Download Individual Differences In Mental Ability and more Lecture notes Psychology in PDF only on Docsity!
CHAPTER 4
Individual Differences In Mental
Ability
Arthur R. Jensen
One of the most obvious "facts of life" to all teachers, at every level of education, is the phe- nomenon of individual differences-in mental abilities, special talents, and traits of personality. Especially salient are those characteristics of pupils that are the most clearly related to the suc- cess of teachers' efforts to impart knowledge and intellectual skills. The most prominent of such characteristics is general mental ability, or intel- ligence. Consequently, the study of individual dif- ferences, especially differences in intelligence, has been one of four major themes of educational psy- chology (along with development, learning, and measurement) ever since this field was formally recognized as a branch of psychology. The ideal of universal education, which first gained impetus and implementation in America, literally forced educators' practical and humane concern with the problem of making formal schooling a successful and rewarding experience for the whole school-age popUlation, which ranges widely in mental abilities and other characteristics that are importantly relat- ed to scholastic performance. This chapter centers its focus on the history of attempts to understand only one of these differen- tial variables-intelligence. The concept of intel-
Arthur R. Jensen • School of Education, University of California, Berkeley, CA 94720.
ligence has a longer, more complex, and much more controversial history than is found for any other theme within the whole purview of educa- tional psychology. The history of the concept of intelligence therefore merits a whole chapter in its own right. Indeed, a large book could well be de- voted to the topic. Other dimensions of individual differences are relatively latecomers to educational research, and their importance, in terms of their relative contribution to variance in scholastic per- formance, is minor in comparison with the role of individual differences in intelligence. Moreover, the basic concepts and methodology of measure- ment and research developed in connection with the study of intelligence have considerable gener- ality, because they have been applied as well to the investigation of other educationally relevant traits, particularly in the domains of personality and motivation. Few psychological phenomena, however, are as highly relevant to education as individual dif- ferences in mental ability. Probably because of the practical consequences of individual differences for scholastic performance and all of its occupa- tional, economic, and social correlates, this sub- ject has had perhaps the most tumultuously contro- versial history of any topic in psychology and education. There is really no argument about the promi- nence or importance of the topic itself. The argu-
62 PART II • THE DEVELOPMENT OF EDUCATIONAL PSYCHOLOGY
ments today involve quite different issues. In the past half-century, millions of school children in America and the Western world have been given tests called "IQ" tests, "intelligence" tests, "scholastic aptitude" tests, and the like. (Quota- tion marks seem advisable early in this discussion, first, to indicate loosely defined popular terms, as contrasted with precisely defined technical terms, and second, to warn against the risk of improper reification of terms that represent abstract COR- cepts.) Whatever these tests "measure," which we will let remain an open question for the time being, two things are now definitely known beyond dispute: (a) The majority of such tests (labeled "IQ," "in- telligence," or "general aptitude") all measure pretty much the same source of variance as indi- cated by high correlations among scores on such tests; correlations typically fall in the range of. to .90, averaging close to .80. (b) No other single item of information that we can obtain about chil- dren is as highly correlated with assessments of scholastic performance as the children's scores on these tests. No other kind of information concern- ing children's background is as highly predic- tive-not the socioeconomic status of the chil- dren's parents, or the parents' education, or occupation, or race, or the national origin of chil- dren's ancestry, or their gender. Children's scores on "IQ" tests account fOi" more of the total vari- ance (i.e., individual differences) in overall scho- lastic achievement than all of these background variables combined, independent of IQ. This ap- pears to be true in every country, for every type of educational system, and for every method of in- struction yet devised. No attempts, by means of varied instructional techniques, to completely overcome the correlation between individual dif- ferences in scholastic performance and scores on "IQ" tests, when a fully representative sample of the school-age population is considered, have come anywhere near success. The fact that this is so is scarcely disputed today. But why this is so and what it means have long been, and still are, questions of intensive inquiry and heated debate in educational psychology. Through it all, the use of numerous tests of "mental abilities" has become widely entrenched in education, in connection with "streaming" or "tracking" pupils, for placement in special class- es, for individual diagnosis of learning problems, for vocational counseling, and for selection for higher education. The lives of countless persons
have undoubtedly been affected to some degree by "mental tests." Just what do such tests actually "measure" that would seem to justify such wide- spread use? Similar tests, often called "aptitude" tests, are also now commonly used with adults outside the school setting, in screening job applicants for in- dustry, and for selection and allocation to various training programs in the armed services. The tests commonly used for these purposes have also been shown, through correlation analysis, to measure much the same individual differences as are mea- sured by the IQ tests administered in schools. There is no longer any question that such tests possess some practical validity for predicting job performance and success in training programs. The persisting question is why such tests have predic- tive validity for so many practical, real-life cri- teria. The fundamental question implied here, we know, has existed long before mental tests were ever invented. Stated bluntly in laymen's terms, it is the simple question: Why are some people smarter than others? Many other questions that need to be answered naturally spring from this sin- gle question, which many persons have viewed as the Pandora's box of psychology. Yet for more than a century, it has remained, and continues to remain, a central question in that branch of psy- chology now known as differential psychology, or the scientific study of individual and group dif- ferences in psychological traits. The current status of research, theories, and controversies on this top- ic is highly complex and perhaps even perplexing to newcomers to this field. It seems likely that the present scene can be more clearly understood when viewed in historical perspective. The history of thought about the nature of individual differences in human abilities should essentially enlighten the question, how did we arrive at our present state of knowledge and theory on this topic? A historical overview might also suggest the most promising avenues for future research. It is the writer's belief that the modern era of research in this field has been evincing lively progress toward addressing, with advanced statistical and laboratory methods, a number of the key questions that have come down from the past. It seems unlikely that a historical survey of the thinkers and their theories and re- searches that have led up to the present state of the field could justify the wiseacre's definition of his- tory as "a chronology of events that never should have happened."
