (Freeman, 1983; Freeman, 1991; Freeman, 1992). Mistakes will be made, but some seem more egregious than others.
While doing my simple counts of Jensen’s work it occurred to me that Jensen would have analyzed “the data” differently. He would have argued that it is imperative to remove redundancy and artifacts, he would have grouped the papers by type, by source of publication, by decade. etc., and he would have thrown much more light on the topic. To use the title of one of his book reviews it would have been “Intensive, detailed, exhaustive”. Indeed these three terms capture much of the flavor of Jensen’s writings. I should also add fair-minded, temperate, and courageous. For someone who has been attacked so vituperatively, both in public and in the published literature, I continue to be astounded at the lack of anger and hostility in his replies and the astuteness with which he dissects the arguments of his critics. To use a psychoanalytic metaphor, I am inclined to believe that he sublimates anger and hostility into mental energy — see his astute discussion of the construct of mental energy in Jensen (1997).
I suspect if you asked other Galtonians what they would recommend as “must reading” the list would be somewhat different from mine. There is so much excellent material to chose from that if only a few lists were combined the final list would virtually exhaust his bibliography.
Jensen’s writings are virtual tutorials on how to write science and how to deal with controversy — stick to the available evidence, put all the evidence in it’s full context, carefully explain the methods, their rationale and the assumptions, acknowledge the lack of evidence when it does not exist and avoid ad hominem arguments. In other words stick to the evidence and be intensive, detailed and exhaustive.
A DIGRESSION ON BERKELEY AND WHO INFLUENCES WHOM
I found Jensen’s description of how, in the psychology department at Berkeley in the 1940′s, genetic influences on individual differences were neglected somewhat misleading. His description should have been tempered by the recognition that Tolman in a very early paper titled, “The inheritance of maze-learning ability rats” (1924) took a clear position on the importance of genetic factors as they influence behavior. Gerald McClearn (1962) provides a concise history of this period at Berkeley. Tolman strongly encouraged Tryon to study genetic influence on behavior and they collaborated to develop a self-recording maze to collect data from the selectively bred animals (Tolman, Tryon, & Jeffress, 1929). Tryon published at least 12 papers on individual differences and genetic influences on learning ability in rats between 1929 and 1941. The first, in 1929, was titled “The genetics of learning ability in rats”. This research program resulted in the famous Tryon maze-bright and maze-dull rat strains. Heron (1935) replicated the Tryon work at Minnesota shortly thereafter. Most psychologists are not aware of the fact that Heron published, with Skinner, (Heron & Skinner, 1940), a paper comparing the rate of bar pressing in the maze-bright and maze-dull rats (the “brights” had a higher rate!). My point here is that the idea of genetic influences on behavior was alive and well at Berkeley when Jensen was there. For some reason it did not “infect” him. I am sad to report that much the same thing happened to me. I entered Berkeley as a sophomore in 1963 and also received an education strongly biased in the direction of experimental (environmental) psychology. As a graduate student in the same department, however, I recall Tryon’s spellbinding introductory psychology lectures — I was a teaching assistant in the course — that incorporated behavior genetic findings. Tryon anticipated Jensen’s work and the arguments of “The Bell Curve” (segregation of cognitive classes in American society) by many years. My collagues at Minnesota tell me that Patterson did also, a claim supported by calls from his students in the 40′s asking me what the fuss about “The Bell Curve” was all about, “Wasn’t it old news?”. The importance of the ideas Tryon was talking about simply did not fully register in my mind. I did not relate them to my own interests in personality and social psychology. Gerald McClearn was also on the Berkeley faculty, teaching Behavior Genetics at this time (he went on to Colorado to found the Institute for Behavioral Genetics), but unfortunately we did not have any contact. Fortunately, Harrison Gough — my advisor — required me to read the first textbook in behavior genetics (Fuller & Thompson, 1960) for my special exams and this gave some sense of the field. I also recall Frank Barron presenting, in a very positive manner, the classic meta-analysis of the IQ literature by Erlenmeyer-Kimling and Jarvik (1963) updated in 1981 by Bouchard and McGue (1981) — to an Institute of Personality Assessment and Research seminar. No one seemed to have been aware of the importance of this paper. Nor did they take it seriously, as IQ was out of style in those days. The intrepid Barron, however, had already carried out an early twin study of creativity (Barron & Parisi, 1976). The importance of work in behavior genetics remained only on the periphery of my consciousness until the appearance of Jensen’s 1969 HER paper. I had taught a course on Human Intelligence at the University of California Santa Barbara using the textbook by Hunt (1961). Even though I had only a rudimentary knowledge of behavior genetics I had found the book very unsatisfactory in its treatment of genetic influences. Jensen’s monograph exploded on the scene like a bombshell and I immediately wrote and asked him for a copy. The 1969 monograph and Jensen’s subsequent writing have changed the field of behavior genetics and individual differences in fundamental ways. I report this long anecdote about the Berkeley psychology department because I believe we really do not know why or how people are influenced by the environmental context in which they find themselves. Why, for example, did Jensen become enamored with Hull’s theory instead of Tolman’s which explicitly recognized the role of heredity and individual differences? Jensen, of course agrees with me on this point. As he succinctly puts it, “It always amazed to see psychologists offering glib explanations of some immensely complicated behavioral individual incident when psychological science has not even provided explanations for comparatively simple phenomenon .”
MEMBERSHIP IN THE LONDON SCHOOL
Now that I have castigated others for the sin of assuming they know how we have been influenced by our environment I will proceed to commit the same sin. One consistent feature of Jensen’s research career is his love of theoretical models with elemental parts and clear quantitative implications. These feature characterize Hullian learning theory, the serial position effect, the verbal learning (experimental) tradition he found himself in at the Human Learning Center in Berkeley, the Level I-Level II theory of group differences, and quantitative behavior genetic theory. This pattern of intellectual interests early on led him to become a member of the The London School. I prefer to call the London School the British Biological-Theoretical Tradition because, a) the latter term puts the origins of the group in a large context (Darwin and Galton came well before the University of London which is the London referred to in the term London School), b) it describes the approach of the group and, c) it provides a nice contrast with what I call the French Clinical-Therapeutic Tradition. The British Biological-Theoretical Tradition has been attacked on a variety of grounds (reductionistic, anti-egalitarian, racist, cold and heartless, etc.) but the most vehement arguments have been against its biological orientation. Consider the following quote, “The interpretation of IQ data has always taken place, as it must, in a social and political context, and the validity of the data cannot be fully assessed without reference to that context. That is in general true of social science, and no amount of biology-worship by behavior geneticists can transfer IQ testing from the social to the biological sciences (Kamin, 1974, p. 2).” Lewontin, Rose, and Kamin (1984) have extended this argument to all behavioral traits including psychopathology. These critics have cut to the heart of the matter. The goal of British Biological-Theoretical Tradition has indeed been, since the time of Galton, to integrate psychology, biology, and genetics (Bouchard, 1996). “The g Factor: The science of mental ability” is a direct descendant of Galton ‘s book “Hereditary Genius” and Spearman’s book “The abilities of man” (1927). It is a brilliant work. It pushes the goal of British Biological-Theoretical Tradition a giant step forward. I challenge the reader to examine Jensen’s magnum opus and decide for him or herself if it has crossed the threshold from the social to the biological sciences. It is worth noting that E.O. Wilson’s recent book, “Consilience: The Unity of Knowledge” (Wilson, 1998), defends a very similar but even broader research program.
While reading “The g Factor” I was struck by Jensen’s detailed knowledge of the lives of many of the important historical figures in the IQ story. His bibliography explains why he is so knowledgeable. In 1984 he wrote bibliographic entries for the Encyclopedia of Psychology on Galton, Pearson, Spearman, and Thurstone and in 1994 he wrote bibliographic entries for Galton, Spearman, and Eysenck for the Encyclopedia of Intelligence. More recommended reading.
Jensen reported in his commentary in this issue the definition of “Jensenism”, taken from current dictionaries, in order to try to free himself from its grip. I don’t think this is possible, consequently, I thought it might be worth preparing a proper and more comprehensive definition that praises Jensen. My definition is anticipatory and includes the effect of his magnum opus, “The g Factor”. My definition mimics the definition of an older heresy — Jansensim (See encyclopedia Britanica 15th Ed., Micropaedia, p. 515)
Jensenism: A scientific movement of unorthodox tendencies (heresy) that appeared chiefly in the United States in the late 1960′s. The movement was scientific (religious) in origin, arising out of the theoretical (theological) problem of reconciling the empirical observation of massive and important individual differences if intelligence as well as a large and persistent black-white difference in intelligence (lack of divine grace) with the belief that all men are created equal (human freedom). Jensenism exalts the influence of the genes (grace) made available by mother nature (Christ the Redeemer). According to the doctrine, genes are capable of explaining most of the differences; and it puts forth the scientific (Augustinian) arguments regarding the necessity of genes for any explanation of the differences, the infallible efficacy of genes, and demonstrates the absolutely arbitrary character of environmental explanations. Consistent with this pessimistic view of man’s nature and freedom are its rigoristic views on scientific method and quantification.
The publication of the manifesto of Jensenism, “The g Factor”, after attempts to censor it, aroused violent controversy. The work was accused, chiefly by Psychologists (Jesuits), of divesting freewill of all reality and of rejecting the universality of the redemption by environmental means. Nevertheless, the Jensenist interpretation of the empirical evidence spread. It was defended by many disciples and it attracted many influential converts.
The establishment, in the pages of the New York Times (Papacy), struck out against Jensenism with the publication of a devastating review of “The g Factor” (the Bull of Cum Occasione) which among other things condemned the five propositions of Jensenism on the relationship between black-white differences in IQ and genes.
Jensensim is a complex movement, based more on a commitment to scientific method (a certain mentality and spirituality) than on specific doctrines. It is an attempt, in line with that of the Reformers, to reform psychology (the Church) in the spirit of early science (Christianity). It opposed what, in its view, was a compromising approach to true scientific method (Christian theology) and practice but was rejected by psychology (the Church) as an exaggerated and unorthodox position.
Costs and Benefits of Defying the Crowd in Science by ROBERT J. STERNBERG, Yale University
Scientists, mirroring the societies in which they live, have devised numerous ways of rewarding conformity and punishing defiance. Some of the mechanisms are reviewed. Scientists who defy the crowd can gain extrinsic reinforcement, but often from sources that promote irresponsibility on the part of these scientists. For the most part, Arthur Jensen has spent his career in defiance of the scientific crowd. Some of this work has made an outstanding contribution to the science of intelligence: other work, I believe, has been regressive. What kind of system might appropriately reward that work which has made a contribution?
On November 27, 1997, the day of the Macy’s Thanksgiving Day parade, my wife and I were in New York City. We had no interest in the Thanksgiving Day parade and were walking down Seventh Avenue when we started encountering noticeable pedestrian traffic walking in the opposed (northward) direction. The farther we walked, the heavier the opposing pedestrian traffic became, and the more visibly annoyed people became that we were walking in the opposite direction. Eventually, we reached a density of human traffic such as we had never seen before. The choice of whether to walk in the direction opposed to everyone else was taken away from us: Police waved us over onto a side street to head toward the east. We could either join the crowd or leave it altogether. We were forbidden to oppose it.
Throughout much of history and much of the world even today, people have had the same choice with respect to their ideas. They have had the option either to join the crowd or, if they are lucky, to leave it, but not to oppose it. Through secret police, inquisitions, kangaroo courts, and even summary execution, people who have chosen either not to be part of the parade or at least to be fellow-travelers with it have been subjected to punishment. Of course, people in the United States like to believe that it is different there. After all, many of those living in the U.S. believe it to be a “free country.”
On the one hand, the freedoms enjoyed by residents of the U.S. are substantial. For example, people can criticize the government or even sue its chief executive, both without being imprisoned or otherwise totally silenced. At the same time, the society has been able to function without a formal government-sponsored “thought police” in part because the members of the society have themselves taken on aspects of the role of a thought police, obviating the need for a formal squadron. The society has devised many ways to punish nonconformers, as any child in a schoolyard has observed. Of course, at other times in the country’s history — most notoriously but not only during the McCarthy era — nonconformers in the U.S. have not been so lucky. Those who have defied the crowd have been vilified or even perished. Scientists are not much different from other people. Scientists, too, have developed a number of ways to ensure that their numbers follow the crowd.
ENFORCING CONFORMITY TO THE CROWD AMONG SCIENTISTS
Thought policing is not limited to politics. It occurs in science as well. Many individuals enter science because they believe it is a calling that encourages free thinking and independent thought. Many of these same individuals soon discover that their idealism bears little contact with reality. As Kuhn (1970) and others have observed, scientists are no more independent-minded or free-thinking than anyone else. If anything, they cherish conformity more than the rest.
Scientists enforce conformity in a number of ways, both formal and informal.
1.Training. For the most part, students learn through their preprofessional training what the current paradigms are and what kinds of work are rewarded and what kinds are not. They are encouraged to do kinds of work that will be rewarded. To a large extent, training is considered “good” to the extent that it teaches students where the rewards are.
2.Publications. Many people who have submitted articles to journals have discovered that the refereeing process is an excellent way to ensure conformity under the banner of quality control. Of course, it is difficult to get articles accepted if they are totally pedestrian; but there is almost always some journal that will take an article, no matter how pedestrian it may be. More difficult is to get articles accepted if they go against the accepted wisdom, as John Garcia discovered in his studies of conditioning and as many others have discovered in their own work. Thus, many people find that the work that is hardest to get accepted is not only their worst work, but also their best.
3.Grants. Grants provide an excellent way to reward conformity. People who work outside established paradigms often find it very difficult or impossible to get funded, so that they are effectively prevented from doing much of what they might have intended to get done. There are many forces that contribute to making granting agencies a conservative force (see Sternberg, 1996a, 1997). First, low selection ratios allow even one negative reviewer essentially to blackball a proposal. Second, programmatic agencies fund work within their established program of research but not outside it. Third, people who are asked to serve on review panels will, for the most part, be those working within established and accepted paradigms. Fourth, those who agree to spend the vast amounts of time it requires to be on such panels may tend even more toward conformity than those who would rather devote the time to their own research. Finally, proposals are expected to make contact with existing paradigms, and if they do not, they can be rejected for this reason alone.
4.Recognitions. Through prizes, awards, organizational offices, and the like, scientists can enforce their set of values, recognizing those who play the accepted game well and failing to recognize those who go outside the accepted limits. In some cases, these views may even have nothing to do with the work for which recognition is being given. For example, a lifetime achievement award to be presented to Raymond Cattell at the annual meeting of the American Psychological Association in 1997 was suspended pending investigation of his religious beliefs!
5.Book Reviews. Books of scholars who go beyond the limits are typically subject to negative reviews, sometimes by people who seem not to have read the books.
6.Graduate Students. “Respectable” researchers do not send their undergraduate students to work with “disreputable” researchers.
7.Informal Networks. Perhaps most importantly, those who work outside accepted networks never make it into the professional in-groups. They are less likely to be asked to serve on committees, write promotion letters, give invited talks at major scientific conferences, give departmental colloquia, and the like.
In sum, the various fields of science construct a system for enforcing conformity. Scientists who do not conform are “out.” But other mechanisms come into play that not only reward scientists for divergent views, but actually encourage the scientists to diverge even more, even to the point of irresponsibility.
ENCOURAGING NONCONFORMITY TO THE CROWD AMONG SCIENTISTS
If scientists received no rewards at all for nonconformity, they might cease to be nonconformists. But there are at least three major sources of rewards for nonconfonnists.
1.Internal Rewards. Scientists who state what they believe and then fight for their beliefs have the satisfaction of knowing that they are saying and doing what they believe in. They can also hope that, in the long term, the scientific establishment will come around to their way of thinking and reward what they are doing. In fact, such changes are not unusual. In their writings, both Sandra Scarr and Robert Plomin have commented on how behavior-genetic work that was devalued in the 1970s came to be valued by the latter half of the 1990s.