64 PART II • THE DEVELOPMENT OF EDUCATIONAL PSYCHOLOGY
types. Plato likened these types, in terms of their rarity, to gold, silver, and brass, and held that the ideal society would assign people to occupations on the basis of this classification. The three main divisions were first the philosophers, who would govern; then the warriors; and lastly, the artisans. But the basis for such a classification of people was not made clear, nor were the means of achiev- ing it. But Plato's idea is probably the first major expression of opinion regarding the recognition of individual differences as being related to a soci- ety,s general welfare. Plato is also credited as the first thinker to sug- gest a hierarchical structure of mental functions- an idea that comes down to the present day. He regarded reason, or intellect, as the highest aspect of the soul, which ideally dominated the lower functions of emotion and drive. In Phaedrus, he depicts intellect as the charioteer who holds the reins, the emotion and drive are likened to the team of horses that draw the vehicle. The charioteer is the cybernetic element, the horses the dynamic ele- ment. Here already we can see some of the basic ingredients of modem psychology. Plato's illustrious student, Aristotle (384- B.C.), was really the first formal psychologist, in that he wrote the first books on the subject, De Anima, De Sensu et Sensili, De Memoria et Remi- niscentia, and On Psyche. Aristotle clearly dis- tinguished various psychological functions, such as sensation, reaction, desire, memory (recogni- tion and recall), knowing, and thinking. Unlike Plato, Aristotle recognized thinking as directly de- pendent upon what he regarded as the lower pro- cesses of sensation and memory. Thought was viewed as deliberation preceding action. Aristotle might also be regarded as the first cognitive theo- rist. He constrasted actual activity with the hypo- thetical capacity or mental activity on which it de- pends; this is the first introduction of the concept of ability as a latent trait, distinct from its behav- ioral expression. Aristotle reduced Plato's threefold classification of the soul to only two broad divisions, which he termed dianoetic (cognitive functions) and oerectic (emotional and moral functions). It was the Roman author, orator, and statesman, Cicero (106- B.C.), who, in translating Aristotle's Greek termi- nology, coined the almost exact equivalent of "di- anoetic" in Latin as intelligentia-hence the ori- gin of the word intelligence. But neither Aristotle nor other ancient Greek philosophers said anything about individual differences in the various psycho-
logical qualities that they propounded. Besides the fact that these qualities were thought of largely as qualities of the soul and hence were exempt from human frailty, the social systems of the ancient and medieval world, consisting of aristocracies and serfdoms, probably afforded little scope for the salience of individual differences in abilities. A person's occupation and station in society were determined by the circumstances of his birth. For- mal education was the privilege of only an elect few, and the great inequality of opportunities for education and vocational choice could largely ob- scure the perception of human differences as repre- senting characteristics that are intrinsic to individ- uals. Indeed, the first clear statement concerning indi- vidual differences in mental abilities came some years following the heyday of Greek philosophy, from the Roman philosopher, Quintillian (A.D. 35-95), who might well be called the first real educational psychologist. He wrote the following advice to teachers, which would not look out of place in a modem textbook of educational psychology.
It is generally, and not without reason, regarded as an excel- lent quality in a master to observe accurately differences of ability in those whom he has undertaken to instruct, and to ascertain in what direction the nature of each particularly in- clines him; for there is in talent an incredible variety, and the forms of mind are not less varied than those of bodies. (As quoted in Stoddard, 1943, p. 79)
It would be a long time, however, before anyone else systematically considered the subject of indi- vidual differences in mental abilities. (Mental means simply that individual differences are not mainly due to differences in sensory or motor ca- pabilities per se.) The mind-body dualism pro- pounded by the early Greek philosophers, and the idea of mind as a spiritual essence or soul indepen- dent of physical or organic cause, was elevated and perpetuated by the Christian scholastics. Most prominent among them was the Catholic the- ologian Thomas Aquinas 0225-1274), who fol- lowed Aristotle in subdividing the functions of mind. The first division was between the intellec- tual and the appetitive functions. The intellectual function was further subdivided into sensation, perception, memory and reproductive imagination, and reasoning and creative imagination. This structure of the mind, with minor variations, per- sisted in philosophical writings down to the 19th century. But throughout this period, these catego-
CHAPTER 4 • INDIVIDUAL DIFFERENCES IN MENTAL ABILITY 65
ries of mind remained philosophic abstractions without being viewed in relation to human dif- ferences in their individual manifestations. That conceptual leap would have to await a major revo- lution in human thought, namely, a fully biological conception of the human species, and of human behavior, as fundamentally continuous with the rest of the animal kingdom, as a product of organic evolution rather than of special creation. Among early philosophers, John Locke (1632-
- has had a lasting influence on this field through his most famous work, the Essay Con- cerning Human Understanding (1690). Essen- tially, Locke brought mind closer to naturalistic explanation. He opposed the notion of innate ideas and viewed the human mind at birth as a blank tablet, or tabula rasa, which is gradually filled with impressions through the avenues of the spe- cial senses. All knowledge, Locke claimed, comes from only two sources, sensation and reflection, or "the association of ideas." He wrote,
Let us suppose the mind to be, as we say, white paper, void of all characters, without any ideas; How comes it to be furnished? Whence comes it by that vast store, which the busy and bound· less fancy of man has painted on it with an almost endless variety? Whence has it all the materials of reason and knowl- edge? To this I answer, in one word, from experience. In that all our knowledge is founded, and from that it ultimately de- rives itself. (Quoted by Boring, 1950, p. 172)
Thus the line was clearly drawn between nativism, or the k~ea that the mind comes equipped with certain built-in qualities, and empiricism, accord- ing to which the properties of the mind are wholly attributable to individual experience. Although there is nothing explicit in this empiricist philoso- phy concerning intelligence and individual dif- ferences, the implications of Locke's tabula rasa conception were that both intelligence and human differences therein must arise entirely from dif- ferences in people's experiences-an idea that has come down to the present day in the research and controversy concerning the relative effects of "nature" and "nurture" (or heredity and environ- ment) on mental abilities and other psychological characteristics. The British philosopher Herbert Spencer (1820-
- was the immediate precursor of the scien- tific era in the study of intelligence and individual differences. He was a Lamarkian evolutionist, who propounded his own pre-Darwinian ideas about evolution. After the publication of Darwin's Ori- gin of Species (1858), Spencer was converted from
Lamarkism to the theory of natural selection, and he became the leading philosopher of the Darwi- nian revolution. Because Spencer was never him- self an empirical scientist, we must assign him to the prescientific era as regards his contributions to psychology. However, his textbook, The Princi- ples of Psychology (1855), was the first psychol- ogy book to resurrect the term intelligence and to pay specific attention to individual differences. Spencer viewed human intelligence as a unitary trait that emerged through the differentiation of adaptive functions in the course of biological evo- lution. Later, with the publication of Darwin's the- ory of natural selection as the explanation of evolu- tion and the "survival of the fittest" as its principal mechanism of evolution, Spencer per- ceived the biological significance of individual dif- ferences as the essential raw material on which evolution depends. Spencer's extension of this line of thought to the human social conditions of his time has been termed "Social Darwinism," often in a pejorative context. However, Spencer's idea of intelligence as a biologically adaptive function for achieving the "adjustment of internal to exter- nal relations" is a progenitor of the detailed mod- em efforts to understand both animal and human intelligence in an evolutionary perspective, as seen, for example, in Harry Jerison's chapter, "The Evolution of Biological Intelligence" in the recent Handbook of Human Intelligence (Stern- berg, 1982). The concept of the phylogeny of intel- ligence, the idea that intelligence increases pro- gressively throughout the phylogenic scale of the animal kingdom, is also attributable to Spencer. His view of the ontogeny, or individual develop- ment, of intelligence in humans, from birth to ma- turity, is that it has three main aspects, (a) an in- crease in the accuracy of inner adjustments to outer demands, (b) an increase in the number of items of simple knowledge, and (c) an increase in the com- plexity of consciousness of the external environ- ment. The idea of accuracy of perceptions was likely a precursor of Francis Galton's (1822-1911) emphasis on sensory discrimination as a measure of intelligence, and the ideas of number and com- plexity were much later relabeled and empirically researched by Edward L. Thorndike (1874-1949) as breadth and altitude of intellect (Thorndike, Bregman, Cobb, & Woodyard, 1927). But it was actually Spencer, rather than Galton, who is so often credited (or blamed) for the concept of intel- ligence as a unitary or general ability. As Guilford (1954) has put it, "The conception of intelligence
CHAPTER 4 • INDIVIDUAL DIFFERENCES IN MENTAL ABILITY 67
and Galton, in England. Until recent years, these two lines have shown only occasional and casual interaction. The one subject on which the "two disciplines of scientific psychology" have finally become focused in a fruitful merger, only within the last decade, is the study of human intelligence. But the threads of this development really go back to Galton in the latter half of the 19th century.