2.Fringe Groups. Fringe groups of scientists may set up their own organizations to reward what they are doing, or may find that their work is ideologically consistent with the priorities of political or social fringe groups and thus accepted and even welcomed by such groups. Such scientists may therefore find themselves having to decide whether to associate with these groups in order to feel extrinsically rewarded. But these groups may in turn encourage the scientists to take positions even more extreme than those they believe in, and perhaps to take positions that are irresponsible.
3.The Media. By far the most powerful ally of the nonconforming scientist in this country can end up being the media. The media thrive on controversy and on the offbeat. Thus if virtually all scientists believe that AIDS is caused by the human immunodeficiency virus (HIV) and a few scientists do not believe this to be the case, the disbelievers may find themselves actually getting more media attention because of the divergence of their views. If most people believe that racial differences in psychometrically measured intelligence are largely environmental and a few scientists believe (or are willing publicly to state) that such differences are probably largely genetic, the views of the minority are likely to attract attention. Moreover, even if most of these scientists’ views are conventional, it is the unconventional part of their views that is likely to attract the media attention.
The situation can become pernicious because media attention tends to be short-lived. Unfortunately, almost the only way for nonmainstream scientists to maintain media attention and the extrinsic reinforcement it brings is either to take new unconventional positions, or to become more extreme in the positions they already have taken. Many of us scientists who have worked with the media have found reporters trying to get us to make statements more extreme than we really believe, simply because such statements make for more interesting press coverage. The reinforcement system thus can turn a nonconforming but responsible scientist into a less responsible or even an irresponsible one. Worse, it may be only through the media that one can gain any coverage of one’s divergent views. Iced out of mainstream science, scientists with nonconforming views may thus turn to the media to get press coverage of their views, not fully realizing the dangerous game into which they are entering. Of course, the press coverage further “turns off’” the so-called respectable scientists, so that what formerly might have been a bad situation with regard to the scientist’s participation in mainstream science becomes an even worse one.
THE ROLE OF DEFIANCE OF THE CROWD IN SCIENCE
Defiance of the crowd in a Thanksgiving parade is rather innocuous. In science, defiance of the crowd has higher stakes. Elsewhere, Todd Lubart and I have proposed that defiance of the crowd is the hallmark of creativity (Sternberg & Lubart, 1991, 1995). Individuals in science or any other field who make the most difference are those who defy the crowd. These individuals generate ideas that, like stocks with low price-earnings (PE) ratios, seem unattractive and even repugnant to others. The individuals work to raise the value of their metaphorical stocks, attempting to convince other people of the value of their ideas. Ultimately, they metaphorically “sell high,” moving on to their next unpopular idea.
In our work, we give numerous examples of how initial receptions to creative ideas are often unfavorable and even patently hostile (Sternberg & Lubart, 1995). Scientists have developed a number of ways to ensure that scientists follow the crowd.
What’s the problem, then? Why not just institute some kind of guarantee that scientists who defy the crowd will be rewarded rather than punished? The problem stems from the fact that creativity is typically defined not only in terms of novelty, but also in terms of quality and appropriateness. In terms of the stock-market analogy, one needs to remember that many and probably most low P-E stocks never do rise much in value. Consider the example of HIV and AIDS.
The scientist who denies that the human immunodeficiency virus causes AIDS takes a large risk. He will be disparaged by other scientists for defying the crowd. But if he can show to their satisfaction or that of others who hold power in the society that he is correct, then he may actually end up being a hero. In the case of the HIV opposition, no such demonstration has emerged. Nor has any credible science emerged from the efforts of proponents of cold fusion. In both cases, novelty without perceived quality has led proponents of offbeat views to be labeled not as creative, but rather, as crackpots. Creative people, of course, are risk-takers, but they tend to be sensible risk-takers (Sternberg & Lubart, 1995). They are willing to take risks that, in the long run, are more likely to pay off. The risks taken by proponents of the theory that HIV does not cause AIDS and of the theory of cold fusion have, to date at least, failed.
THE CASE OF ARTHUR JENSEN
Where does Arthur Jensen fit into the schema that has been set up in this article? First, most of Jensen’s career has been outside mainstream science. Since his article almost 30 years ago in the Harvard Educational Review (Jensen, 1969), Jensen has been viewed by many as outside mainstream science. This fact is ironic, because the overwhelming majority of his articles and books have been within mainstream science. Jensen’s work on reaction time and intelligence (e.g., Jensen, 1982) is solidly within the information-processing tradition. Jensen’s work on test bias (Jensen, 1980) is solidly within the psychometric tradition, as is his work on the g factor (Jensen, 1998). Thus, what constitutes a relatively small proportion of his work has, for the majority of the scientific community, defined him. As noted in an invitation letter to this symposium that “he has received very little official recognition for his work and probably will not in the future” (Detterman, 1997), Jensen’s defiance of the scientific crowd has cost him. Awards and recognitions that he otherwise might have received for influential, highly cited work may never come.
Second, Jensen has been courted both by political fringe groups and by the press. Not despite but rather because of the unpopularity of his views, Jensen has been a media figure to an extent that is rare in mainstream psychology. Few psychologists are as well known, and some who perhaps are, such as the late Richard Herrnstein, are known for much the same reason — not for their mainstream work (in Herrnstein’s case, on animal learning), but for their work on race, heritability, and intelligence.
Scientifically, I disagree with most of the corpus of Jensen’s work for reasons that are not relevant to this article but are discussed elsewhere (e.g., Sternberg, 1985, l996b). But in terms of the criteria by which I believe scientific work should be judged — such as creativity, basis in theory, empirical rigor, and impact — I believe that most of Jensen’s work fares well. The corpus of Jensen’s psychometric work on the general factor and of his information-processing work on reaction time — but not of his behavior genetic work, including work on racial differences.– place him as one of the outstanding leaders in the field of human intelligence. Indeed, few people now alive have had more impact on the field, for better or worse. And few people studying human intelligence have more scientific investigations to their credit. Indeed, much of the highly cited work in the field of intelligence has little or, arguably, no scientific basis at all.
I exclude from this accolade Jensen’s work on behavior genetics and racial differences in intelligence because, for a number of reasons discussed elsewhere (e.g., Sternberg, I 996b), I believe this work to be not only wrong, but wrong-headed. My goal here, though, is not to discuss substantive differences, but rather, how a field should evaluate scientific work that defies the crowd.
Arthur Jensen is, in my opinion, an epitome of the need to change the reward system in science. Suppose, for the sake of argument, that many other scientists believed as I do that Jensen’s work on information processing and on psychometrics has been ground-breaking but his work on behavior genetics has not been, or even has been regressive. How does the reward system function?
Much of the way the academic reward system functions — not just in science — is by the reputation of the academic (Caplow & McGee, 1958). When reputation is viewed unidimensionally or almost unidimensionally, the field may find itself forced into judgments it should not make. If one body of work within a corpus is disfavored, scientists may end up generalizing this disfavor unfairly to other work by the same investigator.
At times, some kind of combination formula with regard to the bases of evaluation is inevitable. For example, when a department has just one available slot for a job and someone must be hired, a hard choice must be made despite the fact that an idiographic model of evaluating candidates might seem much more appropriate than a nomothetic model. But many decisions need not be unidimensional.
Many and probably most major scientific awards are to individuals for the cumulative corpus of their work. As a result, a scientist who has done even one stream of unpopular work may find him or herself iced out of the awards system because this work damages — rightfully or wrongfully — the valuation given to the overall corpus or work. Perhaps a better way to grant recognition would be to a program of work, with the individual rather than the work being seen as incidental. Thus, instead of giving an award to Scientist X for Research Program A, the award would be given to Research Program A — not necessarily the whole corpus of a scientist’s work — with the scientist receiving the award incidentally. The focus would be on the work, not on the scientist. In the case of Jensen, one could recognize the value of his work on reaction time or the general factor without recognizing the value of other work. In the case of Cattell, one would reward the work, say, on the theory of fluid and crystallized intelligence or on the 16-personality factor theory irrespective of what Cattell’s personal beliefs might be.
In some cases, judgments of work are being influenced not by portions of the person’s work, but by judgments of the person’s character with respect to things that arguably have nothing to do with the work. Cattell’s religion is a case in point. A more extreme example is Paul DeMan. The work of Paul DeMan, in particular, and deconstructionism in general, are undergoing a thorough reexamination in light of fairly recent discoveries that DeMan wrote virulently anti-Semitic tracts in his youth. Such tracts certainly may and probably must greatly damage our evaluation of DeMan as a person. But Richard Wagner the composer and Ezra Pound the poet were also virulent anti-Semites. Their work stands as it was, regardless of how personally despicable either or both of them likely may have been. It would probably be a loss to the world if Wagner’s and Pound’s works were ignored because of their despicable personal views or because of their deeply flawed personal characteristics.
If we are to believe Gardner (1993), many creative individuals have had much less than savory personal characteristics. There is good reason to judge people and their work separately, and then to judge people’s distinct programs of work separately. Indeed, almost every creative individual has produced work of which he or she is, at best, not proud, and at worst, ashamed (or should be).
Science has a number of ways of enforcing adherence to the dictates of the crowd. Scientists who choose to defy the crowd can still gain reinforcement, but when it comes from others, it is often in the form of temptations that can lead the scientist down a path to irresponsibility. Scientists would do better if they focused their evaluations not on individuals, but on programs of work within the total corpus of the scientists’ work. In this way, people whose work is viewed as undesirable in some ways are not punished so that neither they nor other work they may do is taken seriously.
In the case of Arthur Jensen, I believe that a large body of his work is deserving of great commendation (although I disagree with most of it). I hope it is for his work on information processing and the general factor that he is remembered, not for his work on behavior genetics, test bias, or racial differences in intelligence and related traits.
If there is anything for which citizens of a country should give thanks on Thanksgiving Day, it is not that they can join a parade, but that they can choose to walk, at the very least, away from it, and at best, in opposition to it.
Acknowledgements: Preparation of this article was supported under the Javits act program (Grant R206R5000 I) as administered by the Office of Educational Research and Improvement, U.S. Department of Education. The findings and opinions expressed in this article do not necessarily reflect the positions or policies of the Office of Educational Research and Improvement or the U.S. Department of Education.
The “Jensen Effect” and the “Spearman-Jensen Hypothesis” of Black-White IQ Differences by J. PHILIPPE RUSHTON, University of Western Ontario
Arthur Jensen’s research on the biological basis of mental ability has culminated in his encyclopedic new work The g Factor (1998) which massively confirms “Spearman’s (1927) hypothesis” that Black-White IQ differences vary systematically as a function of each test’s g loading. More generally, The g Factor consolidates the psychometric, neurophysiological, behavior genetic, and comparative evidence for the existence and importance of g and links it to evolutionary processes. But perhaps Jensen’s greatest legacy to science will be his pioneering method of correlated vectors which subsumes, under a much broader principle, his famous (1969a) hypothesis about the heritability of the Black-White IQ gap and, as Osborne (1980) dubbed it. the “Spearman-Jensen hypothesis” that Black-White IQ differences are greatest on the g-factor. Jensen’s method of correlated vectors demonstrates that g (specifically a test’s g loading) is the best predictor of that test’s correlation with a given variable, in future, when a significant correlation occurs between g-factor loadings and variable X, the result might usefully be called a “Jensen Effect” (for that X variable). because otherwise there is no name for it, only a long explanation of how the effect was achieved. Naming it the “Jensen Effect” would honor one of the greatest psychologists of our time.
A Personal Note
Perhaps I am the only psychologist of my generation who missed the tumultuous appearance of Arthur Jensen’s (1969a) famous Harvard Educational Review article arguing that IQ is heritable and that genetic factors are involved in the Black-White IQ gap. The attendant brouhaha failed to reach my attention in England where I was an undergraduate student at the University of London. Two years later, however, when Hans Eysenck popularized Jensen’s argument in his 1971 book Race, Intelligence, and Education, I was a graduate student at the London School of Economics and Political Science, and Eysenck’s book created such a furor that a small group of us social psychologists decided to study the issue. Jensen’s clearly argued response to seven “replies,” as well as his original exposition (all usefully compiled in an offprint series by the Harvard University Press) led some of us to believe that he might well be right.
Jensenism, described as one of the great heresies of 20th century science, continued to inspire heated debate at the London School of Economics for the next two years, culminating in a physical assault on Professor Eysenck when he came to give us a lecture in 1973 on “The Biological Basis of Intelligence.” I was more than just a horrified witness to this ‘political action’ by a dozen Maoists (proudly sporting red Mao-Tse Tung badges in their lapels). I was even featured in a newspaper photograph in a scrum around Eysenck, energetically pulling at rampaging ‘demonstrators,’ but wearing the fashionably long hair of the time, it might not be obvious from the photograph whose side I was on! The Maoists made no attempt to hide after Eysenck was hustled away, for the police were not to be called and there was an unfortunate sentiment that Eysenck only got what he deserved. “No Enemies on the Left” was a mantra at the L.S.E. in the early 1970s.
The first time I heard Jensen speak in person was at the 1978 annual meeting of the American Psychological Association in Toronto where he (1979) presented “g: Outmoded Theory or Unconquered Frontier?” The science was inspirational, all about reaction-time and speed-of-processing correlates of IQ. The large ballroom was filled to overflow and the audience, rapt with attention, burst into enthusiastic applause when he had finished. If only in contrast to anxious expectations, the ‘infamous Dr. Jensen’ struck me as warm, humane, and giving of one of the most exciting talks I had ever heard.
I eventually met Jensen in early 1981 while spending a term as a Visiting Scholar at Berkeley’s Institute of Human Development. Having just written a book explaining altruism from a social learning perspective (Rushton, 1980),. I was broadening my focus to encompass behavioral genetic and sociobiological viewpoints. Although many of those at the Institute of Human Development had earned international reputations for documenting the early emergence of personality traits and their power to predict social adjustment, few were interested in searching for behavior genetic causes. The reason was not hard to find. At Berkeley, any discussi> ————————————————————————
ervous hop, skip, and a jump away from Jensen’s controversial racial hypothesis.
Jensen occupied an office in the School of Education, one floor up from my office in the psychology department. We easily established rapport. The question of race differences was beginning to fascinate me and on this topic, of course, Jensen was most informative. Over several lunches at Pasand, one of his favorite local Indian restaurants, he sketched out his views and helpfully answered queries. Back at his office he provided reprints. It was clear that Jensen’s defining trait was intellectual curiosity and for him the study of race differences presented an acid test. How could the topic, which loomed so large in education and society, be avoided for ideological reasons if psychology was to be scientific and if the individual scientist was to maintain personal integrity? I came away profoundly influenced and determined to read the relevant literature.
International Distribution of IQ, Brain Size, and Related, Traits
Many researchers were inspired by “Jensenism.” Lynn (1978, 1982) and Vernon (1982) not only pushed the envelope, but extended the ‘outside of the envelope’ and made the race-IQ debate international in scope with their findings that East Asians average higher on tests of mental ability than do Whites, whereas Caribbeans (and especially Africans) average lower. As Lynn’s (1997) and Jensen’s (1998) most recent reviews show, East Asians, measured in North America and in Pacific Rim countries, typically average IQs in the range of 101 to 111. Caucasoid populations in North America, Europe, and Australasia typically average lQs from 85 to 115 with an overall mean of 100. African populations living south of the Sahara, in North America, in the Caribbean, and in Britain typically have mean IQs from 70 to 90. (Blacks in sub-Saharan Africa score about 2 standard deviations [approximately 30 IQ points] below the mean of Whites on nonverbal tests.)
As a budding sociobiologist, I too was inspired by Jensenism. It seemed to me that by its impact on diverse areas of behavioral science, Jensenism might help complete the Darwinian revolution. I began to review the international literature, studying not only IQ, but other behavioral traits like speed of physical maturation and longevity, personality and temperament, family structure and crime, and sexual behavior and fertility, and later brain size (Rushton, 1984a, 1984b, 1988). I have found that on these traits East Asians are slower maturing, less fertile, less sexually active, with larger brains and higher IQ scores than Africans, who tend towards the opposite in each of these areas. Europeans, I found, fell between the other two groups. As Jensen (1984) elaborated (in a commentary on my first review), a network of such related evidence provides more opportunity for finding and testing alternative theories than does any single dimension drawn from the set.
As a now avowed Jensenist, I carried out experiments finding, for example, that the amount of inbreeding depression on 11 sub-tests of the Wechsler Intelligence Scale for Children in Japan predicted the magnitude of the Black-White differences on the same sub-tests in the U.S. (Rushton, 1989). Inbreeding depression, a purely genetic effect, was a sufficiently robust predictor to overcome generalization from the Japanese in Japan to Blacks and Whites in the U.S. There really is no other explanation, other than a genetic one, for the correlation between inbreeding depression and Black-White differences.
I also calculated cranial capacities from external measurements of the head using large archival data sets including a stratified random sample of 6,325 U.S. Army personnel (Rushton, 1992), a sample of tens of thousands of men and women collected by the International Labour Office in Geneva (Rushton, 1994), and a sample of thousands of American children from birth to age seven (Rushton, 1997). After adjusting for the effects of stature, weight, and sex, the cranial capacities consistently averaged higher for East Asians than for Europeans, who averaged higher than Africans, as reviewed by Rushton and Ankney (1996) and Jensen (1998).
Jensen’s The g Factor
All the issues Jensen raised in 1969 are still with us today. Indeed, much of the opposition to IQ testing and heritability would probably disappear if it were not for the stubborn and unwelcome fact that, despite extensive well-funded programs of intervention, the Black-White difference refuses to go quietly into the night.
Jensen’s long intellectual march has culminated triumphantly in his latest book, The g Factor (1998), an exposition of the reality of Spearman’s (1927) seminal concept of g, the general factor of intelligence. Jensen’s tome does not draw back from Jensenist conclusions–that the average difference in IQ found between Blacks and Whites has a substantial hereditary component, that this difference is related mainly to the g-factor, and that it has important societal consequences.
Chapter 11 of The g Factor fully documents how, on average, the American Black population scores below the White population by about 1.2 standard deviations, equivalent to 18 IQ points. This mean difference between Blacks and Whites in IQ scores has scarcely changed over the past 80 years (despite some claims that the gap is narrowing) and can be observed as early as three years of age. Controlling for overall socioeconomic level only reduces the mean difference by 4 IQ points. Contrary to purely cultural explanations, culture-fair tests tend to give Blacks slightly lower scores, on the average, than more conventional tests, as do non-verbal tests compared with verbal tests, and abstract reasoning tests compared with tests of acquired knowledge.
The reason, in fact, that Jensen pursued Spearman’s hypothesis is that it so exquisitely solved a problem that had long perplexed him about test bias with respect to Black-White differences. He had noted that the Black-White differences are markedly smaller on tests of rote learning and short term memory than on tests of reasoning and those requiring any transformation of the information. He initially formalized these observations in his so-called Level I-Level II theory (Jensen, 1968). Level I tasks were those that required little or no mental manipulation of the input to arrive at the correct output. A clear example of Level I ability is Forward Digit Span in which people recall a series of digits in the same order as that in which they are presented. Level II tasks, however, require some mental manipulation of the input in order to arrive at the appropriate response. A clear example of Level II ability is Backward Digit Span in which people recall a series of digits in the reverse order to that in which they are presented. Jensen found that Black-White differences are twice as large for Backward as for Forward Digit Span. As this finding did not readily lend itself to an explanation in terms of cultural bias or in terms of any other theory Jensen knew of except his Level I-Level II notion, he kept thinking about it.
After Jensen re-read Spearman, he realized that his Level I-Level II formulation was only a special case of the more general hypothesis proposed by Spearman. Jensen began testing Spearman’s hypothesis on a wide variety of psychometric tests administered to large representative samples of the American White and Black populations (Jensen, 1985, 1987). The g Factor summarizes the results from 17 independent data sets on a total of nearly 45,000 Blacks and 245,000 Whites derived from 171 psychometric tests. g loadings consistently predict the magnitude of the Black-White difference (r = +.63). Spearman’s hypothesis is borne out even among three-year-olds administered eight sub-tests of the Stanford-Binet. The rank correlation between g loadings and the Black-White differences is +.71 (p <.05>
Spearman’s hypothesis applies even to the g factor extracted from performance on elementary cognitive tasks. In some of these studies, 9-to-12-year-olds are asked to decide which of several lights is illuminated and move their hand to press a button that turns that light off. All children can perform such tasks in less than one second, but children with higher IQ scores perform faster than do those with lower scores, and White children, on average, perform faster than Black children (Vernon & Jensen, 1984). The correlations between the g loadings of these types of reaction time tasks and the Black-White differences range from +.70 to +.8l.
Jensen also applied Spearman’s hypothesis to East Asian-White comparisons, using the same reaction time measures. The direction of the correlation is opposite to that in the Black-White studies, indicating that, on average, East Asians score higher in g than do whites. No one so far seems to have looked at East Asian-White differences on conventional psychometric tests as a function of their g loadings. From the study just mentioned, however, Jensen’s prediction should be clear: One should find the mirror image of Spearman’s hypothesis for Black-White differences. It might be interesting to note, in light of the above, that in an early reply to a charge of “white supremacy,” Jensen (l969b, p. 240) made a remarkably presaging conjecture, He wrote: “. . .if I were asked to hypothesize about race differences in what we call g or abstract reasoning ability, I should be inclined to rate Caucasians on the whole somewhat below Orientals, at least in the United States.”
The Spearman-Jensen Hypothesis
Osborne (1980) suggested that if scientific credit was to be assigned appropriately, the “Spearman hypothesis” that Black-White differences are greater on more g-loaded sub-tests should become the “Spearman-Jensen hypothesis” because it was Jensen who brought Spearman’s hypothesis to widespread attention, and it was Jensen who did all the empirical work confirming it. Jensen (1997) himself has noted that, “Because Spearman himself never presented it as a formal hypothesis, a few people have objected to my crediting it to Spearman. So whenever I say ‘Spearman’s hypothesis,’ I hope you will visualize these words in quotation marks.”
The Jensen Effect
The Spearman-Jensen hypothesis turns out to be readily subsumable under a more general principle that, when resulting in a positive finding, we might call a “Jensen Effect.” Recall that the Spearman-Jensen hypothesis was tested by first extracting the g factor from a variety of cognitive tests, and then relating these scores (a ‘vector’ of scores, i.e., with direction as well as quantity), to the mean Black-White differences on those same tests (a second ‘vector’ of scores). Jensen extended this method of correlated vectors to a variety of variables. Using this procedure, Jensen (1998) showed that the vector of a test’s g loadings is the best predictor of that test’s correlation with a variety of variables, including not only scholastic and work-place performance, but also brain size, brain pH, brain glucose metabolic rate, average evoked potential, reaction time, and other physiological factors. The Jensen Effect can be seen whenever there is a significant correlation between the vector of the sub-tests’ g loadings and the vector of the same sub-tests’ loadings on variable X (where X is some other, usually non-psychometric variable).
This methodological innovation of Jensen’s may be an even greater discovery than the totality of empirical results generated by it, important though these undoubtedly are. His method of correlated vectors is fully explicated in The g Factor (Appendix B) and is also discussed in the opening remarks of this symposium. To honor Jensen’s accomplishments into the future, I propose that when a significant correlation occurs between the two vectors the result be called a Jensen Effect (for that X variable), because otherwise there is no name for it, only a long explanation of how the effect was achieved.
Chapter 12 of The g Factor presents Jensen’s technical arguments for why he believes that race differences are about 50% genetic in origin. He emphasizes the fact that it is precisely those components of intelligence tests that are most heritable and that most relate to brain size which most profoundly differentiate Black from White groups. The heritability data are especially interesting because genetic theory and culture theories of race differences make predictions opposite to each other. Culture theory predicts that differences between races will be greater on those culturally malleable items on which races can grow apart as a result of dissimilar experiences.
The g Factor also cites the evidence of transracial adoption studies. Three studies have been carried out on Korean and Vietnamese children adopted into White American and White Belgian homes. Though many had been hospitalized for malnutrition, prior to adoption, they went on to develop IQs ten or more points higher than their adoptive national norms. By contrast, Black and mixed-race (Black-White) children adopted into White middle-class families typically perform at a lower level than similarly adopted white children. In the well known Minnesota Transracial Adoption Study, by age 17, adopted children with two White biological parents had an average IQ of 106, adopted children with one Black and one white biological parent averaged an IQ of 99, and adopted children with two Black biological parents had an average IQ of 89 (which is not different from that of Black children raised by Black parents in these northwestern states).
The g Factor also devotes a fair amount of space to racial differences in brain size. Chapter 6 reviews the literature which shows that the brain-size IQ relation emerges most clearly using Magnetic Resonance Imaging (r = .44 across eight separate studies). Chapter 12 documents the three-way racial gradient in brain size established by aggregating data from studies using four kinds of measurements: (a) wet brain weight at autopsy, (b) volume of empty skulls using filler, (c) volume estimated from external head sizes, and (d) volume estimated from external head measurements and corrected for body size. East Asians and their descendants average about 17 cm3 (1 in3) larger brain volumes than do Europeans and their descendants, whose brains average about 80 cm3 (5 in3) larger than do those of Africans and their descendants. Jensen (1998, pp. 442-443) calculated an “ecological” correlation (used in epidemiological studies) of +0.998 between median IQ and mean cranial capacity across the three populations of “Mongoloids,” “Caucasoids,” and “Negroids.”
Finally, The g Factor considers the race differences from an evolutionary perspective. Jensen accepts the “Out-of-Africa” theory, that Homo sapiens arose in Africa about 100,000 years ago, expanded beyond Africa after that, and then migrated east after a European/East Asian split about 40,000 years ago. Since evolutionary selection pressures were different in the hot savanna where Africans evolved than in the cold Arctic where Mongoloids evolved, these ecological differences had not only morphological, but also behavioral effects. The farther north the populations migrated ‘Out of Africa,’ the more they encountered the cognitively demanding problems of gathering and storing food, gaining shelter, making clothes, and raising children during prolonged winters. As these populations evolved into present-day Europeans and East Asians, they underwent selective pressure for larger brains.
The g Factor’s strong conclusion about race differences in fact came as something of a suprise to me. In all my discussions with Jensen about race differences since 1981, I had been struck by his careful circumspection. More than once he went so far as to say that he doubted that methods were available for determining whether Black-White differences were heritable (including the methods of behavior genetics). As best I recall, he said something like: “We can never ‘prove’ for certain that the race differences in IQ are heritable in the sense that we can ‘prove’ something in mathematics. All empirical science can do is increase the probability that genetic factors are involved.”
Pushing Out the Envelope Even Further
Science is a never ending journey and Jensenism has traveled far since 1969. With regard to the significance of brain size, for example, early on, Jensen described brain size as unrelated to IQ (1969a, p. 73; 1973, p. 333, 349), and did not cite the literature on racial differences in brain size. Somewhat later, in Bias in Mental Testing (1980), he cited Van Valen’s (1974) re-assessment of the literature showing a +.30 correlation between brain size and IQ along with a Table from Hooton (1939) showing a linear relation between head size and socioeconomic status. By 1984, Jensen cited Ho, Roessmann, Straumfjord, and Monroe’s (1980) autopsy studies showing a Black-White brain weight difference of about 100 grams and outlined a variety of ways to examine relations between race, brain-size, and IQ. By the time of The g Factor, Jensen’s own studies had shown that head size was related to IQ even within-families, that the head size/IQ relationship occurred on the most g-loaded tests, that Blacks and Whites differed in head size, and that the Black and White differences in head size disappeared when Blacks and Whites were matched for IQ.
The conclusion that there are racial differences in average brain size is becoming accepted. For example, Ulric Neisser, Chair of the recent American Psychological Association’s Task Force Report on The Bell Curve (Neisser et al., 1996) acknowledged that, with respect to “racial differences in the mean measured sizes of skulls and brains (with East Asians having the largest, followed by Whites and then Blacks). . . .there is indeed a small overall trend” (Neisser, 1997, p. 80).
From the beginning, Jensenism did not stop with IQ. For example, Jensen (1 969a, p. 86) cited studies showing the early development of motor behavior in Black infants with some Black samples at six months of age scoring nearly one standard deviation above White norms. Paralleling the behavioral precocity, Jensen (1969a, p. 87) reported evidence of faster bone development in Black infants (established using X-rays) and earlier maturation of brain wave patterns (measured using EEGs). Soon after, Jensen (1973: 289-290) suggested that race differences in the production of two-egg twins, being most common among Blacks and least common among East Asians, with Caucasians intermediate, “may be a reflection of evolutionary age.” In a long footnote, he wrote: “[T]he three racial groups lie on a developmental continuum on which the Caucasian group is more or less intermediate. A related fact is that there is an inverse relationship throughout the phylogenetic hierarchy between the tendency for multiple births and the prolongation of immaturity.”
As a committed Jensenist, I pursued these hypotheses with vigor and proposed a gene based “life-history theory” familiar to evolutionary biologists as the r-K scale of reproductive strategy to account for the racial trade-off between brain size and egg-production, and other variables (Rushton, 1995). At one end of this scale r-strategies emphasize high reproductive rates while at the other K-strategies emphasize high levels of parental investment. This scale is generally used to compare the life histories of widely disparate species, but I used it to describe the immensely smaller variations within the human species. Following Jensen’s trail I went on to hypothesize that Mongoloid people are, on average, more K-selected than Caucasoids, who in turn are more K-selected than Negroids. My book Race, Evolution, and Behavior documents the reality of racial differences in over 60 physical and behavioral traits.
In recent years, the equalitarian dogma has run headlong into some very bad karma. In the wake of the success of The Bell Curve (Herrnstein & Murray, 1994), and other recent books that provide race-realist answers to the question of differential group achievement, there has been an intense effort to get the ‘race genie’ that Jensen’s l969 Harvard Educational Review paper loosed safely back in the bottle, to squeeze the previously tabooed toothpaste back in the tube. By firmly establishing the psychometric, neurophysiological, behavior genetic, and comparative evidence for the existence and importance of Spearman’s g, Jensen’s The g Factor makes it near certain that such obscurantist efforts will end up shredded by Occam’s razor.
On Arthur Jensen’s Integrity by SANDRA SCARR, University of Virginia, emerita
Few psychologists have engendered the controversy or endured the abuse that Arthur Jensen has in the past three decades. His adamant adherence to a hard-edged science and an uncompromising personal integrity have led to notoriety. Although these virtues might be rewarded, if applied to less controversial topics. Art Jensen has been vilified because he applied his standards to the most important and painful social issues of our day. In this article, I admire his ethics but trace the negative reactions he evoked. His legacy to psychological science goes beyond important studies on choice reaction times and intelligence, environmental effects on intelligence, and race differences in mental development: Art Jensen set a standard for an honest psychological science.
For more than 40 years, Arthur Jensen has unflinchingly strived to make psychology an honest science. My emphasis is on both words, honest, and science. For this alone, I would admire him enormously, but there is much more to admire about Art’s lifework, which continues unabated by his official retirement. Besides his intellectual mentor, Robert Thorndike, and a few other pillars, such as Lee Cronbach, Robert Woodworth, and Lewis Terman, Arthur Jensen’s contributions tower above educational psychology and psychometrics.
As his own essay (this issue) demonstrates, Art relentlessly pursues a hard-edged, hypothetic-deductive science that treads on a more emotional, humanistic psychology. Art has no sympathy for mushy thinking. For him, impressions and feelings are not data and have no place in psychology, beyond perhaps the hypothesis-formation stage. Art is ruthlessly scientific: If hypotheses derived from a theory cannot be tested by logical experimentation and data analysis, the theory does not deserve to be called psychological science.
Art rejects convenient compromises and politically expedient obfuscation. These virtues have not been universally appreciated. I have never known him to evade a controversy or mollify an opponent, when the intellectual stakes are high. Outspoken and bloodlessly calm in the face of threats, Art confronts the most emotional critics with logical argument and polite disdain. He remains agnostic where data do not drive him to a conclusion, and his agnosticism on matters of test bias, IQ testing, and racial differences in “g” has cost him dearly. Even with his back to the wall, he continues to proclaim the facts, as he sees them.