Sir Francis Galton Galton was born the same year as Gregor Men- del (1822-1884), the father of modern genetics, and he died the same year as Alfred Binet (1857- 1911), the inventor of the first practical test of intelligence. Interestingly, Galton was the first in- vestigator of the genetics of intelligence and the first to attempt the objective measurement of abilities. Galton was born into a wealthy English family. A half-cousin of Charles Darwin (1809-1882), they both were grandsons of the philosopher, phys- iologist, and poet, Erasmus Darwin (1731-1802). Galton was a prodigy who could read and write by the age of three. After attending medical school and earning a degree in mathematics at Cambridge University at 21, he fell heir to a family fortune that allowed him freely to pursue his extremely wide and varied scientific interests for the rest of his long life, without need to earn a living. He used his fortune to travel, to finance his research, to found journals (Biometrika and Annals of Human Genetics, which are still in existence today), to endow a chair in genetics (occupied by such lumi- naries as Karl Pearson and Sir Ronald Fisher) and the famous Galton Laboratory at the University of London. He also founded the Eugenics Society, which still exists. Galton was one of the greatest scientific dilet- tantes of all time. Because he was also a genius, he made original contributions to a variety of fields: exploration and geography (of Africa), mete- orology, photography, fingerprint classification, genetics, statistics, anthropometry, and psychome- try. His prolific achievements and publications brought him worldwide recognition and many hon- ors, including knighthood, Fellow of the Royal Society, and several gold medals awarded by vari- ous scientific societies in England and Europe. 3
3The chief sources on the life of Galton are Galton's Memoirs (1908). Pearson's (1914-1930) three-volume biography, and a modern bIography. containIng also a complete bibliography of Galton's publications, by Forrest (1974).
What is Galton's legacy to the psychology of individual differences? Above all, he vigorously promoted the idea of objective measurement and quantitative analysis of data, whether by mere counting, or by ranking, or by true measurement. His favorite motto was, "When you can, count." He acted accordingly, some would say, to an al- most eccentric extreme. He applied this predilec- tion for quantification mainly to the study of human variation in just about every physical and mental characteristic that was within his power to count, rank, or measure. Unlike Wundt, the father of experimental psychology, who saw individual differences as a nuisance to be overcome in the search for general laws, Galton regarded human variation as of paramount importance and as per- haps the most interesting of all phenomena for sci- entific study in its own right. Hence the "two dis- ciplines of scientific psychology, " stemming respectively from Wundt and Galton. As a result of Galton's pursuit, he was led to invent a number of the statistical and psychometric concepts and methods familiar to all present-day researchers, including the bivariate scatter dia- gram, regression and correlation, multiple correla- tion, percentile ranks, standardized or scale-free scores, rating scales, the use of the normal, or Gaussian, distribution as a basis for the interval scaling of traits, and the use of the median and geometric mean as measures of central tendency. But the details of these contributions more prop- erly belong in the history of measurement and sta- tistics per se. Galton's main substantive contributions, which depended heavily on his quantitative inventions, are found essentially in two works: Hereditary Ge- nius: An Inquiry into Its Laws and Consequences (1869), his most famous and most influential work, and Inquiries into Human Faculty and Its Development (1883). The second work is of in- terest from our standpoint for its descriptions of the odd assortment of "tests" Galton invented for measuring human capacities. Successful or not, they were the very first objective "mental" tests. Like every scientific innovator, Galton was also a product of his time. This is reflected in his choice of "tests." The prevailing doctrine at the time was faculty psychology, which traces back to the an- cient Greek philosophers, who conceived of the mind as consisting of a number of distinct and separate powers or faculties, such as sensation, discrimination, perception, memory, and reason. And the chief techniques of experimental psychol- ogy at the time were the so-called brass instrument
68 PART II • THE DEVELOPMENT OF EDUCATIONAL PSYCHOLOGY
apparatuses of W undt' s laboratory, gadgets for measuring various types of sensory discrimination and speed of reactions. In keeping with the psy- chology of his time, Galton believed that because all the contents of intellect must come through the sense organs, the capacity for fineness of sensory discrimination was one of the two main aspects of mental ability; the other, because of its supposed adaptive evolutionary significance, was sheer speed of reaction to an external stimulus. In Human Faculty (1883), he argued,
The only information that reaches us concerning outward events appears to pass through the avenue of our senses; and the more perceptive the senses are of difference, the larger is the field upon which our judgment and intelligence can act. (p. 19)
Hence, Galton's battery of tests consisted mostly of devices for measuring auditory, visual, and kin- esthetic discrimination, short-term memory span, as well as simple reaction time to visual and au- ditory stimuli. These various tests, along with a number of physical measurements, were obtained during the brief period between about 1884 and 1890, on more than 9000 individuals, who paid threepence apiece to be run through all the tests in Galton's .. Anthropometric Laboratory" in the South Kensington Science Museum. Galton ex- pressed his notion of the aim of such tests as follows:
One of the most important objects of measurement ... is to obtain a general knowledge of the capacities of a man by sink- ing shafts, as it were, at a few critical points. In order to ascertain the best points for the purpose, the sets of measures should be compared with an independent estimate of the man's powers. We thus may learn which of the measures are the most instructive. (Quoted in Anastasi, 1965, p. 25)
Galton's idea was quite sound, and presages the modern psychometric concept of external validity. Unfortunately, however, Galton's particular collection of tests of sensory discrimination and reaction time did not prove to be very fruitful in his own day. Such simple tests could often distinguish the mentally deficient, but differences among per- sons of normal and superior intelligence, as judged by educational and occupational attainments, were generally so slight and seemingly unreliable as to afford scarcely any evidence for the claim that they measured intelligence. At least so it seemed at the time. Mere visual inspection of the data yields an unpromising picture. Reliability theory had not yet been conceived, and modern analyses of Galton's data reveal exceedingly low reliability of many of his tests. The reaction time tests, for example, were based on only a few trials and therefore
yielded measurements with an average reliability of only 0.18 in the total sample. Tests with such low reliability could hardly show impressive cor- relations with any criterion, and mean differences between different age groups and occupational cat- egories look unimpressive to casual inspection. Unfortunately, multiple regression analysis and statistical tests of significance had not yet been invented. When, in recent years, modern statistical analyses have been applied to Galton's old data, there were found to be highly significant mean differences by age group and by five occupational categories (ranging from professional to unskilled) on many of Galton's measurements. 4 Still, Ga- lton's simple tests, at least in their original primi- tive form, proved to be practically useless for indi- vidual assessment. The first practically useful test for mental ability was still waiting to be invented by Alfred Binet, some 15 years later. It was not until almost a century after Galton's failed attempt that psychologists have looked with renewed interest at Galton's ideas in search of more refined techniques for fathoming the nature of individual differences in mental abilities. One of the leading modern cognitive theorists, Earl Hunt, has stated, "We believe that Galton, not Binet, had the right approach. Measurement in science should be dictated by theory. What is needed is a better theory" (Hunt, Frost, & Lunneborg, 1973, p. 195). The statement is somewhat reminiscent of John Dalton's comment to the effect that the most important thing for a scientist is not necessarily to be right, but to have the right idea. And Galton had the right idea. But he lacked the necessary tech- nical and statistical apparatus to make it work. Galton's ideas about the nature of intelligence were not very formalized as a theory in the usual sense. Deeply impressed by Darwin's theory of evolution and the central role of individual varia- tion in natural selection and "fitness for survival," Galton thought of intelligence as having developed in the course of evolution as a general, heritable fitness trait in the Darwinian sense, attaining its highest development in Homo sapiens, while still evincing variation between individuals and be- tween various subspecies, or races. (One chapter
4Nearly all of Galton's original data had been secured by Pro- fessor Gerald McClearn, while at the University of Colorado's Institute of Behavior Genetics. Various specialists in genetics and psychometrics are in the process of analyzing the data with modem statistical techniques. The information reported here was provided by one of those who are reexamining Galton's data, Professor Ronald Johnson of the University of Hawaii.
70 PART^ II^ •^ THE DEVELOPMENT OF EDUCATIONAL PSYCHOLOGY
the Galtonian "brass instrument" variety and, in particular, a study published in 1901 by one of his Ph.D. students, Clark Wissler (1870-1947), led to the early demise of Galtonian methods of mental testing in America. Wissler, working in Cattell's lab, administered to between 90 and 252 Columbia College under- graduates a battery of Galtonian tests measuring various simple sensory and motor capacities, dis- crimination, short-term memory, color-naming speed, and simple visual and auditory reaction time, as well as several physical measurements. These simple measures were correlated with class standing and grades in classics, foreign language, and mathematics courses, which were assumed to reflect individual differences in general mental ability, or intelligence. Pearson ian correlations were calculated between each of the "mental tests" and the academic "riteria. It was the very first use in psychology of lhe product-moment co- efficient of correlation, invented in 1896 by Karl Pearson (1857-1936), protege of Galton. Few of Wissler's correlations significantly exceeded zero. Unfortunately, Wissler's results, interpretation, and conclusions largely reflected psychometric and statistical naivete. With the clarity of hindsight, modem students can easily see that the deck had been strongly stacked against finding significant or substantial correlations. Each test score was based on an average of only three to five measurements, which we now know would result in exceedingly low reliability; the "range of talent" was very re- stricted in this highly selected group of Ivy League students, a fact that greatly attenuates correlations; and the reliability and validity of course grades as a measure of intelligence leave much to be desired. (The best present-day IQ tests generally show cor- relations of less than .50 with grades in selective colleges.) Wissler's and Cattell's disappointing re- sults, coming from the most prestigious psycho- logical laboratory in America, cast a pall over the whole Galtonian approach to studying individual differences in abilities. Galton's methods might have survived this blow and been developed fur- ther, however, had it not been for a momentous development in France, just 4 years later.
Alfred Binet Binet (1857-1911) was France's greatest psy- chologist, an investigator of remarkably broad in- terests, insight, and ingenuity. 5 Trained in experi-
SThe best account of Binet', life and work is the biography of Binet by Theta H. Wolf (1973).