He exposes intellectual dishonesty in whomever he finds it, and there is plenty of intellectual dishonesty to find among our Politically Correct colleagues. Art is an important player in battles against the kind of naive environmentalism that has squashed constructive, scientific contributions from psychologists to the most important educational issues of our time, from Head Start to special education to university entrance requirements. Although we are both infamous for exposing naked Emperors, I may be just a tiny bit more tolerant of the bleeding-hearts among us — a weakness that has saved me from much of the abuse he has suffered.
Art’s own studies of learning processes and “g” unwaveringly follow models derived from physical sciences. Psychological science consists of rigorous experiments, psychometrically credible tests, and sophisticated data analyses. He is an unapologetic reductionist, who believes that complex processes will always be explainable in simpler, component terms.
For Art, mind is no more than brain chemistry. In this belief, he clearly rejects systems theories and cognitive theories of mind, in favor of mechanistic, physical models. For those who believe that the whole may be more than the sum of its component parts, especially in biological systems, and that experience is constructed by minds, Art’s strict adherence to physical science model may seem anachronistic.
Determined and persistent, Art followed several lines of research on learning and intelligence. To my mind, his three most important research contributions are:
1.The elegant series of studies on reaction times in complex, choice tasks;
2.His studies of older and younger siblings in California and Georgia to test competing genetic and environmental hypotheses about racial differences in IQ; and
3.The clever construct validity studies, matching the performance of younger White children to that of older Black children on tasks where Black-White difference are most prominent.
In the series of studies on reaction times, he showed that brain functions — speed, reliability, and capacity — can be measured in seemingly simple reaction time tasks that are importantly related to psychometric “g” and by extension to many forms of academic and other life achievements. Despite carping by critics from the narrow world of experimental psychology, Art showed real-life implications for laboratory tasks that heretofore ha4 gone unnoticed, except among laboratory psychologists. (I was there in Britain where the mocking of some learning researchers was extremely distasteful to all but the nastiest high-table fools.) In characteristic fashion, Art ignored the ad hominem slurs and persisted to show how important their seemingly trivial tasks really were. Art succeeded in giving psychometric “g” some important physical correlates (he might say physical bases, but I won’t go that far). That line of research has many more miles to go.
Closer to my interests, in the second example, Art saw an opportunity for a naturally occurring experiment — the comparison of older and younger siblings, as a test of competing theories about the origins of racial differences in IQ. He reasoned logically (as always) that if environmental deprivations were responsible for lower test scores of Blacks, then the longer children were exposed to such environments, the more they would lag behind test norms; that is, the lower their IQ scores would become. Older siblings have longer exposure to such deprivations; hence, they ought to score lower on IQ and standardized achievement test than their younger sibs. If, on the other hand, genetic differences were primarily responsible for Black-White differences, then no older-younger sibling differences should be observed. Among Berkeley, CA school children, no older-younger sibling differences on tests were observed. In poverty-stricken, rural Georgia, however, the environmentally predicted declines in test scores were found.
More developmental psychologists are embarrassingly glib on racial differences; Any observed Black-White difference must be due to “racism,” social disadvantage, and other neighborhood and school features, because they correlate with IQ. By using sibling comparisons, Art showed that such excuses (I refuse to call them explanations) were not true in Berkeley, where exposure to the mainstream culture is extensive for even the poorest minority children, whereas in rural Georgia, restriction of learning opportunities explained the sibling IQ differences. These studies showed that in really deprived rearing circumstances, even Art Jensen can find environmental effects! Kidding aside, these studies of sibling differences in IQ are all the more important because Art did them. One can only hope his critics will remember to attribute them to him.
The studies of sibling IQ differences in California and Georgia helped me to think about what kinds of environments have negative effects on intellectual development and which do not. Our own adoption studies found that children adopted in infancy into working class families achieved IQ levels as high as adoptees reared in privileged professional families, whereas biological offspring of such families differed by 10 IQ points, on average. Clearly, genes were the major cause of social class differences in IQ, not whether parents take their children to ball games or museums, or whether they listen to Country & Western tunes or to Mozart (take that, Art-the-music-snob). These results, and Art’s sibling studies, led me to stand up for “good-enough parents,” who provide loving support and learning opportunities, but not necessarily those the intelligentsia value most. My proposal, that most parents are “good enough” at child rearing to support their children becoming the best they can be, provoked PC colleagues to attack me as anti-child welfare, because surely every child needs to have parents just like them to become the best (their self-serving snobbery is appalling and unrecognized).
Since our working class Midwesterners were doing as good a job with their adopted children as their highly educated compatriots, my conclusion about “good enough” parents is logically inescapable. So is the conclusion from Art’s research; to wit, the African-American families in California did expose their children to learning opportunities sufficient to maintain their intellectual growth over the school years. The fact that their IQ test scores lagged behind those of Whites is not likely to be explained by differences in learning opportunities.
An interesting parallel to this work is our longitudinal study of interracial adoptees. At the average of 7 years, the African-American adopted children scored 106.1 on IQ tests. By the average age of 18 however, their IQ scores had declined to 96.8. Children with one White and one Black parent scored, on average, 109.0 at age 7 and 98.5 at age 18; children with two Black parents (and later adoptive placements) scored 96.8 at age 7 and 89.4 at age 18. The test performance of the Black/Black adoptees was not different from that of ordinary Black children reared by their own families in the same area of the country. My colleagues and I reported the data accurately and as fully as possible, and then tried to make the results palatable to environmentally committed colleagues. In retrospect, this was a mistake. The results of the transracial adoption study can be used to support either a genetic difference hypothesis or an environmental difference one (because the children have visible African ancestry). We should have been agnostic on the conclusions; Art would have been.
A less recognized line of research, and one with great implications for developmental psychology, is Art’s use of younger White children to model the test performance of older Black children. By showing that response and error patterns of Black children matched, on average, those of White children two years younger, Art did more than challenge the test-bias literature. He showed that differences in test performance among age-matched White and Black children can be most simply explained as differences in rates of mental development. The implicit analogy to physical growth is powerful: Slower growth rates over the same length of time lead to lesser final attainments, whether one is speaking of height or of intelligence. The implications of these studies are truly frightening, but Art does not flinch. I have yet to see these findings incorporated into introductory psychology textbooks or developmental texts, however, so the wrath of Politically Challenged has not rained down on him yet.
Among his many works, those that will be most widely cited and remembered are his rigorous reviews of data on test bias, evidence for the “g” in general intelligence, and reviews of research on group differences in IQ and achievement. In scholarly yet accessible prose, Art tells coherent stories that make the best sense of complex theories and data. Along the way, he refutes the many ad hoc claims about test bias, disposes of theories of multiple intelligences, and lays waste to naive environmental theories of race and social class differences in educational achievements. In a dozen impressive books and hundreds of articles, spanning 30 years, Art has brought uncompromising logic and scientific rigor to the most controversial topics of our age.
In my last term at the University of Virginia, I taught an undergraduate course on intelligence. The text was Bias in Mental Testing. At first, some students were surprised and even alarmed that many of their assignments were drawn from a book by that infamous Dr. Jensen. But they came to appreciate the serious nature of the book and its helpful chapters on testing, validity, reliability, and potential biases in mental tests. By the end of the semester, they felt they had accomplished several feats — to have read nearly all of the 700+ pages and to have passed tests on the content. Another accomplishment was their open minds about the content and the author, whom they came to admire. It’s a splendid book.
Art seems to have been genuinely surprised by the notoriety he attained from his writings on race and IQ. Others cannot understand his surprise. When one lobs hand grenades at the intelligence and potential achievements of others, one should anticipate a violent reaction. For Art to say that only 5% of the Harvard Education Review article concerned racial difference in IQ is like saying the only problem Lincoln had in the time he attended Ford’s theater was the split second he was shot. Somehow, the percentage is not the critical issue in either case.
Anticipated or not, the consequences of his notoriety were severe and prolonged. Few can claim to be, or to have been, as sorely tested as Art has been in defense of psychology as a science. I have witnessed his steadfastness in the face of a screaming, unruly mob who disrupted his lecture on learning and intelligence and threatened his personal safety. I learned what it was like to be spat upon and to put my body on the line to get Art out of a University of Minnesota auditorium. It was shocking and frightening, as surely the radicals intended, but it was most of all infuriating, because no disciplinary actions were taken against those who assaulted us. Those were the wonderful 1970s.
As he mentions in his essay (this issue), his automobile tires were slashed, police had to open his mail, and his office at the University of California-Berkeley was stripped bare to protect him from a potential bomb. Art’s office at Berkeley was more like a San Quentin cell than a typically cluttered faculty office. His family was threatened, and his personal freedoms seriously compromised — all because he reported his conclusions about genetics and IQ, based on a serious scientific review of the research literature.
By his own account, he is no extravert. Nor, I may add, did warmth and humor soften the acrimonious exchanges he had with hostile audiences. One might also observe that insight into his violent, enraged opponents was lacking. The logical, unemotional Dr. Jensen would never behave in such an uncivilized manner, nor comprehend those who do.
Art Jensen has also endured abuse from thugs with pens instead of megaphones. Personally, I have no empathy for politically driven liars, who distort scientific facts in a misguided and condescending effort to protect an impossible myth about human equality (= identity). Art believes he understands the motives of the Marcus Feldmans, Steven Jay Goulds, and Leon Kamins of the intellectual world. They seem to speak his language, albeit with forked tongues. I find them despicable, because they have the knowledge and intellect to know that they deliberately corrupt science. To deny falsely the scientific evidence that nearly all measurable human traits are moderately to highly heritable is to deny parents and policy makers essential knowledge to run their own lives and the society as a whole. Self-appointed saviors of the equality myth are far more dangerous to an honest psychological science than a hundred outraged groupies who don’t know that the lecture was supposed to be about, anyway.
All in all, with clear conscience, Art stands up for data, searches for the most logical and supportable explanations, and rejects all of the ad hominem garbage thrown his way.
I did observe a humorous episode with the notorious Arthur Jensen. While at York University, we took a little stroll to a neighborhood shop, where another customer asked me if we were from the conference on intelligence. She had heard that the terrible Arthur Jensen was there. “I can’t understand how they could have let him in the country!,” she proclaimed. With Art standing mutely at my side, I told her that Dr. Jensen was indeed present. “Is he as awful as they say?,” she asked. “Oh yes,” I said, “dreadful!”At least that’s the way I recall it.
Art Jensen’s contribution to psychological science are enormous, and they continue to mount. His work includes the impeccable tome on test bias, the most thoughtful research on learning and intelligence, and some critical studies on race and environment. The massive body of work will persist for generations of psychologists. Yet, I believe that his most important contribution is intellectual honesty and integrity to a psychological science that is threatened with Politically Correct corruption. Art has not known how to be politically expedient, or to couch his ideas in soothing terms, so that he has often suffered academic rejection. But most people heard you, Art, and they remember, even if they did not like the message. Both inside and outside of academia, your intellectually honest legacy will prevail.
Jensen on “Jensenism” by Arthur R. Jensen
University of California, Berkeley
“Though Jensenism is a term listed in several dictionaries, Arthur Jensen has produced a more extensive body of work than suggested by the dictionary entry. To the public, he is mainly known for his work on the genetics of intelligence. This article discusses the work that is publicly less well known. Work discussed includes studies in learning, memory. the cumulative deficit hypothesis, Spearman’s hypothesis, and speed of information processing, to name a few. The publicly better known work is also discussed. A bibliography of Jensen’s publications is included in an appendix. (Abstract written by D. Detterman)”
To discover that one’s name has entered the dictionary as an “ism” is both flattering and embarrassing, and is cause for reflection. I know because it happened to me. Recent editions of a number of dictionaries contain the word “Jensenism.” The Random House and Webster’s Unabridged Dictionaries, for example, contain the following entry:
Jensenism, n. the theory that an individual’s IQ is largely due to heredity, including racial heritage. [1965-1970]; after Arthur R. Jensen (born 1923), U.S. educational psychologist who proposed such a theory; see -ism] — Jensenist, Jensenite, n., adj.
For those who understand the meaning of heritability in quantitative genetics, the wording is rather inept and the “theory” attributed to me has been around at least since the time of Francis Galton (1822-1911), whose Hereditary Genius (1869) predated the very article that led the popular press to label me a “hereditarian” by exactly one century. The dictionary definition can’t be overly derided, however, as it is quite true that, in 1969, I did present a fairly comprehensive review of the evidence that IQ is substantially heritable and had stated that it is a reasonable hypothesis that genetic as well as environmental factors are involved in the well documented Black-White average difference in IQ. Also, I like to think that I was partly responsible for getting Galtonian thinking back on track in differential psychology after it had been derailed in the behavioral sciences for at least a generation following World War II (the period dominated by what Sandra Scarr once referred to as “naive environmentalism”).
However, the more serious disadvantage of having one’s name turned into an “ism” is that, from that moment on, one is liable to be identified only as the “ism” in the dictionary. The rest of one’s research activity can be unfairly eclipsed, and findings and formulations that are unique and perhaps even fundamentally more important are forgotten. One of my aims here is to forestall this threatened eclipse of other aspects of my research and shine some light on how that which got me labeled as an “ism” fits into the larger orbit of my lifetime’s work.
Essentially, I have always been a differential psychologist. Human idiosyncracies and individual differences in behavior interested me before I had ever heard of psychology. The first book I read on the subject, more or less by accident while in high school, was J.B. Watson’s Psychology From the Standpoint of a Behaviorist (1929). It was probably the main reason I chose to major in psychology in college, after reluctantly but realistically deciding not to pursue a career in music. Though I became acquainted with some well-known psychologists, such as Edward Tolman and Egon Brunswik, as an undergraduate psychology major at Berkeley, the one psychologist whose work most captured my attention (but whom I never saw in person) was the then Sterling Professor of Psychology at Yale, Clark L. Hull, a latter-day Watsonian and Pavlovian behaviorist. One could say that I became a Hullian, and I recall writing a long term paper for one of my courses extolling Hull’s theory of learning — excessively so, according to the comments of the TA (one of Edward Tolman’s graduate students) who graded my paper. Primed, I suppose, by Watson, I was especially attracted to Hull’s purely mechanistic system for explaining behavior, as spelled out in his Principles of Behavior (1942). B.F. Skinner’s Behavior of Organisms (1938) was also appealing but lacked the systematic theoretical system that made Hull’s approach seem more promising to me.
I was totally unaware at the time that these now classic works in psychology, and indeed my whole undergraduate education in psychology, neglected individual differences and the influence of genetic factors on behavior. These topics were scarcely admitted as part of the field of psychology, at least as it was presented at Berkeley in the 1 940s. Experimental psychology dominated the department at that time, and the implicit assumption of experimental psychology was that individual differences in the behavioral realm originated entirely outside the organism, through its exposure to different environmental contingencies, and they could be explained, if one were at all interested in doing so, in the purely stimulus-response-reinforcement terms of conditioning and learning. In its focus on discovering general laws or principles of behavior, experimental psychology traditionally treated individual differences as a nuisance variable, or as merely error variance in the statistical analyses of its data.
This limited perspective of my undergraduate courses in psychology was extremely implicit and so completely taken for granted that it did not enter my consciousness until some years later. I occasionally meet psychologists even today who think of individual differences as error variance or as purely a product of environmental diversity. I was still largely operating on this assumption in 1964 when I wrote a major paper that attempted to explain social class differences in scholastic learning entirely in terms of the then current S-R theories and principles of verbal learning (67). Ironically, the publication of that paper was so long-delayed that it appeared after my position on the major basis of individual and group differences had changed in a hereditarian direction. Large differences in the publication lag of one’s articles and book chapters during certain periods may even create a false impression of contradictory vacillations in one’s theoretical stance. The publication dates of one’s articles are not always perfectly correlated with the actual chronology of one’s changing position on theoretical issues.
Thanks to the beautiful “recreational reading” room (the Morrison Library) on the Berkeley campus, where I spent most of my evenings, I believe I got as much or more of my undergraduate education from entirely self-selected extracurricular reading as I got from my courses and textbooks. The most lasting influence I recall are works by M.K. Gandhi, Bertrand Russell, G.B. Shaw, Havelock Ellis, H.G. Wells, Aldous Huxley, Alfred Korzybski, and biographies of famous musicians, scientists, and philosophers. They instilled a certain critical sense as well as a humanistic idealism that, in the long run, made a greater impression on me than did most of the relatively uninspiring textbooks I was required to read in my courses. To make more time for the reading I most liked, it was my policy never to read anything in my college textbooks more than once.
It was my extracurricular reading, probably more than anything else, that led me to look for the ways psychological science might have practical applications that could benefit individuals and society. Some years later when I decided to enter graduate school to work toward a Ph.D., I examined various university catalogues to see what they offered in applied psychology. I recognized more of the names of psychologists whose works in applied areas, such as clinical and educational psychology, that I had already come across in my reading on faculty of Teachers College, Columbia University than in any other university catalogue. Egon Brunswik’s course on the history of psychology had also left me with a distinct impression of Columbia as having one of the pioneer departments of psychology, shaped by such luminaries as James McKeen Cattell, E.L. Thorndike, and Robert Woodworth. (When I arrived at Columbia, Woodworth was still lecturing at age 87, and I audited his two courses.) The fact that Columbia University is located in New York City, home to Carnegie Hall, Toscanini and the NBC Symphony, the New York Philharmonic, and the Metropolitan Opera, provided a powerful added attraction. The musical capital of America, New York is visited each year by many of the world’s greatest orchestras, conductors, and virtuosos. And my interest in music has never been second to my interest in psychology, though I have necessarily devoted more time to the latter, of course, since it has been my livelihood. When I wasn’t on the Columbia campus, chances are I was hanging out in Carnegie Hall, either at a concert or a rehearsal.
At Columbia’s TC I studied educational, clinical, and personality psychology. My Ph.D. dissertation (under Professor Percival Symonds) was on the Thematic Apperception Test as a measure of aggression (2,3). 1 found my three years as Symonds’ research assistant much to my liking. However, my interest in the subject of his research at that time, based on the psychoanalytic or “dynamic” interpretation of various projective techniques (8, 20), proved short-lived. Though Symonds was a man of noble character and in many ways a fine mentor, my three years at TC were probably influenced more by the lingering shadow of the psychologist who had been Symonds’ mentor but who had died three years before I arrived at TC — Edward L. Thorndike, probably America’s greatest psychologist. Thomdike’s influence and his conception of psychology still pervaded the intellectual atmosphere at TC during my tenure and was repeatedly reinforced by an imposing portrait of the great man that hung on the wall above the card-catalogues in the TC library. I felt compelled to read some of Thorndike’s books and I liked them a lot, especially for their clear thinking and their objective and empirically anchored approach to the remarkably broad range of subject matter in psychology with which he dealt.
It is amazing how much of what today is viewed as established fact in psychology was either discovered or presaged by E.L. Thorndike. As he was one of the leading pioneers of psychology as a natural science, he became the first of my “heroes” in psychology; the other two (Galton and Spearman) I discovered a few years later. These are the three psychologists whose key works I return to and re-read for their wealth of hypotheses, original and insightful ideas, and inspiration, always to be rewarded. If there have been any authentic geniuses in the history of scientific psychology, in my estimation they include at least these three. (I have written about Galton [238, 352, 3831 and Spearman [239, 353, 383].)
During my clinical internship at the University of Maryland Psychiatric Institute in Baltimore, I examined a great variety of psychiatric patients, using all of the prevailing techniques of clinical psychology, and typically wrote a clinical report on each patient. During my training experience in psychotherapy, I quickly came to realize that I was less satisfied and less effective working with people directly than in working with data. I did not enjoy many of the routine aspects of clinical work, probably because I am quite low in extraversion. Hence I welcomed collaboration with one of my clinical supervisors in some research we did on the Rorschach as an index of pathological thinking that completely eschewed the traditional systems of scoring Rorschach protocols and was solely based on characteristics of the subject’s verbal expression (5; see also 39, 46).
It was also at this time that I began seriously reading the books and articles of Hans J. Eysenck, who was then a maverick personality researcher and the professor of psychology in the University of London’s Institute of Psychiatry. Most of my evenings that year were spent reading every book and article by Eysenck that I could find in the university library, as well as many of the references he made to the influential work of others, particularly Galton, Spearman, and Thurstone. This provided a much needed antidote to the predominantly Freudian or psychoanalytic concepts that informed my clinical work. It was not so much the specific aspects of Eysenck’ s own theories or his research, but rather his general approach to psychology as a natural science that provided my first real sense of finally having discovered my true vocation. I felt I was no longer groping for the path that I should take to make my life seem worthwhile. I believed that research and teaching in the field of differential psychology, broadly conceived, was exactly the path for me. So I wrote to Eysenck for his permission and applied to NIMH for a postdoctoral fellowship to spend a year in Eysenck’s department in London. Luckily, both were granted and, with a year’s extension of my fellowship, I had the good fortune to spent two full years with Eysenck. (Six years later I returned to his department as a Guggenheim Fellow for another full year during my first sabbatical leave from Berkeley [detailed in 149, 378]).
I emphasize my postdoctoral work with Eysenck, because I believe it planted the seeds of virtually everything I have done since then. It put me on the path that I have followed, in one way or another, for all of my later research. Although each of the many subsequent byways could not have been anticipated, they all led more or less consistently in one general direction — what came to be known as the London School of differential psychology, originated by Galton and with Spearman, Burt, and Eysenck successively as its leading exponents (283, 376, 377). (I knew personally only Eysenck  and Burt [126. 225, 326, 367].) The London School is not really a school or even a doctrine or a theory. Rather, it is a general view of psychology as a natural science and as essentially a branch of biology.
Its central concern is variability in human behavior. It is Darwinian in that it views both interspecies variation and an important part of intraspecies variation (both individual and group differences) in certain classes of behavior as products of the evolutionary process. It is behavior-genetic in that the evolutionary process depends upon genetic variation and selection, and the neural basis of behavioral capacities is subject to these evolutionary mechanisms the same as other physical characteristics. It is quantitative in that it emphasizes the objective measurement and taxonomy of behavior and the operational definition of latent traits or hypothetical constructs. It is analytical in that it subjects quantitative data to mathematical formulation and statistical inference. It is experimental in that it typically obtains measurements, both behavioral and physiological, under specifically defined and controlled conditions. It is reductionist in that it aims theoretically to explain complex phenomena in terms of simpler, more elemental processes. It is monistic (as opposed to dualistic) in that it neither posits nor seeks any explanatory principle that does not consist of strictly physical processes; it views complex psychological phenomena as emerging solely from interactions among more elemental neurophysiological processes and their past and present interactions with environmental conditions.
Besides the extensive reading, studying (courses in multivariate statistics with Patrick Slater and factor analysis with A.E. Maxwell), and writing (4, 6, 7, 9, 14) that occupied my time as a postdoc, I undertook laboratory research on individual differences in the effects of massed and distributed practice in selective stimulus-response learning and I devised a special apparatus for the directly measuring individual differences in reactive inhibition independent of any form of learning per se (51). The specific hypotheses I tested derived from Eysenck’s theory of the basic neural processes responsible for individual differences in extraversion-introversion (or E, as it was called), which had been established as a continuous unitary trait by the factor analysis of personality inventories, behavioral ratings, and objective behavior measurements derived from certain laboratory techniques. Eysenck’ s theory of E at that time brought me back to Clark Hull’s theory of learning, which had first fascinated me as an undergraduate. I became a born-again Hullian, this time around becoming more thoroughly versed in every facet of Hull’s theory and most of the theoretical and empirical literature related to it, including Pavlov’s classic work on conditioning.
Eysenck’s theory held that the basis of E is the rate of build-up of a hypothetical neural process called reactive inhibition, or IR (as defined in Hull’s system). The theory contends that trait extraversion reflects a more rapid build-up and a slower spontaneous dissipation of IR under the conditions in which IR is hypothesized to be manifested in behavior, such as the experimental extinction and spontaneous recovery of conditioned responses, the effect of massed trials versus spaced trials in serial rote learning, and the reminiscence effect in motor learning (as on the pursuit-rotor). It was this aspect of Eysenck’s research program that led me into theories of learning and the experimental psychology of human learning, which soon completely overshadowed my interest in personality research. I saw the study of individual differences in learning in its own right as a more fundamental and scientifically researchable subject than the study of personality. The last postulate in Hull’s behavior theory in its final form (in 1952) states that individual differences in learning, or excitatory potential (sEr), are a net product of individual differences in each of the hypothetical constructs in his system, such as habit strength (sHr), drive (D), reactive inhibition (IR), conditioned inhibition (sIr), sensory limen or threshold of stimulus activation (sLr), and spontaneous oscillation of reaction potential (sOr). I thought this approach provided the basis for a rigorous, quantitative and experimental approach to the study of individual differences in human learning. I later elaborated on this idea in a paper that, in retrospect, strikes me as an overly ambitious and practically unfeasible program for research on individual differences in human learning (59). Since the largest part of the individual differences variance in the forms of learning that are important for education and the acquisition of many other real-life skills is factorially indistinguishable from Spearman’s g, or general intelligence, I now believe a program of research on the nature of g to be probably more fruitful than focusing on learning per se (189, 301).
But before going on with my story, I should point out what may not be well known to younger readers, that Hull’s system, which dominated the learning field from about 1940 to 1960 (he died in 1952), waned rapidly in the early Sixties and became virtually extinct by 1970. Since then, Hull’s citation index has hovered close to zero. This is a remarkable fate for one who, for over a period of at least twenty years, many considered the leading theorist in scientific psychology. In marked contrast, E.L. Thorndike, 48 years after his death, remains among the 100 most frequently cited psychologists in recent decades, and the number of citations of Charles Spearman has increased in each decade since his death (in 1945) and risen most rapidly since 1970.
Thorndike’s and Spearman’s intuitions, hypotheses, theories, and the phenomena on which their interests were focused, mainly learning and cognitive ability, were evidently more important, more original, and scientifically more productive than Hull’s precisely formulated theory of learning, however impressive his achievement seemed in its day. The problem, I think, was not Hull’s in use of the hypothetico-deductive method, which I believe was exemplary, but that the many interrelated parts of his whole grand theoretical edifice, its postulates (as Hull called them), were erected on too slim a foundation of empirical studies. Hence the subsequently growing number of experiments inspired by the theory and devised to test it increasingly failed to confirm its predictions. Though modifications and additional ad hoc principles were proposed to meet the explanatory demands imposed by new empirical evidence, Hull’s system gradually collapsed beyond repair and was eventually discarded, much like the geocentric theory in astronomy and the phlogiston theory in chemistry. In the history of science, of course, this is a perfectly respectable demise for a theory. The really fatal shortcoming of Hullian theory, however, was its nonbiological behaviorism, a position that was bound ultimately to leave it theoretically barren.
Rather early in my career, while still a Hullian, I tried to modify Hull’s theory to make it accommodate some of the contradictory experimental evidence by proposing a fundamental mathematical reformulation of the treatment of reactive and conditioned inhibition within the Hullian framework (18). But this kind of ad hoc doctoring could not save Hull’s system any more than postulating retrograde motion of the planetary epicycles could preserve Ptolemy’s geocentric theory. Though I gradually lost interest in Hull’s theory, my interest in human learning, particularly in its individual differences aspect, was undiminished.
Now that the stage is set, with a backdrop of the values and attitudes against which all my later activity can be more understandable, I will provide a brief account of the specific studies that I believe mark the key points in my research activity, and how and why I moved from each point to the next. Studies never arise from thin air, of course, but also one does not have to go looking for things to research. Each new project, it seems, is absolutely compelled by the preceding ones, or by one’s purposeful and critical reading of the literature or by one’s self-criticism and others’ criticisms of one’s previous work. The completion of each study always leave some loose ends. Problems abound and one continually searches for what seems the most fruitful path toward each problem’s solution. Given the pages allotted me, this account is necessarily quite telegraphic, referencing only my main publications on each topic. A perusal of my bibliography (see Appendix) indicates that my publications fall into six main categories: (i) clinical and personality, (ii) human learning, (iii) behavior genetics, (iv) racial-cultural differences, (v) test bias, and (vi) mental chronometry and g theory. (I will ignore the first category, with some dozen or so articles, which in retrospect I consider of much less importance or interest.)
Human Learning [deleted for brevity]
The Behavior-Genetics of Intelligence
When, in 1966, I was invited to spend a year at the Center for Advanced Study in the Behavioral Sciences, I had enough research material on the learning characteristics of culturally disadvantaged children, as they were called at that time, that I thought I could best spend my year at the Center writing a book about my findings. I took all of my research material with me and began work at the Center, a wonderfully undisturbed and heavenly atmosphere for study and writing, with a most helpful staff and the intellectual companionship of the many other fellows at the Center.
A reasonably comprehensive book about the educationally disadvantaged children, I thought, should contain one short chapter addressing the issue of the inheritance of intelligence, if only to show that this line of explanation for individual and group differences in scholastic performance could be dismissed as outmoded and scientifically discredited. I had never given this topic much thought and knew shamefully little about it at that time. It had never been touched upon in my entire education to that point and the subject was generally either unmentioned or scorned by virtually everyone I knew working in the field of learning and the educationally disadvantaged. My first exposure to it was in 1957 during my postdoc in London, when I attended a lecture on “the inheritance of mental ability” by Sir Cyril Burt. Though I was highly impressed by Burt’s lecture as a brilliant tour de force, its subject didn’t really capture my interest at that time. Burt’s lecture was later published in the American Psychologist (1958, 13, 1-150), and it seemed a good place to start what became my program of reading virtually the entire world literature on the genetics of mental ability at that time. I hadn’t expected to go that far into the subject, but the more I read, the more I realized it couldn’t be dismissed and had to be taken seriously.
In order to be able to evaluate much of what I read, I had to tackle the technical aspects of quantitative genetics. Luckily, a professor of ethology and genetics was also a fellow at the Center that year and was a most helpful tutor and guide to the literature on quantitative genetics. I felt most resentful of the fact that I had reached that stage of my education and of my career and had not been exposed to the existing scientific knowledge on the genetics of mental ability. I was even more dismayed to realize that my case was all too typical of those working in most branches of psychology, particularly experimental, educational, and clinical. All human variation in abilities was attributed to the learning opportunities afforded by different environmental and cultural circumstances to which individuals were exposed. Though at that time the literature on behavior genetics was but a fraction of its present volume, what there was seemed sufficient to call in question the prevailing 100 percent environmentalism of the l950s and ’60s. My task was cut out for me: to help dispel the ignorance that generally prevailed in educational psychology concerning the role of genetic factors. In reading E.L. Thorndike, the father of American educational psychology, I found that he was on the right track in his intuition about the importance of genetic factors in individual differences, but his line of thought on this subject rapidly went out of fashion shortly after World War II, for no good scientific reason.
Therefore, during my year at the Center, I wrote several articles that stemmed from my new-found interest in the genetics of mental ability and its implications for education (61, 62, 63, 64, 68, 70). The most frequently cited of these articles is based on my examination of the famous Hoizinger formula for estimating heritability from the difference between the correlations between MZ twins and between DZ twins. I showed that Hoizinger’s formula, which was virtually the only one ever used in studies of the heritability of intelligence up to that time, did not estimate heritability as it is defined in quantitative genetics, nor did it take account of the effect of assortative mating on the estimation of heritability from twin data (6 1; see also 178). I provided a new formula that not only accorded with the meaning of broad heritability as defined in genetics but also took account of assortative mating. (The formula could also be generalized for estimating heritability with other kin-ships besides twins, such as full siblings and half siblings.) I used this formula to recalculate heritability coefficients for IQ on every published study of MZ and DZ twins.
Although the articles I wrote that year emphasized the evidence for the substantial heritability of individual differences in IQ, I thought (and wrote) that it was unnecessary to invoke genetic causes for the observed racial differences in IQ, which I thought could be explained in terms of cultural bias in the tests and poor environmental opportunities for acquiring the particular knowledge and skills called for by conventional tests. One of my articles written at the Center (63), which originated as an invited address at a convention was titled “How Much Can We Boost IQ and Scholastic Achievement?” It came to the attention of the editors of the Harvard Educational Review, and in 1968 (when I was back at Berkeley) they asked me to expand it into a more comprehensive article for the Spring, 1969 issue of the Review. They even provided an outline of the topics they wanted me to deal with in the article, including my view on the heritability of race differences (which I had not previously mentioned). I gladly accepted the editor’s invitation, as an opportunity to consolidate what I had been studying and thinking about during my year at the Center. The result was a 200-page typescript which became a 125-page article in the Harvard Educational Review (HER) titled “How Much Can We Boost IQ and Scholastic Achievement?” (76). Though unexpected at the time, I suppose it was the article that forever changed my life, for better or worse.