mental and physiological psychology, as well as in medicine, Binet was the first major figure in our field of interest who could be called a clinical psy- chologist, who thought and acted like a clinician in the best sense of that term. All his predecessors perceived themselves either as philosophers or as natural scientists. Binet was not a strong theorist, and he developed no formal theory of intelligence; but his numerous writings afford a fairly clear im- pression of his conception of intelligence, and his methods of developing the first practically useful intelligence test have provided many followers, as well as critics, grist for theoretical inference about the nature of intelligence as conceived by Binet. Binet was already eminent when he was drawn to the study of intelligence. The story is well known, how he and his co-worker, Theodore Si- mon (1873-1961), a psychiatrist, were commis- sioned in 1904 by the Minister of Education, to devise a practical, objective means for assessing mental subnormality in primary school children. Contrary to some of the later lore that has grown up about Binet, largely through the interpretations of American followers who wished to sharpen the contrast between Binet and the Galtonian school in Britain, Binet, in fact, greatly admired and was profoundly influenced by the British evolutionists Darwin, Spencer, and above all, Galton. The idea that Galton and Binet were at opposite poles is false, although their disciples have often been at odds. Binet accepted Galton's idea of intelligence as a general ability that enters into "nearly all the phenomena with which the experimental psychol- ogist has previously concerned himself-sensa- tion, perception, memory, as well as reasoning," and Binet also distinguished special abilities, which he termed "partial aptitudes" (Binet & Si- mon, 1905a). Binet was also a hereditarian regard- ing the basis of individual differences and claimed that his intelligence scale was expressly devised to reflect innate differences, in contrast to "ped- agogical scales" that measure specifically educa- tional attainments (Binet & Simon, 1905b). It was when Binet actually set about devising a test of intelligence that he became truly innovative, taking a quite different approach from the one sug- gested by Galton. Binet was well informed of the unimpressive results obtained using the Wundtian and Galtonian "brass instrument" techniques of measuring simple processes as a means for assess- ing intelligence. In looking around for more promising measures, Binet was impressed by a new sentence completion test devised by the German psychologist Hermann
CHAPTER 4 • INDIVIDUAL DIFFERENCES IN MENTAL ABILITY^71
Ebbinghaus (1850-1909), who is best remem- bered for his experimental studies of verbal learn- ing and memory. The completion test consisted of sentences with missing words that the subject had to fill in with words selected so as to make good sense of the incomplete sentence. This was proba- bly the first successful test of higher mental abili- ties; it quite clearly discriminated between primary school pupils when they were classified by their teachers as being good, average, or poor in scho- lastic standing. (A sentence completion test is still in use today, for example, as part of the well known Lorge-Thorndike Intelligence Test; and it generally shows a higher correlation with the total IQ than any other type of subtest.) Ebbinghaus emphasized the importance of complexity of a task's cognitive demands as being essential for the assessment of the higher mental functions thought of as intelligence. Complexity thus became a key idea in Binet's effort. He abandoned Galton's and Cattell's simple sensorimotor tests (except Gal- ton's test for discriminating weights) and devised instead a large number of single-item "tests" based, not on laboratory apparatus, but on brief tasks children could perform with such com- monplace. things as pencil, paper, coins, blocks, pictures of familiar objects, and the like. Each task posed a problem involving attention, adaptability, memory, judgment, reasoning, or some common item of information. Binet's most original contribution was the con- cept of mental age as a device for selecting and scaling items so as to permit a meaningful in- terpretation of the child's performance. As it was obvious to Binet that children's mental capability increases with age, he used age as a criterion for selecting and grading his test items. By calibrating items in terms of the percentage of each normative group of children sampled at one-year age intervals from age 3 to 15 years who passed the item, it was possible to express a child's raw score (i.e., number right) on the whole battery of items in terms of mental age. A 6-year-old who got as many items right as the average 8-year-old, for example, would be said to have a mental age of 8 years. It was the German psychologist, William Stem (1871-1938), who suggested dividing the child's mental age (MA) by his chronological age (CA) in order to express his relative standing, in comparison with other children, in rate of mental development. The ratio of MA/CA (x 100, to re- move the decimal), was termed the "mental quo- tient" by Stem, and was later translated by Lewis M. Terman (1877-1956) as "intelligence quo-
tient," or IQ. The Binet-Simon intelligence scale, consisting of a graded series of heterogeneous items, was the prototype of virtually all subsequent tests of intelligence down to the present time. Binet never attempted to develop a consistent or unified theory, or even a formal definition, of in- telligence, but from his voluminous writings one can discern Binet's implicit conception of intel- ligence. This effort, however, may be a bit like describing a Rorschach inkblot, with different writers emphasizing different aspects of Binet's rather unsystematic views. Those aspects of Bin- et's ideas about intelligence that show the least similarity to the Galtonian and British lines of thought have been the most emphasized by Binet's followers in America. Although at times Binet writes of intelligence as a general ability, at other times he emphasizes its heterogeneity, which seemingly (but mistakenly) justifies the hetero- geneous item content of his test. General intel- ligence, in Binet's thinking, is not a single func- tion, but the resultant of the combined effects of many more limited functions, such as attention, discrimination, and retention. In his later writings, he put greater emphasis on the more complex men- tal functions-logical processes, comprehension, jUdgment, and reasoning-as the sine qua non of intelligence. He argued that intelligence could be measured efficiently only by using a great variety of items that "sample" these higher processes. As Tuddenham (1962) has aptly put it: "Regarding intelligence as a product of many abilities, Binet sought in his tests to measure not an entity or sin- gle dimension- 'general intelligence' -but rather an average level- 'intelligence in general''' (p. 489). Tuddenham's characterization of Binet's view probably represents the prevailing conception of intelligence among the majority of American psy- chologists and especially among clinical psychol- ogists. But there are also serious theoretical and psychometric problems with this B inetian view, as first pointed out by the first really important the- oretical successor to Galton, Charles Edward Spearman (1863-1945). The question of whether intelligence is a unitary process or is a resultant of the complex interaction of a great many different, more specialized processes is one of the chief is- sues of contention by contemporary theorists. But before bringing in Spearman, who begins a whole new line of investigation, this would seem the right place to mention Binet's main intellectual heirs in America. There is not much that needs to be said about them in the present context, however, be-
CHAPTER 4 • INDIVIDUAL DIFFERENCES IN MENTAL ABILITY 73
la), and derived the formula for the nonparametric rank-order correlation coefficient. But his greatest methodological contribution was the invention of factor analysis, a methodology that has developed and dominated the study of human abilities ever since it was first introduced by Spearman in 1904. Spearman came to psychology relatively late in life. After a career as a British Army officer, from which he retired, at age 34, with the rank of major, he began a new career by earning a Ph.D. degree in psychology at the University of Leipzig, under Wundt. He then joined the psychology faculty at the University of London, and soon thereafter he was appointed successor to William McDougall as professor and head of the psychology department, a chair he held for 25 years. In terms of the impor- tance of the topics he researched, his great origi- nality, and his enduring influence, Spearman was unquestionably Britain's greatest psychologist. Besides his intellectual brilliance and mathe- matical talent, the traits that characterized his ca- reer were his clear, no-nonsense, scientific style of thinking about psychological problems and his un- alloyed impatience with armchair philosophizing and speculation. This hard-nosed attitude led Spearman into conflict with much of the psycho- logical thought of his day. In his autobiography, Spearman (1930a, p. 330) described his career as "one long fight." For the present purpose, unfor- tunately, it is impossible to do more than summa- rize Spearman's contributions rather too briefly and hence inevitably with considerable simplifica- tion. Spearman's major works, however, are still worth reading, as many of the issues he raised are still very much alive in contemporary research on intelligence (Spearman, 1904, 1923, 1927, 1930b; Spearman & Jones, 1950). Spearman's most fa- mous book, in which he most completely explica- tes his main contributions, is The Abilities of Man (1927). It still ranks near the top of the list of "must" reading for students of individual dif- ferences. Virtually all the basic questions that con- tinue to occupy contemporary researchers and theorists of human ability were first clearly posed by Spearman. When Spearman began his career in psychol- ogy, the doctrine of formal faculties was the gener- ally accepted view of individual differences in abilities. Persons differ in the powers of the many distinct "faculties" that constitute the mind, such as perception, discrimination, memory, recollec- tion, attention, reason, common sense, language, imagination, invention, comprehension, motor control, kinesthetic sense, visualization, and so