My Year of Turmoil
Based on a review of the empirical literature, my HER article made four main claims:
(i) experimental attempts to raise the IQ of children at risk for low IQ and poor scholastic performance by various psychological and educational manipulations had yielded little, if any, lasting gains in IQ or scholastic achievement; (ii) individual differences in IQ have a high heritability (.70-.80, corrected for attenuation), but environment also plays an important part; (iii) most of the exclusively cultural-environment explanations for racial differences in IQ and scholastic achievement were inconsistent and inadequate, so genetic as well as environmental factors should be considered; and (iv) certain abilities, particularly rote-learning and memory (i.e., Level I ability) have only a weak relationship to IQ, which suggests that these Level I abilities might be used to compensate to some extent for low IQ (i.e., Level II ability) and thereby make school instruction more beneficial for many children, regardless of their racial or social class background, who are below average in Level II but are average or above in Level I. (Pupils with this pattern of abilities constitute the majority of those who are most at risk for failure under traditional classroom instruction.)
Viewed as a whole, it seemed quite reasonable. But it was the few pages on race differences in IQ and achievement (about 5% of the article) that aroused so much sound and fury, most of it focused on the one sentence that violated what I later came to realize is the greatest taboo in the latter half of the twentieth century. Here is what I wrote concerning the Black-White difference in IQ: “The preponderance of the evidence is, in my opinion, less consistent with a strictly environmental hypothesis than with a genetic hypothesis, which, of course, does not exclude the influence of environment or its interaction with genetic factors” (76, p. 82).
That one aspect of the article was blown up by the mass media, with feature articles in TIME, Newsweek, LIFE, U.S. News & World Report, the New York Times Magazine, and many other newspapers and magazines, as well as radio and TV programs. The Berkeley campus was in an uproar for weeks (and sporadically for months and even years thereafter) with bands of demonstrators disrupting my classes, slashing all the tires on my car, and painting swastikas on my office door. The student paper, The Daily Cal, carried many denunciations and only a few defenses of my position, and there were demands from dissident groups that I be fired. The campus police assigned plainclothes bodyguards to accompany me whenever I left my office, and for several months the campus bomb squad handled the screening and opening all of my mail, even some of the unidentified mail received at my home. There were telephoned and mailed threats on my life and on my family; phone calls were routed (and recorded) through the local police station. A number of the calls that came in over one period of several days so worried the police that they urged me and my family to spend a week away from our home at some unknown location, as the police could not provide 24-hour protection. (We stayed with friends in a neighboring suburb; an inconvenience, but as they had a lovely swimming pool, it was a pleasant diversion.) Worst of all, from my standpoint, was that my on-going research in the Berkeley schools was immediately terminated and permanently proscribed by the Berkeley school officials (128). When I asked one official for an explanation, he remarked, “The Berkeley schools are a political unit, not a research institute.”
Many other harrowing incidents followed, some taking place when I was lecturing on other college campuses, both in the United States and abroad, even when my lectures didn’t touch on the subjects of genetics or race. The largest, most tumultuous demonstration I ever experienced was, surprisingly, at the University of Melbourne, in 1977, where about fifty policemen had to rescue me from a madding mob. The unprovocative topic of my [undelivered] lecture: The relationship between intelligence and learning [189; see also 301]. The very next day the same thing happened in the same setting to Hans Eysenck. His topic: the relationship between personality and learning. (I have written at greater length in the Preface to my Genetics and Education about the bizarre events following the publication of my HER article [112; see also 149].)
The really important consequence of the HER article for my subsequent activity was that it raised a number of questions and issues concerning subjects that called for fuller explication or further research. In many articles (from #77 on) and three books (143, 144, 206), I consolidated my position on these subjects as best as empirical research permitted at that time and launched new research on the remaining unanswered questions and speculative hypotheses. Some people advised me to get out of this controversial area altogether. One eminent psychologist friend warned that if I scorned the Zeitgeist, it would in turn scorn me. However, rather than duck for cover, which I peculiarly felt would be disgracefully un-Gandhian, I resolved not to be whipsawed by the prevailing orthodoxy in the social and behavioral sciences, but to do whatever I could to reform the social sciences. And I believe that at least the scientific community, if not the media and the political establishment, has indeed changed its mind if not its voice over the past 30 years, with an almost total collapse of naive environmentalism and an increasing recognition, at least in the pages of academic journals, of the importance of genetic factors and of environmental factors with biological effects on the development of human mental ability. The well-known survey by Synderman and Rothman (1988) of over 600 psychologists in the relevant fields showed that their modal response on every question that involved the very issues considered heretical in my HER article agreed with the position I then stated. To what extent my own work may have helped usher in the new perspective would be impossible to estimate, but I believe I have played a role. Many other influences, of course, have brought about the demise of doctrinaire environmentalism and advanced the biological orientation of mainstream behavioral science.
Bias in Mental Testing
At about the time of my HER article, the question of culture bias in mental tests was frequently brought up. I was familiar with the early resarch on social class bias in standard tests, pioneered by Kenneth Fells (who had been one of my professors), but I found rather little more than speculation in the literature regarding test bias with respect to racial or ethnic groups. Yet that question was crucial. I reviewed what little evidence existed on the subject in the mid-1960s (69, 99), but found it inadequate and largely unconvincing. Besides the educational, social, and economic unfairness of using tests that are differentially and systematically inaccurate for different racial, ethnic, and social-class groups in our population, I considered also the scientific importance of test bias for the field of psychology in its own right. Psychometrics — the science and technology of mental measurement — is of course basic to many fields in psychology, as indeed reliable and valid measurement is essential for the development of any science or technology. To the degree that the standard psychometric instruments then in use were biased, either by culture, social status, or gender, basic research in differential and educational psychology as well as the practical applications of testing in educational placement, in college admission, in personnel selection, and in assigning recruits to various training schools in the armed services were all compromised by having to operate with deficient tools.
I thought it imperative to devise methods for detecting the presence of various kinds of psychometric bias. This became the main focus of my research effort for the next few years (109, 153, 176, 179, 181, 182, 289). It culminated, in 1980, with the publication of my Bias in Mental Testing (199; see also 202,203,217), an 800-page tome which was then (and may still be) the most comprehensive work on the subject. Research on criteria of bias based on a test’s so-called external validity, that is, its practical predictive validity (i.e., both the regression [and correlation] of criterion measures on test scores) in different sub-populations had already been quite well investigated by psychometricians during the period between 1970 and 1980. Though I fully explicated this work in Bias, my own research contributions emphasized internal indicators of bias, such as whether different groups, (e.g., Black-White, male-female) differ significantly in various psychometric features such as the test’s reliability, the test items’ rank order of difficulty, the test scores’ correlation (and regression) with chronological age, the relative frequency of choosing various distractors (i.e., error responses) in multiple-choice tests, the groups’ similarity in the factor structure, and the groups’ similarity in kinship correlations and heritability values for the test in question.
A methodoligica1 innovation I introduced was the use of what I termed “pseudo-race age groups.” For example, when I found significant differences between Black and White school children in their specific choices of error distractors (in the Raven Progressive Matrices test), I created two “pseudo-race” groups composed entirely of White children, the groups differing in age such that the younger group and the older group had the very same mean difference in total test score as the mean difference between age-matched Blacks and Whites. I discovered that the same-age Black-White differences in the frequencies of selecting a particular distractor (i.e., a wrong answer) among the several distractors for each item on the Raven test were virtually identical to the differences between the two groups of White children that differed in age by almost two years (approximately ages 8 years and 10 years). Applying this method to a variety of tests, including Gesell’s Figure Copying test (a good predictor of scholastic performance in the primary grades), free drawing, and several Piagetian tests, we found that in every feature of test performance, age-matched groups of Black and White children differed in exactly the same way as did “pseudo-race” groups of different-age White children when the age of the younger group is about 80% that of the older group.
Our findings with pseudo-race groups suggested that the observed racial differences in performance were not attributable to test bias, but reflected a developmental difference in rate of mental maturation, with Whites (and more so, Asians), on average, having a steeper trajectory and a higher asymptote. From my own studies and my review of the total empirical literature on test bias, I concluded in Bias: “…the currently most widely used standardized tests of mental ability — IQ, scholastic aptitude, and achievement tests — are, by and large, not biased against any of the the native-born English-speaking minority groups on which the amount of research evidence is sufficient for an objective determination of bias, if the tests were in fact biased” (p. ix). Essentially the same conclusion was announced independently two years later in a joint report by the National Research Council and the National Academy of Sciences (Wigdor & Garner, 1982), which had chosen a panel of nineteen leading experts in psychometrics to review the evidence.
The Cumulative Deficit Hypothesis
One hypothesis proposed in the 1960s to explain the Black-White difference in the trajectory of raw scores on mental tests and on scholastic achievement across grades 1 to 12, held that the increasing racial disparity in test performance with increasing age is the result of a cumulative deficit in learning, such that failure to learn particular knowledge or skills thoroughly at one grade level hinders the ability to learn more advanced material in later grades. Because Black children, on average, begin school having learned less of the prerequisites for learning in the primary grades, they fall further and further below national norms in scholastic achievement with each additional year. Progressive learning deficit is thought to act cumulatively, like mounting credit card debt. This hypothesis was popular in the 1960s and provided much of the rationale for Head Start and other compensatory education programs (54; see also 158, 162, 304, 314, 380).
My investigation of the phenomenon, however, found the evidence for it ambiguous, at best. The divergence between Black and White test scores with increasing age or grade level in school was fully apparent when looking at raw scores, but there was little, if any, evidence for a divergence of Black and White mean scores when the scores are expressed as standardized scores. That is, when measured in age-standardized scores, the mean Black-White difference of about one standard deviation remains constant from kindergarten to 12th grade, because the standard deviation within each group also increases proportionally with age.
I reasoned that if there were a true cumulative deficit effect for IQ, and if it was the cause of Blacks’ lower average IQ, and if IQ declined the longer a child stayed in a culturally disadvantaged environment, then Black children at any given age should have, on average, a lower IQ than that of their younger siblings. The cumulative deficit theory predicts that the positive difference between the ages of the older (O) and the younger (Y) sibling is positively correlated with the Y-O difference in IQ. There should be a substantial such correlation among Blacks (i.e., the older sib should have a lower IQ than the younger sib), but this effect should be negligible or nonexistent for middle-class and upper-class Whites. A significant correlation for Blacks would support the favored environmental explanation of the cumulative deficit, because there is nothing in genetic theory which would predict a systematic difference, on average, between the genotypes of full siblings for any given characteristic, as each sib receives a random set of half of each of its parent’s genes.
I tested this prediction with several hundred pairs of Black siblings and White siblings, all of school age (156). Despite the large samples, the hypothesis was not born out, although the theory-predicted correlation was significant (p
This null outcome made me wonder if the Black population in Berkeley, California, despite its typically lower IQ compared to Whites and Asians in the same schools, was somehow atypical of the general Black population, perhaps being less environmentally disadvantaged. The cumulative deficit might occur only in children whose environmental disadvantage falls below a critical a threshold necessary for normal phenotypic development of the individual’s genotypic potential. I realized, therefore, that another study using exactly the same methods would have to be done in an area where there could be no question that the vast majority of the Black school children lived in a conspicuously impoverished environment.
I found the necessary data for this study in a school district in one of the poorest counties in the rural South. The IQ of the Black pupils was 71, averaged over kindergarten to 12th grade the average IQ of the White pupils was 101. All of the full siblings, White and Black, enrolled in all of the schools of this rural county were included in the study. The findings were startling. The White school population showed no evidence of an age-related decline in IQ, in this respect being like the White sample in my Berkeley study. The Blacks, however, showed a marked age-decrement in IQ, as indicated by the younger minus older sibling IQ difference — a decrement of about one IQ point for each year of the Y-O sibling age difference. In other words, with family background controlled (by the sibling design) these Black children declined, on average, about one IQ point per year throughout their time in school (180). One might have argued that this was not necessarily an environmental effect but a racially genetic difference in the trajectory of the mental growth curves for Blacks and Whites. The California data, however, seemed to rule out this interpretation, as they evinced no such effect for Blacks. If the effect observed in Southern rural Blacks were a genetic racial characteristic rather than an environmental effect, it should have shown up, at least to some degree, in the California Blacks as well. It therefore seems most likely that some substantial part of the IQ deficit for Blacks in the poorest environments is a result of environment, most probably environmental factors that have biological consequences, such as unfavorable prenatal conditions, poor nutrition, and childhood illnesses, which can limit mental development.
While re-reading Spearman’s major work, The Abilities of Man (1927), I came across a brief passage (p. 379) that had not previously captured my attention. On second reading, however, it made a major impact. On the basis of one slight study (by American psychologists) of Black-White differences on a variety of cognitive tests, Spearman conjectured that variation in the magnitude of the B-W difference across various tests is directly related to the size of each test’s g loading. (Because the article presented only the tests’ means but not their intercorrelations, Spearman could not extract the g factor with which to test his conjecture.) I henceforth referred to this conjecture’ as “Spearman’s hypothesis.” It struck me as of quintessential importance, because, if true, it is a much more general hypothesis than my Level I-Level II formulation, which appears to be just a special case of Spearman’s hypothesis. Spearman’s hypothesis also seemed to explain why the size of the B-W difference (in standardized units) varied so widely across different cognitive tests. This question had always been a stumbling block to the prevailing environmental theories, which were a plethora of piecemeal, ad hoc, inconsistent, and unconvincing explanations. Spearman’s hypothesis, if true, would mean that the B-W difference was essentially a difference in g. Therefore, if we are to understand the phenotypic B-W difference in measurements of cognitive ability, it would be necessary to understand the nature of g itself. First, Spearman’s hypothesis had to be put to a rigorous empirical test. This called for representative samples of Blacks and Whites measured on as wide a variety of mental tests as could be found. I tested Spearman’s hypothesis on a large scale (224, 256, 266, 267, 268, 288,290, 296, 324, 325, 339, 375). By publishing my analysis of much of the evidence as a target article in The Behavioral and Brain Sciences (266), its was subjected to commentaries by over thirty experts in psychometrics and cognitive psychology. In brief, the total evidence strongly bears out Spearman’s hypothesis. It is no longer a hypothesis, but must now be regarded as an empirical fact, as much so as Galton’s Law of Filial Regression or Thorndike’ s Law of Effect. (The most recent comprehensive summary of the methodology and evidence on Spearman’s hypothesis is provided in Chapter 11 of my recent book, The g Factor .)
Speed of Information Processing and g [deleted for brevity]
Non-psychometric Correlates of g
Critics of the g construct have argued that g is merely an arbitrary artifact of the way psychometric tests are constructed and inherent in the mathematical procedure of factor analysis. If this were truly the case, I reasoned, the g factor should not be related to variables other than psychometric tests and should tend to disappear when using different factor analytic methods, assuming, of course, that such methods (like varimax rotation) are not specifically devised to scatter the g variance among a number of uncorrelated group factors.
Method Invariance of g. First, I treid to determine whether the g factor of a correlation matrix of diverse psychometric tests is more or less invariant when the g factor is extracted by any of the several quite different methods that have been used by various researchers throughout the history of factor analysis, from Spearman to the present day. By applying each of the main methods of factor analysis to real data and also to artificial data for which the g loadings of the “tests” were known exactly, it was found that g is remarkably similar across all of the different methods, as shown by congruence coefficients averaging over +.99 (360).