on. One theorist even listed as many as 48 distinct mental faculties, including "sense of the ridicu- lous. " Spearman questioned whether the numerous list- ed faculties were truly distinct components of the mind. Are "memory" and "recollection" really different abilities, or "imagination" and "in- vention," or "reason" and "comprehension"? If so, mental ability could be objectively measured only by devising special tests for each of the many faculties. But there were endless armchair debates among psychologists concerning the number and names of the faculties. Spearman saw an objective solution to this problem by the use of correlation. If two (or more) nominal faculties were claimed to be distinct, it should be possible to devise tests of each one, to administer the tests to a group of persons who show individual differences in the power of the faculties in question, and show that the measurements of the different facuIties are uncorrelated. Spearman performed this type of study with school children, using tests, examination marks, and teacher ratings on a variety of variables, in- cluding classics, French, history, geography, mathematics, "common sense," musical talent, and measures of auditory, visual, and kinesthetic (weight) discrimination. The matrix of correlations among all of these tests revealed all positive inter- correlations, suggesting to Spearman that all of the measures reflect a common factor, that is, a com- mon or unitary source of the covariance among the variables. Individuals who scored exceptionally high on anyone variable tended to score above average on all the others as well. Moreover, the correlation matrix displayed a quite regular varia- tion among the sizes of the correlation coefficients, such that by arranging the variables in the matrix in the order of their average correlation with every other variable, the correlations displayed what Spearman referred to as a hierarchy, that is, the correlations in the matrix decreased regularly in both the horizontal and vertical directions from the diagonal, going from the upper left to the lower right comer of the matrix. It especially impressed Spearman that in this hierarchical pattern of cor- relations there was no clear discontinuity between the scholastic measures (classics, etc.) and the measures of musical ability and of sensory dis- crimination. This observation seemed to confirm Galton's notion that discrimination ability is a basic aspect of general intelligence. Spearman showed mathematically that such a hierarchical correlation matrix could be "explained" in terms
74 PART II • THE DEVELOPMENT OF EDUCATIONAL PSYCHOLOGY
of a single factor (Le., source of variance) that every test in the matrix has in common. He later assigned the label g to this general factor, which he identified with general intelligence. Spearman hypothesized that every type of cognitive test mea- sured g in addition to one other source of variance (besides error), labeled s (for specific). The s is entirely specific to a particular test (or a very nar- row class of highly similar tests). This hypothesis became known as Spearman's "two-factor theo- ry" of ability, according to which the total true- score variance (a;} on any test is expressed as the sum of two components, g variance (a~) and s
variance (a;>, hence a~ = a~ + a;.
Spearman invented a method, now known as factor analysis, but actually a rather simple fore- runner of the modem techniques under this name, that made it possible to determine precisely the proportion of g variance in each of the variables that are entered into a correlation matrix. The square root of this proportion can be interpreted as the test's correlation with the hypothetical ability represented by g; this correlation between a test and a factor is commonly termed the loading of the test on a given factor (in this case g). Much of Spearman's subsequent research con- sisted of determining the g factor loadings of nu- merous diverse tests. As many as 94 various tests were factor analyzed in one study (Spearman & Jones, 1950, Chap. 8). Various tests differed widely in their g loadings, even when the loadings were corrected for attenuation, ranging from slightly greater than zero up to .80 and above. Spearman regarded the differences in tests' g load- ings as a basis for discovering the essential nature of g. He attempted to do this by comparing high and low g-loaded tests for their similarities and differences. The types of tests with the highest g loadings, he found, were those that require induc- tive or deductive reasoning and have a quality of abstractness. In general, the g loadings of tests were found to increase, going from tests of simple sensorimotor abilities, to tests of rote and asso- ciative memory, to tests involving the grasping of conceptual or abstract relationships, as typically found in verbal and figural analogies tests. Hence, Spearman characterized g, or general intelligence, as the "eduction of relations and correlates," that is to say, inductive and deductive reasoning. But this is merely a description of the types of tests that best measure g.. these are tests requiring fairly complex mental manipulations in order to arrive at the correct answer. But this empirical observation
can hardly be called a theory of g. It does not tell us what g is, independently of the very mathe- matical operations of factor analysis, by means of which we have determined the "existence" of g and the extent of its loading in various tests. Nor does the description of g in terms that characterize the most highly g-loaded tests tell us why even tests that involve no reasoning or conceptual con- tent, such as pitch discrimination and choice reac- tion time, also have some moderate g loading. Spearman fully admitted that factor analysis does not, and logically cannot, permit a declaration of the nature of g, but can only point to those tests that measure it best. This "defining of g by site rather than by nature," he wrote, is a "way of indicating what g means... just as definite as when one indicates a card by staking on the back of it without looking at its face" (1927, p. 76). Spearman (1927, Chap. 7) considered many dif- ferent speculative hypotheses of the nature of g. He settled on the hypothesis of a unitary mental energy. This "energy" was deployed to whatever specific "engines" or brain processes were in- volved in different mental tasks, some tasks requir- ing more energy, and some less, and hence their different g loadings. In Spearman's view, this uni- tary source of energy enters into every kind of mental task, and the observed positive correlation between all tests is a result of individual dif- ferences in the amount of mental energy that peo- ple brought to bear on the tests. The specificity peculiar to different tests was attributed to lo- calized or specific energies. "Successful action would always depend partly on the potential ener- gy developed in the whole cortex and partly on the efficiency of the specific group of neurons in- volved" (1923, p. 6). The main problem with Spearman's theory of g as "mental energy" is not that it is necessarily wrong, but that no means have been found to test it empirically. Theories are scientifically useful only when opposing theories can be pitted against one another in an empirical test. Thus, without an em- pirical means of being tested, Spearman's theory of g remains only speculative and problematic to this day. The g factor itself, however, remains se- cure as an established empirical phenomenon, summarizing the observation that virtually all men- tal tests that are scorable according to an objective standard of performance are positively intercorre- lated in an unrestricted sample of the general population. The application of Spearman's method of factor
76 PART II • THE DEVELOPMENT OF EDUCATIONAL PSYCHOLOGY
Test Z (Factor :n:)
3 ~B
2 C
B
Test X &. y (Factor 1)
C
3
Figure 2. Representation of the rank order of three persons on three tests in a hypothetical two-dimensional (i.e., two factors) test correlation matrix.
A two-factor (2-dimensional) case:
Person A B C
X 2 3 Test Y 2 3 Z 3 2
Test X Y Z
X 1.0 0. Y 1.0 0. Z 0.5 0.
A 2-dimension space is needed to represent these data (see Figure 2).
A three-factor (3-dimensional) case:
A
X Test Y 1 Z 2
Person B
2 3 3
c
3 2 1
Test
X Y Z
X Y Z
And a 3-dimension space is needed to represent these data (see Figure 3). One can go on adding dimensions, although it becomes impossible to depict more than three di- mensions graphically, and the geometry of n-di- mensional space can be treated only in purely mathematical terms. The scientifically desirable economy of factor analysis as a means of describ- ing the "structure" of a correlation matrix results from the fact that most of the covariance among a large number of tests can be accounted for in terms of a relatively much smaller number of factors, because many different tests share some of the same factors in varying degrees. It is important to recognize just what factor anal- ysis does and does not tell us. It tells us which tests "go together," that is, it parsimoniously describes the correlations among a number of diverse tests in terms of a limited number of uncorrelated common sources of individual differences variance, called factors, that are shared by all mental tests (in the case of g) or by particular groups of tests (in the case of group factors). Thus factor analysis is es-
CHAPTER 4 • INDIVIDUAL DIFFERENCES IN MENTAL ABILITY 77
3 B
Test Y
(Factor II) 2 C
A B 2
Test X
C 3
f>.
2 l fest -<$>-") 'I.<l~ l«~C.
B
(Factor I)
Figure 3. Representation of the rank-order of three persons m a hypothetical three-dimensional (i.e .• three factors) test correlation matrix.
sentially descriptive. It is said to describe the struc- ture of abilities. It is not an explanatory theory. It does not explain why various tests are correlated as they are, or why various tests show quite different average correlations with all the other diverse tests in a battery. Factors merely afford a systematic description of phenomena with unknown causes. Factors themselves are not the causes of anything; they are simply descriptive abstractions. The basic empirical phenomena from which factors are de- rived are individual differences in test scores and their intercorrelations among diverse tests. It is these phenomena, and consequently the factors to which they give rise, that are in need of scientific explanation in causal terms. If we accept g, the largest common factor, as a working definition of intelligence, then a major aim of a theory of intelligence is the explanation of g. This boils down to an explanation of why differ- ent tests are correlated with one another and why some tests are correlated more highly than others. As already noted, Spearman put forth a unitary or monistic explanation of g in terms of a hypo- thetical "mental energy." He hoped that future neurophysiological research would discover indi- vidual differences in some form of general neural energy in the cerebral cortex. Spearman's monistic
theory of g as mental energy was soon challenged by rival theories.
Edward Lee Thorndike. The leading Amer-
ican educational psychologist, Thorndike (1874-
- was best known for his studies of learning. But he also played a major role in the development of intelligence tests and was the first American to espouse a theory of intelligence, the elements of which were borrowed directly from his theory of learning as the formation of new stimulus-re- sponse (S-R) bonds under the influence of reward, or positive reinforcement. For Thorndike, learning was a process of "selecting and connecting"; hence his term connectionist theory. An indi- vidual's behavioral and intellectual repertoire was made up, basically, of innumerable S-R connec- tions in the nervous system, the specific connec- tions being acquired through experience in the en- vironment. Thorndike's theory of intelligence was set forth in his major contribution to this field, The Measurement of Intelligence (Thorndike et ai., 1927), which is also one of the major classics of this field that is still rewarding to read. According to Thorndike, individual differences in intelligence reflect the number of S-R bonds that persons ac- quire by a given age. He hypothesized that persons differ innately in the number of potential neural
CHAPTER 4 • INDIVIDUAL DIFFERENCES IN MENTAL ABILITY^79
they require more mental energy. From the view- point of sampling theory, however, the factors re- vealed by factor analysis really describe the char- acteristics of tests rather than factors of the mind. Although "sampling theory," as it later came to be known, originated with E. L. Thorndike, it was formalized mathematically by the British psycho- metrician and educational psychologist, Sir God- frey H. Thomson (1881-1955), who had spent a year (1923-24) at Columbia University working with Thorndike. Thomson's (1951) "sampling theory" of g was seen as a challenge to Spear- man's "mental energy" theory. It has gained con- siderable popUlarity among psychometricians, es- pecially in the United States. Although the "sampling theory" has been around since at least 1914, when first introduced by Thorndike, it has never given rise to any empirical research that could put it to a significant test. Its appeal is en- tirely intuitive. The typical criticism of Thorn- dike's and Thomson's sampling theory has been cogently expressed by Jane Loevinger (1951):
The sampling theory hardly qualifies as a true theory, for it does not make any assertion to which evidence is relevant. Perhaps the large number of adherents to this view is due to the fact that no one has offerei evidence against it. But until the view is defined more sharply, one cannot even conceive of the pos· sibility of contrary evidence, nor, for that matter, confirmatory evidence. A statement about the human mind which can be neither supported nor refuted by any facts, known or conceiv· able, is certainly useless. Bridgman and other philosophers of science would probably declare the sampling theory to be meaningless. (p. 595)
Louis L. Thurstone. The leading American psychometrician and factor analyst, Thurstone (1887-1955) developed a method of "multiple factor analysis" (Thurstone, 1947) that facilitated the extraction of a number of factors from a cor- relation matrix of numerous diverse tests, and along with it he proposed an objective criterion for the "rotation" of the factor axes that he called simple structure, intended to yield psychologically interpretable factors. Rotation of the factor axes to the simple structure criterion maximized the load- ings of certain tests on particular factors and mini- mized the tests' loadings on other factors, making it relatively easy to describe the various uncorre- lated factors in terms of the particular tests on which they had the largest loadings. Ideally, each factor would load only on certain tests and each test would be loaded on only one factor, in which case it could be called a "factor pure" test. Applying his method of multiple factor analysis to large batteries of tests, Thurstone (1938) ex-
tracted a number of factors that he termed primary mental abilities: verbal fluency, verbal com- prehension, numerical, spatial, reasoning, percep- tual speed, and associative memory. There was no g factor in this structural model of abilities, for the simple reason that the criterion of simple structure mathematically precludes the extraction of a gener- al factor. This limitation of Thurstone's method became a point of considerable contention between British and American psychometricians. The ap- propriateness of the simple structure criterion in the domain of human abilities was soon chal- lenged. It was noted that a good simple structure could not be achieved with orthogonal (uncorre- lated) factor rotation; allowing oblique rotation of the factor axes, so that the axes were at less than right angles and were thus oblique, or correlated, factors, permitted a much closer approximation to the ideal simple structure. Thurstone himself re- solved the conflict with Spearman. By factor ana- lyzing the intercorrelated primary factors, Thur- stone showed that the g factor emerged as a sec- ond-order factor, or superfactor. Thurstone's method of multiple factor analysis with orthogonal rotation to simple structure had merely scattered the large g factor among the so-called primary fac- tors. When Eysenck (1939) reanalyzed Thur- stone's correlation matrix of more than 50 diverse tests, using a method of factor analysis that allows the appearance of a general factor and various group factors, he found that the g factor accounted for more of the total variance in all the tests than the variance accounted for by all of the remaining group factors combined. In fact, it has proved im- possible to construct factor-pure tests of Thurstone's primary mental abilities that do not also measure Spearman's g, and usually each test is more highly loaded on g than on the primary factor it was specially devised to measure. At best, so-called factor-pure tests measure g plus the one primary factor they were devised to measure.