Meta-analysis of Physical Correlates. In 1930, long before the invention of meta-analysis, Donald 0. Paterson published his classic work, Physique and Intellect, which reviewed all of the then existing studies on the correlation between physical features and measures of intelligence. Most of the correlational studies were based on rather small samples, and as the correlations between physical measures and IQ are typically small, they were usually nonsignificant statistically. Paterson simply compared the number of significant and nonsignificant correlations and usually concluded that the null hypothesis could not be rejected. His conclusions of essentially “no physical correlates of IQ” have become a dogma in psychology textbooks, and the vast majority of psychologists even today will tell you, for example, that there is no correlation between head size or brain size and IQ. As I generally doubted many of Paterson’s conclusions, I decided to review all of the studies of physical correlates done since 1930 and, when possible, to combine the results of various studies by the methods of meta-analysis. The result was that the null hypothesis (i.e., no correlation) could be rejected at high levels of confidence for most of the physical characteristics that had been examined in relation to IQ; these include body size, head size, brain size, blood types, ocular characteristics, and other anatomical and physiological variables (341). The significant correlations between psychometric scores and such a wide variety of physical traits argues forcefully that the population variance on standard mental tests, such as IQ, reflects latent traits that are profoundly enmeshed with organismic variables in complex ways.
An innovative feature of my meta-analytic review was that, where possible, it examined three different types of correlation of a given physical trait with IQ: (i) the correlation within individuals (WI), (ii) within families (WF), and (iii) between families (BF). This methodology, based on sibling data (202), is an analytically important tool that helps determine the probable cause of the observed correlation. For example, failure to find a WF correlation, even when there is a significant WI correlation in the general population, rules out pleiotropy (i.e., two or more distinct phenotypic characteristics being the result of the same gene). Height and IQ show a WI (and BF) correlation, but they do not show a WF correlation (200). Head size and IQ show a WF correlation (358), as do myopia and IQ (299), suggesting that these two physical traits are pleiotropically related to IQ. These findings aid the search for the specific gene loci responsible for variance in IQ or g and may also provide clues to the precise physical basis of IQ variance.
The Methods of Correlated Vectors. IQ is highly g-loaded, but it is typically saturated with other factors as well. To determine whether a given nonpsychometric variable is related to g per se, rather than to any other factors or specific sources of variance in test scores, I invented the method of correlated vectors. Essentially, it consists of factor analyzing a large psychometric battery of highly diverse subtests to obtain the g loadings of each subtest. This column vector of the subtests’ g loadings is termed Vg. Each subtest is then correlated with some non-psychometric variable, X. The column vector of these correlations is Vx. Controlling for differences in the subtest reliabilities (by disattenuating or partialing out the subtests’ reliability coefficients), a significant correlation between the parallel column vectors Vg and Vx shows that g and X are related. It tells us that the larger a subtest’s true g loading, the larger is its correlation with variable X.
I have examined a number of variables (X) by this method. The correlation between Vg and the various measures (i.e., Vx) is shown in brackets: in brackets): spouse correlation [.90]; heritability [.60-.80]; inbreeding depression [.80]; cerebral glucose metabolic rate [-.79]; brain intracellular pH [.63]; head size [.60-.70]; choice and discrimination RI [.70-.80]; average evoked potential habituation amplitude [.80]; and AEP waveform complexity [.95]. All of these correlations are significant; the particulars on the studies of each variable are given in (226, 258,282, 356, 383 [Chapters 6-9]). No other factor shows anywhere near the same degree of relationship to non-psychometric variables as does g. My research shows conclusively that psychometric g, far more than any other factor, reflects individual differences in certain biological and developmental properties of the brain that govern its speed, consistency, and capacity for information processing. Though manifested overtly in many ways that can be described in behavioral terms, g itself cannot be described or explained in behavioral or psychological terms. The g factor per se does not reflect any particular achievements, knowledge, or skills, but rather the information processing capacity for acquiring and using the knowledge and skills necessary for achievement.
I have pointed out a crucial conceptual distinction, namely, that the construct (in this case g) and the vehicle used for measuring the construct (in this case, a psychometric test and the scores it yields) are not one and the same; they are conceptually and empirically distinct. Though the rank order of individuals’ scores on any highly g-loaded test can be accounted for largely in terms of individual differences in the level of g, the absolute level of the individuals’ raw scores on any such test also reflects the particular composition of the test items (332). This fact has important consequences for the interpretation of test scores and the secular trend in the population mean for any particular vehicle of g (319, 368, 383, Chapter 10).
In addition to showing that g is correlated with various biological variables, I have also amassed empirical evidence (based largely on the method of correlated vectors) to show that it is g itself that accounts for most of the practical predictive validity of tests used in educational placement and selection and in personnel selection in industry and the Armed Forces. When the validity coefficient is based on a multiple correlation, typically the increment in predictive power contributed by all other factors (and by test specificity) independent of g is remarkably small (383, Chapter 9).
I see basic research on human mental ability, particularly g and the major group factors, as advancing in each of two directions, which I think of as the horizontal and the vertical. Both are necessary and each can be scientifically rigorous.
Horizontal research on g explores the whole nexus of behavioral, social, and economic correlates and consequences of individual and group differences in the level of g. I believe g plays a greater role in these spheres than most educators, sociologists, criminologists, economists and social policy analysts presently realize. But serious consideration of this probability seems to be strongly resisted in some circles. The generally nihilistic reaction of the mass media to The Bell Curve by Herrnstein and Murray (1994), which, more than any previous scholarly work, examined the relationship between g and a number of social variables of national concern, is a case in point. It is a reasonable supposition that in the global economy and the competitive technological and information-intensive world of the twenty-first century, a nation’s chief natural resource will be its population’s overall level of educability, in which the distribution of g-loaded abilities will inevitably be a major determining factor. Researching the extent and the ramifications of the g factor in the nexus of societal variables is the province of a budding new field named the “sociology of intelligence” by sociologists Robert Gordon and Linda Gottfredson. (See the a special issue of INTELLIGENCE, 1997,24, No. 1, for example.)
Vertical research digs down in search of the causal basis of g. Being analytical and reductionistic by nature, I personally find this is the more interesting aspect and the one to which I expect to devote my efforts in the future. By definition, an underdeveloped field is one in which many of the findings, and even some of the classic experiments, have not been replicated and many of the supposed facts not fully consolidated. I am happy therefore to see research on “mental speed” in relation to psychometric g being actively pursued in a number of laboratories around the world. The new information that I see coming in, almost every month, is most valuable. However, there is still confusion, contradiction, and many unanswered questions. More standardized apparatus and procedures are called for (as the same standardized reagents are used in every chemical laboratory) and much more importance must be accorded to replicating the theoretically crucial findings across different laboratories. The measurement of individual differences in g by means of mental chronometry is as close to the interface between brain and behavior as we are likely to come. The consolidation of the knowledge gained at this interface is important for vertical advancement, that is, identifying the basis of g in the structural and functional features of the brain itself.
A few hypothesis-generating steps have already been taken by showing g-correlates of direct brain measurements obtained with evoked potentials, magnetic resonance imagining (MRI), positron emission tomography (PET scan), and intracellular brain pH levels. Many researchers in the neurosciences, however, have ignored g as a subject for brain research, mostly, I fear, because a crucial distinction has not been made sufficiently clear.
In recent articles (374, 384), 1 have emphasized the distinction between (i) the neural circuitry or design features of the brain possessed by all neurologically intact members of a species that are responsible for that species’ characteristic behavioral capacities, and (ii) the properties of the brain of a given species that cause intraspecies variation (i.e., individual differences) in that species’ characteristic behavioral capacities. We know from research in behavior genetics that intraspecies variation in many behavioral capacities is not entirely the result of experiential differences and learning. But I have found no good reason to believe that the design features of the brain (which are undoubtedly crucial determinants of interspecies variation in behavioral capacities) are necessarily involved in intraspecies variation. The latter may well be due to an entirely different set of causes than neural circuitry or other designs features of the brain, but rather involve such within-species factors as: differences in blood supply (via the richness of the capillary network), the degree of myelination of axons (which affects nerve conduction velocity), the neuroglial cells (which nourish neurons), the brain chemistry of neurotransmitter (which affect synaptic transmission), and individual differences in the number of neurons involved in the various brain modules. We now know quite conclusively from MRI studies, for example, that IQ is correlated with brain size, but we still don’t know what precisely it is about brain size that causes this correlation.
All but an exceedingly few neuroscientists today are interested in intraspecies variation in behavioral capacities. They may well find discovering the brain’s general operating principles daunting enough, without having to explore the causes of individual variation in the functional efficiency of the essential design features of the brain and their general operating principles. As Francis Crick (1994) has pointed out, neuroscience cannot yet explain even how the human brain sees things in the environment, much less how it performs the complex functions we call intelligence. The very existence of the g factor (like all other psychometric factors) is only revealed by examining intraspecies individual differences in each of the broad and diverse class of functions we regard as constituting intelligence — discrimination, generalization, learning, memory, insight, abstraction, problem solving, and the like. But before we can begin to research the physical basis of g, do we first need to discover all of the brain’s design features that make these functions possible? I don’t think so. The question of what causes the various cognitive functions of the brain to be positively correlated is a very different question from that of understanding the specific operating mechanisms of each of these functions.
As a heuristic proposition to encourage research in this “vertical” search for the neurological causes of g, I propose the following working hypothesis: Individual differences in human behavioral capacities do not result from differences in the brain’s structural operating mechanisms per se, but rather are the result of other aspects of cerebral physiology that modify the sensitivity, efficiency, and effectiveness of the basic information processes that mediate the individual’s responses to certain aspects of the environment.
I’m placing my bets on the search for those aspects of brain physiology responsible for g as most likely to generate the next path-breaking discoveries in differential psychology and human biology. I have been told by experts that the technology to do this already exists. So, looking ahead, I see my principal endeavor to be sparking the interest of qualified scientists in the brain sciences and helping them solicit the necessary resources to pursue this “vertical” investigation of g.
A New Twist on Jensenism by ALAN S. KAUFMAN, Yale University School of Medicine
Jensenism is redefined, not in terms of his notions about the genetics of intelligence, but in terms of the personal qualities and beliefs that have made Jensen a researcher of note: going against convention. tackling controversial topics with empiricism, refusing to be intimidated by threats and picket lines, and being flexible enough to modify his beliefs. These aspects of Jensen are illustrated by the author’s personal experiences with the man and his research.
My own, personal dictionary, not to be found in any bookstore or publishing house, has a different definition of Jensenism, one that more accurately portrays the man and his work:
Jensenism, n. (I) the art of going against the grain of conventional psychological wisdom; (2) the belief that no topic is too holy or taboo to reside beyond the grasp of scientific, empirical inquiry; (3) the ability to remain steadfast in one’s beliefs–sometimes with a touch of arrogance–despite threats, accusations, denunciations, and attacks; (4) the flexibility to allow one’s own strictly held beliefs to be overturned by new empirical discoveries; after Arthur R. Jensen (born 1923), U.S. educational psychologist, who has practiced each of these tenets during his impressive research career [1955-present].
I have long been an admirer of Arthur Jensen’s approach to research. I disagree with some of his conclusions, especially those concerning genetics and race differences, but I applaud his commitment to the scientific process. And, quite simply, the man is brilliant.
I remember when I first came face-to-face with Jensen’s brilliance. The Journal of Special Education had organized a special issue in 1984 devoted to the controversial test that I co-authored with my wife Nadeen, the Kaufman Assessment Battery for Children (K-ABC; Kaufman & Kaufman, 1983). Many luminaries in psychology were invited to write articles about some aspect of the K-ABC, and I was asked to read through all of the articles and write a rebuttal article. The package of 13 articles arrived at my home just before I was to drive to the airport for a cross-country trip. I took the package with me and spent the next five hours reading each article and feverishly taking notes for my rebuttal. Though the group of contributors included Anne Anastasi, J.P. Das, and Robert Sternberg, among others, I had a good feeling as I read through the first dozen articles. I was not worried about rebutting the key points made by the various authors; I was feeling confident, even a bit cocky.
Then I got to Jensen’s article, which was at the bottom of the pile. As I began to read his criticisms of the K-ABC, I began to sweat. One line from a movie kept weaving in and out of my consciousness: As the Sundance Kid (Robert Redford) and Butch Cassidy were being pursued by some relentless unknown enemies, Sundance asked repeatedly, “Who are these guys?” As I read page after page of Jensen’s insightful critique, involving cognitive complexity, Spearman’s hypothesis, and indifference indicators, I kept subvocalizing the words, “Who is this guy?” Of course, I knew quite well who he was; it just never occurred to me that he was so familiar with my work and that he would start his attack with smoking guns. The other authors wrote articles filled with text, opinion, mid sometimes emotion. Jensen buttressed his text with original data analyses that occupied four new tables and six new figures. He used these analyses to challenge and provoke, to some extent, but mostly to inquire, to seek the truth.
He used the K-ABC subtests to assess the validity of Spearman’s hypothesis (i. e., the notion that the magnitude of black-white differences in a set of tasks is positively correlated with the tasks’ g loadings). He obtained a correlation of .58 for the 13 K-ABC sub-tests, virtually identical to the value of .59 for a larger group of 121 cognitive tasks (Jensen, 1984), impelling him to conclude that “the K-ABC tests cannot be regarded as at all atypical; they conform to Spearman’s hypothesis at least as well as many other tests” (p. 395). That finding supported his overall perspective about the K-ABC, and he might have left it at that. But he proceeded to point out a finding that was opposite to his arguments: “The regression line for the K-ABC tests. . .falls significantly below the regression line. . .for all 121 tests. . . .That is, the K-ABC tests show considerably smaller differences than would be predicted from their g loadings. This phenomenon poses what may be the major puzzle of the K-ABC” (Jensen, 1984, p. 395). And, indeed, Jensen enjoys solving puzzles. He delighted in formulating thought-provoking hypothesis after hypothesis to attempt to solve this puzzle. I disagreed with most of his ruminations, on Spearman’s and other hypotheses, citing data or facts that I perceived to be contrary to his notions (Kaufman, 1984). Yet, I couldn’t help but tip my hat to his objectivity and insight: “The tactics for writing the articles. . . vary quite a bit, ranging from the brilliant, data-based, meticulous critique of certain key aspects of the K-ABC expounded by the noted Arthur Jensen [to the emotional responses of some others]” (Kaufman, 1984, p. 410).
About a year after being so thoroughly impressed by Jensen’s empiricism, wisdom, and sense of fairness, I had the chance to see him in action, and get to know him personally, at the 1984 American Psychological Association meeting in Anaheim. Jensen was giving a paper on several topics that included Spearman’s hypothesis, his K-ABC research, and black-white differences in IQ and achievement. The media, naturally, was not far behind, and a circus atmosphere developed, with reporters, picketers, video cameras, and security guards everywhere. Jensen was escorted into the large, packed room by the guards and he showed not the slightest trace of intimidation. He began his address with blunt remarks about the large achievement discrepancies between blacks and whites, differences that were not explainable by simple environmental variables. I was taken aback by the directness of his statements and his interpretations on a topic that invariably invites hemming, hawing, apologetic statements, and back-stepping. I guess I shouldn’t have been so surprised in view of his writings on the topic, but I had never heard him speak before and was expecting a small dose of political correctness. Instead, I witnessed something more akin to sublime self-confidence blended with in-your-face arrogance.
That evening, Nadeen and I joined Jensen, Cecil Reynolds, and one or two other psychologists for dinner. Our Son, James, then 10 years old and now a Ph. D. student of Robert Sternberg’s at Yale, had been watching TV that afternoon and saw the demonstrations against Jensen the psychologist and threats against Jensen the man. James, out of fear, pleaded for us not to join Jensen for dinner, but settled for a solemn promise that we wouldn’t sit too close to him. Dinner was thoroughly enjoyable as Jensen, though a bit uncomfortable with the spotlight, entertained everyone with tales of harassment and intrigue stemming from his notoriety. I recall him telling of a TV talk show host’s duplicity; the host (I believe it was Mike Wallace) made Jensen look foolish when he changed the questions that actually aired from the actual questions that were asked during the taped interview.