Contemporary Theorists The two leading contemporary factor analysts of the abilities domain are Joy Paul Guilford (b.
- and Raymond Bernard Cattell (b. 1905). Guilford (1959, 1966, 1967, 1977) has pro- posed a complex scheme, or "facet" model, for the classification of abilities that he has called the Structure of Intellect (SOl) model. The hypo- thetical abilities of the SOl model represent the intersections of 5 different mental operations (cog-
80 PART II • THE DEVELOPMENT OF EDUCATIONAL PSYCHOLOGY
nition, memory, divergent production, convergent production, and evaluation) x 5 different types of contents (visual, auditory, symbolic, semantic, and behavioral) x 6 different types of products (units, classes, relations, systems, transforma- tions, and implications), making for 5 x 5 x 6 = 150 abilities in all. Guilford regards each of the SOl abilities as unique, or factorially distinct from all the others. The SOl model thus suggests a pos- sible 150 types of tests, and from year-to-year new tests are reported as having been devised to mea- sure stilI a few more of the abilities suggested by this model. The number of such tests must now exceed 100. If all these tests were subjected to a type of factor analysis that does not mathe- matically prohibit the extraction of a general fac- tor, it seems virtually certain that a large g would emerge. Yet the SOl does not admit a g factor. A model with 150 hypothesized unique abilities, however, is actually beyond the reach of factor analysis for all practical purposes, and so the 150 abilities have not come anywhere near being sub- stantiated by factor analysis. The testability of the SOl model poses such staggering problems that it seems unlikely that it will ever be able to face the challenge of empirical verification (Undheim & Hom, 1977). Scientifically, the SOl model has not really advanced beyond a purely formal system (one of many possible rational systems) for the generation and classification of mental tests. Al- though Guilford's SOl is apparently a quite com- prehensive and fine-grained system of categories into which an extremely great variety of tests may be classified, it is highly arguable whether it actu- ally tells us anything about the nature of intel- ligence. It completely evades the central question: Why are all tests correlated with one another, thereby giving rise to g? Cattell (1963, 1971) has distinguished two as- pects of g, which he has termed fluid (gf) and crystallized (gc). Tests based on specific knowl- edge and cognitive strategies acquired prior to tak- ing the test, such as general information, vocabu- lary, arithmetic, scholastic knowledge and skills, and the like, are most heavily loaded on the gc factor. Tests with little or no knowledge content but that depend on short-term memory for novel material presented in the test situation (e.g., digit span memory) and novel problem solving involv- ing reasoning about figural materials (e.g., figure analogies, matrices, series completion) are the most heavily loaded on the gf factor. People reach their peak power on gr in their late teens or early
twenties, whereas g,. gradually increases until old age, provided persons are not entirely cut off from experiences that afford opportunities for new leaming. The gc factor can be interpreted as re- flecting the knowledge and skills acquired through the individual's investment of gfin specific forms of learning and experience. Consequently, indi- vidual differences in gf and gc will be more or less highly correlated depending on the degree of sim- ilarity in people's educational experience and in the cultural values that influence the types of expe- rience in which gf will be invested. The correlation between gf and gc again yields the superfactor g. Recent studies (Gustafsson, 1984; Undheim,
- based on a hierarchical type of factor analy- sis of collections of tests well representative of fluid and crystallized abilities suggest that gr is "absorbed" into the g (a "neo-Spearmanian" g) at the top of the factor hierarchy; that is, when g is partialled out of gp the residualized gf is reduced to zero, and hence it is concluded that gf is the same factor as Spearman's g (or vice versa). The gc factor remains as one of two or three second-order factors in the hierarchy. In contrast to the factor analytic school, a quite different approach, clinical and qualitative, to the study of intelligence was taken by the noted Swiss child psychologist Jean Piaget (1896-1980). In his major work on this subject, Piaget (1950) viewed intelligence as a biological process of adjustment between the conscious organism and its physical and social environment. The term intelligence in- dicates the forms of organization or equilibrium by which the organism cognitively structures its sen- sory and motor experiences. The complexity of the cognitive structures increases and changes qualitatively through different stages of the child's mental growth. Piaget's descriptions of the stages of mental growth developed from his observations of children when confronted by various problems cleverly devised by Piaget to reveal the "logic" of children's thinking at different stages of their men- tal development. Briefly, Piaget viewed the mental development of the child as going through four main stages, which are invariant in sequence for all children: (a) the sensorimotor stage (onset from birth to about 1 year) is the first phase of intellec- tual development, in which knowledge and thought are intimately tied to the content of specif- ic sensory input or motoric activity of the child; it includes conditioning, stimulus-response learn- ing, reward learning, perceptual recognition, and associative or rote learning and memory. (b) The