I have always enjoyed Jensen’s ability to revisit the old in psychology and come away with something new. His research on reaction time is excellent and thought provoking. A simple twist on Galton’s initial tasks for measuring intelligence, long relegated to historical status within the field of psychology and long believed to be an “off target” attempt to measure cognitive ability, and Jensen was able to produce an apparently valid measure of IQ. A sophisticated empirical treatment of the g factor, likewise dismissed by mainstream psychology as a concept that is primarily of historical importance, and Jensen re-established g as an invariant construct of potential value. One might disagree with the meaningfulness of g and dispute the theoretical basis or practical utility of the g factor, but the quality of Jensen’s research on the general factor demands that his findings and conclusions not be taken lightly or dismissed cavalierly.
The net result of his research and writing on these formerly historical topics is to reinterpret history. Galton’s original intelligence test focused on sensori-motor abilities and, though powerfully influential in the emergence of the IQ concept in the United States and Europe, Galton’s test was ultimately declared invalid and was replaced by Binet’s IQ test. Binet had the insight to allow errors of measurement to invade the science of intelligence and relied conceptually on Spearman’s g factor in providing a rationale for his choice of cognitive tasks for his battery. Subsequently, the g approach was replaced by Wechsler’s multi-score instruments, and theories from Thurstone to Guilford to Horn that greatly downplay g; even the latest revision of the Stanford-Binet yields subtest and area scores and relies on a blend of Thurstone and Horn-Cattell as a theoretical foundation. But the careful research by Jensen on reaction time and the g factor forces historians and those involved in the clinical assessment of intelligence to rethink the contributions of Galton, Spearman, and Binet, and to realize that the early instruments may have been more on-target than initially believed. Galton, in particular, may have missed by a few inches instead of a few miles. Intriguingly, Jensen’s research on reaction time and g is not only tied to the past; it also may provide a link to the future as assessment enters the computer-based stage of examining EEGs and CT scans for, potentially, increasingly objective measurement of intelligence (Eysenck & Barrett, 1985; Jensen, 1985).
As a trainer of school and clinical psychologists in intellectual assessment for a quarter-century, I have found some of Jensen’s research and theorizing quite valuable. The research that impressed me for its simplicity, yet far-reaching implications, was his work on Wechsler’s Digit Span (Jensen & Figueroa, 1975). So much had been written on the potential impact on intelligence test scores of environmental variables such as motivation and perseverance and of cultural variables such as the relevance of the stimuli used in test questions. Evaluating the importance of these variables is difficult because of their multifaceted complexity. Yet, Jensen’s Digit Span research was ingenious because it effectively held these variables constant. Why would an examinee be more or less motivated or persevering when responding to Digits Forward versus Digits Backward, tasks that utilize the identical, culturally neutral stimuli? The research results that revealed quite different individual variation on the repetition of digits in the forward versus backward sequence were provocative and could not be easily dismissed by proponents of the key role of motivation or culture loading in accounting for group differences.
Jensen’s hierarchical Level I–Level II theory of intelligence, simplistic as it is, provides clinicians with a valuable method for interpreting profile fluctuations when more conventional explanations (such as verbal–nonverbal) do not suffice. When interpreting intelligence test profiles, I have always considered it to be a mark of “intelligent testing” to be able to apply diverse theories as explanations for a child’s or adult’s subtest fluctuations (e. g., Kaufman, 1990, 1994). For the purpose of profile interpretation, complex theories are usually less practical or effective than simple theories for grouping subtests into alternative, relevant patterns that may reveal a person’s cognitive strengths and weaknesses. Jensen’s memory–reasoning dichotomy has proved to offer a useful alternative interpretive strategy for understanding a person’s profile when the fluctuations in that profile suggest that the test’s global scales cannot be meaningfully interpreted. For example, memory-reasoning sometimes fits the data better than Wechsler’s Verbal-Performance split, the K-ABC’s Sequential-Simultaneous division, or the Stanford-Binet IV’s fluid-crystallized dichotomy. In view of the fact that of the alternative theories mentioned, only Jensen’s is hierarchical, the application of Level I-Level II theory to profile interpretation adds an additional dimension to the mix, a dimension that sometimes has implications for educational applications.
My own research disputes some aspects of Jensen’s most controversial statements regarding the intellectual abilities of blacks versus whites. On the K-ABC, for example, one of the Achievement subtests is Faces & Places, a test of general information that uses a visual-vocal instead of an auditory-vocal format; thus, instead of responding orally to a question such as “Who is Martin Luther King, Jr.,” the child must respond orally to a picture of Dr. King. This K-ABC subtest assesses range of general knowledge, as does Wechsler’s Information subtest. Yet, unpredictably, the apparently culture-loaded Faces & Places subtest produces trivial black-white differences whereas Wechsler’s information subtest yields among the largest racial differences observed on conventional intelligence tests (Kaufman, 1994; Kaufman & Kaufman, 1983). My interpretation is that the legacy of large racial differences has maintained when the tests have been the same old tests used since the time of Binet and World War I. When a new task is tried, even when it is an apparent shift of a kaleidoscope such as Faces & Places relative to Information, then the racial differences may disappear. In fact, the black-white difference was also small (as was the Hispanic-white difference) for adolescents and young adults on an adult analog of the general information task called Famous Faces (Kaufman, McLean, & Kaufman, 1995). Furthermore, a new fluid reasoning task called Four-Letter Words, clearly a Level II task from Jensen’s system, also produced much smaller than predicted race differences for a large sample of adolescents and adults (Kaufman, Chen, & Kaufman, 1995). These findings reinforce the notion that the so-called “constant difference” of about one standard deviation between test scores of whites and blacks may be largely a function of the limited selection of traditional tasks that defined virtually all tests of intelligence from the past. Race differences on the new breed of intelligence tests that has proliferated in the past two decades, many of them theory-based, may not conform so closely to the findings of tasks from the Binet-Wechsler tradition.
Yet, despite my disagreements with some aspects of Jensen’s research and writing, I remain steadfast in my admiration for his stubborn insistence that no topic is too holy to be scrutinized by empirical analysis; that no interpretation of data is too politically incorrect to permit a straightforward expression of one’s scientific opinion; that threats and intimidation are not effective methods for thwarting creativity and expression; that some of the best inspirations for research can be found in the historical annals of psychology among discarded and disavowed ideas; and that one should be ready and willing to abandon strictly held beliefs if new, compelling data should come along to suggest that the old ideas may be wrong. To me, Jensen is the quintessential scientist.
Jensen and Intelligence by NATHAN BRODY, Wesleyan University
Jensen’s Contributions to the study of intelligence are discussed. The paper considers his writing on the topic of racial differences in scores on tests of intelligence. The paper concludes with a discussion of his research on the correlates of the g vector.
JENSEN AND RACE
“. . . it is a peculiar sensation this double consciousness, this sense of seeing oneself through the eyes of others, of viewing one’s soul by the rape of a world that looks on in amused contempt and pity. One ever feels his twoness — an American, a Negro:…”
W.E.B. DuBois, Souls of Black Folks
This quotation, taken from a book published in 1903, is a doubly apt introduction to Jensen’s work on race and intelligence, it reminds us of the effects of beliefs about racial differences on African-Americans who experience “the sense of seeing oneself through the eyes of others. . .by the tape of a world that looks on in amused contempt…” It is also descriptive of the perception of Jensen’s work by many hi the field of Psychology. Just as those he writes about are forced to see themselves through his lens, others see him through the lens of someone whose views about racial differences they may abhor or reject.
Jensen’s contributions to an understanding of individual differences in intelligence extend far beyond a discussion of racial differences, but it is his work on race that often serves to define his contributions. Therefore, a discussion of his work on race is an apt beginning to an evaluation of his overall contributions. Jensen (1974; 1977) published one of the best studies demonstrating that extremely poor schooling could result in a cumulative deficit in the intellectual functioning of African-Americans. He used a sibling control design to demonstrate that African-American children attending schools in the segregated south in the 1950s exhibited a cumulative decline in intelligence relative to the intelligence of their younger siblings. He also found that this effect was not present for African-American children attending schools in Berkeley, California. These studies are illustrative of Jensen’s imaginative ability to obtain data that address a critical issue. Jensen’s results are buttressed by an analysis of the consequences of deprivation of formal education associated with the decision of the Prince Edward County School Board in Virginia to avoid compliance with a court ordered desegregation plan (Green, Hoffman, Morse, Hayes, & Morgan, 1964). Green et al. found that African-American children who were deprived of the opportunity to attend public schools exhibited declines in intelligence of approximately six points per year for each year of deprivation of formal schooling. Jensen’s results and the results of the Green et al. analysis are probably the two most convincing studies in the literature indicating that educational influences can reduce the intellectual functioning of African-Americans.
Robert Sternberg once wrote that he did not know why Jensen used his formidable psychometric knowledge and talent to address this particular issues (Sternberg, 1985). The choice of any of our research topics is mysterious and not illuminated by somewhat simplistic and reductionist analyses of political motives. I rather think, perhaps wrongly, that my interest in the field of intelligence derives in part from a personal and moral imperative I feel that compels me to differ with Jensen with respect to his views on race and intelligence. Nevertheless, I believe that anyone who wishes to write about the issue of race and intelligence must acknowledge Jensen’s formidable contributions to this topic and his comprehensive knowledge of this area of research. Jensen’s book on bias in testing is an extraordinarily thorough and scholarly analysis of the issue of test bias (Jensen, 1980). I like to compare this book with another book that I admire greatly, Paul Meehl’s monograph on Statistical vs. Clinical Prediction (Meehl, 1954). Both books serve to define the principal issues that must be understood in addressing the topics that they consider. Both books develop their arguments with unusual clarity and sophistication. And, to a remarkable extent, the conclusions reached in both books have stood the test of time and become part of the canon of empirically established generalizations that define our knowledge of important topics. Jensen established what is now close to the received wisdom of knowledgeable students of intelligence — tests of intelligence are equally valid indices of the performance of individuals who differ with respect to their racial identification. In several technical senses of the term, they are not biased — a conclusion endorsed in the recently published report of the American Psychological Association’s task-force on intelligence composed of individuals with diverse views of the field (Neisser et al., 1996).
I do not agree with Jensen’s argument, developed in great detail in his forthcoming book on g, that genetic differences contribute to differences in performance on tests of intelligence between African-American and other racially identified groups (Jensen, 1998). I believe that his argument in favor of a genetic hypothesis is not well grounded and I hope to publish an analysis of my reasons for not accepting his arguments. It is easy for those who know little about Jensen’s views or the detailed analysis of research he presents in support of his views to dismiss his arguments out of hand. It is hard to dismiss his arguments (but I believe possible to do so) if one reads him carefully and is informed about the literature. I believe that the reasons for group differences in scores on tests of intelligence can not be ascertained from the available data. Whether a determination of the reasons for the group differences in scores would be theoretically or socially useful is hard to know — it may depend in part on the reasons for the difference and what we can do to remediate the difference or to minimize its impact. And, whatever our differences may be about this issue, there is at least one belief about race and intelligence that we all share — within group racial differences are larger than between group differences.
Race does not define a person’s score on a test of intelligence (or, for that matter, any other characteristic other than race). I remember reviewing a paper by Jensen dealing with an analysis of the relationship between head circumference and the g vector that included data derived from two different racial groups (Jensen, 1994). In my review I noted that his discussion of his findings was not well-supported by his analyses and I suggested that he needed to rewrite his discussion to present a somewhat more cautious and weaker conclusion than he had presented. Jensen, on this occasion, agreed with me, and wrote a very generous letter to the editor of the Journal thanking me for my suggested recommendations and changed the article to reflect my criticisms. I think that this episode is illuminating. Jensen is not an ideologue or a person who is not able to respond to criticism in a fair way. He is a scientist with formidable technical skills who strives for an understanding of the topics that he addresses. In this regard, his work is a model of scientific decorum. We should all strive to emulate his ability to test our beliefs against a recalcitrant reality that often is resistant to our ability to represent it in distorted ways. In the long run, if we are clever and honest, it will impose its structure and truth on us rather than ours on it.
In my opinion, Jensen’s most important contribution to the field is contained in his new book on the g factor (Jensen, 1998). In the first paper dealing with g, Spearman attempted to determine the g loadings of different measures of intelligence (Spearman, 1904). For much of this century, it has been understood that tests differed in their g loadings and there was a consensus about the kinds of tests that had the highest g loadings. Carroll’s comprehensive re-analysis of the canon of correlation matrices derived from diverse measures of intelligence provides ample support for the proposition that tests with high loadings on gf have higher g loadings than other tests (Carroll, 1993). So, too, Marshalak, Lohman, and Snow’s multidimensional scaling analysis of tests of ability demonstrates that tests with high loadings on gf such as the Ravens have higher g loadings than other tests (Marshalek, Lohman, & Snow, 1983). An examination of the contents and intellectual processes required for correct solution of tests that have high g loadings provides a basis for speculations about the nature of g.
Jensen (1998) has taken the analysis of g beyond the realm of metaphorical speculation. He derives g loading values for test batteries and then uses the vector of g loadings as a parametric index that can be related to other measures. These analyses provide a nomological network of laws and relations surrounding g that serves to specify the theoretical meaning of g construed as a hypothetical construct that is a variable component of different measures of intelligence.
Jensen (1998) links the g vector to several biologically relevant vectors. He notes that Pedersen et al. (1992) obtained heritability values for different tests in a battery of tests of intelligence administered to a sample of older Swedish adult MZ and DZ twins reared together and apart. The vector defining the heritability of the tests is correlated with the vector defining the independently ascertained g loadings, r = .77. Jensen provides additional evidence based on Wechsler sub-test g loadings indicating that the vector of g loadings is correlated with the vector of heritability values for Wechsler sub-tests.
Jensen reports other results indicating that the g vector is linked to biological indices. He analyzed data on head size and intelligence and obtained a vector for different tests of intelligence that represented the correlations between measures of head size and scores on different tests of intelligence (Jensen, 1994). This vector was correlated with the g loading vector. Head size is an imperfect index of brain size and the relationship between head size and intelligence indicates that intelligence is related to brain size. This establishes that the g vector is linked to a biological index of intelligence.
Jensen (in press) reanalyzed the data obtained from a French adoption study reported by Capron and Duyme (1989). This study used a complete cross-fostering design to study the effects of variations in social class background of biological and adopted parents on the IQ of adopted children. Previous analyses of these data indicated that children’s IQ was influenced in an additive manner by the social class background of both adoptive and biological parents. The latter influence was found to be stronger than the former. Jensen obtained a vector defining the magnitude of the difference in Wechsler sub-test scores for adopted children reared in high and low social class families. He also obtained another vector defining the difference in sub-test scores of the Wechsler test for adopted children whose biological parents differed in social class background. This latter vector correlated with the vector defining g loadings for the sub-test scores, r = .53. The comparable correlation between the g loading vector and the vector of differences in sub-test scores defined by the social class background of an adopted child’s adopted parents was .01. These data indicate that the nature of the influence of biological and adopted parents on an adopted child’s IQ is different. The former influence varies with the g loading of the test and the latter influence does not, apparently influencing components of variance in an IQ test that are unrelated to g. This highly original analysis adds to the evidence suggesting that the g vector is a biologically influenced component of the variance in diverse measures of intellect and this analysis provides evidence that the nature of the influence on IQ of biological and adopted parents is both qualitatively and quantitatively distinct.
Jensen’s analyses of the g vector also include studies relating the vector to vectors defining the predictive validities of sub-test scores on the Wechsler tests for measures of academic performance. He obtained correlations between g vector scores and the vectors of correlations between Wechsler sub-test scores and high school student’s rank in class and college student’s grade point average. The correlation with the g vector for the high school sample was .53 and the comparable correlation for the vector derived from the college student sample was .83. These analyses indicate that the predictive validity of a test of intelligence for a measure of academic success is related to the g loadings of the test.
Jensen’s analyses of the correlates of g vectors provide the quantitative underpinning for what has long been apparent — g is a biologically influenced heritable component of the commonality among diverse measures of intellect that is related to the ability of individuals to acquire knowledge in formal academic contexts. Perhaps we have always known this, but following Jensen’s highly original use of analyses of the correlates of g vectors we know this with a kind of quantitative precision not heretofore available.
Jensen’s work on the correlates of the g vector reveals some of his best attributes — an ingenious ability to develop quantitative analyses that address fundamental issues in highly original ways that advance our knowledge of critical issues in the field.